DOGUŞ UNIVERSITY PUBLICATIONS Publication Number, 3 Proceedings of the International Conference on Prospects for Research in Transport and Logistics on a Regional - Global Perspective February 2009, Istanbul Dogus University Publications, 3 1st ed., February 2009 ISBN 978-9944-5789-2-9 © Dogus University, 2009 Cover and Page Design by Özay ÖZAYDIN CIP - Doğuş University Library Prospects for Research in Transport and Logistics on a Regional : Global Perspective (I : February 2009 : İstanbul : Turkey) First international conference on prospects for research in transport and logistics on a regional : global perspective February 2009, İstanbul / Edited by Özay Özaydın.— 1st. ed.— İstanbul : Doğuş University, 2009. 380 p. : ill. ; 27 cm.— (Doğuş University publications ; 3) ISBN 978-9944-5789-2-9 1. Logistics -- Congresses. 2. Transportation -- Congresses. I. Özaydın, Özay. II. Title. 658.5 - dc22 Distribution Dogus University 34722 - Acibadem, Kadikoy, Istanbul, TURKEY Tel. : +90 216 544 55 55 Fax. : +90 216 544 55 35 e-mail : info@dogus.edu.tr URL : http://www.dogus.edu.tr Print Akar Matbaacılık INTERNATIONAL CONFERENCE on PROSPECTS for RESEARCH in TRANSPORT and LOGISTICS on a GLOBAL - REGIONAL PERSPECTIVE International Steering Committee Füsun ÜLENGIN - CHAIR (Doğuş University) Yücel CANDEMİR (Istanbul Technical University) Güldem CERİT (Dokuz Eylül University) Mehmet Şakir ERSOY (Galatasaray University) Haluk GERÇEK (Istanbul Technical University) Yoshitsugu HAYASHI (University of Nagoya) Ali HUZAYYIN (Cairo University) Kaan ÖZBAY (RUTGERS, The State University of New Jersey) Werner ROTHENGATTER (Universität Karlsruhe) Mehmet TANYAŞ (Okan University) Eddy VAN DE VOORDE (University of Antwerp) International Advisory Board Kenneth BUTTON (George Mason Univ., USA) Ralph GAKENHEIMER (MIT, USA) Adib KANAFANI (UC Berkeley, USA) H. Baha KARABUDAK (ETU, Turkey) Andreas KOPP (The World Bank, USA) Samer MADANAT (UC Berkeley, USA) Çetin NUHOĞLU (TOBB, Turkey) Mark WARDMAN (Univ. of Leeds, UK) Local Steering Committee Füsun ÜLENGIN - CHAIR (Doğuş University) Ekrem DUMAN - Shuttle Services and Transportation Organization (Doğuş University) Yasemin Claire ERENSAL - Volunteer Relations (Doğuş University) Cengiz GÜNGÖR - Technical Tour Organization (Istanbul Technical University) Yeliz EKİNCİ - Participant Relations, Sponsor Relations (Doğuş University) Kıvanç ONAN - Technical Tour Organization, Printed Material Design (Doğuş University) Özay ÖZAYDIN - General Design, Webmaster, Editor (Doğuş University) A. Talha DİNİBÜTÜN (Doğuş University) Mesut KUMRU (Doğuş University) Ahmet Nuri CERANOĞLU (Doğuş University) Mustafa KARAŞAHİN (Süleyman Demirel Ataç SOYSAL (Doğuş University) University) Berrin AĞARAN (Doğuş University) Müjde EROL GENEVOIS (Galatasaray Burcu KULELİ PAK (Doğuş University) University) Dalga Derya TEOMAN (Doğuş University) Özgür KABAK (Istanbul Technical University) Emel AKTAŞ (Istanbul Technical University) Şule ÖNSEL (Doğuş University) Gönül YENERSOY (Doğuş University) Temel ÖNCAN (Galatasaray University) Ilker TOPCU (Istanbul Technical University) Yıldız Esra ALBAYRAK (Galatasaray Mehmet TANYAŞ (Istanbul Technical University) University) WORDS of WELCOME and GRATITUDE Stimulated by the globalization and the new economic trends, the value of global trade has been increasing twice as fast as the global production. This considerable increase in the volume and value of global trade and production must be carried and delivered by physical means. This indicates that transportation and logistics are the most promising industries for the world economy. Companies today have been offered a world of opportunities in terms of production and sales thanks to the globalization but they still need logistics in order to deliver the right product or service to the right place under proper conditions. The rise in worldwide trade and the increasing interaction between countries previously separated by trade barriers have spurred a significant increase in logistics flows at all geographical levels. Whereas virtual trade is possible virtual logistics and transportation is not. In a study titled “Logistics Benchmarking in EU” it is underlined that the critical factors that play important role in the evolution of logistics in a country are transportation infrastructure, knowledge and communication infrastructure, transportation and logistics education potential, regulations, and effectiveness in national and international coordination. We tried to reflect the effects of such changes in the presentations of RTraL 2009. That is why, we aimed to bring together researchers and practitioners in global, regional as well as country-based perspective and stimulate the exchange of ideas in the fields of logistics and transportation as defined. The attention is given to the recent theoretical and practical developments in all related fields and the respective concepts, models and methodologies. The conference consists of research papers and application case sessions about regional and global issues in logistics and transportation, public policies, education and training, transport modeling, traffic engineering, supply chain management, transportation infrastructure and investment appraisal, transportation planning and economics, and sustainable transport policies. Particular emphasis will be given to recent developments in all related fields, their methodologies, concepts and implementation details. In this Conference, the call for papers attracted 106 papers. The papers were reviewed by three referees using double-blind process. Among the full papers of the registered authors, 8 are rejected, 48 are accepted without revision while 50 are conditionally accepted and then revised according to the invaluable suggestions of referees. If this Conference proves to be a successful one, the contribution of the referees will be one of its most valuable assets. Therefore, RTraL 2009 Conference Organizers owe much gratitude to the members of International Program Committee for their valuable reviewing work. Special thanks are due to the keynote speaker Werner Rothengatter - University of Karlsruhe and to the panelists Eddy Van de Voorde - University of Antwerp, Ali Huzayyin - University of Cairo, Yücel C andem ir- Istanbul Technical University, Werner Rothengatter - University of Karlsruhe. We also want to draw your attention to our sponsors list. We would like to thank the sponsors of the conference for their generosity and their investment in current and future transportation and logistics researchers. Füsun ÜLENGİN RTraL '09 Chair, Dogus University, Istanbul, Turkey INTRODUCTION “International Conference on Prospects for Research in Transport and Logistics on a Global - Regional Perspective” has undertaken the challenge to host very important experts and practitioners of Transport and Logistics from a large spectrum of countries. In our opinion, the conference has fulfilled the purpose of establishing an International Society; “Eurasian and Eastern Mediterranean Institute of Transportation and Logistics Association (EMIT)” that is expected to have a very promising role in the Eurasian and Eastern Mediterranean countries. The purpose of the Association is to contribute to establishing and developing the exchange of research work between all parts of the world in all fields of transportation and logistics. This proceedings book consists of 13 chapters, grouping the contributed papers into the following categories: Global Issues in Logistics and Transportation (3 papers), Regional Issues in Logistics and Transportation (2 papers), Education and Training in Logistics and Transportation (2 papers), Supply Chain Management (3 papers), Sustainable Transport Policies, Traffic Engineering (4 papers), Evaluation of Public Policies, Network Models and Environment (4 papers), Contemporary Topics in Transport and Logistics (7 papers), Transport Planning and Economics (3 papers), Planning, Operations, Management and Control of Transport and Logistics (3 papers), Transport Modeling (5 papers), Freight Transportation and Logistics Management (7 papers), Transport and Land Use (3 papers), Transport Infrastructure and Investment Appraisal (2 papers) It can be readily seen from this volume of selected papers that all papers do elaborate on rather timely problems in the fields of expertise related to Transport and Logistics, which have a considerable global importance. Füsun ÜLENGİN RTraL '09 Chair, Dogus University, Istanbul, Turkey CONTENTS Committees i Words of Welcome and Gratitude ii Introduction iii Chapter 1 Global Issues in Logistics and Transportation 1 Potential to Reduce GHG through Efficient Logistic Concepts 3 Werner Rothengatter A methodological framework for the evaluation and prioritisation of multinational transport projects: the Case of euro-asian transport linkages 21 Dimitrios TSAMBOULAS, Angeliki KOPSACHEILI Container port throughput performance - case study: Far east, north west european and mediterranean ports 29 Vesna DRAGOVIC-RADINOVIC, Branislav DRAGOVIC, Maja SKURIC, EmirĞIKMIROVlC and Ivan KRAPOVIC Chapter 2 Regional Issues in Logistics and Transportation 35 Logistics service providers in turkey: A panel data analysis 37 Emel AKTAŞ, Füsun ÜLENGİN, Berrin AĞARAN, Şule ÖNSEL Milestones in the process of survey preparation for the logistics sector: case study for Istanbul, Turkey 43 Evren POSACI, Darçın AKIN Chapter 3 Education and Training in Logistics and Transportation 51 Education in transport and logistics in an age of global economy 53 Yücel Candemir The role of education and training in the supply chain sector 59 David Maunder Chapter 4 Supply Chain Management 64 Modeling reverse flows in a closed -loop supply chain network 67 Vildan ÖZKIR, Önder ÖNDEMİR and Hüseyin BAŞLIGİL Strategic analysis of green supply chain management practices in T urkish automotive industry 73 Gülçin BÜYÜKÖZKAN and Alişan ÇAPAN A new framework for port competitiveness: the network approach 79 Marcella DE MARTINO, Alfonso MORVILLO Chapter 5 Sustainable Transport Policies, Traffic Engineering 87 Clean transport: innovative solutions to the creation of a more sustainable urban transport system 89 Ela BABALIK-SUTCLIFFE Effects of urban bottlenecks on highway traffic congestion: case study of Istanbul, Turkey 95 Darçın AKIN and Mehtap ÇELİK Establishing an effective training module for IMDG code in MET institutions 105 Kadir CICEK, Metin CELIK An investment decision aid proposal towards choice of container terminal operating systems based on information axioms 109 Metin CELIK, Selcuk CEBI Chapter 6 Evaluation of Public Policies, Network Models and Environment 115 Possibilistic linear programming approach for strategic resource planning 117 Özgür KABAK, Füsun ÜLENGİN A structural equation model for measuring service quality in passenger transportation 125 G.Nilay YÜCENUR and Nihan ÇETİN DEMİREL Analysis of potential gain from using hybrid vehicles in public transportation 133 ¡rem DÜZDAR and Özay ÖZAYDIN Optimization of e-waste management in Marmara region - Turkey 141 İlke BEREKETLİ, Müjde EROL GENEVOIS Chapter 7 Contemporary Topics in Transport and Logistics 1 47 Future prospects on urban logistic research 149 Rosario MACÂRIO, Vasco REIS An analyze of relationship between container ships and ports development 155 Branislav DRAGOVIC, Vesna Dragovic-Radinovic, Dusanka Jovovic, Romeo Mestrovic and Emir Ğikmirovic A holistic framework for performance measurement in logistics management 161 Yasemin Claire ERENSAL Heuristics for a generalization of tsp in the context of PCB assembly 167 Ali Fuat ALKAYA and Ekrem DUMAN Premium e-grocery: Exploring value in logistics integrated service solutions 173 Burçin BOZKAYA, Ronan De KERVENOAEL and D. Selcen Ö. AYKAÇ Travelers response to VMS in the Athens area 179 Athena TSIRIMPA and Amalia POLYDOROPOULOU Regional airports and local development: the challenging balance between sustainability and economic growth 189 Rosârio MACÂRIO and Jorge SILVA Chapter 8 Transport Planning and Economics 195 How financial constraints and non-optimal pricing affect the design of public transport services 197 Sergio R. Jara-Diaz and Antonio Gschwender Revenue management for returned products in reverse logistics 203 Mesut KUMRU Intra-city bus planning using computer simulation 211 Reza AZIMI and Amin ALVANCHI Chapter 9 Planning, Operations, Management and Control of Transport and Logistics 217 A review of timetabling and resource allocation models for light-rail transportation systems 219 Selmin D. ÖNCÜL, D. Selcen Ö. AYKAÇ, Demet BAYRAKTAR and Dilay ÇELEBİ An approach of integrated logistics HMMS model under environment constraints and an application of time scale 225 Fahriye Uysal, Ömür Tosun, Orhan Kuruüzüm Freight transport planning with genetic algorithm based projected demand 231 Soner HALDENBILEN, Ozgur BASKAN, Huseyin CEYLAN and Halim CEYLAN Chapter 10 Transport Modeling 239 Inverse model to estimate o-d matrix from link traffic counts using ant colony optimization 241 Halim CEYLAN, Soner HALDENBILEN, Huseyin CEYLAN, Ozgur BASKAN The impact of logistics on modelling commercial freight traffic 251 Ute IDDINK and Uwe CLAUSEN A comparative reviewof simulation-based behavior modeling for travel demand generation 257 Seda Yanık, Mehmet Tanyaş An efficiency analysis of turkish container ports using the analytic network process 269 Senay OĞUZTİMUR, Umut Rıfat TUZKAYA A multi-objective approach to designing a multicommudity supply chain distribution network with multiple capacities 277 Gholam Reza Nasiri, Hamid Davoudpour and B.Karimi Chapter 11 Freight Transportation and Logistics Management 283 Evaluation of turkey’s freight transportation 285 Burcu KULELİ PAKand BaharSENNAROĞLU Short sea shipping, intermodality and parameters influencing pricing policies in the Mediterranean region: The Italian context 291 Monica GROSSO, Ana-Rita LYNCE, Anne SILLA, Georgios K. VAGGELAS Relevant strategic criteria when choosing a container port - the case of the port of Genoa 299 Monica Grosso, Feliciana Monteiro Determination of optimum fleet size and composition - A case study of retailer in Thailand 307 Terdsak RONGVIRIYAPANICH and Kawee SRIMUANG New container port development: forecasting future container throughput 313 Dimitrios TSAMBOULAS, Panayota MORAITI Sea port hinterland flows and opening hours: the way forward to make them match better 319 Marjan BEELEN, Hilde MEERSMAN, Evy ONGHENA, Eddy VAN DE VOORDE and Thierry VANELSLANDER International road freight transport in Germany and the Netherlands driver costs analysis and French perspectives 327 Laurent GUIHERY Chapter 12 Transport and Land Use 335 Land rent and new transport infrastructure: How to manage this relationship? 337 Elena SCOPEL Effects of pavement characteristics on the traffic noise levels 345 Aybike ONGEL and John HARVEY Fuzzy medical waste disposal facility location problem 351 Yeşim KOP, Müjde EROL GENEVOIS and H. Ziya ULUKAN Chapter 13 Transport Infrastructure and Investment Appraisal 357 Agents’ behavior in financing Italian transport infrastructures 359 Paolo BERIA Free trade agreements in the mediterranean region: a box-cox analysis 367 Matthew KARLAFTIS, Konstantinos KEPAPTSOGLOU and Dimitrios TSAMBOULAS Chapter 1 Global Issues in Logistics and Transportation Research in Transport and Logistics 2009 Proceedings Do u University Werner Rothengatter Potential to Reduce GHG through Efficient Logistic Concepts Abstract Saving GHG is the biggest challenge of the forthcoming decades. The EU has decided to re- duce CO2 by 20% until 2020 (compared with 1990), if other countries follow, even by 30%. In freight transport it will be difficult to achieve such figures, because freight transport is growing faster than GDP and energy efficiency is improving only slowly. Most studies as- sume that energy savings in freight transport in the first instance stem from more efficient propulsion technology. But technical progress on this field will by far not achieve the ambi- tious targets of CO2 reduction. Therefore, the logistic concepts and processes have to be ana- lyzed with respect to their potential to contribute to CO2 savings. The German Ministry of Economic Affairs has launched the project LOGOTAKT (2007) which is on developing a logistic concept to save energy and reduce climate gases of freight transport. LOGOTAKT presents a new concept of scheduled logistic services which is organized as a virtual moving platform, serving the entry points and points of transshipment in defined time intervals. It is organized as an open network for which different companies have access and can participate. For long-haul transport railway service can be included. For this purpose a subset of marshalling yards is changed to logistic centers to bundle freight which can be de- composed into pallets on the last mile to the clients. A logistic concept will be presented which uses all of the above options. Simulation exercises show that the concept may save about one third on CO2 emissions in the affine sectors which makes about 10% for the total freight transport market. This is possible in a self-financing regime without state subsidies. Together with other measures it seems to be possible to achieve CO2 reduction values of 20- 30% in the sector of freight transport and logistics and to meet the global EU reduction target. 3 Research in Transport and Logistics 2009 Proceedings Do u University Potential to Reduce GHG through Efficient Logistic Concepts 1 Introduction Saving GHG is the biggest challenge of the forthcoming decades. The EU has decided to re- duce CO2 by 20% until 2020 (compared with 1990), if other countries follow, even by 30%. In freight transport it will be difficult to achieve such figures, because freight transport is growing faster than GDP and energy efficiency is improving only slowly. As a result most of the CO2 savings would have to be provided by better logistics concepts. The logistic concept LOGOTAKT will be presented which is based on a virtual moving logis- tic platform and provides deliveries regularly on a fixed schedule. Railways are integrated, supposing that major marshalling yards can offer high level logistic facilities for to consoli- date large consignments from pallet units. The time table of the moving platform might be adjusted to the frequency and volumes of delivery. Simulated tests are run on a 6, 12 and 24 hrs schedule which seems to meet the requirements of many shippers on long-haul relation- ships. The paper will elaborate furthermore on the institutional requirements for an open network to bring logistic alliances together, the potential transaction cost and the problems of privacy. It will come out with the conclusion that better logistic concepts can contribute more than 10% savings of energy/CO2 and that the challenge to reduce 20-30% of CO2 emissions in freight transport until 2020 is not unrealistic, at least for the key segments of the market. This is un- derlined by reporting on voluntary actions of the industry, e.g. DHL has announced to reduce the CO2 emissions by 30% until 2020. If this is possible in the most complex market of parcel service then it should also be possible in market segments with more simple operations, plan- ning and management. 2 Contribution of Freight Transport to Global Warming The transport sector is responsible for 24% of the world-wide CO2-emissions. In the OECD countries this share has already grown to 30% and a further increase is highly probable. 4 Research in Transport and Logistics 2009 Proceedings Do u University benötigt. Figure 1: Development of CO2 Emissions in OECD Countries by Sector Source: ECMT, 2007 Figure 1 shows that the emissions of freight transport in OECD countries are about on third of the total transport emissions and that road freight emissions are increasing at a higher speed compared with the emissions of road passengers. Reason is that the growth of passenger transport is tending towards a saturation level in industrialized countries (declining popula- tion, modest growth of income for low and middle income classes). Furthermore the rising energy prices have revealed a high technological potential for fuel savings (smaller, lighter, energy efficient cars), i.e., a trend towards less powered cars and a reduction of SUV sales can be observed. It can also be concluded from the Figure that rail and inland waterway shipping do not play a dominant role in land-borne freight transport. Reasons are (1) the low share of the freight transport market which is held by the railways, and (2) the comparatively low emissions of CO2 per tonkm of transport which is about 15-30% of the specific emissions of road freight transport, on the average (see INFRAS/IWW, 2004). When it comes to freight transport the potential for energy savings appear much lower com- pared with passenger transport. Reasons are: - Firms plan their logistics by and large efficiently, - Firms buy the vehicles according to economic advantages (not according to prestige arguments as in the case of passenger car purchase), - Energy is a cost factor and the industry will try to develop least cost strategies for energy consumption, - The growth rate of freight transport (in terms of tonkm) is much higher than in passenger transport. In the last decades it was higher than the growth rate of GDP in the OECD countries. 5 Research in Transport and Logistics 2009 Proceedings Do u University Therefore, a stabilization of CO2 emissions of freight is already regarded a success in the lite- rature and the EU Commission is expecting only a modest contribution to CO2 reduction, much lower than the targeted average of 20% to 30% until 2020 compared with 1990. 3 Elements of a CO2 Reduction Strategy in Freight Transport General Action List The most important elements of more intelligent logistics are (see EC, 2008; German MoT, 2008; Manheim, 1999): (1) Use of better technology, modern engines with less fuel consumption Until now the improvements of environmental quality in freight transport has focused on toxic exhaust emissions (NOx, Particles) and noise reduction. This reflects in a rising share of EURO 4 and EURO 5 trucks on the highways. Parts of EURO 6 technology (particle filters or traps) are already on the market although EURO 6 is not yet defined. But there is little progress with the reduction of fuel consumption. More technical intelligence was invested in the increase of horsepower to keep speed high in mountainous areas. Presently there is no specific action at sight within the EU or some of the member countries which could be consi- dered comparable to the effort on the side of passenger cars (120 g/km limit planned, to be enforced by a penalty system). As fuel consumption is an important factor in competition, because the share of fuel costs has increased drastically in the past years, one can assume that the potential of fuel savings through better technology will not be very high in the next years to come. (2) Adjusted movement of vehicles (lower speeds, continuous movement) Driver education to save energy has become popular and it is reported that this measure alone can lead to reductions of about 5%. It is achieved through smooth and precautious driving, minimizing the acceleration and deceleration cycles and lowering speeds. Considering the latter one has to add that usually trucks go higher speeds than allowed because overrunning speed limits is not penalized in most countries if the overrun is less than 10% of the limit. Furthermore, bypassing of slower vehicles is reduced in energy efficient driving as small speed differentials bring little time advantage. (3) Better loading of vehicles Better loading of vehicles means in many cases that the fixed time tables of departure are sub -Paris, loads. (4) Improved planning of tours and organizing return loads One way to increase efficiency and energy use is to reduce empty running of trucks. In many cases tours can not be paired in a way that tour and return tour are symmetrical. This is shown by the example in Figure 2: The company collects the load from south-west parts of Germany 6 Research in Transport and Logistics 2009 Proceedings Do u University with destinations in London, Hull and Manchester. For the return tour load is collected partly in Liverpool and south-west to London and carried to Paris. Parts of the load are unloaded in Paris and added to the truck load of a French haulage company which serves southern France. The rest is carried to Stuttgart where parts are loaded on a truck of a third company serving southern Germany. The example underlines that the intelligence of the business lies in the first instance in planning the tour in a way that the moving capacity is optimally loaded. Figure 2: Tour and Return Tour of a Stuttgart Haulage Company Source: Liedtke, 2006 (5) Less just-in-time and direct delivery operations Just-in-time delivery has in the past years dominated the logistic requirements of the shippers. The hauliers have adjusted to these performance requirements successfully as long as the ma- jor input resources, as there are personnel and energy inputs, were cheap. Low wages through hiring drivers from Bulgaria, Poland or Romania, and low fuel prices were together with improved logistics the reasons for a drastic decrease of road freight transport cost in an or- der of magnitude of about 40% after starting the EU liberalization on this market. In the present phase of the economy wages are going up and in particular the energy prices are bit- ing. Together with user tolls on highways and motorways, which have for instance been in- troduced in Switzerland, Austria and Germany and increased in Italy, Spain or France this has lead to a substantial increase of variable costs of trucking. As just-in-time service and direct delivery are in many cases causing an inefficient use of loading capacity one can observe that just-in-time is successively substituted by just-in-sequence delivery. This means that the 7 Research in Transport and Logistics 2009 Proceedings Do u University schedule of industrial workflow is determining the frequency of delivery, and often this can be adjusted in a way that wider time windows are opened for the transport service. (6) Integration of rail, inland waterways or coastal shipping in long-haul supply chains While just-in-time requirements in general could only be fulfilled by road transport the just- in-sequence principle also allows for integrating rail, inland waterways or coastal shipping. The crucial performance requirement is not so much the total delivery time rather than the reliability and punctuality of the service. This is in the case of railway service highly influ- ence by the number of transshipment operations on the tour and the number of changes of borderlines. The railways still have big difficulties with fulfilling this essential performance requirement, but first examples show that there is a big potential. In the New Opera project of the EU it is estimated that the railways could triple their transport volume (in tonkm) until the year 2025 if they could exploit the inherent potentials in their organization. (7) Better pairing of transport alongside busy corridors On the trans-european corridors there is presently a dominance of North-South transport be- cause the ARA ports (North Sea) are very strong, attracting most freight volume and trans- shipping the freight to the south, south-west and south-east. Analyzing the gravity power of ports one comes to the conclusion that the Mediterranean ports have much more potential than they are using presently. The reasons are in most cases missing efficiency and missing links to the land-borne networks. In the future this picture might change and one can observe that the major players in the logistic business increase their activities in the Mediterranean ports. As a result the long-haul transport alongside the North-South corridor could be better balanced and the transport capacity better used. This concerns the railways in the first instance, which pre- sently show a poor loading on international return tours. (8) Better consolidation at bundling points through establishing alliances of shippers The example in Figure 2 demonstrated how firms can increase the transport efficiency through co-ordinated logistics. Such co-operations have only developed in the small but not in the large, i.e., bringing the logistics of big players together. Main reason is that the big com- panies are afraid of losing privacy of their individual strategies. Logistics is seen as a part of market. Question is whether the increasing pressure on variable costs will create incentives to start co-operative solutions. It is a basic proposition of this paper that most of the logistic po- tential for reducing energy consumption and CO2 emissions lies on the field of bundling activ- ities of the market players. Development of Marshalling Yards to Railports Marshalling yards are traditionally used as points to block and make up freight trains. The new idea is to use these points as centers for freight processing and eventually additional val- ue adding services. This presupposes that the volume of freight is high enough, i.e. pallets of unitized goods have to be consolidated to generate wagon loads, wagons are composed to wagon groups with identical destinations and wagon groups are blocked to form full trains. If such a process can be organized in a competitive way then this could be the point of departure towards a revival of single wagon operations in a rail system. From the perspective of railway companies this would lead to a new and innovative field of activity with strong support of communication systems (tracking and tracing of wagons and eventually pallets), automatic 8 Research in Transport and Logistics 2009 Proceedings Do u University processing at railports and automatic blocking of trains (automatic coupling technology). Fig- ure 3 gives an example for the allocation of railports and transshipment points from the view- point of the Deutsche Bahn AG. TSP Railport Figure 3: Example for a Network with Railports and Transshipment Points (TSP) Source: LOGOTAKT, 2008 4 The LOGOTAKT Concept The General Concept LOGOTAKT stands for a strictly scheduled logistic service, trying to consolidate the freight consignments at bundling points and using the appropriate means of transport for every seg- ment of the transport chain. Contrasting the usual logistic schemes which try to optimize the system and the processes for a single firm LOGOTAKT tries to bring together the transport needs of many players to consolidate their consignments. This means that in a logistic envi- ronment which is characterized by a trend of diminishing size of individual consignments LOGOTAKT can be thought as a virtual moving logistic platform with fixed entry and deli- very times which serves many players and aims at substantially increasing the volumes of transport on the major corridors. This can be achieved through an open network to which every firm has access, in principle. Therefore it is not sufficient to develop the logistic concept, only. Furthermore, a business model has to be created to set the conditions clear under which the firms can participate, de- termine a pricing/cost allocation system for the transport services and establish a concept 9 Research in Transport and Logistics 2009 Proceedings Do u University which guarantees privacy for the participating players such that no party has to sacrifice inter- nal business information. Figure 3 shows the important elements of the concept. A Z B D B B Y X Open Networks Intemodality for main run Robustness Scheduled transport activity Figure 3: Basic Elements of LOGOTAKT Source: Logotakt, 2008 There are a number of pre-conditions for the acceptance and the success of the concept: - Use of modern sensor and communication technology. - Application of automatic processing for consolidation, loading/unloading and train blocking. - Crossdocking for time sensitive freight from light to heavy goods vehicles or rail wagons. - Reliability and robustness of the processes, optimization of buffering. - Savings of inventory holding. At the end of the day the participation in the virtual moving platform has to be profitable for the firms. This does not mean that the transport processes as such have to be cheaper or faster compared with the stand-alone optimization of firms. It is the total costs of the logistic systems and the logistic processes which have to be compared to each other. There is some probability that LOGOTAKT will be a profitable system if firms are facing the following situation with and without: 10 Research in Transport and Logistics 2009 Proceedings Do u University Without (stand-alone) With (participation in the network) High volumes per transport activity (e.g. Smaller volumes per transport activity trucking tour) High speed of vehicles to guarantee just-in- Lower speeds feasible, if schedules not time delivery disturbed Big time intervals between transport Smaller time intervals between transport activities activities, e.g. 6 hrs. service Relatively high minimum inventory holding Reduction of inventory holding for for preliminary goods preliminary goods Relatively high minimum stock of produced Lower minimum stock of produced articles articles Relatively high variance of volumes, Smaller variance of volumes, higher therefore special direct tour necessary reliability, less special tours High buffering and physical swaps Controlled and low buffering, no physical swaps Table 1: Conditions for Profitability of LOGOTAKT Some Scientific Challenges LOGOTAKT is generating a number of new scientific challenges for different disciplines of economics, engineering and computer science. Although focus is laid on the economics part one can immediately see from Figure 4 that the economic approaches are closely related to the contributions of the other disciplines or intrinsically interdisciplinary (such as micro and macro-modelling). The following aspects deserve particular attention: (1) Micro-modelling of logistic and transport activities Micro-modelling means in this context that the behavior of single shippers, forwarders, transport companies and recipients (producers, consumers, retail businesses) is simulated individually. While there are a large number of logistic optimization models existing trying to find out optimal solutions for specified logistic problems of a single firm by using optimization techniques and graph theory there exist only a few approaches which can be with randomly distributing firms in space, assign production activities to these units and generate transport needs. Transport orders are traded on a simulated market and assigned to forwarders which hire transport firms for performing the tours. Consignments are allocated to vehicles and the ours are allocated to the infrastructure networks. In this step it is important to disaggregate the transport needs in enough detail to simualte the typical transport patterns related to commodity groups. This simulation process, carried out with the software INTERLOG, allows for aggregating individual transport activities using distributions and constraints derived by aggregate statistics. As a result the aggregate statistics are reproduced by a pattern of individual activities. In principle there are an infinite number of individual activity patterns which can generate the aggregate picture. Comparable to the theory of traffic distributrion in space one can identify a pattern of individual activities which has the highest probability to occur. 11 Research in Transport and Logistics 2009 Proceedings Do u University Network- Market model L O G O T A K T Model Firm 5 Micro Model Firm 1 Firm 4 Firm 1 Firm 3 Firm 2 Alternatives JIT Inventory Holding Milk-Runs Direct Non-regular Tours Figure 4: LOGOTAKT Modelling and Checks for Improvement Source: Logotakt, 2008 Micro Model of Firms in Space Aggregation of Transport Activities Source: Liedtke, 2006 Until now such a micro-macro-modelling is only possible for about 60% of the freight transport activities in Germany. Analyzing the remaining market segments one can assume that they are less affine to the LOGOTAKT service (e.g.: bulk cargo, (frozen) food, fruit). Figure 5 shows that the aggregated micro flows reproduce the traffic loads by trucks on the 12 Research in Transport and Logistics 2009 Proceedings Do u University German motorway and highway system fairly well. To make the picture complete, a macro simulation for freight and passenger flows is applied and co-ordinated with the aggregated micro flows. The complete macro flow pattern is used to derive the resistance parameters for the network links (time and operating costs) which are inputs for the micro-simulation. This feedback modelling mechanism is applied for time intervals of the day to take the different traffic conditions into account. The foundation for this micro/macro-modelling procedure has been laid through the OVID project for the German Ministry of Research and Education (Ovid, 2003). Preliminary Results for Selected Companies Before the main phase of the study was started a rough estimation of the LOGOTALT potential was carried out for two selected companies, for which the data base for all operations for a selected time perios were made available. Four scenarios were run: - Scenario 0: Transport by trucks, only (Base Case) - Scenario 1: Transport by rail on the main carriage - Scenario 2: Transport by rail on the main carriage, access/egress organized by individual firms) - Scenario 3: Least cost organization of transport using the LOGOTAKT concept (open access). - Scenario 4: Least cost organization of transport using the LOGOTAKT concept (closed shop). (1) Potential for a large company Cost comparison (%) Vehicle km Transport Transshipment Total Total km Delta km Total km cost % Cost % Cost % road Road % rail Scenario 0 95.3 4.7 100 2,471,291 Scenario 1 93.9 7.3 101.2 411,593 -83.3 2,439,645 Scenario 2 92.2 5.0 97.2 458,935 -81.4 2,439,645 Scenario 3 86.3 5.1 91.4 1,496,892 -39.4 1,194.973 Table 1: Effects of Different Scenarios on Traffic Activity and on Costs (2) Potential for a medium sized company Cost comparison (%) Vehicle km Transport Transshipment Total Total km Delta km Total km cost % Cost % Cost % road Road % rail Scenario 0 100 100 6,334,000 Scenario 3 65.7 4.8 70.6 2,305,000 -63.6 2,425,000 Scenario 4 91.6 5.0 96.6. 6,579,000 +3.9 645,000 Table 2: Reduction Potential for Vehicle kms and Transport Costs (2 Selected Companies) Source: LOGOTAKT 13 Research in Transport and Logistics 2009 Proceedings Do u University The time base has been chosen differently for the companies such that only the relative figures matter. With respect to these relative figures it is necessary to underline that this exercise was to quantify the potential and not the expected value of realization, considering all barriers to introduce the concept properly. With these restrictions for the interpretation of the figures in mind one can derive the following results: (1) The potential of cost savings and of integrating the railways in the long-haul transport chain seems unexpectedly high. Of course this presupposes that the railway companies become commercially efficient and reliable partners, which until now is underlined by many studies (see for instance the New Opera study for the EU Commission, which forecasts a tripling of rail freight transport until the year 2025 and derives this figure from a substantial improvement of technical and organizational capability of the railways.) However, only a few railway companies are actually on the way to achieve the necessary logistic performance. (2) Just-in-time and direct service can easily be shown inferior to a scheduled moving platform, because the latter needs less inventory holding, additional tours with suboptimal loads and physical swaps. (3) Introducing LOGOTAKT as a closed system for a single firm leads to an improvement of logistic efficiency but uses by far not the potential which is offered by an open system with a set of players. Single optimization with LOGOTAKT might lead to more traffic although the total costs are decreasing. This underlines that an incomplete realization of the concept can lead to counterproductive results with respect to congestion and environmental impacts. (4) It is very important to identify the critical mass for the financial viability of the service. Below the critical mass considerable losses might occur because of idle the incremental costs of access for additional firms are strictly decreasing. Approach to Estimate the Critical Mass for LOGOTAKT LOGOTAKT affine segments can be found in the market for general cargo transport. This includes for instance consumption and investment goods, preliminary goods, food, building materials or liquid chemicals. Other markets are less relevant. Regional transport is organized differently, in many cases on a day-to-day basis, Special and unitized transport tasks require special vehicles or swap bodies, have to be organized on an indiviudal time schedule and thus can hardly be integrated in a general logistic platform. 14 Research in Transport and Logistics 2009 Proceedings Do u University Segmentation of Transport Market General 34% Regional 43% Special Unitized 17% 6% Figure 6: Segmentation of the German Freight Transport Market Source: LOGOTAKT, 2008 For the LOGOTAKT affine market segments a multi-agent model has been established to check for the thresholds which are relevant for the profitability of the concept for the companies. Without referring to the details of the model and the data output derived from first applications one can draw some general conclusions: (1) There exists lower and upper bounds for profitability with respect to the volume of freight, measured by the number of pallet loads. (2) The relevant range is determined by - value of the consignments - fluctuation of production - probability of extra tours (in addition to the planned tours) - cost of inventory holding. Depending on the combination of influencing factors the lower bound is in the order of magnitude of 2-3 pallet while the upper bound widely differs. In exceptional cases it is about 4 pallets (low value, low fluctuation, little probability of extra tours, low cost of inventory holding) int other cases more than 40 pallets. Figure 7 gives a sketch of the principle shape of the cost functions for stand-alone and LOGOTAKT logistic regimes. While the existence of a lower bound is easily explained through the fact that a minimum volume of transport is necessary to justify highly frequent shipments of goods at low costs the existence of upper bounds is not intuitively understandable. But as a matter of fact each company has the alternative to ship their goods with own or contracted vehicle capacity and can derive increasing returns from higher volumes of shipment. As soon as the integration of the railways into the supply chain is possible at reasonable costs the upper bound goes up substantially and even may vanish. But note that the first comparisons have been carried out under restrictive assumption such that it is not possible to give more exact numerical estimators. The estimation of the total transport volumes of companies which are LOGOTAKT affine gives a high certainty that there is a market for the system. The minimum turnover which is 15 Research in Transport and Logistics 2009 Proceedings Do u University necessary to run the system with a low frequency (one service per 12 hrs) is overrun 15 times in the simulation which justifies the expectation that a considerable subset of entry and transshipment points can be served even on a 4 hour basis. This would increase the potential and attract freight transport with higher logistic requirements. Total cost per pallet present logistic regime Logotakt logistics 2 pallets 3 pallets 4 pallets .... n pallets Figure 7: Example for an interval of profitability (sketch of principle results for consignment values of about 50,000 Euro/ton Source: Pischem, 2008 5 Effects on road transport and CO2 emissions For the firms analyzed the savings of external costs can be calculated on the base of the Handbook (CE et al, 2008) for marginal costs or of the study FACORA of Infras/IWW (2004) for average costs. As it is more easy to compare and aggregate average figures we take the latter cost basis. The average values for different modes of freight transport are exhibited in Figure 8. 16 Research in Transport and Logistics 2009 Proceedings Do u University Figure 8: External Cost of Freight Transport Source: Infras/IWW, 2004 Figure 9: External Cost Comparison in the EU Handbook Source: CE et al., 2008 17 Research in Transport and Logistics 2009 Proceedings Do u University The recently published Handbook is based on the marginal cost philosphy which has dominated the EU financed studies internalisation of externalities since 1998, when the White Paper on Fair and Efficient Pricing was published by the Commission. Although there are a number of literature pieces which criticize the strong and abstract assumptions of neo- classical welfare theory it still can be regarded the main stream of economic approaches worked out for the Commission (see Rothengatter, 2003). The consequence of this approach is that the weight of external congestion costs (time and operating costs exceeding the optimum) is relatively high while the weights of external costs of accidents, air pollution and climate change is relatively low. In the comparison table of Infras/IWW congestion costs is not considered because it represents a different type of externality. Congestion costs are time and operating costs. But they are widely internal to the set of all users such that they do not affect the competitive balance between transport modes negatively for the competitors. Keeping this in mind an intermodal comparison between different modes of transport including congestion costs is not meaningful. Therefore we restrict to the Infras/IWW picture. Infras/IWW have used two different values for the ton of CO2, 20 Euro (low) and 140 Euro high). Evaluating the recent IPCC reviews and the Stern review, followed by a series of high level political activities to reduce CO2 emission world-wide, only the high value fits into this scenario of high political concern. Starting from these preliminaries we can give a rough estimation of the effect of introducing the LOGOTAKT system. Taking the figures from Figure 8, adjusting them to the tonkm scale and weighing them with the values of Table 2 we result in a reduction of 33% for firm 1 and 56% for firm 2. As the estimation for the potential of firm 2 seems to be very optimistic we take only the first reduction figure and apply it to the LOGOTAKT-affine market share of 34 %. Result is a savings of external costs and of CO2 in an order of magnitude of 11%. Taking into account a reduction potential from technology (engines, aerodynamics, tyres) of 5% and from driver education of 5% we result in a total reduction potential of 20% in the next years. It is highly probably that also in the non-affine market segments, this is two thirds of the total, there is a some potential to save external costs and CO2 through better logistics. But other concepts will be needed to quantify this potential. One example is the market for parcel service which is fastly growing. In the past this was a market for air cargo and trucking, using different sizes of vehicles. It is hard to discover a potential for energy reduction, looking at the logistic operations from an outside view. However, as soon as the incentives are set right, the management of firms might discover unexpectly high potentials. One example is the DHL company, which has set the target to reduce energy consumption and CO2 emissions by 30% until the year 2020, based on 2007. The main instruments are the integration of railway service on a main corridor (Frankfurt-Leipzig, served by a high-speed parcel train by night), optimal stucture of the truck fleet, education of drivers and optimal planning of milk run tours. A second example is the Webasto company which is producting roofs for Volkswagen cabriolets in Portugal. In a first round of optimisation they optimised milk runs for the German supply industry, consolidation of tours and routing from Germany to Portugal, which saved about 25% of transport costs, and energy consumption in a similar range. In a second round they are looking for partners among the other Volkswagen suppliers in Portugal. This would be a frist step towards a closed LOGOTAKT system. In a third step, under the assumption that there is enough transport volume, the railways could be integrated, constructing a supply corridor from Germany to Spain (2 Volkswagen plants) and Portugal (1 Volkswagen plant) together with other companies in an open network. This underlines that 18 Research in Transport and Logistics 2009 Proceedings Do u University y logistic service with open access the decision situations can change in the future. 6 Potential to Reduce CO2 Emissions of Freight Transport Through Efficient Logistics It is a big challenge to make freight transport so energy efficient that the global EU reduction targets can be met. Although this market seems to be growing with high rates in the medium future a high reduction potential can be identified. Looking at the two companies which have been analysed in some detail with respect to their present logistic patterns and the reduction potential through implementing the LOGOTAKT concept one results in figures between 33 and 56% for the LOGOTAKT-affine market segments. Assuming that the 56% reduction simulated for the medium sized firm are too optimistic and the 33% simulated for the large firm are more realistic, and weighing this figure with the market share of LOGOTAKT-affine transport one results in an average figure of 11% for the CO2 reduction potential. This is achieved by a high frequently turning logistic platform and an open network to foster co- operation of the players. If we add 5% for technological improvement until 2020, which is according to trend development and a further 5% which can be achieved through driver education and assistant systems to smoothen driving cycles one results in an average figure of more than 20% which the sector freight and logistic can contribute to CO2 reduction in industrialised countries. Last but not least there is some probability that also in the market segments, which are not LOGOTAKT-affine (66% of the total) there is at least a small potential which can be added to the above figure. One can argue that this estimation is theoretical and derived from simulation exercises. But there are already best practice examples from the industry which point to the same order of magnitude, i.e. 20-30% (see the cases of the companies Webasto and DHL). A central condition has to be met, however: Energy prices have to go up further, not so dramatic as in the past three years, but observable for the market players. This will stimulate a lot of decentral incentives in the companies to save energy and increase the acceptability to form logistic alliances or to participate in an open logistic network. As a result there is no reason to exclude the logistics sector from the global EU CO2 emission reduction targets. Despite the expected growth of freight transport in the future there is a high potential to achieve substantial energy and CO2 reductions by intelligent logistics. 19 Research in Transport and Logistics 2009 Proceedings Do u University Literature CE et al., 2008: Handbook on Estimation of External Costs in the Transport Sector. Brussels. EC, 2007: European Commission. Freight Transport Logistic Action Plan. Brussels. COM 2007. 607. Final. Brussels. ECMT, 2007: Cutting Transport Emissions. Paris. Logotakt, 2007: Project for the German Ministry of Economic Affairs. Led by D-Logistics Liedtke, G., 2006:An Actor Based Approach to Commodity Transport Modelling. PhD Thesis. Karlsruhe. Manheim, M. L., 1999: The Next Challenges in Transportation Research: Enhancing the Application of the Mind. In: Meersman, H., Van de Voorde, E. and W. Winkelmans (eds.): Proceedings of the 8th World Conference on Trnasport Reserach. Vol. 1. Oxford. New Opera, 2008: Project for the European Commission. Coordintated Action of the European Railway Industry. Paris. Karlsruhe. Karlsruhe. Pischem, M., 2008: Wirtschaftlichkeitsnachweis eines Transportnetzwerkes mittels Multi- Agenten-Simulation. Master Thesis. Karlsruhe. Rothengatter, W., 2003: How good is first-best? Marginal Costs and Other Principles. In: Transport Policy.10. 4. 345-358. 20 Research in Transport and Logistics 2009 Proceedings Do u University A METHODOLOGICAL FRAMEWORK FOR THE EVALUATION AND PRIORITISATION OF MULTINATIONAL TRANSPORT PROJECTS: THE CASE OF EURO-ASIAN TRANSPORT LINKAGES Dimitrios TSAMBOULAS 1, Angeliki KOPSACHEILI 2 Abstract - With the globalization of trade and the opening of borders, transport infrastructure projects are becoming more multinational than country specific. Moreover, several countries encounter scarcity of resources to finance them. On the other hand consultants, organizations and authorities are often obliged to perform evaluation of transportation infrastructure projects within a short time span and with limited–data environment. Thus, the need arises for a simplified, flexible and transparent methodological framework to prioritize investments in transport infrastructure located in more than one country. This paper presents such a framework, developed in three simple and consequent phases in order to be used at strategic level by policy makers and cope with limited availability and quality of data. The framework has been applied to prioritize 230 transport infrastructure projects of multinational character, comprising 18 countries from the Euro-Asian region, working together under a UN Project. Keywords - Project evaluation, transportation infrastructure, multinational investment INTRODUCTION Transportation infrastructure is a pre-requisite –though by no means a guarantee- of economic development [1]. This necessitates the improvement of transportation infrastructure and services, especially in less developed countries, which encounter scarcity of resources to finance them. Simultaneously, with globalization and opening of countries’ borders, infrastructure projects are characterized more multinational than country specific. This is also verified by the continuous development of initiatives such as Trans- European Networks (TEN-T), Trans European Motorway and Railway networks (TEM, TER), Pan-European Corridors, Euro-Asian Transport Linkages (EATL) etc. In view of this strategic role of transportation and the limited funding sources (national and/or international), the need arises to evaluate alternative infrastructure investments of multinational character and establish priorities among the different projects. The tendency today is to deviate from the so far conventional evaluation methods that tend to focus on a relatively limited set of impacts i.e. Cost-Benefit Analysis (CBA) [2]. Decision-makers need to know more than just construction costs and traffic performance; they need information on long-term and indirect impacts on society’s mobility as well as the ability to serve diverse needs [3]; and recently they need a better understanding of the social and political consequences of transportation infrastructure projects. Thus, the process of evaluation for selecting a project or a portfolio of projects may reflect social and political issues in addition to technical or more "rational" considerations, which in some cases are not considered critical in the selection process [4]. In addition transportation infrastructure projects can have several goals due to their more international/global character [2]. However, transport infrastructure development can benefit all regions concerned if a proper evaluation method is employed to incorporate all the diverse objectives and interests across the regions. Under this concept, formalistic evaluation methods might not be appropriate. Moreover, consultants, organizations and authorities are often obliged to perform evaluation of transportation infrastructure projects in short time spans or in a limited–data environment, so comprehensive information might be difficult to be collected and the development of better and sounder simplified evaluation methods is sought. These necessitate the use of Multi-Criteria Analysis (MCA) methods, the criteria of which are formed and refined through observations, discussions, experimentations and mistakes/ corrections. 1 Dimitrios Tsamboulas, Professor, National Technical University of Athens, School of Civil Engineering, Department of Transportation Planning and Engineering,dtsamb@central.ntua.gr 2 Angeliki Kopsacheili, Civil Engineer – Attiko Metro S.A., akopsacheili@ametro.gr 21 Research in Transport and Logistics 2009 Proceedings Do u University Literature review provides a wide range of evaluation methods and/or frameworks for transportation infrastructure projects [5]-[9], listing also various sources of error and bias in them. Consequently evaluation process cannot be a black box; critical judgment is required to probe the importance of every link in a investments prioritization framework. The development of coherent, well structured, flexible and simplified – but not simplistic- evaluation framework with “societal principles”, for transportation infrastructure projects, is quite useful. Few research initiatives exist that adopt an MCA for international transportation infrastructure investment initiatives {[0], some of them combining also a CBA [11]-[13]. This paper presents such a simplified methodological framework, structured in three levels (identification, assessment and prioritization) to secure the inclusion of all proposed projects, by employing sufficient limited criteria reflecting the societal values, the priorities and the available resources of the regions concerned, as well as the viability of the projects and their global/ international character. It is developed in order to be used at strategic level by policy makers and cope with limited availability and quality of data in countries concerned. The framework has been applied to prioritize 230 transport infrastructure projects of multinational character, comprising 18 countries from the Euro-Asian region, working together under a UN Project of ECE and ESCWA. METHODOLOGICAL FRAMEWORK Overview The objective of the proposed methodological framework is to identify within a group of projects, the ones for which the investments have to be made in a medium and long-term time horizon, regardless of their geographical location and the region/ country where are physically located. The framework comprises the following phases: - PHASE A - IDENTIFICATION: The identification phase concerns the collection of readily available information/ data regarding projects and identify the prospective projects to be further evaluated based on their funding possibilities and the common-shared objectives of the responsible authorities. - PHASE B - EVALUATION: The evaluation component of the framework is based on a combination of very well-known multi-criteria approaches such as: direct analysis of criteria performance, AHP (Analytical Hierarchical Process) and MAUT (Multi Attribute Utility Theory). MAUT employs sufficient but limited criteria reflecting, among other things, the transport policy priorities of the countries concerned, the available financial resources, the financial and economic viability of the projects and their international dimensions. AHP contributes to overcome subjectivity in deriving criteria priorities using pair-wise comparisons that make difficult to promote open biases towards specific criteria. Direct analysis of criteria performance is employed for deriving criteria scores and can perform relatively well when a limited amount of information is available. - PHASE C - PRIORITIZATION: In the final phase, the prioritization one, the selection of the better performing (in the evaluation) projects takes place and based on these performances, projects are prioritized in four Priority Categories (I, II, III and IV) over a specified time horizon. Analysis of Phases Phase A - Project Identification In this phase the distinction of projects in two major categories, those with committed funding and those without committed funding, takes place. Projects that have already secured necessary funding are not further evaluated and they are directly prioritized as Priority Category I. Projects without committed funding are further “screened” based on common objectives of the responsible authorities (i.e. national government, local authorities, funding bodies, banks, private sector organizations etc.). The screening levels are three and concern projects’ “relevance”, “readiness” and “viability”. All three levels are simple and easy to apply, so that the projects (local, national and international) that merit evaluation and consequently funding and implementation can be selected from the ones proposed by National Plans. The “relevance” level, expresses the relevant importance of a project within a group of projects as it concerns the international perspective, and thus the generic criteria used in this level can be grouped under three broad headings: (a) Related to international transportation policies and agreements 22 Research in Transport and Logistics 2009 Proceedings Do u University (b) Related to national transportation policies and objectives (c) Dealing with elimination of cross-border transportation problems (bottlenecks, missing links etc.) The “readiness” level concerns the maturity of the project in terms of planning and evidence of authorities commitment, and thus the generic criteria used in this level can be grouped under two broad headings: (a) Related with project status (existing studies, allocation of work among the responsible stakeholders, time plan for elaboration) (b) Related with planning organization’s and implementation authority’s commitment to the project The “viability” level concerns the expected transportation, economic and social benefits of the project, and thus the generic criteria used in this level can be grouped under three broad headings: (a) Related to financial and economic impacts and benefits (b) Related to societal and environmental impacts and benefits (c) Related to traffic impacts and benefits Thorough evaluation is performed in the next phase only for projects passing all the three screening levels. Finally, projects without committed funding that didn’t pass all of the screening levels were automatically classified in the last Priority Category (IV), which lists all projects to be implemented at the latest stage of the time horizon. Last, a secondary objective if this phase is to collect the necessary data in the right format, to serve as input in the evaluation phase. Phase B – Project Evaluation The objective of this phase is to derive project’s scores (degree of performance) and compile a record of project’s scores for use in the prioritization phase. The existence of several different types of projects as well as national objectives, favor the use of a method of multicriteria character, than simple monetary methods like CBA. Such a method will allow available information to be taken into account, even at the very preliminary level of project definition. At the same time some specific elements of particular interest for the decision- makers may be introduced. Thus Phase B, includes the following components: (a) Definition of criteria (b) Measurement of criteria (c) Criteria weighting (d) Derivation of total score per project (a) Definition of Criteria The specific evaluation criteria were developed in two “dimensions”: - the horizontal dimension called “Functionality/ Coherence” expresses the role of the project in the functionality and coherence of the network to be developed. - the vertical dimension called “Socio-economic Efficiency/ Sustainability” expresses the socio-economic return on investment as well as the strategic/ political concerns of the national political authorities, and of international ones in case of co-financing by them (e.g. EU, EIB, World Bank) Under these two fundamental orientations of the evaluation process, the following criteria have been introduced, which are aimed at covering all of the objectives and specifics relating to the evaluation of transport projects of international character. Horizontal Dimension: Functionality/ Coherence Criteria (CA) - Serve international connectivity (reaching a border crossing point or provide connection with a link that is border crossing); (CA1) - Promote solutions to the particular transit transport needs of the landlocked developing countries; (CA2) - Connect low income and/or least developed countries to major European and Asian markets; (CA3) - The project crosses natural barriers, removes bottlenecks, raises substandard sections to meet international standards, or fills missing links in the EATL; (CA4) 23 Research in Transport and Logistics 2009 Proceedings Do u University Vertical Dimension: Socio-economic Efficiency and Sustainability Criteria (CB) - Have high degree of urgency due to importance attributed by the national authorities and/or social interest; (CB1) - Pass economic viability test; (CB2) - Have a high degree of maturity, in order to be carried out quickly (i.e. project stage); (CB3) - Financing feasibility (CB4) - Environmental and social impacts (CB5) (b) Measurement of Criteria Criteria were measured first in a “physical scale” either by direct classification according to available data/measurable characteristics and/or by quality attributes, provided by preference judgment from the involved national authorities. The physical scale was chosen to be a simple five point with threshold values based mainly on criterion’s nature. As an example the physical scale/measurement of criterion of the extent the project is expected to increase traffic (Criterion CB2) is presented: A: By more than 15%, B: 10-15%, C: 5- 10%, D: less than 5%, E: Will not affect traffic. To make the various criterion scores compatible it was necessary to transform them into one common measurement unit or else transform “physical scale” measurement into a common “artificial scale” measurement. Criteria quantification was not be based on a sophisticated utility function but on a simple crooked linear function which connects threshold values of an artificial scale with threshold values of a physical scale. The artificial scale chosen is: A = 5, B = 4, C = 3, D = 2, E = 1, with 5 the highest value. Therefore: CJi 1,5 1 Where: J = A, B (representing the criteria dimension) i = 1,..,5 (representing the number of criteria in each dimension) (c) Criteria Weighting At this stage, for establishing the criteria weights Saaty’s Analytical Hierarchy Process (AHP) was used, because it is simple, transparent and widely accepted procedure. In addition, the existence of Eigen vector method in AHP provides fast and reliable weights: fast in expressing the short time necessary for its application; and reliable in minimising the subjectivity of weights’ values. It should be noted here that the resulted criteria weights should add up to unity. WJi 0,1 2 WJi 1 (3) Where: J = A, B (representing the criteria dimensions) i = 1,..,5 (representing the number of criteria in each dimension) (d) Derive Total Score per Project The total score of each transportation project was calculated by (4), which is based on multi-attribute utility theory, following the work of Keeney and Raiffa [14]. C 5 T.S.Project = CJi *WJi (4) J A i 1 Where: CJi [1,5] WJi [0,1] 24 Research in Transport and Logistics 2009 Proceedings Do u University J = A, B (representing the criteria dimensions) i = 1,..,5 (representing the number of criteria in each dimension) Phase D – Project Prioritization The objective of this phase is to provide a prioritization of all considered projects, based on their total scores and thus to assist decision-makers to realize the time-order of projects implementation in the desired time horizon. The classification of priority categories is: - I: immediate implementation of projects - II: short term implementation of projects - III: medium term implementation of projects - IV: long term implementation of projects The total score of each project (resulting from the application of equation 4) puts it in one of the four priority categories. Hence, if: - the project already has committed funding, it belongs to priority category I. - the project scores between 4-5 then it belongs to priority category II. - the project scores 3 -4 then it belongs to priority category III. - the project scores 1 -3 then it belongs to priority category IV. APPLICATION CASE STUDY Brief Description of the Case Study The project of Euro Asian Transport Linkages Development aims at prioritisation of transport infrastructure projects (road, rail, maritime, inland waterway, inland/border crossing) along the adopted Euro- Asian transport routes. More specifically the objectives are: - to identify and formulate international transport linkages and corridors between Europe and Asia, including highways, railways, inland water routes and seaport connections for multimodal transport operations; - to identify and promote major international transport facilitation conventions to enhance capacity to improve and harmonize national transport legislation and transport facilitation measures; - to assist in the establishment and effective functioning of national transport development bodies which are responsible for formulation and implementation of national action plans on transport facilitation and development; - to establish a database of experts and institutions by each Regional Commission, in consultation with its Member States, other development agencies, and relevant officials of the UN system; to create a website for the project to disseminate information about experts, institutions and project progress; and - to organize national and regional workshops and expert group meetings for promoting the project’s objectives. The involved 18 countries are: Armenia, Azerbaijan, Belarus, Bulgaria, China, Georgia, Islamic Republic of Iran, Kazakhstan, Kyrgyzstan, Moldova, Romania, Tajikistan, Turkey, Ukraine, Uzbekistan, Afghanistan, Russian Federation, Turkmenistan. The number of infrastructure projects (road, rail, maritime, inland waterway, inland/border crossing) considered is 230. Application of Methodological Framework Application of Phase A - Identification Out of the 230 projects considered 133 projects were directly categorized as Priority Category I, since they had committed funding and from the rest 97, all managed to pass the three screening levels and were considered for further evaluation. For the latter mentioned projects -that passed the three screening levels-, data collection was performed, based on specific project fiches/ templates. 25 Research in Transport and Logistics 2009 Proceedings Do u University Application of Phase B - Evaluation (a) Definition of Criteria Measurement Horizontal Dimension: Functionality/ Coherence Criteria (CA) - To what extent does the project improve international connectivity (for example, by reaching a border- crossing point or providing connection with a link that is border crossing; (Criterion CA1)? A: Greatly, B: Significantly, C: Somewhat, D: Slightly, E: Does not improve connectivity. - To what extent will the project promote solutions to the particular transit transport needs of the landlocked developing countries (Criterion CA2)? A: Greatly, B: Significantly, C: Somewhat, D: Slightly, E: Does not. - Will the project connect low income and/or least developed countries to major European and Asian markets (Criterion CA3)? A: Greatly, B: Significantly, C: Somewhat, D: Slightly, E: Does not. - Will the project cross a natural barrier, alleviate bottlenecks, complete a missing link or raise substandard sections to meet international standards along a Euro-Asian Transport route (Criterion CA4)? A: Greatly, B: Significantly, C: Somewhat, D: Slightly, E: Does not. Vertical Dimension: Socio-economic Efficiency and Sustainability Criteria (CB) - Does the project have a high degree of urgency due to importance attributed by the national authorities and/or social interest (Criterion CB1)? The project is... A: In the national plan and immediately required (for implementation up to 2008). B: In the national plan and very urgent (for implementation up to 2010), C: In the national plan and urgent (for implementation up to 2015), D: In the national plan but may be postponed until after 2015, E: Not in the national plan. - To what extent is the project expected to increase traffic (Criterion CB2)? A: By more than 15%, B: 10-15%, C: 5- 10%, D: less than 5%, E: Will not affect traffic. - At what stage is the project (Criterion CB3)? A: Tendering, B: Feasibility study, C: Pre-feasibility study, D: Planning, E: Identification. - What is the financing feasibility of the project (Criterion CB4)? A: Excellent (IRR > 15%), B: Very Good (IRR between 13% to 15%), C: Good (IRR between 10% to 13%), D: Medium (IRR between 4,5% to 10%), E: Low (IRR less than 4,5%) - To what extent does the project have potentially negative environmental or social impacts (pollution, safety, etc) (Criterion CB5)? A: No expected impact, B: Slight impact, C: Moderate impact, D: Significant impact, E: Great impact. Based on the criteria measurement described above, each criterion score was calculated for each project. (b) Criteria Weighting According to priorities set out from the national authorities pair wise comparisons of all criteria were made. The measurement of preference is done according to Saaty’s 9-points scale, where 1 implies the base factor is equally equivalent in importance to the other factor, and 9 implies the base factor is overwhelmingly more important than the other factor. For each country different weight are produced, which they are averaged. The resulting final weights for each criterion are (the subscript denotes first the criterion dimension and then the criterion number in each dimension): - Horizontal Dimension: Functionality/ Coherence Criteria (CA) WA1 = 3,13%, WA2 = 9,38%, WA3 = 19,79%, WA4 = 17,71%, - Vertical Dimension: Socio-economic Efficiency and Sustainability Criteria (CB) WB1 = 12,67%, WB2 = 12,67%, WB3 = 3,33%, WB4 = 7,33%, WB5 = 14% Application of Phase C - Prioritisation The priority categories for Euro-Asian Transport Linkages were: - I: projects, which have funding secured and are ongoing or planned and are expected to be completed in the near future (up to2010). - II: projects which may be funded and implemented rapidly (up to 2015). 26 Research in Transport and Logistics 2009 Proceedings Do u University - III: projects requiring some additional investigations for final definition before likely financing (up to 2020). - IV: projects requiring further investigations for final definition and scheduling before possible financing Results of the Application The application of the methodological framework for prioritization produced the following results for the 230 considered projects. - 133 projects were classified in the Priority Category I - 16 projects were classified in the Priority Category II - 10 projects were classified in the Priority Category III - 71 projects were classified in the Priority Category IV CONCLUSIONS The evaluation and prioritization of transportation infrastructure projects at multinational level requires decisions of public investment that have to be done jointly by the countries where the projects will be implemented as well as international organizations if funding is required by them. Thus, different objectives and priorities, as well as available resources have to be considered, rendering conventional evaluation methods not so useful. The aim of this paper is to offer decision-makers a methodological framework to prioritise projects in an international context. It is a coherent, well structured, flexible and simple –but not simplistic- with “societal principles”, for the evaluation and prioritization of multinational transportation infrastructure projects. It is structured in three levels (identification, assessment and prioritization) to secure the inclusion of all projects, as proposed by the countries, to employ sufficient but limited criteria reflecting the societal values, the priorities and the available resources of the countries concerned. On the other hand the viability of the projects and their international character are included as well. The first basic advantage of the framework is that it “saves time and money” in project evaluation procedure, since the first level rules out projects with insufficient information available and accepts for further evaluation only “mature” projects in terms of funding commitment, data availability and “political” commitment. The same benefit arises from the limited but sufficient number of criteria. In addition, the limited data requirements and the easiness in each application, renders the framework useful for the decision makers in countries with different levels of development. The application of the framework for the elaboration of Euro-Asian Transport Linkages Development Project has proved its applicability and its ability to produce results that coincide with the national plans and at the same time promote the international transportation connections. AKNOWLEDGEMENT The present paper is based on research carried out with the financing from UNECE for the Euro-Asian Transport Linkages Development, a UN Development Account Project, with ultimate purpose to identify the main Euro-Asian transport routes for priority development and cooperation. REFERENCES [1] Adler, H.A. (1987) Economic Appraisal of Transportation Projects: A Manual With Case Studies. Economic Development Institute of the World Bank. [2] Nijkamp, P., S.A. Rienstra, J.M. Vleugel. (1998) Transportation Planning and the Future. Wiley, England. [3] Victoria Transportation Policy Institute, Online TDM Encyclopedia, Updated June 2004 http://www.vtpi.org/tdm/ [4] Zanakis, S.H., T. Mandacovic, K. Gupta, S. Sahay, S. Hong. (1995) A Review of Program Evaluation and Fund Allocation Methods Within the Service and Government Sectors, Socio-Economic Planning Science, Vol. 29, No. I pp. 59-79. 27 Research in Transport and Logistics 2009 Proceedings Do u University [5] Nash, C.A. (1993) Cost-benefit Analysis of Transportation Projects. In Efficiency in the Public Sector: The Theory and Practice of Cost-Benefit Analysis. Edward Elgar, Aldershot. [6] Little, I.M.D. and J.A. Mirlees.(1994) The Costs and Benefits of Analysis. Project Appraisal and Planning Twenty Years On, In Cost-Benefit Analysis. Cambridge University Press, Cambridge. [7] Mackie, P. and J. Preston. (1998) Twenty-one Sources of Error and Bias in Transportation Project Appraisal, Transportation Policy, Vol. 5, pp. 1-7. [8] Giorgi, L. and A. Tandon. (2002) Introduction: The Theory and Practice of Evaluation. Project and Policy Evaluation in Transportation. Ashgate, England, pp.1-13. [9] Hartgen, D. T., and L. A. Neuman. (2002) Performance (A TQ Point/Counterpoint Exchange With David T. Hartgen and Lance A. Neumann). Transportation Quarterly, Vol. 56, No.1, pp. 5-19. [10] Tsamboulas D. A. (2007) A Tool for prioritizing Multinational Transport Infrastructure Investments”, Transport Policy, Volume 14, Issue 1, Elsevier, Netherlands [11] Sinha, K.C., and Li Zongzhi. Methodology for Multicriteria Decision-Making in Highway Asset Management. Presented at the 83rd Annual Meeting of the Transportation Research Board, Washington, D.C., 2004. [12] Nellthorp, J., S. Grant-Muller, H. Chen, P. Mackie, S. Leleur, D. Tsamboulas, A. Pearman, J. Latkinson. Comparing the Economic Performance and Environmental Impact of Trans-European Road Networks: the EUNET Project and Assessment Tool. Presented at the 2nd European Road Research Conference, Brussels, 7 - 9 June 1999. [13] Leleur, S. (1995) Road Infrastructure Planning: A Decision-Oriented Approach. Polyteknisk Forlag, Denmark. [14] Keeney, R.L., H. Raiffa. (1976) Decisions with Multiple Objectives; Preferences and Value Tradeoffs, Wiley, New York. 28 Research in Transport and Logistics 2009 Proceedings Do u University CONTAINER PORT THROUGHPUT PERFORMANCE – CASE STUDY: FAR EAST, NORTH WEST EUROPEAN AND MEDITERRANEAN PORTS Vesna DRAGOVIĆ-RADINOVIĆ1, Branislav DRAGOVIĆ2, Maja ŠKURIĆ3, Emir ČIKMIROVIĆ4 and Ivan KRAPOVIĆ5 Abstract ⎯ This paper tries to examine this issue from two viewpoints, the theoretical and the empirical ones. The first job is to estimate the port capacity performance, and the make empirical studies of major terminal operators in Asia and Europe, to find out their differences. The results shown there are huge differences between major Far East, North West European and Mediterranean terminals and ports. The study is based on data published on the web site of Containerization International (downloaded in the 2005 and 2006) and data from the Yearbook of Containerization International. For each port and terminal data are referred to the waterside operation of berths and QCs are the more important determinants of productivity. Keywords ⎯ Port throughput performance, Port productivity measures, Throughput optimization 1. INTRODUCTION We present effect on container terminals capacity performance with numerical results and computational experiments which are reported to evaluate the efficiency of Major Far East, North West European and Mediterranean Ports. The container throughput and terminal capacity performance in the 4 leading port ranges are analyzed in this study (East Asia, South - East Asia, Northern Europe and North - East Asia). Latterly three of the four top port ranges belong to Asia while Le Havre – Hamburg range presents European main port area. To test the calculation result of ports and terminals, with the assistance of one major container shipping line, the terminal operation data of most ports is collected for the purpose of comparison. The information collected covered the ports and terminals throughput for each port, berth characteristics, terminal performances and so on. Meanwhile, the terminal information in these ports is collected from both the Yearbook of Containerization International and websites, to find both the terminal and berth throughput. The Far East, North West European and Mediterranean ports and terminals selected for analyze are showing in subsections 2.1, 2.2 and 2.3. The comparison between these ports is made by terminal basis, and Far East terminals and ports are listed in Table 1. Apart from the classical theoretical references (e.g., [6], [7]), used for develops and describes methodology to study the container port capacity performance in this paper, it was necessary to review some segments of papers (e.g., [5], [8], [9] and [10], in which some individual elements of calculation of the various capacity performances have been considered. This paper is organized as follows. In Section 2 we give a brief description of the world container port throughput. Also, in this Section we present the Major Far East, North West European and Mediterranean ports selected for capacity performance analysis. Section 3 compares various measures of productivity between Far East (FE) and North West European ports (NEW), as well as Mediterranean ports (including Contship Italia Group – Eurogate (CIG-E)). Comparison of container terminals with leading terminal operators in selected ports is reported in Section 3, also. This implies a visual impact of capacity performance and their influence to ports and terminals productivity. In Section 4, we present effect on container terminal performance with numerical results and computational experiments which are reported to evaluate the productivity of the Hamburg, Hong Kong, Busan port layouts and Contship Italia Group - Eurogate. The final section concludes the paper. 1 Vesna Dragović-Radinović, University of Montenegro, Maritime Faculty, Kotor, Montenegro, bdragovic@cg.yu 2 Branislav Dragović, University of Montenegro, Maritime Faculty, Maritime Transport & Traffic Department, Kotor, Montenegro, bdragovic@cg.yu; branod@cg.ac.yu 3 Maja Škurić, University of Montenegro, Maritime Faculty, Kotor, Montenegro, majask@cg.yu 4 Emir Čikmirović, University of Montenegro, Maritime Faculty, Kotor, Montenegro, ciki@cg.yu 5 Ivan Krapović, University of Montenegro, Maritime Faculty, Kotor, Montenegro 29 Research in Transport and Logistics 2009 Proceedings Do u University 2. WORLD LEADING CONTAINER PORTS THROUGHPUT The world container port ranking changed significantly between 1996 and 2006, see Figure 1 (Si – Singapore, HK – Hong Kong, Sh – Shanghai, Bu – Busan, Ko – Kaohsiung, Sz – Shenzhen, Ro – Rotterdam, Ha – Hamburg, LA – Los Angeles, LB – Long Beach, An – Antwerp, Du – Dubai, Yo – Yokohama). The total container traffic volume of the 20 top ranking world container ports reached 205.3 million TEU in 2006. These ports increased their handling volume by 10.4 per cent compared with results in 2005, representing approximately 50 per cent of the total world container traffic. New container ports gained momentum (like Tanjung Pelapas) and other ports established their role as international load centers. 30000 Yo Du An Sz LA LB Ha Ro Sh Bu Ko Si HK 25000 20000 15000 10000 5000 0 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005 2006 FIGURE 1 FIGURE 2 The world container port ranking: 1996 – 2006 [2]-[4] Major Asian Container Ports in 2006 [1]-[4] 2.1. Far East Ports Thanks to the development of global economy, especially the rapid expansion of Asian economy including China, the container volume is steadily increasing year after year. Every port makes efforts in its own way in order to make contributions to the national economy by way of creating more value-added. For this study, major Far East container ports have been included in this survey in an effort to secure validity and objectivity. The target ports and container terminals surveyed are as illustrated in the Table 1. In Figure 2 are presented 16 Far East container ports with container traffic development from 2002 to 2006. All of them belong to 20 Top World Ports in 2006. Some improvements increase production incrementally, by 5-10%, and other improvements make a quantum jump, by 10-20%. This paper deals merely with the increasing of port terminal productivity, as a part of logistic network, due to automation of some ports subsystems like as leading Asian ports. TABLE 1 Selected Far East container ports and terminals Ports Terminals Terminal operators Remarks Singapore All together (Tanjong Pagar; Keppel; Brani & Pasir Panjang) PSA Web site Kwai Chung-Tsing Yi Different 1; 2; 5 & 9 South MTL Web site Hong 4; 6; 7 & 9 North HIT (Containerisation Kong 8 East COSCO-HIT International) 3 DP World Shanghai Pudong Int’l CT HPH-COSCO Web site Shanghai Shanghai CT HPH-COSCO (Containerisation (SPICT + SCT + Shanghai Midong CT) Different International) PECT Shinsundae Container Terminal DONGBU Logistics Div. of Dongbu Corporation Direct survey Busan Hutchison Busan CT HPH Gamman Different (PECT + Dongbu + HBCT + Gamman) Different Our study reveals that over the past six years from 2000 to 2006, the throughput per berth of each container terminal in Far East ports are steadily increasing year after year, ranging from minimum 300,000 TEU per berth to maximum 700,000 TEU per berth, and showing an average throughput per berth of around 500,000 TEU. This study also shows that during the past six years from 2000 to 2006, the average throughput in TEU per meter of berth length of the container terminals in Far East ports is steadily increase every year 30 Throughput in 000 - TEU Research in Transport and Logistics 2009 Proceedings Do u University from minimum 1,300 TEU to maximum 2,100 TEU per berth meter, and showing an average throughput in TEU per m of berth length of around 1,700 TEU. 2.2. European ports In an effort to find out the throughput of European container ports, this study has made a survey of the throughput of all the European container ports from 2001 to 2006, updating the existing data of each item by way of first and second data checkup and spot survey. The European ports selected for analysis are: Rotterdam (ECT and APM), Hamburg, Antwerp, Bremen, Gioia Tauro, Algeciras, Felixstowe, Valencia, Le Havre and Barcelona, see Figure 3 [8]. The comparison between these ports is made by total terminal basis, and these ports are presented in Figure 3. Figure 3 presents 11 European Major Ports with development functions of their throughput. The information collected covers the throughput (in TEU) for each port, berth characteristics, terminal performances and so on. FIGURE 3 FIGURE 4 Top 11 European container ports with container traffic Mediterranean container ports in 2006 [2], [3] and [8] development: 2001-2006 [8] In the two relationship analyses – correlation between throughput performance and terminal land occupied, and correlation between berth throughput and terminal land utilization, we can find out the suggestion related to the optimal throughput calculation. That is, the following 4 factors should be taken into consideration: a) the traits of container traffic (in TEU), b) the traits of berth facilities (the length of berth, number of QC (quay crane) and productivity), c) the traits of container yard - CY (size, number of stack, and storage period) and d) customer service level. 2.3. Mediterranean ports Container throughput growth for South Europe ports including all Mediterranean and Adriatic containerports had felled from 11.9% in 2000 to 6.1% in 2001 and remained close to this level in 2003. Thereafter throughputs increased for about 9.0% in 2004 and the same trend is expected in 2005. In 2001, the Mediterranean region generated 20.9 million TEU, while in 2010 this will have risen to about 40.0 million TEU [3] and [4]. If the Mediterranean ports were not competitive, then it would be reasonable to assume that productivity levels - as measured in terms of facility utilization - would be poor and that there would be little evidence to suggest that they were improving. It is also relevant to contrast the current level and development of terminal productivity. However, Figure 4 summarizes the position of leading Mediterranean ports and their throughput in 2006. 3. COMPARISONS OF PRODUCTIVITY BETWEEN SELECTED REGIONS It is relevant to consider the current level and development of terminal productivity with the situation in broadly comparable regions in the world. That implies the comparison between Far East, European and Mediterranean ports. Considering this, local conditions make direct comparisons with other port markets. 31 Research in Transport and Logistics 2009 Proceedings Do u University Figure 5 presents the position of major ports in Europe, also the high volume ports in Far East and the overall average noted in selected Asian ports as a whole and partially. Port operation efficiency is examined from the viewpoint of the TEU per berth meter and TEU per hectare that is measured by ports productivity. These performances are shown in Figure 5. The average performances of the port operation efficiency are examined in Figure 5 with parameters for MECP (Major European Container Ports) and MACP (Major Asian Container Ports). On the other hand, Figure 6 shows the relationship between the TEU per berth meter and hectare per berth of major European and Asian ports. These results present the comparison between land utilization and the berth throughput (as illustrated in Figures 5 and 6). FIGURE 5 FIGURE 6 The TEU per hectare and TEU per berth meter of Major The TEU per berth and ha per berth of Major European and European and Far East Ports [4] Far East Ports [4] FIGURE 7 FIGURE 8 The TEU per hectare and TEU per berth meter of the FE, The TEU per berth and ha per berth of the FE, NWE and NWE and CIG-E ports/terminals CIG-E ports/terminals FIGURE 9 FIGURE 10 The TEU per QC and Average QCs per berth of the FE, The TEU per sq m and Total terminal area in sq m per NWE and CIG-E ports/terminals berth meter of the FE, NWE and CIG-E ports/terminals Legend: In Figures 5 and 6, above mentioned terminals PECT, DONGBU, HBCT make part of Busan port. In Figures 7 – 10 are shown next terminal operators: HHLA-Hamburg Hafen und Lagerhausgesellschaft; EUROGATE-Eurogate Container Terminal Hamburg; APM-Terminals in Rotterdam; HPH-Terminal Operator in Busan; HNN-Hesse-Noord Natie in Antwerp; SPICT, SCT,-Shanghai; COSCO-HIT-Hong Kong; DONGBU, HPH-Busan. Figures 7 – 10 shows a summary of the terminal operation parameters of major terminals in Rotterdam, Hamburg, Antwerp, Singapore, Hong Kong, Shanghai, Busan, including Contship Italia Group - Eurogate. Container Terminal Operators own terminals with operation efficiency that is examined from the viewpoint of the TEU per berth meter, TEU per hectare or square meter, TEU per QC and TEU per berth that are measured 32 Research in Transport and Logistics 2009 Proceedings Do u University by ports productivity. The average performances of all their terminals are also present in Figure 7 – 10. The results suggest the differences between the TEU per berth, TEU per berth meter, TEU per hectare or square meter, TEU per QC and average QCs per berth. In addition, these Figures present average value expression for all terminals and whole ports. The position for quay and land utilization is further detailed in Figures 7 and 9. In all these Figures 7 – 10, MECTO and MACTO present the average values for the specified terminals and ports. All of them are measured under berth basis, with 300 meters of quay length. In the past few years we have seen a process of concentration in ownership of container shipping lines and have also observed the development of relatively long-lasting consortia between some of the major shipping lines. When these trends are considered in connection with the steady increase in ship sizes that has been recorded, it is apparent that the size of stevedoring contracts has increased sharply. The market is also forecast to expand at growth rates of between 5.4 - 10 per cent per annum up to 2010 and then between 4 - 6 per cent in the following period. Within this total, the deep-sea and transshipment sectors will expand at a considerably more rapid pace. This means that not only will significant initial capacity have to be provided but, also, a port must be able to offer capacity to meet rapidly expanding requirements for large customers. The service level provided by a port is a function of numerous factors - ship lines in port, container dwell time, handling systems and port efficiency, etc. It is far from clear that the insistence of multiple terminals in a port would have any positive effects on these issues. Clearly, a fragmented container port would more likely, result in additional port stay costs, higher intra-terminal transit traffic, costs from consolidating full barge and rail loads, etc. This would have the effect of decreasing the competitive position of the port. 4. COMPARISONS OF PRODUCTIVITY BETWEEN SELECTED PORTS For selected Container Terminals from Port of Hamburg, Port of Busan, Port of Hong Kong and Contship Italia Group, operation efficiency is examined from the viewpoint of the TEU per QC, TEU per berth, TEU per hectare and TEU per berth meter that are measured by ports productivity. The performances of these Container Terminals are shown in Figures 11 – 14. The results reveal significant differences between the TEU per berth meter, TEU per hectare and TEU per berth. The position for quay and land utilization is further detailed in Figures 12 – 14. The average performances of these terminals and ports are also presented in Figures 11 – 14. The Port of Hamburg achieved a throughput of 6.1 million TEU in 2003, 7.003 in 2004, 8.088 in 2005 and 8.862 in 2006, an increase of 14.2%, 13.0%, 13.5% and 8.8% over the previous year, respectively. In the same time the Port of Busan reached 12.075 million TEU or an average increase of 6.22% per year while the Port of Hong Kong has over 23 million TEU in 2006. Here we consider all terminals in the Port of Hamburg for 2005 and selected terminals from Table 1 for Busan and Hong Kong in 2006. Eurogate is Europe´s leading container-terminal and logistics group. Furthermore, jointly with Contship Italia, it operates sea terminals in the Mediterranean region. Six container terminals in Gioia Tauro, La Spezia, Livorno, Ravenna, Salerno and Cagliari, plus the intermodal network of the transport company Sogemar, are combined under the umbrella of Contship Italia S.p.A. of Milan (Italy). The largest terminal of them is Gioia Tauro, which handled 2.873 million TEUs in 2006 and 3.3 million TEU in 2007 as the leading transshipment centre in the Mediterranean. In the same time, the Port of La Spezia achieved a throughput of 0.99 million TEU while the Port of Livorno reached 0.4 million TEU in 2006. The Port of Salerno and Port of Cagliari had 0.24 and 0.65 million TEU in 2006, respectively. Contship Italia Group has developed large transhipment hubs as well as flexible regional gateway ports. 250000 1200000 200000 1000000 800000 150000 600000 100000 400000 50000 Hong Kong 200000 Hong Kong Busan 0 BusanHamburg 0 L Hamburg T IT Mediterranean portsM L H IT I T IT Mediterranean ports T / / -H P / M H ITC BU O D K / C / H CT U O- H PI / P E / B D K ON G S nO ma / B PE/ NG / SC n a / H TA D / C m H A O C O m / B /C / T a L/ TG /C / D / m H T a /H GT T B C / T H C L /C H T G CT B / G C / C TB /T C / / H C T S TB /T S/ C S S/ C L L FIGURE 11 FIGURE 12 TEU per QC of selected ports TEU per berth of selected ports 33 TEU/QC TEU/berth Research in Transport and Logistics 2009 Proceedings Do u University 80000 4000 70000 3500 60000 3000 50000 2500 40000 2000 30000 1500 20000 1000 10000 Hong Kong 500 Hong Kong Busan Busan 0 0Hamburg Hamburg L T T Mediterranean ports TL Mediterranean ports M I IT IT / / H -H PI / M / H ITH I CT BU CO T - P E / D K C BU O D K P G n / H PE G C / / H/ N A O CO S a B / N nO O S a m / B CT / D / T am m / A C / L/ H /C T H / D / am L/ H T TH T T G G C B C / G G C CT B /T C/ / H T C/ / H C TB S/ TC TB T S/ S/ C /CS L L FIGURE 13 FIGURE 14 TEU per ha of selected ports TEU per m of berth length of selected ports 5. CONCLUSIONS After analyzing the collected data, the results show great differences between selected ports and terminals. Far East berth and terminal productivity is significantly higher than in major European ports. In addition, containerization is well established in comparative regions. Average utilization of European ports and Contship Italia Group are lower than Far East ports as a whole. On the other hand, productivity is high and increasing in major northern European container terminals. Furthermore, relationship between container handling and land performance and current throughput of container terminals may be used to find the optimal capacity performance. In summary, productivity is high and increasing in major European container terminals. This is a manifestation of the highly competitive nature of the business, with standards of operation forced upwards by the requirements of the shipping line customers. It cannot be said that productivity in the region is a manifestation of any lack of competitive pressures. China with a high throughput performance presented through the huge improvements in ship operations over the past ten years came as a surprise, as did the performance achieved by an emerging new port such as Shanghai. Compared with major Far East ports, Chinese ports are relatively showing a steep growth and high throughput performances. This rapid increase of container volume is stemmed from the economic expansion and growing markets of China. From the perspective of resources per container terminal of China, each terminal has: above 3.5 units of QC (quay crane) per berth with around 300 m of quay length, 5.5 YT yard truck per QC, and 4 units of RTGC per QC. In addition, short storage period and cheap labor costs are also important factors to her high handling performance. But the possibility of productivity reduction coming from low service level due to the rapid container volume increase and the conspicuous obsolescence of facilities should be taken into consideration. REFERENCES [1] Containerisation International, viewed 15 December 2007; http://www.ci-online.co.uk [2] Yearbook of Containerization International, 2004 and 2005. [3] Dragović, B. and Ryoo, D.K., 2007, “Container ship and port development: A review of state-of-the-art”, Proceedings of the Ninth International Conference on Fast Sea Transportation, FAST 2007, Shanghai, China: 31-39. [4] Dragović, B., Park, N.K., Radmilović, Z., 2008, “Container port capacity performance – Case study: Major European and Asian ports”, Proceedings of Annual Conference – The International Association of Maritime Economists ’08, Dalian, China: 1-16. [5] Le-Griffin, H.D., Murphy, M., 2006, “Container terminal productivity: Experiences at the ports of Los Angeles and Long Beach”, NUF Conference: 1-21, viewed 7 November 2007; http://www.metrans.org/nuf/documents/Le-Murphy.pdf [6] MARAD., 1986, “Improving productivity in us marines container terminals (IPUSMCT)”, National Academy Press, Washington D.C. [7] MARAD., 1998, “Improving Productivity in US Marine Container Terminal”, viewed 10 October 2007. http://www.marad.dot.gov [8] Park, N.K., Seo, C.G., Choi, H.R., Dragović, B., Radmilović, Z., Chen, T., Lee, Y.C., Lim, S.Y., Cho, K.S., Lee, J.Y., 2006, “A study on improvement of calculation system of optimal throughput per berth in Korean Container Terminal”, Final Report for MOMAF, Korea. [9] Talley, W.K., 1988, “Optimum throughput and performance evaluation of marine terminals”, Maritime Policy and Management, 15(4), 327-331. [10] Talley, W.K., 2006, “An economic theory of the port”, Port Economics, Research in Transportation Economics, Volume 16, 43–65. 34 TEU/ha TEU/berth meter Chapter 2 Regional Issues in Logistics and Transportation Research in Transport and Logistics 2009 Proceedings Do u University LOGISTICS SERVICE PROVIDERS IN TURKEY: A PANEL DATA ANALYSIS 1 2, Berrin 3, 4 Abstract The companies competing in today's business environment are forced to re-engineer their supply chain management in order to meet the increasing needs of the customers. Today’s trend for industrial firms is to have a variable cost system by receiving logistics services through outsourcing and focusing on core competencies. This study aims to analyze the logistics service providers sector in Turkey comprehensively and to reveal the sector profile clearly by comparing the data collected in year 2001 and 2007. For this purpose, initially, an empirical research study was carried out to assess the profile of companies operating as logistics service providers and the logistics services already being purchased by real sector and the nature of sector-specific services. The companies to participate in the research were selected so as to form a homogeneous distribution with respect to their turnover, number of employees and geographical locations, therefore aiming to achieve a complete portrait of the Turkish Logistics Providers Sector as a result of the research. The field study involves face-to-face interviews with 71 companies for the year 2001, and 101 companies for the year 2007. The results indicate that although the number of 3rd party logistics providers (3PLs) increase in total, when the number of different sectors they are providing services is analyzed, it is found that especially for the top served sectors, the number of logistics service providers is significantly decreased. This can be interpreted as the 3PL companies are focusing on a limited number of different sectors to provide services. Keywords 3rd party logistics providers, outsourcing, logistics activities, survey, Turkey INTRODUCTION Organizations have been increasingly turning to outsourcing in an attempt to enhance their competitiveness, increase profitability and refocus on their core business. In the academic and practitioner literature, emphasis has shifted from outsourcing parts, components, and hardware subsystems towards the even greater unexploited potentials that intellectual systems offer. The motivations for outsourcing in any industry are driven by an ever-greater organizational pursuit to ensure cost discipline, whilst improving quality of service and delivery capability. However, as the outsourcing has become a popular mechanism for differentiation by contracting out the non-core activities, the differences in the motivations for outsourcing have emerged. This has been ignited by the debate as to what is core and what is non-core function. Outside vendors are regarded as specialists who can provide similar or better level of service at a lower cost than available in-house. However, through outsourcing, firms can also generate various non-financial benefits such as responding to environmental uncertainty in ways that do not increase costs associated with internal bureaucracy. Moreover, they can also focus on building their core competencies, while outsourcing the noncore activities to specialist vendors for both one-off and continual improvements. This is because firms are reported to have limitations as to the depth of specialist knowledge possessed by the suppliers [1]. Fierce competition in today’s global markets, the introduction of products with short life cycles and the heightened expectation of customers have forced manufacturing enterprises to invest in and to focus attention on their logistics systems. This, together with improvements in communications and transportation technologies, has resulted in continuous evolution of the management of logistics systems [2]. The new century has shifted the importance of organizational functions and today’s trend for industrial firms is to outsource those products and activities, which are not the company’s core business. The 1 Acibadem, Istanbul, Turkey, aktasem@itu.edu.tr 2 Turkey, fulengin@dogus.edu.tr 3 Turkey, bagaran@dogus.edu.tr 4 Turkey, sonsel@dogus.edu.tr 37 Research in Transport and Logistics 2009 Proceedings Do u University international outsourcing has been referred to as one of the drivers that have made the world “flat” and with increase in international outsourcing, the sourcing debate has moved from what and how to outsource to what and where to outsource [3]. The importance of outsourcing varies among sectors. Outsourcing has grown by 52% for the period 1993–2003 for medium–high-tech sectors while the increase for low-tech sectors is much lower, only 19% [4]. Bendor-Samuel (1998) asserts that outsourcing provides certain power that is not available within an organization’s internal departments [5]. This power can have many dimensions: economies of scale, process expertise, access to capital, access to expensive technology, etc. Another possible benefit is that outsourcing provides companies with greater capacity for flexibility, especially in the purchase of rapidly developing new technologies, fashion goods, or the myriad components of complex systems [6], [7]. Likewise, by outsourcing logistics activities, firms can save on capital investments, and thus reduce financial risks. Investment on logistics assets, such as physical distribution centers or information networks, usually needs large and lump sum costs, which involves financial risks. Furthermore, the 3PL provider can spread the risks by outsourcing to sub-contractors. As the world becomes more global and the boundaries between countries and cultures disappear, many developing countries, including also Turkey, are turning into attractive centers for international firms because of the geographical locations, low working fees, and high potential for market extensions. However a previous study shows that, in Turkey, outsourcing is still solely based on transportation [8]. As can be seen from this research many Turkish firms understand logistics services as taking the transportation order from the manufacturer and delivering the goods to destination points, without thinking about the warehouse design, the best location of the warehouse or inventory management. Such way of thinking concerns only one side of the subject and reduces the logistics services to a narrow transportation perspective. This study aims to determine the current situation of outsourcing logistics activities in Turkey, which has a great potential for logistics activities among the surrounding continents because of its geographical location. An empirical research study was carried out to determine the types of logistics activities that are most frequently provided by the 3rd party logistics firms and to reveal the changes in the conjuncture if there are any. A questionnaire was prepared to examine the current situation as well as the future plans of Turkish 3PL firms in terms of logistics activities. Results indicate that most of the firms provide services for more than one industry; apparel, automotive and chemistry industries being the most frequently served. Another perspective of the study highlights the changes in the sector between 2001 and 2007. THE FRAMEWORK OF THE STUDY AND RESEARCH METHODOLOGY This research presented here reveals the results of a subgroup belonging to a large logistics sector survey. The survey includes the four main groups of players operating in the logistics sector. These groups are: Logistics Service Providers, Logistics Service Customers, Logistics Equipment and Hardware Providers, and Information Systems Providers (Figure 1). This study focuses on the first subgroup, logistics service providers survey. A field study involving face-to-face interviews with the companies operating in the logistics sector as service provider was performed for the research. In the field study, face-to-face interviews were preferred, rather than sending questionnaires by mail. The main reasons for this are the low rates of return for studies performed via mail, the lack of possibility to correct misunderstandings and the loss of the opportunity to obtain information that can only be achieved during an interview. 1 - Logistics 2 - Logistics service service providers customers survey survey 3 - Equipment and 4 - Logistics hardware providers information system survey providers survey FIGURE 1 Turkey logistics sector survey 38 Research in Transport and Logistics 2009 Proceedings Do u University The companies to participate in the research were selected so as to form a homogeneous distribution with respect to their turnover, number of employees and geographical locations. The questions in the survey can be grouped under two groups, namely, profile related questions and logistics related questions. Profile related questions include industry in which the firm operates, duration of operation, number of employees, existence of a foreign partnership, and sales turnover. Logistics related questions include the cities that the 3PLs have offices and distribution centers, the services they provide, the sectors that the 3PLs provide special services (i.e. rack transportation for apparel sector), warehouse information, and perception about size of the logistics sector. RESEARCH FINDINGS The questionnaire has sections on company profile, logistics services, warehouses, and number of employees (which can be analyzed under company profile). When two different respondent sets from 2001 and 2007 are compared with respect to how long they have been in business, it can be said that there had not been a significant change (p=0.975) in the composition (Table 1). Approximately 70% of the firms have survived in the market at least 8 years or more. TABLE 1 Operating year, status and capital structure, range of employees comparisons Questions Categories 2001 2007 0-1 years 1.4% 2.0% 1-2 years 8.5% 4.0% Operating year 2-4 years 4.2% 9.9% 4-8 years 15.5% 14.9% More than 8 years 70.4% 69.3% International company 38% 48% Firm status Local company 55% 48% Partner with an international company 7% 5% Single partner 23% 18% Capital structure Multiple partner 72% 80% Public company 6% 2% 1-25 14% 28% 26-50 17% 22% Range of 51-100 14% 17% employee numbers 101-250 15% 15% 250-500 17% 6% More than 501 17% 13% When the firm status is analyzed it has been found that 55% of the participant companies operating in the Turkish logistics sector have local status, 38% of the participants are international companies and 7% are companies with international partnerships in the year 2001. The fact that the re-engineering process of the Turkish logistics sector has started recently shows that Turkey is an attractive market for foreign companies. The ratio of international companies introduced to the Turkish market via partnerships with a local company, or fully independently, has reached 48% in a short time (see Table 1). Although the ratio of international companies increased to 48% in 2007 the partnership status of the participants indicates no significant difference in the percentage of firms that are international / local / partner between 2001 and 2007 (p=0.209). neither the increase in percentage of international companies from 38% to 48% nor the decrease of local companies from 55% to 48% is found statistically significant. Similarly, the change in the percentage of companies that are partner with an international company is insignificant. 39 Research in Transport and Logistics 2009 Proceedings Do u University Another finding of the research on capital structures of the companies interviewed is that 72% in 2001 and 80% in 2007 have multiple partners. The ratio of single-partner companies is 23% in 2001 and 18% in 2007, whereas the ratio of public logistics companies is 6% in 2001 and 2 % in 2007. When the capital structure of the firms is compared regarding two survey years, there is no statistically significant (p=0.881) change in the percentage of firms that are single partner, multiple partner or public company. The majority of the firms are multiple partners (see Table 1), which can be a consequence of high first investment costs of warehouses, distribution centers, and cargo fleet. When the range of employee numbers of the firms are concerned regarding the scale given in Table 1, there has been a significant change in the range of employees of the firms from 2001 to 2007. The percentage of firms which have less than 100 employees is increased from 48% in 2001 to 66% in 2007 (p=0.016). It can be concluded that the firms prefer to work with less employees in 2007. While the percentage of firms with 251-500 employees is 18% in 2001, it is found that only 6% of the firms in 2007 belong to this range (p=0.014). Meanwhile, the percentages of small scaled firms (with 1-25 and 26-50 employees) have also increased from 2001 to 2007. However, the only statistically significant change is observed for firms with 251-500 employees. The most frequently provided services in 2001 are international land transportation Truck Load (TL) and Less than Truck Load (LTL), domestic land transportation (TL), warehouse, and distribution to customer warehouse. While no changes is revealed in the rate of provided services in 2007, the Project Transportation where transportation is designed according to customer’s needs gains more weight and replaces warehouse in the rank of occurrence (Table 2). TABLE 2 The most frequently provided services Service 2001 2007 International land transportation (TL) 86% 83% International land transportation (LTL) 77% 67% Domestic land transportation (TL) 72% 66% Warehouse 72% 49% Distribution to customer warehouse 70% 53% Project Transportation 65% 53% A further analysis has also been conducted to reveal the changes and as well as their direction in the services provided by the 3PLs. Table 3 shows that the provided services changed significantly from 2001 to 2007. A decrease in all these services is observed which can be interpreted as the 3PLs are now more focusing on providing core services that they are good at rather than providing numerous services to various sectors. TABLE 3 Significant changes in the services provided Service provided 2001 2007 Significance Distribution to customer warehouse 70% 53% p=0.025 Domestic land transportation (LTL) 70% 44% p=0.000 Ship transportation 58% 40% p=0.019 Air transportation 69% 34% p=0.000 Distribution center 52% 28% p=0.001 Cross docking 51% 26% p=0.001 Reverse logistics 46% 25% p=0.003 Bonded warehouse 65% 38% p=0.000 Warehouse 72% 49% p=0.002 Palletization 59% 33% p=0.025 Shrinking 56% 30% p=0.000 Labeling 56% 33% p=0.002 Packaging 55% 30% p=0.001 Quality control 37% 19% p=0.009 Full export-import operations 58% 36% p=0.004 Customs clearing 59% 42% p=0.023 Operational reporting 62% 37% p=0.001 The services that have not changed significantly are; distribution to final consumption location, international land transportation (LTL and TL), domestic land transportation (TL), project transportation, 40 Research in Transport and Logistics 2009 Proceedings Do u University container transportation, railroad transportation, intermodal transportation, light assembly/disassembly, vendor managed inventory, collaborative forecasting and collaborative planning, and e-procurement. Sectors that 3PLs provide specific services, and the respective specific services: Considering the sectors on which companies providing sector-specific services focus, it is obvious that the apparel sector predominates. The apparel sector is followed by automotive, food retail, chemistry, and medicine / health sectors, respectively. When we look at the top five sectors that the logistics service providers are providing special services (such as rack transportation for apparel industry of frigorific transportation for food industry), no change has been observed from 2001 to 2007. The only sector facing a considerable change in terms of sector specific services is apparel, in which the number of 3PLs providing special services to this sector is significantly decreased. Parallel to this, the rack transportation services have also decreased significantly from 32% in 2001 to 15% in 2007 (p=0.006). Sectors that the 3PLs are providing regular services: In 2001, apparel, automotive, chemicals, machinery, computers/electronics sectors are indicated to be the sectors offered services by the great majority of the participants. In 2007, we see that computers/electronics sector is not anymore in the top five list being replaced by the construction materials sector. This result is not surprising considering the construction boom observed in those years between 2001 and 2007. The ranking of the sectors has also changed in 2007 as machinery, construction materials, apparel, automotive and chemicals respectively. When the data is analyzed to find out whether the change in the number of 3PLs is significant or not, it has been found that there is a substantial decrease in apparel, automotive, chemistry and computers/electronics sectors. Although there has been an increase in the percentage of 3PLs for constructing materials sector from 62% in 2001 to 70% in 2007, this increase is not statistically significant (see Table 4). For computers/electronics sector, a statistically significant (p=0.003) decrease is found for the percentage of 3PLs providing services to this sector (from 69% in 2001 to 46% in 2007). Similarly apparel and automotive sector has witnessed a significant decrease (see Table 4 for significance values). TABLE 4 Top sectors served and significant changes Top sectors for 2001 and 2007 2001 2007 Significance Apparel 83.10% 63.35% p=0.010 Automotive 80.28% 60.40% p=0.005 Chemicals 76.06% 56.44% p=0.008 Machinery 71.83% 71.29% p=0.938* Computers/electronics 69.01% 46.53% p=0.003 Construction materials 61.97% 70.30% p=0.256* In order to calculate the size of the logistics sector and its growth rate during recent years, the participants were asked about their sales turnovers, the rates of change of turnover relative to previous year and turnover expectations for the next year. The reluctance for providing sales turnover information has somewhat decreased from 2001 to 2007; i.e. 66% of the firms reported their sales turnover in 2001, while 88% of the firms report their sales turnover information in 2007. Once the outliers have been discarded, the average of sales turnover for 2001 is 27,665 YTL, and the average of sales turnover for 2007 is 36,125 YTL. However, this difference is not found statistically significant (p=0.510). Another question directed to the participants was about their estimates on the size of current Turkish logistics market. The estimates of the participants have risen from 2-4 billion in 2001 to 12-14 billion USD in 2007. The firms are also compared with respect to their strategic behavior. In this section they are asked about whether their vision and mission is determined, their strategy is reviewed regularly, and their strategic goals are documented. In 2007, an additional question on whether all employees are informed about the mission, vision and the strategy of the company or not is asked as well. The answers to these questions are given on a 1-5 scale, depicted in Table 5. TABLE 5 Strategy related answer options Answer Point Not implemented 1 Planning to implement 2 In preparing stage 3 Partly implemented 4 Fully implemented 5 41 Research in Transport and Logistics 2009 Proceedings Do u University When the question on whether vision and mission are determined is analyzed, it is seen that in 2001, the firms mentioned that their vision and mission are partially implemented (average: 4.26), while in 2007, they answered as they are at the preparation stage (average: 3.87). There has been found statistically significant difference (p=0.043) between these answers. In fact, since the logistics sector market is enlarged in 2007 when compared to 2001, many new players which are at an earlier stage of strategic planning studies entered the market. When the answers to “strategy is reviewed regularly” question is analyzed, it has been found that there is no statistically significant difference between firms that are at the partly implementation stage between 2001 and 2007 (average: 4.26 in 2001 and 4.00 in 2007 respectively, p=0.162). Similarly there has not been found a statistically significant difference for documentation of strategic goals, where firms are both at partly implementation stage (average: 3.91 in 2001, and 3.49 in 2007, p=0.063). CONCLUSION AND FURTHER SUGGESTIONS The changing nature of work reflects a major shift in the way work has traditionally been done. To remain competitive and to ensure continued survival amidst such ‘hypercompetitive environment’ firms are attempting to devise new strategies. Research has found that under such circumstances firms disintegrate their business functions and increase outsourcing Error! Reference source not found.. The use of outsourcing as a strategic device has been structured on the idea that certain functions such as data handling, customer relations management and information processing are common activities among different industries and thus can be decoupled from their respective value chains. Consequently, firms can focus on their core competencies to develop superior capabilities in order to outcompete other firms in the same industries while externalizing the decoupled or disintegrated functions. In the literature outsourcing has been identified as one of the most important components of ‘flexible’ firms that can respond quickly to unanticipated threats and opportunities of the market. There are abundant examples in the computer and apparel industries, where industry leaders such as Microsoft, Dell Computers, and Reebok have established the advantages of outsourcing peripheral functions while gaining flexibility and speed through their flatter organizational forms. The competition in the logistics sector is increasing and causing 3PLs to provide a limited number of services. Similarly, the number of different cities that a single 3PL has offices as well as distribution centers is decreasing dramatically which is also an indicator of more focusing on regional markets rather than providing services for the whole country as well as the European Union. However, since the logistics market is growing in size there are more players in the market, and hence, they have not completed their strategic planning issues yet (or they are at early stages of the strategic planning on the average). REFERENCES [1] Burdon S., Bhalla A., 2005. Lessons from the Untold Success Story: Outsourcing Engineering and Facilities Management, European Management Journal, 23(5), 576–582. [2] Bramel J., Simchi-Levi D., 1997. The logic of logistics, Springer. [3] Kedia B.L., Lahiri S., 2007. International outsourcing of services: A partnership model, Journal of International Management 13, 22–37. [4] Cadarso M.A., Gomez N., Lopez L.A., Tobarra M.A., 2008. The EU enlargement and the impact of outsourcing on industrial employment in Spain, Structural Change and Economic Dynamics 19, 95–108. [5] Bendor-Samuel, P., 1998. The brave new world of outsourcing, http://www.outsourcing- journal.com/issues/may1998/html/everest.html. [6] Carlson, B., 1989. Flexibility and theory of the organization, International Journal of Industrial Organization, 7 (1), 189-203. [7] Harrison, B.T., 1994. Lean and mean: the changing landscape of corporate power in the age of flexibility, Basic Books, New York. [8] Ulengin F., Ulengin B., 2003. Impact of Internet on supply chain activities: the case of Turkey, The International Logistics Congress 2003 30 June, 1 July 2003. 42 Research in Transport and Logistics 2009 Proceedings Do u University MILESTONES IN THE PROCESS OF SURVEY PREPARATION FOR THE LOGISTICS SECTOR: CASE STUDY FOR ISTANBUL, TURKEY Evren POSACI1 2 Abstract This paper presents the stages, from beginning to the end, of the process of survey preparation in order to collect related data for the analyses of the logistics sector in Istanbul, Turkey as a case study. For such a study, we first determined the study scope which had two aspects: One was the determination and classification of the actors in the logistics sector. The other was the determination of the survey scope which included the basic information to be collected regarding the actors in the logistics sector and the design of survey questionnaires for targeted recipients. Later, survey methodology was determined which included a) establishment surveys which are face-to-face interviews with representatives of the actors of the logistics sector, b) vehicle classification counts for 3-day-24-hour exiting and entering at the gates the establishments, c) truck driver surveys at the same vehicle classification locations as well as at outbound external stations (4 highway segments and 2 ports’ gates), and truck driver diaries for freight/goods distribution companies. In addition, trip diaries are distributed to truck drivers of some of the actors of the sector. Sampling framework and survey instruments were designed considering almost all parties of the logistics sector. The system operations among these actors were determined by investigations through many interviews, meetings and interactions with the sector’s members. The study scheme was designed with the input of all these information and data regarding the system operations of the logistics sector in Istanbul. By this scheme, various types of questionnaires were designed for different groups of the system’s actors. Questionnaires were distributed first to a group of actors of the sector in order to evaluate the quality and understandability of the survey questions. Keywords— logistics planning, traffic counts, truck/goods survey, urban planning, Istanbul 1. INTRODUCTION Trucks are not allowed to operate along major arterials and highways of the city of Istanbul during the peak-hours, especially along the two bridges crossing the Bosporus which divides the city into two parts, i.e., East and West parts. This application affects manufacturers, commodity transporters and goods forwarders since it reduces the daily crossing capacity through the city between the two sides. The reason behind this decision given by the city management is that the truck/freight traffic will severely affect the passenger traffic along the already congested highway network during rush hours. However, no such analysis has been made to support or deny this claim. For this reason, we think that local freight and truck movement data collection efforts can provide more representative and accurate data to support or deny such claims as well as to perform the freight/goods movement analysis and to support urban planning process for the future of the city [1]. Forecasting goods movement and truck volumes are a necessary prerequisite for the development of travel demand models to evaluate urban and regional transportation plan alternatives [2]. Freight transportation planning for urban areas is not performed as usual as passenger transportation planning since most planners and engineers working for public agencies are not trained in freight planning and related analyses [3]. The only comprehensive source of information on urban goods movement published in recent years belongs to a book by Ogden [4]. 1.1. Reasoning of the Study There has been a growing interest in the consideration of freight/goods movements in transportation planning processes because of the critical role played by freight transportation in 1 -212-245-9900, ext. 1382, Fax: +90-212-245-9891, evrenposaci@yahoo.com. 2 . Tel: +90-262-605- 1642, Fax: +90-262-653-8495, dakin@gyte.edu.tr. 43 Research in Transport and Logistics 2009 Proceedings Do u University regional economic growth and development as well as the sectors’ capacity for fulfilling the regional employment needs [2], [3]. During the last two decades, freight demand models have been developed and documented [3], [5]-[8]. No credible data source regarding commodity and goods movements have been available for the city of Istanbul, Turkey until recently. As well in abroad there have been very few truck surveys that collect trip data for internal freight/goods movements [2]. Both due of the lack of inventorial database and the lack of attention to the subject, logistics and goods movements have not been thoroughly studied from the perspective of urban or city planning. It is a pity that though a logistics system greatly affects urban transportation and land use systems, which are the basics elements of urban systems, the subject has not been studied and paid attention to at all, rather than the definition of zonal land uses with respect to warehouses, ship-dockyards, container lots, and so on. Without having such data regarding logistics and goods movements, urban transport decisions cannot be applied successfully. In order to include the analyses of logistics and goods movements in urban planning and also in order to do logistics planning, data with respect to goods and truck movements need to be available. But, this has not been the case for the city of Istanbul until recently. No data with respect to logistics and freight planning have been available for Istanbul until recently. For this reason, a comprehensive study was planned and initiated for the first time in Istanbul, Turkey, in 2007 in order to analyze all logistics and freight movements within and through the city. By this study, an x-ray of major logistics movements has been taken and an introduction to logistics planning has been realized using these data in the Istanbul’s 2023 Strategic Regional Plan with 1/100.000 and 1/25.000 scales. These data with proper updates can also be used for future planning studies. Those logistics analyses did not find their own places in urban plans in the past will hopefully be the basic inputs to all planning studies from now on. 1.2. Purpose of the Study The purpose of this study is to document the stages, from beginning to the end, of the process of survey preparation in order to collect related data for the analyses of the logistics sector in Istanbul, Turkey as a case study. By this effort, a comprehensive inventorial database with respect to goods and logistics movements in Istanbul by time of day, day of week, type of goods and truck, goods size/weight, and the land use that generates logistics movements was aimed to create. 1.3. Methodology Types of data collection methodologies were utilized in the study named as follows: 1. Logistics sectors’ establishment surveys, 2. Vehicle classification and passenger counts, 3. Roadside intercept surveys for all vehicles at major external stations (highway and freeway locations) of the study area, 4. Survey of truck drivers intercepted at the logistics establishments, and 5. Travel diary surveys for truck drivers and trucking company surveys. Before determining activities and capacities regarding goods movements of the members of the logistics sector, questionnaires were prepared and distributed for pre-testing to some sector members. After this pre-test, forms were completed by the randomly selected sector members by conducting a face-to-face interview by the trained interviewers. Vehicle classification and passenger counts (included pedestrians and bicyclist) were done by trained individuals at the gates of a sample of establishments during the same time period for 24 hours. Though 3- or 7-day counts were planned at the beginning of the study, budget constraints confined us to be fine with one-full day. To overcome this limitation and extend one-day observation to a several- day traffic flow profile, we developed a new-count sheet to record “stated traffic flows” based upon partial observations for the counted and the other two uncounted days. Based upon the existence of a meaningful correlation between the actual and the stated traffic flows, the stated traffic flows for the other two uncounted days were factorized to replicate actual traffic counts. Roadside intercept surveys and truck driver surveys were carefully done by trained persons. Roadside intercept surveys were done from Wednesday noon until Friday midnight at six external 44 Research in Transport and Logistics 2009 Proceedings Do u University stations on exiting lanes though truck driver surveys were at the gates of a sample of establishments for 24 hours during the time of vehicle counts. Lastly, truck driver surveys were done by trained individuals counting private cars, buses, motorcycles and cargo vehicles including minivans, small and large trucks and trailers with or without goods in the vehicle entering or exiting at the same gates where traffic surveys were done at the same time period for 24 hours. Cargo vehicles were sampled by their arrivals; i.e., the first vehicle stopped at every 10 minutes during a 30-min. period and the driver was asked for vehicle’s odometer, plate number, origin or destination for the load related to the surveyed land use, departure time from his origin and approximate arrival time to his destination, and the type and size of the goods. 1.4. Scope This study is designed to be quite comprehensive since its purpose is to create a detailed inventorial database with respect to all goods movements and the land use generating these movements within the city of Istanbul (size=5712 km2). Within this context, it is extremely important to identify the actors who are creating these goods movements; mainly, manufacturers, freight forwarders, shippers, storages, wholesalers and retailers. These actors are displayed in a scheme of the system as seen in Figure 1. There are many elements constituting the urban logistics system; actually, all activities can be considered as a part of the logistics system. However, if we want the draw a flow chart of this system at a basic level, at the top of the system manufacturers are located on organized industrial and agricultural campuses or spread all over the city. Later, goods were transported to transfer centers, such as storages of shippers, warehouses or wholesalers from where they were manufactured. Following that, goods and products were transported to transfer centers such as ports and harbors, customs, cargo terminals. Final destinations are the markets where they reach to customers at retailers. It is highly important to correctly determine the major elements of the logistics system, which we call the actors of the study. These actors include organized industrial campuses, light industrial areas, boat construction shores, port and harbors, shippers, wholesale markets, transport depots, warehouses, container lots, customs, bonded warehouses, distributers, and retailers. 2. STUDY DESIGN The study design used in this paper is described in the following sections. 2.1. Analysis of the Logistics System: Actors of the System In this section, the issue is to correctly describe the relationships among the actors, and between them and the market. This is extremely important since by this relationship it is possible to lead the study to the right direction and make the descriptions of the sector’s actors and the distinctions among them properly. The identification how the logistics system created by the system’s actors is operating in Istanbul was made possible by a comprehensive study. This identification was achieved by proper correspondences with the major actors of the system. By these correspondences, we were able to learn how the system operates and what the actors do in detail. Thus, main components of the goods movements in Istanbul were determined and study scope was defined definitely as possible. By the definition of the system’s operation, it was possible to stratify the actors by business and by their positions taken in the system. Then, actors were placed into groups by these definitions. By these stratifications, response rates to survey questions were observed at very high percentages. 2.2. Sampling Frame: Determination of Populations and Samples Another important issue is to define populations for each actor groups as correct as possible. By this way we can know the rate of representation of the sampling group in Istanbul. In this case, we used various databases created for the city of Istanbul by various departments of the metropolitan municipality of Istanbul such as housing, land use, mapping, transportation, recreational parks, city and planning departments as well as the departments of the central government such as the Ministry of Education. These databases include the data regarding housing, business and service sectors, industrial sector, and parks and cemeteries. 45 Research in Transport and Logistics 2009 Proceedings Do u University Spatial distribution of the system’s actors in Istanbul is considered by designing questions. Many actors have branches or warehouses in several parts of the city. Braches belonging to same actors are included in the database as either separate entities or one entity. For example, an enterprise X has a management office at one location, at a different location X has a factory that has production lines, and at different locations it has four warehouses. This X is included in the database as either one entity or six entities (one management office, one factory and four warehouses). Also, the functions of these entities are different from each other. The office building has managerial activities, the factory has production or manufacturing activities and the warehouses have storage activities. It is very possible to have the X in three different lists or in just one or two lists. It is again very possible to have the X unlisted at all but was encountered at field screenings. All these issues were handled properly while preparing population lists. For this reason, we designed the questionnaires to ask if the selected entities have other branches, warehouses or managerial offices at other parts of the city or the country. By considering the possibilities of compound or individual entities, for example some factories have warehouses and some others do not. The issue is whether these compound entities or others do business significantly different regarding the movements of goods and logistics. If there are differences between such different units, then all branches of a single entity must be surveyed separately. However, this increases the time and personnel expenditures. Another approach is to increase the sample size in order to lessen the effects of such discrepancies within the sample. To keep it in mind, by questionnaires information regarding the capacity, type and size of the business is collected only for the visited locations. In the case of the entity of X, if we visited only the factory, the questionnaire is filled out for the factory, not for the ware houses or the managerial office. If we preferred to receive information for all entities belonging to the X, for some cases the questionnaire is filled out only for the visited locations and some other cases for all entities, and in this case we could not know which information belongs to which unit. To help to clarify such cases, we developed and used the third type of questionnaire at the managerial locations. By this questionnaire, information regarding capacity, type and size of the business is collected for the visited location as well as for one entity for the similar entities, such as at one of the four warehouses and the information for the visited warehouse is extended to the rest three by using size and capacity information related to the business. 2.3. Survey Design Considering various types of actors, two approaches were considered for the design of the survey questionnaires. First approach was to prepare a long and comprehensive questionnaire book and have the actors fill out the same questionnaire from beginning to end. In this case, the sections and questions that are not related to some actors can be left blank or checked with a choice of “not applicable.” However, such unanswered questions with lack of information can be mixed with others that are really “not applicable.” If every questions has a last choice of “not applicable,” this can make the response time of interviewees longer, which can limit the possibility of the survey’s application. An advantage of this type of survey applications was its easiness for the analysis of answers by having a single type of data coding system. If we do not want all interviewees read all questions, we might have more directions to advise skipping some questions for some type of interviewees. In this case, interviewers might observe a difficulty to apply the forms and a strict training must be given to interviewers frequently during the course of the study. In the second approach, the system’s actors are stratified and grouped by their businesses and by their positions taken within the systems. For each groups, different survey forms are designed to ease the above difficulties mentioned in the first approach. By this approach, the forms are easily completed because unrelated questions are not included in the forms, and by having fewer questions, responses of the interviewees are taken under high enthusiasms. In this study, the second approach was utilized in order to make this stage successful. 2.3.1. Grouping the System’s Actors In the analysis of the current logistics system, we have observed that some actors get together at certain locations by their decisions, if necessary, with the changes in planning decisions by the planning authority. These actors do their business by their own management policy. This is just the case in the transport depots and shippers in Topkapi. This is one campus and holds 140 firms acting as 46 Research in Transport and Logistics 2009 Proceedings Do u University freight forwarders to in and out of Istanbul. These firms were shown this location as an urban planning decision and future planning decisions will be enforced to all of them and information regarding all these individual firms can be received by their management offices. Thus, such land uses are considered to be managed by “organized management offices.” The management of organized industrial campuses is one of such examples to these organized management offices. For such campuses, we have developed two types of survey questionnaires: 1) the first one is for individual firms and units within the organized campuses, and 2) the second is for their management offices. By the first questionnaire, we have collected information regarding the business and capacity of individual firms and units. The second from targeted the information regarding the whole campus and the businesses and capacities of their members as a whole. Lastly, there is the third type of the sector’s actor, to which the first type of survey cannot be applied. For those, we have developed a third type of questionnaire form in order to collect information regarding the business and capacity of them. These actors are not manufacturers, wholesalers, or retailers but organizational institutions of them, customs offices, freight forwarders or goods distributors within the city. 2.3.2. Who Will Answer the Surveys? After defining and groping of the actors, the issue is who answer the survey and what their positions will be. The questionnaires were answered by general managers, vice managers and logistics managers. If these people are not present at the time of visit, another appointment is taken. However, at some cases these positions are taken by less educated people, and in one case during pretests a representative of a warehouse did not know whether he was working at a bonded warehouse. Thus, explanations are written in an informal language as possible and academic or formal language is not used when the opposite is possible. 2.3.3. Design of the Forms Design of the surveys forms included grouping the same type of questions and creating a logical flow from beginning to end. For example, questions seeking information regarding the identification are placed at the beginning and some others seeking financial information are placed at the end of the forms as is usually known that financial information might not be disclosed to other competitors. Thus, easy questions came first and the difficult or sensitive ones came later in order not to let the interview discontinue earlier. Flow charts of the questionnaires helped a lot to decide how to use the information sought by each questions and how to relate them each other, if necessary. By this, some questions sought detailed information and some other did not. Also, each question is evaluated from the perspectives of the benefits that will be gained if asked, such as more information to make detailed analyses; and of the losses that will be lost if not asked, such as longer times and more personnel to complete the surveys, and tiring interviewees and rejecting to continue. Especially these losses might harm the dependability of the surveys. Thus, gains and losses by each question are discussed by professionals before and during the surveys based upon the answers received. The flow chart of the questionnaires included the following groups of questions: 1. General information regarding the entity, a. Name, address, phone and fax number, email and website. b. Date it was established at the site visited c. Field of business that is involved in i. Major field ii. Minor fields 2. Size of land use a. Types and size of buildings and land uses b. Capacity use for each building and land use 3. Number and size of freight/goods movements a. By month of year, day of week, and time of day b. Type and size of vehicle, and freight/goods type 4. Number and size of vehicle movements a. By month of year, day of week, and time of day 47 Research in Transport and Logistics 2009 Proceedings Do u University b. Type and size of vehicle, and freight/goods type 5. Capacity related to the sector a. Quality of the business b. Satisfaction of customers 6. Current employment and future personnel needs a. By education, profession and experience 7. Financial Situation during the last year a. Income and expenditures b. Benefit/Cost ratio 2.3.4. Traffic and Truck Surveys After the logistics sector (establishment) surveys were completed, vehicle classification and truck surveys were conducted. Traffic and truck surveys were conducted over a smaller sample than the establishment surveys were. The sample size for traffic and truck surveys is about 10 percent of the sector surveys. 2.4. Difficulties Observed During the Study Some difficulties were observed during the study because the logistics sector is very huge and extensive. Though the efficiency of the sector is not truly within the scope of this study, some questions to disclose the time of duration of trucks for loading and/or unloading within each land use are designed and this design is failed, for example, at wholesale markets because some trucks spent more than 24 hours, even 2-3 days at such land uses. These cases greatly increased standard deviations of the asked variables. However, by this question though it is aimed to disclose the time spent for the operations of loading and unloading and to investigate reasons of such inefficient operations, such unusual cases distorted the averages of the asked variables. By analyzing the reasons of inefficient operations, the logistics system will be commented on in more details and necessary precautions will be taken for the optimization of the system. Though possible answers for the question asked are expected to be a couple of hours, the answers to it received at the wholesale markets are 2-3 days. 2.5. Solutions for Successful Designs First of all, type of data needed for the study must be determined. Based on the type of data required, volume of data must be decided. Secondly, time, personnel and cost must be projected for such type and volume of data. Types of equipment and human resources required to perform data collection dictates the total cost of the study. Sometimes, automated data collection methods such as loop or radar detectors hinder details of trip and vehicle classification data. For this reason, manual data collection methods might be chosen. If the study is conducted by the regional planning agency, such as the metropolitan municipality’s transport department as in this study, other local as well as central government agencies cooperate easily to make the study successful as desired. For the success of the surveys, level of detail must not exceed beyond the scope of the study. This increases the cost and makes the interviews real burden on the shoulder of interviewees and they may give up answering the questionnaires as long and complex questions keep coming 3. CONCLUSIONS AND RECOMMENDATIONS FOR FUTURE RESEARCH This paper summarized the purpose, scope, methodology, and stages of the study of logistics and goods movements as well as the things that need to be careful about while completing such a study. As a conclusion, such a comprehensive study that included analyses for urban planning is the pioneer of its kind as well as the comprehensive as possible. Since it is the first of its kind, possibly it includes some possible faults and errors in its measurements and analyses. Some of them can be avoidable by future updates. Some errors come from the methods utilized, which are, for example, due to human errors in traffic counts and some other are from the nature of questions taking place in the questionnaires. However, by the study an inventorial database of the logistics in Istanbul will be documented and the analysis of the data will be used for the purpose of urban planning. Even, regional 48 Research in Transport and Logistics 2009 Proceedings Do u University logistics plans are considered to be conducted. This study will be a good example for the same studies in the future. As a recommendation for future research, we strongly suggest electronic vehicle counting by the analyses of video images. In this study, we used the video images and visually extracted data by trained personnel in order to control manual traffic counting. Secondly, we opted out recording plate number of the vehicles during counting since it increased the time and human resources incredibly, well beyond existing resources of the study. However, via video imaging methods, which are currently under development by a company of the Istanbul’s municipality, it is possible to track vehicles entering the boundaries of the land uses that are the subject of the study but not directly related to this location can be excluded from the analysis. Also, by tracking the vehicles which are related to these locations studied, it can be possible to determine how long they occupied the parking locations within the land uses. ACKNOWLEDGEMENT Authors acknowledge Istanbul Metropolitan Municipality, General Directorate of Transportation, Department of Transportation Planning by creating the opportunity for the authors to work on the project of Trip Generation Handbook and Logistics Study and let them present their experiences at the International Conference on Prospects for Research in Transport and Logistics on a Regional-Global Perspective, Feb., 12-14, 2009, Istanbul, Turkey. In addition, several specialists and academicians contributed greatly to the development of survey questionnaires named in this paper. Authors would like to acknowledge their contributions and cite their names here: Dr. Aykut Toros (professor of emeritus and project leader of the Research group in IMP), Dr. Gungor Evren (professor of emeritus of transportation at ITU), Ihsan Hadi Karadeniz (Vice Manager, Department of Transportation Planning, IBB), Ahmet Kocaslan (chief researcher of the Research group in IMP). REFERENCES [1] Beagan, D., Fischer, M. and Kuppam, A., 2007, “Quick Response Freight Manual II”, FHWA- HOP-08-010, U.S. Department of Transportation, Federal Highway Administration, Cambridge Systematics, Inc. MA 02140. [2] Kuppam, A., Livshits, V., Vallabhaneni, L., Zmud, M., Wilke, J., Elmore-Yalch, R., and Fischer, M., 2008, “An Approach for Collecting Internal Truck Travel Data: Lessons Learned from Maricopa Association of Government’s Internal Truck Travel Study”, TRB 2008 Annual Meeting, Washington, DC. USA. [3] Chatterjee, A., 2004, “Freight Transportation Planning for Urban Areas”, ITE Journal, December 2004, 20-24. [4] Ogden, K.W., 1992, “Urban Goods Movement: A Guide to Policy and Planning”, Ashgate Publishing Company. [5] Pendyala, R.M. and Shankar, V.N., 2000, “Freight Demand Modeling: Synthesis of Research and Development of Multi-Level Conceptual Frameworks”, 9th IATBR Conferenc, Sdney, Australia. [6] Fischer, M.J and Han, M., 2001, “Truck Trip Generation Data”, NCHRP Synthesis 298, Transportation Research Board, Washington, DC. [7] Cambridge Systematics, 2003, “Methods for Forecasting Statewide Freight Movements and Related Performance Measures”, NCHRP Project 8-43, Transportation Research Board, Washington, DC. [8] Cambridge Systematics, 2004, “Accounting for Commercial Vehicles in Urban Transportation Models”, U.S. Department of Transportation, Federal Highway Administration, February 2004. 49 Chapter 3 Education and Training in Logistics and Transportation Research in Transport and Logistics 2009 Proceedings Do u University EDUCATION IN TRANSPORT AND LOGISTICS IN AN AGE OF GLOBAL ECONOMY. 1 Abstract - It appears that any country which ignores research and, therefore the competent human resources to carry out it, is doomed to be the underdog in the coming age of the 21st Century. Apart from other sectors of the economy, the logistics and transport sectors are to be scrutinised as specific ping part of the world, on a comparative basis. The aim of this paper is to discuss this. Keywords Education - Knowledge Capital - Knowledge economy Globalisation - Transport System 1. WORLD ECONOMY AT THE THRESHOLD OF NEW MILLENNIUM Since last quarter of 20th Century the world economy is passing through a transformation process to be portrayed with a strong stress on technological development on the one hand and serious environmental problems at unprecedented levels on the other. These two elements, technology development and the cost of development in general, environmental deterioration, surpass all the other factors in designing development policies today. These twin developments take place simultaneously within a context of the so-called process of Globalization is an ongoing trend. Although the benefits of globalization for developing countries are sometimes doubted, globalization is still seen as a prerequisite for further development. It can be interpreted as the growing economic interdependence of countries worldwide through the increasing volume of cross-border transactions in goods and services and variety in traded goods and services, through larger international capital flows and also through the more rapid and widespread diffusion of technology. It affects trade patterns, capital flows and location choices of firms at a regional and global level. 2 In an age of intensified global movement, accelerated technological development and challenges of intrinsic environmental troubles (global warming, climate change and the changing demographic structure in the world as well as the widening gap between the richest and the poorest), the designing and implementation of appropriate (sustainable) development policies gain a high level of priority in Capitalism is undergoing an epochal transformation from a mass-production system where the principal source of value was physical labour to a new era of innovation-mediated production where the principal component of value creation, productivity and economic growth is knowledge and intellectual capabilities. Capitalism in this new age of innovation-mediated production will require deep and fundamental changes in the organization of enterprise, regions, nations and international economic and political institutions. Survival in this new era will require the development of new organizational forms and systems, such as teams and new incentive systems, which decentralize decision making, mobilize intellectual capabilities, and harness the knowledge and intelligence of all members of the organization 3 As the very movement of globalisation itself has happened before, today we face a new stage in this process. The last wave is distinctive in that 1 2 ARJAN LEJOUR, Globalization and the Global Environment: Four Quantitative Scenarios. CPB Netherlands Bureau for Economic Policy Analysis. 10 February 2003 3 R. FLORIDA & M. KENNEY, The New Age of Capitalism. Innovation-mediated production. Futures. July-August 1993. p.637. 53 Research in Transport and Logistics 2009 Proceedings Do u University (a) First, and most spectacularly, a large group of developing countries broke into global markets. (b) Second, other developing countries became increasingly marginalised in the world economy and suffered declining incomes and rising poverty. (c) Third, international migration and capital movements .... have again become substantial 4 As an indication of the effect(s) of globalisation process we can perhaps look at the change in world trade on a comparative basis. Between 1950 and 2004 world trade has grown 22-fold, from $ US 375 billion to $ US 8 164 billion.5 However, a global financial crisis at unprecedented dimensions gaining momentum makes the analysis of the world economy a controversial one. As it is really difficult to forecast the future of the globalisation clearly, the very presence of the globalisation movement with its existing domain is a fait accompli which we cannot / should not overlook. 2. PROSPECTS FOR THE FUTURE: ECONOMIC TRANSFORMATION Whatever the developments into the future may take direction we can safely predict that the importance and effectiveness of one principal instrument of economic development, alongside with others, will keep its legacy: the technological competence of the countries. Economic development results from an assessment of the development objectives with the available resources, core competencies, and the infusion of greater productivity, technology and innovation, as well as improvement in human capital, resources, and access to large markets. Here, with an ever increasing impetus, technology production and technology development play a distinctive role in the process. Economic development transforms a traditional dual-system society into a productive framework where markets can function for a betterment of the whole sections of the society. This picture turns into a more promising one as we transfer it from a national / regional level into a global one. In this process research and knowledge / knowledge economy are the keywords to solve the puzzle. If we refer to the goals set by the European Union to face the challenge(s), knowledge, innovation and education turn out to be the tools of EU policies. the most dynamic and competitive knowledge-based economy in the world capable of sustainable economic growth with more and better jobs and greater social cohesion, and respect for the environment by 2010 Supporting knowledge and innovation been put in the first place by the EU Commission among the (eight) key measures with a high European value-added. In parallel with the Barcelona European Council which reviewed progress towards the Lisbon goal in 2002, the European Research Area, ERA, has contributed to better and To reach the Barcelona objective, research investment in Europe should grow at an average rate of 8% every year, shared between a 6% growth rate for public expenditure and a 9% yearly growth rate for private investment. This is ambitious yet realistic given the strong support given to the objective 6 3. KNOWLEDGE ECONOMY AND THE HIGHER EDUCATION SYSTEM At the dawn of the new millennium, the roles attributed to the sciences seem to be changing hands. There is in biology at the moment a sense of barely contained expectations reminiscent of the physical sciences at the beginning of the 20th century. It is a feeling of advancing into the unknown and that where this advance will lead is both exciting and mysterious. 7 In a sense, it will not be wrong to claim that there will not be a room in the new map of the world in the near future for the nations who cannot produce and develop technology. This will lead us to the conceptual boundaries of a knowledge society / economy. 4 The World Bank, Globalization, Growth, and Poverty. A WB Policy Research Report. Washington, DC 2002. 5 WTO, International Trade Statistics, 2005 6 Investing in research: an action plan for Europe. COMMUNICATION FROM THE COMMISSION. Brussels, xxx COM(2003) yyy final 7 The Economist, June 16th 2007. 54 Research in Transport and Logistics 2009 Proceedings Do u University The Knowledge Economy is emerging from two defining forces: the rise in knowledge intensity of economic activities, and the increasing globalisation of economic affairs. The rise in knowledge intensity is being driven by the combined forces of the information technology revolution and the increasing pace of technological change. Globalisation is being driven by national and international deregulation, and by the IT-related communications revolution. However, it is important to note that the term the overall economic structure that is emerging, not to any one, or combination of these phenomena 8 Referring again to EU policies, the official documents of this giant underline the importance of In the long run, however, the economic performance of countries is also strongly determined by knowledge-related factors (e.g. technical change and human capital). In particular, R&D and technological innovation have contributed substantially to the strong US economic performance over recent years. More generally, the contribution of knowledge investments and activities to employment, productivity and economic growth has been emphasised in many studies .9 The figure below taken from the same study depicts the interaction between different institutions in a country that individually and jointly contribute to the production, diffusion and utilisation of knowledge. In this system, science, technology/innovation and industry are central but not sufficient to ensure economic growth, competitiveness and job creation. The education and training system, human resources and the labour market, and the financial system all have a substantial impact on the -Technology- Further in this sphere of interest, another Commission Report10 The creation of a Europe of knowledge has been a prime objective for the European Union since the Lisbon European Council of a source of opportunity, but also of major challenges. Indeed universities go about their business in an increasingly globalised environment which is constantly changing and is characterised by increasing competition to attract and retain outstanding talent, and by the emergence of new requirements for 8 John Houghton and Peter Sheehan, A Primer on the Knowledge Economy. Centre for Strategic Economic Studies Victoria University 9 European Commission, Key Figures 2007. EUR 22572 EN 10 Commission of the European Communities, The role of the universities in the Europe of knowledge. Brussels, 05.02.2003 COM(2003) 58 final 55 Research in Transport and Logistics 2009 Proceedings Do u University which they have to cater. ... To implement the Lisbon agenda, the European Union has embarked upon a series of actions and initiatives in the areas of research and education. One example is the European area of research and innovation, to achieve which fresh perspectives have just been opened up3 and, in this context, the objective to increase the European research and development drive to 3% of the Union's GDP by 2010. All these documentation serve to put forward that education in general and the higher education in particular is a sine qua non for instituting a knowledge economy for our age. 4. THE ROLE OF TRANSPORT AND LOGISTICS NETWORKS IN THE NEW ECONOMY As global movement and global competition dominates the agenda of the world economy, the need for an efficient transport and logistics design and management is a growing theme for the foreseeable strategic decision-making come to the forefront as a major theme in transport an Strategic decision-making in this sense represents a complex set of trade-offs of key challenges and associated (perceived) benefits and costs to the business. Active market and institutional forces changing at a fast pace present challenges to the fields of transport and logistics management 11 The current growth trends in the transport sector have to be radically changed and reversed. For this to happen, there needs to be a strong commitment to the objectives of sustainability at all levels of transport policy-making. The transport sector has to go through a substantial policy shift if it is to comply with the objectives of sustainable development. Such a policy shift has taken place in many parts of the world, with sustainable transport becoming an integral part and an indispensable objective of most national and supranational policy documents. There is broad agreement now that a more balanced transport system needs to be put in place, which requires a significant modal shift and inter-modality on major transport corridors. This requires not only a good level of policy knowledge but also advanced management skills in order to encourage an interconnected, intermodal, interoperable transport system, which can both help reduce petrol- dependency, emissions, congestion and bottlenecks, and at the same time ensure and strengthen competitiveness. In addition, urban transport policy is another major area where improvements are essential for better accessibility, mobility, environment, and quality of life. Public transport systems and non-motorised modes of transport are to be improved in order to encourage the use of these more environmentally friendly modes and to promote equity in transport opportunities that is to ensure accessibility for all. Such improvement schemes are increasingly being introduced together with restrictions on car usage, such as congestion charging, parking control and management, pedestrianisation, car-free areas, etc. While such transport policy measures have become widespread, there is also increasing awareness that these problems can be solved through land-use policies that can influence travel behaviour and attitudes towards car usage. In addition to macro policies for urban growth management, such as the compact city, policentricity, etc., neighbourhood design approaches to influence travel behaviour received significant emphasis, with much research on designing transit- oriented or pedestrian-friendly development and less car-dependent neighbourhoods. The distance between production and (intermediate and final) demand determines the size of trade, and in turn, the demand for transport. As globalisation affects the volume of consumption, production and the place of production, it has great impact on the volume of transport. 5. HIGHER EDUCATION IN PROMOTING THE EFFICIENT TRANSPORT & LOGISTICS NETWORKS However, the relationship between investment in knowledge and performance is complex and non- linear. What factors can explain the differences in innovative performance across countries with rather similar levels of knowledge investment? An important source of diversity between industrialised 11 D.A. Hensher & Brewer, A.M., Transport. An economics and management perspective. Oxford Univ. Press. p.2. 2004 56 Research in Transport and Logistics 2009 Proceedings Do u University economies relates to the respective roles of the main actors (i.e. firms, universities, and government and other public research institutions) in the process of knowledge production, diffusion and utilisation, as well as to the forms, quality, and intensity of their interactions. These actors are influenced by a variety of factors that exhibit some degree of country specificity: industry structure, the education and training system, the human resources and labour market, the financial system, etc. the highly skilled and flexible hu dynamic global markets a major factor in development and is critical to information and communications technology. Returning once more to the EU policies on this sphere, an important point is the so-called Bologna Process. This process refers to the signing of 29 European ministers of a Declaration (at the University of Bologna), the Bologna Declaration which aimed to undertake a European Higher Education Area. Bologna reforms are refocusing on higher education institutions, now that the legislative framework is largely in place. Universities must accept their responsibility to drive forward and urge governments to accept that the process needs time, and financial and human resources, to ensure long-term sustainability. In short, as the world enters the 21st century, the quality of education continues to be a major factor in ability to succeed and excel. In the 1950s and 1960s, education endeavours in transportation were focused at the college level, mainly through engineering programs. Development in the field now comes about because of continuing demands and commitments at several levels. First and foremost, within the educational establishment, formal academic programs have moved beyond traditional lines (e.g., engineering and science). As a result, transportation education programs are a formal area of study as well as an adjunct to or support for other established academic disciplines. As the transportation education system grows, the focus is changing in several ways. From an academic standpoint, additional policy areas as opposed to technical areas become apparent. For example, students and professors broaden their scope to examine communication between public and private interests, strategic management of human and capital resources, environmental impacts, and most recently, the impacts of computerization and technology. The paradigm in transportation policy is shifting away systems. The shift is being accomplished through the use of enhanced management systems and intelligent transportation systems. At the same time, there is a growing realization that transportation education needs to broaden its focus beyond academic offerings. To interest future leaders (including traditionally underrepresented populations) in transportation careers, elementary and secondary curricula are being developed and revised. 12 Here, one of the major factors to affect the context, philosophy and the pathway of the transportation education is the very fact of globalisation movement. 6. WHAT SHOULD BE IN THE AGENDA FOR THE FUTURE? Today, transportation issues became more complex and more dependent on the developments in other sectors of the economy than in the past. Transportation issues were part of the academic agenda but did not have their own framework. It is no longer sufficient to have a technical background and it is and will continue to be a multidisciplinary and lifelong endeavour. In the 21st century, four areas will have a crucial effect on future innovation: globalization, technology, changing demographics and curriculum development. Furthermore, we need to review our educational systems in terms of (a) the actors in education systems 12 P.A. MANNING, Transportation Education. http://onlinepubs.trb.org/onlinepubs/millennium/00125.pdf 57 Research in Transport and Logistics 2009 Proceedings Do u University (b) the identification of the policy issues to which these systems are designed and implemented. As the uncertainty for the future of the global economy draws points of interest on behalf of the policy-makers as well as the actors of the transport industry and curriculum developers, there are efforts to deal with these elements of blur. Ubbels, Rodenburg and Nijkamp for instance develop a scenario analysis to tackle this problem.13 For them One way to deal with the (uncertain) impacts of globalisation is to construct several scenarios. The construction of these future images can be a useful tool for exploring the uncertain future. In the context of GITAGE,14 the globalisation issues are identified in four scenarios. These address in various ways the process of globalisation, its effects on the various regions, political cooperation between regions, the pace of technological progress (see Nederveen et al., 1999), changing consumption patterns, and developments in the transport sector. 15 In order to be competitive in the globalising knowledge economy, the developing countries need to invest in their innovation systems at the national and level, perhaps much more so than their developed counterparts. As the globally competitive countries are turning their production towards value-added segments and knowledge-intensive products and services, there is greater dependency on access to new technologies, knowledge and skills. And, with the parallel processes of globalisation and localisation, the local availability of knowledge and skills is becoming increasingly important. In the design and development of educational models the very fact of the vitally important interactive working of transport and logistics networks should not be underestimated. This interaction is too costly to be ignored. Any educational programme should take into account the interactive format of freight transport and logistics networks. This brings us to the perspective of intermodal education. Intermodalism, in its simplest notion, involves the transport of both people and goods. Depending on the findings of a TRB Conference Workshop16, we have to draw serious lessons from Many transportation professionals are not graduates of university transportation programs logistics networks. New applications for transportation and intermodalism do not come from a textbook; they come from real-life experiences - therefore, we need to educate the educator; Private companies need to support transportation education by providing financial support, contributing teachers, and participating in courses; and More research is needed on intermodalism, and these research results need to be incorporated into education and training. 17 These facts urge us to suggest / propose that (1) A new and well-worked study has to be planned to work out the feasibility and comparable stand covering a well-defined area for search. (2) The involvement and participation of a large spectrum of higher education institutions must be secured as well as public and private sector cooperation. (3) as a preliminary to search the existing transportation and / or logistics or transportation & logistics programmes in terms of intermodalism, interdisciplinary aspects, course contents, faculty involvement and programmatic inadequacies to see the inventory of the area. (4) The collaboration of an international parental organisation should be sought. This is important to safeguard the quality and competence of the work to be done. 13 B. Ubbels & C. Rodenburg & P. Nijkamp, A Multi-layer Scenario Analysis for Sustainable International Transport. Tinbergen Institute Discussion Paper. TI 2000-102/3 14 GITAGE, Globalisation, International Transport and the Global Environment 15 ibid 16 Intermodal Transportation Education and Training, TRB Conference Proceedings # 17, 1998. 17 R.D. Krebs, Intermodal transportation education and training, Keynote address. TRB Conference Proceedings # 17, 58 Research in Transport and Logistics 2009 Proceedings Do u University THE ROLE OF EDUCATION AND TRAINING IN THE SUPPLY CHAIN SECTOR David Maunder1 Abstract The Chartered Institute of Transport [CILT] was founded in 1919 after the First World War where the logistics problems were magnified. The logistics and transport problems of the war and new modes of transport such as shipping and rail made it obvious that transport as a new subject needed to be studied, learned and developed if resources were to be used efficiently and effectively. In 1926 CIT was granted a Royal Charter to promote, encourage and coordinate the study and advancement of the science and art of transport in all its branches. In July 2001 the CILT became the Chartered Institute of Logistics and Transport [CILT]. There are currently over 30 sections and branches worldwide with over 30,000 members. From the beginning, the CILT's role has been education in the broadest sense: to spread knowledge and to be a source of authoritative views for communication to governments, industry and the community. The paper will explain the aims and objectives of CILT, its membership grades and describe its education courses which are aimed at different levels of the transport/logistics sector. The courses are designed to promote the highest standards of professional expertise and practice of those already working in the sphere of logistics and transport and lead to membership or chartered membership of the institution. The CILT holds charitable status. Keywords: CILT, education, institute, logistics, NPO, supply chain management, transport 1. INTRODUCTION The Chartered Institute of Transport [CILT] was founded in 1919. During the First World War, both motorized transport and aviation had become a reliable means of movement of both men and freight. There were thus two new modes of transport to be added to those of shipping and rail. The logistical problems of pursuing a war combined with the establishment of the new modes made it obvious that transport as a subject needed to be studied, learned and developed, if resources were to be used as efficiently and effectively as possible. On November 26th, 1926 the Chartered Institute of Transport was granted a Royal Charter [a] to promote encourage and coordinate the study and advancement of the science and art of transport in all its branches, to initiate, foster and maintain investigation and research into the best means and methods of and appliances for transport, transit and locomotion and the problems that are involved and their most satisfactory solution, to extend, increase and disseminate knowledge and exchange information and ideas in regard to all questions connected therewith and to assist and further in all practicable ways the development and improvement of transport, transit and locomotion in the best interest of the community. [b] to hold meetings of the Institute for reading and discussing communications bearing upon traffic and transport. [c] to provide for the delivery of lectures, the holding of classes and examinations the awarding of certificates, medal, scholarships and diplomas in connection with traffic and transport [d] to hold or promote exhibitions of appliances apparatus or systems of transport [e] to form a library. On July 18th 2001 the Institute of Logistics amalgamated with the Chartered Institute of Transport and the Institute was renamed the Chartered Institute of Logistics and Transport [CILT]. From the beginning, the role of the CIT and latterly CILT has been education in the broadest sense: to spread knowledge and to be a source of authoritative views for communication to governments, industry and the community. 2. AIMS AND OBJECTIVES 1 David Maunder, Chartered Institute of Logistics and Transport International, London, UK. david.maunder@ciltinternational.org 59 Research in Transport and Logistics 2009 Proceedings Do u University The efficient movement of people and goods is vital to the quality of life enjoyed by society and the effective functioning of trade, the economy and essential services. Thus, it is desirable that all those involved in the planning, operation and management of logistics and transport should be well qualified for their work. This also applies to those engaged in allied activities, such as logistics and transport education, consultancy and research. They all make a crucial contribution to the effectiveness of the entire transport sector. Throughout its history the CILT has constantly strived to achieve three objectives: To ensure an adequate supply of qualified personnel through an internationally recognized education and training programme/system To keep qualified individuals up to date throughout their professional career To play an active and influential role in shaping logistics and transport policies for the future development of the sector In seeking to achieve its three objectives, the CILT aims can be summarized as follows: To promote the study of the science and art of logistics and transport To provide an educational programme with assessment and examinations leading to professional qualifications and membership grade within CILT To encourage members’ active participation in continuing professional development To cooperate with the educational sector and fellow professional institutions with the objective of raising standards To foster investigation and research into the development and improvement of logistics and transport by every practical means To initiate regular meetings and the exchange of ideas through the medium of conferences, workshops, seminars, discussion groups, lectures, visits and other professional activities. 3. INTERNATIONAL ORGANISATION The CILT has always been an international organization, indeed over one third of the worldwide 30,000 membership are based outside of the UK. From Africa to Asia, Canada to Australia, Ireland to Hong Kong membership is growing. In recognition of the ever growing importance of its international status and activities, the CILT revised its executive structure in 1990. One of the resulting benefits is greater autonomy for and recognition of the larger local membership. There are now ten National Councils: in Australia, Hong Kong, India, Ireland, Malaysia, New Zealand, North America, Pakistan, Singapore and the UK. All administrative matters in these countries are handled locally, enabling a fast efficient and effective service to be provided for members. In addition to National Councils there are Independent Sections and Branches in at least 20 other countries in Africa and SE Asia. In these countries some administrative functions are handled locally but many are handled by CILT International. All National Councils, Sections and Branches belong to CILT International which is the ‘umbrella’ organization for CILT worldwide and holds the Royal Charter- all members pay a small annual membership fee to CILT International. There is close liaison between the local CILT organization and Ministry’s of Education and Transport, universities and colleges training providers and major employers and employer associations. In this way, local needs where appropriate are sure of being recognized and incorporated into examination arrangements. 4. MEMBERSHIP GRADES OF CILT There are four membership grades namely: Student/Affiliate Member Chartered Member Chartered Fellow All members are actively encouraged to progress through the different grades via experience of the logistics and transport industry, educational qualifications and continuing professional development [CPD]. By becoming a member of CILT members gain a lifelong professional anchor, a unique source of support, 60 Research in Transport and Logistics 2009 Proceedings Do u University knowledge and networking and an international membership qualification of a professional institute recognized throughout the world. Student/Affiliate membership A student, already working in the logistics and transport sector can join as an affiliate member with the intention of studying and taking CILT qualifications at either Certificate or Diploma level. Otherwise anyone already working in the sector can join at this level but who do not aspire or qualify for higher membership grades. Member The grade of Member is open to those who have relevant qualifications such as at CILT Certificate or Diploma level and have at least 3 years experience within the logistics and transport sector. A Member of CILT can use the post nominals MILT after his/her name. Chartered Member Chartered membership is open to those who have successfully passed the CILT Advanced Diploma course [or exempting equivalent qualifications] and have at least 5 years relevant experience in the logistics/transport sector including 3 years at senior level. In addition, Chartered Membership can be applied for via the “professional route” where a candidate who lacks educational or professional qualifications provides a complete CV identifying his/her experience to date within the sector including where applicable evidence of papers submitted at conferences and seminars and presents a paper at a local CILT Section. Chartered Members can use the post nominals CMILT after their names and this is the grade that most will aspire to during their career. Chartered Fellow This is the highest grade of membership. Fellowship is reserved for those who hold positions of high responsibility within the logistics and transport sector. Fellowship [FCILT] of the CILT is open to Chartered Members who have a minimum of at least 7 years senior management experience. In exceptional circumstances Fellowship grade is granted to individuals who have attained a position[s] of eminence in logistics and transport. 5. EDUCATION AND TRAINING The education and training processes of the CILT are independent, integrated and designed to promote the highest standards of professional expertise and practice of those working in the logistics and transport sector. The Institute now offers a three-tier professional qualification structure which deals with all major aspects of logistics and transport, following the prevailing industry best practices. The three-tier structure includes: The International Certificate in Logistics and Transport which is designed for first line supervisors and managers already working in the field of logistics and transport. Those who are successful in the Certificate course meet the knowledge and standards for membership [MILT] subject to a minimum of 3 years’ experience in the logistics and transport sector. The International Diploma in Logistics and Transport which is a higher level qualification aimed at those already working in the industry/sector at a middle management level. Those who are successful at the Diploma level course meet the knowledge standards for membership [MILT] subject to a minimum of 3 years experience in the logistics and transport sector. The International Advanced Diploma in Logistics and Transport which is an advanced course aimed at developing strategic decision making skills in logistics and transport for aspiring senior managers. Those who are successful in the Advanced Diploma course meet the knowledge standards for Chartered membership [CMILT] subject to a minimum of 5 years experience at a senior level in the logistics and transport sector. A number of universities and colleges offer degree and diploma courses in logistics and transport and many of these are recognized by the CILT as satisfying all, or part, of the educational requirements either for Membership [MILT] or Chartered membership [CMILT]. Some of these institutions also offer courses for the CILT examinations. At most of these institutions, there are Chartered Members on the academic staff who in addition to their experience in education and research have invaluable practical experience in the logistics and transport sector. Some National Councils are able to offer their own courses leading to CILT qualifications. 61 Research in Transport and Logistics 2009 Proceedings Do u University These courses may have local variations in content to meet local market requirements but all are based on key knowledge areas for membership in line with CILT International courses and qualifications. 6. THE INTERNATIONAL CERTIFICATE IN LOGISTICS AND TRANSPORT This qualification is designed for front line supervisors and managers working in the logistics and transport industry. It is intended to increase knowledge, skills and confidence as part of the managerial role for a junior manager within the industry. The course consists of 7 units or modules of which 3 are mandatory for all students to study. Students then have a choice of one unit from a selection of 4. Either at the end of each unit or when all 4 units have been studied students take examinations with the pass mark set at 50%. The 3 core subject units that all students have to take are the following: Managing the Logistics and Transport Environment Managing Resources Customer Service and Quality Management in Transport Operations Optional Units comprise the following: Warehousing and Store Operations Freight Operations Fleet Management Passenger Transport As an example of the extent of the course content the topics included for the three core subjects that have to be studied by all students is shown below: Managing Resources comprises 3 topics namely: People, finance and technology. Managing the Logistics and Transport Environment comprises: Internal structures and processes, external factors, stakeholder interests, marketing and commercial needs, legal and environmental issues. Customer Services and Quality Management comprises topics such as: Customer service and competition, customer service culture, developing, maintaining and improving service quality, setting and monitoring service standards, analysing quality problems and service failures, TQM, MIS, benchmarking, legal requirements, admin systems and communications. The elective modules are equally extensive in terms of topics covered. The syllabus is set out into units which are then broken down into topics of knowledge and competencies. Key learning points and coverage are also provided for each unit. Each unit/module takes approximately 60 hours of study so the entire course takes 250 hours of which 100 hours will be taught in the classroom. A Training Provider needs to be accredited by CILT International to run the course following a recommendation from the local CILT Section. Accreditation depends on meeting a set of criteria including facilities and staff experience to teach the courses. Upon accreditation the teaching materials are provided on a CD for the local training provider who then needs to localize/ customize the material to the local environment. The teaching material is both student and teacher friendly and is ideal for self tuition. So the course is taught locally using internationally provided teaching materials and examinations are set locally and marked locally but moderated by CILT International to ensure quality control and standards are maintained at similar levels throughout the CILT international community. The course meets the knowledge standards required for Membership of CILT subject to a minimum of 3 years experience in the industry. 7. THE INTERNATIONAL DIPLOMA IN LOGISTICS AND TRANSPORT The CILT International Diploma builds on knowledge gained from studying the Certificate course or from the working environment itself. It is set at a higher level than the Certificate and will consequently take longer – possibly 500 hours of study of which 200 will be classroom based and the rest in self tuition. The qualification is basically for those already in middle management posts who wish to develop a strategic view of logistics and transport operations and be capable of reviewing and modifying operational activities. 62 Research in Transport and Logistics 2009 Proceedings Do u University It is a modular course comprising 6 units all of which have to be studied. The units comprise: Managing Transport and Logistics Operations Resource Management Transport Economics and Finance Supply Chain and Logistics Inventory Management Passenger Transport For each unit detailed Power Point Presentations are provided as teaching material which is user friendly for lecturers and easily understandable for students. As with the Certificate course however, lecturers need to localise the presentations by incorporating additional material to cover the local economy, transport and logistics sector and environment. Topics for three of the six units are shown below as an example of the entire course. The Transport Economics and Finance unit comprises topics such as: Transport supply and demand, externalities, economics and finance, costs and budgetary measures, decision making and evaluation and statistical research methods. The unit Managing Transport and Logistics Operations includes topics such as: Organisational structure and relationships, transport characteristics, logistics and supply chain, customer service, information technology, the business environment and planning, controlling and decision making. Passenger Transport comprises: Movement of people, origins, destinations and routes, handling of passengers, modes of passenger transport and services. As with the Certificate course training providers have to be accredited by CILT International to run the course but upon accreditation the teaching materials are provided. All six units are examinable with the examination papers and marked scripts moderated by CILT International to ensure quality control and standards. Depending on the candidates experience and knowledge students do not need to study the Certificate course before starting the Diploma. However, the course is set at a higher level and so candidates need to have a good basic knowledge of logistics and transport before commencing the course. The course meets the knowledge standards required for Membership of CILT subject to a minimum of 3 years experience in the industry. 8. THE INTERNATIONAL ADVANCED DIPLOMA IN LOGISTICS AND TRANSPORT This advanced course which is the equivalent of a degree is primarily aimed at developing strategic decision making skills in logistics and transport. Modern companies in their efforts to cope with an ever changing and challenging environment use two key processes to build and sustain their future. The first is strategic planning which enables top management to determine what business direction it wants to achieve. The second is logistics and transport planning which enables the company to proceed in a systematic way to identify and turn specific opportunities into revenue streams and profitability. The course presents the conceptual idea of integration of logistics and transport planning into the strategic planning process. Students who study for the Advanced Diploma are expected: a) To become familiar with certain elementary analytical concepts useful to analyse logistics and transportation environmental trends, customers, organization, finance, and cost dynamics. b) To develop an understanding of logistics and transportation strategic planning methods including strategic management and leadership, strategic environment, and organization level planning. c) To integrate and apply the concepts discussed into their daily workplace. d) To be aware of ethical issues, sustainability and minimizing the impact of the organization and its operations on the global environment e) To develop skills in research methodology and produce a work-based project. Teaching material has not been developed at the present time though Standards or a Syllabus has been finalised. Therefore, a Training Provider who wishes to offer the course needs to create teaching material based on the Standards provided by CILT International. It is expected that students will study for at least a 2 year period before completing the course. Four core Units comprise the course which all students have to study and they comprise: 63 Research in Transport and Logistics 2009 Proceedings Do u University Strategic Management and Leadership Strategic Environment for Logistics and Transport Organisation Level Planning in Logistics and Transport Research Methodology which includes a work based research project. The course content comprises the following: Strategic Management and Leadership topics include: organisational culture, leadership, collaboration and ethical management. The Strategic Environment for Logistics and Transport comprises: society and sustainability, government and politics, economics, finance and risk and contingency planning. The topics included in Organization Level Planning in Logistics and Transport comprise: vision and strategy, transportation planning, delivering customer service, innovation and change, supply chain performance Research Methodology and Work-based Project includes topics such as: nature of business and management research, research approaches, sampling and data and a work based project. The first 3 units are examinable and moderated as for the Certificate and Diploma courses and the work based project is assessed locally but moderated by CILT International to maintain standards. Those who are successful in the Advanced Diploma course meet the knowledge standards for Chartered membership [CMILT] subject to a minimum of five years experience at a senior level in the logistics and transport sector. 9. CONCLUSIONS The Chartered Institute of Logistics and Transport is the leading professional body associated with the logistics and transport industry. With over 30,000 members in 30 countries worldwide CILT holds unparalleled professional recognition. Established in 1919 and receiving its Royal Charter in 1926, the Institute has a traditional history but is constantly adapting to stay consistent and relevant to today’s worldwide logistics and transport issues. Close links are still retained today with the British Royal household as they have throughout CILT’s history. In 1922 HRH the Prince of Wales was made the Honorary President whereas currently HRH the Princess Royal is the CILTUK Patron and Her Royal Highness Queen Elizabeth is the CILT International Patron. The principal objective of the Institute is ‘to promote and encourage the art and science of logistics and transport’. The Institute achieves this objective through both its membership and its educational qualifications. Membership provides a professional identity and international recognition to those already employed in the logistics and transport sector. The Institute’s professional qualifications educate not only those already in the sector but also those wishing to enter it. The CILT holds charitable status and so resources are invested for the benefit of its members throughout the CILT worldwide community. 64 Chapter 4 Supply Chain Management Research in Transport and Logistics 2009 Proceedings Do u University MODELING REVERSE FLOWS IN A CLOSED –LOOP SUPPLY CHAIN NETWORK Vildan ÖZKIR1, 2 and 3 Abstract This paper seeks to describe features of establishing a closed-loop supply chain for the collection of End-of-Life Products. To address this issue, the problem is handled through reverse logistics and is modeled as a capacitated facility location problem. The main purpose of the study is to describe a general closed loop supply chain including customers, collection points, reverse centers, plants and recovery facilities. In this context, we propose a mixed integer nonlinear programming model for this problem in which product recovery processes can be viewed in three steps. First step is to determine the number and location of collection points where returned products from end- customers are collected. In the second step, these products are sorted and classified in order to properly recover these items in reverse centers (disassembly, de-manufacturing or refurbishing options). Ultimately in the third step, returned products are recovered as refurbished products in recovery facilities. We developed a numerical example to illustrate the usefulness of the proposed model and obtained the solution by using GAMS software. Keywords Closed Loop Supply Chain, Product Recovery Options, End of Life Product. INTRODUCTION Supply chain management has emerged as one of the major areas for companies to gain a competitive edge. Managing supply chains effectively is a complex and challenging task, due to the current business trends of expanding product variety, short product life cycle, increasing outsourcing, globalization of businesses, and continuous advances in information technology [Lee, 2002]. A supply chain has traditionally been considered as a line, starting with the movement of goods from suppliers to manufacturers, and going ahead with wholesalers, retailers, and finally reaching to consumers through these distribution channels. Nowadays, this line is continuously getting closed, namely a circular chain occurs. Hence, complex industrial relationships prove the existence of material flows not only going downstream but also going upstream during the production, distribution, and consumptions stages. Reverse logistics has been spreading worldwide, involving all layers of supply chains in most of the industry sectors. Take back obligations, environmental concerns and customer pressure forced reverse logistics to become a key component in modern supply chain. The first known definition of reverse logistics was published by the Council of Logistics Management (CLM): ‘… the term often used to refer to the role of logistics in recycling, waste disposal, and management of hazardous materials; a broader perspective includes all activities relating to logistics carried out in source reduction, recycling, substitution, reuse of materials and disposal.’ Forward supply chain management covers all business functions and decisions related with the flow of goods starting from supplier to end customer while reverse supply chain management covers the opposite flow which oriented towards recovery processes starting with collection of returns from end customers. Forward supply chain associates with reverse supply chain by closing the material flow regarding environmental and economic concerns. The attempts for management of forward and reverse flows of goods are being researched in the area of the closed loop supply chain management. Closed loop supply chain studies in the literature especially investigate the design and recovery concepts. 1 al Engineering Department, Turkey, cvildan@yildiz.edu.tr 2 Önder Öndemir, Northeastern University, Lab for Responsible Manufacturing, Department of MIE, Boston, MA, U.S.A., ondemir.o@neu.edu 3 Turkey, basligil@yildiz.edu.tr 67 Research in Transport and Logistics 2009 Proceedings Do u University This study aims to represent a model framework for designing the product recovery process in a closed loop supply chain. This framework represents a novel example for multi-level, multi- commodity and capacitated product recovery network in the whole CLSC. This paper is organized as follows. Section 2 reviews existing research and approaches for the design strategies of a supply chain. In section 3, we introduce a closed loop supply chain (CLSC) structure and material flows in it. In section 4, we identify the main formation of proposed CLSC model. In section 5, we present the results of research and, we conclude in section 6. SUPPLY CHAIN DESIGN Supply chain design studies aims to establish an effective and efficient system for the flow of all materials in the supply chain regarding environmental concern. The success of supply chain depends on its design; the success of the chain design depends on how close the chain is. The closeness of the loop is measured by the amount of waste which influences environmental concern. So, it can be said that supply chain gets closer if and only if it generates less the waste. Hence waste disposal is another key factor for a successful chain. The unique principle when designing a supply chain is related directly to cost minimization issue. Besides this, CLSC has additional design principles emanated from the requirement of sustainability. Krikke, Pappis, Tsoulfas ve Bloemhof-Ruward (2001) summarized the additional principles - matching network design with recovery options, enhancing quality and rate of return, enhancing design for recycling and etc.- that these principles generally serves as environmental principles. The more waste the chain generates, the more the chain have to dispose of. Consequently, the design of the CLSC has to aim reducing disposal costs, improving less waste generating designs and increasing the usage of more recyclable components. Environmental issues are mostly related with recovery, reuse, remanufacture, and recycle operations. Studies regarding environmental concern mostly charged with exploring the best opportunity for closing the supply loop. On the other side, typically, supply chain design studies strive to find the best set of inventory, transportation, and production decisions. Therefore, the problems confronted during these studies are of higher complexity and have multi-dimensional characteristics. Supply chain design is a complicated project which needs an extreme effort to balance the environmental parameter(s) and profit parameter(s). Environmental parameters are smart indicators of sensitive behaviors for minimizing environmental damage by reducing the impacts of wastes and respectively waste disposals that occur in the lifecycle of a product(s). Profit parameters are cost dependent indicators involving the costs generated from production, transportation, warehousing, etc., and revenues related with price and sales. The success of supply chain is directly related with the perfection of its design. The perfection of the design can be revealed by measuring the performance of the design. Performance measurement process and key performance indicators are widely studied in the literature (Hill, 1993; Agarwal et al., 2006; Beamon, 1999; Gunasekaran, Patel and McGaughey 2004; Lai, Ngai and Cheng 2002; Kleijnen and Smits 2003). CLOSED LOOP SUPPLY CHAINS (CLSC) Closed Loop Supply Chain (CLSC) practices, recoverable product environments, and the design of these products and materials, have become an increasingly important segment of the overall push in industry towards environmentally conscious manufacturing and logistics [Zhu, Sarkis and Lai, 2008]. CLSC management comprehends all business functions and hence decisions regarding the adaptation of business strategy, marketing, quality management, information systems, logistics and so on in view of closing material flows, thereby limiting emission and residual waste, but also providing customer service at low cost [Krikke et al., 2001]. Both the forward chain and the reverse chain are considered, since there is a strong interaction between the two of them. 68 Research in Transport and Logistics 2009 Proceedings Do u University Raw Material Component Assembly Store Use Suppliers FL Producers FL Distribution FL Customers Centers Recovery Facilities RL Reverse Centers RL Collection Points RL End of Repair & Refurbish Inspect, Sort, Classify Collection Life Products Product Recovery Options Product Recovery Material Recovery Component Recovery Reusable Reusable Reusable FL: Forward Logistics Products Materials Components RL: Reverse Logistics Figure 1: Closed Loop Supply Chain Flows Forward supply chain includes all movement and storage of raw materials, work-in-process inventory, and finished goods from point-of-origin to point-of-consumption while reverse supply chain considers the reverse flow of any type of returns from point of consumption to the point of recovery process ends. Reverse supply chain investigates how to establish a physical reverse channel and information flow to support the strategy (including standardized processes and procedures for minimizing the causes of returns and facilitating the efficient physical flow of planned returns from customer to final disposition). Besides this, it also investigates how to design the best reverse logistics system and return disposition strategy that minimizes cost and liability, maximizes reverse velocity, profitability, asset value recovery, and customer loyalty. Tibben-Lembke and Rogers (2002) presented a unique study examining the similarities and differences between reverse and forward logistics. CLSC refers to an integrated framework including both forward and reverse supply chain. But, it is essential to understand in what respects closed loop supply chain design will be affected from forward supply chain design and/or reverse supply chain design. CLOSED LOOP SUPPLY CHAIN MODEL The main purpose of proposed model is to describe a general closed loop supply chain including customers, collection points, reverse centers, plants, distribution centers and recovery facilities. We formulate a multi-product, multi-echelon, multi- period and profit maximizing closed loop supply chain model covering activities from suppliers to recovery facilities. Thus, we present a mixed integer linear programming model regarding possible product recovery options in the context of closed loop supply chain. The optimization model aims to determine the optimal locations for collection points, reverse centers, plants and recovery facilities while maximizing the profit. Proposed CLSC model can be summarized as follows: Profit = Total Sales Revenue - Total Fixed Cost - Total Setup Cost - Total Purchasing Cost - Total Transportation Cost - Total Processing Cost - Total Holding Cost - Total Penalty Cost. Total Sales Revenue can be calculated by summation of recovered product Sales, new Product Sales and Recovered Material Sales. Total Fixed Cost can be calculated by summation of all fixed costs in collection centers, reverse centers, plants, recovery facilities, and distribution centers. 69 Research in Transport and Logistics 2009 Proceedings Do u University Total Setup Cost can be calculated by summation of all setup costs in collection centers, reverse centers, plants, recovery facilities, and distribution centers. Total Purchasing Cost can be calculated by summation of procured components in plants and retrieving cost in collection centers. Total Transportation Cost can be calculated by summation of all transport cost related with all closed loop supply chain echelons. Total Processing Cost can be calculated by summation of total disassembly cost, total de- manufacturing cost, total disposal cost, total production cost, total sorting and classifying costs and total refurbishing costs. Total Holding Cost occurs in reverse centers and distribution centers. Reverse centers handles products and components but distribution centers handles products in two forms: recovered products and new products. Total Penalty Cost occurs on dissatisfied customer demand situations. The proposed optimization model maximizes profit function subject to these constraints: The optimization model for collection center location decision minimizes investment (fixed and running costs of facilities as well as transportation costs) subject to following constraints: (1) Goods collected from customers must go to an open (available) collection center, (2) All goods collected from a customer must go to exactly one collection center, (3) Minimum number of collection points (CP), reverse centers (RC), plants (P), distribution centers (DC) and recovery facilities (RF). (7) Capacity constraints. (5) Balance equations for material flows. (6) All parameters and variables are non-negative. Quantity of any customer demand for each product type in a particular time period, unit sales prices, distance between arrival and destination nodes, unit transportation cost for products per unit distance and fixed costs, setup costs, unit processing costs are some of the important input parameters used in this model. The binary decision variables decide whether a facility is opened during a particular time-period and also ensure that a particular return goes to one open facility only. The optimization model decides the number of each facility type and where these facilities are located while maximizing profit of the whole closed loop supply chain. NUMERICAL APPLICATION In order to illustrate the application of the proposed model, we consider a multi-stage, multi- period, and multi-commodity supply chain design. The example is a medium size problem that given parameters are customer demands per period, unit sales price of products, component quantities and component types in a product type, unit processing costs, unit transportation costs and alternative location destinations. All the other input data are available upon request and all the instances are modeled with GAMS. We can summarize the stages in the closed loop supply chain as follows: In every period; Customer demands occur in two forms: o Demand for New Products: Brand new products have higher prices with life long guarantee. o Demand for Recovered Products: Lower prices with limited life long guarantee. Customer returns occur for two reasons: o Retrieve Price: To improve the quality and the quantity of returns, company offers a financial incentive for each return. In this case, we assume the amount of incentive offered by the company is equal for every quality level and for every type of used product returns. o Environmental Sensitiveness or Environmental Legislations. Distribution Recovered Product Demand Returned ProductsCustomers Collection Centers Centers New Product Demand Collection of Returns: Used product returns are collected from customers by collection points. The unique responsibility of collection points (CP) is to collect used products from customers 70 Research in Transport and Logistics 2009 Proceedings Do u University and to send them to reverse centers for inspection, sorting and deciding the best recovery option. Reverse Center Operations: Product returns are inspected in reverse centers. Inspected returns are sorted according to inspection results. Inspection results includes the decision of recovery option. The main purpose of reverse centers (RC) is to explore the best recovery option. In this case, we assume that reverse centers are able to disassemble (or demanufacture) a returned product. Product recovery decisions includes: o Product Recovery Option: Returned product requires only small repairing or refurbishing process. Each component of product corresponds to the requirement of the minimum life-time for related product design. In our case, to faciliate mathematical formulation, we defined Pref(t) as the ratio for product recovery option which is given for each period t. o Component Recovery Option: If any component of returned product does not work or does not correspond to the requirement of the minimum life-time, this returned product is disassembled to its components. If any component of a returned product works, this component can be used in the production of recovered product. In our case, we defined Pdis(t) as the ratio for component recovery option including disassembly process. o Material Recovery Option: If any component of a returned product does not work or does not correspond to the requirement of the minimum life-time, this component can only be recovered by demanufacturing process. Demanufacturing process includes transforming products (or components) to their materials that they are formed. We defined Pdem(t) as the ratio for material recovery option after disassembly process. We show the decision procedure in figure .... for product recovery process. Non-Recoverable Componenrts Returns Pref(t) INSPECTION Product Recovery Decision REFURBISHING Pdis(t) Recoverable Products Returned Products are sent to Recovery Facility Component Recovery Decision DISASSEMBLY Non-Recoverable Componenrts Components Pdem(t) INSPECTION Material Recovery Process DEMANUFACTURING Recoverable Components are sent to plants Recovered Materials are sold Figure 2: Product Recovery Process Recovery Facility Operations: Recovery facilities are responsible with repairing and refurbishing operations. Repaired or refurbished products are sent to distribution centers after packaging process. Distribution Center Operations: Distribution centers serves customer in two types. In our example, system works with two types of products and these two types of products may be in two forms: recovered products or new products. The supply channel of recovered products is recovery facilities while the supply channel of new products is plants. In our example, closed loop supply chain occurs four types of candidate location for any type of facility (CP, RC, RF, P, DC) and five major customer locations. Costs for setting up a facility and costs for fixed processing are given. Hence, the model decides where to locate each type of facility in each period. We restricted model by a constraint which guaranties an open facility remains open for the rest of periods. The facility locations and the quantities are decided by the model regarding total 71 Research in Transport and Logistics 2009 Proceedings Do u University transport cost, total setup cost and total fixed cost. Maximum profit is calculated as 189.800 USD and setup decisions are summarized in Table 1. Setup Decisions Plants DCs CPs RCs RFs Final Locations P3 D2, D3 J1,J2,J3,J4 R1 F2 Table 1: Set up decisions SUMMARY AND CONCLUSION Environmental legislations and concerns leave no choice to companies but to improve their responsibility on product life time. They are responsible with exploring the best opportunities for consuming fewer natural resources by evaluating more recovery options for end of use and/or end of life products. Closed-loop supply chains open up a new and interesting set of issues to be addressed by industry. They also offer opportunities for meaningful academic research that takes into consideration the interdisciplinary nature of the problems and the need to develop frameworks and models based on an integrated-business-process perspective (Guide, Harrison and Van Wassenhove, 2003). In this study, we propose a framework for CLSC model for strategic decisions related to forward supply chain and reverse supply chain operations from the view of product recovery. The developed model allows the comprehensive description and analysis of the system operations taking into account the “green image” for supply chain design. Especially, proposed framework includes a detailed product recovery process and the goal of the proposed model is maximization of profit function. The world is changing very rapidly, and companies, in cooperation with academics, must quickly develop supply chains that can handle coordinated forward and reverse flows of materials. They must develop methods of showing managers the benefits to be obtained by developing reverse supply chains. If managers cannot quantify the potential financial (and nonfinancial) benefits, they are unlikely to consider return flows as anything other than a nuisance, for which they must minimize losses(Guide, Harrison and Van Wassenhove, 2003). Overall, the model developed here can help practitioners to design better CLSCs having the ability to respond quickly to consumers need and determine the right policies to manage the operations efficiently. An organization can economically benefit from savings by effectively managing a supply chain. Thus, we proposed a MILP model for a multi-stage closed loop supply chain system which provides insights in formulating the problem. REFERENCES [1] Agarwal., A., Shankar. R., and Tiwari, M.,K., 2006, "Modeling the metrics of lean. agile and leagile supply chain: An ANP-based approach", European Journal of Operational Research, Volume 173, Issue 1, Pages 211-225. [2] Beamon, B.,M., 1999, "Measuring supply chain performance", International Journal of Operations & Production Management, Volume 19, Issue 3, Pages 275 – 292. [3] Guide, V.D.R., Harrison, T.,P., Van Wassenhove, L.,N., 2003," The Challenge of Closed-Loop Supply Chains", Interfaces, Volume 33, Issue 6, Pages 3-6. [4] Gunasekaran, A., Patel, C., McGaughey, R.,E., 2004, "A framework for supply chain performance measurement", International Journal of Production Economics, Volume 87, Issue 3, Supply Chain Management for the 21st Century Organizational Competitiveness, Pages 333-347. [5] Hill, T., 1993, "Manufacturing Strategy: The Strategic Management of the Manufacturing Function", (2nd Edition), MacMillan, London. [6] Kleijnen, J., P., C., Smits, M., T., 2003, "Performance metrics in supply chain management", Journal of the Operational Research Society, Volume 54, Issue 5, Pages 507-514. [7] Krikke, H., Pappis, C., P., Tsoulfas G., T., and Bloemhof-Ruwaard, J., 2001, Design Principles for Closed Loop Supply Chains, ERIM Report Series Reference No. ERS-2001-62-LIS. [8] Lai, K., Ngai, E., W., T., Cheng, T., C., E., 2002, "Measures for evaluating supply chain performance in transport logistics", Transportation Research Part E: Logistics and Transportation Review, Volume 38, Issue 6, Pages 439-456. [9] Lee. H.,L., 2002, "Aligning Supply Chain Strategies with Product Uncertainties", California Management Review Volume 44, Issue 3, Pages 105-119. [10] Tibben-Lembke, R.,S., and Rogers, D.,S., 2002, "Differences between forward and reverse logistics", Supply Chain Management: An International Journal, Volume 7, Issue 5, Pages 271-282. [11] Zhu, Q., Sarkis, J., Lai, K., 2008, "Green supply chain management implications for closing the loop", Transportation Research Part E: Logistics and Transportation Review, Volume 44, Issue 1, Pages 1-18. 72 Research in Transport and Logistics 2009 Proceedings Do u University STRATEGIC ANALYSIS OF GREEN SUPPLY CHAIN MANAGEMENT PRACTICES IN TURKISH AUTOMOTIVE INDUSTRY Gülçin BÜYÜKÖZKAN1 and ÇAPAN 2 Abstract Environmentally responsible manufacturing, green supply chain management (GSCM), and related principles have become an important strategy for companies to achieve profit and gain market share by lowering their environmental impacts and increasing their efficiency. This paper examines the main components and elements of GSCM and how they serve as a foundation for an evaluation framework in Turkish Automotive Industry. The identified components are integrated into a strategic assessment and evaluation tool using analytical network process (ANP). The dynamic characteristics and complexity of the GSCM analysis environment make the ANP technique a suitable tool for this work. The supplied case studies provide additional insights for research and practical applications. Keywords Analytic network process, green supply chain management, strategic decision making, Turkish automotive industry. INTRODUCTION As competition has intensified and globalized over the last decade, supply chain management has received greater attention by manufacturing organizations. Firms increasingly rely on their supply network to handle more complex technologies and higher customer expectations. Among these expectations, increasing attention is devoted to suppliers' social responsibility with a particular focus on fair and legal use of natural resources. Green supply chain management (GSCM) has emerged as an important new broad-based innovation that helps organizations develop ‘win–win’ strategies that seek to achieve profit and market share objectives by lowering their environmental risks and impacts, while increasing their ecological efficiency [1]. Environmental impacts occur at all stages of a product's life cycle, from resource extraction, to manufacturing, use and reuse, final recycling, or disposal. Interdisciplinary research has integrated the efforts of management, engineering, physical and social sciences to investigate the issues relevant to this topic. Similarly, multifunctional groups within organizations and external stakeholders have a role in decisions related to organizations and the natural environment. When organizational environmental decisions are to be made they will necessarily be strategic and usually more complex for this reason. These decisions will have internal and external implications for the management of an organization. Green supply chain decisions are one of the latest issues facing organizations with strong internal and external linkages [2]. One approach to model the dynamic nature of business and its relationship to the natural environment into a decision framework is a technique that is capable of considering the multidimensional qualitative and strategic characteristics. This paper is one that addresses this issue. Our aim is to identify and structure the primary strategic and operational elements for a framework that will aid managers in evaluating green supply chain alternatives. The structure that is developed in this paper is a network hierarchy that can be used to evaluate the alternatives. The technique for analyzing the decision is based on the analytical network process (ANP) approach first introduced by Saaty [3]. The dynamic characteristics and complexity of our decision environment, which is true for most strategic decisions, makes the ANP technique a suitable tool. Ford Otosan, Hyundai Assan and Toyota are selected as case companies in this study for the evaluation of GSCM alternatives. The supplied case studies provide additional insights for research and practical applications. The organization of the paper is then as follows. The paper begins with a presentation of the suggested approach. Then, we attempt to structure various main components of the GSCM into an evaluation framework. After a brief review of ANP, the next section includes the illustration of the methodology through the cases of three automotive companies. As a conclusion, the paper gives some concluding remarks. 1 Gülçin Büyüközkan, Galatasaray University, Faculty of Engineering and Technology, Industrial Engineering Department, Ortaköy, Istanbul, Turkiye, gulcin.buyukozkan@gmail.com, gbuyukozkan2003@yahoo.com. 2 Institute of Sciences, Industrial Engineering PhD. Program, Ortaköy, Istanbul, Turkiye, alisancapan@gmail.com. 73 Research in Transport and Logistics 2009 Proceedings Do u University AN EVALUATION FRAMEWORK FOR GSCM GSCM is gaining increasing interest among researchers and practitioners of operations and supply chain management. Three drivers (economic, regulatory and consumer pressure) drive GSCM worldwide. It integrates sound environmental management choices with the decision-making process for the conversion of resources into usable products. GSCM has its roots in ‘environmental management orientation of supply chains’. It has emerged as an effective management tool and philosophy for proactive and leading manufacturing organizations. The scope of GSCM practices implementation ranges from green purchasing to integrated life-cycle management supply chains flowing from supplier, through to manufacturer, customer, and closing the loop with reverse logistics. A number of definitions of GSCM exist [4], [5]. The research in GSCM addresses a variety of issues ranging from organizational research and practice in GSCM [6-9] to prescriptive models for evaluation of GSCM practices and technology [2], [9-11]. This study proposed an analytic approach based on the ANP methodology to assist in GSCM strategic decisions. The following two main phases are realized. GSCM Conceptualization Phase: • Determine and classify the factors that influence the success of GSCM with a literature review. • Through discussion with industrial experts, if there exists, add different analysis dimensions or views in evaluation framework. • Determine possible interactions and/or relationships among factors and criteria; and develop an evaluation model. GSCM Analysis Phase: • Calculate the relative importance weights of the factors by means of pair wise comparisons supplied by industrial experts of the case company. • Form a supermatrix (i.e. a two dimensional matrix composed from the relative-importance-weight vectors) and normalize it so that the numbers in every column sum up to one. • Calculate the converged (“stable”) weights from the normalized supermatrix. • Calculate the final score of each green supply chain alternative for the case companies. GSCM CONCEPTUALIZATION PHASE A detailed literature search with the concepts related to “green supply chain” is realized. We can find some concepts and elements which can be served as the foundation for a decision framework for prioritizing or selecting systems by the organization that will aid in managing green supply chains [2], [4], [5], [9], [12- 19]. They are summarized as follows. New Product Development Phases : New product development (NPD) is the complete process of bringing a new product or service to market. Green new product development (GNPD) is defined here as product development into which environmental issues are explicitly integrated in order to create one of the least environmentally harmful products a firm has recently produced. The main difference between GNPD and traditional NPD is that GNPD is more focused on products post-use, and particular design for the ‘‘Five R’s’’ of repair, reconditioning, reuse, recycling, and remanufacture [20]. The NPD process consists of several phases. Following of Rosenthal [21], this study considered four steps NPD process model which consists of ideas generation and conceptual design, definition and specification, prototype and development, and commercialization. Design for Environment : Design for Environment (DFE) encompasses the process by which environmental issues are incorporated into the product realization process. DFE might be the most preferential way to improve green supply chain because changes in the design can cost-effectively leverage environmental effects both upstream and downstream. Valentincic [22] characterized DFE by a simultaneous design of a product and its manufacturing process in order to achieve the best outcome and consequently optimize the overall costs. Major elements of the DFE dimension include design for manufacturing, design for assembly and design for packaging. Green Logistics Dimension: A more tactical set of organizational elements that will influence how the supply chain is to be managed, either internally or externally, can be described by green logistics dimension of an organization. Major elements of the green logistics dimension typically include procurement, production, distribution, reverse logistics and packaging [2], [14], [23-24]. From procurement to reverse logistics these 74 Research in Transport and Logistics 2009 Proceedings Do u University steps encompass whole the life cycle of the products. Although packaging is not viewed as a lifecycle element, it is also affect GSCM. Green Organizational Activities Dimension : The major five green organization activities dimensions are reduce, reuse, remanufacture, recycle and disposal [2], [25]. Extended Producer Responsibility (EPR) requires firms for developing these activities. The responsibility of the environmental impact of products throughout their entire life cycle embraces also the end of life effects. Manufacturing Priorities : Manufacturing priorities can be defined as a set of coordinated objectives and action programs applied to a firm’s manufacturing function and aimed at securing medium and long term, sustainable advantage over that firm’s competitors. In this study, manufacturing priorities are identified as quality, delivery, efficiency, flexibility and customer service. Green Supply Chain Alternatives : In this study, we focus on three supply chain alternatives according to the framework created for the purpose of evaluation of the GSCM. In the framework, alternative 1 (ALT1) is identified as JIT (Just in Time), alternative 2 (ALT2) as environmental focused supply chain management and alternative 3 (ALT3) as current SCM system of the case company. It is obvious that every SCM system has different advantages and disadvantages. In consequence they have different environmental effects. The next step of the study is to justify the proposed structure and derive the interactions and/or relationships in the identified components. This step is realized through review of literature on GSCM, through discussion with industrial experts and our expertise. Finally, the final shape of the conceptualization phase and the relevant components are structured as given in Figure 1. (B) New Product (E) Development Phases (J) Design For Environment (F) (D) (C) Manufacturing (I) Priorities Green Logistics Dimensions (A) (L) (G) Improved GSCM Operations (H) System Alternatives For GSCM Green Organizational (K) Activities Dimensions FIGURE 1 Graphical representation of dimensions and relationships of decision framework for evaluating GSCM. The GSCM system is evaluated on five different dimensions (levels or clusters); the Green Logistics, the Green Organizational Activities, the New Product Development Phases, the Design For Environment and the Manufacturing Priorities. In this study, the interdependence or feedback type relationship occurs between both the green logistics and the new product development dimensions and the new product development and the design for environment as represented by two reverse arrows among those levels. The other arrows in the model indicate a one-way relationship. In addition, the interdependency relationships of green logistics elements and the new product development phases are shown by looped arcs in Figure 1. The capital letters from A to L in parentheses in Figure 1 represent the weight matrices used in the supermatrix construction which is described in details in following section. In the framework, the current supply chain systems of the case companies are compared to two other alternatives (ALT1 and ALT2). The end result or the model will indicate which system best meets the needs of the decision maker based on interaction between five different dimensions. 75 Research in Transport and Logistics 2009 Proceedings Do u University GSCM ANALYSIS PHASE Analytic Network Process Selection of a suitable methodology that can decode the high-level relationship model presented in Figure 1 in order to evaluate possible GSCM alternatives is a critical issue. This methodology should be able to use quantitative, qualitative, tangible, and intangible factors pertaining to the decision of whether and which alternative should be evaluated. With this trend, ANP is capable of taking the multiple dimensions of information into the analysis. ANP is a general form of the analytical hierarchy process (AHP) first introduced by Saaty [3]. While the AHP employs a unidirectional hierarchical relationship among decision levels, the ANP enables interrelationships among the decision levels and attributes in a more general form. The ANP uses ratio scale measurements based on pair wise comparisons; however, it does not impose a strict hierarchical structure as in AHP, and models a decision problem using a systems-with-feedback approach. The ANP refers then to the systems of which a level may both dominate and be dominated, directly or indirectly, by other decision attributes and levels. The ANP approach is capable of handling interdependence among elements by obtaining the composite weights through the development of a ‘‘supermatrix’’. Saaty [3] explains the supermatrix concept similar to the Markov chain process. The supermatrix development is defined in the next sub-section. Although the number of ANP related works has increased in the recent years, there are only four studies realized by Sarkis to our knowledge that are used for environmentally related problems. After presenting a model for evaluating environmentally conscious business practices [12], Sarkis [26] proposed a methodological framework for evaluating environmentally conscious manufacturing programs. This second research integrated the ANP and data envelopment analysis. More recently, Meade and Sarkis [27] used ANP to evaluate and select third party reverse logistics providers while Sarkis [2] used ANP in a decision framework for GSCM. Application of the Evaluation Framework in the Case Companies In the strategic analysis of the GSCM framework, Hyundai Assan (www.hyundai.com.tr), Ford-Otosan (www.ford.com.tr) and Toyota (www.toyotatr.com) are analyzed. These firms are selected because they are world wide firms and they have severe applications on GSCM. -Assan), Vedat Okyar (Ford- -Otosan) and Mustafa Diker (Toyota) offered guidance for the application of the GSCM framework. Pair wise comparisons and the calculation of relative importance weights: Eliciting preferences of various components and attributes will require a series of pair wise comparisons where the decision maker will compare two components at a time with respect to an upper level 'control' criterion. In ANP, like AHP, pair wise comparisons of the elements in each level are conducted with respect to their relative importance towards their control criterion [3]. Saaty [28] has suggested a scale of 1 to 9 when comparing the two components, with a score of 1 representing indifference between the two components and 9 being overwhelming dominance of the component under consideration (row component) over the comparison component (column component). If a component has a weaker impact on the control criterion, the range of scores will be from 1 to 1/9, where 1 represents indifference and 1/9 an overwhelming dominance by a column element over the row element. When scoring is conducted for a pair, a reciprocal value is automatically assigned to the reverse comparison within the matrix. That is, if aij is a matrix value assigned to the relationship of component i to component j, then aji is equal to 1/aij (or aij * aji = 1). The weightings have been obtained from our industrial experts by asking a series of paired comparison questions. Once all the pair wise comparisons are completed, the relative importance weight for each component is determined. Given that A is the pair wise comparison matrix, the weights can be determined by expression: A w = max w, where max is the largest eigenvalue of A. Saaty [28] provides several algorithms for approximating w. In this paper a two- stage algorithm was used that involved forming a new n n matrix by dividing each element in a column by the sum of the column elements and then summing the elements in each row of the resultant matrix and dividing by the n elements in the row. This is referred as the process of averaging over normalized columns. In the assessment process, there may occur a problem in the transitivity or consistency of the pair wise comparisons. For an explanation on inconsistencies in relationships and their calculations see Saaty [28]. The priority vectors for each pair wise comparison matrix will be needed to complete the various supermatrix submatrices. We will need a total of 50 priority vectors to complete our supermatrix. This 76 Research in Transport and Logistics 2009 Proceedings Do u University requirement means that 50 pair wise comparison matrices must be completed for each case company. The pair wise comparison matrix results used were all tested for achieved the consistency goals. Supermatrix formation: ANP uses the formation of a supermatrix to allow for the resolution of the effects of the interdependence that exists between the clusters within the decision network hierarchy. The supermatrix is a partitioned matrix, where each submatrix is composed of a set of relationships between two clusters in the graphical model. A generic supermatrix of our model is shown in Figure 2, with the notation representing the various relationships from Figure 1; for instance, ‘‘A’’ is the submatrix representing the influence relationship between green logistics dimension elements’ and control factor of the goal of selecting a green supply chain system. The solution procedure: The supermatrix M is a reducible matrix with a multiple root, as defined by Saaty [3]. To solve for the values of the alternatives, Saaty [3] recommends that the values of M be column stochastic. That is, the sums of the columns should be normalized to equal a value of 1. To complete this task, each of the columns may either be normalized by dividing each weight in the column by the sum of that column. The final step in the process is to obtain a priority ranking for each of the alternatives. We will determine this ranking by calculating the influence of each of the alternatives on the objective of improving the GSCM. Saaty states that a simple hierarchy and the additive solution approach is appropriate if strong dependencies among the criteria do not exist [3]. But, in this case the dependencies are considered to be strong. In addition, Schenkerman [29] has shown that the supermatrix approach is capable of reducing the occurrence of rank reversal, thus providing more accurate portrayals of decision-maker preferences. Saaty recommends a simple solution technique to solve this problem by raising the supermatrix M to a large power until convergence occurs [3]. The relative influences of the alternatives on the objective of improving the GSCM are shown in the “Goal” column. The results are obtained by using Super Decisions 1.6.0 Software (www.superdecisions.com). ALT DFE GL GOA GOAL MP NPD Alternatives 0 0 0 K 0 L 0 Design For Environment 0 0 0 0 0 0 F Green Logistics 0 0 G 0 A 0 D Green Organizational Activities 0 0 H 0 0 0 0 GOAL 0 0 0 0 0 0 0 Manufacturing Priorities 0 0 I 0 0 0 J New Product Development 0 E C 0 0 0 B FIGURE 2. General submatrix notation for supermatrix. The results for Hyundai-Assan show that ALT 2 has a higher priority value (with score of 0.434) than the current situation (with score of 0.294) which is better than the ALT1 (with scores of 0.272). The results for Ford-Otosan (Figure 4) show that ALT 2 has a higher priority value (with score of 0.406) than the current situation (with score of 0.354) which is better than the ALT1 (with scores of 0.240). The results for Toyota (Figure 5) show that ALT 2 has a higher priority value (with score of 0.400) than the current situation (with score of 0.344) which is better than the ALT1 (with scores of 0.255). The obtained results show that analyzed case companies’ SCM systems are more environmentally conscious than JIT system. On the other hand, it is obvious that they have to implement some adjustments to improve their environmental performance. Hyundai-Assan’s SCM system takes the lowest ANP priority value when it is compared with other companies. So, we can say that Hyundai-Assan has to show more effort to improve his current SCM system. It is remarkable that Ford-Otosan’s and Toyota’s current SCM systems take close ANP priority values, but unfortunately they have some insufficiencies to fully adopt environmental focused supply chain management. For this reason, if Ford Otosan and Toyota manage to overcome these insufficiencies, they will be reference companies for GSCM in Turkey. 77 Research in Transport and Logistics 2009 Proceedings Do u University CONCLUDING REMARKS This paper addresses the need for a strategic analysis model to assist management in evaluating a number of alternatives for improving GSCM initiatives. Through this line, an evaluation model is developed based on a literature survey and refined with industrial experts. The proposed model is implemented in Hyundai-Assan, Ford Otosan and Toyota. The obtained results provide additional insights for research and practical applications. The ANP method used in this study offers a more precise and accurate analysis by integrating interdependent relationships, but requires more time and effort (additional interdependency relationships increase geometrically the number of pair wise comparison matrices). For this reason, an application of the ANP approach, as proposed in this study should be targeted at more strategic decisions, especially for long- term profit and long-term competitiveness considerations. ACKNOWLEDGMENT The authors would like to express their deep gratitude towards the industrial experts of Hyundai-Assan - and Toyota (Mustafa Diker). Gülçin Büyüközkan acknowledges also the financial support of the Galatasaray University Research Fund. REFERENCES [1] van Hoek, R.I., 1999, “From reversed logistics to green supply chains”, Supply Chain Management, 4, 3, 129-134. [2] Sarkis, J., 2003, “A strategic decision framework for green supply chain management”, Journal of Cleaner Production, 11, 4, 397-409. [3] Saaty, T.L., 1996. “The Analytic Network Process”. RWS Publications, Pittsburgh, PA. [4] Zhu, Q. and Sarkis, J., 2004, “Relationships between operational practices and performance among early adopters of green supply chain management practices in Chinese manufacturing enterprises”, Journal of Operations Management, 22, 3, 265–289. [5] Srivastava, S.K., 2007, “Green supply chain management: a state-of-the-art literature review”, International Journal of Management Reviews, 9, 1, 53–80. [6] Hall, J., 2000, “Environmental supply chain dynamics”, Journal of Cleaner Production, 8, 6, 206–225. [7] Geffen, C. and Rothenberg, S., 2000, “Suppliers and environmental innovation: the automotive paint process”, International Journal of Operations and Production Management, 20, 2, 166–186. [8] Theyel, G., 2001, “Customer and supplier relations for environmental performance”, Greener Management International, 35, 61–69. [9] Sarkis, J., 1995, “Supply chain management and environmentally conscious design and manufacturing”, International Journal of Environmentally Conscious Design and Manufacturing, 4, 2, 43–52. [10] Faruk, A.C., Lamming, R.C., Cousins, P.D. and Bowen, F.E., 2002, “Analyzing, mapping, and managing environmental impacts along the supply chain”, Journal of Industrial Ecology, 5, 2, 13–36. [11] Handfield, R., Walton, S. and Sroufe, R., 2002, “Applying environmental criteria to supplier assessment: a study of the application of the analytical hierarchy process”, European Journal of Operational Research, 141, 70–87. [12] Sarkis, J., 1998, “Evaluating environmentally conscious business practices”, European Journal of Operational Research, 107, 1, 159-174. [13] Sheu, J-B., Chou, Y-H. and Hu, C-C., 2005, “An integrated logistics operational model for green-supply chain management”, Transportation Research Part E: Logistics and Transportation Review, 41, 4, 287-313. [14] Sarkis, J., Meade, L.M. and Talluri, S., 2004, “E-logistics and the natural environment”, Supply Chain Management: An International Journal, 9, 4, 303-312. [15] Zhu, Q., Sarkis, J. and Lai, K-H., 2007, “Initiatives and outcomes of green supply chain management implementation by Chinese manufacturers”, Journal of Environmental Management, 85, 1, 179-189. [16] Zhu Q., Sarkis J., and Lai K., 2007, “Green supply chain management: pressures, practices and performance within the Chinese automobile industry”, Journal of Cleaner Production, 15, 1041-1052. [17] Sarkis J., 2006. “Greening the Supply Chain”, Springer-Verlag, London. [18] Beamon B.M., 1999, “Designing the green supply chain”, Logistics Information Management, 12, 4, 332-342. [19] Simpson D., Power D., and Samson D., 2007, “Greening the automotive supply chain: a relationship perspective”, International Journal of Operations & Production Management, 27, 1, 28-48. [20] Pujari D., Wright G., and Peattie K., 2003, “Green and competitive: Influences on environmental new product development performance”, Journal of Business Research, 56, 657–671. [21] Rosenthal, S.R., 1992. “Effective product design and development - how to cut lead time and increase customer satisfaction”. Business one Irwin, 21-30. [22] Valentincic J., Brissaud D., and Junkar M., 2007, “A novel approach to DFM in toolmaking: a case study”, International Journal of Computer Integrated Manufacturing, 20, 1, 28 – 38. [23] Handfield, R.B., Walton, S.V., Seegers, L.K. and Melnyk, S.A., 1997, “Green value chain practices in the furniture industry”, Journal of Operations Management, 15, 4, 293-315. [24] Zhu, Q., Sarkis, J. and Lai, K-H., 2008, “Green supply chain management implications for closing the loop’’, Transportation Research Part E: Logistics and Transportation Review, 44, 1-18. [25] Meade, L. and Sarkis, J., 2002, “A conceptual model for selecting and evaluating third-party reverse logistics providers”, Supply Chain Management: An International Journal, 7, 5, 283-295. [26] Sarkis, J., 1999, “A methodological framework for evaluating environmentally conscious manufacturing programs”, Computers & Industrial Engineering, 36, 4, 793-810. [27] Meade, L. and Sarkis, J., 2002, “A conceptual model for selecting and evaluating third-party reverse logistics providers”, Supply Chain Management: An International Journal, 7, 5, 283-295. [28] Saaty, T.L., 1980. “The Analytic Hierarchy Process”. New York, McGraw-Hill. [29] Schenkerman, S., 1994, “Avoiding rank reversal in AHP decision support models”, European Journal of Operational Research, 74, 3, 407–419. 78 Research in Transport and Logistics 2009 Proceedings Do u University A NEW FRAMEWORK FOR PORT COMPETITIVENESS: THE NETWORK APPROACH Marcella DE MARTINO1, Alfonso MORVILLO 2 Abstract - In the last decade, the port economics literature has given great emphasis to the Supply Chain Management approach as the new paradigm for the definition of port competitiveness. Nevertheless, the application of SCM approach to the port is particularly complex given the traditional hostile relationships between port actors. In the effort to overcome such complexity, some authors have considered ports as Logistics Service Providers interpreting their role within supply chains through the integrative practices undertaken by Shipping Companies and Terminal Operators in the supply of integrated logistics services. Even tough these actors are crucial for the port competitiveness, they determine a passive role of port in the new competitive scenario. The definition of a potential and pro-active role of port in the supply chains is the objective of this paper that presents, through a literature review on SCM and port competitiveness, a new framework for port competitiveness. Keywords - Supply Chain Management, Port competitiveness, Network approach INTRODUCTION In the last decade, the port economics literature has given great emphasis to the Supply Chain Management (SCM) as the new paradigm for port competitiveness. SCM, as managerial philosophy, supports the development of partnerships between actors of the supply chain and considers the integration of activities and resources of these actors along business processes, the source of competitive advantage. Accepting this view, some authors have reconsidered recent contributions and conceptual categories of the managerial literature on Supply Chain Management (SCM) to re-define strategic positioning and port strategies [1] - [2] - [3] - [4]. In particular, a new systematic view is being affirmed, by which the competitiveness of a port, although still strongly related to structure-type variables (geo-economic context, institutional model and the infrastructures of connection), is increasingly dependent on software components of the port business, that include the range of services offered, the ICT systems for the exchange of information between actors in the same and in other port communities, the know-how of the maritime industry, the level and intensity of relationships between actors; all elements that define the quality of a port [2] - [4] - [5]. Through a literature review on SCM and port competitiveness, this paper proposes a new theoretical framework for port development. The paper is structured in two main parts. The first part focuses on logistics integration and SCM issues in the port environment. This section of the paper is particularly outstanding, as each concept has not been dealt with univocally in literature; from an examination of these studies, it becomes evident that there is an absence of frameworks to represent the entire port community in the process of creating value. Because the port is a complex reality, involved in a series of supply chains, each of which is a specific entity with particular needs to meet, networks make it possible to describe the relationships that are created between port actors in the process of customer satisfaction. As a result, in the second part of the paper a new model is proposed that considers determinants for port competitiveness the interorganisational relationships created between the different network actors to manage various business activities and to oversee critical resources in order to satisfy the clients. LOGISTICS AND SCM IN THE PORT LITERATURE In the recent debate on Port competitiveness, a growing number of papers have focused on the issue of integration of port activities in the companies’ supply chains, in the attempt to frame unitarily the port in the process of creating value for the end customer [4]. The difficulties that have since been encountered in the use of a systematic approach in the port – aimed at determining the contribution of the whole port community in the supply chains - is mainly due to the lack of a “competitive community spirit” among the actors, both public and private. The concept of integration in the port context has essentially concerned intermodality and 1 National Research Council (CNR), Institute for Service Industry Research (IRAT), Naples, Italy, m.demartino@irat.cnr.it 2 National Research Council (CNR), Institute for Service Industry Research (IRAT), Naples, Italy, a.morvillo@irat.cnr.it 79 Research in Transport and Logistics 2009 Proceedings Do u University organizational integration undertaken by global carriers aimed at responding to the changing requirements of industrial and commercial enterprises and, at the same time, improving their own internal efficiency [6] - [7]. [8] - [9] - [10] - [11] - [12] [13]. Even though these two areas of study extend the field of analysis to the role and integration modes of ports within global transport systems, on the other hand they underline the passive role traditionally played by ports in relation to the strategic choices of other port community operators, with obvious effects on their bargaining power and, in general, on their development strategies. In this respect, it has already been shown that ports are mere “pawns in the game” within global transport systems and that the power of liner shipping affects port development [14]. In re-defining port positioning and strategies, Robinson [1] tries to systemise recent contributions in the SCM literature by proposing to create a new paradigm, based on the conceptual categories of the value constellation [15]. In the paper, the port is considered to be a Third Party Logistics (TPL) provider that intervenes in a series of different companies’ supply chains. Specifically, the port is a market focused firm, where all businesses are managed to provide superior value to target customer. Moreover, port’s strategic positioning choices involve two key issues [16]: what supply chain resources a port should own or control and how it should defend its ability to accumulate the value so derived. The proposed framework, innovative for the conceptual categories on which it is based, offers interesting insights on port value creation process in the supply chain; in its exemplification, the framework focuses on integrative practices undertaken by shipping companies for the supply of complex logistical services, from intermodal transport to the handling of goods. Paixão and Marlow [3] analysed the application of the agile strategy to the port sector, by focusing on the new managerial strategies undertaken by port players (mainly terminal operators) aimed at tackling the high levels of market uncertainty. By analyzing external integration, the authors, in contrast to Robinson [1], identify the role of the port in the supply chain through the contribution of the “multi-port operator”, i.e. large terminal operators, in the process of creating value. In a empirical work, Carbone and De Martino [2] have analyzed the contribution of the port of Le Havre in the Renault supply chain, according to the SCM approach. The model applied [17], was based on the assumption that a better integration between the port actors leads to a higher competitiveness of the supply chain. The research proved that relational component of port’s offer is crucial in determining port competitiveness. Finally, Bichou and Gray [4] try to conceptualise the port system from a logistics and supply chain management perspective, aiming to define a new framework to measure port performance. As to the results of their exploratory investigation (carried out on a sample of 100 experts in the port field), it is interesting to note that respondents belonging to the port community showed a lack of familiarity with logistics and SCM concepts, especially those related to logistics integration. This implies that all the models described above are still in the phase of theoretical elaboration and that their empirical implementation calls for a “cultural leap” especially with reference to Port Authority and other actors of the port community. PORT COMPETITIVENESS: A NEW FRAMEWORK From the aforementioned literature review on port competitiveness, it becomes apparent that there is a lack of frameworks that are able to represent the complexity of the port community in the process of creating value. The Robinson’s [1] interpretation of port as Third Party Logistics provider correctly combines the creation of value to the concept of value constellations and his work represents a fundamental contribution to the definition of a model of analysis of port competitiveness. In his exemplification, however, the framework deals with value exclusively from the perspective of the shipping companies. This interpretation of port value creation seems more appropriate to describe the contribution of one single port operators in the supply chain and it neglects the advantages offered by promoting the different interactions that can exist between port actors of the supply network. Although shipping companies have a fundamental role in the port competitiveness, they cannot represent the only subject of the development options, because the port should build its success by developing activities and resources that not exclusively favour the shipping company’s distribution network, but also others port users, including the manufacturing companies of its own hinterland. This consideration proves even closer to reality if we consider the port - marketplace relationship, i.e. the function of centrality and inter-connection that the port can play [18]. In particular, the concept of centrality refers to the flow of goods that originates in the regional economic system to which the port belongs (local hinterland). The concept of inter-connection, instead, refers to the position of the port within the intermodal routes [19]. In this case, the flow of goods does not depend on the conditions of the regional economic system, but on the relative position of the port in the distribution network of large shipping companies and multi- modal transport operators [20]. In the light of these considerations, we [21] believe that the SCM determines 80 Research in Transport and Logistics 2009 Proceedings Do u University the differential competencies and services’ features that the port must foster, so as to promote the economic development of its own hinterland (port performing the function of centrality). Moreover, given the state of the art of SCM and port competitiveness, and considering that the port represents a complex reality, we believe that models based on a network perspective, adopting the concept of value chain constellation, can set the framework to determine port competitiveness. In particular, these models take all the possible modalities of interaction into consideration among a multiplicity of network actors, by analyzing the development of inter-organizational relationships in the management of business activities and resources in the process of creating value for the client. Moreover, according to these models, a major factor in strategic positioning is the relational capacity of the company. Mobilizing and combining critical resources and activities is connected to company’s abilities to combine its own resources and activities with those of others [22]. Relationships are the mean by which companies access to, mobilize and combine critical resources and processes, promoting innovation and productivity in the network but also within the company. SCM MODELS ACCORDING TO A NETWORK PERSPECTIVE According a growing number of authors, most of SCM models emphasise the linear relationships between a focal firm, a supplier, and a customer. One major issue overlooked is that any focal organisation is normally part of several supply chains and therefore effort to optimise individual supply chains, without considering interdependencies among chains, may hamper the efficiency elsewhere in the network [23]. Recognising the existence of different interdependencies [24] between the supply chains of a network allows the company to integrate activities and resources at different levels, including that of the network. The sequential (or serial) interdependence suggests that there is a unidirectional relationship between the activities of network actors; the output of an activity represents the input of the next one [25]. Making use of these interdependencies, a company can create economies of integration. The pooled and reciprocal interdependencies between activities, resources and actors allow us to properly understand how value is co-produced by actors in a supply network [23]. In particular, the pooled interdependence between two activities shows that they are connected to a third one or that they share the same resources, and that they are thus indirectly interdependent. Exploiting pooled interdependencies can lead to economies of scale or scope: economies of scale, in the case of two activities being the same, economies of scope when the two tasks are similar. Reciprocal interdependence is the mutual exchange of input and output between two parties. This form of interdependence is closer and stronger, because every variation in activity of one party can be made if, and only if, there is a reciprocal change in activities of the other party. The advantages that can be obtained through these reciprocal interdependencies are associated to the combined capabilities of each party to resolve problems and utilise efficiently and effectively the resources of the network. Exploiting reciprocal interdependencies means “being innovative, agile and reactive to changes”. The supply network structure is a field of study of considerable importance [26], as it includes the identification and description of the relationships between actors of the same network and the process of creating value for the end customer, through different interdependences [27]. Every company is a member of a set of supply chains with different roles and decision-making power. In this environment the inter-organisational relationships are considered to be the most relevant strategic resources [28], “bridges of value”, as they give companies access to other actors’ resources in the network and they strongly contribute to the value co-production [15]. Studies on the network have highlighted the aspects of interdependence and interaction between different actors involved in the creation of value. In particular, they attempt to overcome what is considered to be a traditional vision of value creation, based on the Porter value chain concept, and try to develop alternative models of interpretation, such as, value net, value network, or value creating network [29] - [30] - [31]. According to this view, the value for the client is not created by a single company, but it is co-produced in combination with other actors in the network. The greater the degree of interdependence between the network’s actors, the more potential co-produced value there will be [32]. THE APPLICATION OF DUBOIS’ MODEL From the examination of the SCM models according to a network perspective, the work presented by Dubois [33] proves to be particularly useful in representing the creation of value for the port in supply chains [34]. Dubois [33] in an effort to analyse the complex patterns of interrelated chains (a supply network), suggest a framework consisting of: products, activities and resources, firms (or business units) and relationships. Within such a framework, there are two major assumptions: (1) individual firms try to optimise their respective sets of resources and activities by taking interdependencies across boundaries into account; 81 Research in Transport and Logistics 2009 Proceedings Do u University and (2) the relationships between firms provide them with means to coordinate their activities and to interact in the development of the resources activated by, and of the products resulting from, their respective activities. Source: Dubois [33] FIGURE 1: A complex pattern of interrelated supply chains (Supply network) Firms involved in these chains will perceive products differently (Figure 1). For example, for firm F there are three products that will follow three different supply chains and will end up in three different end products (b, c and d). For firm F there are also other products of relevance; those supplied by firm E as inputs. Firm F’s activities (utilising a common resource) result in three different products that are further refined by G, H, I and J into end products b, c and d. By this way of organising the activities, the three supply chains utilise resources also activated in other chains within, for example, firms B, E, F and K. Hence, in this simplified supply network there are a number of interdependence within and among the supply chains that the actors should take into account in order to be efficient. Each firm in the network will have different perspectives on how to organise and manage their activities, resources. This model, based on the value chain constellation concept already applied by Robinson [1], allows us to view the port as a network of actors who co-produce value by promoting different interdependencies (sequential, pooled and reciprocal) between supply chains. In fact, the port is involved in a number of supply chains (that represent the different entities to satisfy) and, contextually, the services supply in a specific supply chain is generated by different port actors (the port is a network of companies) that pursue a common strategy in satisfying the specific needs of the customer. To apply this model to the port’s environment, two components of the concept of integration must be distinguished on the basis of the responsible actor for strategic decisions: firstly, the Port Authority that determines the quality of hardware components (infrastructure and their inter-connections to the market place) of the port’s offer; secondly, the software component which is defined by the capabilities or the development of distinctive competencies of other port actors in managing various activities in the supply chain. Even if closely interdependent [35], the software component is considered the main determining factor in the port competitiveness, at least for industrialised countries [36]. Port Authorities – mainly in charge of the hardware dimensions of port development - have to ensure and enable the efficient management of supply chains through policy actions. These should be aimed at improving port infrastructures and their connections within existing transport systems, keeping in mind the criteria of environmental, social and economic sustainability; allowing free competition between port operators through concessions of terminals and spaces for the supply of value added services; enhancing the collaboration and coordination of port activities through IT systems; promoting the development of its own hinterland by creating economical, relational and social connections between the port and the market place. The ability of the “port” to recognise and exploit interdependencies within and between different supply chains will determine its capability to create value in supply chains. In our framework, the port is represented as a network of actors that carry out a number of activities in the supply chain in close collaboration, sharing different resources (Figure 2). The higher the level of collaboration (integration) between actors the greater the benefits that they will perceive in promoting the pooled and reciprocal interdependencies between the various supply chains. In this way, the features of the supply chain play an key role in defining the port development 82 Research in Transport and Logistics 2009 Proceedings Do u University policies, because they determine the importance of the resources to be controlled and the activities to carry out in the port. Only through an understanding of these needs, can the port exploit the chance of becoming an active part of the supply chains to which it belongs and thus, gain the advantages of better integration. A S W F H C SC D H MA TO E LO S RO B W Firm Resource Relationship Activity Legenda: SC = Shipping Company; TO = Terminal Operator; MA = Maritime Agency; S = Stevedore company; F= Forwarder; H= Haulier; RO= Railway Operator; W= Warehousing; LO= Logistics Operator Source: Our elaboration FIGURE 2: Port’s value creation within supply network In our specific case, firms A and C have to export their products to firms D and E, respectively. From port perspective, each supply chain represents different entity that may or may not be in conflict with each other for what the demand attributes are concerned (reliability, security, speed, destinations, etc.). In the port, many firms are involved in performing activities through different resources. For exemplification, let consider three main actors: SC, a shipping company, F a freight forwarder and TO, a terminal operator. On the basis of their level of inter-organisational relationships, the port can be viewed as a fragmented entity – in which each operator performed its own activity separately form those performed by the others actors – or a single unit – where port operators performed different activities in close synergy by sharing resources. For example, there may exist pooled interdependencies between the actors SC and F, if they use common logistics resources, such a terminal or communication systems, in order to satisfy the requirements of firm A and B. Moreover, there may exist pooled interdependencies between the two supply chains involving firm A and B, as they jointly use the port to export their products. The analysis of activities, resources, and inter-organisational relationships allows determining port actors’ interaction in the management of port activities and the relative resources exploited in the value generation process. Naturally, resources and activities are completely intertwined, because resources are necessary for the undertaking of activities and have no value unless they are activated. And the way in which resources are “activated” and activities performed depends largely on the inter- organisational relationships among port operators and others actors of the supply network. CONCLUSION This paper starts with the consideration that in today’s competitive context, ports should define strategies and competitive positioning on the basis of a new paradigm that views the ports as actors in the supply chains. The objective is to design port development strategies within the conceptual categories of SCM, so as to define a pro-active role of the port, which is no longer linked passively to the strategic choices of the Global Players, but is a dynamic player in the competitive scenario. From the literature review, it is clear that there is a substantial “gap” between the theoretical frameworks available in the specific literature and the possibility of implementing such frameworks in the port business environment, at least in the short run. Some of the reviewed contributions, in fact, pointed out the scarce sensitivity of port operators and public authorities towards the pursuit of new forms of integration entailing cooperation with other actors and hence mutual trust. In order to overcome such a limitation, the supply network perspective allows us to correctly interpret the creation of value by the port according to a SCM approach, because the current competitive scenario requires 83 Research in Transport and Logistics 2009 Proceedings Do u University flexibility and an integrated approach between different activities and other players in the process of satisfying the end customer. Within such a perspective, the port has to build its competitive advantage on the development of competencies that go beyond the traditional efficient movement of cargo, but include the effective supply of value added logistics services. A determining role must therefore be played by the Port Authority that is called to identify those resources that allow the port to provide a higher quality infrastructures and services, i.e. the so called “critical assets”. As the port is part of a variety of different supply chains (that represent different entities to satisfy) and at the same time the offer of services is generated by various port players (the port is a network of companies), the work output is the definition of a framework to identify the creation of value for the port. In accordance to the new paradigm, by which competition is not played by individual companies but between supply chains, port competitiveness is determined by the quality of the entire port, both in terms of infrastructure and links to the transport systems (hardware components) and in terms of services (software components). In this context, development strategies of the Port Authority has to foster, through the creation and implementation of infrastructures of connection and access to international communication and business networks, the development and promotion of new logistics and transport businesses that are able to support different production and distribution chains. Resources play an important role as they can promote at different levels, the development of inter-organisational relationships between the various players in the supply network. In this scenario, terminals represent only one element of the Port development policies, because they cannot promote the creation of new added value logistics activities on their own. ICT systems, modal-interconnections and the acquisition of new areas, represent critical assets for the port to succeed in becoming the centre of attraction and development of new logistics and transport businesses. On the other hand, port operators perform their own activities within a supply network by using and sharing resources; in such environment, each supply chain represents a specific entity with specific needs to be satisfied. Port operators confront different requirements and their ability to exploit interdependencies among supply chains will determine extent to which port, as a complex net, crates value in such a supply network. In conclusion, the concept of value constellations represents more appropriately the value configuration approach to the diagnosis of port’s competitive advantage. Activities, resources and the level of inter-organisational relationships between the players in the network are therefore critical and essential in the port’s value creation process. In this scenario, the framework presented, that subsequently needs to be validated, represents a useful tool for the Port Authority to define new core-businesses in the port, to specify relevant core and threshold competencies that need to be developed and therefore to decide its strategic positioning in this new competitive scenario. REFERENCES [1] Robinson R., 2002, “Ports as elements in value-driven chain systems: the new paradigm”, Maritime Policy and Management, vol. 29, n. 3, pp. 241-255. [2] Carbone V. and De Martino M., 2003, The integration of port operators in the automotive supply chain: the port of Le Havre and Renault, Les Collections de l’INRETS, n. 251, Paris. [3] Paixao A.C. and Marlow P.B., 2003, “Fourth generation ports- a question of agility?”, International Journal of Physical Distribution & Logistics Management, vol. 33, n. 4, pp. 355-376. [4] Bichou K. e Gray R., 2004, “A logistics and supply chain management approach to port performance measurement”, Maritime Police & Management, vol. 31, n. 1 pp. 47-67. [5] CNEL, 2004, La competitività della portualità italiana, CNEL Report. [6] Morash E. A. and Clinton S. R., 1997, “The role of transportation capabilities in international supply chain management”, Transport Journal, vol. 36, n. 3, pp. 5-17; [7] Janic N. and Reggiani A., 2001, “Integrated transport systems in the EU”, Transport Review, vol. 21, n. 4, pp. 469- 497. [8] Slack B. and Frémont, A., 2005, The Transformation of Port Terminal Operations – From the Local to the Global, Transport Reviews, Volume 25, Number 1/January 2005, pp. 117-130. [9] Notteboom T.E. and Winkelmans W., 2001, “Structural changes in logistics: how will port authorities face the challenge?”, Maritime Policy & Management, vol. 28, n. 1, pp. 71-89. [10] Cariou P., 2001, “Vertical Integration Within the Logistic Chain: Does Regulation Play Rational? The Case for Dedicated Container Terminals”, Transporti Europei, 7: 37-41. 84 Research in Transport and Logistics 2009 Proceedings Do u University [11] Heaver T., Meersman H., Moglia F., Van de Voorde E., 2000, “Do Mergers and Alliances Influence European Shipping and Port Competition?” Maritime Policy and Management, 27: 363-373. [12] Evangelista P. and Morvillo A., 1999, “Alliances in Liner shipping: an instrument to gain operational efficiency or supply chain integration?”, International Journal of Logistics: Research and Application, vol. 2, n. 1. [13] Notteboom T.E. (2004), “Container Shipping and Ports: an overview”, Review of Network Economics, vol. 3, issue 2, pp. 86-106. [14] Slack B. (1993), “Pawns in the game: ports in a global transport system”, Growth and Changes, vol. 24, pp. 579-588. [15] Normann R. and Ramirez R. (1993), “From Value Chain to Value Constellation: Designing Interactive Strategy”, Harvard Business Review. [16] Cox A. et al. (2002), Supply Chain, Markets and Power, Routledge, London and N.Y. [17] Lambert D.M. (2001), “The supply chain management and logistics controversy”, in Brewer A.M., Button K.J. and Hensher D.A. (ed.), Handbook of logistics and supply chain management, vol. 2, Pergamon, pp. 99-125. [18] Genco P. (2000), “Globalizzazione dei mercati e nuove dimensioni della concorrenza interportuale”, Economia e Diritto del Terziario, n. 1, pp. 7-37. [19] Siviero L. (2005), Economia dei trasporti intermodali e innovazione logistica, FrancoAngeli. [20] Notteboom T.E. and Rodrigue (2005), “Port Regionalization: towards a new phase in port development”, Marittime Policy and Management, 32, n. 3, pp. 297- 313 [21] De Martino M. and Morvillo A. (2005), Logistics integration for the development of the National and Local Economic System: the case of the port of Naples, Logistics and Research Network 2005 Conference, 7-9 September, University of Plymouth – UK, pp. 125-130. [22] Tunisini A. and Snehota I. (2003), The Role of Supply Networks in Strategic Positioning, Annual IMP Conference, Lugano 4-6 September. [23] Hakansson H. and Persson G. (2004), “Supply Chain Management: the logic of supply chain and networks”, The International Journal of Logistics Management, vol. 15, n. 1, pp. 11-26 [24] Thompson J. D. (1967), Organisations in action, McGraw-hill, New York [25] Porter M.E. (1985), Competitive Advantage: Creating and Sustaining Superior Performance, Free Press. [26] Hakansson H. and Snehota I. (1995), Developing relationships in business networks, Rutledge, London. [27] Lazzarini S. G. et al. (2001), “Integrating supply chain and network analyses: the study of netchain”, Chain and Network Science, vol. 1, n. 1, pp. 7-22. [28] Hakansson HY. (1982) (eds), International Marketing and Purchasing of Industrial Goods: an interaction approach, Wiley, Chichester. [29] Stabell B.C. and Fjeldstad O.D. (1998), “Configuring value for competitive advantage: on chain, shops and networks”, Strategic Management Journal, vol. 19, pp. 413-437 [30] Heumer L. (2004), “Value creation strategies in supply networks: the case of logistics service providers”, IMP Conference 2004, Dijon, France, 5-7 September. [31] Holzle K., Ritter T., Walter A. (2002), “Relationship function and customer trust as value creators in relationships: a conceptual model and empirical findings for the creation of customer value”, paper presented at the 18th Annual IMP Conference, Dijon, September. [32] Hakansson H and Waluszewski A. (2001), Co-evolution in technological development – the role of friction, paper presented at the 17th IMP Conference, Oslo [33] Dubois A. et al (2003), “Interdependence within and among supply chain”, paper presented at 12th International IPSERA Conference – Budapest. [34] De Martino M. and Morvillo A., (2007), Supply Chain Management e competitività portuale: nuove prospettive di analisi, Economia e Diritto del Terziario, n 1, pp. 93-118 [35] Normann R. (1991), Service Management, John Weley & Sons, New York. [36] Huybrechts M. et al. (2002), “Port Competitiveness: an economic and legal analysis of the factors determining the competitiveness of seaports“, Editions De Boeck Ltd 85 Chapter 5 Sustainable Transport Policies, Traffic Engineering Research in Transport and Logistics 2009 Proceedings Do u University CLEAN TRANSPORT: INNOVATIVE SOLUTIONS TO THE CREATION OF A MORE SUSTAINABLE URBAN TRANSPORT SYSTEM Ela BABALIK-SUTCLIFFE1 Abstract Environmental problems at both urban and global levels have become major issues in the field of transport planning. A significant policy change took place in national and supra-national transport policy documents, which emphasize the need to improve, and increase the usage of, clean and sustainable modes of urban transport. In the past decade, transport planning in many cities around the world experienced a radical change in approach, placing increasing emphasis on public transport and non-motorized modes as well as various innovative programs for changing mobility patterns, travel behavior, and eventually raising awareness on clean transport. This paper provides an overview of a Europe-wide project on policies and implementations of individual cities regarding clean and sustainable urban transport. The project has helped to gather a large number of best-practice cases, a selection of which are presented in this paper in order to provide a better understanding of the wide spectrum of possible plans, projects and programs that local governments can implement to create a cleaner and more sustainable urban transport system. Keywords Sustainable transport, clean transport, best-practice, ELTIS INTRODUCTION The transport sector is one of the main contributors to environmental problems, such as global warming, greenhouse effects and the rapid consumption of non-renewable resources. It is clear that the way we plan and operate transport systems has to change if we are to comply with the objectives of sustainable development [1]. Such a policy change has taken place in many parts of the world, with sustainable and clean transport objectives becoming an integral part of most national and supranational policy documents. With the Brundtland Report [2] in 1987, the Earth Summit [3] in Rio de Janeiro in 1992, and the Kyoto Protocol [4] in 1998, the need to attain a more sustainable development and the need to minimise the negative effects of development on the environment became universally-acknowledged objectives. The Habitat II meeting in Istanbul in 1996 and the OECD Conference [5] on sustainable transport in the same year in Vancouver brought arguments for sustainability to the forefront of urban planning. The 2001 Habitat document [6] on the role of urban transport on sustainable human settlements development helped to highlight the importance of the transport sector in attaining sustainability goals, while the World Bank Urban Transport Strategy [7] published in 2002 provided a framework for urban transport planning with its focus on the effects of transport on urban development, the environment, and poverty reduction, as well as the importance of mass rapid transportation, public road passenger transport, non-motorised transport, and techniques for demand management, traffic management and pricing. The 2001 EU Transport White Paper also emphasised the need to create a more balanced transport system in which clean transport modes should play an increasingly important role. The 2007 EU Green Paper on urban mobility [8] further highlighted the importance of clean transport and the potential provided by non-motorised modes. These policy documents helped to shape urban transport policies in most European countries; and as a result, reducing petrol-dependency, emissions, congestion and bottlenecks, and at the same time improving accessibility, environment and quality of life became priority objectives. There is broad agreement that public transport systems and non-motorised modes of transport must be improved if the use of these more environmentally-friendly modes is to be encouraged and accessibility ensured for all. It is also important that such improvement schemes are introduced together with restrictions on car usage, such as congestion charging, parking control and management, pedestrianisation, car-free residential areas, etc., with a view to increasing their impact on mode choice and traffic reduction [9]. The governments of many European countries have introduced policy documents to steer policymaking and implementations towards these approaches for more sustainable transport systems. UK policy documents are noteworthy in this regard: the Planning Policy Guidance on Transport [10] provides an explicit emphasis on the need to promote more 1 -Sutcliffe, Middle East Technical University, Faculty of Architecture, Department of City and Regional Planning, Ankara, Turkey, ebaba@arch.metu.edu.tr 89 Research in Transport and Logistics 2009 Proceedings Do u University sustainable transport choices and to reduce the need for travel, especially by car. The 1998 UK White Paper on transport had a clear emphasis on sustainable and clean transport, as well as the need to abandon road programmes; while the White Paper in 2004 highlighted strategies for clean transport, i.e. making walking and cycling a real alternative for local trips, and improving public transport. The government also provided guidelines for local authorities on ways to encourage walking [11]. While such national policy documents may be fewer in number in other European countries, there has been, nevertheless, a significant shift in policies towards clean transport, affecting the nature of projects and regulations in urban transport. This paper provides an overview of such projects implemented in European cities, representing the radical change from the road- and car-oriented transport programmes of the past. These are examples of innovative solutions in creating a sustainable clean urban transport system, and the cases are selected from a vast list of best-practice cases gathered through a Europe-wide project on clean urban transport. The project, which aims to encourage the sharing of practices and gains of various cities in their attempt to transform their transport planning approach, have helped to gather a large number of best-practice cases, largely from Europe but also from other parts of the world, with the underlying objective of helping other cities towards more sustainable solutions in urban transport, and raising the awareness of city governments, as well as citizens2. The case studies presented in this paper are those that the Middle East Technical University’s working team, as the Turkish section of the mentioned project, had focused upon: firstly, a number of best-practice cases from Europe were selected to be shared and disseminated among local governments in Turkey; and secondly the best-practice cases from Turkey were gathered to be included in the project. Through these two sets of case studies, the paper also helps to facilitate a comparison of urban transport planning approaches in European and Turkish cities. BEST-PRACTICE CASES FROM EUROPE Innovative solutions for a sustainable green transport system cover a wide and diverse set of policies and projects. The cases gathered as best-practices in the project are categorised under such project headings as: clean and energy-efficient vehicles; public transport improvements; cycling; walking; demand management and pricing; mobility management and travel awareness; transport and land-use planning, etc. In this section, innovative projects that represent contemporary planning approaches in urban transport are presented. Projects for Clean and Energy-Efficient Vehicles The introduction of public transport vehicles that run on alternative “green” fuels is one of the major areas for intervention in the creation of sustainable urban transport. Diesel, natural gas and biogas are some of the alternatives available to help reduce the air pollution created by public transport buses. Hydrogen buses were introduced in Reykjavik, Iceland; Trolleybuses running partially on batteries were introduced in Landskrona, Sweden; not only public transport buses, but also taxis and private company vehicles were encouraged, through incentives such as free parking and a priority taxi lane, to convert to natural gas in Gothenburg, Sweden. Similarly, electric cars were offered as rental vehicles in the city centre in Trento, Italy; and electric scooters and hybrid passenger cars were introduced as part of a trial project to test the affects on air quality in Thessalonica, Greece. All these projects were aimed at encouraging the use of renewable energy sources and reducing emissions and air pollution. Projects to Improve Public Transport Systems Most cities worldwide are investing in their public transport systems to increase the service quality with a view to attracting more passengers, and in particular to create a modal shift from private cars to public transport. New metros, light rail and tram systems have been introduced in many cities, not only in Europe, but in all parts of the world. Such urban rail investments have been extremely popular with local governments; however, there is also an increasing recognition that the improvement of bus systems, through the introduction of busways, buslanes and bus priority schemes, can be equally effective and at much lower costs. In addition to these infrastructure investments, a relatively new approach that has been seen in several European cities is the introduction of information or mobility centres, which provide travel information and travel plans to both residents and visitors. Such centres not only provide information on the transport services of the cities, but 2 These individual experiences and cases are listed in the wide case study database of the European Local Transport Information System [ELTIS] webpage at www.eltis.org, which is one of the main intended outcomes of the project. 90 Research in Transport and Logistics 2009 Proceedings Do u University also raise awareness on sustainable alternatives in traveling and encourage users to choose public transport, cycling and walking in their trips. Cycling Cycling is considered as not only a sustainable travel alternative in urban transport, but also a means of introducing regular exercise into the daily lives of a country’s population. In this respect, cycle lanes and networks have been introduced and expanded in cities in Belgium, the United Kingdom, Bulgaria, Croatia, the Czech Republic, Denmark, the Netherlands, Finland, France, Hungary, Italy, Poland and Serbia. Taking this one step further, free or rental bicycle points have been introduced throughout the urban areas (and especially at urban rail station sites) of many cities, including Paris and Lyon in France; all of the Ranstad area in the Netherlands; Munich, Berlin, Frankfurt, Cologne, Stuttgart and Karlsruhe in Germany; Leuven in Belgium; Bern, Genf, Lausanne, Thun and Zurich in Switzerland; and in a number of cities of Italy and Spain. The latter measure aims to incorporate cycling into commuter trips in integration with public transport journeys. In addition, many cities have launched competitions and campaigns to encourage their citizens to use cycling as a daily transport mode to increase their health values. Walking Pedestrianisation schemes and the improvement of pedestrian routes and zones have become widespread in many European cities. Cities of Belgium, France, Denmark and the Netherlands have played a pioneering role in the pedestrianisation of their city centres. However, schemes to promote walking do not have to bring about full pedestrianisation: improvements of sidewalks in terms of their width, pavement quality, cleanliness, lighting and better accessibility through ramps have been regarded as effective measures. In several cases, such improvement schemes have been integrated with ‘Safe Routes to School’ concepts, which help to create a safe and pedestrian-friendly network in accessing schools. In such schemes, as well as in other parts of urban areas, traffic calming measures have been implemented to reduce traffic speeds and create safer roads for walkers and cyclists. The ‘Home Zones’ scheme was launched in the UK, inspired by the Woonerf concept, with a view to reclaiming the streets in residential neighbourhoods and creating safe open spaces for children to play and for communities to gather. In addition to all of these improvement schemes, campaigns, education programs and games for schoolchildren have been introduced to encourage people to walk more. Demand Management and Pricing While improvements in public transport, cycling and walking are important in sustainable transport planning, alone they are not effective in creating a modal shift from the private car to these more sustainable modes of travel. It is essential that such schemes are supported by measures that make driving into the city centre more difficult, expensive, and therefore less attractive and convenient. Closing certain roads and zones to motorised traffic is one way of making driving more difficult; and hence pedestrianisation schemes can help discourage car usage in central urban areas. In addition, European cities have opted increasingly for measures to make it more expensive to drive into the city centre. Parking fee schemes have been introduced in to increase the cost of driving into the city centre, backed by the introduction of park and ride systems. More importantly, congestion charging has come to be considered as an effective measure: the introduction of such a scheme in London, UK, in 2003 helped to raise revenues for public transport and decreased traffic in the central zone, although causing overcrowding on the public transport system. Provided that public transport is also improved, congestion charging can indeed result in significant reductions in traffic, and such schemes have also been adopted in Stockholm, Sweden in 2006; and in Valletta, Malta in 2007. Awareness-Raising Campaigns Awareness-raising campaigns are not a new concept in urban transport, with campaigns for traffic safety, driver behaviour and the wearing of seatbelts being widespread around the world. However, today there are an increasing number of awareness-raising campaigns to encourage citizens to make sustainable travel choices, aimed at increasing the awareness of all target groups (schoolchildren, parents, university students, residents, etc.) of the consequences of their travel decisions and transport mode choices. Street festivals and one-day closures of selected residential streets have been carried out in many cities, particularly in conjunction with Home Zones projects. In addition, education programs and games directed at schoolchildren have become widespread: The Snakegame [12] was tried out in Belgian schools to encourage schoolchildren to make 91 Research in Transport and Logistics 2009 Proceedings Do u University sustainable travel choices within a game; Mobility Week workshops [13] and orienteering games were carried out in schools in Switzerland; a short movie competition [14] was held among university students in Belgium to allow them express their thoughts on transport, the environment and sustainability; and similarly a marketing competition [15] was held among university students in the Netherlands to stimulate them to discover ways of promoting and ‘selling’ sustainable transport alternatives. Competitions and health programs have also been launched in cities to encourage citizens to cycle and walk to their places of work or education with a view to improving health. Furthermore, many cities in Europe encourage, and in some cases oblige, companies and schools to prepare ‘Travel Plans’ that inform employees and students about sustainable travel alternatives. Transport and Land-Use Planning While there are various transport measures that can help reduce traffic in towns, it is increasingly recognised that land-use planning is one of the most effective tools in changing travel behaviours and encouraging more walking, cycling and public transport usage. It is well documented that urban rail investments cannot be successful unless there is vigorous planning action by local governments to increase development densities along these systems, and create transit-oriented and pedestrian-friendly developments around the stations [16, 17]. The integration of light rail investment and land-use planning in Portland, Oregon, US, has long been a best-practice case, illustrating that traffic reduction can be achieved when the light rail line is planned as a major urban development corridor, and when transit-oriented design approaches are implemented at station sites. In addition to using land-use planning in support of public transport investments, good public transport access can be a tool in creating land-use patterns that can eliminate car traffic. ‘Car-Free Settlements’ have been developed in a number of European cities: ‘car-reduced living’ projects have been introduced in Berlin in Germany and Zurich in Switzerland; and car-free residential areas have been planned and successfully implemented in Bremen, Cologne, Freiburg and Hamburg in Germany; as well as in Amsterdam in the Netherlands; Camden, Edinburgh and Surrey in the UK; Gent in Belgium; Vienna in Austria; etc. Experiences in these cities reveal that car-free and car-reduced residential areas can be successful and popular, provided that they are well supported with good public transport systems. EXPERIENCE OF TURKISH LOCAL GOVERNMENTS WITH SUSTAINABLE AND GREEN TRANSPORT Projects for Clean and Energy-Efficient Vehicles Buses that run on natural gas have been introduced in many cities. Ankara introduced both diesel and natural gas buses as a measure to reduce air pollution in the city. In Istanbul, a ‘Clean Fuels/Clean Vehicles’ project has been launched to investigate transport-based emissions and introduce mitigation measures. The first phase of the project was a transport emissions inventory; based on the outcomes of this phase, plans are now being made to reduce transport-based emissions in the city. Projects to Improve Public Transport Systems In terms of public transport improvements, perhaps similar to many cities in the world, there is growing interest in urban rail systems in Turkey. Konya and Kayseri have all opted to introduce urban rail systems, however not all these cases can be considered as best-practices, as there is very little interest in integrating some of these systems with land-use planning, or supporting them with measures to restrict and discourage car usage. is one of the best-practice cases, with its high ridership throughout the day, and its positive impact on mobility and accessibility in the city centre. Two of the city centre’s main boulevards have been allocated to the tram system, and have thus been closed to motorized traffic. This has been complemented by further pedestrianisation projects in the centre, creating a pleasant, although limited, network of car-free streets. In addition, regeneration and redevelopment projects were implemented at the same time; and the Porsuk River, which runs through the centre of the city, has been rehabilitated and its surroundings integrated into the pedestrian system. All these projects altered the overall physical appearance of the city centre, resulting in reduced traffic and emissions, and increased walking and public transport usage. The Izmir light rail system is also a best-practice case, being introduced as part of a wider package of projects under the ‘Transformation in Transportation’ Project. The ferries, which were previously under- 92 Research in Transport and Logistics 2009 Proceedings Do u University utilized, underwent renewal and were integrated with the new light rail stations and bus stops; bus routes were restructured and feeder buses were introduced for the light rail system; and a new fare and ticketing system was launched. As a result, public transport usage has increased significantly: ridership of the ferries and light rail has increased, while bus travel, which decreased initially due to the extensive restructuring, has again increased, which is testament to the potential of a well-integrated public transport system. Busways and buslanes are the least common forms of public transport improvement in Turkish cities, despite their proven success in a number of European cities. It would appear that the popularity of urban rail alternatives overshadow busway options. It is only in Istanbul that a busway project has been implemented, and recently extended, along an important traffic demand corridor: the system provided a high quality service, and more importantly significant time savings in this heavily-congested corridor. Cycling Projects to introduce bicycle lanes and networks are rather limited in Turkish cities. Konya is the only city with an extensive cycle lane network, currently measuring over 60 km and planned to be further expanded in the near future. Konya enjoys a rather flat topography, which has resulted in a cycling tradition among its citizens; however, the local government has also been promoting cycling in the city, both through the network expansion, and through annual cycling festivals. In addition to Konya, the Nilufer district in Bursa has also launched a project to create an extensive cycling network, with plans to introduce cycle parks at LRT stops. Walking Walking accounts for more than 20 percent of trips in most metropolitan cities in Turkey, and more in smaller cities. Nevertheless, the current infrastructure for walking, i.e. pedestrian road network, pedestrian zones, and even the physical quality of sidewalks are extremely poor, with no plans for improvement evident among the loc corridors in the city centre have been closed to motorized traffic, becoming tram-only roads with wide pedestrian areas. In has been applied to widen and enhance the physical quality of the sidewalks: a busy and congested road has been converted into a one-way street, allowing for a significant expansion of the pedestrian sidewalks. In addition, the road has been cobbled to decrease vehicle speeds, and on-street parking has been restricted. After the scheme, pedestrian volume significantly increased, particularly among women and the elderly, which is testament to the sense of increased safety and security. These examples, however, remain as exceptions, rather than the rule. Most metropolitan cities choose to increase the capacity of vehicle roads in the city centres, resulting in narrower sidewalks, making walking an inconvenient and insecure mode. Turkey has no example either of ‘Home Zones’ or ‘Safe Routes to School’. Demand Management and Pricing Parking fees in the city centre appear to be the only management and pricing measure in Turkish cities, although this would appear to be aimed more at raising revenue than a means of discouraging motorists from driving into the city centre. Park and ride schemes are also limited. Awareness Raising Campaigns Awareness raising campaigns for sustainable transport are also limited. Besides the annual cycling festival in Konya, the most important event has been ‘The Streets Belong To Us”, held in Istanbul throughout 2007, in conjunction with the ‘Towards Carfree Cities Conference’ held in Istanbul in September 2007 by the World Carfree Network. Organized by the Mimar Sinan Fine Arts Academy in Istanbul, the Turkish Association of Road Traffic Accident Victims, and the World Carfree Network, each month a residential street in Istanbul was closed to traffic for a day (on Sundays); and various activities were held to show the residents how they can benefit from their streets as a safe open urban space. The event proved to be very popular, creating a safe open area for a day in which children could play and the local community could gather. Transport and Land-Use Planning While urban rail systems are proposed by many cities, they are seldom integrated into the metropolitan and regional plans, and once the rail systems are put in place, it is seldom that local plans are formulated to support these investments. The Ankara metro system was designed alongside the metropolitan plan of the city, 93 Research in Transport and Logistics 2009 Proceedings Do u University and therefore took place along the main development corridor of the city. However, after its construction, the location of the metro line was not taken into account in the local development system was implemented alongside with urban regeneration projects; however, its integration into the wider metropolitan development policies and projects of the city has been rather limited. In addition, no local governments have experimented with ‘car-reduced’ or ‘carfree’ living, aside from the Princes’ Islands in Istanbul, which have been transformed into carfree settlements, revealing how the elimination of motorized traffic can create a smoke-free, noise-free and clean environment. However, these are islands, which make the implementation of such a scheme easier, and so far there have been no attempts to transfer this experience to new residential development areas. COMPARISON AND CONCLUSIONS The review of European best-practice cases shows that urban transport planning has indeed gone through a significant transformation, and that many projects and schemes that were once considered radical and unacceptable are now being implemented in many cities with promising results. Urban transport systems and policymaking are required to go through a significant shift in order to comply with the objectives of sustainability; and it appears that this shift has already taken place in a number of countries in Europe. In the case of Turkey, it is evident that some best-practice cases have been introduced, mostly notable in the development of new urban rail systems. However, the experience of local authorities in other measures towards sustainable urban transport is inadequate. Projects for improving and encouraging cycling and walking, the two most important modes for sustainable transport, remain limited, along with demand management and awareness raising campaigns. Nevertheless, there have been some promising examples of local campaigns in this regard, among which can be listed ‘The Streets Belong to Us’ in Istanbul, and the annual cycling festival in Konya. The ELTIS project can be considered as an inventory study, and is an important platform for the presentation of contemporary planning approaches in sustainable urban transport. The comparison of the experiences of Turkish local authorities against this inventory is helpful in showing the strengths as well as weaknesses and shortfalls in urban transport planning and policymaking in Turkey. REFERENCES [1] Low, N., 2003, Is urban transport sustainable? In N. Low and B. Gleeson (eds) “Making urban transport sustainable” (New York: Palgrave Macmillan), pp. 1 – 22 [2] World Commission on Environment and Development, 1987, Our Common Future. The Brundtland Report. [3] Earth Summit, 1992, Rio Declaration on Environment and Development. [4] Kyoto Protocol, 1998, Convention on Climate Change. [5] OECD, 1996, Towards Sustainable Transportation. The Vancouver Conference. [6] UNCHS (Habitat), 2001, The Role of Urban Transport in Sustainable Human Settlements Development. Background Paper no.7 [7] World Bank, 2002, Cities on the Move: A World Bank Urban Transport Strategy Review. Washington DC. [8] EU, 2007, Towards a New Culture for Urban Mobility, Green Paper. [9] Goodwin, P.B., 2001, Traffic Reduction, in Button K.J. and D.A. Hensher (eds) Handbook of Transport Systems and Traffic Control, Pergamon, pp. 21-32 [10] UK Department for Communities and Local Government, 2001, Planning Policy Guidance 13: Transport. [11] UK Department of the Environment, Transport and the Regions, 2000, Encouraging Walking: Advice to Local Authorities. [12] http://www.eltis.org/study_sheet.phtml?study_id=784&lang1=en [13] http://www.eltis.org/study_sheet.phtml?study_id=1150&lang1=tr [14] http://www.eltis.org/study_sheet.phtml?study_id=821&lang1=en [15] http://www.eltis.org/study_sheet.phtml?study_id=1573&lang1=en [16] Banister, D., 1999, “Planning more to travel less: land use and transport.” Town Planning Review, 70(3), pp. 313-338 [17] Bernick, M. And R. Cervero, 1997, Transit Villages in the 21st Century. McGraw-Hill 94 Research in Transport and Logistics 2009 Proceedings Do u University EFFECTS OF URBAN BOTTLENECKS ON HIGHWAY TRAFFIC CONGESTION: CASE STUDY OF ISTANBUL, TURKEY Dar n AKIN1 and Mehtap ÇEL K2 Abstract— This paper presents findings of a study which analyzed the effects of highway bottlenecks on urban traffic congestion as a side product of a large-scale project in which a “speed-delay study” was conducted for the development and calibration of a “Travel Demand Model (TDM)” for the city of Istanbul, Turkey. Speed-travel time data were collected on- and off-peak travel times of a weekday as well as weekends using the maximum-car method on toll roads, highways and major arterials of the city. The analysis of speed-travel time data revealed that the worst delays and longest queues were observed at crash locations (travel time index, TTI, varied between 4.20 and 17.8) since they were unexpected incidents with little chance to avoid without an advanced drivers information system in order to warn and re-route the upstream traffic. The worst case among the observed incidents caused long delays was the accident occurrence on a divided highway, D100 (E-5) at a 6-lane section (TTI=17.8). Capacity decreases due to road constructions was also as bad as the traffic flow blockages due to accident occurrences (TTI is varied between 3.82 and 4). Key words: bottleneck, floating/maximum car method, Istanbul, traffic congestion, travel time.. 1. INTRODUCTION Economic consequences of the urban traffic congestion are reached to tremendous points in terms of external costs due to lost time in daily traffic. The annual total cost of traffic congestion in Istanbul is estimated to be 3.12 billion US$. The annual average traffic delay per capita in Istanbul is 73.9 hours whereas this figure is 51 hours in Los Angeles which has the same size as Istanbul in terms of the urban population and experience a severe traffic congestion in the urbanized area. A recent study revealed that the annual cost of the traffic congestion in Istanbul is estimated to be 144,713,430 vehicle-hour delay, 1,034,701,022 person-hour delay, and additional fuel consumption of 441,217,731 liters [1]. As for the world-wide consequences of traffic congestion, the 2004 Annual Mobility Report of Texas Transportation Institute [2] has reported concerning results. Regarding the 85 major urban areas in the US, urban congestion has been increasing everywhere in areas of all sizes and thus, longer portions of the day as well as more travelers and goods are affected by congestion. While 32 percent of the total peak period was congested in the 85 major urban areas in 1982, this ratio has increased to 67 percent in 2002 by an increase more than 100% in 20 years’ time. Therefore, the total time that is spent stuck in traffic has risen from 0.7 billion hours in 1982 to 3.5 billion hours in 2002. Likewise, additional fuel consumption due to traffic congestion was 1.2 billion gallons in 1982 while it increased to 5.7 billion gallon in 2002. Besides, the total annual cost of traffic congestion, which was estimated to be 14.2 billion US$ in 1982, increased by more than four times and reached to 61.0 billion US$ in 2001 [3] and then increased further by 3.6 per cent to 63.2 billion US$ in 2002. Moreover, the total traffic delay in the 85 major urban areas in 2002 was 3.5 billion hours, which is annually equivalent to a total delay of approximately 400,000 years or 46 hours per individual [2]. Another significant effect of the remarkable increase in traffic congestion is that the average weekday peak period trip takes 37 percent longer than the same trip during the off-peak period whereas it used to be only 12 percent longer in 1982. 45 percent of the traffic congestion is led by recurring problems including insufficient capacity, improper signal timing and inadequate utilization of intelligent transportation systems. Table 1 summarizes the causes of the traffic congestion and clearly depicts that bottlenecks is the number one source of traffic congestion in the US with 42 percent [4]. 1 Department of City & Regional Planning, Gebze Institute of Technology, Kocaeli, Turkey, dakin@gyte.edu.tr 2 Mehtap Çelik, Masters Students, Department of City & Regional Planning, Gebze Institute of Technology, Kocaeli, Turkey, mcelik@gyte.edu.tr 95 Research in Transport and Logistics 2009 Proceedings Do u University TABLE 1 Major Sources of Traffic Congestion in the US [4] Source of traffic congestion Percent Bottlenecks 42 Traffic incidents 26 Bad weather 16 Highway repair and maintenance 11 sources Improper signal timing 5 This study researched the effect of bottlenecks on travel times on major highways and arterials of the city of Istanbul. Travel time data collected on- and off-peak times of day using the maximum-car method. Bottlenecks are defined as the roadway segments where the number of lanes serving the traffic suddenly decreased, and as the toll booths on the access-controlled road segments (toll roads) where vehicles have to slow down while passing through automated passing booths (called OGS, Automated Passing System, in Turkish) or come to a complete stop at cash (manual) or magnetic card reader (called KGS, With Card Passing System, in Turkish) booths. Besides, temporary bottlenecks that are created by the incidents such lane drops or lane shifts due to road or infrastructure constructions and by crashes are also included in the analysis. The paper is organized as follows: The next section describes the methodology of the research study. Data analysis and results are offered in section 3. Summary and conclusions are offered in the last section. 2. METHODOLOGY Floating car method is used to collect the travel time data. A compact sedan car is run to collect travel times during different time periods between major intersections and interchanges, and then the speed is calculated based upon the distances obtained from the highway network seen in Figure 2. 2.1. Data Collection Travel time data collected at the bottlenecks which are defined in the next section fall into one or more of the following time periods starting 24th of April, 2006 until the schools are closed for summer on 16th of June, 2006: FIGURE 2 Highway Network of the City of Istanbul. 1. Morning peak period (07 am-10 am) 2. Morning off-peak period (10 am-12 pm) 3. Noon peak period (12 pm-2 pm) 4. Afternoon off-peak period (2 pm-5 pm) 5. Afternoon peak-period (5 pm-8 pm) 6. Evening period (8 pm-10 pm) 7. Weekend (Saturday or Sunday or both) A simple Visual Basic code written in MS Excel was operated on a portable computer to record travel times between major interchanges and at-grade intersections while traveling along 96 Research in Transport and Logistics 2009 Proceedings Do u University predetermined routes during the defined time periods above. A view of the Excel sheet created to facilitate the management of data collection is shown in Figure 3. FIGURE 3 Excel Sheet to Record Travel Times Along Predetermined Routes. 2.2. Bottlenecks There are two major bottlenecks in the middle of the city connecting one side to another; namely, Bogazici (Bosporus) and Fatih Sultan Mehmet (FSM) Bridges. Temporary bottlenecks created by incidents such lane drops or shifts due to road/interchange constructions and by accident occurrences were also included in the analysis. They are itemized as follows: 1. Major bottlenecks a. Bogazici (Bosporus) Bridge b. Fatih Sultan Mehmet (FSM) Bridge 2. Temporary bottlenecks a. Lane drop/shift due to construction i. new bridge construction on westbound Highway D100 (E-5) ii. reconstruction of Gulsuyu Interchange on eastbound Highway D100 (E-5) iii. repavement on Northbound Beykoz Shore Road (Minor Arterial) b. Accident occurrences i. on southbound Cekmekoy-Sarigazi Interchange (Minor Arterial) ii. on westbound Minibus (Bagdat) St (Minor Arterial) iii. on westbound Highway D100 (E-5) at Maltepe Interchange. 3. DATA ANALYSIS AND RESULTS 3.1. Major Bottlenecks Major bottlenecks studied in this paper are the two major bridges connecting both sides of the city; namely, Bogazici (Bosporus) and Fatih Sultan Mehmet (FSM) bridges. 3.1.1. Bogazici (Bosporus) Bridge Bogazici or Bosporus bridge also called 1st Bridge opened in 1973 creates two bottlenecks: one for the flow from Europe to Asia side (toll booths exist at the Asia side of the bridge), and the other for the flow from Asia to Europe side (toll booths do exist at neither side of the bridge) as seen in Figure 4. 97 Research in Transport and Logistics 2009 Proceedings Do u University FIGURE 4 Bogazici (Bosporus) Bridge-2x3 Lanes (the 1st bridge connecting both sides of the city opened in 1973). Speeds observed on Bogazici Bridge during different time periods are given in Table 1. It should be noted that the major flow on the bridge Europe to Asia occurs in the afternoon periods. During the morning peak period (7 am-10 am) the speed of upstream flow (77 km/hr) was only 9.4% lower than that of the downstream flow (85 km/hr). The speed on the bridge was fairly high as 69 km/hr. These measurements although descriptively fell into the morning peak-period, they were taken after 9:45 am, so close to the end of the morning peak-period. Thus, the travel time at the upstream section was only 10% higher than that of the downstream (travel time index or TTI =1.10, which is defined as the downstream speed divided by the speed on the section) and it would not be wrong if data are considered to be the morning off-peak data. However, during the afternoon off-peak period (2pm- 5pm) TTI at the upstream varied between 3.64 and 6.17. For the flow from Europe to Asia, the most critical time period was afternoon off-peak (2 pm-5pm) because travelers expecting high congestion after 5 pm tend to cross the bridge earlier. The second most highly traveled period was TABLE 1 Speeds Observed on Bogazici (Bosporus) Bridge During Different Time Periods Direction Time Period Section (No. of Ave. Speeds Travel Speed Travel Time observations) Measured Index* Index+ (1) (2) (3) (4) (5) (6) Morning (off-) Upstream (1) 77 0.91 1.10 Peak On the bridge (1) 69 0.81 1.23Downstream (1) 85 1 1 Afternoon off- Upstream (2) 12 ; 22 0.16 ; 0.28 6.17 ; 3.64 peak On the bridge (2) 65 ; 61 0.88 ; 0.76 1.14 ; 1.31Downstream (2) 74; 80 1 ; 1 1 ; 1 Upstream (2) 30 ; 24 0.47 ; 0.45 2.13 ; 2.21 Europe to Asia Afternoon Peak On the bridge (2) 30 ; 48 0.47 ; 0.91 2.13 ; 1.10 Downstream 2() 64 ; 53 1 ; 1 1 ; 1 Upstream (1) 15 0.21 4.73 Evening On the bridge (1) 43 0.61 1.65 Downstream (1) 71 1 1 Upstream (1) 92 0.99 1.01 Weekend On the bridge (1) 99 1.06 0.94 Downstream (1) 93 1 1 Upstream (3) 21 ; 21 ; 21 0.26;0.25;0.26 3.81 ; 4 ; 3.81 Morning Peak On the bridge (3) 53 ; 55 ; 53 0.66;55;0.66 1.51;55;1.51 Downstream (3) 80 ; 84 ; 80 1 ; 1 ; 1 1 ; 1 ; 1 Asia to Europe Upstream (2) 55 ; 58 0.58 ; 0.98 1.73 ; 1.02Noon Peak On the bridge (2) 57 ; 51 0.60 ; 0.86 1.67 ; 1.16 Downstream (2) 95 ; 59 1 ; 1 1 ; 1 Afternoon Peak Upstream (1) 55 0.58 1.73On the bridge (1) 40 0.42 2.38 98 Research in Transport and Logistics 2009 Proceedings Do u University Downstream (1) 95 1 1 Upstream (2) 48 ; 33 0.69 ; 0.52 1.46 ; 1.94 Weekend On the bridge (2) 53 ; 61 0.76 ; 0.95 1.32 ; 1.05 Downstream (2) 70 ; 64 1 ; 1 1 ; 1 *: Travel Speed Index equals to the speed on the section divided by the downstream speed. +: Travel Time Index equals to the downstream speed divided by the speed on the section. evening time (after 8 pm). Travelers definitely choose this bridge to cross the Bosporus during off- peak periods because it is the shorter path from Europe to Asia side of the city for the commuters whose origins or destinations are close to the route of the Highway D100 (E-5) which is connected to Bosporus Bridge. During afternoon peak period, TTI varied between 2.13 and 2.21. Travelers who expect high congestion during the peak period, they tend to use FSM Bridge to cross the Bosporus. For the flow from Asia to Europe side, the major flow occurs in the morning periods as expected the TTI of the upstream flow varied between 3.81 and 4. In the afternoon peak period, the bridge was more congested than the upstream section. 3.1.2. Fatih Sultan Mehmet (FSM) Bridge Fatih Sultan Mehmet or FSM Bridge also called 2nd Bridge creates two bottlenecks: one for the flow from Europe to Asia side (toll booths exist at the Europe side of the bridge), and the other for the flow from Asia to Europe side (toll booths do exist at neither side of the bridge) as seen in Figure 5. FIGURE 5 Fatih Sultan Mehmet (FSM) Bridge- 2x4 Lanes (the 2nd bridge connecting both sides of the city opened in 1989). Speeds observed on FSM Bridge during different time periods are given in Table 2. It should be noted that the major flow on the bridge Europe to Asia occurs in the afternoon peak period (5 pm-8 pm). There are five speed measurements during the period and TTI varied between 1.19 and 12.67. The highest two values (11.38 and 12.67) belong to the two days when heavy vehicles are unusually allowed to cross the bridge during that period. Other times heavy vehicles are held up by the police until the end of the peak-period (8 pm). For the flow from Asia to Europe side of the city, TTI at FSM Bridge (2.33) was lower than that at Bosporus Bridge (3.81-4) during the morning peak period as expected because Bosporus Bridge provides the shorter route to cross the Bosporus Strait and most travelers choose that one. 3.2. Temporary Bottlenecks 3.2.1. Lane Drop/Shift Due to Construction 3.2.1.1. New Bridge Construction on Westbound Highway D100 (E-5) A lane drop/shift 1300 m after Pendik Interchange on westbound Highway D100 (E-5) between Pendik and Kartal interchanges as seen in Figure 6 caused a tremendous decrease in travel speed. Speeds observed on the site are given in Table 3. 99 Research in Transport and Logistics 2009 Proceedings Do u University TABLE 2 Speeds Observed on FSM Bridge During Different Time Periods Direction Time Period Section (No. of Ave. Speeds Travel Speed Travel Time observations) Measured Index* Index+ (1) (2) (3) (4) (5) (6) Upstream (5) 69;25;75;8 ;6 0.82 ;0.74 ; 0.84 1.22 ; 1.36 ; 1.19 0.09 ; 0.08 11.38 ; 12.67 Europe to Asia Afternoon Peak On the bridge (5) 76;55;75;35;63 0.90 ;1.62; 0.84 1.11 ; 0.62 ; 1.19 0.38 ; 0.83 2.60 ; 1.21 Downstream (5) 84;34;89;91;76 1 ; 1 ; 1 ; 1 ; 1 1 ; 1 ; 1 ; 1 ; 1 Upstream (1) 24 0.43 2.33 Morning Peak On the bridge (1) 65 1.16 0.86 Downstream (1) 56 1 1 Upstream (1) 79 0.80 1.24 Noon Peak On the bridge (1) 82 0.83 1.21 Asia to Europe Downstream (1) 99 1 1 Afternoon off- Upstream (1) 111 1.14 0.87 peak On the bridge (1) 102 1.05 0.95Downstream (1) 97 1 1 Upstream (1) 108 0.84 1.19 Weekend On the bridge (1) 104 0.81 1.23 Downstream (1) 128 1 1 *: Travel Speed Index equals to the speed on the section divided by the downstream speed. +: Travel Time Index equals to the downstream speed divided by the speed on the section. These number denote the speed measurements between the toll booths and the bridge where the vehicles pass through 16 toll booths and then get fit into 4 lanes on the bridge. FIGURE 6 Temporary Bottleneck on Westbound Highway D100 (E- 5) due to New Bridge Construction Between Pendik and Kartal Interchanges. TABLE 3 Speeds Observed on Westbound Highway D100 (E-5) Between Pendik and Kartal Interchanges Time Period Section (No. of Ave. Speeds Travel Speed Travel Time observations) Measured (km/hr) Index* Index+ (1) (2) (3) (4) (5) Morning off-peak Upstream (1) 16 0.25 4Downstream (1) 64 1 1 *: Travel Speed Index equals to the speed on the section divided by the downstream speed. +: Travel Time Index equals to the downstream speed divided by the speed on the section. 100 Research in Transport and Logistics 2009 Proceedings Do u University The incident (capacity decrease due to lane drop/shift) of new bridge construction on westbound Highway D100 (E-5) resulted the Travel Speed and Travel Time Indexes at the upstream location as 0.25 and 4, respectively. This means while the speed is decreased to 16 km/hr, the travel time is increased by 4 times at the upstream. 3.2.1.2. Reconstruction of Gulsuyu Interchange on Eastbound Highway D100 (E-5) A lane drop/shift on eastbound Highway D100 (E-5) as seen in Figure 7 caused a significant decrease in travel speed. Speeds observed on the site are given in Table 4. FIGURE 7 Temporary Bottleneck at Gulsuyu Interchange on Eastbound Highway D100 (E-5). TABLE 4 Speeds Observed at Gulsuyu Interchange on Eastbound Highway D100 (E-5) Time Period Section (No. of Ave. Speeds Travel Speed Travel Time observations) Measured (km/hr) Index* Index+ (1) (2) (3) (4) (5) Evening Upstream (3) 17 0.25 4.06Downstream (3) 69 1 1 *: Travel Speed Index equals to the upstream speed divided by the downstream speed on the section. +: Travel Time Index equals to the downstream speed divided by the upstream speed on the section. The incident (capacity decrease due to the reconstruction of Gulsuyu Interchange) on westbound Highway D100 (E-5) resulted the Travel Speed and Travel Time Indexes at the upstream as 0.25 and 4.06, respectively. This means that while the speed is decreased to 17 km/hr, travel time along the 1700 m long section is increased by 4.06 times at the upstream of the incident location. 3.2.1.3. Repavement on Northbound Beykoz Shore Road (Minor Arterial) A lane drop on northbound Beykoz Shore Road caused a significant decrease in travel speed. Speeds observed on the site are given in Table 5. TABLE 5 Speeds Observed on Northbound Beykoz Shore Road (Minor Arterial) Time Period Section (No. of Ave. Speeds Travel Speed Travel Time observations) Measured Index* Index+ (1) (2) (km/hr) (4) (5) (3) Noon-peak Upstream (1) 11 0.26 3.82Downstream (2) 42 1 1 *: Travel Speed Index equals to the upstream speed divided by the downstream speed on the section. +: Travel Time Index equals to the downstream speed divided by the upstream speed on the section. 101 Research in Transport and Logistics 2009 Proceedings Do u University The incident (capacity decrease due to lane drop) on northbound Beykoz Shore Road, which is a minor arterial, resulted the Travel Speed and Travel Time Indexes at the upstream as 0.26 and 3.82, respectively. This means while the speed is decreased to 11 km/hr, the travel time is increased by 3.82 times at the upstream of the incident location. 3.2.2. Accident Occurrences 3.2.2.1. Southbound Cekmekoy-Sarigazi Interchange (Minor Arterial) An accident on southbound Cekmekoy-Sarigazi Interchange resulted a decrease of 76% in travel speed or an increase of 317% in travel time. Speeds observed on the site are given in Table 6. TABLE 6 Speeds Observed During Accident on Southbound Cekmekoy-Sarigazi Interchange Time Period Section (No. of Ave. Speeds Travel Speed Travel Time observations) Measured (km/hr) Index* Index+ (1) (2) (3) (4) (5) Evening Upstream (4) 18 0.24 4.20Downstream (4) 75 1 1 *: Travel Speed Index equals to the upstream speed divided by the downstream speed on the section. +: Travel Time Index equals to the downstream speed divided by the upstream speed on the section. The incident (accident occurrence) on southbound Cekmekoy-Sarigazi Interchange, which is a rural arterial, resulted the Travel Speed and Travel Time Indexes at the upstream as 0.24 and 4.20, respectively. This means while the speed is decreased to 18 km/hr, the travel time is increased by 4.20 times at the upstream of the incident location. 3.2.2.2. Westbound Minibus (Bagdat) St (Minor Arterial) An accident on westbound Minibus St (Bagdat St) resulted a decrease of 79.3% in travel speed or an increase of 383% in travel time. Speeds observed on the site are given in Table 7 as follows: TABLE 7 Speeds Observed During Accident on Westbound Minibus (Bagdat) St Time Period Section (No. of Ave. Speeds Travel Speed Travel Time observations) Measured Index* Index+ (1) (2) (km/hr) (4) (5) (3) Evening Upstream (2) 6 0.21 4.83Downstream (4) 29 1 1 *: Travel Speed Index equals to the upstream speed divided by the downstream speed on the section. +: Travel Time Index equals to the downstream speed divided by the upstream speed on the section. The incident (accident occurrence) on westbound Minibus St (Bagdat St), which is a secondary urban arterial, resulted the Travel Speed and Travel Time Indexes at the upstream as 0.21 and 4.83, respectively. This means while the speed is decreased to 6 km/hr, the travel time is increased by 4.83 times at the upstream of the incident location. 3.2.2.3. Westbound Highway D100 (E-5) at Maltepe Interchange An accident on westbound Highway D100 (E-5) located 800 m after Maltepe Interchange resulted a decrease of 94.4% in travel speed or an increase of 1680% in travel time. Speeds observed on the site are given in Table 8. 102 Research in Transport and Logistics 2009 Proceedings Do u University TABLE 8 Speeds Observed During Accident on Westbound Highway D100 (E-5) 800m After Maltepe Interchange Time Period Section (No. of Ave. Speeds Travel Speed Travel Time observations) Measured (km/hr) Index* Index+ (1) (2) (3) (4) (5) Evening Upstream (2) 5 0.056 17.8Downstream (2) 89 1 1 *: Travel Speed Index equals to the upstream speed divided by the downstream speed on the section. +: Travel Time Index equals to the downstream speed divided by the upstream speed on the section. The incident (accident occurrence) on westbound Highway D100 (E-5) at Maltepe Intersection resulted the Travel Speed and Travel Time Indexes at the upstream as 0.056 and 17.8, respectively. This means while the speed is decreasing to 5 km/hr, travel time along the 900 m long section is increased by 17.8 times at the upstream of the incident location. 4. SUMMARY AND CONCLUSIONS This research paper presents the findings of a study investigated the effect of bottlenecks on urban traffic congestion in the city of Istanbul, Turkey. Travel time data collected on- and off-peak times of day using the maximum-car method on major highways and arterials of the city. Bottlenecks are defined as the roadway segments where the number of lanes serving the traffic suddenly decreases and as the toll booths on the access-controlled road segments where vehicles have to slow down (at automated passing, OGS, booths) or come to a complete stop (at cash or magnetic card reader booths). Besides temporary bottlenecks created by incidents such lane drops or shifts due to road or infrastructure constructions and by accident occurrences were also included in this analysis. Main conclusions from the analysis of travel time data and travel time indexes (TTI) collected during different time periods along major highways and main arterials are as follows: The worst travel time values are observed at accident occurrences (TTI is varied between 4.20 and 17.8) since they are unexpected incidents. The worst case was the accident occurrence on Highway D100 (E-5). One of the main reasons for observing such a long delay was the poor “accident management” as well as an illegal act of drivers for using the shoulder as a travel lane. The accident management is generally poor in terms of the arrival of emergency vehicles as well as the police in Istanbul because an average of 642 property damage only (PDO) accidents occur everyday in Istanbul. However, as of April 1st of 2008 drivers involved in a PDO accident do not have to wait for the police to arrive at the accident scene and do the report if they agree upon the occurrence of the accident. This new application was questioned by many with a great suspicion. However, the results after the first month of the application showed that it was accepted by many drivers and the number of cases in which drivers were agreed upon the occurrence of the accident was 66% in Istanbul, %48 in Ankara, and %50 in Izmir, the three biggest cities of Turkey in the order from the largest to the smallest. Capacity decreases due to road constructions was also as bad as the road blockages due to accident occurrences (TTI is varied between 3.82 and 4). Travel time data as well as the TTI index showed that Bogazici (Bosporus) Bridge was the fist choice of commuters to cross the Bosporus since it provides the shorter route compared to Fatih Sultan Mehmet (FSM) Bridge. The worst time period to cross the Bosporus over Bogazici Bridge from Europe to Asia side of the city was afternoon off-peak with the TTI between 3.64 and 6.17. From Asia to Europe side, the worst time was morning peak period with the TTI varied between 3.81 and 4. 103 Research in Transport and Logistics 2009 Proceedings Do u University While crossing the Bosporus via FSM Bridge during afternoon peak period, it was seen that heavy vehicles created big troubles that increased the TTI to the highest ever (11.38 and 12.67). For weekend trips, travelers always choose Bogazici Bridge over FSM Bridge to cross the Bosporus. Remedies for such long delays along major highways and main arterials are as follows: For crossing the Bosporus, more crossing alternatives must be offered to travelers. One less costly one regarding the infrastructure cost is providing more capacity for ferry boat routes. Second better as well as more efficient alternative for crossing the Bosporus is to build a new highway bridge accommodating a high capacity public transportation line. For other incidents causing longer delays to urban travelers, one good remedy is doing construction works as fast as possible by working 24 hours a day and doing the maintenance works at nights. After talking about the magnitude of the delays observed along the highways and main arterials of the city, we will comment on how to handle the solution of corridor improvement without further disturbing the traffic during traffic incidents as follows: As a number one action, the poor “accident management” must be replaced by a better one; such as dynamic message signs located at some strategic locations over the major highway network links must convey information on traffic incidents like time of occurrences, duration, speed and travel time conditions upstream of the incidents so that drivers be diverted to alternative routes. Such a system can suggest or command alternative routes to drivers if they stuck in traffic due to traffic incidents in order to minimize delays especially along major routes. Advanced (by VMS, internet, TV, cellular phones, billboards, etc.) or conventional (by radio) traveler information systems can help drivers beware about the incidents and take actions such as postponing trips, taking alternative modes, and so on. REFERENCES [1] Ergün, G., Implementation of Transportation and Congestion Management Strategies: Strategic Planning Studies Department, 2005. [2] Schrank, D., and T. Lomax. The 2004 Urban Mobility Report. Texas Transportation Institute and the Texas A&M University System, 2004. [3] Schrank, D., and T. Lomax. 2003 Annual Urban Mobility Report. Texas Transportation Institute and the Texas A&M University System, 2003. [4] Cambridge Systematics. Traffic Congestion and Reliability: Linking Solutions to Problems. 2004. 104 Research in Transport and Logistics 2009 Proceedings Do u University ESTABLISHING AN EFFECTIVE TRAINING MODULE FOR IMDG CODE IN MET INSTITUTIONS Kadir CICEK1, Metin CELIK 2 Abstract The International Maritime Dangerous Goods (IMDG) Code is a mandatory framework for all aspects of handling dangerous goods and pollutants in maritime transportation. Due to its significance, the IMDG Code has also recommended additional training to seafarers who involve carriage of dangerous goods onboard merchant ships. The main aim of this paper is to design an effective training module towards IMDG Code in Maritime Education and Training (MET) institutions. Moreover, this paper explores the links between shipping casualties associated with dangerous goods and training requirements. The outcomes of this paper contribute IMDG training contents for seafarers and undergraduate/graduate maritime students. Keywords IMDG Code, Maritime transport, Maritime training, MET institutions INTRODUCTION The transportation of dangerous goods by air, marine, rail, and road is a highly complex issue which requires maintaining safety and environmental related precautions. Besides the enormous numbers of researches on evaluating and controlling the hazards [1-5], the ongoing efforts about developing regulatory framework have continued to ensure safer carriage of dangerous goods using different transportation modes. Furthermore, advance-training requirements on dangerous good transportation are appeared in order to reduce the potential role of relevant operators in casualties. For example, the International Maritime Dangerous Goods (IMDG) Code [6] regulates the all aspects of handling dangerous goods and pollutants in maritime transportation. Principally, it is intended to provide for the safe transportation of hazardous materials by ships, protect operators and to prevent marine pollution. Due to its significance, the IMDG Code has also recommended additional training to seafarers who involve carriage of dangerous goods onboard merchant ships. This paper targets to develop a training module for the IMDG Code to hold in Maritime Education and Training (MET) institutions. It follows the feedbacks from shipping transportation industry to underline the interruptions in carriage of dangerous goods by merchant ships. TRANSPORTATION OF DANGEROUS GOOD ONBOARD SHIPS Classification of Dangerous Goods Maritime laws and regulations require using and handling of hazardous goods in respect to the material specifications. For example, one set of requirements may apply to their use in the workplace while different requirements may apply to spill response, sale for consumer use, or transportation. The most widely applied regulatory scheme about the transportation of dangerous goods is so called United Nations Committee of Experts on the Transport of Dangerous Goods (UNCOETDG) issues Model Regulations on the Transportation of Dangerous Goods. In this Model Regulation for all types of transport (sea, air, rail, road and inland waterways) the classification of dangerous goods, by type of risk involved, has been drawn up [7-9]. Most regional and national regulatory schemes for hazardous materials are harmonized to a greater or lesser degree with UN Model Regulation. For instance, the International Civil Aviation Organization has developed regulations for air transport of hazardous materials that are based upon the UN Model but modified to accommodate unique aspects of air transport [7]. Similarly, the International Maritime Organization (IMO) has developed the Dangerous Goods Regulations for transportation on the high seas. Table 1 illustrates the classification of dangerous goods according to IMO Dangerous Goods Regulations which includes formal UNCOETDG’s classification. 1 Kadir Cicek, Istanbul Technical University, Faculty of Maritime, Marine Engineering Department, Tuzla, Istanbul, Turkiye, cicekk@itu.edu.tr 2 Metin Celik, Istanbul Technical University, Faculty of Maritime, Maritime Transportation & Management Engineering Department, Tuzla, Istanbul, Turkiye, celikmet@itu.edu.tr 105 Research in Transport and Logistics 2009 Proceedings Do u University TABLE 1 Classification of Dangerous Goods according to IMO Dangerous Goods Regulations Class 1 Explosives Division 1.1 Substances and articles which have a mass explosion hazard Division 1.2 Substances and articles which have a projection hazard butnot a mass explosion hazard Substances and articles which have a fire hazard and either a Division 1.3 minor blast hazard or a minor projection hazard or both, but not a mass explosion hazard Division 1.4 Substances and articles which present no significant hazard Division 1.5 Very insensitive substances which have a mass explosionhazard Division 1.6 Extremely insensitive articles which do not have a massexplosion hazard Class 2 Gases Division 2.1 Flammable gases Division 2.2 Non-flammable, non-toxic gases Division 2.3 Toxic gases Class 3 Flammable Liquids Class 4 Flammable solids; substances liable to spontaneous combustion; substanceswhich, on contact with water, emit flammable gases Division 4.1 Flammable solids, self-reactive substances and soliddesensitized explosives Division 4.2 Substances liable to spontaneous combustion Division 4.3 Substances which in contact with water emit flammableGases Class 5 Oxidizing substances and organic peroxides Division 5.1 Oxidizing substances Division 5.2 Organic peroxides Class 6 Toxic and infectious substances Division 5.1 Toxic substances Division 5.1 Infectious substances Class 7 Radioactive material Class 8 Corrosive substances Class 9 Miscellaneous dangerous substances and articles* MHB Materials hazardous only in bulk** * Marine pollutants which are not of an otherwise dangerous nature are listed in class 9 ** The regulations for materials hazardous only in bulk are not applicable to these materials when they are carried in closed freight containers, however, many precautions may have to be observed Links to Shipping Accidents Shipping accident is a term generally used for any accident results in financial loss, in either life or property or both [11]. The reasons for shipping accidents are many and complex. In addition, several crucial causes play a role in occurrence of shipping accidents. Natural conditions, technical failures, route conditions, ship-related factors, human errors, cargo-related issues are the mostly highlighted factors [10]. Any shipping accident, whatever in nature, is every seafarer’s nightmare but the effects of ship accident carry dangreous goods on nature and people is much greater. Usually, any accident might have more than one cause. Nevertheless, investigations and statistical analyses on the basis of the main causal trends explicitly reveal that human errors continue to be the major cause for all shipping accidents. For example, on 14 April 2007, the UK registered product tanker, was carried dangerous good, Motor Tanker (MT) Audacity was involved in a collision with the Panama registered general cargo ship Motor Vessel Leonis, in very poor visibility, in the precautionary area at the entrance to the River Humber. Both vessels sustained damage to their bows. Fortunately there were no injuries and no pollution was occurred [12]. The communication between two ships 106 Research in Transport and Logistics 2009 Proceedings Do u University involved was unclear and prone to misunderstanding, and the use of standard marine phrases was not practised, the pilots and bridge teams, on both vessels, did not make a full assessment of the risk of collision and VTS procedures for managing traffic in the precautionary area were insufficient are the reasons of this accident [12]. One other accident was occurred on 10 December 2006, chemical and petroleum product tanker, carry clean petroleum product, Prospero was approaching No. 2 Jetty, of the SemLogistics terminal, Milford Haven, when the master suddenly and without warning lost control of the vessel’s podded propulsion system [13]. This caused the vessel to make contact with the jetty’s infrastructure, resulting in material damage to both the jetty and the vessel before control was regained. The lack of in-house maintenance procedures, inadequate system knowledge by ship’s officers and shore staff, and weak safety management system (SMS) and onboard system documentation, overlaid on a propulsion system for which, when introduced, no dedicated technical standards existed, resulted in a vessel whose resilience to defects and emergencies was significantly weakened are the reasons of this accident. In both two accidents the reasons, especially regarding with the human errors and technical failures based on lack of maintenance [13]. The common viewpoint for prevention of shipping accidents associated with dangerous cargo is to enhance the competency and knowledge of seafarers especially about the key aspects of routine operations (i.e. cargo handling, tank inspection, gas-freeing, tank cleaning, tank purging, etc.) in operational level satisfactorily. REVIEW OF IMDG CODE REQUIREMENTS IMDG Code is accepted as an international guide to the transport of dangerous goods by sea and is recommended to governments for adoption or for use as the basis for national regulations. It is intended for use not only by the mariner but also by all those involved in industries and services connected with shipping, and contains advice on terminology, packaging, labeling, placarding, markings, stowage, segregation, handling, and emergency response. The code is updated and maintained by the IMO every 2 years. IMDG Code is divided into seven parts and includes index, appendices, and supplement; Part 1: General provisions, definitions, training Part 2: Classification Part 3: Dangerous good list and limited quantities exceptions Part 4: Packing and tank provisions Part 5: Consignment procedures Part 6: Construction and testing of packagings, IBC, large packagings, portable tanks and road tank vehicles Part 7: Provisions concerning transport operations DESIGN OF IMDG TRAINING MODULE This section describes a design concept for the IMDG training is held in MET institutions. Training design has enough detail for evaluating training module in the light of criteria established. IMDG training module consists of three levels which are correspondingly explained in the further subsections. Level I: Theoretical Aspects In this level, each trainee attends some courses, which provide familiarity with the general provisions of IMDG Code. This level include a description of the classes of dangerous goods, labeling, marking, packing, stowage, segregation, compatibility provisions and a description of available emergency response applications, precautions and documents and provides an overview of the shipping accidents to identify the contributing factors related to dangerous cargo. The feedback is then used to modify the training module that carry out. In addition, this level gives information about related codes and publications for function-specific training. Level II: Rule-based Shipboard Implementations The successful application of IMDG Code is greatly dependent on the appreciation by all trainees concerned with the risks involved and on a detailed understanding of the regulations. This can be achieved by properly planned and maintained rule-based implementation concerned with the transport of dangerous goods. The main aim of the rule-based implementation is to give detailed training concerning specific dangerous goods transport provisions which are applicable to the function that person performs and to constitute a form 107 Research in Transport and Logistics 2009 Proceedings Do u University of thinking logical area. At that point in this level, the regulatory basis of IMDG Code is introduced, regulations concerning with transport of dangerous goods are extracted and analyzed and after that try to gain effective problem solving and decision making abilities in limited times to trainees with implement correct rules. Level III: Health and Safety Precautions in Operational Level The health and safety precautions in operational level are another aspect within the developed module. In this level, the trainees receive training on emergency response procedures and information and correct using of them, methods and procedures for accident avoidance, such as correct marking, labeling, placarding requirements and appropriate methods of stowage of dangerous goods, general dangers presented by various classes of dangerous goods and immediate procedures to be followed in the event of an unintentional release of dangerous goods including any emergency response procedures for which the person is responsible and personal protection procedures to be followed. CONCLUSION AND DISCUSSIONS This research develops a fundamental framework of training module on IMDG Code. It mainly explores the shortfalls in IMDG Code practice in marine transportation to eagerly link up with the proposed training module. Based on the industry based survey of carriage of dangerous goods by ships, the urgent needs for the training of crew members are seemed especially for the following topics: (1) Preparing dangerous goods loading/stowage plan, (2) Load/unload dangerous goods into/from ships, (3) Handle dangerous goods in transport. To satisfy these needs, training modules in three levels are then structured in broad form. The further research proposal can be addressed as the design a detailed syllabus for the developed IDGM training module. REFERENCES [1] Fabiano,B, Currò, F. Palazzi, E., Pastorino R., 2002, “A framework for risk assessment and decision-making strategies in dangerous good transportation”, Journal of Hazardous Materials, Volume 93, Issue 1,pp.1-15. [2] Jones J.C., (2005). Commentary on recent proposals for classification of dangerous goods for transportation, Journal of Hazardous Materials, Volume 122, Issues 1-2, 30 June, pp.61. [3] Gheorghe, A.V. Birchmeier, J., Vamanu, D. Papazoglou, I., Kröger, W. (2005). Comprehensive risk assessment for rail transportation of dangerous goods: a validated platform for decision support, Reliability Engineering & System Safety, Volume 88, Issue 3, June, pp.247-272. [4] Chang, Y-H., Yeh, C-H, Liu, Y-L. (2006). Prioritizing management issues of moving dangerous goods by air transport, Journal of Air Transport Management, Volume 12, Issue 4, July, Pages 191-196. [5] Berman, O., Verter, V., Kara, B.Y. (2007). Designing emergency response networks for hazardous materials transportation, Computers & Operations Research, Volume 34, Issue 5, May, pp.1374-1388. [6] Uehara,Y., Nakajima T. (1985). Proposal of a new test method for the classification of oxidizing substances, Journal of Hazardous Materials, Volume 10, Issue 1, February 1985, pp.89-106. [7] http://en.wikipedia.org/wiki/Hazardous_material [8] http://www.searates.com/reference/imo/ [9] United Nations Committee of Experts on the Transport of Dangerous Goods, 2007, “UN Recommendations on the Transport of Dangerous Goods. Model Regulations”, Fifteenth revised edition. [10] Akten, N., 2006, “Shipping accident: a serious threat for marine environment”, Journal of Black Sea/Mediterranean Environment, Volume 12, pp. 269-304. [11] Istanbul, Ekim 1982, 1. [12] Marine Accident Investigation Branch, 2008, “Report on investigation of the collision between Audacity and Leonis” [13] Marine Accident Investigation Branch, 2007, “Report on investigation of the loss of control of product tanker Prospero” 108 Research in Transport and Logistics 2009 Proceedings Do u University AN INVESTMENT DECISION AID PROPOSAL TOWARDS CHOICE OF CONTAINER TERMINAL OPERATING SYSTEMS BASED ON INFORMATION AXIOMS Metin CELIK1, Selcuk CEBI 2 Abstract Choice of an operating system is a critical decision-making process in order to manage optimum layouts for container terminals. The recent marine technology allows utilization of tractor/chassis system, direct straddle carrier system, relay straddle carrier system, yard gantry-crane system, and front-end loader system. This paper proposes a methodology based on information axioms to compare strengths and weakness of terminal operating systems in respect to critical attributes quantitatively. The proposed methodology enables decision aid for both redesigning activities in existing terminals and forthcoming projects of new maritime enterprises. Keywords Container handling equipment, Container terminal layouts, Information axioms TECHNICAL BACKGROUND Container terminals, the complex investment projects, require great levels of technical infrastructure and equipment [1]. On the other hand, the reliability and efficiency of the operations in container terminals has vital importance to ensure the sustainability of transportation network [2]. The significance of container terminals have been appreciated by the relevant practitioners to maintain the ongoing efforts on optimization of various decision problems such as berth allocation [3], quay crane scheduling [4], yard operations [5], and transfer operations [6]. In modelling of mostly optimization problems [7], the terminal infrastructure and operating system are recognized as common critical constraints. Among these issues, choice of operating systems is a critical decision-making process in order to manage optimum layouts for container terminals. This paper develops a methodological framework to compare the current terminal operating systems under multiple criteria. It aims to identify the most suitable operating system based on the service requirements of a container terminal. The paper is organized as follows: Section 2 introduces the utilization of information axioms in decision-making. Section 3 states the proposed model and it demonstrates the quantified findings on different operating systems. At the end of the paper, potential contributions of this research in both academic manner and industrial level are discussed. RESEARCH METHODOLOGY: DECISION-MAKING USING INFORMATION AXIOMS An information axiom (IA) is the second axiom of an axiomatic design (AD) proposed by Suh (1990). It establishes a scientific basis to improve design activities based on logical and rational thought process [8]. The primarily goal of an AD approach is to provide reference points to improve design stages [9]. An AD states that the independence of functional requirements (FRs) must always be maintained, where FRs are defined as the minimum set of independent requirements that characterizes the design goals [10]. Furthermore, an IA states that the design having the smallest information content is the best design among those designs that satisfy the independence axiom [9]. At this insight, an IA is recognized as a conventional method, which facilitates the selection of proper alternative that has minimum information content. Recently, the IA has been adapted to solve different multiple criteria decision-making problem cases in literature [11-18]. The theoretical concept of an IA deals with minimizing the information content (I) of the design. It targets to reach minimum value of I which is the best design alternative among the feasible solutions that satisfy the independence axiom. In IA application, a multiple criteria selection process, which states an alternative has the highest probability of FR success, it is so-called as the best design [10]. Apart from the traditional applications of an IA, this paper follows an ideal FR (IFR) [19] to rank the terminal operating systems. 1 Metin Celik, Istanbul Technical University, Faculty of Maritime, Maritime Transportation & Management Engineering Department, Tuzla, Istanbul, Turkiye, celikmet@itu.edu.tr 2 Selcuk Cebi, Istanbul Technical University, Faculty of Management, Industrial Engineering Department, Macka, Istanbul, Turkiye, cebisecuk@gmail.com 109 Research in Transport and Logistics 2009 Proceedings Do u University The mathematical background of an IA approach can be introduced as follows: If the probability of success for given FR is p, the I is calculated using Eq.1: I i log 1 2 (1) pi If there is more than one FR, the I is calculated as follows: I system log 2 P(m) (2) m I system log 2 ( Pi ) (3) i 1 I m msystem log 2 Pi log 2 (1/ Pi ) (4) i 1 i 1 This paper developed an investment decision aid approach to assist the relevant maritime enterprises in terminal operating system choice. The proposed decision aid system also involves a group aggregation, which is also accomplished using a Borda score method. It gives consistent ranking of terminal operating system alternatives. The proposed decision aid towards choice of terminal operating system in accordance with the expectations of maritime enterprises is given in Figure 1. Step 1. Specify the potential terminal operating system alternatives Step 2. Define the evaluation attributes and set IFR values for each attribute respectively Step 3. Assign linguistic judgments on each alternative The judgments of an expert # 1 The judgments of an expert # 2 Step 3.Transform linguistic judgments into TFNs Step 4. Compute an I value for each alternative Alternative rank order based on Alternative rank order based on expert # 1 judgments expert # 2 judgments Step 5. Apply Borda score approach to obtain overall ranking of terminal operating systems Step 6. Choice of suitable terminal operating FIGURE 1 Proposed investment decision aid for container terminals 110 Research in Transport and Logistics 2009 Proceedings Do u University DEMONSTRATION OF PROPOSED MODEL This paper follows the research methodology based on fuzzy information axioms to model the operating system selection problem for container terminals. Problem Statement The container handling system selection problem is particularly addressed in literature [20]. In this paper, it is determined to compare the similar alternatives. Therefore, the proposed model incorporates the five different container terminal operating systems as follows: (A1) Tractor/chassis system (A2) Straddle carrier direct system (A3) Straddle carrier relay system (A4) Yard gantry system (A5) Front-end loader system On the other hand, the alternatives are compared based on following critical attributes: advantages in land utilization (b1), automation potential (b2), safety and reliability in cargo handling operations (b3), capital cost (c1), operating and maintenance cost (c2), manning level (c3). Assigning Fuzzy Linguistic Judgments on FRs & Alternatives The decision-making methodology with information axioms requires assigning of judgements about expectations from alternative operating systems in terms of identifying the IFR values. The linguistic scale and corresponding TFNs are given in Table 1. TABLE 1 Linguistic scale Linguistic Term Abbreviation TFN Very Low VL (0.0,0.0,0.3) Low L (0.0,0.3,0.5) Medium M (0.3,0.5,0.7) High H (0.5,0.7,1.0) Very High VH (0.7,1.0,1.0) Considering the evaluation attributes, Table 2 illustrates the IFR values in TFNs form. These judgements represent the expectations of an enterprise from container terminal operating systems. Hereby, to demonstrate the proposed decision aid, the IFR values are randomly assigned. Those values can be modified in respect to the different projections of maritime enterprises. TABLE 2 Definitions of IFR values on attributes in TFNs form Functional requirements Evaluation attributes Cost/Benefit Linguistic terms TFNs Advantages in land utilization (b1) Benefit At least (0.5,1.0,1.0) Automation potential (b2) Benefit At least (0.3,1.0,1.0) Safety and reliability in cargo handling operations (b3) Benefit At least (0.5,1.0,1.0) Capital cost (c1) Cost M At most (0.0,0.0,0.7) Operating and maintenance cost (c2) Cost At most (0.0,0.0,0.7) Manning level (c3) Cost At most (0.0,0.0,0.7) Furthermore, the judgements of the two different experts on alternative operating systems are given in Table 3 and Table 4. In this research, the expert profile mainly consists of professional managers from private container ports in Turkish maritime industry. 111 Research in Transport and Logistics 2009 Proceedings Do u University TABLE 3 Judgements of Expert #1 on alternative operating systems Evaluation attributes Alternatives b1 b2 b3 c1 c2 c3 (A1) Tractor/chassis system M L H L VL VL (A2) Straddle carrier direct system H H H H H M (A3) Straddle carrier relay system M M H M M L (A4) Yard gantry system VH H H H H H (A5) Front-end loader system VH H H H H H TABLE 4 Judgements of Expert #2 on alternative operating systems Evaluation attributes Alternatives b1 b2 b3 c1 c2 c3 (A1) Tractor/chassis system M M H L L VL (A2) Straddle carrier direct system M VH M H M M (A3) Straddle carrier relay system M VH M H M M (A4) Yard gantry system VH VH VH H H M (A5) Front-end loader system H H M M M M Findings The information content values are computed based on decision-making algorithm with fuzzy information axioms. To illustrate this algorithm, a sample calculation is provided for judgements of Expert #2 on straddle carrier direct system (A2) with respect to advantages in land utilization (b1) as follows: FIGURE 1. Judgements of Expert #1 for straddle carrier direct system with respect to land utilization (b1) Ii log 1 2 pi I11 log2 0.678(x2 x1 ) 1 ( 1 2 )(x2 x1) Finally, Table 4 and 5 represents the total I values for each alternative operating system with respect to the assessments of experts. According to the findings, the rankings of the alternatives are as follow; A1>A3> A2 A4=A5 for Expert # 1, A5>A1>A4>A3 >A2 for Expert # 2 and A1> A5>A3>A4> A2 with respect to Borda function. There is a slight difference between the ranking orders of each expert. If we use the additive procedure for the total IA values the rank also is found as A1> A5>A3>A4> A2. Since the ultimate goal of the decision aid is to offer an operating system for container terminal, Tractor-trailer system is considered as the most feasible operating system alternative for this enterprise. However, the rank reversals can be possible if the preferences of maritime enterprises are changed in terms of modifying the IFR values. 112 Research in Transport and Logistics 2009 Proceedings Do u University TABLE 4 Calculated information content values on alternative operating systems according to Expert 1 Information contents Alternatives b1 b2 b3 c1 c2 c3 Total (A1) Tractor-trailer system 2.809 3.491 0.678 0.286 0 0 7.264 (A2) Straddle carrier direct system 0.678 0.286 0.678 3.491 3.491 1.169 9.793 (A3) Straddle carrier relay system 2.809 1.169 0.678 1.169 1.169 0.286 7.28 (A4) Yard gantry system 0 0.286 0.678 3.491 3.491 3.491 11.44 (A5) Front-end loader system 0 0.286 0.678 3.491 3.491 3.491 11.44 TABLE 5 Calculated information content values on alternative operating systems according to Expert 2 Information contents Alternatives b1 b2 b3 c1 c2 c3 Total (A1) Tractor-trailer system 2.809 1.184 0.678 0.286 0.286 0 5.243 (A2) Straddle carrier direct system 2.809 0 2.809 3.491 1.169 1.169 11.45 (A3) Straddle carrier relay system 2.809 0 2.809 3.491 1.169 1.169 11.45 (A4) Yard gantry system 0 0 0 3.491 3.491 1.169 8.151 (A5) Front-end loader system 0.678 0.286 0.678 1.169 1.169 1.169 5.149 CONCLUSION AND DISCUSSIONS This paper developed an investment decision aid towards a techno-commercial problem on container terminal operating system selection. The advantage of the proposed methodology based on IA is seemed especially definition of IFR values to express the expectations and preferences of an enterprise from terminal operating systems. Therefore, utilization of this approach ensures a great flexibility to relevant decision makers in order to choice an adequate operating system in respect to the strategic and operational priorities of maritime enterprises in container terminal investment projects. The changes in IFR preferences can slightly influence the ranks of alternative operating systems. Hence, the proposed decision aid ensures consistent solutions for the different projections belong to enterprises in maritime sector. The model outcomes also aid to redesign activities of current operating systems in order to create a combined operating system in container terminals. ACKNOWLEDGEMENT The authors are grateful to Captain Alparslan TAVAS and Captain Hakan DENIZKUSU from operation department of KUMPORT for their technical contributions to this research. REFERENCES 1. Allahviranloo, M., Afandizadeh, S., 2008. Investment optimization on port's development by fuzzy integer programming European Journal of Operational Research, 186 (1), 423-434. 2. -Jentsch, D., 2002, Transport Policies for the Euro-Mediterranean Free-Trade Area: An Agenda for Multimodal Transport Reform in the Southern Mediterranean , World Bank Technical Paper, 527. 3. Wang, F. and Lim, A. 2007, A stochastic beam search for the berth allocation problem , Decision Support Systems 42: 2186-2196. 4. Sammarra, M., Cordeau, J. F., Laporte, G. and Monaco, M. F.2007, A tabu search heuristic for the quay crane scheduling problem , Journal of Scheduling 10: 327 336. 5. Cordeau, J. F., Gaudioso, M., Laporte, G. and Moccia, L., 2007, The service allocation problem at the Gioia Tauro maritime terminal , European Journal of Operational Research, 176: 1167 1184. 6. Lee, Y. and Hsu, N. Y. (2007). An optimization model for the container pre-marshalling problem, Computers and Operations Research 34: 3295 3313. 113 Research in Transport and Logistics 2009 Proceedings Do u University 7. Steenken, D., Voss, S. and Stahlbock, R. Container terminal operation and operations research - a classification and literature review, OR Spectrum 26: 3 49. 8. Suh, N.P. 1990, The Principles of Design , Oxford University Press Inc., NY. 9. Suh, N.P. 2001, Axiomatic Design: Advances and Applications , Oxford University Press, NY. 10. Suh, N.P. 2005, Complexity Theory and Applications , Oxford University Press, NY 11. Kulak, O. and Kahraman, C., 2005a, Fuzzy multi-attribute selection among transportation companies using axiomatic design and analytic hierarchy process Information Sciences, 170, 191-210. 12. Kulak, O. and Kahraman, C., 2005b, Multi-attribute comparison of advanced manufacturing systems using fuzzy vs. crisp axiomatic design approach , International Journal of Production Economics, 95, 415-424. 13. Kulak, O., Durmusoglu, M.B., and Kahraman, C., 2005, Fuzzy multi-attribute equipment selection based on information axiom , Journal of Materials Processing Technology, 169, 337-345. 14. Kulak, O., 2005, A decision support system for fuzzy multi- Expert Systems with Applications, 29( 2),310-319. 15. Coelho, A.M. and Moura J.F., 2007 Axiomatic design as support for decision-making in a design for manufacturing context: A case study , International Journal of Production Economics 109, 81 89 16. Celik, M., Kahraman, C., Cebi, S., and Er, I. D., Fuzzy Axiomatic Design-Based Performance Evaluation Model for Docking Facilities in Shipbuilding Industry: The Case of Turkish Shipyards , Expert Systems with Applications, (in pres). 17. Celik, M., Cebi, S., Kahraman, C., and Er, I. D, Application of axiomatic design and TOPSIS methodologies under fuzzy environment for proposing competitive strategies on Turkish container ports in maritime transportation network , Expert Systems with Applications, (in pres). 18. Celik, M., Cebi, S., Kahraman, C., and Er, I. D, An Integrated Fuzzy QFD Model Proposal on Routing of Shipping Investment Decisions in Crude Oil Tanker Market Expert Systems with Applications, (in pres). 19. Kahraman, C. and Cebi, S., A new multi-attribute decision making method: Hierarchical fuzzy axiomatic design , Expert Systems with Applications, (in pres). 20. Ahmad, M.Z., Abdul Kader, A. S.,Ahmat, A. N., Idris, J. 2006, A fuzzy application on a development planning model for a container terminal Jurnal Teknologi Siri A, 45 (A). pp. 13-29. 114 Chapter 6 Evaluation of Public Policies, Network Models and Environment Research in Transport and Logistics 2009 Proceedings Do u University POSSIBILISTIC LINEAR PROGRAMMING APPROACH FOR STRATEGIC RESOURCE PLANNING Özgür KABAK1, 2 Abstract Strategic resource planning decisions are very important for the medium and long term planning success of manufacturing companies. The inputs of the resource planning models face environmental and system uncertainties. In this paper, a fuzzy logic- based modeling is proposed to deal with those uncertainties For this purpose, a possibilistic linear programming (PLP) model is used to make strategic resource planning decisions using fuzzy demand forecasts as well as other inputs such as costs, yield rates etc. The objective of proposed PLP is to maximize the total profit of the enterprise. The model is designed to guide the supply chain managers to decide on strategic problems such as “which resources are used for which product”, “which resources are outsourced”, “which products should be produced/outsourced and how much”, and “demand of which markets are satisfied and how much”. The model is finally applied to Mercedes-Benz Turk, one of the largest bus manufacturing companies in the world, and the results are evaluated. Keywords Fuzzy modeling, Possibilistic linear programming, Supply chain planning, Uncertainty. INTRODUCTION The interest in supply chain planning (SCP) has recently increased due to the fact that the opportunity of an integrated planning of the supply chain (SC) can improve the profitability, reduce production and outsourcing costs and enhance customer service levels. As a result, the enterprises can cope with increasing competitiveness introduced by the market globalization [1][2]. Based o the detailed analysis of the papers in Science Direct database for the 2004-2008 period, the applied papers or mixed (applied and theoretical) papers are less than the theoretical papers. Similarly, the studies dealing with two-tier SCs are higher in number rather than the ones that models multi-tier, serial, and network SCs. However, the use of network structure simplifies the representation of SC units and/or functions as well as the interrelationships among them. Consequently, the network structures could be easily transformed to mathematical models. This property is important for the improvement of a SC system (e.g. [3][4]) as well as for taking the operational and strategic decisions (e.g. [5][6]) or designing a new product (e.g. [7][8]). The network based models developed so far have several drawbacks, the most important of which is their deterministic nature. However, in SCP models there are several uncertainties that should be taken into account. Especially in specific problems that necessitate future projections such as new product design or strategic planning, there will be several parameters that cannot be estimated deterministically. Therefore, the fuzzy logic is an important tool to model those types of uncertainties. There are several SC models that assume fuzzy parameters such as demand (e.g. [9][10][11][12][13][14][15]), operation time (e.g. [11]), price (e.g. [14]). In the fuzzy models, generally, the objective function coefficients (see [16][12][17][15]), the constraint parameters (see [18][17][15]), the satisfaction of the objective function (see [19][20][21]), and the satisfaction of constraints (see [19][21]) are modeled with fuzzy modeling. In the majority of the fuzzy logic-based SC models, all the possible uncertainties are not taken into account. For example Hsu and Wang [16] consider only demand as uncertain. Mula et al. [19] treat the realization of the constraint as fuzzy but do not model the fuzziness of the objective function and of the constraint parameters. Another problem encountered in fuzzy logic based models is identified in the defuzzification process. In several researches the defuzzification of the fuzzy parameters is made at the very beginning of the solution (e.g. [14][21]). Although such an approach facilitates the solution, it cannot properly reflect the fuzziness of the results. In the network-based SC models developed so far, another drawback is the assumption of centralized decision (e.g. [4][5]). In fact a centralized approach may provide an efficient approach for a system-wide improvement. However, in real life situation the firms in the SC may not accept such a centralized approach and may prefer a system where their own decisions will also play an important role [22]. Thus, an efficient SC should allow the possibility of considering the decisions at the different levels of the hierarchy in the SC. Additionally the network-based models encountered in the literature are generally very difficult to apply to a real life cases (e.g. [3][4][8][6][23]). Although some 1 Özgür Kabak, Istanbul Technical University, Management Faculty, Industrial Eng. Dept., Macka, Istanbul, Turkey, kabak@itu.edu.tr 2 . Dept. 117 Research in Transport and Logistics 2009 Proceedings Do u University heuristics are proposed to solve the models, generally, there have no guarantee of optimum solution (see [10][11][18][12][17][24][21]). In those papers, generally small hypothetical examples are used to show the applicability of the models and the possibility of their application to large scale problems is not discussed. According to these results it can be concluded that a model that is network-based and that treats the SC as a system of interrelated functions, includes as many efficient members of different hierarchy levels as possible and uses the fuzzy logic to reflect the fuzziness of the planning decision will be an important improvement in the related research area. Additionally the solution approach of the proposed model should be easily applied to large scale real life cases. For this purposes a possibilistic linear programming approach is proposed to model the SC in strategic resource planning perspective. In the following section the proposed model is given. PROPOSED MODEL The main idea of the proposed model is to allow uncertain and therefore flexible decisions to cope with the uncertainties revealed in strategic SCP. In SCP problems, demand has been the most important and extensively studied source of uncertainty (e.g. [1][25][26][21][27][16][12]). Given the fact that effectively meeting customer demand is what mainly drives most SCP initiatives, it is appropriate to put emphasis on incorporating demand uncertainty in the planning decisions. Under this uncertain environment to attempt to make crisp decisions may cause irrelevant or irreversible long term decisions that will need important revisions in medium or short term. Therefore fuzzy decisions are suggested for strategic SCP planning in the proposed model. In order to make fuzzy decision in SCP; a PLP is utilized. In fuzzy linear programming problems, the coefficients of decision variables are fuzzy numbers while decision variables are crisp ones. On the other hand, in PLP formulation the coefficients of decision variables are crisp and/or fuzzy numbers while decision variables are obtained as fuzzy numbers [28]. The problem examined in the paper is motivated from Lakhal et al. [3]. The model accepts that the below given configuration is given at the beginning of the model development.(1)A SC that is the integration of the focal enterprise, its current suppliers and customers, as well as the potential suppliers and customers, and related products, semi-products and raw materials (in the rest of the paper “product” will be used for these three concepts), (2)Resources used to produce the products as well as their costs and capacity levels, (3)Outsourced products and other outsourcing opportunities, as well as their costs, (4)Production and outsourcing yield rates of products. Based on the given data, the proposed model helps the enterprise make decisions about the following strategic questions: 1) Which product should be produced internally? 2) Which resources should be utilized to the production of which product? 3) Which products should be outsourced, and how much? 4) Demands of which market should be satisfied? Additionally, the model can be used to predict the changes of decisions on the change of different input values based on what-if analysis. The Proposed Possibilistic Linear Programming Model The decisions that are sought to made in the proposed model are the production (U~ p ), outsourcing ( D ~ p ) and sales amounts ( S~p ) of the products as well as the outsourcing amount of the resources ( D ~Kr ) where p and r represent products and resources, respectively (p P, r R). In order to deal with the uncertainties in SCP all of the decision variables are designed to be fuzzy and it is represented by the sign of ( ~ ) over their symbols. All the products and the resources can be outsourced theoretically. However in the real problems, firms may limit the outsourcing and production decisions of some products or resources. For his purpose the set of products that should be outsourced are labeled as PD, and the set of products that should be produced are labeled as PU and the following constraints are included into the model: U~ p 0 , p PD (1) D~ p 0 , p PU (2) Outsourcing amounts of the products may be limited based on the capacity of the suppliers or according to a strategic decision of the firm. The following constraint is added to the model to introduce the capacity limit. D~ ~p DC p , p (3) Where DCp is the capacity limit related to product p and ~ is the fuzzy less than sign. In the proposed model system uncertainty is represented by “yield rates”. For the production uncertainty, production yield rate (V~U p ) is used that is the ratio between actual and planned production amounts. Similarly, outsourcing yield rate (V~Dp ) is proposed for outsourcing uncertainty. 118 Research in Transport and Logistics 2009 Proceedings Do u University For each product, total amount of production and outsourcing should be greater than the sum of amount of the product used for the other products and sold amount of the product. The following constraint is offered to guarantee this inequality that also includes the yield rates: V~U U~ V~p p Dp D ~ ~ ~ ~ p BOM pu U u S p p P (4) u Where BOMpu is the bill of materials that represents the amount of product p required to produce product u. , , ~ are fuzzy summation, fuzzy multiplication, and fuzzy greater than signs, respectively. In the model, the production amount is constrained to the resources. During the production of a product some resources are consumed. All the resources including machine times, labor times, transportation and warehousing resources that are used in the production process should be included in the model. KKpr represents the amount of resource r used to produce product p. The company has capacity for the resources that can be increased by rent and outsourcing. Outsourcing is also constrained to the capacity of the suppliers. KCr, is the capacity of resource r. DKCr is the outsourcing capacity. The following constraints are introduced to relate production and resources and the capacities. U~p KK ~ pr KCr D ~Kr , r (5) p D~K ~r DKCr , r (6) The main source of uncertainty~ in SCP is the demands. For this reason the demands related to product p are represented by fuzzy numbers: Tp . There are two ways to model network structured systems in mathematical programming: (1) maximization of the profit, (2) minimization of the cost. The last echelon of the network is limited to some value in the former model while the last echelon is greater than some value in the latter. In the proposed model profit maximization is the objective; therefore the sales amounts are limited to the demands. S~ ~p T ~ p , p P (7) Revenues and costs should be defined to find the profit. Total sales revenue of the firm is calculated as follows, where Fp, is the sales price of the product p.~ Revenue= S p Fp (8) p P Total cost is composed of product outsourcing cost, source consuming cost, and source outsourcing cost which are calculated as given in the following: Product outsourcing cost = DM D~p p (9) p P Source consuming cost = KM r U ~ p KK (10) pr r R p P Source outsourcing additional cost= D~K r DKM r (11) r R Where DMp, is the unit outsourcing cost of product p, KMr is the unit consuming cost of source r, and DKMr, is the additional cost of outsourcing source r. As a result the basic objective of the model (Z1) is constructed as follows: max Z1 S ~ p Fp (-) KM U ~ KK ~ ~ (12) r p pr DKr DKM r DM p Dp p P r R p P p P All decision variables of the model are fuzzy numbers. If the profit maximization is designed to be the only objective then the fuzziness of decision variables as well as the objective function can not be controlled and the results of the model may be inapplicable. For this reason an additional objective is defined to minimize the fuzziness of the profit. Controlling the fuzziness in the profit is also controls the fuzziness in the decision variables as the profit is composed of the decision variables. The concept of entropy is used to measure the fuzzin~ess of the fuzzy sets and numbers. The objective of the fuzziness of profit (Z2) is defined as follows where H( A ) indicates the entropy of fuzzy number A~ . Min Z 2 H (Z1 ) (13) As a result a multi-objective possibistic linear programming model is obtained. Details of the solution procedure of this model are given in the following section. 119 Research in Transport and Logistics 2009 Proceedings Do u University Solution of the proposed PLP model There are different approaches to solve PLP problems. Buckley and Feuring [30] suggested an evolutionary algorithm to solve the fuzzy flexible program that is used to explore the whole nondominated set to the multi-objective fuzzy linear program where all coefficients and decision variables are fuzzy. Tanaka et al. [28] proposed transformations of fuzzy constraints characterized as interval, triangular, and exponential possibility distributions to linear constraints according to predetermined possibility levels. Disadvantage of the former is complex solution procedure while the latter needs predetermined possibility levels. In this model, triangular fuzzy numbers (TFNs) are chosen to represent all fuzzy parameters and variables since TFNs are enough to represent the uncertainty in demands and yield rates as well as the decision variables and it is easy to apply mathematical operations for TFNs. A TFN, A~ , is defined by its left support (L), medium value (M), and right support (R). In the solution procedure, normalization of the multiple objectives is suggested to aggregate them in the same scale. In order to normalize the fuzzy objectives the upper and lower bounds of objective values are found. Initially the following LP model is suggested to find the upper and lower bounds of the profit. (LP-1) (definitions of the parameters and decision variables are introduced in Table 1) Objective function Max Z S p * Fp - KM r * U p * KK pr DKr * DKM (14)r DM p Dp p r p p Subject to U p 0 , p PD (15) Dp 0 , p PU (16) D p DK p , p (17) VU *U , p. (18) p p VDp * Dp BOM pu *U u S p u U * KK KC DK , r (19) p pr r r p DKr DKCr , r (20) S p Tp , p (21) U p , D p , S p 0 , p (22) All decision variables and the constraints in LP-1 are defined by crisp numbers. The objective function of LP1, which is defined in (14), is the defuzzified version of the first objective in PLP. The defuzzification of the fuzzy parameters in PLP while converting it to LP-1 is realized according to the result that is aimed to be reached. For instance the yield rates and the demands are considered to be at their highest level (i.e. right supports of the corresponding TFNs) to find the upper bound of the profit while they are considered to be at their lowest level (i.e. left supports of the corresponding TFNs) to reach the lower bound of the profit. When the upper bound ( Z1 ) and lower bound ( Z ) of the profit is calculated by running the LP-1 twice with different 1 parameters, the normalization of the first objective of the PLP is made according to the following formula: Z M1 Z1 (23) Z1 Z1 Where Z M1 is the medium value of the profit, which is obtained as a TFN. On the other hand, the following formula is used to normalize the second objective of the PLP: Z1 Z Z R 1 1 Z L 1 (24) Z1 Z1 Where Z R L1 and Z1 are the right and left supports of the profit. Here, the entropy of a TFN is calculated by the difference of the two supports. When the normalization formulas for the objectives are defined, the PLP model can be converted to LP-2. 120 Research in Transport and Logistics 2009 Proceedings Do u University TABLE 1 Definitions of the parameters and the decision variables in LP 1 and LP 2 Indices: p,u Products; p,u P R Resources; r R i Critical points of TFN, i = L,M,R Parameters: P The set of products R The set of resources PD The set of products that should be outsourced, PU The set of products that should be produced, PU PD P P BOMpu Bill of materials matrix (amount of product p KKpr Amount of resource r that is required for product that is required for product u) p KCr The capacity of resource r KMr Unit cost of resource r DKCr Outsource capacity of resource r DKMr Unit outsourcing cost of resource r DMp Unit additional cost of outsourcing product p DKp Capacity of outsourcing amount of product p Tp Demand of product p Fp Price of product p VU p Production yield rate of product p VDp Outsourcing yield rate of product p VDi ith critical point of outsourcing yield rate of i ithVU critical point of production yield rate of p product p p product p T i ithp critical point of demand of product p Z1 , Z1 Upper and Lover bounds of the profit Decision Variables: U p Production amount of product p D Outsourcing amount of product pp S p Sales amount of product p DK r Outsourcing amont of resource r i ithU critical point of production amount of p D i ith critical point of outsourcing amount of p product p product p S i ith critical point of sales amount of product p DK i ith critical point of outsourcing amount of p r resource r Z i ith critical point of profit Minimum level of the the normalized objectives1 The signed ( X ) parameters and decision variables are defuzzified versions of the similar variables of PLP model. (LP-2) (definitions of the parameters and decision variables are introduced in Table 1) Objective Function Max Z = (25) Subject to; Z M1 Z1 Z1 Z1 (26) Z R L1 Z1 Z1 Z1 Z1 Z1 (27) Z M S M * F - KM * U M * KK DK M * DKM DM * D M (28) 1 p p r p pr r r p p p r p p Z L S L * F - KM * U R * KK DK R * DKM DM * D R (29) 1 p p r p pr r r p p p r p p Z R R1 S p * Fp - KM r * U L p * KK pr DK L r * DKM r DM L (30) p * Dp p r p p U Rp 0 , p PD (31) D Rp 0 , p PU (32) D Rp DC p , p (33) VU i i i i i i , (p,i) (34) p *U p VDp * Dp BOM pu *U p S p u U i * KK KC DK i , (r,i) (35) p pr r r p DK Rr DKCr , r (36) 121 Research in Transport and Logistics 2009 Proceedings Do u University S ip T i p , (p,i) (37) U L Mp U p U R p , p (38) D Lp D M p D R p , p (39) S L S M S Rp p p , p (40) DK Lr DK M r DK R r , r (41) U pi , Dpi , S pi 0 , (p,i) (42) The aim of LP-2 is to maximize the minimum level of normalized objectives (defined as ) of PLP. is obtained through (26) and (27) while three critical values of the profit is calculated by using (28)- (30). The constraints (31)-(37) are the conjugate of (1)-(7) after the application of TFN operations. Constraints (38)-(41) prevents having illogical TFNs for decision variables (for TFN, A~(L, M, U) , L M U should be satisfied). In (42) sign restrictions are given. APPLICATION IN AUTOMOTIVE INDUSTRY Mercedes- It produces several types of buses with capacity of 4000 per year. The proposed methodology is applied in the body manufacturing of the plant. Particularly, it is used to make strategic resource planning of the year 2009 for one kind of body, coded as “Integro 15/R”, which is sold to Mercedes, Germany. The body is composed of 2178 product, sub-product and raw materials, 823 of which are raw materials. There are two main groups of raw materials, which are metal tubes and metal sheets. While applying the proposed model to the body production, first of all, the resources are determined. The main resource of the production is labor. On the other hand, there are numerous different kinds of work benches and work stations in the production process. It could be very difficult to model the system by including all different benches/stations since some of them are substitutes of each other, processing times are not clear etc. Therefore main sections of the production process are considered as the resources of the system. The production process is composed of three sections: tube processing, sheet processing and welding. The related parameters about the resources are given in Table 2. TABLE 2 Parameters related to resources Resource The Resource Capacity Unit cost ID (Year 2009) (€) Outsourcing capacity Additional unit cost of outsourcing (€) 1 Labor 4422600 man-hours 0,216670 2063880 man-hours 0,1083300 2 Tube processing 762048 hours 0,022159 76205 hours 0,0110795 3 Sheet processing 394632 hours 0,080577 39463 hours 0,0402885 4 Welding 3129840 hours 0,004029 312984 hours 0,0020145 One of the important inputs of the model is production yield rates. Yields occur in the production system according to products’ production process. The yield rates are specified for the product groups (see Table 3). TABLE 3 Production yield rates Product group Pessimistic value of Most possible value of Optimistic value of the yield rate the yield rate the yield rate Products of tube process 0.988 0.992 0.995 Products of sheet process 0.982 0.988 0.993 Products of welding 0.976 0.984 0.990 Other products 0.994 0.996 0.997 Yield rates related to outsourced products are specified according to the type of the raw material as given in Table 4. The demands of the body is specified thorough aggregating several statistical and judgmental forecasts as T~1 =(180, 202.8, 225.6). In order to solve the proposed PLP that is structured according to the given parameters, first of all, upper and lower bounds of the profit are calculated with LP-1. For the upper bound of the profit the optimistic values of the fuzzy parameters are taken into consideration while pessimistic values are considered for the 122 Research in Transport and Logistics 2009 Proceedings Do u University lower bound. As a results of two runs of LP-1, it is found that Z1 = 788024 and Z1 = 565761, which indicates that the maximum fuzziness level of the profit is 222263. Then LP-2 is employed to reach the final solution. TABLE 4 Outsourcing yield rates Product group Pessimistic value of Most possible value of Optimistic value of the yield rate the yield rate the yield rate Metal Tubes 0.993 0.997 1.000 Metal sheets 0.990 0.995 1.000 Other raw materials 0.995 0.998 1.000 LP-2 is solved in GAMS software with CPLEX solver. According to the results, the profit, which is specified as a TFN, is (527968, 654528, 661464). This result indicates that the most probable value of the profit is € 654538 while the pessimistic and optimistic profits are equal to 527968 and 654528, respectively. and the normalized values of the objectives are equal to 0.399. The sales amount of the body “Integro 15/R” is (177.600, 186.037, 186.037). The results show that the capacities of the resources are adequate for determined production level. According to most possible values of the production, 86.27% of the labor, 99.94% of the tube processing, 92.86% of the sheet processing, and 81.12% of the welding sources are utilized. Results related to production/outsourcing amounts indicate that there is no uncertainty in the outsourcing amounts (i.e. the three critical values of the related TFNs are equal). On the other hand, production amounts of the 66% of the products are uncertain where the level of uncertainty varies from %0 to %12. CONCLUSIONS In this paper a possibilistic linear programming model is proposed to make strategic resource planning decisions in SC context. The proposed model is solved through two LP models. Furthermore the model is used to analyze the resource utilization of a bus manufacturing company. The results show that the own resources of the company is adequate for meeting the current demands. However, if the demand increases then the capacity will be insufficient. Under these circumstances, the model suggests to outsource some products that are currently produced by the company itself even for a price that is higher than their current costs. In its current form, the proposed model is expected to provide an important guide to SC managers in their strategic plans, taking into account the fuzziness of the long term plans. The proposed model can be improved by offering more detailed methodologies for determining the fuzzy inputs. The application of the model has to be integrated with the enterprise resource planning software of the companies. REFERENCES [1] Guillén, G., Mele, F.D., Bagajewicz, M.J., Espuña, A. and Puigjaner, L., 2005, “Multiobjective supply chain design under uncertainty”, Chemical Engineering Science, 60, 1535 – 1553. [2] Koutsoukis, N.-S., B. Dominguez-Ballesteros, C.A. Lucas, G. Mitra, 2000. “A prototype decision support system for strategic planning under uncertainty”. Int. J. of Phy. Dist. & Log. Man., 30, 640-660. [3] Lakhal, S., Martel, A., Kettani, O., Oral, M., 2001. “On the optimization of supply chain networking decisions”, European Journal of Operational Research, 129, 259-270. [4] Lakhal, S.Y., 2007. “An operational sharing and transfer pricing model for network-manufacturing companies”, European Journal of Operational Research, 175, 543-565. [5] Sabri, E.H. and Beamon, B.M., 2000. “A multi-objective approach to simultaneous strategic an operational planning in supply chain design”, Omega, 28, 581-598. [6] Altiparmak, F., Gen, M., Lin, L. and Paksoy, T., 2006. “A generic algorithm approach for multi-objective optimization of supply chain networks”, Computers & Industrial Engineering, 51, 197-216. [7] Graves, S.C. and Willems, S.P., 2001, “Optimizing the supply-chain configuration for new products”, Technical Paper, Leaders for Manufacturing Program and Sloan School of Management, MIT, Cambridge, [8] Wang, J. and Shu, Y.-F., 2007. “A possibilistic decision model for new product supply chain design”, European Journal of Operational Research, 177, 1044-1061. 123 Research in Transport and Logistics 2009 Proceedings Do u University [9] Petrovic, D., Roy, R., and Petrovic, R., 1998. “Modelling and simulation of supply chain in an uncertain environment”, European Journal of Operational Research, 109, 299-309. [10] Petrovic, D., Roy, R. and Petrovic, R., 1999. “Supply chain modelling using fuzzy sets”, International Journal of Production Economics, 59, 443-453. [11] Wang, J. and Shu, Y.-F., 2005. “Fuzzy decision modeling for supply chain management”, Fuzzy Sets and Systems, 150, 107-127. [12] Wang, R.-C. and Fang, H.H., 2001. “Aggregate production planning with multiple objectives in a fuzzy environment”, European Journal of Operational Research, 133, 521-536. [13] Chen, C.-L. and Lee W.-C. 2004. “Multi-objective optimization of multi-echelon supply chain networks with uncertain product demands and prices”, Computers & Chemical Engineering, 28(6-7), 1131-1144. [14] Torabi, S.A. and Hassini, E., 2008. “An interactive possibilistic programming approach for multiple objective supply chain master planning”, Fuzzy Sets and Systems, 159, 193-214. [15] Xu, J., Liu, Q., and Wang, R., 2008. “A class of multi-objective supply chain networks optimal model under random fuzzy environment and its application to the industry of Chinese liquor”, Information Sciences, 178(8), 2022-2043. [16] Hsu, H.-M. and Wang, W.-P., 2001. “Possibilistic programming in production planning of assemble-to-order environments”. Fuzzy Sets and Systems, 119, 59-70. [17] Wang, R.-C. and Liang, T.-F., 2005, “Applying possibilistic linear programming to aggregate production planning”, International Journal of Production Economics, 98, 328–341. [18] Maity, K. and Maiti, M., 2007. “Possibility and necessity constraints and their defuzzification – A multi-item production-inventory scenario via optimal control theory”, Eur. J. of Op. Res., 177, 882-896. [19] Mula, J., Poler, R. and Garcia, J.P., 2006. “MRP with flexible constrains: A fuzzy mathematical programming approach”, Fuzzy Sets and Systems, 157, 74-97. [20] Wang, R.-C. and Liang, T.-F., 2004. “Application of fuzzy multi-objective linear programming to aggregate production planning”, Computers & Industrial Engineering, 46, 17–41. [21] Liang, T.-F., 2006. “Distribution planning decisions using interactive fuzzy multi-objective linear programming”, Fuzzy Sets and Systems, 157, 1303-1316. [22] Ryu, J.-H., Dua, V. and Pistikopoulos, E.N., 2004. “A bilevel programming framework for enterprise-wide process network under uncertainty”, Computers & Chemical Engineering, 28, 1121-1129. [23] Roghanian, E., Sadjadi, S.J. and Aryanezhad, M.B., 2007. “A probabilistic bi-level linear multi-objective programming problem to supply chain planning”, Applied Mathematics and Computation, 188, 786-800. [24] Chen, C.-L. and Lee W.-C. 2004. “Multi-objective optimization of multi-echelon supply chain networks with uncertain product demands and prices”, Computers & Chemical Engineering, 28(6-7), 1131-1144. [25] Petrovic, D., 2001. “Simulation of supply chain behaviour and performance in an uncertain environment”, International Journal of Production Economics, 71, 429-438. [26] Das, S.K. and Abdel-Malek, L., 2003. “Modeling the flexibility of order quantities and lead-times in supply chains”, International Journal of Production Economics, 85 (2), 171-181. [27] Leung, S.C.H., Tsang, S.O.S., Ng, W.L. and Wu, Y., 2006. “A robust optimization model for multi-site production planning problem in an uncertain environment”, Eur. J. of Op. Res., 181(1), 224-238. [28] Tanaka, H., Guo, P. and Zimmermann, H.-J., 2000. “Possibility distribution of fuzzy decision variables obtained from possibilistic linear programming problems”, Fuzzy Sets and Systems, 113, 323-332. [29] Leon, T. and Vercher, E. 2004. “Solving a class of fuzzy linear programs by using semi-infinite programming techniques”, Fuzzy Sets and Systems, 146, 235-252 [30] Buckley, J.J. and Feuring, T., 2000. “Evolutionary algorithm solution to fuzzy problems: fuzzy linear programming”, Fuzzy Sets and Systems, 109, 35-53. 124 Research in Transport and Logistics 2009 Proceedings Do u University A STRUCTURAL EQUATION MODEL FOR MEASURING SERVICE QUALITY IN PASSENGER TRANSPORTATION G.Nilay YÜCENUR1 2 Abstract With the improvement of the customer based management approaches, the firms have to consider the costumers’ needs and wants in their activities and also they have to look the events and products with customers’ perspective. In this paper, the variables are analyzed which have effects on customers’ loyalty such as service quality, sacrifice, service value and customer satisfaction in passenger transportation with airways. Also the interrelationships are analyzed among these variables. The model which was proposed by Demirel et al. (2006) is used for predicting customers’ loyalty in airway sector. In an application section there are 224 questionnaires from 6 different airway companies in Turkey for domestic flights. The data was analyzed with SPSS 12.0 and LISREL 8.80 packet programs. SPSS 12.0 was used for analyzing reliability and normality and LISREL 8.80 was used for analyzing model fit and structural equation model analysis such as path diagram. Keywords Airways, loyalty, satisfaction, service quality, service value, structural equation INTRODUCTION Loyal customers are very important for all firms in all sectors but especially for passenger transportation sectors. In airways, the firms’ main structure are based on customers’ needs and wants. In this sense, the firms have to support the high quality products and services in their activities. Service quality levels affect the firm’s competitive advantage and they determine market share and profitability. The key variables normally considered when modeling passengers’ decision-making processes include service quality with expectation and perception, sacrifice, service value, customer satisfaction and customers’ loyalty. In this paper, the variables are analyzed which have effects on customers’ loyalty such as service quality, sacrifice, service value and customer satisfaction in passenger transportation with airways. Also the interrelationships are analyzed among these variables. The model which was proposed by Demirel et al. (2006) is used for predicting customers’ loyalty in airway sector. In an application section there are 224 questionnaires from 6 different airway companies in Turkey for domestic flights. The data was analyzed with SPSS 12.0 and LISREL 8.80 packet programs. SPSS 12.0 was used for analyzing reliability and normality and LISREL 8.80 was used for analyzing model fit and structural equation model analysis such as path diagram. With the model which is used in this paper service quality can be measured in passenger transportation. Our study both synthesizes and builds on the efforts to conceptualize the effects of service quality, sacrifice, service value and customer satisfaction on customers’ loyalty. The rest of this study is structured as follows: The first part describes literature review about service quality in airway services.Next part discusses the Procedure, methodology, the research model, criteria and results of empirical study. The final results of the empirical study are presented and discussed in the final section. THEORETICAL BACKGROUND Recent marketing research defined loyalty as a deeply held commitment to repurchase or repatronize a preferred product or service consistently in the future. Customer loyalty research has provided theoretical justification for viewing satisfaction as an important antecedent to loyalty, and has empirically showed significantly positive relationships. Prior research frequently suggests that loyal customers are likely to provide new referrals through positive word of mouth. They buy more products and resist competitive ressures. Guest loyalty was used as an intervening variable that has a time dimensional effect on repeat 1 Istanbul, Turkey, nserbest@yildiz.edu.tr 2 Nihan Çetin Demirel, Istanbul, Turkey, nihan@yildiz.edu.tr 125 Research in Transport and Logistics 2009 Proceedings Do u University purchase and word of mouth. Benefits and trust are the most important antecedents to guests’ loyalty and loyalty results in increased product use [2]. In the airline industry context the problem is whether management can perceive correctly what passengers want and expect. Moreover, expectations serve as standards or reference points for customers. In evaluating service quality, passengers compare what they perceive they get in a service encounter with their expectations of that encounter. Assessing passenger expectations is not a static exercise as passengers are becoming increasingly sensitive to quality. However, not all service dimensions are equally important to all passengers, because no two passengers are precisely alike, especially when demographics; purposes of travelling and ethnic background is considered [3]. The key variables normally considered when modelling passengers’ decision-making processes include service expectation, service perception, service value, passenger satisfaction, and airline image. Understanding what consumers expect from a service organization is important because expectations provide a standard of comparison against which consumers judge an organization’s performance. Airlines need to understand what passengers expect from their services. To date, the effect of air passengers’ expectations on service perception and passenger satisfaction has not been fully investigated, even though it is an important commercial consideration [4]. In the passenger airline industry, only the customer can truly define service quality. The quality of airline service is difficult to describe and measure due to its heterogeneity, intangibility and inseparability. Never theless, quite a few conceptual and empirical studies have been devoted to investigate the service quality issues in the passenger airline industry. Various schemes for defining service quality dimensions or attributes have been proposed from the perspective of passengers. Most of these schemes are presented as quality measures for examining the relationships between service quality and related issues such as airline choice, customer satisfaction, customer loyalty, passenger type, airline type, airline class, aircraft type, productivity changes in quality levels over time, total transportation service offering, assessment group and attribute dependency [5]. For this paper our variables are service quality, service value, sacrifice, customer sarisfaction and customer loyalty. Service Quality Service quality can be defined as a consumer’s overall impression of the relative efficiency of the organization and its services. Customer satisfaction can be defined as a judgement made on the basis of a specific service encounter. The importance of the relationships between airline service quality, passenger satisfaction, and behavioural intentions have been examined. Although the direction of airline service quality and passenger satisfaction has been studied empirically, the causal order between airline service quality and passenger satisfaction, and the exact relationship between airline service quality, passenger satisfaction and behavioural intentions, is still a matter of debate because the direction may vary depending on context [4]. Service Value Value can be defined as a customer’s overall assessment of the utility of a product based on perceptions of what is received and what is given. Service value has been identified as an important variable of customer satisfaction and behavioural intentions. Even though studies have looked at service quality and value, the relationship between them still remains unclear. In spite of the importance of perceived service value as a form of assessment of services, there has also been only limited analysis of the exact nature of service value and its influence on customer behaviour. Previous airline service studies have often ignored service value and few have investigated the effect of service value on passenger behaviour [4]. Sacrifice Customer perceived sacrifice, which helps to integrate extant research and provides a more comprehensive picture about how customer value can be influenced. Sacrifice refers to what is given up or sacrificed to acquire a product or service. In fact, many customers count time rather than monetary cost as their most precious asset. Therefore, generally speaking, it is clear that there are two broad kinds of sacrifice: monetary costs and non-monetary costs. The former can be assessed by a direct measure of monetary cost of 126 Research in Transport and Logistics 2009 Proceedings Do u University the service or product and the latter can be defined as the time, effort, energy, distance and conflict invested by customers to obtain products or services [6]. Customer Satisfaction Customers have service expectations of an organization. Organizations are obliged to serve their customers. In the eyes of the service-profit chain theorists and champions of organizational quality, the closer the organization behaves in terms of what is expected of it by its customers, the more effective the organization. Similarly, and directly related to this research effort, the effectiveness of internal organizational service units can be measured by the degree of satisfaction of its performance by its internal customers (role set members). When members of internal organizational units satisfy the needs of their unit's internal customers, they also are enabling their internal customers to perform their tasks. By doing so, the network of organizational units are more apt to work effectively together to accomplish the overall customer service aims of the organization [7]. Customer Loyalty Early literature on loyalty involved a dual perspective with the views held by the researcher influencing the perspective taken. For example, researchers holding a deterministic view of loyalty advocated the need to consider loyalty from an attitudinal perspective while researchers holding a stochastic view considered loyalty from a behavioural perspective [8]. Traditionally, customer loyalty has been defined as a behavioral measure. These measures include proportion of purchase, probability of purchase, probability of product repurchase, purchase frequency, repeat purchase behavior, purchase sequence, and multiple aspects of purchase behavior. In the retailing context, following measures of customer behavior are commonly applied by practitioners – share of purchase (SOP) that measure the relative share of a customer’s purchase as compared to the total number of purchases and share of visits (SOV) that measure the number of visits to the store as compared to the total number of visits. Other commonly used measures in the industry include Share of Wallet (SOW) – that is expenditure at a specific store as a fraction of total category expenditures which is analogous to share of purchase (SOP); Past Customer Value (PCV) – based on the past profit contribution of the customer; Recency, Frequency and Monetary Value (RFM) – measure of how recently, how frequently and the amount of spending exhibited by a customer [9]. THE RESEARCH MODEL AND HYPOTHESIS DEVELOPMENT Service quality can be regarded as a composite of various attributes. It not only consists of tangible attributes, but also intangible/subjective attributes such as safety, comfort, which are difficult to measure accurately. Customer Loyalty H5 H Service Value6 H7 H1 H2 H4 Service Quality Sacrifice Customer Satisfaction H3 FIGURE 1 The Hypotheses of Relationships between Model Variables 127 Research in Transport and Logistics 2009 Proceedings Do u University Different individual usually has wide range of perceptions toward quality service, depending on their preference structures and roles in process (service providers/receivers). In Figure 1 the research model and the relationships between model variables are represented with hypothesis. To measure service quality, conventional measurement tools are devised on cardinal or ordinal scales. Most of the criticism about scale based on measurement is that scores do not necessarily represent user preference. This is because respondents have to internally convert preference to scores and the conversion may introduce distortion of the preference being captured [10]. For this conceptual model we can develop seven different hypotheses. H1 Sacrifice has a direct effect on service value. H2 Service quality has a direct effect on service value. H3 Service quality has a direct effect on customer satisfaction. H4 Service value has a direct effect on customer satisfaction. H5 Service value has a direct effect on customer loyalty. H6 Service quality has a direct effect on customer loyalty. H7 Customer satisfaction has a direct effect on customer loyalty. TABLE 1 Demographic Information of the Respondents (1) Alternatives Amount Percentage (%) Gender Female - Male 147 - 97 % 60.2 - % 39.8 Age 20 – 30 / 31 – 40 178 - 66 % 73.0 - % 27.0 High School 10 % 4.1 Education Before University 10 % 4.1University 157 % 64.3 Master 67 % 27.5 Under 1000 YTL 5 % 2.0 1000 – 3000 YTL 143 % 58.8 Monthly income 3001 – 5000 YTL 76 % 31.1 5001 – 7000 YTL 10 % 4.1 7001 – 11000 YTL 10 % 4.0 TABLE 2 Demographic Information of the Respondents (2) Alternatives Amount Percentage (%) Which airway firm did Turkish Airlines 164 % 67.2 you fly with in the last Onur Air 20 % 8.2 time? AtlasJet Air 35 % 14.4SKY Air 25 % 10.2 How many times have 1 71 % 29.1 you flied with the same 2 – 5 123 % 50.4 airway firm? 6 – 10 15 % 6.111 - Above 11 35 % 14.4 What was the reason of Work 96 % 39.4 your travel? Personal 53 % 21.7Holiday 95 % 38.9 Yourself 199 % 81.6 Who chose the airway Secretary 10 % 4.1 firm? Travel agent 20 % 8.2Family 5 % 2.0 Other 10 % 4.1 Service quality 25 % 10.2 Low price 50 % 20.5 What was the reason of Timing 79 % 32.4 your choice? Casual choice 10 % 4.1 Trust 60 % 24.6 Other 20 % 8.2 1 40 % 16.4 How many times did you 2 – 5 91 % 37.3 fly in last year? 6 – 10 68 % 27.9 11 - Above 11 45 % 18.4 128 Research in Transport and Logistics 2009 Proceedings Do u University METHODOLOGY This research proposes an integrative model based on established relationships among service quality, sacrifice, service value, customer satisfaction and customer loyalty, and tests it in the context of airways firms in Turkey. The questionnaires were used for research. Subjects were asked to assess items of different constructs such as factors viewed as antecedents of customer loyalty such as service quality, sacrifice, customer value, and customer satisfaction in terms of their perceptions, based on a seven-point scale. A seven- point Likert-type response format ranging from “strongly disagree” to “strongly agree” was used for all items. Sample and Procedures The sample was chosen from 6 different airways firms from Turkey. There were 217 passengers’ questionnaires. Before the main research there was a pilot application to 60 respondents from different 3 airways firms. Respondents answered the questions with face to face pollster. The questionnaire contained six parts. Those parts were: perceived service quality, sacrifice, service value, customer satisfaction, and customer loyalty; there was an additional section for customer’s demographic information. Table 1 and Table 2 represent demographic information of the respondents. Results of Structural Equation Modeling and Hypothesis Testing Structural equation modeling was performed to investigate the relationships between the criterion variable of behavioral intention and the respective predictor variables of service quality, sacrifice, service value, and customer satisfaction. Figure 2 displays the structural model parameters and summarizes the degree to which the data fit the model. FIGURE 2 The Path Analysis Results 129 Research in Transport and Logistics 2009 Proceedings Do u University TABLE 3 Fit Indices for Measurement and Structural Models Index Value x2/df 13.244 Goodness-of-fit (GFI) 0.35 Adjusted goodness-of-fit (AGFI) 0.29 Root mean squared residual (RMR) 0.24 Root mean squared error of approximation (RMSEA) 0.19 Normed fit index (NFI) 0.63 Relative fit index (RFI) 0.61 Comparative fit index (CFI) 0.65 The GFI estimates the amount of variance explained by the model, and the AGFI adjusts this estimate by taking into account the degrees of freedom. Both of these estimates can vary from 0 to 1 [11]. The simplest fit index provided by LISREL is RMR. This is the square root of the mean of the squared discrepancies between the implied and observed covariance matrices. RMR has a lower bound of 0 and an upper bound of 1. Similarly to the RMR, the RMSEA is based on the analysis of residuals. On the other hand, the NFI ranges from 0 to 1. An NFI of 0.95 means that the model is %95 better fitting than the null model. The RFI ranges between 0 and 1, with values approaching unity indicating a good fit to the data. CFI is based on the noncentral x2 distribution. The CFI also ranges between 0 and 1 [11]. The evaluation of this conceptual path model indicated a moderate goodness of fit (x2/degree of freedom= 13.244, GFI= 0.35, AGFI= 0.29, RMR= 0.24, RMSEA= 0.19, NFI= 0.63, RFI= 0.61, CFI= 0.65). The goodness of fit index (GFI) value was lower than 0.8 and also the adjusted goodness of fit index (AGFI) value was 0.29. The AGFI value appeared lower. The RMSEA value was higher (>0.05). This may indicate a level of discrepancy between the sample observed and the hypothesized correlations in the theoretical model. Although the normed fit index (NFI) value was lower than the commonly accepted value of above 0.90. Researchers have recommended comparative fit index (CFI) as a better fit index than NFI. Similarly, the CFI value for the current model was clearly a lower than the commonly accepted value of above 0.90 [11]. Since those indices all had values close to or below the level for superior fit, the proposed theory model was believed to have achieved a good model fit but not an excellent model fit. The path coefficients are shown on the arrows in Figure 2. The results obtained from our analyses indicate that sacrifice has a direct negative effect on service value (-0.14) therefore the data does not support H1. This value is not significant. Because of its t value is equal -2.69. The t value must be above 1.96 [12]. Service quality has a direct and significant effect on service value (0.62) and thus H2 is confirmed. As confirmed in Figure 2, service quality has a direct effect on customer satisfaction (0.94), thus H3 is confirmed. Service quality has a significant effect on customer loyalty (0.18) and therefore H6 is supported. Service value has a direct negative effect on customer satisfaction (-0.011) therefore the data does not support H4. This value is not significant. Because of its t value is equal -0.37. Service value has a direct effect on customer loyalty (0.039) and thus we confirm H5. Figure 2 shows that customer satisfaction has a direct effect on customer loyalty (0.18), thus H7 is confirmed. Thus, on one hand service quality has a direct significant effect on customer satisfaction; on the other hand it has direct and positive effects on service value and customer satisfaction. Similarly, on one hand service value has a direct and positive effect on customer loyalty. CONCLUSIONS This paper has presented a model includes service quality, sacrifice, service value, customer satisfaction, and customer loyalty. By analyzing a set of 244 airways passengers, the study uses the structural equation model with LISREL 8.80 software to show that the sacrifice has a direct effect on service value, and has indirect effects on customer satisfaction and customer loyalty. Depending on this service value has direct effects on customer satisfaction and customer loyalty. Finally, the service quality has direct effects on service value, customer satisfaction and customer loyalty. Overall, this research highlights the important role of three dimensions on customers’ loyalty. Our findings indicate that service quality, service value and customer satisfaction lead to customers’ loyalty. In 130 Research in Transport and Logistics 2009 Proceedings Do u University airway services managers should place special emphasis on the fact that of customers’ perception of servcie quality, sacrifice, service value and customer satisfaction. In airways services, in all airways firms service quality and service value is effective on customer loyalty. In airways transportation, the level of perceived service quality and service value are the major components of customer loyalty, besides in the assessment of the airways passengers may consider other factors such as price or reputation. In future research, value scales of demographic features could be viewed as a variable in a new model. REFERENCES [1] Demirel, N.Ç., Serbest, G.N. and 2006. “A Proposed Model to Measure Service Quality in Passenger Transportation”, 4th International Logistics and Supply Chain Congress – ILC2006, 322-327. [2] Kim, W.G., Lee, C. and Hiemstra S.J., 2004. “Effects of an online virtual community on customer loyalty and travel product purchases”, Tourism Management, 25, 343–355. [3] Gilbert, D. and Wong R.K.C., 2003. “Passenger expectations and airline services: a Hong Kong based study”, Tourism Management, 24, 519–532. [4] Park, J.-W., Robertson, R. and Wu, C.-L., 2004. “The effect of airline service quality on passengers’ behavioural intentions: a Korean case study”, Journal of Air Transport Management, 10, 435–439. [5] Chang, Y.-H. and Yeh, C.-H., 2002. “A survey analysis of service quality for domestic airlines”, European Journal of Operational Research, 139, 166–177. [6] Wang, Y., Lo, H.-P. and Yang, Y., 2004. “An Integrated Framework for Service Quality, Customer Value, Satisfaction: Evidence from China’s Telecommunication Industry”, Information Systems Frontiers 6: 4, 325–340. [7] Gilbert, G.R., 2000. “Measuring internal customer satisfaction”, Managing Service Quality, 10, 178-186. [8] Rundle-Thiele, S., 2005. “Elaborating customer loyalty: exploring loyalty to wine retailers”, Journal of Retailing and Consumer Services, 12, 333–344. [9] Kumar, V. and Shah, D., 2004. “Building and sustaining profitable customer loyalty for the 21st century”, Journal of Retailing, 80, 317–330. [10] Tsaur, S.-H., Chang, T.-Y. and Yen, C.-H., 2002. “The evaluation of airline service quality by fuzzy MCDM”, Tourism Management, 23, 107–115. [11] Kelloway, E.K., (1998), “Using LISREL for Structural Equation Modeling a Researcher’s Guide”, Sage Publications Inc. [12] Schumacker, R.E., and Lomax, R.G., (2004), “A Beginner’s Guide to Structural Equation Modeling”, Second Edition, Lawrence Erlbaum Associates Inc. Publishers, London. 131 Research in Transport and Logistics 2009 Proceedings Do u University ANALYSIS OF POTENTIAL GAIN FROM USING HYBRID VEHICLES IN PUBLIC TRANSPORTATION 1 and Özay ÖZAYDIN2 Abstract: In the last two decades, the pollution in major cities was a growing concern. A significantly large contribution was made by road transportation sources to this problem. EU regulations on emissions and improvements in fuel quality sustained an improvement in air quality through reduction in pollutants. Latest trend taking the environmental issues into consideration in consumer car industry is hybrid vehicles, which use a conventional inner combustion engine using petroleum or diesel fuel, alongside with an electric engine, selectively using the necessary one depending on the current speed and power needs. As the vehicles equipped with this technology create less emission and consume less fuel, investigation of the potential gain when used public transportation is inevitable. In this study, a comparison-based study will be done between traditional busses using fossil fuel and new technology busses that the Istanbul Municipality selected as candidate fleet, utilizing energy, finance, environment and social responsibility aspects. Keywords: Public transportation, Hybrid vehicles, Environment I. INTRODUCTION Public transport, public transportation, public transit or mass transits are transport systems that transport members of the general public, usually charging set fares. The terms generally taken to include rail and bus services, and wider definitions might include scheduled airline services, ferries, and taxicab services. A restriction sometimes applied is that transit should occur in continuously shared vehicles, which would exclude taxis that are not shared-ride taxis. Public transport may be regulated as a common carrier and usually provides scheduled service on fixed routes on a non-reservation basis, although share taxis provide an ad-hoc form of flexible public transport and demand responsive transport provides a pre-bookable form of public shared transport. Taxicabs and other vehicles for hire are generally fully flexible. The majority of transit passengers are traveling within a local area or region between their homes and places of employment, shopping, or schools. II. EFFECTS OF PT ON ENVIRONMENT The people moves from one place to another all their lives for various reasons. The distance moved is getting bigger as the cities growing. And they have to use vehicles to reach their destinations. In the developed countries well educated people use the public transportation vehicles. Public transportation has many of the advantages. It is more economical for the people, for the country and for the environment. It yields less noise, less pollution, less heat emission for the atmosphere, less money for fuel and vehicle, less area for roads and parking. [1] III. NEXT STEP: HYBRID VEHICLES Why would anyone buy a hybrid electric vehicle (HEV)? First mass introduction to markets happened in 1999, and was a radical change in motor vehicle technology. Currently available HEV models achieve fuel economy increases of 10-40% [2]. compared to similar-sized vehicles, with commensurate reductions in greenhouse gas emissions. Some of HEVs’ benefits are collective. Millions of people buying less-polluting and more fuel-economical cars can produce cleaner air and reduce risk of climate change, but no single HEV buyer can have much impact on these problems. 1 Turkey, iduzdar@halic.edu.tr 2 Turkey, oozaydin@dogus.edu.tr 133 Research in Transport and Logistics 2009 Proceedings Do u University Further, the private benefits of HEVs are unclear; reduced expenditures on fuel are routinely shown to be lass than the vehicle purchase price premium of the hybrid vehicle over an assumed non-hybrid alternative. If private benefits are illusory and if collective benefits are not achievable by individual customers, why would anyone buy an HEV? Hopes that HEVs (or any new technology that promises collective benefits) can be successfully mass-marketed depend on answers to such a question [4]. An answer HEV not only provides its owner with transportation, but also provides symbolic meanings that owners can incorporate into better stories about themselves [3]. The symbolic meanings associated with HEVs are multiple and multi-layered, including widely recognized ideas like preserving the environment, opposing war, saving money, reducing support for oil producers, and owning the latest technology. But these denotations are linked to more personal connotations, such as concern for others, ethics, maturity, national independence, or individuality [4]. IV. HYBRID VEHICLES AS A PUBLIC TRANSPORTATION TOOL Hybrid-electric buses are being developed to answer specific challenges faced by today's public transport operators, range, fuel economy, emissions, and safety. Today's conventional buses still exhibit relatively poor fuel economy and moderately high emission levels while today's battery-electric buses cannot handle the demands of most transit duty cycles, specifically with respect to vehicle range. Hybrid buses are being developed in response to these challenges. Many hybrid-electric vehicles have evolved over time from initial work in the electric bus arena with assistance from both government and private programs [1]. Additionally, many urban areas where transit buses operate experience air quality problems that are also driving the decision to adopt alternative technologies. These include both alternative fueled traditional buses and hybrid buses. A hybrid vehicle can be set up as a parallel configuration, where both the electric motor and the mechanical engine can drive the wheels, or as a series configuration, where only the electric motor drives the wheels. Different advantages and tradeoffs exist for each combination of components (Table 1 and Table 2) TABLE 1 - SERIES HYBRID [1] Benefits Trade-offs Fuel cell compatible Greater energy losses as more energy passes through the energy storage device than in parallel hybrid Reduced emissions – engine rarely idles and Maximum power at high speeds may only be tends to operate in a narrow peak efficiency available with both the APU and energy band storage device operating Improved low speed acceleration – all the Lower steady state efficiency as generator is power is routed through the electric motor required to convert engine energy to electric providing high torque at low speeds energy and back again into mechanical energy Numerous component layout options, simpler Increased weight – smaller APU will increase packaging reliance on batteries and may require more batteries (weight) and may shorten battery life Most of the countries now import more than half of the oil it consumes annually. The trade imbalance created by oil imports poses a major threat to the nation's economy should foreign sources of fuel become disrupted. Governments have directed federal agencies to develop programs to reduce the nation's consumption of imported fuels. One of the primary strategies has been to support the development of advanced vehicle technologies, like hybrid-electric, that reduce fossil fuel consumption overall and can utilize domestically produced fuels such as natural gas. 134 Research in Transport and Logistics 2009 Proceedings Do u University TABLE 2 - SERIES HYBRID [1] Benefits Trade-offs More overall power – both engine and Generator rotor must spin when using motor can supply power simultaneously energy from engine, adding more mass and rotating inertia Reduced weight – smaller energy storage Less capable of capturing all available capacity possible compared to series hybrid regenerative braking energy Greater energy efficiency during steady Not fuel cell compatible state operation – not all energy must go through generator and electric motor as with series hybrid Greater battery life as batteries are used primarily for regenerative braking and acceleration assist, not for primary acceleration A diesel engine is currently the most efficient power supply available to the transit industry. In a conventional bus, current diesel engines are about as efficient (~30% overall) as they can get, although direct fuel injection will be used in the future to force some small efficiency gains and emission reductions. The reason this increase is expected to be small for conventional buses is that engine efficiency is dominated by the operating cycle and excessive idle time and not necessarily by the peak efficiency of the engine. In a hybrid-electric vehicle, however, the engine is not coupled mechanically to the wheels and can be operated more efficiently or, in some cases, turned off completely. In slow urban drive cycles, nearly 50% of the energy expended by the vehicle is utilized for acceleration while the remaining energy goes to auxiliary systems and road load. If all the kinetic energy could be captured it would potentially double current transit bus fuel economy? Real world system limitations usually result in a maximum capture of about 50% of the available kinetic energy during regenerative braking. By recapturing 50% of the total kinetic energy (25% of the total expended energy), fuel economy is increased by 33%. [5] The European Commission (2001) states in their White paper (European transport policy for 2010), that logistics can contribute to one of the objectives through: reducing the environmental impact of transport (e.g. improved vehicle utilization). Two general approaches for reducing the environmental impact can be identified. The first is to rely on new, more energy efficient technology, which for public transport has proven to be insufficient. The second is to rely on companies to restructure their processes. In the public transport literature (the micro perspective) two methods to reduce the environmental impact are to either introduce more energy efficient technology, or to organize in a different way. However, it is not enough to introduce new technology to stop the development, e.g. more energy efficient engines. There is a need for larger structural changes in public transport. A hybrid-electric bus is defined as carrying at least two sources of energy on board the vehicle, with an electric drive to provide partial or complete drive power to the vehicle's wheels. In most cases the two sources of energy will be an electrical energy storage device and an APU (auxiliary power unit). In a conventional bus, the diesel engine generates power that is mechanically transferred to the wheels through the transmission and differential. In a hybrid-electric drive bus, the engine produces electricity and may not be coupled to the wheels. Power is electrically transmitted to the wheels by a combination of the engine generator set and traction batteries. Although simple, the components of an electric drive system may only be well understood by transit agencies that operate electric trains or trolley buses. Owning Costs Here we will try to provide an understanding of the cost factors involved with owning and operating hybrid electric drive buses. Since the technology is new it is impossible to use hard numbers on the cost to own and 135 Research in Transport and Logistics 2009 Proceedings Do u University operate these buses. Recent buyers have paid between 840.000 and 440.000 USD/bus. But in 1999 New York City Transit paid 385.000 USD/bus for 125 hybrid buses. Fuel By definition HEV’s consumes at least two different fuels, and these different fuels costs must be considered to estimate the operating costs. One is diesel or natural gas, and the other is lead acid batteries. TABLE 3. HYBRID BUS FUEL COSTS [1] Conventional Drive Hybrid Fuel Cost (($/Gal)*1000) 720,00 720,00 Fuel Economy (mpg*100) 350,00 450,00 Annual$/bus 5.554 4.320 ($/mile)*1000 210,00 160,00 $/bus life (*) 83.314 64.800 Cost ($/ACkWh) * 1000 80,00 80,00 Annual MWnr/bus 0,00 2,60 Annual/bus 208 $/mile 0,01 $/bus life (*) 3.120 Replacement Cost($/pack) 50.000 Annual/bus 3.333,00 $/mile 0,12 $/bus life (*) 50.000 Annual/bus 5.554 7.861 $/mile 0,21 0,29 $/bus life (*) 83.314 117.920 (*) Based on 15 years life and 27000 miles per year. $ 140.000 120.000 100.000 80.000 60.000 40.000 Cost Components 20.000 0 Conventional Drive Hybrid FIGURE 1. COMPARISON OF HYBRID BUS FUEL COSTS The hybrid-electric drive definition problem is getting more unclear by the fact that the technology takes many forms and different labels to describe them. For instance, there are series and parallel hybrids, engine-dominant and battery-dominant hybrids, charge-sustaining and charge-depleting hybrids and dual- mode hybrids. All of these are currently under development and/or deployment and each has its advantages. Many perceived advantages are actually a function of the electric drive, which results in more available torque at low speed, smoother acceleration, and efficient regenerative braking. In either parallel or series 136 Total Traction Fuel Cost 720,00 Fuel Batteries Electrical Diesel 350,00 Annual$/bus 5.554 210,00 $/bus life (*) 83.314 80,00 Annual 0,00 $/mile Replacement $/mile Annual/bus 5.554 0,21 $/bus life (*) 83.314 Research in Transport and Logistics 2009 Proceedings Do u University configurations the short-term benefit is increased efficiency due to the capture of kinetic energy through regenerative braking. Hybrid-electric drive systems on transit buses are being aggressively investigated as a means of improving fuel economy, reducing emissions, and lowering maintenance and operating expenses. Several major federally funded research and development projects are testing the viability of these drive systems on buses. With the rapid pace of development and improvement of hybrid-electric drive technology, more transit agencies are being asked to evaluate the potential for hybrid-electric drive systems in their fleets. It is intended to provide transit managers with a better understanding of the technology, including benefits, challenges, and life-cycle costs. In this paper, alternative fuels are considered for their potential to displace the oil is the main and only source of transport fuel. But we want to analyze only the hybrid electric buses among twelve different alternative fuel buses. TABLE 4. AVERAGE VALUES OF WEIGHTS [1] Criterion Manufacture Academic Research Bus Institute Organization Operator Average Energy Supply 0,0357 0,0314 0,0340 0,0249 0,0313 Energy Efficiency 0,1040 0,0943 0,1020 0,0748 0,0938 Air Pollution 0,1355 0,2090 0,1595 0,1605 0,1661 Noise Pollution 0,0452 0,0697 0,0532 0,0535 0,0554 Industrial Relationship 0,0923 0,0357 0,0480 0,0757 0,0629 Employment Cost 0,0900 0,0680 0,0343 0,1393 0,0829 Maintenance Cost 0,0300 0,0227 0,0114 0,0464 0,0276 Road Facility 0,1373 0,0953 0,1827 0,0803 0,1239 Vehicle Capability 0,0827 0,0590 0,1520 0,0283 0,0805 Speed of Traffic 0,1520 0,2420 0,1400 0,2637 0,1994 Sense of Comfort 0,0957 0,0730 0,0833 0,0523 0,0761 The evaluation of alternative-fuel buses should be considered from various aspects as; energy supply, energy efficiency, air pollution, noise pollution, industrial relationship, implementation costs, maintenance costs, capability of vehicle, road facility, speed of traffic flow, comfort. The AHP is used to determine the weights of evaluation criteria. The decision makers of the related industries as bus producers and operators assessed their subjective relative importance for each of the criteria. Table 4 shows the average values of weights. Then the next step is to evaluate the alternatives. 0,3000 Values 0,2500 0,2000 0,1500 0,1000 0,0500 Criterion 0,0000 Manufacture Academic Institute Research Organization Bus Operator Average FIGURE 2. COMPARISON OF CRITERIA VALUES 137 Energy Supply Energy Efficiency Air Polluttion Noise Pollution Relationship Empolyment Cost Maintenance Cost Road Facility Vehicle Capability Spped of Traffic Sense of Comfort Research in Transport and Logistics 2009 Proceedings Do u University TABLE 5. VALUES OF CRITERION FUNCTIONS [1] Alternatives Energy Energy Air Noise Ind. Empoly. Maint. Road Vehicle Speed Sense Supply Eff. Poll. Poll. Rel. Cost Cost Facility Capab. of of Traffic Comfort Diesel Bus 0,82 0,59 0,18 0,42 0,58 0,36 0,49 0,79 0,81 0,82 0,56 Hybrid-Gas 0,77 0,63 0,63 0,52 0,66 0,63 0,65 0,67 0,70 0,80 0,74 Hybrid- Diesel 0,77 0,63 0,51 0,58 0,66 0,63 0,65 0,67 0,70 0,80 0,74 Hybrid- CNG 0,77 0,73 0,80 0,48 0,63 0,66 0,65 0,67 0,71 0,62 0,78 Hybrid- LPG 0,77 0,73 0,80 0,48 0,63 0,66 0,65 0,67 0,71 0,62 0,78 Values 0,90 0,80 0,70 0,60 0,50 0,40 0,30 0,20 0,10 Criterion 0,00 Energy Energy Air Polluttion Noise Empolyment Maintenance Road Facility Vehicle Spped of Sense of Supply Eff iciency Pollution Relationship Cost Cost Capability Traff ic Comfort Diesel Bus Hybrid-Gas Hybrid-Diesel Hybrid-CNG Hybrid-LPG FIGURE 3. COMPARISON OF CRITERION FUNCTIONS The major force behind the development of hybrid drive buses is emissions reduction. Hybrid buses are cleaner and emit fewer grams of pollutants per mile than conventional diesel buses. Hybrids have been shown to reduce particulates and NOx by as much as 50% during testing. Hybrid buses may be very desirable for highly industrialized countries [5]. TABLE 6. HYBRID BUS EMISSION CREDITS [1] Conventional Drive Hybrid PM g/mile 0,24 0,12 ton/year/bus 0,01 0,00 ton/year/100 buses 0,71 0,36 NOx g/mile 30,10 19,20 ton/year/bus 0,89 0,57 ton/year/100 buses 89,40 57,20 VOC g/mile 0,14 0,08 ton/year/bus 0,00 0,00 ton/year/100 buses 0,42 0,24 CO g/mile 3,00 0,10 ton/year/bus 0,09 0,00 ton/year/100 buses 8,91 0,30 CO2 g/mile 2.779,00 2.262,00 ton/year/bus 82,54 37,18 ton/year/100 buses 8.253,63 6.718,14 138 Research in Transport and Logistics 2009 Proceedings Do u University 3.000,00 2.500,00 2.000,00 1.500,00 1.000,00 500,00 0,00 PM (g/mile) NOx g/mile VOC g/mile CO g/mile CO2 g/mile Conventional Drive Hybrid FIGURE 4. COMPARISON OF EMISSION CREDITS Dual mode hybrids may allow limited zero emission range for operation in highly sensitive areas. For example, crowded downtown, tourist or historic areas might particularly benefit from pure electric (zero emission) propulsion mode of operation for short distances. In Boston, a new underground bus transit way is being constructed that will require buses to travel a 1.1 mile tunnel without burning fuel. Dual-mode hybrid buses might be particularly well suited to this type of application. The major impetus behind alternative fuels is interest in reducing harmful emissions, improving public health, and conserving energy. Political, social, and environmental pressures are being placed on transit agencies to adopt alternative fuels and technologies including hybrid-electric drive. Normal operating conditions for diesel buses range from 10°C to 85° C. Hybrid buses can perform in similarly demanding conditions; however, this will likely require some type of thermal management of the batteries. Lead acid batteries become less efficient when ambient temperatures drop below 45° C while other batteries such as nickel metal hydride may require active cooling at warmer temperatures. Regardless, hybrid drive system developers are designing thermal management into their battery management systems. Hybrid transit bus maintenance costs are not well understood at this time. There is not yet enough operating experience with hybrids to realistically quantify the hours and costs associated with maintaining them. In practice, transit agencies experimenting with early hybrid prototypes report higher than normal maintenance costs which are typical for a new technology. Hybrid technology can also have noise reduction benefits over conventional buses. Diesel bus noise standards typically specify that noise levels not exceed 83 dBA at any seat location in the bus at 50 km/h. Comparative noise data for hybrid bus was not available, but hybrid buses are quieter than standard diesel buses in service. V. RESULTS Hybrid-electric drive technology introduces new opportunities and challenges in bus operation and maintenance. Hybrid drive offers operational advantages such as smoother and quicker acceleration, more efficient braking, improved fuel economy, and reduced emissions. Maintenance requirements may initially increase due to energy storage system requirements; however, these may go away once the technology develops. In the long term, hybrid bus maintenance may be less difficult than conventional mechanical technology due to savings associated with transmission and brakes. Infrastructure modifications are expected to be minor. Mechanical and safety retraining is needed in light of high voltage components. Transit providers must understand the issues and risks involved in deploying hybrid drive technology. 139 0,24 0,12 30,10 19,20 0,14 0,08 3,00 0,10 2.779,00 2.262,00 Research in Transport and Logistics 2009 Proceedings Do u University Successful introduction of a new technology into the bus fleet ultimately depends on both the commitment from management and a successful management plan. Hybrid drive will present public transportation managers with new challenges in the areas of operation, maintenance, labor, cost, and public relations. A management plan must address each of these issues and receive the full support of the agency's leadership. Addressing the nuts and bolts of infrastructure, hardware, and operating schedules will ensure success. Other human factors must be addressed as well, such as driver and mechanic acceptance. Employee acceptance can be won through careful training and rewards. Customer acceptance and public relations must also be carefully addressed. Hybrid technology may win supporters among public transportation users, elected officials and environmental constituencies for being quieter, smoother, and better for the environment. REFERENCES [1]Transportation Research Board National Research Council. (2000). Hybrid-Electric Transit Buses: Status, Issues, and Benefits. TCRP Report 59, National Academy Press, Washington, D.C. [2] US Department of Energy. (2007). Fuel Economy Guide. Report DOE/EE*0314, Office of Energy Efficiency and Renewable Energy, Washington, DC. [3]Kurani, K., Turretine, T., & Heffner, R. (2006). Narrative self-identity and societal goals: automotive fuel economy and global warming policy. In D. Sperling, J. Cannon, & J. Cannon (Ed.), Driving Climate Change. Burlington: Elsevier. [4] Heffner, R. R., Kurani, K. S., & Turrentine, T. S. (2007). Symbolism in California's early market for hybric electric vehicles. Transportation Research Part D , 12, 396-413. [5] Environmental and Energy Study Institute. (2007). Hybrid Buses Costs and Benefits. Washington, D.C. 140 Research in Transport and Logistics 2009 Proceedings Do u University OPTIMIZATION OF E-WASTE MANAGEMENT IN MARMARA REGION - TURKEY 1, Müjde EROL GENEVOIS2 Abstract Electrical and electronic equipments have already begun to accumulate at the garbage dumps. This garbage accumulation brings big danger to the environment and to the human health. That’s why we should look for exploring the ways to dispose of these wastes in an environmentally friendly manner. In this research, the recycling and disposal-landfill waste management strategies are chosen as the ways of treating WEEE. A model based on Badran and El-Haggar study with the conception of collection stations for WEEEs is proposed. The aim of this study is to select the places of collection stations by minimizing the e-waste management system cost for different scenarios. The criterion for choosing the optimal scenario is the minimum value of the objective function, which corresponds to minimum cost. Mixed integer programming is used to model the proposed system and its solution is performed by using LINDO. Keywords Mixed Integer Programming, Recycling, Separation Station Selection, Waste Management, WEEE (Waste Electrical and Electronic Equipment). INTRODUCTION Nowadays, technological developments increase day after day and we all see plenty of hi-tech samples in the global market. These frequently sold products created shorter life cycles for them. Hence, electrical and electronic equipments, which are a subset of technological equipments, have already begun to accumulate at the garbage dumps. The UNEP (United Nations Environment Programme) estimates that the world produces up to 50 million tons of Waste Electrical and Electronic Equipment-WEEE (or e-waste for short) per year. The current annual production of e-waste is 1.8 million tons in Germany, 1.5 million tons in France, and roughly 6 million tons in Europe. Unfortunately, quantity of e-waste in Turkey cannot be stated clearly because of the lack of statistical data. However, Turkey's annual e-waste production is assumed to be around 1 million tons, with reference to Germany's e-waste quantity. This assumption is made on the similarity between Turkey’s and Germany’s population and consumption habits. The European Commission identified the need for legislation to address the escalating problem of wastes, especially WEEE at the Community level, and this has taken the form of the WEEE Directive [1]. The EU’s WEEE Directive (waste of electronic and electrical equipment) obligates manufacturers of electronic and electrical equipment to take back old equipment from customers free of charge and to dispose of these wastes in an environmentally sound manner [2]. Turkey, as one of the candidates for EU membership, is also on the way to prepare a new regulation about WEEE for the purpose of adaptation to EU’s legislation. In the near future, this new regulation will be promulgated by Ministry of Environment and every EEE producer in Turkey should reform its organization and take the responsibility of its wastes. The objective of this paper is to propose a model for e-waste management in Marmara region – Turkey, which incorporates the concept of separation stations, by taking base Badran and El-Haggar’s (2006) study [3]. The proposed model will be optimized by minimizing the e-waste management system cost using mixed integer programming. The best location for separation stations is selected from the given candidate locations so that system cost is minimized. In the remaining part of the paper, waste management and its sub-topic e- waste management is mentioned. Section 2 includes some definitions and principles. The mathematical model to optimize e-waste management systems and to select the separation stations is represented in Section 3. Section 4 contains the scope of the application. Section five gives the results of the study. Finally in Section 6, conclusions driven from the research are provided. WASTE MANAGEMENT 1 Galatasaray University, Faculty of Engineering and Technology, Industrial Engineering Department, Turkey, ibereketli@gsu.edu.tr 2 Müjde Erol Genevois, Galatasaray University, Faculty of Engineering and Technology, Industrial Engineering Turkey, merol@gsu.edu.tr 141 Research in Transport and Logistics 2009 Proceedings Do u University Waste management is the collection, transport, processing, recycling or disposal of waste materials. The term usually relates to materials produced by human activity, and is generally undertaken to reduce their effect on health and the environment. Waste management is also carried out to reduce the materials' effect on the environment and to recover resources from them. The purpose of waste management is to reduce the amount of waste being produced, thus reducing the disposal costs and the impact on the environment. This means that reducing waste should be the primary concern. There are two principles in reducing waste, firstly to reduce the quantity generated and secondly to adopt an effective system to manage unavoidable waste [4]. Reuse and recycling are the best methods of dealing with waste that is unavoidable according to the hierarchy. Reduction and reuse are preferable to recycling, which requires waste to be re- processed before it can be reused [5]. The next best disposal route is incineration, which can also be used as a means of generating energy, although it does have some undesirable environmental effects. Finally, disposal in landfill sites is the least desirable option [6]. WEEE and WEEE Management Electrical and electronic equipment or ‘EEE’ means equipment which is dependent on electric currents or electromagnetic fields in order to work properly and equipment for the generation, transfer and measurement of such currents and fields falling under the categories set out in Annex IA and designed for use with a voltage rating not exceeding 1000 Volt for alternating current and 1500 Volt for direct current. [7] Waste electrical and electronic equipment is a generic term embracing various forms of electric and electronic equipment that have ceased to be of any value to their owners, or a waste type consisting of any broken or unwanted electrical or electronic appliance. There is, as yet, no standard definition [8]. EU’s Directive 2002/96/EC defines ‘waste electrical and electronic equipment’ or ‘WEEE’ as follows: WEEE means electrical or electronic equipment including all components, subassemblies and consumables which are part of the product at the time of discarding [7]. In this paper, we will accept the definition of EU’s Directive 2002/96/EC. Waste generated from electrical and electronic equipments is potentially hazardous due to heavy metal (zinc, cadmium, mercury, chromium and copper) effluent content. As such, they require special treatment and disposal techniques. At present, most waste arising from the industrial sector is released into the surrounding atmosphere, discharged into adjacent water bodies (rivers, streams, sea), stored on site, disposed of in privately owned landfills, incinerated in the open, or dumped haphazardly. [9] There are generally three different ways of treating of WEEE: reuse, recycling and disposal (such as incineration, landfill) [10]. It is already said in section 2 that disposal in landfill is the least desirable option [6]. However, in Turkey’s conditions, the reuse and recycling methods are much expensive and underdeveloped because the lack of technologies and know-how. So unfortunately, disposal in landfill is, for now, preferable to the other ways of treating e-wastes. In this search, the focus will be on recycling and disposal – landfill in order to reduce the disposed waste quantity and to dispose the rest in an environmentally friendly manner. MODELING Separation stations are included in the proposed model for the e-waste management in Marmara Region. A schematic diagram of the waste flow of the system is shown in Figure 1. Six potential separation station sites in the chosen districts of Marmara are evaluated by the model to choose the best locations in order to minimize the system cost. QSij QCSLjl Source i Collection station j QCSjk Landfill area l QCk Recycling centre k FIGURE 1 Schematic diagram of the waste flow of the system 142 Research in Transport and Logistics 2009 Proceedings Do u University The objective is to select between the different potential sites for separation stations to minimize the total cost of the entire e-waste management system. The total cost is dependent on the facilities and is minimized by the mixed integer model. Assumptions The assumptions made by Badran and El-Haggar are also accepted in this study, only with one exception. In their seventh assumption Badran and El-Haggar said that average vehicles speeds are determined as 25 km/h for the collection vehicles and 35 km/h for the transfer vehicles. In this paper, the speeds for both collection and transfer vehicles are assumed as 40 km/h. In addition, in this study traffic jam, busy hours are not considered for the duration of the route in hours while transportation cost is calculated. Distances between sources, separation stations, recycling centre and landfill areas are measured by Euclidean metric rule. E-wastes are transferred to separation centres only from the sources in the same city. The only exception is the third separation centre. It is in Istanbul’s Asian side however it accepts also e-wastes from Kocaeli. Formulation The mathematical formulation of the mixed integer model will be as follows: Objective function: J J I K J L J M iF n j D jCZ V S j CQ i j SSV k QC j k C V l SQC j j 1 j 1 i 1 k 1 j 1 l 1 j 1 I J J K J L T i Qj iS j S T j QkC j k CS T S j QCl j lSC L S L i 1 j 1 j 1 k 1 j 1 l 1 Balance constraints: J Q Si j Gi for i = (1, ….., I) (1) j 1 J L Q j kC 1 RSk Q k l C0 for k = (1, …., K) (2) j 1 l 1 J I Q j Ck r1 S Q i j S0 for k = (1, …., K) (3) j 1 i 1 J I Q j Cl (1 rS1 ) L Q i j 0S for l = (1, …..,L) (4) j 1 i 1 Capacity constraints: I Q i jS C Cj DSj for j = (1, …., J) (5) i 1 J Q Cj k S C kC for k = (1, …., K) (6) j 1 J Q C jSl LCl for l = (1, …..,L) (7) j 1 143 Research in Transport and Logistics 2009 Proceedings Do u University Q iSj M jD 0 for i = (1, ….., I); j = (1, …., J) (8) J D j T for j = (1, …., J) (9) j 1 Non-negativity constraints: Q Sij 0; Q C Sj k 0; Q C Sj lL 0; for i = (1, ….., I); j = (1, …., J); k = (1, …., K); l = (1, …..,L) (10) D j 0 or 1 for j = (1, …., J). (11) Decision variables: There are four decision variables: QCSjk : amount (in ton) of monthly e-waste to be transferred from separation station j to recycling centre k (j = 1,. . . ,J; k = 1,. . . ,K) QCSLjl : amount (in ton) of monthly e-waste to be removed from separation station j to landfill l (j = 1,. . . ,J; l = 1,. . . ,L) QSij : amount (in ton) of monthly e-waste to be removed from district i to separation station j (i = 1,. . . ,I; j = 1,. . . ,J) Dj : a variable of zero or one, which is to take one if a separation station is to be set up at candidate location j (j = 1,. . . ,J). Parameters: There are thirteen parameters involved in the system in order to calculate the decision variables: CCk : monthly capacity of the recycling centre CCSj : monthly capacity of the separation stations CLl : monthly capacity of the landfill Gi : amount of monthly waste generated at source i FCSj : fixed cost of the separation station represented as monthly fixed cost VCk : variable cost of the recycling centre represented per ton processed VCSj : variable cost of the separation station represented per ton processed VCl : variable cost of the landfill represented per ton processed TCSjk : transportation cost per ton of waste from the separation station j to the recycling centre k (j = 1,. . . ,J; k = 1,. . . ,K) TCSLjl : transportation cost per ton of waste from the separation station j to the to landfill l (j = 1,. . . ,J; l = 1,. . . ,L) TSij : transportation cost per ton of waste from the source i to the separation station j (i = 1,...,I; j = 1,...,J) T : number of separation stations Rk : the percentage of the waste entering the recycling centre that is rejected. r1 : ratio of WEEE transported from waste separation centre to the recycling centre. (r1 = 0.1 ) APPLICATION Required data The required data used in the application and their explanations are as follows: 144 Research in Transport and Logistics 2009 Proceedings Do u University i : the index which indicates the waste sources. ( i = 1, ….., 15) In Marmara region, Istanbul, Kocaeli and Bursa are the cities which generate the highest quantity of e- waste. For this study, we chose the districts of those cities as sources which produce the largest amount of e- waste and the organised industrial zones. In Istanbul seven districts, in Bursa five districts, and in Kocaeli three districts are determined. The waste generation sources are considered as point sources, located at the centroid of the respective district. The chosen districts can be seen on Fig.1. j : the index which indicates the separation centres. (j = 1, ….., 6.) In separation centres, collected e-wastes are controlled. If they are recyclable, they are sent to recycling centre. If not, e-wastes are transferred to the landfill. By the assumption mentioned before in introduction, it can be said that Turkey's annual e-waste production is assumed to be around 1 million tons. In this study, the optimization is realized on monthly base. Therefore it can be assumed that approximately 83.000 tons of e-waste is generated every month. In addition, regarding the population of Marmara region, it can be assumed that the ¼ of this e-waste quantity, ~20.000 tons, is emerged here. This information is needed for the second data, separation station data. The separation station monthly capacity for e-waste is assumed to be 4000 tons. The location of these separation stations is determined according to adjacency to the most e-waste consuming districts. The location of the separation stations are shown in the map in Figure 2. k : the index which indicates the recycling centre. (k = 1) This data is required for the recycling centre. In Marmara region, there is only one company which recycles the e-wastes. This company, EXITCOM (www.exitcom.com), is located in Kocaeli. In this company, computers, printers, phones, mobile phones, cameras, electronic equipments, monitors, televisions, and home appliances are treated and recycled in the scope of e-waste management. Its e-waste recycling ratio is approximately 90%. Therefore Rk is given 10%. l : the index which indicates the landfill areas. (l = 1, 2) In Marmara region, there are two different landfill areas to dispose of wastes. One is in Kocaeli and the other is in Gebze. The last data are required for the transportation costs. The transportation cost per ton for each route is divided in two parts. One is the transportation cost based on the distance traveled and the other is based on the duration of the trip. FIGURE 2 The chosen districts (red marks) and possible separation stations (blue marks). Transportation cost per ton = Transportation cost per ton per hour duration of the route in hours + Transportation cost per ton per km distance in km. In order to calculate the transportation cost, the fuel cost is taken as 3,45 YTL/liter, the capacity of every long vehicle as 10 tons, vehicle cost as 100.000 YTL, and its useful life as 15 years. The duration of the trip will be calculated based upon average vehicle speed of 40 km/h for all routes. Results The proposed model is solved using LINDO 6.01. The results of the modeling, total costs and selected separation stations are summarized in the Table 1. 145 Research in Transport and Logistics 2009 Proceedings Do u University Table 1 The results of the modeling Total cost (YTL) Selected Separation Stations 85.629.568 D1, D3, D5 CONCLUSION Nowadays, the depletion of resources and the pollution are two serious problems in the world. WEEE, or e-waste for short, is one of the most influential factors on these problems. E-waste is a generic term embracing various forms of electric and electronic equipment that have ceased to be of any value to their owners, or a waste type consisting of any broken or unwanted electrical or electronic appliance. While the hazardous substances in e-waste pollute the environment, their shorter life cycle causes the fast consumption and the continuous overproduction; hence the depletion of resources increases. To cope with these problems, governments should make law about waste treatment and the producers should take the responsibility of e- wastes. In this direction, e-waste management in Marmara region is modelled with the similar concept of having location stations as Bardan and El-Haggar’s model. The model is solved using LINDO and optimal solution is obtained. Hence three of six separation stations are selected. In this paper, the biggest problem is the lack of data about the detailed distance measures between the locations. The distance between different places is assumed to be the distance between their respective centroids and is measured by Euclidean distance. Additionally, traffic jam is not taken into account for the duration of the route in hours while transportation cost is calculated. Then again, the statistical data about the amount of e-waste generation is not known precisely and is assumed according some other data. The result of the presented model is based on some assumptions in the absence of the available data and could change with more accurate data. For the future work, another objective, such as minimizing the possessing time, can be added to the model and the goal programming approach can be used for the selection problem. ACKNOWLEDGEMENT This research has been financially supported by Galatasaray University Research Fund. REFERENCES [1] Darby L., L. Obara, 2005. Household recycling behaviour and attitudes towards the disposal of small electrical and electronic equipment, Resources, Conservation and Recycling, 44, 17–35. [2] Schmidt M., 2005. A production-theory-based framework for analysing recycling systems in the e-waste sector, Environmental Impact Assessment Review, 25, 505-524. [3] Badran M.F., El-Haggar S.M., 2006. Optimization of municipal solid waste management in Port Said- Egypt, Waste Management, 26, 534-545. [4] Teo, M., Loosemore, M., 2001. A theory of waste behaviour in the construction industry. Construction Management and Economics 19, 741–751. [5] Lingard, H., Gilbert, G., Graham, P., 2001. Improving solid waste reduction and recycling performance using goal setting and feedback. Construction Management and Economics 19, 809–817. [6] Pitt M., Smith A., 2003. Waste management efficiency at UKairports, Journal of Air Transport Management 9, 103–111. [7] Directive 2002/96/Ec of the European Parliament and of the Council of 27 January 2003 on waste electrical and electronic equipment (WEEE), Official Journal of the European Union, (13.2.2003), L 37/24- 38. [8] Widmera R., H. Oswald-Krapf, D. Sinha-Khetriwal, M. Schnellmann, H. Böni, 2005. Global perspectives on e-waste, Environmental Impact Assessment Review, 25, 436-458. [9] El-Fadel M., Zeinati M., El-Jisr K. and Jamali D., 2001. Industrial-waste management in developing countries: The case of Lebanon, Journal of Environmental Management 61, 281–300 [10] He W., G. Li, X. Ma, H. Wang, J. Huang, M. Xu, C. Huang, 2006. WEEE recovery strategies and the WEEE treatment status in China, Journal of Hazardous Materials, B136, 502-512. 146 Chapter 7 Contemporary Topics in Transport and Logistics Research in Transport and Logistics 2009 Proceedings Do u University FUTURE PROSPECTS ON URBAN LOGISTIC RESEARCH 1, Vasco REIS2 Abstract - urban logistics is a vital activity for the existence of urban regions. Yet, the very activity that maintains urban regions living is also endangering it. Solutions are therefore required to maintain and develop the current level of urban logistic activities without jeopardising economic and social development. Academia plays, in this sense, an important role on the design of those solutions. In this paper the authors discuss the main challenges currently affecting urban regions and discuss the principal solutions put into practice so far. Yet, urban logistic research is only now giving its first steps, and therefore suffers from the symptoms of a non-mature research area: lack of models and tools, absence of consensual definitions, lack of reliable data and data gathering methods. The authors in this paper identify seven research streams that if achieved would conduct the field of research on urban logistics to a next level of development and maturity. Keywords challenges, road map, urban logistics INTRODUCTION Nowadays in Europe, urban areas concentrate, around 80% of the population, similar figures are found on other developed regions. In the developing regions, which traditionally had a larger rural population, people has been emigrating towards cities and urban regions in search for better employment and living conditions. Social activity generates a great amount of demand for transport services, either related with human living, such as food, clothing, toys, furniture or waste transport; or related with business and industrial activity. The growth of urban regions has resulted in a continuous increase of freight transport activity [1][2]. Yet, the progressive increase of freight activity is no longer compatible with the current expectation and quality patterns of the population, particularly in developed countries and regions. Urban logistics activities hree dimensions of the concept: environmental, economic, and social. The main source of the problems is the physical flow of cargo within the urban region. Nowadays the large majority of those flows are done by road vehicles. A major problem is the worsening of congesting problems of the urban road network, particularly in the old town centres were road tends to be narrower. Road infrastructure is a scarce resource that private cars, public transportation vehicles and freight vehicles have to share. Congestion reduces quality of life and negatively affects economic development. The increase of freight transport activity natural increases the number of conflict with other users. A second factor is the visual and sound intrusion of freight transport, and loading and unloading activities. Heavy trucks are noisier than private cars and therefore disturbing the quality of urban areas. This problems is worsened by the lack of suitable parking spots in many urban regions forcing truck drivers to parking vehicles anywhere possible (even going against the law), which is the source of many conflicts with both walking citizens and other traffic. Urban freight transport activities are nowadays essentially based on road services, which have to share transport infrastructure with other users private cars. Furthermore, freight transport road vehicles tend to provoke more damage on infrastructure than private vehicles (both because are heavier and tend to have longer stops), increasing the public budget efforts in road maintenance. A growing number of complains and demands have been raised against the urban freight activities, searching for new solutions, less intrusive and damaging. Although, initiatives have been implemented and research efforts have been put into practice, yet so far with limited results. The truth is that knowledge on urban freight is still incipient and research in its childhood. So, some time will be required until positive measure could be implemented. DEFINITION OF URBAN LOGISTICS An evidence of the initial stage of research and knowledge on a certain field is the lack of a consensual and universal definition on the object of analysis [3]. And, this situation is visible in urban logistics. Several 1 , Instituto Supe , Technical University of Lisbon, Department of Civil Engineering and Architecture, Lisbon, Portugal, rosariomacario@civil.ist.utl.pt; 2 Vasco Reis, , Technical University of Lisbon, Department of Civil Engineering and Architecture, Lisbon, Portugal, vreis@civil.ist.utl.pt 149 Research in Transport and Logistics 2009 Proceedings Do u University definitions may be found on the literature. The Organisation for Economic Co-operation and Development, through the Work Group on Urban Logistics, proposed in 2003 the following definition for urban logistics [4]: The delivery of consumer goods (not only by retail, but also by other sectors such as manufacturing) in city and suburban areas, including the reverse flow of used goods in terms of clean waste . Several conclusions may be drawn from this definition. Firstly, urban logistics is restricted to urban areas. Therefore, freight flows that either cross; or start or end outside those regions are only within the scope of urban logistics while are part of longer logistic chains that either start or end within (or across) urban regions. Secondly, every freight flow produced within urban areas is considered as being urban logistics. Therefore, urban logistics is a very wide range concept, including a large diversity of situations with not necessarily any common point with the exception of being conveyed in an urban region. Thirdly, urban logistics also include the growing field of reverse logistics. Yet, this definition has some limitations. Firstly, no mention is made to the agents in the context of urban logistics. And one of the challenges when researching on this area is precisely the large variety of agents involved. A definition on urban logistics should refer the agents. In this paper, all agents (both public and private) that have interests in urban logistic activities, are considered. Secondly, the definition reduces logistics to the physical flows. Yet, similarly to what happens in logistics, other flows occur in parallel: information flow which corresponds to the exchange of information amongst agents, and contractual flow which corresponds to the liability and contractual terms established between agents. The very existence of these flows is recognised in the definition proposed by the Council of Logistics Management CLM [5]: is a part of the supply chain management process responsible for the planning, implementation and control of the efficient flow and storage of goods, and related services and information, . In this paper, as a result of previous research work3, the authors have adopted the following wide scope definition for urban logistics activity: urban logistics is the process of distributing goods(including not only retail, but all economic sectors within cities and urban regions, including reserve flows of waste or not used goods. The process of distribution entails three types of flows: physical, logical and contractual. Finally, urban logistics agents are all those (private and public) agents that have any sort of interest, either commercial, social or other. MAIN CHALLENGES Urban logistic activity is inherently complex. Such nature raises several challenging difficulties which makes difficult the implementation of solutions. Firstly, urban logistics is essentially a private sector, where agents are profit driven. They will seek to maximise profits and competitive positioning within market. Therefore, logistic companies have no incentive in engaging for more sustainable or less intrusive solutions, if that represents an increase of costs. By the very opposite they will strive to pass to others (mainly Society) all possible costs, generating externalities to the society as a whole. And, indeed, many of the urban logistic related problems occur because the responsible for causing it, are not paying the full costs of its action. Therefore, they have no incentives to change behaviour. This raises the need for regulation, so that externalities are properly incorporated in the activities that are provoking them. However, public intervention in urban logistics has had so far a rather narrow scope, being [2]. Moreover, there is a strong belief that an increase of costs in logistic activities would result in a decrease of economic competitiveness for the companies and, ultimately, of the whole regions. Secondly, urban freight logistic problems are mainly dealt at local level, with local political authorities (such as municipalities). And, although the problems generated by urban freight activity being specific of the region, a national or supra-national would yield better results. This because certain solutions may require the intervention of higher levels on the political hierarchy, like for instance: changes in national legislation, or major urban interventions. Moreover, local authorities often lack the resources (equipment, budget) and knowledge to study and implement potential solutions. The absence of national or supra-national political intervention is patent in Europe, where there is a European Common Transport Policy that only deals with inter urban transport flows. 3 Done by the authors to the National Science Foundation in Portugal. 150 Research in Transport and Logistics 2009 Proceedings Do u University Thirdly, urban freight sector is made of complex labyrinth of agents with often divergent perspectives and goals. The main agents are: transport companies, shippers, clients, citizens, local authorities. Each group is by itself very complex, for example: transport companies range from global logistic companies (such as FedEx or UPS) to small delivery companies with few vehicles. In what the clients, these can be local shoppers that receive a delivery per week, until major retail shopping centres that receive several trucks per day. The key for the success on the implementation of solution relies at large extend on building consensus on the various groups of agents. Otherwise, if one or more groups do not agree or are suspicious on the intension of a certain measure, the most natural reaction is boycotting those initiatives [6]. Fourthly, urban logistic activity is very complex, and remains to some extend unknown, which is a consequence of the multiplicity of agents. Logistic operators have been developing tailored logistical solutions [6]. Such evolution results in the existence of multiple flows of freight in an urban area, ranging from a single transport to a certain location within urban area, to complex routes with several stopover where some unloading or loading activity occurs, like for example: post delivery. Therefore, there are no universal solutions because each type of flow requires a particular way of approach. Fifthly, the actual structure of an urban region is the result of a dynamic evolutions spanning in many case several centuries. Each step of development was done upon the existent infrastructure and with a certain purpose. An incremental process with different degrees of irreversibility at each step of the process. Naturally, in the past, neither the motorisation levels nor the sustainability awareness was as relevant as it is nowadays. Therefore, many cities lack suitable infrastructure for the realisation of urban logistic operations (namely, parking spots, terminals, sidewalks for small vehicles, etc.). Therefore, it will be necessary to redesign urban spaces or find creative solutions to overcome these physical limitations. Sixthly, the validity and applicability of any research study depends upon the existence of reliable and useful information and data from the real world. Yet, there is scarcely available (and reliable) data on urban logistics activities, to large extend because this sector is done on a private basis. And, even when exists, data series tend to be recent, and normally do not coincide between regions (as this is normally a local initiative). This situation is in sharp contrast with inter-urban logistics (and freight transport) activity, where for almost every country (and even supra-national institutions such as the European Union, World Bank, OCED) there is a long tradition of collecting data. It will take some time until a coherent and adequate body of data would be built. WHAT HAS BEEN DONE SO FAR GOOD PRACTICES In order to tackle the challenges presented above, rationalisation of the distribution process (from the economic, spatial and temporal perspectives) is required. This means reducing the flow of goods yet keeping been pointed out in several European cities. These solutions are not only aimed at the transport activity, but to the organization of the whole logistic chain. applied, i.e., several measures were implemented in a combined form, as part of a broad political strategy for urban logistics. TABLE 1 Measures implemented to solve the urban logistics problems Type of Measure Examples Legislative and Cooperative logistic systems, encouraging night deliveries, public-private organizational measures partnerships, intermediate delivery depots. Access restriction Access restrictions according to vehicle characteristics (weight or volume), measures conditioning access to pedestrian areas, urban tolls, periodic restrictions. Territorial management Creation of loading and unloading areas, of load transfers, and mini logistic measures platforms. Technological measures GPS, track and tracing systems, route planning software, intelligent transport systems, adoption of non polluting vehicles and vehicles adapted to urban characteristics (size and propulsion). Infrastructural measures Construction of urban distribution centres, and peripheral storing facilities, use of urban rail for freight (freight trams), underground freight solutions. 151 Research in Transport and Logistics 2009 Proceedings Do u University Generally, the different authors try to group the adopted solutions in accordance to different criteria [6][7][8]. In this work, we decided to group the measures identified according to their focus of application. This systematization can be found in Table 1, where they were placed according to the degree of intervention needed for their implementation (i.e., Some of the measures identified are very ambitious, while others lack the needed coherence and fail to achieve the goals for which they have been designed. This happens for a number of reasons, like the relative novelty [9] of the introduction of this subject on the urban management agenda, and also the lack of knowledge about the implementation processes involved in urban logistics and hence on the appropriate way to tackle the collateral problems accruing from these processes. Nevertheless, it is already possible to draw some notes on the nature of initiatives already taken. Cooperative logistic systems These correspond to cooperative schemes arranged amongst logistic agents, with the purpose of optimising (both in spatial and temporal terms) their freight distribution networks. The cooperation may be achieved at several levels, like for example: sharing of infrastructure or vehicles, or sharing of information [5]. Cooperative systems may result in a reduction of the costs of operations, due to both an increase of efficiency on the utilisation of resources and on the existence of economies of scale. On the other hand, this Yet, these arrangements reduce the individual visibility of each agent, because the operations and customer relations are carried out by one of the agents. Moreover, there is an increase of communication costs, and complexity on the management of logistic operations. The need of aligning and coordinating the operations of several individual agents can be a difficult task. Night delivery Night delivery has the main advantage of making use of road infrastructure in periods of lower or no congestion. Yet, freight operations tend to be noisy which can disturb sleeping residents. This drawback may be overcome with the adoption of proper technological solutions. Other problem is the increase of costs to the clients, because they will have to have workers to process the cargo. Public private partnerships Public private partnerships initiatives are particularly attractive in those situations that require large investments and simultaneously have high risks. Public authorities and private investors agree on the share of risks and level of investment of each party. Often this is the solution to attract private agents into projects that otherwise would have to be managed by public authorities. The participation of private agents is positive not only because normally higher levels of efficiency if obtained, but also to promote economic development. Urban logistical centres and urban terminals The implementation of medium to small size terminals within urban regions if used by more than one agent can generate important benefits [10]. Flows can be concentrated in larger trucks (thus reducing the number of vehicles in transit) to terminal and local distribution can be done using lighter vehicles (non- motorised). Moreover, these facilities increase the flexibility of logistical systems in the sense these facilities can be used as drop off and collection points for clients. Yet, the creation of these facilities often require the involvement of public authorities owing to the larger initial and maintenance investments, which can be overcome through public private partnerships. Legal Measures Simple changes on the legal framework for the access to road infrastructure may yield important benefits, namely, reduction of conflicts with other users; reduction of noise, emissions and congesting; or optimisation of public space. The drawbacks are an increase on the operations costs on both logistical agents that may have to change the logistical schemes or acquire new equipment and clients because they may have to extend their working period to receive cargo. Moreover, the success of these measures depends on an adequate enforcement. 152 Research in Transport and Logistics 2009 Proceedings Do u University These measures normally include the imposition of restrictions on the access. These restrictions can be either temporal (when logistic activities are banned during dome periods of the day), spatial (when logistic activities are prohibited in some urban zones) or technological (when only certain vehicles can access a specific zone). Technological Measures Technologies measures can provide important solutions on the reduction of the environmental footprint of urban logistics activities. Technological measures include intelligent transport systems (ITS), information technologies (IT) and new transport vehicles (NTV). ITS and IT enable higher integration and seamlessly of flows amongst agents. Logistical systems may be optimised being possible to reduce operational costs. H with a reduction on the number of necessary vehicles. NTV concern the utilisation of new types of vehicles. These vehicles can have a different design, especially made for urban utilisation, or use new propulsion solutions, such as hybrid or electrical. NTV can help reduction pollution and noise problems on urban regions, yet, they still contribute for congestion and other problems. WHAT IS MISSING - ROAD MAP FOR FUTURE RESEARCH Urban logistic sector is at a crossroad, on the one side, demand is raising with more and more need for logistical services, on the other side, populations and governments have been demanding a reduction of the burden generated by those very logistical activities. Although, we may envisage a tense relationship amongst these two parties, a dynamic equilibrium has to be found, in order to avoid jeopardising both urban living standards and economic development. Academia at large and research in particular may provide an important help by depicting new solutions. Research on urban logistics is still in its childhood and a long way lies ahead until this field of research could reach the current level of development of other areas such as logistics. The authors have identified various needs of research, in order to help close the gap of knowledge and, in this way, promote the development of better solutions. Modelling and simulation tools are nowadays a fundamental tool on the development and testing of potential solution in several fields of knowledge. However, very few steps have been achieved so far on the development of robust models. As already written one of the problems lies on the youth of urban research logistics, while the other lies on the lack of suitable data. Yet, owing to the large variety of actors and functions on urban logistics activities, the task of gathering data is daunting. Therefore, one main stream of research is the development of techniques for collection of data on urban logistics activities. Data should comparable across regions and activities, so that multiple analyses could be carried out. Other stream of research is naturally the development of models for the analysis of urban freight flows, and the evaluation of possible solutions. Once again, the multiplicity of urban logistic activities made the development of a universal model virtually impossible. Therefore, specific business case models should be developed, and only afterwards integration can be sought. Horizontal models that could analyse all types of traffics would be very important, so that conflict analyses could be carried out. In this sense, the relevance of the applicability of the traditional four steps models, whose philosophy is still nowadays widely used in the transportation field (people and logistics flows), should be adequately validated in the field or urban logistics. Several doubts have been put on the applicability of this model to urban logistics, namely, the fact that flows are generated in terms of origin and destination, while a large part of urban logistics have one origin and several destinations, along a path. This problem leads us to a next area of research which is related with the generation of flows of goods. There is a major lack of robust models to generate flows of goods within urban regions. This absence limits the validity of any evaluation of a certain solution. Another line of research concerns the integration of urban logistics activities with inter urban logistic activities. This urban part of long logistic chains is normally called as last mile, at present a major problem for both private companies and public authorities. The level of congestion and lack of suitable infrastructure makes delivery times unreliable and difficult to manage. Thus companies have to implement redundant systems or increase the number of vehicles to ensure the maintenance of quality levels, which ultimately results in an increase of the costs. A higher number of vehicles or illegal parking increases the burden over 153 Research in Transport and Logistics 2009 Proceedings Do u University urban environment. Therefore, a seamless integration and management of the last mile along with the rest of the transport chain is a win-win situation. Urban logistic activities are, as already written, essentially private. Therefore, the success of any initiative depends on the maintenance of the competitive level and profit of the private agents. Nowadays, there is little knowledge on how to design urban logistics solutions that do not affect the competitiveness of private agents [6]. A deeper understanding on the actual strategies, competitive forces and business models of private agents would be of utmost importance for the development of business models for the implementation of urban logistic activities that would be neutral for the competitive balance between the agents. A final stream of research concerns the regulation of urban logistic sector. Regulation and legislation may be used to promote the adoption of better urban logistic schemes, for instance, laws may define maximum levels of emissions, type of contracts, or nature of service. Yet, no specialised knowledge has been created on this area. In summary, urban logistics is a vital activity for the existence of urban regions. Yet, the very activity that maintains urban regions living is also endangering it. Solutions are therefore required to maintain and develop the current level of urban logistic activities without jeopardising economic and social development. Academia plays, in this sense, an important role on the design of those solutions. In this paper the authors discuss the main challenges currently affecting urban regions and discuss the principal solutions put into practice so far. Yet, urban logistic research is only now giving its first steps, and therefore suffers from the symptoms of a non-mature research area: lack of models and tools, absence of consensual definitions, lack of reliable data and data gathering methods, and lack of a sound research tradition. The authors in this paper have identified seven research streams that af achieved would conduct the field of research on urban logistics to a next level of development and maturity. REFERENCES [1] . Transportation, 35, 269-284 Transportation Research Part A: Policy and Practice, 41, 280-285. [3 ? A review of intermodal rail - 34. [4 st Paris. ISBN: 92 64 10280 9. [5 Master Thesis, Technical University of Lisbon. [6] Beelen , M., Meersman, H., Van de Voorde, E., Vanelslander, T., Vergauwen, B. and Verhetsel ew concepts in city logistics and distribution: which are the promising best practices? [7 ties, 22, 15-28 [8 Oporto. [9 Inter- and Intra- City Freight Distribution Networks Work Package 1: Final report Comparative ogistics and freight transport policy in urban areas: a case study of Berlin-Brandenburg / Germany European Planning Studies, 12, 1035-1053 [11] Hensher, D. and S. Puckett (2005) Refocusing the modelling of freight distribution: Development of an economic- based framework supply chain behaviour in response to congestion charging . Transportation, 32, 573-602. 154 Research in Transport and Logistics 2009 Proceedings Do u University AN ANALYZE OF RELATIONSHIP BETWEEN CONTAINER SHIPS AND PORTS DEVELOPMENT Branislav DRAGOVIĆ1, Vesna Dragović-Radinović 2, Dušanka Jovović3, Romeo Meštrović4 and Emir Čikmirović5 Abstract ⎯ This paper focuses on the most important trends affecting the deep-sea container shipping industry with reflecting on significant increase in size of the ships employed as well as container port and terminal development. Such changes in the size and capacity of the largest containership were happened due to the strong growth in world container flows and ports throughput, the economies of scale in container ship operations, increasing competitive pressures in the liner shipping and port industry. The evolution in supply chains and logistics models urges liner shipping and container terminals to reconsider their function in the logistics process described with the trends in mega ships and mega container cranes in ports. The strategic analysis regarding the latest development of ULCC’s, deployment patterns, and its implications in terms of shipping structure, impact on port authorities, infrastructure requirements and container terminal operational efficiency is given in this paper. Keywords ⎯ Containerports and ships, Changes in ship and port environment, Relationship between ship and port 1. INTRODUCTION Container ships have reached a fifth or sixth generation development. Although some ports and terminals may face the problem of installing advanced systems for handling large numbers of containers per ship call, most ports will be providing for more modest service demands where present ship-to-shore container cranes and ground storage of containers will suffice. While these large containerships will give a low cost per ton mile if they are running at capacity, their efficiency is impaired if their mode of operation is to call at several ports at each end of the journey to accumulate cargo. Container terminals can generate substantial amounts of land traffic mode, particularly when compared with logistics activity at regular container service. The liner shipping and port industry has been reshaped by logistic integration, containerization, deregulation and globalization. Liner shipping companies and container terminal operators have become the most important part in the international supply chains, which are complex and logistics models. The evolution in supply chains and logistics models urges liner shipping and container terminals to reconsider their function in the logistics process described with the trends in mega ships and mega container cranes in ports. These performances require the new ideas and concepts in container terminal planning in order to keep pace with the development of the mega container vessels. All before mentioned is summarized on Figure 1 [6] and [7]. This figure describes the close connection between trends in container shipping and development of container maritime and intermodal logistics. Figure 2 gives maximal container ship sizes by year of built [6] and [7]. This paper is organized as follows. In Section 2 we give a brief description of the container ship development. In Section 3 we review the related world container port and discuss main topics regarding the world maritime container trade. Section 4 evaluates and compares various cases of relationship between port and ship systems. This implies a visual impact of what has happened to ship-to-shore cranes (QC – quay crane), container ships and terminals in 45 years. Thus, we present effect on container terminals with numerical results and computational experiments which are reported to evaluate the efficiency of the Pusan East Container Terminal (PECT) layout. Finally, we conclude with the consideration about future container port trends in the coming years. 1 Branislav Dragović, University of Montenegro, Maritime Faculty, Maritime Transport & Traffic Department, Kotor, Montenegro, bdragovic@cg.yu; branod@cg.ac.yu 2 Vesna Dragović-Radinović, University of Montenegro, Maritime Faculty, Kotor, Montenegro, bdragovic@cg.yu 3 Dušanka Jovović, University of Pristina – Kosovska Mitrovica, Faculty of Law, Department of Economics, dusankaj@cg.yu 4 Romeo Meštrović, University of Montenegro, Maritime Faculty, Mathematical Department, Kotor, Montenegro, romeo@cg.ac.yu 5 Emir Čikmirović, Maritime Faculty, Kotor, Montenegro, ciki@cg.yu 155 Research in Transport and Logistics 2009 Proceedings Do u University FIGURE 1 FIGURE 2 Background of container ships operation Maximum container ship size by year of built 2. CONTAINER SHIPS DEVELOPMENT The capacity of world container fleet is more than 10 million TEU in operations and about 3.0 million TEU on order in the 2007 and 2008. World cellular fleet capacity is around 9.6 million TEU. According to the carrying capacity cellular ships take part with more than 75% in the structure of the world container fleet. This represents the increase of more than 15% than in 1995. So, in the period of 1995 to 2007, the carrying capacity of cellular ships increased for more than 3.5 or from 2667602 TEU to 9610813 TEU. Also, average TEU per ship in 1995 was 1548 and, in 2007, an approach to 2399 TEU per ship or it is increasing of 64%. While, at the end of 1995 there were 392 cellular ships with the carrying capacity more than 2500 TEU which presents the total capacity of 1296635 TEU, at the beginning of 2007, there were 1128 ships with the carrying capacity above 3000 TEU with total of 5764656 TEU or almost 60% of total cellular fleet capacity. The largest mega container ships in service today with capacities between 8680 and 14500 TEU, now represent 2.5% of the world fleet of container ships, 9% by total seagoing capacity of cellular ships. Also, the largest of the ships currently on order have capacities of more than 12000 TEU, and there is clearly scope for even larger ships. Figure 3 shows container ship development in relation to generation of ships [5]-[7], [10] and [11]. The composition of cellular fleet is presented in Figures 4 and 5 [5] and [6]. Figures 6 and 7 demonstrate projection of containership fleet growth from 2006-2008 [5], [6], [11] and [12]. 800 700 2500000 600 2000000 500 1500000 400 300 1000000 200 500000 100 0 0 <499 500- <499 1000- 01.01.2007 500- 01.01.2007 999 1500-1000- 1499 2000- 01.01.20061500- 1999 3000- 4000-999 1499 2000- 3000- 01.01.2006 2999 3999 5000- >7500 1999 49992999 4000- 74993999 5000-4999 >75007499 FIGURE 4 FIGURE 5 Deliveries of CSs by number of Deliveries of CSs by carrying ships per ship classes capacities per ship classes 800 700 2500000 600 2000000 500 1500000 400 300 1000000 200 500000 100 0 0 <499 <499 500- 01.01.2007500- 1000- 01.01.2007 1000- 999 1500- 1500- 1499 2000- 3000- 01.01.2006999 1499 2000- 3000- 01.01.2006 1999 2999 4000- 1999 4000- 3999 5000- 4999 >75002999 3999 5000- 74994999 >75007499 FIGURE 6 FIGURE 7 FIGURE 3 Number of ships per ship Carrying capacity per ship Container ships development classes 2006 & 2007 classes 2006 & 2007 3. WORLD CONTAINER PORTS DEVELOPMENT World container throughput had almost 6 times in the last twelve years and reached 417 million TEUs in 2006. The world top six busiest ports in 2006 are the East Asian ports surrounding China, which accounts for 26% of the world container throughput (see Figure 8). 156 Number of ships Number of ships TEU TEU Research in Transport and Logistics 2009 Proceedings Do u University FIGURE 8 FIGURE 9 Major World Container Ports in 2006 with container traffic World Maritime Container traffic and 20 development: 2002-2006 [1], [2], [6], [7], [9], [10], [13] and [14] Top World Ports: 1973-2006 [2], [6] and [7] Over time, the use of oceanborne containers to transport international trade has affected the distribution of total maritime trade among world container ports. The increasing number of container shipments causes higher demands on the seaport container terminals, container logistics and management, as well as on technical equipment. So, Fig. 9 shows trend functions of world maritime container traffic in last 33 years. Also, container traffic of the 20 largest seaports in the world from 1973 to 2006 is shown on Fig. 9. Five decades ago, the containership revolution started in the world, changing how the world handles international freight transportation. In 2006, world maritime container traffic was estimated at 417 million TEU, 10 percent more than the 378 million TEU transported in 2005. The world container port ranking changed significantly between 1995 and 2006. The total container traffic volume of the 20 top ranking world container ports reached 205.3 million TEU in 2006. These ports increased their handling volume by 10.4 per cent compared with results in 2005, representing approximately 50 per cent of the total world container traffic. New container ports gained momentum (like Tanjung Pelapas) and other ports established their role as international load centres. Today the largest world container ports are located in Asia. Like in previous years especially Chinese ports saw a constant growth at a very high level. Even though the container traffic of the Chinese container ports grew on average by 25 per cent, the growth did not reach last years values. All in all, the strong growth of Chinese container ports is dominating port traffic not only on a regional but on a worldwide level. 4. AN ANALYSE OF RELATIONSHIP BETWEEN CONTAINERSHIP AND TRANSSHIPMENT HUB PORT Generally speaking, the competitive environment for container ports has changed drastically in recent years. Particularly, in the area of container shipping, a port will gain a significant share of the business only when it can demonstrate a combination of rates, facilities, and inland connection that create a clear competitive advantage for an identified group of customers. Figure 10 gives a visual impact of what has happened to ship-to-shore cranes (QC – quay crane), container ships and terminals in 45 years [6] and [7]. Structural engineers have had the task of keeping up with size, weight reduction, fatigue life, etc. The civil engineers keep driving more piles to support what the structural engineers develop. This figure is a base reference here for illustration and comparison. It shows three basic subsystems: QCs, container ships and terminals. A way has been found to calibrate the combined effect of container shipping (ship development), port container terminal (terminal development) and ship service time with handling costs (average cost per ship service) on the main container port link (berth-ship including different layout of container yard for container storage) for main four generation of ship, terminal and container quay crane evolution, by making certain assumptions which are described in Figure 10. The relationship between ship size with average cost per ship served and terminal system costs is available from Figures 10-i and 10-l, while Figure 10-j shows that the shorter the route length, the flatter is the line graph showing total shipping cost per TEU (as one would expect, this implies that the economies of ship size are of greater benefit on longer routes (see, [3] and [4]). 157 Research in Transport and Logistics 2009 Proceedings Do u University FIGURE 10 The relationship between ship size, container quay crane and terminal layout development Graphs 10-i,j,k and l provide quantified sets of curves for average ship cost in port, total port or terminal system cost, container ship cost per TEU related to size of ship in relation to a number of different lengths of voyages and ship service time in hours vs. lifts per hour. They are based from previous investigations [3], [4] and [5]. Figure 10-a shows the layout of the New Basin Container Terminal, comprising four berths offering berthing space to four second-generation container ships or, alternatively, to three third-generation container ships. It can be seen from drawing that the port opted for a pier-type configuration. Figure 10-a shows the layout of the New Basin Container Terminal, comprising four berths offering berthing space to four second-generation container ships or, alternatively, to three third-generation container ships. It can be seen from drawing that the port opted for a pier-type configuration. This type of terminal focuses attention to some areas of terminal performance for improvement of general layout from perspective continuous evolution of ship size. Consequently, Figures 10-b – 10-d give a visual impact of what has happened to terminal layout in 30 years. The new concept port layout regarding efficiency handling system can be explained in two variants (see Figures 10-c and 10-d). Container quay cranes can be either renovated or developed as new concept that are different from existing methods in order to facilitate improved productivity in berth (Figures 10-e – 10-h). 4.1. Effect on container terminals development The results of this study imply that the economies of container ship operations are now, and are likely to be, such that terminal operators must provide excellent service guaranteeing safety, on-time service, and accuracy. To do that a hub ports facilitate adequate port facility, equipment, and handling system. In addition, average cost per ship service and any costs involved in transshipment must be minimized. 158 Research in Transport and Logistics 2009 Proceedings Do u University This subsection gives a ship-berth link modeling methodology based on statistical analysis of container ship traffic data obtained from the PECT. Implementation of the presented procedure leads to the creation of a simulation algorithm and analytical methodology that captures the ship-berth link performance well. The efficiency of operations and processes on the ship-berth link has been analyzed through the basic operating parameters such as average service time, average QC productivity and average number of QCs per ship. PECT is big container terminals with a capacity of 2008573 twenty foot equivalent units (TEU) in 2005. There are five berths with total quay length of 1500m and draft around 14-15 m, Figs. 11 and 12 [8]. Ships of each class can be serviced at each berth. The input data for the both simulation and analytical models are based on the actual ship arrivals at the PECT for the ten months period from 1 January 2005 to 31 October 2005 (Fig. 11) and 1 January 2006 to 31 October 2006 (Fig. 12), respectively [8]. This involved approximately FIGURE 11 FIGURE 12 1225 ship calls in 2005 and 1285 in PECT layout, 2005 PECT layout, 2006 2006. The ship arrival rate was 0.168 ships/hour in 2005 and 0.176 in 2006. Total throughput during the considering period was 1704173 TEU in 2005 and 1703662 TEU in 2006. Also, the berthing/unberthing time of ships was assumed to be 1 hour. The ships were categorized into the following three classes according to the number of lifts: under 500 lifts; 501 – 1,000 lifts; and over 1,000 lifts per ship. Ship arrival probabilities were as follows: 24% for first class, 40% for second and 36% for third class of ships in 2005 and 30% for first class, 38% for second and 32% for third class of ships in 2006 [8]. 25 90 3.5 2006 2006 2006 3.3980 2005 82.91 3 3.1420 200520.06 70 2005 77.39 2.5 2.69 17.73 60 62.19 2.52 15 50 256.66 1.93 12.5 13.47 1.76 10 40 40.96 1.5 7.85 8.64 30 37.55 1 5 20 0.5 10 0 0 <500 lifts 500 - 1000 lifts >1000 lifts 0 <500 lifts 500 - 1000 lifts >1000 lifts <500 lifts 500 - 1000 lifts >1000 lifts FIGURE 13 FIGURE 14 FIGURE 15 Average service time of ships Average QCs productivity QCs assigned per ship Container terminals in Busan Port, especially PECT, are trying to expand capacity and increase performance at a maximum of investments. Often the container terminal operations are changing to meet increased customer demands as well as to adapt to new technologies. Reasons for the decrease of the average cost per ship served with the introduction of new container berth, QCs, container yard area and automated stacking cranes (ASC) include that waiting time of ships and the average time that ships spend in port decrease with the advanced handling systems improving the operations procedures (see results from Figures 13, 14 and 15) [6, 7 and 8]. A complete description of the simulation and analytical models is provided in [8]. 5. CONCLUSIONS Containerization continues to grow as it has done for a long time. While the entire world merchant fleet grew by only 1%, the containership fleet expanded by more than 10% and this continuous growth has also led to the development of very large containerships. Numerous studies indicate that mega container ship construction is not only feasible but may be a necessary development if this market expansion is to be accommodated in the most cost effective manner. After deploying above 12000+ TEU ships, how should ocean carriers reorganize the shipping network design? Will the bigger vessels lead to a reduction in port calls and how will the terminal operators respond? The impact of next generation ships on shipping lines and container terminals is examined. 159 Average service time of ships in hours Average QCs productivity per ship per hour Average QCs asigned per ship Research in Transport and Logistics 2009 Proceedings Do u University The challenges facing the shipowner and the port operator are certainly real. But, it may be concluded that container shipping technology has undergone changes mainly in the scale of ships and terminals used but very little change in the technology itself. Changes in containership technology will probably not occur in the next 5 years, but radical changes in port, terminal and feeder operations must be expected to happen soon. The objective of most changes will be to: Reduce transshipment handling cost; Reduce mainline and feeder vessel turnaround times at transshipment and other terminals; Introduce driverless automated guided vehicles and continuous container conveyers; Introduce floating yard barges for containers; Introduce large outreach gantries spanning mainline and feeder vessels. The liner shipping and port industry has been reshaped by logistic integration, containerization, deregulation and globalization. Liner shipping companies and container terminal operators have become the most important part in the international supply chains, which are complex and logistics models. The evolution in supply chains and logistics models urges liner shipping and container terminals to reconsider their function in the logistics process described with the trends in mega ships and mega container cranes in ports. These performances require the new ideas and concepts in container terminal planning in order to keep pace with the development of the mega container vessels. The trend towards globalization of trade, together with the uncoupling of production sites and markets, has significantly increased the demand for containerized marine shipping. Due to the increased volume of container traffic and especially the sizes of container ships, container terminals have become important components of sea and logistics networks. These terminals serve as hubs for the transshipment of containerized goods from ship to ship or from ship to other modes of transportation, e.g., rail and trucks. The face of the container shipping industry continues to change. Timing of introduction of new generations of container ships is very difficult. However, an attempt has been made for the next few years. The trend towards larger containerships also makes it more difficult to choose between hub port and feeder port strategies. This trend is driven by the continued growth in container shipping and increased deployment of mega-ships on major trade routes. The time-sensitive operating practices of such mega-ships mean that they require full loading capacity so that they can efficiently call at major hub ports with minimal dwelling time. By identifying demand growth, and making an assumption as to the market share of further demand could be approved by these ships, a range of possible fleet development can be estimated. REFERENCES [1] Containerization International, 2001, 2002, 2003, 2006, January - December. [2] Containerization International Yearbook, 1985, 2001 and 2002, 2005, London. [3] Cullinane, K., and Khanna, M., 1999, “Economies of scale in large container ships”, Journal of Transport Economics and Policy, Vol. 33, 185-208. [4] Cullinane, K., and Khanna, M., 2000, “Economies of scale in large containerships: optimal size and geographical implications”, Journal of Transport Geography, Vol . 8, 181 - 195. [5] Dragović, B., Radmilović, Z., Maraš, V., 2003, “Containership evolution and mega-container carriers development”, Proceeding of the Seventh International Conference on Fast Sea Transportation, FAST 2003, Naples, Italy, Vol. III - Session D1, 51- 58. [6] Dragović, B., Ryoo, D.K., Park, N. K. and Radmilović, Z, 2007, “Container ship development: A review of state-of-the- art”, Proceeding of Annual Conference – The International Association of Maritime Economists 2007, Athens, Greece, Conference Proceedings on CD, 1-25. [7] Dragović, B. and Ryoo, D.K, 2007, “Container ship and port development: A review of state-of-the-art”, Proceedings of the Ninth International Conference on Fast Sea Transportation, Proceedings of the Ninth International Conference on Fast Sea Transportation, FAST 2007, Shanghai, China, 31-39. [8] Dragović, B., Park, N. K. and Radmilović, Z, 2007, “Anchorage-ship-berth-yard link modeling in container port“, Proceedings of the 11th Conference on Transport Research, WCTR 2007, University of California, Berkeley, CA, USA, Paper No. 558, Conference Proceedings on CD, pp. 1 - 21. [9] Notteboom T., 2004, “Container shipping and ports: an overview”, Review of Network Economics, Vol. 3, 2004, 86-106. [10] ISL Market Analysis World Port Development, 2006, www.isl.org/products_services/publications/pdf/Comm_11-2006-short.pdf [11] BRS Alphaliner Fleet Report, June 2006 and March 2007, http://www1.axsliner.com/WWW/research_files/liner_studies/nofleet/BRS-AlphaForecast.xls [12] Biggest ships in the world, listed by TEU capacity, http://en.wikipedia.org/wiki/Container_ship [13] SSMR June (2006) World seaborne container trade and port traffic, www.isl.org/products_services/publications/pdf/COMM_6-2006-short.pdf [14] Hamburg Harbour, http://www.hafen-hamburg.org/en/ 160 Research in Transport and Logistics 2009 Proceedings Do u University A HOLISTIC FRAMEWORK FOR PERFORMANCE MEASUREMENT IN LOGISTICS MANAGEMENT Yasemin Claire ERENSAL1 Abstract -The purpose of this paper is to introduce and describe an approach to performance measurement of logistics processes. The paper has four main parts. First, an introduction to logistics processes is provided. Then, the key performance indicators of the logistics processes are outlined. Third, a new improvement oriented performance measurement framework of logistics is presented. The originality/value of this paper is that it uses balanced objective matrix methodology (BOMAX) for examining key issues of logistics performance measurement at multiple levels within the organization. In this manner this article presents a comprehensive framework in constructing a strategic performance measurement system of logistic processes. It provides a multi-perspective approach which is more focused on the alignment of performance measurement framework of logistics process with company strategies and addresses the consolidation issue of these multiple viewpoints in a single, consolidated value. Keywords- Performance Measurement, Logistics and Supply Chain Management, 1. INTRODUCTION Nowadays logistics is seen as a value-adding process that directly supports the primary goal of the enterprise, which is to be competitive in terms of a high level of customer service, competitive price and quality, and flexibility in response to market demands. During the last decade logistics has gained much attention in increasing efficiency and flexibility of organizations as logistic costs make up a significant part of total production costs. The entire logistic process, from the acquisition of raw materials to the distribution of end-customer products, makes up a logistic chain consisting of multiple actors. Logistic activities within an enterprise can be divided into; -feed-forward flow of goods, including transportation, material handling and transformation (manufacturing, assembly, packaging, etc.);-feed-back flow of information, including information exchange regarding orders, deliveries, transportation, etc., and ; -management and control, including purchasing, marketing, forecasting, inventory management, planning, sales and after-sales service. Stevens [1] defines a logistic chain as a system whose constituent parts include suppliers of materials, production facilities, distribution services and customers, all linked together via the feed-forward flow of materials and the feed-back flow of information. All these logistic processes are performed by using resources in the form of equipment, manpower, facilities and financial assets. In order to have a value-adding logistics process that directly supports the primary goal of the enterprise, which is competitive, an organization must position its basic strategic elements and core competencies of logistics process to adjust rapidly to critical changes in the environment. That ability implies that the organization has a measurement system in place for reviewing frequently the strategic performance of its logistics process. In this respect strategic performance management and measurement is critical to the success of any organization and needs to reflect the aims and the strategies of an organization that have been developed to achieve those aims. Through performance measurement systems managers can be able to encapsulate and ‘take hold of’ information about strategy, core competencies and future competitive ambitions of its logistic process in a tangible way. This includes monitoring results, comparing to benchmarks and best practices, evaluating the efficacy and efficiency of the logistics process, controlling for variances, and making adjustments as necessary. With sufficient and carefully selected information provided by performance measurement systems, it becomes possible for companies to understand better what is going on and what is about to happen in logistics. It is important to realize that what is not understood cannot be managed. Managers need clear, timely and relevant signals from their internal information systems to understand root causes or problems in logistics process, to initiate correction action, and to support decisions at all levels of the organizations. Therefore a new measurement concept is needed which is consistent and compatible with this process perspective. This study provides new insight into understanding the success and hindering factors of logistics management. The extensive literature 1 Turkey, yerensal@dogus.edu.tr 161 Research in Transport and Logistics 2009 Proceedings Do u University review and case studies provide academics and managers a macro picture of the goals, challenges, and strategies for implementing an effective logistics management and performance measurement. 2. THE PROCESS VIEW IN LOGISTICS AND SUPPLY CHAIN MANAGEMENT Process view is one of the key elements in logistics and supply chain management. According to Christopher [2]: Logistics is a process of strategically managing the procurement, movement and storage of materials, parts and finished inventory (and the related information flows) through the organization and its marketing channels in such a way that current and future profitability are maximized through the cost- effective fulfillment of orders. Andersen [3] has argued that several issues have stressed the logic of the transition from viewing the company as a number of departments to focusing on the business processes being performed: __Every process has a customer, and focusing on the process ensures better focus on the customer. __The value creation with regard to the end product takes place in horizontal processes. __By defining process boundaries and the customers and suppliers of the processes, better communication and well-understood requirements can be achieved. __By managing entire processes that run through many departments rather than managing individual departments, the risk of sub optimization is reduced. __By appointing so-called process owners, who are responsible for the process, the traditional fragmentation of responsibility often seen in a functional organization is avoided. __Managing processes provides a better foundation for controlling time and resources. Many of these elements are based on the fact that every single process has both a supplier and a customer. A main point is that any business process has a customer, either external or internal. Based on this definition, almost all activities within a company can be seen as a business process or part of a business process, including the processes related to logistics. Performance measurement of logistics describes the feedback or information on logistics activities with respect to meeting customer expectations and strategic objectives. Performance measurement systems should answer two simple questions [4] a. Are functions and departments doing the right things? b. Are they doing them well? Performance measures are used to measure and improve the efficiency and the quality of the logistics processes, and identify opportunities for progressive improvements in logistics process performance. Traditional measures, however, are usually ineffective barometers of performance because they do not isolate non-value-added costs. In addition, most measures overlook key non-financial performance indicators [4].Performance measures are classified in several ways in the literature. When describing and measuring the performance level in a business process, a number of parameters might be used. It is pivotal to employ a balanced set of measures in order to understand the performance of the process and be able to identify improvement areas. Typical dimensions for describing and measuring performance are [5]: __Qualitative and quantitative measures. __‘Hard’ versus ‘soft’ measures. __Financial versus non-financial measures. __Result versus process measures. __Measures defined by their purpose (result, diagnostic, and competence). __Efficiency, effectiveness, and changeability. __The six classic measures (cost, time, quality, flexibility, environment, and ethics). All areas should be considered when developing performance measures. It should be emphasized that these dimensions overlap. In order to diagnose the ‘health status’ of an organization one should ideally employ a balanced combination of measures. Fagerhaug [5] has developed a criteria sheet. The sheet provides the name and a short description of the process/structure, as well as a number of text-based and number-based measures belonging to each of the five categories mentioned above. Based on this criteria sheet the author of this paper would argue that a number of measures could be used to enhance the performance of the logistics processes (Table 1). As listed in Table 1, several concrete measures are introduced to show how logistics process performance can be measured in practice. It should be emphasized that the measures are examples rather than a final set. Different measures are needed in different levels of organization. There should be information available for strategic management purposes at the company level. On the other hand, information is needed also for 162 Research in Transport and Logistics 2009 Proceedings Do u University operational management at the work shop level. Measures can be used mainly on three levels. Firstly, logistics companies can analyze general environment and their own performance at company level. These measures are global in nature, covering a wide scope of activities. Global measures provide top management with a sense of whether strategic objectives are being achieved. They are monitored month-to month or quarter-to-quarter. In a sense they keep management in touch with the outside world. Secondly, they can measure individual projects' performance. Finally, they can focus on processes and departments. These types of measures are more specific to the internal workflow. They represent day-to-day measures of operating effectiveness and efficiency [4]. It is better for a small company to pick only a couple of measures and start with them. This means that they can find out the current performance and improvement potential of the firm. 3. THE INTEGRATION OF OBJECTIVE MATRIX (OMAX): AN APPROACH IN DETERMINING THE PARTIAL AND OVERALL PERFORMANCE INDEXES OF LOGISTICS PROCESS - THE BALANCED OMAX (BOMAX) Productivity in a narrow sense has been measured for several years. In 1978 an enlarged method, the POSPAK method, was introduced. This method indicates specific measures in order to improve the overall productivity of an enterprise [6]. One of the first approaches to performance measurement was published by Sink and Tuttle [7]. The model claimed that the performance of an organizational system is a complex interrelationship between seven criteria. In 1993 Hronec [8] published the book ‘Vital Signs’, where he described how to use quality, time, and cost performance measurements to chart the company’s future. In 1995 Rolstadås [9] edited the book ‘Performance Management’. It sought to provide the reader with a detailed overview of the modern enterprise by focusing on performance evaluation and measurement and performance improvement techniques. Since 1995 a number of books and papers on performance measurement and management have been published. An example of one of these published materials is the so called BOMAX performance measurement system, which was developed by the researcher self [10]. Through performance measurement, the various performance level of the business should be monitored. Based on this business model, BOMAX has suggested three levels of hierarchy for defining performance indicators. Each performance indicator is a function of two or more performance measures. The three levels of hierarchy for defining performance indicators are: ‘Enterprise Level’, ‘Process Level’ and ‘Functional Level’. BOMAX emphasize that the self-assessment process allows the organization to discern clearly its strengths and areas in which improvements can be made and culminates in planned improvement actions which are then monitored for progress. Based on the self-assessment, improvement planning should be performed and initiated. As shown in the figure, performance measurement provides input for the improvement planning, choice of improvement tools, as well as for the self-assessment process. As it was mentioned the chapter before, there is needed a method of indexing performance measures, and calculating an overall, multi-factor, performance index. Several techniques are available for this purpose. The thesis of this article is that the BSC [11] and OMAX [12]. Philosophies complement each other quite effectively. BOMAX is one of the few techniques capable of integrating the whole gamut of strategic measures into a single coherent summary as a consolidated value. Had OMAX alone been used, a manager would have trouble making tradeoffs among alternative strategic objectives. In contrast, had the BSC alone been used, the connection between financial and non- financial criteria would have been less robust. By integrating both OMAX and the BSC, the organizations could be able to create synergies which overcame the weaknesses of the individual methodologies. By combining the strengths of the two, we end up with a stronger, more robust framework with increased predictive power, the so-called ‘Balanced Objective Matrix (BOMAX).’ Through BOMAX method performance measures are normalized and an overall, multi-factor performance index is calculated. An index is a composite number that is created by mathematically combining several individual measures. While concentrating just on only one strategic performance index simplifying the decision making of managers and avoiding the confusion caused by dealing with many performance indicators at once which usually yields only a vague general perception. The single number resulting from the BOMAX will tell management if the organization’s strategic performance qualifies as excellent, unsatisfactory or just mediocre. This single index is the indicator of how well the organization is doing against the preset target or evaluated along with other measures. The four main components of BOMAX are: the performance measures of each of the logistics processes (the scaled KPI) (Mi), the weights (wi), the performance scale (Li), and the performance index (PIi). 163 Logistics BSC Perspectives –Key Performance Indicators (KPI) Processes Learning-Growth Internal Process Customer Satisfaction Financial ML1, ……….. MLn MP1, ……….. MPn MC1, ……….. MCn MF1……….. MFn Warehousing ML1 Commitment /Staff MP1 Packaged pcs/year MC1 Failurlessness (failure rate) MF1 Pers. Costs/ commitment receiving, put- away, MP2 Utilization of loading surface (%) MC2 Politeness MF2 Pers.costs/Volume, unit, value replenishment, cycle-counting, MP3 Utilization of loading space (%) MC3 Reliability MF3 Construction and op.costs/Yearpicking, packing, shipping, kitting, ML2 Man-facilities relation (ergonomy, MP4 Vol.loaded in/out/Staff MC4 Advising MF4 Managing costs/commitment returns environment ,protection)ML3 Optimum combination of jobs (tasks) MP5 Packaged pcs/staff MC5 Quick response on MF5 Total man.costs/total warehouse cost and facilities complaints MP6 Handled volume/Staff MC6 Damage events/total MF6 Stor.,pack.,handl., costs/volume, unit, activities value MP7 Handled volume/year MP8 Physical performance time/Processing time Inventory ML4 Optimum combination of jobs (tasks) MP9 Av.inventory (volume, value) MC7 Restorability (av.restoring MF7 Pers. Costs/commitment turns, safety- and facilities time, total break-down stock, cycle-stock, time) transit-stock, ML5 Commitment MP10 Av.storage time (day, hour, MC8 Politeness MF8 Pers.costs/ strategic-stock, Staff minute) Volume, unit, value replenishment frequency, MP11 MC9 Reliability MF9 Construction and op.costs /Year consignment- minute) stock MP12 Rotation frequeny /year MC10 Advising MF10 Interest/ Volume, unit, value MP13 Commitments/year MC11 Quick response on MF11 Total stor.&inv.costs/total log.costs complaints MP14 Average time of comm.(hour, minute) Transportation ML5 Commitment /staff MP15 Tr.ed volume/ MC12 Restorability (av.restoring MF12 Total tr. Costs/total log.costs rating, routing, Staff time, total break-down tracing and time) tracking, sourcing, MP16 No of transports/ MC13 Politeness MF13 Tr.costs/volume, tkm, km auditing and year payment, MP17 Ton km/Vehicle and year MC14 Reliability MF14 Tr.cost/time&vehicle management MP18 Av.Vol.Transport/ MC15 Trucking&tracing MF15 Tr.costs/commitment Av.time. Transport MP19 Average distance (km) MC16 Quick response on complaints MP20 Utilization of vehicles(%) related MC17 Damage events/total to time, load capacity and tkm activities MP21 Error delivery/total delivery commitments Table 1: Key Performance Measures of Logistics Process in respect to BSC Research in Transport and Logistics 2009 Proceedings Do u University The top row of the matrix includes the KPI which are assessed by BSC approach and defined for each BSC perspectives. The perspectives and their abbreviations are Learning and Growth (L); Finance (F); Customer-Citizen (C) and Internal Process (I). In BOMAX methodology dissimilar measures can be compared, and combined to produce an overall-global performance index. In order to be able to compare dissimilar measures, to see the inter-relationships between them, it is possible to index the measures scores, and so convert them to the same scale. The middle section holds eleven rows with different outcomes for the specific performance measures. These rows are ranked or scaled from zero to ten. The index scale is created by establishing a target value for each performance measure based on their current performance. The bottom of the target range is defined as the minimum level that can be permitted. Matching the levels of performances with the level of the rows in a way that an outcome of ten will be the most desirable and zero the least desirable and typical outcomes of each indicator are aligned with a score of three. The initial baseline called the ‘as is’ performance level for each performance measure is determined and assigned to level 3. The scaling should be conducted in the way that grade 10 could be achieved with excellent performance at least in five years time horizon. That means the time horizon should be defined in BOMAX very carefully and can typically be about a year or less for short-term goals or spans several years for long-term goals. The possible outcomes-the intermediate values- of the performance measures are found in the body of the matrix and are calculated for scores between these ranges (Li). The objective of these arrangements prevents the awarding of high grades to mediocre or normal performance, and embraces the notion of stretching the work force to a superior performance and betterment of the system. Ranks in the matrix’s body should be clearly marked so they do not provide a margin of doubt when assigning a score. The bottom of the matrix weights the categories for the aggregation process. For each BSC perspective measures on BOMAX weighted according to their importance that is felt would create the biggest challenge in terms of achieving the strategic targets. These weights are determined by management and add up to one hundred. The score in each column is multiplied by the weight, producing a final number or ‘index’ that represents the ‘grade’ of the strategic activities for that specific period (PIi). Performance indexes are not an exact measure of achievement but rather provide an indication of business performance. To be useful, performance indexes must exhibit certain characteristics: appropriateness, relevance, accuracy, timeliness, completeness and comprehensiveness. When indexes begin to move outside the threshold limits, the performance measurement system can alert management, who then attempt to diagnose the problem and address its causes. The development of performance indexes is not an end in itself but rather one part of a structure of governance and accountability. They can indicate whether strategic planning has been undertaken and is well focused on the reason for the organization existing. This method of monitoring the strategic measures enables to identify the current levels of strategic performance, and where action is needed to improve them. It will provide feedback of the effects of the actions, and ensure that the strategic performance continually improves. The BOMAX can be interpreted separately in three different ways: each performance measure as an unit, the performance sub- index as an index of a BSC perspective and lastly as a sum up measure off all BSC categories pulled together in one final index as departmental and company level (Figure 1). BOMAX provide insight into different departments or levels of analysis. Most help desks have various sub sections like front desk and solution providers, etc., which all contributing to overall success of a help desk. The proposed framework, allow managers to do just that. To gauge exactly how well a section in a help desk is performing the overall performance index will allow the managers to capture and report specific data points from each section within the organization providing a ‘snapshot’ of performance. Performance Index of each Logistics Process (PPI) of each in respect of BSC perspectives at period t (Eq.1); LGi wGi LPi wPi LCi wCi LFi wFi PPI i 1,...n i 1,...n i 1,...n i 1,...n ………(1) (t ) PPI (t 1) i=1,………n number of measures, j=1,………m number of logistics processes Mi =Performance measures Li = Performance scales of Mi xijj=The current values of measures(Mi) wi= weights of measures wi 100% i 1,...n 165 Research in Transport and Logistics 2009 Proceedings Do u University CRITICAL SUCCESS FACTORS (CSF1,………CSF1N) Learning&Growth Internal Process Customer Finance Performance Measure ML1 ….. MLn MP1 ….. MPn MC1 ….. MCn MF1 ….. MFn Actual value XL1 …. XLn XP1 ….. XPn XC1 ….. XCn XI1 …… XIn 10 L10 9 L9 8 L8 7 L7 6 L6 5 L5 4 L4 3 L3 2 L2 1 L1 0 L0 Level LL1 … LLn LP1 … LPn LC1 … LCn LI1 … LIn Weight wi =100% wL1 …. wLn wP1 …. wPn wC1 …. wCn wI1 …. wIn Performance Value LL1.wL1 …… LL1.wLn LP1.wP1 …… LPn.wPn LC1.wC1 …… LCn.wCn LI1.wI1 …… LIn.wIn FIGURE 1: THE DEPARTMENTAL BOMAX OF A HELP DESK 4. CONCLUSION This paper has sought to give an introduction to a new approach for measuring and improving performance of logistics processes. In order to describe the approach, an introduction has been given to performance measurement. A performance improvement framework has also been introduced. This method can be employed at certain intervals, for instance annually may also be used for short periods of time. The developed performance measurement systems are both focused on the results and the processes of logistics. We would argue that the use of the approach would enhance the performance of logistics processes. References (1) Stevens (1989). G.C. Integrating the supply chain. International Journal of Physical Distribution and Materials Management 19 8 (1989), pp. 3–8. (2) Christopher, (1992) Logistics and Supply Chain Management-Strategies for Reducing Costs and Improving Services, Pitman, London, 1992. (3) Andersen, B., (1999), Business Process Improvement Handbook (ASQ Quality Press, Milwaukee, USA). (4) Lynch C.(1991) Measure Up!- Yardsticks for Continuous Improvement, Blackwell Business, Oxford. (5) Fagerhaug, T., (1998), Practical Productivity Measurement, Proceedings from 10th Working Seminar on Production Economics,Innsbruck/Igls,Austria. (6) Sjøborg, E.R., (1984), Totalproduktivitet etter POSPAK-metoden (Bedriftsøkonomisk Forlag, Oslo, Norway). (7) Sink, D.S. and Tuttle, T.C., (1989), Planning and Measurement in Your Organization of the Future (Industrial Engineering and Management Press, Norcross, USA). SJØBORG, E.R., 1984, Totalproduktivitet etter (8) Hronec, S.M., (1993), Vital Sign. Using Quality, Time, and Performance Measurements to Chart Your Company’s Future (AMACOM/American Management Association, New York, USA). (9) Rolstadås, A. (1995), Performance Management: A Business Process Benchmarking Approach (Chapman & Hall, London, United Kingdom). (10) Erensal, Y.C. (2007) ‘Designing a Corporate Strategic Performance Measurement System: A Conceptual Framework with a Holistic Approach’ ed., Erkollar, A., Enterprise & Business Management : A Handbook for Educators, Consulters and Practitioners 1. Aufl.. - Marburg : Tectum Press, Marburg, 94-116 (11) Kaplan, R.S., Norton, D.P., (2001). The Strategy-Focused Organization: How Balanced Scorecard Companies Thrive in the New Business Environment. Harvard Business School Press, Boston, MA. 32 (4), pp.38–55. (12) Riggs, J. and Felix. G., (1983). Productivity by Objectives. Englewood Cliffs, NJ: Prentice-Hall. 166 Research in Transport and Logistics 2009 Proceedings Do u University HEURISTICS FOR A GENERALIZATION OF TSP IN THE CONTEXT OF PCB ASSEMBLY Ali Fuat ALKAYA1 and Ekrem DUMAN2 Abstract Traveling Salesman Problem (TSP) is one of the most well-known NP-Hard combinatorial optimization problems. Adding new constraints to the problem yields different generalizations to the problem and each new generalization forms the basis of a new research area. In this study, we propose new techniques for a generalization of the TSP. In this problem, the cost of traveling between two cities does not only depend on the distance between these cities, but also on the visiting sequence. We analyzed the problem under different conditions; the first and last points (nodes) are set fixed or they are free and for solving the problem we propose several heuristics. After analyzing constructive heuristics, improvement heuristics are applied. As improvement heuristics, we implemented pair-wise exchange procedure (PEP) and record-to-record travel with local exchange moves (RTRLEM). Comparison of these approaches together with their parameter fine tuning are given. Keywords Heuristic, printed circuit board, sequence, traveling salesman problem INTRODUCTION Traveling Salesman Problem (TSP) is one of the most well-known NP-Hard combinatorial optimization problems. Adding new constraints to the problem yields different generalizations to the problem and each new generalization forms the basis of a new research area. Mostly known generalizations are Asymmetric TSP, Vehicle Routing Problem (VRP) and its variants. TSP is observed in many research areas and among these we can easily pronounce logistics and transportation. After a small survey on TSP, one can easily argue that there is uncountable number of studies towards solving it since it is introduced to the literature, ranging back to at least the late 1940’s. The TSP is the focus of interest for many research disciplines (mostly computer scientists and mathematicians) because, even after about half a century of research, the problem has not been completely solved. This is because any programmable efforts to solve such problems would grow super-polynomially with the problem size. These categories of problems became known as NP-hard [1]. We observe the spirit of the traveling salesman problem within many practical applications in real life. For example, a mail delivery person tries to figure out the most optimal route that will cover all of his/her daily stops, a network architect tries to design the most efficient ring topology that will connect hundreds of computers, a manufacturing engineer tries to design the shortest sequence for assembling components on a printed circuit board (PCB). In all of these instances, the cost or distance between each location, whether it be a city, building or node in a network, is known (we use the terms “vertex”, “node” and “point” interchangeably in this study). With this information, the basic goal is to find the optimal tour. In this study, we analyze a generalization of TSP in which, the cost of traveling between two cities does not only depend on the distance between these cities, but also on the visiting sequence. This problem is observed in PCB assembly machine optimization. We firstly define the problem, and then propose some heuristics for solving the problem. The performance analyses of the proposed heuristics are also done within the study. In the next section, classical TSP definition and formulation is given together with a literature survey of proposed techniques and heuristics towards optimizing it. Also definition of the generalization of the TSP is given. The constraints to the problem are given with its definition. In the following section, proposed solution heuristics are defined in detail. The proposed heuristics are compared on randomly generated PCB data and the results are given in Results and Discussion section which is followed by a conclusion. 1 Ali Fuat Alkaya, Marmara University, Faculty of Engineering, Computer Science and Engineering Department, Göztepe, Istanbul, Turkey, falkaya@eng.marmara.edu.tr 2 Turkey, eduman@dogus.edu.tr 167 Research in Transport and Logistics 2009 Proceedings Do u University PROBLEM DEFINITION, CONSTRAINTS AND LITERATURE SURVEY The TSP is stated as, given a complete graph, G=(V,E), where V is a set of vertices, E is a set of edges, and a cost, cij, associated with each edge in E. The value cij is the cost for travelling from vertex i V to vertex j V. Cardinality of V is denoted by n, i.e. |V|=n. Given this information, a solution to the TSP must return the cheapest Hamiltonian cycle of G. A Hamiltonian cycle is a cycle in which each vertex in a graph is visited exactly once. In order to obtain a solution to TSP, several solution methodologies are developed. Exact algorithms are guaranteed to find an optimal solution and to prove its optimality for every instance of a class of combinatorial optimization problems. The run-time, however, often increases dramatically with the instance size, and often only small or moderately sized instances can be practically solved to provable optimality. We can categorize the exact algorithms as branch-and-bound, cutting planes and branch-and-cut [2]-[4]. Another approach for solving TSP-like problems is developing heuristic methods. In heuristic methods we sacrifice the guarantee of finding optimal solutions for the sake of getting good solutions in a significantly reduced amount of time. Among the basic heuristic methods we usually distinguish between constructive methods and local search methods. Constructive algorithms generate solutions from scratch by adding—to an initially empty partial solution—components, until a solution is complete. They are typically the fastest heuristic methods, yet they often return solutions of inferior quality when compared to local search algorithms. Local search algorithms start from some initial solution and iteratively try to replace the current solution by a better solution in an appropriately defined neighborhood of the current solution. Among the constructive methods developed for TSP we can list Convex-Hull and Nearest Neighbor algorithms [5]-[7]. Several local search methods are also developed for TSP. Croes developed first 2-opt algorithm, not after a decade Lin generalized the 2-opt concept to r-opt [8]-[9]. Or developed Or-Opt algorithm where he relocates chains of length three, two and one [6]. Recently, Babin et al. showed that improved Or-Opt + 2-opt is an excellent combination and is easy to implement [10]. All of these local search methods put promising results for reaching optimality in a reasonable time. For a complete survey and summary about TSP, we direct the reader to [7]. A generalization of the TSP In this study, we deal with problem which is a generalization of the TSP. In classical TSP, the cost between two points is fixed and known a priori. Thus the cost function, C1(x,y), implies that the cost increase linearly with distance, d(x,y) (3). The distance may be either Euclidean or Chebyshev metric. In Euclidean metric, the distance is directly calculated by the formula d (x, y) x y 2 x y 21 1 2 2 (1) whereas in Chebyshev metric the distance is calculated by using d (x, y) max x1 y1 , x2 y2 . (2) Chebyshev metric is usually encountered in placement machines. In this study, we calculate cost in time units, so we define C1(x,y) mathematically as follows: C1(x, y) d (x, y) (3) v where v is a predefined constant. However, in our problem the cost of travel between cities depends also on the visiting sequence and some other predefined parameters. We face this generalization of TSP while optimizing the operations of a component placement machine in PCB manufacturing industry [11]. Nevertheless, as we pointed out before, TSP and its variants can be observed in a number of cases in real life, and we believe that optimization techniques developed for this type of problem can be applied in other research areas. In this generalization of TSP, the cost function is stated as C2(ti , x, y) max ti ,C1(x, y) (4) Observe that this function is dependent on a parameter ti and it is known at compile time. What makes the problem complicated is that there are a number of ti values and they are used in different positions of the placement sequence. In our case, there are four different ti values and t1?@#A1 B #C $?D ? 7  =  EF G  B  - #  7  H    < "   A  =  E F  G   B   -  # 2   ?  =  FA   H    + , =  I  D I7J  F  G   (      =  -  #     H 7  173 Research in Transport and Logistics 2009 Proceedings Do u University          &!=                 - .      #              -                 #                  ?              0     1   -                                            #  *        -               / 1             ;       2               $1    $8<%    <1   $%%K  #       3,5  $ <       G $ <     / 1                             21-#                    1                    1           -         )                              0              1         .           #        /             ) 0  1              #                          04"51 0 ,   /      , >,  7    , %  > ,  ;  ,   3 3  /    .      9,   /   , 174 Research in Transport and Logistics 2009 Proceedings Do u University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esearch in Transport and Logistics 2009 Proceedings Do u University -        *    0 1       /                                 #              7      *  G    0$1   ,              0<1081                   0&1             091                  0A     1    0:1    0      7     1               ,           0'1       )                     0 %1                               0 1                    0 $1                     0 <1                      0 +1                       0 810 &1            0 91        #                 #  =     4"5                O-= '<               P      6            34"5#6  #               - 6     /             -      7                          6              #        -         6        - G           G      0  1     0 ) 1     0  ) 1       0 1   G-OF"(  5   -   -                                                          )    6  <     '          $#      :   176 Research in Transport and Logistics 2009 Proceedings Do u University     G 6      7 Q %% %R       P                              -                                           #      0 1        -        $8 )0<'R1      7 Q $%L Q $% %R   Q %% 8%R      Q$% 6  G$  $        1    4"5   1       6  <                          $      -              Q       Q %:%     Q+8%        0     Q $ Q8   <   <%   1    G-OF"($G-OF"($? =   $S5T <  L          #                   7                      -               Q &9%9        Q<$% 6                  7                         #      7  R    0        1A     7                    $                  K  7     # $  #?(  3            ! %&'&  "#$  ! P   : $8 P  $ 8 3    Q  Q  G    Q $ Q $ #   .  8&&  +9&: P         : 3           Q8 #  3  Q:%&  Q$+9&:0U $V$K8&& V 1 0U $V8K V +9&:K8V:1 5      :% $8% 5     $8% #    :% 8%% P  Q''<' Q$8$<$  177 Research in Transport and Logistics 2009 Proceedings Do u University #?($ P         K  7    (')*'"+)!,'- . /. 0. 1. 2. 32 /.. %R 09$%91 0+9%91 0$$%91 $'< $9'< '%+< 8$'< $8R 0<+891 $9'< '%+< 8$'< $ 8+< <9 &: 8$9'< 8%R $9'< 8$'< $99'< +%$'< 8$9'< :+%+< 8$'< 98R '%+< $99'< +&8+< &8$'< :+%+< <%' : 999'<    L                                                  7 (                       .               (                                  -                                   )            #                            0      1 4   ! E I$%%9S5  C WE ?XXT3  B  EE $!   , E A $%%: S                  Y    .   /#  T(  Z   L  "  :&88+898 x1) X1 remanufacturing capacity of the manufacturer X2 refurbishing capacity of the manufacturer Ki sensitivity parameters ( > 0) ( i = 1, 2, 3) d cost of disposal of a refurbished product R expected revenue from remanufactured and refurbished products In this model, the probability of selling of a remanufactured product is given by (1-(p1/P1))(1- (x1/K1X1)) where P1 is the maximum price that can be charged at which the probability of selling becomes zero. Hence, demand or the expected number of units of remanufactured product sold is given by (1-(p1/P1))(1-(x1/K1X1))x1. It is seen from the expression that demand is a linearly decreasing function of price for a given number of available remanufactured units. Also, it can be shown that demand is a concave function of availability for a given price and the maximum of the function occurs at K1X1/2. If we restrict K1 2, we can ensure that within the range of x1 demand is an increasing function of availability with decreasing returns to scale, which means demand grows more rapidly with availability in the initial phase of introduction of remanufactured products. But the growth rate tapers off as more and more remanufactured products become available in the market [13]. To model the probability of selling of refurbished products, the same logic can be applied. However, there is an issue that needs to be addressed. According to the problem definition, an unsold remanufactured unit can be disposed of at the price of a refurbished unit, and hence this would impact the probability of selling of refurbished products. From the expression for demand for remanufactured products, it is seen that the number of unsold remanufactured products increases linearly with price, but exponentially with availability. Hence, the impact of availability of remanufactured products would be much more effective than the impact of their price on the probability of selling of refurbished products. In fact, it is assumed that the increase in the number of unsold remanufactured units due to increase in its price is absorbed by the manufacturer by bundling the units with separate service provisions. The effective prices of remanufactured units thus sold approximately equal those of refurbished units, and the attractive offers are promptly lapped up by customers Since the prospective customers of refurbished units are price-sensitive and they look for cost-effective utilization of these products. It can be assumed that the increase in price of remanufactured units would virtually have no impact on the probability of selling of refurbished units. On the other hand, the increase in availability of remanufactured units would definitely have an impact. Accordingly, the probability of selling of refurbished products is defined as (1-(p2/P2))(1-(x2/K2X2))(1-(x1/K3X1)) where P2 (?  /  ?  @    : )   ; .? > ;A  =  %    +    .   )    B?<  /   ?     ? @A    ( :   C  ?       =  %     4     :      4     +  4 >?  /  ?   ?  @   - : D      = %   4    :    4    + 4 >?  /   ?   @   219 Research in Transport and Logistics 2009 Proceedings Do u University 7                       (  7               $      ((        -     #  -   !"#$   ?              ?                           =  #                  (!     9        !                *       5      ,  & 3    (      E                                &         ?          (  ( 9 5 (                            #                             #           #        01"#        3F.                    &A         #      #  A              0 1)    ,                       G                            # A      H   4A                   I                     /             0(1 J ?   (       #   #      K 0 1               L01  2                  H2  A      &   #                      3        %                                             ##  ?#            ?      # ?          #       ? /           #                                 #   3      %        4       3    3      ?#          #               #                                 3      /       KM              3  M      #3  K             .  # 2   A          A          % &'" ( () • 6     K              (  • N     K7          •       • 6    • 4 & # ?    • 4 &        • G #  K4    #                &     (  • 4#    K4#          • 4 &#    K4 &#          220 Research in Transport and Logistics 2009 Proceedings Do u University • 4  K4          # #       5 • / #    K6            #         • E #    K6            #         • "              • 6   K.    &     ' 3   % &'" ( !") •   &K  &      (  • )  &K)  &#           (( •      K     H  • .     K.           (( • :      K:             (( • O K      # #  ! • 7      K:      H  * • 6       K6                      • 6         K6                       • C          K    ?         ! #  •   K:        •  # ?    K     # ? •       K          • O        K4               %% &'"&#)( #) 7                                     /    # # /CC 3                        6                    &                                   ( .                        -      #                     #        &#   $ . H                       #   5     # #         #      9                #      #                   ,  # ?         & # ?  *            &    !J             #                   #                                                               221 Research in Transport and Logistics 2009 Proceedings Do u University ( # ?             # ?            ? - 7                                         #              #           %*(+ " !,+-)            #                             #             #         %     #                  #     #       #          #   #  :       #             ?             *  2     A            01 A       E2  K4A           01                %.))/0&#)    :              3                      J                    #    #     (              #                 #                  # 0"  " !!$1 - /     ? #     $ /         # ?       *     1 / #             ?    P                               2    ?                       )                /          3         #          #   #    /                          /#      2           2  0A      1     2  0A    1 %                     1 +.!- 7 9 ( *(" % 1/) + 3& ,) :(+'+  :)  '  " /+(  - & &'+ &+( - - " +/ % & (& (% %(+)-% ;(+0 3& ,) (& (% /) !  222 Research in Transport and Logistics 2009 Proceedings Do u University   . 3 ?..F  E +34. !!(Q7               7   R7    ($)  )(,53(*  . 7 .  ?7***Q+ . 8    R+ "  "  7  N -$3$ (. GI***Q.     # ?    RS   T 4    +  N$9,3,, -C   F 4  "6 I   F7 8?   ): !!, Q.               K   F ?  R "  U E  O   N ($  5$3 ,! $C ? +I"#7/"  ? 7 C  FN !!Q.4   .   6  O  7  . /   N$0(1**3 - 5C ? +I+ 7   ) !! QO         R+  S    E  O  N-! 553 ,! 9C ?  +I    ? 4 !!$ Q)    "      R 7.. !!- 86") (55 )  3 N  C G   93(( ,"  47 7 4N !!-Q."         # ?         '      R+  S   E  O  N$(5$39* *" )" "P" !!$Q%      3  #   R/    ) #  N-9$9$3$,$ ! " )"  %  G8 *** QF    .             R (  /    "  I#  3C  /  /      -* 3-*$  " )"I " +"G **, Q" # )     8 O    G IK %  4   /   R" E O  N $01,,93,*-   "#7I  F G 8 !! Q./    G    F   . .     )  7  R7    + " N ,$3*! ( :  %:     +G   J.N ) !!9QOO  . :  O  3"  7 2 R: )   N(,01$3(( - + 3J   7 ? 4***QC 7  )  . %  .  " #)   7  RS   G   N$-*3-(5 $ % A V :?  P !!( Q/   )     7     N        G  F    . R+  S   E  O  N59( !W(($ 5 %   F  4  4 !!9 QO     #    R +   S     E  O  N,!$*3$( 9 F  4" "   8 .   / !!5 Q  " #      K . &      #     R.  E O  N--3(  , G " )  ) !!9Q.           '   R" U E O  N(-**3*(( * G :I 8FO 48 N  4S"4 !!$QE O      #     R)  6   N$*0-1-593-*9  ! / 4P )4F !!QG            3     #    K     R /+++             3  K .    #N(01,93*$   I   I ** Q.  #          ?#  3       #    R7    O   "  $*355   8 # O !!9 Q.        3            =       R EO )  )  C  223 Research in Transport and Logistics 2009 Proceedings Do u University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esearch in Transport and Logistics 2009 Proceedings Do u University AN APPROACH OF INTEGRATED LOGISTICS HMMS MODEL UNDER ENVIRONMENT CONSTRAINTS AND AN APPLICATION OF TIME SCALE Fahriye Uysal1, Ömür Tosun2, Orhan Kuruüzüm3 Abstract Concern for the environment has led many firms to define policies that protect the environment within which they operate. This concern is reflected in all the activities of the product life cycle, both in those of direct logistics as well as reverse logistics. The aim of the paper is to investigate how environment policies, in the form of emission charges or emission limits, affect the logistics decisions of a firm. Time Scale calculus is used to analyze the integrated logistics model of HMMS under environment constraints. Keywords Integrated Logistics, Reverse Logistics, Time Scales INTRODUCTION With the increased environmental concerns and stringent environmental laws, integrated logistics has received growing attention throughout this decade. Integrated logistics includes the planning, implementation, and control of the flow and storage of raw materials, in-process inventory, finished goods, services, related information, and payments among suppliers and consumers from the production of raw materials to the final recycling or disposal of finished goods [1]. The integrated logistics process features are inbound logistics, manufacturing logistics, outbound logistics. In this study the integrated logistics is considered in two stages, forward and reverse logistics. Whereas forward logistics is generally the movement of product from one origin to many destinations, the reverse movement of a product is the opposite, from many origins to one destination. Reverse logistics can be defined as the process of moving goods from their typical final destination for the purpose of capturing value, or proper disposal [2]. Moving goods from their point of origin towards their final destination has been the focus of logistics systems. A reverse logistics system incorporates a supply chain that has been redesigned to manage the flow of products or parts destined for remanufacturing, repairing, or disposal and to effectively use the resources [3]. The reasons behind promoting integrated logistics practices are of both economic and environmental kind. Among the economic motives we find the recovery of the value still incorporated in the used product and the important savings in material and components. From the environmental viewpoint, we might cite concern regarding solid waste pollution [4], landfill saturation [5] or the scarcity of raw materials [4], [6]. THE MODEL The integrated logistics model is considered in two stages. The first stage is forward logistics, whereas the second one is the reverse logistics (Figure 1). 1 Akdeniz University, Faculty of Economics and Administrative Sciences, Department of Administration Turkey, +90-242-3101837, fahriyeuysal@akdeniz.edu.tr 2 Akdeniz University, Faculty of Economics and Administrative Sciences, Department of Administration Turkey, +90-242-3106418, omurtosun@akdeniz.edu.tr 3 Akdeniz University, Faculty of Economics and Administrative Sciences, Department of Administration Turkey, +90-242-3101931, orhank@akdeniz.edu.tr 225 Research in Transport and Logistics 2009 Proceedings Do u University p(P(t)) Manufacturing I1(t) P1(t) Store S(t) 1 Market P2(t) P2(t) I2(t) Remanufacturing Store 2 Disposal Forward Logistics Reverse Logistics FIGURE 1 Integrated Logistics Material Flow of the Model HMMS model first appeared in the book by Holt, Modigliani, Muth and Simon [7 ] in the discrete version. In[8], the author studied the continuous version of the HMMS model by taking nonnegative discount rate ( ) into account. Time scale calculus is a new and exciting mathematical theory first introduced by Stefan Hilger in his Ph D. thesis [9], which unites two existing approaches to dynamic modeling – difference and differential equations – into a general framework called dynamic models on time scales. Since time scale calculus can be used to model dynamic processes whose time domains are more complex than the set of integers or real numbers, dynamic modeling in economy will provide a flexible and capable modeling technique for economists [10]. Where the quantities are defined as Cp = regular time cost per unit Pi t / P(t) = production rate at time t (state variable) i = 1,2 P(t) = production rate goal level at time t Cl = cost of holding a unit for one period of time Ii t / I(t) = inventory level at time t (control variable) i = 1,2 Î (t) = inventory size goal level at time t St / S(t) = demand rate at time t; positive and continuously - differentiable T = length of planning period = non-negative discount rate, a constant a = inventory holding cost coefficient, positive constant b = production cost coefficient, positive constant I0 = initial inventory level = linear charge per unit pollution p(P(t)) = pollution: a convex, monotonically increasing, continuously differentiable, and p’ is continuously - differentiable In stage 1, pollution costs are added into the minimization of the production and inventory costs of the forward logistics. T The Discrete Model: C = [C (P ˆ 2 ˆ 2p 1t P) CI (I1t I ) pP t ] (1) t 1 Subject to the constraint I1t I1t 1 P1t S1t , where t = 1,2,…,T T The Continuous Model: C= e t [ (I (t) Iˆ1 (t)) 2 b (P (t) Pˆ (t))2 pP t ]dt 0 2 2 1 226 Research in Transport and Logistics 2009 Proceedings Do u University Subject to the constraint I (t) P1(t) S1(t) , and I(0) = I0 (2) [8] (T ) The Time Scale Model: C eˆ ( (s),0) a (I Iˆ)2 b (P Pˆ )21 1 p P s s (3) 0 2 2 The cost given above is optimized with subject to I (t) P1 (t) S1 (t) and with the initial values of I(0) = I0 and P(0) = P0. At this point, two cases will be considered: Case 1. Let p P t P (linear). If the pollution function p(P(t)) is linear, then the conditions of optimality will be simple. Let us assume that p(P(t)) = P(t) . The first derivative of the pollution is equal to one, and second derivative is zero. a The coefficient matrix A t 1 v t b 1 v t is a constant matrix if v(t) is a constant, or if = 1 0 0. Otherwise A is a non-constant matrix function. For this system, we have the following statement: I – b v(t)A(t) is invertible if and only if v t for all t T. a ˆ ˆ P1 (t) P aI bP ˆ eˆ (t, t ) 0 t ˆ PA 0 eA (t, ( )) b(1 (t)) (4) I1 (t) I t0 0 S(t) Case 2. Let p P t 0,5P 2 (quadratic). If the pollution function p(P(t)) is quadratic, then the conditions of optimality will be simple. Let us assume that p(P(t)) = 0.5 P2(t) . The first derivative of the pollution is equal to P(t), and second derivative is zero. b a The coefficient matrix A t b 1 v t b 1 v t is a constant matrix if v(t) is a 1 0 constant, or if = 0. Otherwise A is a non-constant matrix function. For this system, we have the b following statement: I – v(t)A(t) is invertible if and only if v t for all t T. a P (t) P aI ˆ bPˆ b t Pˆ1 eˆ 0A (t, t0 ) eˆA (t, ( )) b(1 (t)) b (5) I1 (t) I t0 0 S(t) In stage 2, HMMS model is applied to the reverse logistics. T The Discrete Model : C = [C p (P2t Pˆ ) 2 CI (I ˆ2t I ) 2 ] (6) [7] t 1 Subject to the constraint I 2t I t 1 P2t S2t , where t = 1,2,…,T T b The Continuous Model: C= e t [ (I (t) Iˆ(t))2 (P (t) Pˆ (t))22 2 ]dt (7) [8] 0 2 2 227 Research in Transport and Logistics 2009 Proceedings Do u University Subject to the constraint I (t) P2 (t) S2 (t) , and I(0) = I0 (T ) The Time Scale Model: C eˆ ( (s),0)(a (I Iˆ2 ) 2 b (P2 Pˆ ) 2 ) s (8) [10] 0 2 2 The above cost function is optimized with subject to I (t) P(t) S(t) and with the initial conditions of I(0) = I0 and P(0) = P0. ˆ ˆ P2 (t) P aI bP t ˆ P ˆ eA (t, t 0 0 ) eˆA (t, ( )) b(1 (t)) (9) I t2 (t) I 0 0 S(t) SOLUTION OF THE MODEL The integrated HMMS model of the reverse logistics is considered in two stages and solved with time scales analysis. Since we want to compare our results and graphs with the existing results for the time scale, R, in our model we will consider the same quantities as have been taken by [8] : TABLE 1 Parameter Specification for Examples Description Specification Demand S t 1 0.05t 1 0.2sin 2 t Product rate goal level P1 t 1 0.05t Inventory size goal level I1 t 0.2 0.2sin 2 t Production cost coefficient c = 4 Inventory holding coefficient h = 20 Rate of discount = 0.05 Pollution – linear P – quadratic 0.5P2 Pollution tax coefficient - linear = 40 - quadratic = 5 Planning horizon T = 5 Graphics of the solutions are given below. Graphic 1 and 2 are the production and inventory graphics with the linear pollution function. They represent the first stage of the reverse logistics integrated HMMS model. 1.6 1.5 1.4 1.3 1.2 1.1 1 2 3 4 5 0.9 GRAPHIC 1 The Graph of Production P(t) 228 Research in Transport and Logistics 2009 Proceedings Do u University 0.2 0.15 0.1 0.05 1 2 3 4 5 GRAPHIC 2 The Graph of Inventory I(t) In graphic 3 and 4, the structure of the forward logistics with a quadratic pollution function is seen. 2000 1500 1000 500 1 2 3 4 5 GRAPHIC 3 The Graph of Production P(t) 1.4 1.3 1.2 1.1 1 2 3 4 5 0.9 GRAPHIC 4 The Graph of Inventory I(t) 229 Research in Transport and Logistics 2009 Proceedings Do u University The two cases show, that the pollution charge in quadratic case reduces the inventory level, and the production rate is smoother then without charge. CONCLUSION As it can be seen from the solution graphics, when there is a linear pollution function a direct change is seen between pollution and production and inventory costs, but when the pollution function becomes quadratic the relation in production costs becomes parabolic with t. But the changes in inventory costs are mostly in a sinusoidal structure. The solution of the integrated logistic HMMS model proposed a social effect. This effect is more distinguished in the quadratic form than the linear form. REFERENCES [1] Committee on the Impact of Academic Research on Industrial Performance (CB). 2003. The Impact of Academic Research on Industrial Performance. Washington DC, USA: National Academies Press:147. [2] Rogers DS, Tibben-Lembke RS.,1999, Going backwards: R L trends &practices. Pittsburgh, PA: Reverse Logistics Executive Council. [3] Dowlatshahi S., 2000, “Developing a theory of Rev. Logistics.”, Interfaces, 30(3), 143–55. [4] Ginter P.M., Starling, J.M., 1978, “Reverse distribution channels for recycling,”, California Management Review, 20 (3), 72-82. [5] Kroon, L., Vrjens., G., 1995, Returnable containers: An example of reverse logistics.”, International Journal of Physical Distribution and Logistics Management, 25(2), 56-68. [6] Gonzales-Torre, P.,L., Adenso-Diaz, B. and Artiba, H., 2004, „Environmental and reverse logistics policies in European bottling and packaging firms.“ International Journal of Production Economics, 88, 95-104. [7] Holt, C. C., Modigliani, J.,F., Muth, H.A. and Simon, H.A., 1960, Planning Produciton, Inventories, and Work Forces, Prentice-Hall, Englewood Cliffs, NJ. [8] Dobos, I., 1998, “Production-inventory control under environmental constraints, International Journal of Production Economics, 56, 123-131. [9] Hilger, S., 1988, “Ein Masskettenkalkül mit Anwendung auf Zentrumsmanningfaltigkeiten, Ph. D. Thesis, Universitat Würzburg. [10] Atici, F.,M. and Uysal, F., 2008, “A production-inventory model of HMMS on time scales, Applied Mathematics Letters, 21, 236-243. 230 Research in Transport and Logistics 2009 Proceedings Do u University FREIGHT TRANSPORT PLANNING WITH GENETIC ALGORITHM BASED PROJECTED DEMAND Soner HALDENBILEN 1, Ozgur BASKAN 2, Huseyin CEYLAN 3 and Halim CEYLAN 4 Abstract- Road freight transport needs to be estimated using the current status of the transport using the socio-economic and transport related indicators and then the planning may be carried out in order to achieve the goals of sustainable transport system. Estimation of road freight tone-km/year is carried out using population, gross domestic product and number of vehicles i.e. lorries using genetic algorithm (GA) approach. Based on Genetic Algorithm (GA) approach, four forms of the Freight Transport Demand Model (GAFTM) are proposed. Best fit model to historical data is selected for future estimation. The estimated road freight tone-km is transferred to railway freight. Transformations are made to obtain one equivalent value of lorry that carries how much freight and correspondingly one train. Net-tone/km for train and lorry is obtained. After that three scenarios are proposed to control road freight tone-km by transferring the 1%, 2% and 3% of the road freight to railway freight. Extra income that is obtained by diverting road freight traffic to railway is calculated. Results showed that about $180 million dollar can be gained and 84000 lorries may be saved from road in 2025 if scenario 3 is applied. Keywords- Genetic algorithm, road freight tone-km, railway freight, freight transport planning INTRODUCTION Road and rail freight transport provides transport and environmental policy with some of its most intractable problems. Lorries are visually very intrusive, noisy, polluting and responsible for much of the impetus behind road building strategies. They are the most visible component of a relatively new and sophisticated production and distribution system that has evolved in a way that weakens local production and consumption links and encourages longer distance supply lines. Over time the distances over which freight moves have lengthened and the amount of dependence on distant sources and complex road freighting operations has increased. In order to understand the forces that currently mould road freight operations we have to be aware of the importance of the spatial distribution of manufacturing and the geographical location of raw material and intermediate product inputs into a final manufactured product. Such awareness can reveal the beginnings of a new strategy that will move freight transport operations in the direction of sustainable development. Reference [1] has made these processes much more transparent and revealed the opportunities provided by substituting "near" for "far" in sourcing decisions. Freight transport strategies have to be alive to a number of influences. They must recognize the importance and growing importance over time of emissions from this sector. These emissions have well recognized negative impacts on human health and even though lorries form a relatively small part of the total number of vehicles their impact on emission inventories is disproportionately large. Freight transport strategies must recognize the commercial importance of moving goods around and satisfying the transport demands from other economic sectors. This will require careful negotiation with interested parties and careful management of all transport modes and all possibilities for local sourcing. Freight transport strategies must reflect the importance of environmental and sustainable development objectives. Forecasts of future levels of demand in road freight transport vary enormously. European Union (EU) documentation refers to a doubling of road freight [2]. More analytical studies with a well defined time framework have produced a percentage increase in tone kilometers of road freight of up to 149 [3]. 1 Soner Haldenbilen, Pamukkale University, Engineering Faculty, Department of Civil Engineering, Denizli, Turkey, shaldenbilen@pau.edu.tr 2 Faculty, Department of Civil Engineering, Denizli, Turkey, obaskan@pau.edu.tr 3 Hüseyin Ceylan, Pamukkale University, Engineering Faculty, Department of Civil Engineering, Denizli, Turkey, hceylan@pau.edu.tr 4 Halim Ceylan, Pamukkale University, Engineering Faculty, Department of Civil Engineering, Denizli, Turkey, halimc@pau.edu.tr 231 Research in Transport and Logistics 2009 Proceedings Do u University References [3]-[4] predicted a growth of 58% in tone kilometers over the period 1990-2010. In the same period fuel consumption will rise by 23-57% even taking into account improvements in energy efficiency. Forecasts of heavy goods vehicle traffic in Great Britain [5] are based on a constant relationship between GDP and road tone kilometers. The forecast of vehicle kilometers (all heavy goods vehicles) for the period 1988-2025 is for a low growth rate of 67% and a high growth rate of 141%. Forecasting is a very inexact science and past forecasts have underestimated the size of the growth in both passenger kilometers and tone kilometers. Current transport policies are discriminating against rail, coastal shipping and waterways. There is no such thing as a level playing field and the mythology of a free market in transport could not be further from the truth. There is no market mechanism guiding the flow of funds into road building programmes. Freight transport modelling and planning prevents traffic congestion in rural roads and reduces the resource allocation for building new highway network and improved safety. In order to make a good decision making for future prospects of road and rail freight transport, demand for freight in both highways and railways needs to be estimated with mathematical methods. One of the new methods, Genetic Algorithms (GA), first developed by [6]-[7] proposed in this study. It is a quite new method to estimate demand for freight in rural roads. Based on Genetic Algorithm (GA) approach, Freight Transport Demand Models (GAFTM) are developed that use the population, the Gross Domestic Product (GDP) and the Number of Vehicles (NoV) as inputs. One of the main reasons for choosing the GA approach is that the socio-economic and transport related indicators may affect the freight demand in non-linear behaviour. MODEL DEVELOPMENT The GAFTM models use the GA notion that has been developed by [6]. Reference [7] applied its notion to the engineering problems. It is a iterative process that involves reproduction, crossover and mutation. The main advantage of GAs is their ability to use accumulating information about initially unknown search space in order to bias subsequent searches into useful subspaces. GAs differ from conventional nonlinear optimization techniques in that they search by maintaining a population (or data base) of solutions from which better solutions are created rather than making incremental changes to a single solution to the problem. Definition of the GA and its application to transport demand modelling may be obtained in [8]-[10]. The four forms of the GAFTM models are developed in the following way. Exponential form of the GAFTMexp model is: GAFTM w3exp w1 w2 X1 w4 X w5 w X w72 6 3 (1) Quadratic forms of the GAFTM quad models are: GAFTM quad 0 w1 w2 X 1 w3 X 2 w4 X 3 w5 X 1 X 2 (2) w6 X 1 X 3 w7 X 2 X 3 w8 X 1 X 2 X 3 GAFTM quad1 w1 w2 X w3 1 w X w5 4 2 w w7 6 X 3 w8 X1X 2 w9 X1X 3 (3) w10 X 2 X 3 w11X1X 2 X 3 GAFTM quad 2 w1 (w2 X1 w X w5 3 2 w4 X 3 ) (4) where X 61 is the population (10 ), X2 is the GDP (109$) and X3 is the NoV (105). After applying the GAFTM models to estimate road freight transport using data on Table 1, the following weighting parameters are obtained. GAFTM 0.77 0.26X 0.44 0.00X 0.31X 1.35 2exp 1 2 3 R = 0.92 (5) GAFTM quad 0 0.978 0.00X 1 0.00X 2 0.392X 3 0X 1 X 2 R2= 0.91 (6) 0.016X 1 X 3 0X 2 X 3 0X 1 X 2 X 3 GAFTM 0 9.041X 0.001quad1 1 0.00X 2 0.0X 3 0X1X 2 R2= 0.91 (7) 0.0189X1X 3 0X 2 X 3 0X1X 2 X 3 GAFTM quad 2 4.88 ( 1.47X1 0.16X 2 3.34X ) 0.98 R23 = 0.91 (8) 232 Research in Transport and Logistics 2009 Proceedings Do u University The validation of the four forms of the GAFTM models may be obtained in [8]. Tests were carried out based on the minimum relative error in testing period. The minimum error obtained in GAFTMexp model, thus it is selected for future freight transport demand estimation. TABLE 1 Used data for road freight demand estimation Population (106) GDP(109$) Goods transport Years NoV(105) tone-km (109) X1 X2 X3 GAFTM 1980 44.74 69.75 33.07 37.51 1981 45.86 72.78 34.46 39.01 1982 47.00 65.94 35.95 40.57 1983 48.18 62.19 37.67 42.19 1984 49.38 60.76 39.58 43.88 1985 50.66 68.20 41.80 45.63 1986 51.78 76.46 44.19 54.02 1987 52.92 87.73 45.94 58.83 1988 54.08 90.97 47.49 65.46 1989 55.27 108.68 49.00 68.24 1990 56.47 152.39 52.08 65.71 1991 57.50 152.35 55.43 61.97 1992 58.55 160.75 59.53 67.70 1993 59.61 181.99 65.98 97.84 1994 60.70 131.14 68.82 95.02 1995 61.81 171.98 71.92 112.52 1996 62.93 184.72 77.61 123.67 1997 64.08 194.36 88.34 124.34 1998 65.24 205.98 99.72 135.27 1999 66.43 187.66 107.19 134.41 2000 67.64 201.48 118.87 141.82 2001 68.59 144.00 122.97 151.42 2002 69.82 181.00 127.44 150.91 2003 69.93 238.53 137.85 152.16 2004 70.85 301.53 190.73 156.85 2005 71.76 359.97 215.20 166.83 2006 72.67 380.62 240.52 177.40 Sources: General Directorate of Turkish Highways [11], State Planning Organization [12] ROAD FREIGHT TRANSPORT DEMAND IN FUTURE When road and railway freights are analyzed, the road freight is increased about 4 times, but railway freight is not considerably changed within last 21 years for the period of 1985-2006. These trends will continue if efficient freight plan is not made. Figure 1 shows the general trend of road and railway freight between 1985 and 2006 indexed at 1985 as 1. Road freight transport demand is forecasted under different scenarios using these values in this study. Estimation of road tone-km is carried out after forecasting the socio-economic and transport related indicators. The estimation of population, GDP and NoV is carried out in the following way. 1. Population: State Planning Organization (SPO) [12] plans and controls the population growth rate in Turkey according to the 5 years National Development Plans (NDP). This plans show that the growth rate of population is separated into two categories. One is the real growth and the second is the targeted growth rate. It indicates that the population growth rate increase with a decreasing trend especially within last 15 years. Therefore, it may be better to estimate the population of Turkey in 2025 according to the two-case: One is the current population growth rate that can be obtained from 233 Research in Transport and Logistics 2009 Proceedings Do u University the observations and it is named as Case I, and the second is the targeted growth rate according to the NDP named as Case II. 4.50 Railway Highway 3.50 2.50 1.50 0.50 1985 1990 1995 2000 2005 Year FIGURE 1 The trend of road and railway freights for the period of 1985-2006 (fixed at 1985=1) [11]-[13] 2. GDP: The observed GDP is in the fluctuating trend. Therefore, it could be better to take the average growth rate of the GDP under various cases. The cases can be explained as: Case I: Take the average growth rate of the observed period for the GDP as a future growth rate (i.e. annual average growth rate is 4% within last 20 years), Case II: Assume that the GDP of Turkey by means of per capita will meet the EU average in 2025 (i.e. 6% annual growth rate). The projected GDP can be obtained in [9] based on the Case I and Case II. 3. NoV: Number of light goods vehicles (LGV) and trucks show linear increase within the last 30 years. Therefore, the following linear equations are used to predict NoV in the future. y = 55667x + 173990 R2 = 0.94 (9) y = 11657x + 139389 R2 = 0.99 (10) where y is number of lorries and x is the time series where 1985=1, 1986=2…. Projected number of vehicles (NoV) is given in Table 2. The NoV for goods transport is 2.8*106 and the total NoV is 22.5*105 in 2025. TABLE 2 Projected NoV Years 2010 2015 2020 2025 NoV (105) 178.81 212.47 246.13 279.79 Expected road freight tone-km may be analyzed under four combinations as: Population GDP NoV I. Combination Case I Case I Table 2 II. Combination Case I Case II Table 2 III. Combination Case II Case I Table 2 IV. Combination Case II Case II Table 2 Application of GAFTMexp model for road freight tone-km under four combinations can be seen in Figure 2 for the period of 2007-2025. The lowest and highest estimated road freight tone-km is about between 530 and 234 Index Research in Transport and Logistics 2009 Proceedings Do u University 575*109 tone-km/year for combinations I and II. Therefore, combination I and II is selected for future analysis. 600 Combination I Combination II 500 Combination III Combination IV 400 300 200 100 Years FIGURE 2 Estimated road freight tone-km under four combinations SCENARIOS AND FREIGHT DEMAND ANALYSIS The forecasted road freight transport is transferred to the railways under three scenarios. It is assumed that until 2010 there are no policy changes for all scenarios. All the analyses are carried out after 2010. If some part of the road freight is transferred to a railway, then it requires some extra time to plan its coming extra demand from road freight. Scenario 1: Each year 1% of the road freight is transferred to railway and it steadily increases, meaning that 15% decrease on road tone-km in 2025. Scenario 2: Each year 2% of the road freight is transferred to railway and it steadily increases, means that 30% decrease on road tone-km in 2025. Scenario 3: Each year 3% of the road freight is transferred to railway meaning that and 45% decrease on road tone-km in 2025. TABLE 3 Transferred road freight tone-km to railway for the period of 2011 to 2025 Expected road Transferred road freight (106 tone-km) to railway Year freight demand freight (106 tone-km) Scenario 1 Scenario 2 Scenario 3 2011 303.53 3.04 6.07 9.11 2012 319.02 6.38 12.76 19.14 2013 334.92 10.05 20.10 30.14 2014 351.26 14.05 28.10 42.15 2015 368.03 18.40 36.80 55.20 2016 385.00 23.10 46.20 69.30 2017 402.41 28.17 56.34 84.51 2018 420.26 33.62 67.24 100.86 2019 438.57 39.47 78.94 118.41 2020 457.33 45.73 91.47 137.20 2021 476.33 52.40 104.79 157.19 2022 495.80 59.50 118.99 178.49 2023 515.75 67.05 134.09 201.14 2024 536.19 75.07 150.13 225.20 2025 557.13 83.57 167.14 250.71 235 Tone-km (109) 2007 2009 2011 2013 2015 2017 2019 2021 2023 2025 Research in Transport and Logistics 2009 Proceedings Do u University Forecasted road freight tone-km and transformations to railway freight can be seen in Table 3. As can be seen, the minimum values of 84*106 tone-km and 250*106 tone-km can be transported by train according to scenario 1 and 3, respectively. After transferring the road freight tone-km to railway will lead to change the railway freight net-ton-km and it creates some extra income. Current level of railway revenues for only freight transport can be seen in Figure 3. Figure indicates that there is a floating trend in observed revenues between 1985 and 2000. After 2001 observed revenues slightly increased. It is expected that increasing trend of freight income will continue in future years. 0.06 0.05 y = 4E-05x2 - 0.0009x + 0.0182 0.04 0.03 0.02 0.01 0.00 Year FIGURE 3 Observed and expected revenues for railway freight net-tone/km. In order to calculate the equivalent unit value of road freight tone-km, there is a need to estimate how much the average value of one lorry that moves with a goods and similarly one train. It is obtained that one lorry carried 3 tones and one train carries 170 tones. Using these values, transformations from road freight to railway freight are made. Results can be seen in Table 4. Expected extra revenues from rail transport are about $60*106 to $180*106 in 2025 for scenario 1 and 3, respectively. TABLE 4 Divided railway freight demand and extra revenue for railways Year Net-tone/train-km (106) Extra revenues ($106) for railways Scenario 1 Scenario 2 Scenario 3 Scenario 1 Scenario 2 Scenario 3 2011 55 110 166 1.03 2.06 3.09 2012 116 232 348 2.28 4.56 6.84 2014 255 511 766 5.58 11.16 16.74 2015 335 669 1004 7.72 15.43 23.15 2016 420 840 1260 10.23 20.46 30.69 2017 512 1024 1536 13.18 26.37 39.55 2018 611 1223 1834 16.63 33.25 49.88 2019 718 1435 2153 20.63 41.25 61.88 2020 832 1663 2495 25.24 50.49 75.73 2021 953 1905 2858 30.54 61.08 91.63 2022 1082 2163 3245 36.61 73.21 109.82 2023 1219 2438 3657 43.52 87.04 130.56 2024 1365 2730 4095 51.37 102.75 154.12 2025 1519 3039 4558 60.26 120.52 180.78 236 Income ($/netton/km) 1985 1988 1991 1994 1997 2000 2003 2006 2009 2012 2015 2018 2021 2024 Research in Transport and Logistics 2009 Proceedings Do u University CONCLUSIONS This study deals with estimation of road freight transport demand indicators in Turkish rural roads and analysis the railway extra revenues if some part of the road freight is transferred to railways. The GAFTM models are developed using population, GDP and number of vehicles. The road freight demand is projected with four cases under four combinations of the cases. GA approach is selected as a methodology so that road freight tone-km may be better estimated by the non-linear form of the mathematical expressions. Among the four forms of the GAFTM models, the best of the GAFTM model is selected in terms of minimum total average relative errors in testing period. The following results can be drawn from this study. The analysis has shown that the potential for reducing the number of lorry movements in Turkey is very large and much larger than previously recognized. This reduction can be achieved in part by transferring freight from road to rail It is clear, however that a transfer of this kind cannot represent a fundamental solution to the problems of rising ton-km of road freight. Freight transport offers a number of attractive options for building alternatives. Establishing the importance of regional and local production/consumption links and reducing the basic demand for freight transport is one of these alternatives. The case for fundamental demand reduction in road freight transport is a strong one and the time has arrived when continuing to develop along the same path as the last 20 years is no longer acceptable and is in clear conflict with sustainable development objectives. There is a way forward and sustainable development is a stimulus to innovation and experimentation that will chart a new course. Analysis showed that minimum of 1000 lorries discarded from road traffic according to scenario 1 in 2011 and 83570 lorry will be discarded from road traffic if scenario 3 is applied in 2025. This means that improved road safety and environmental pollution and extra revenue for railways. ACKNOWLEDGEMENT The work was partially supported by the Scientific and Technological Research Council of Turkey (TUBITAK) under Grant 104I119 and Scientific Research Foundation of the Pamukkale University with the project number 2007-FBE-003. REFERENCES [1] Böge, S. 1994. The Well Travelled Yoghurt Pot: Lessons for new freight transport policies and regional production. In World Transport Policy and Practice, MCB, Bradford. [2] CEC, 1994. Road Freight Transport: the single European Market. Report of the Committee of Enquiry, July 1994, DGVII, Commisson of the European Communities, Brussels. [3] Hey, C, Hickmann, G, Geisendorf, S., Schleicher-Tappeser, R., 1992. Dead End Road: Freiburg, Germany. [4] Bleijenberg, A., 1993. Fiscal measures as part of a European Transport Policy in ECMT, Transport Policy and Global Warming, OECD, Paris, 141-154. [5] HMSO, 1989. National Road Traffic Forecasts (Great Britain), Department of Transport. [6] Holland, J. H, 1975. Adaptations in Natural Artificial Systems. University of Michigan Press, MI. [7] Goldberg. D.E., 1989. Genetic Algorithms in Search, Optimization and Machine learning. Addison-Wesley, Harlow, England. [8] Haldenbilen, S. and Ceylan H., 2005. “The development of a policy for road tax in Turkey, using a genetic algorithm approach for demand estimation”. Transportation Research 39A, 861-87. [9] Haldenbilen, S. and Ceylan H., 2005. “Transport Demand Estimation based on Genetic Algorithm Approach” Transportation Planning and Technology, V(28), No:6, 403-426. [10] Haldenbilen, S., 2003. Investigation of indicators of sustainable transportation systems using genetic algorithm approach: a case study of Turkey”. PhD Thesis, Pamukkale University, Turkey. [11] GDTH, General Directorate of Turkish Highways, 2008. Statistical Yearbook on Turkish Highways 1990-2006. The Republic of Turkey, Ministry of Public Works and Settlement General Directorate of Highways, Planning division, Ankara, Turkey. [12] SPO, State Planning Organization, 2008. Available at http//.www.dpt.gov.tr. [13] TR, Turkish Railways, 2008. Available at http//.www. tcdd.gov.tr. 237 Chapter 10 Transport Modeling Research in Transport and Logistics 2009 Proceedings Do u University INVERSE MODEL TO ESTIMATE O-D MATRIX FROM LINK TRAFFIC COUNTS USING ANT COLONY OPTIMIZATION Halim CEYLAN1 Soner HALDENBILEN2 Huseyin CEYLAN3 Ozgur BASKAN4 Abstract- Estimation of Origin–Destination (O-D) trip table has long been carried out with maximum entropy, generalized least squares, bi-level programming methods etc. All optimization models developed so far are either calculus based and mathematically lengthy. All approaches are prone to local optima. Therefore, there is a need for a new method which estimates O-D trip table and solves the traffic assignment models simultaneously. The main objective of this study is therefore to develop a model to estimate O-D trip table from link traffic counts using ant colony optimization (ACO) method based on inverse modelling technique. The ACO has been recently developed, as a population based meta-heuristic that has been successfully applied to several NP-hard combinatorial optimization problems. The core of ant’s behavior is the communication between the ants by means of chemical pheromone trails, which enables them to find shortest paths between their nest and food sources. Ant Colony optimization O-D Estimation (ACODE) model is formulated as simultaneous optimization problem, the O- D trip matrices and stochastic user equilibrium (SUE) path and link flows are obtained simultaneously. The ACODE model is applied to an example transportation network which has 13 nodes with 19 links, 25 routes and 4 O-D pairs. O-D trip table is estimated using proposed ACODE model from given link traffic volumes. Results showed that inversely applied ACODE model for O-D matrix estimation from link traffic counts may estimate the O-D trips under SUE assumption. Keywords-O-D estimation, link traffic counts, ant colony optimization, stochastic user equilibrium INTRODUCTION The Origin–Destination (O–D) trip table estimation is an essential element of network based traffic models. Estimating O-D matrix from traffic counts on road links is of considerable importance. The O-D is also an essential ingredient in a wide variety of travel analysis and planning studies [1]. Over the past several decades, a considerable number of methods for O-D estimation have been reported in the literature. O-D matrix is the basic data for the traffic planning and management. It is a demand for traffic that flows from origins to destinations, which is expressed as a matrix to explore the movement of space flow. Statistical techniques have become popular in the estimation or updating of O-D matrix from traffic counts. The traditional way of estimating O-D from home-interview survey data is expensive [2]. Hence, generally, the estimates are based on small sample of home- interview data and thus the accuracy of the estimates suffers. This led the researchers to estimate the O-D from a variety of other data sources among which O-D estimation from link traffic counts has attracted lot of interest as the required data collection is simple and routine. The O-D demand matrix estimation methods in the literature and its advantages and disadvantages are given in the next section. This paper is structured as follows. The next section reviews the O-D estimation matrix estimation methods. An improved ant colony optimization method and its solution procedure are proposed in Section 3. The algorithm defined to estimate the O-D matrix using improved ant colony optimization with inverse model from the link traffic counts is given in Section 4. In Section 5, a numerical example is carried out to present effectiveness for proposed algorithm. Finally, our conclusions can be seen in the last section. 1 Halim Ceylan, Department of Civil Engineering, Engineering Faculty, Pamukkale University, Denizli, 20017, Turkey, halimc@pamukkale.edu.tr 2 Soner Haldenbilen, Department of Civil Engineering, Engineering Faculty, Pamukkale University, Denizli, 20017, Turkey, shaldenbilen@pamukkale.edu.tr 3 Hüseyin Ceylan, Department of Civil Engineering, Engineering Faculty, Pamukkale University, Denizli, 20017, Turkey, hceylan@pamukkale.edu.tr 4 Department of Civil Engineering, Engineering Faculty, Pamukkale University, Denizli, 20017, Turkey, obaskan@pamukkale.edu.tr 241 Research in Transport and Logistics 2009 Proceedings Do u University LITERATURE REVIEW O-D matrix estimation has been studied by a few researchers and notable developments have been achieved at this concept of transportation network design. Reference [3]-[5] were interested in finding probable O-D movements in terms of link flows by deterministic assignment equation. Reference [6] was the first to attempt the equilibrium assignment based O-D estimation. The formulation, however, does not ensure a unique O-D solution because the formulation is not strictly convex in the O-D variables. Uniqueness of the solution is ensured if a target O-D is used [7]-[9]. These models require a complete set of link counts, a target O-D matrix and that the observed flows be in equilibrium. Two significant methods which do not use entropy formulation and include generalized least squares estimation were developed by [10] and [11]. Reference [2] developed another method which also uses generalized least squares and also allows the explicit use of data describing the structure of the O-D. Reference [12] proposes a model which inferences about an O-D matrix from a single observation on a set of link flows. Two problems are discussed in this study; the first, the problem of reconstructing the actual number of O-D trips, and the second, estimation of mean O-D trip rates. A fast constrained recursive identification (CRI) algorithm is proposed to estimate O-D matrices by [13]. The basic idea of the CRI algorithm is to estimate intersection O-D matrices based on equality-constrained optimization. A Fuzzy inference based assignment algorithm to estimate O-D matrices from link volume counts is proposed by [1]. Reference [14] proposes a new model which has been formulated by using a new approach called the calibration and demand adjustment model (CDAM) based on bi-level programming which simultaneously estimates an O–D matrix and the parameters for the nested logit model. The algorithm iterates between the network equilibrium problem and that which is used to obtain a set of paths when equilibrium is attained, and the CDAM is restricted to the set of previously generated columns. The computational tests on the algorithm have been carried out using data from a multi-modal network in Madrid. In this study, an improved ant colony optimization (IACO) based algorithm which is called ACODE is proposed to estimate O-D demands on transportation networks. The ACODE model considers Stochastic User Equilibrium (SUE) conditions for modeling drivers’ route choice perceptions. The methodology is given in the next sections and the model is applied to a test network. ANT COLONY OPTIMIZATION Ant Colony Optimization (ACO) belongs to the class of biologically inspired heuristics. The ACO is the one of the most recent techniques for approximate optimization methods, was initiated by [15]. The core of ant’s behavior is the communication between the ants by means of chemical pheromone trails, which enables them to find shortest paths between their nest and food sources. The Improved algorithm for ACO (IACO) that is proposed in this study is based on each ant searches only around the best solution of the previous iteration with coefficient . It is very important for improving IACO’s solution performance. IACO differs from other ACOs in that its feasible search space (FSS) is reduced with coefficient and its best solution obtained using information on the previous iteration. At the core of IACO, ants search randomly the solution within the FSS to reach optimum or near-optimum values. At the end of the each iteration, only one of the ants is near to global minimum. After the first iteration, when global minimum is searched around the best solution of the previous iteration using the IACO will quickly reach to the global minimum. IACO is performed by modifying the algorithm proposed by [16]. The algorithm can be defined in the following way. At the beginning of the first iteration, all ants search randomly best solution of the problem within the FSS. At the end of the first iteration, FSS is reduced by and best solution obtained of the previous iteration is kept. Optimum solution is then searched in the reduced search space during the steps of algorithm progress. IACO reaches to the global minimum as ants find their routes in the limited space. guides the bounds of search space throughout the IACO application. The main idea of proposed algorithm is given in Figure 1. Main advantageous of IACO is that the FSS is reduced with coefficient and it uses the information taken from previous iteration. For example, consider a problem of five ants represents the formulation of the problem. Five ants being associated five random initial vectors. Only one of the solutions which were obtained at the end of the first iteration is near to 242 Research in Transport and Logistics 2009 Proceedings Do u University global minimum. After the first iteration, FSS is reduced according to coefficient and best solution (Ant 1 is the best solution, as shown in Fig. 1(a)) of the previous iteration. FSS is getting smaller during iteration progress as shown in Fig. 1(b). Coefficient has been chosen according to the size of search space and constraints the problem in order not being trapped in bad local minimum. In IACO, let number of m ants being associated with m random initial vectors (xk (k 1,2,3,......m)) . Quantity of pheromone ( t ) only intensifies around the best objective function value. The solution vector of the each ant is updated using (1). xk (new) xk (old )t t (t 1,2,....., I ) (1) (a) (b) Ant 5 Ant 1 Ant 3 Ant 2 Ant 5 Ant 1 Global Ant 3 Ant 2 minimum Ant 4 Ant 4 FSS (Ant 1-ßt ) v. The minimum annual transportation cost is achieved when the derivative of C(v) with respect to v is zero. Thus, C(v) is minimal when ncF cV m cH m 0 (5) Hence, the optimal fleet composition can be determined by requiring that m n cF (6) cH cV Sensitivity Analysis Optimum fleet size in terms of cost may not result in the desired service quality. Variation of demand over the year results in peak period, in which demand can not be served by the optimum fleet size. In such case, on time delivery can not be achieved. Thus, we will examine the effects of fleet size on the on time performance as measured by a percentage of time periods, in which demand can be served by available fleet. RESULTS Vehicle Routing Figure 2 shows the number of required vehicles over the year 2007 as obtained from the actual operation and saving algorithm. Daily demand is variable, thus the number of required vehicles also varies over the year. Nevertheless, it is clear that efficient vehicle routing can reduce fleet size as shown by a decrease in the average number of required vehicles from 70.4 to 59.4 vehicles daily – a decrease of 15 percent. Figure 3 shows the effects of change in vehicle routing on fleet utilization. Owned fleet is utilized more efficiently as shown by an increase in the average running distance from 615 to 705 km per day. On the other hand, hired vehicles are utilized less as shown by a decrease in the average running distance from 316 to 265 km per day. 90 80 80 70.4veh / day 70 59.4veh / day 70 60 60 50 50 40 40 30 30 20 20 10 10 0 0 45 50 55 60 65 70 75 80 85 90 95 35 40 45 50 55 60 65 70 75 80 85 (a) Actual operation (b) Saving algorithm FIGURE 2 The Number of Required Vehicles Daily in 2007 309 Research in Transport and Logistics 2009 Proceedings Do u University 800 700 600 Actual operation 500 400 Saving algorithm 300 200 100 0 Owned Fleet Hired vehicles FIGURE 3 Comparison of Vehicle Utilization Optimum Fleet Composition Optimum fleet composition can be determined by using (6). Daily operation is considered as a time period, resulting in n = 365. cF, cV and cH are based on actual data of the case study with the values of 4.2, 7.84 and 15.99 baht/km respectively [8]. Optimum fleet composition for the case study is equivalent to 53 owned and 7 hired vehicles, compared to 49 owned and 22 hired vehicles employed in the actual operation. Due to lower cost of owned fleet (cF + cV < cH), the annual transportation cost can be saved by 4.2 percent or equivalent to 7.3 million baht yearly. Sensitivity analysis Table 1 summarizes the on time performance with respect to the number of hired vehicles. It can be seen that the optimum fleet size produces 63 percent on time delivery. If 85 percent on time delivery is to be achieved, the number of hired vehicles must be increased to 14. Consequently, annual cost saving is reduced to just 2.8 percent. It is clear that on time performance is very sensitive to fleet size. Therefore, it must be incorporated into the decision on fleet size and composition. TABLE 1 On Time Performance under Different Fleet Size Number of hired vehicles Total truck (Veh/day) On time delivery (%) (Veh/day) Cost saving (%) 0 29.6 53 5.6 6 50.0 59 4.4 7* 63.0 60 4.2 14 (+1.70 ) 85.0 67 2.8 16 (+1.95 ) 90.0 69 2.5 19 (+ .31 ) 95.0 72 1.9 *optimum number of hired vehicles **standard deviation of number of required vehicles ( ) = 8.11 vehicles/day CONCLUSIONS Management of freight transportation for retail business, which employs both owned fleet and hired vehicles, is a challenging task. The operating cost of owned fleet is lower than that of hired vehicles. However, fixed cost of owned vehicle is high, unless it is efficiently utilized. From this viewpoint, the goal would be to maximize owned fleet utilization and outsource excess load to third party logistics provider. TOPS supermarket, which is a leading retailer in Thailand, was chosen as a case study. Vehicle routing with time windows was employed to ensure that vehicles are utilized in the most efficient manner. It is proved that efficient vehicle routing can save the average number of required vehicles compared to the actual operation. The effects of change in vehicle routing on fleet utilization were also observed. 310 Average distance covered (km) Research in Transport and Logistics 2009 Proceedings Do u University The study also examined the optimum fleet size and composition for the case study. The result shows that efficient route sequencing and fleet size optimization can reduce the annual transportation cost by 4.2 percent or equivalent to 7.3 million baht yearly. Sensitivity analysis indicated that logistics manager needs to trade of cost saving against desired service quality. REFERENCES [1] Abshire and Premeaux, 1991, “Motor carrier selection criteria: perceptual differences between shippers and carriers”, Transportation Journal, 31 (1), 31-35. [2] Ghiani. G., Laporte, G. and Musmanno, R., 2004, “Introduction to Logistics Systems Planning and Control”, John Wiley & Sons Ltd., UK. [3] Clarke, G. and Wright, J., 1964, “Scheduling of vehicles from a central depot to a number of delivery points”, Operations Research 12 (4), 568-581. [4] A.M. Campbell and Martin Savelsbergh, 2004, “Efficient insertion heuristics for vehicle routing and scheduling problems”, Transportation Science, 38 (3), 369-378 [5] Braysy, O., Gendreau, M., 2005. “Vehicle routing problem with time windows. Part II: Metaheuristics”, Transportation Science 39(1), 119-139. [6] Hansen, P., Mladenovic, N., 2003. “A tutorial on variable neighborhood search”, Les Cahiers du GERAD, G-2003- 46, HEC Montreal, Canada. [7] Yang, J., Jaillet, P., Mahmassani, H., 2004 “Real-time multi-vehicle truck load pickup-and-delivery problems” Transportation Science 38(2), 135-148. [8] Kawee Srimuang, 2007, “Fleet Size Determination in Freight Transportation for Retail Business”, Master Thesis, King Mongkut’s University of Technology North Bangkok. ACKNOWLEDGEMENT This study was supported by a research grant from Thailand Research Fund and King Mongkut’s University of Technology North Bangkok. The authors would like to express their gratitude to TOPS supermarket for providing actual data on truck scheduling and operating costs. 311 Research in Transport and Logistics 2009 Proceedings Do u University NEW CONTAINER PORT DEVELOPMENT: FORECASTING FUTURE CONTAINER THROUGHPUT Dimitrios TSAMBOULAS1, Panayota MORAITI 2 Abstract Container ports constitute nowadays important nodal points in the entire global logistic chain of containerized freight. Forecasting the container throughput to a new port is a challenging task. This paper presents a simple methodology for carrying out an initial forecasting of future container traffic through a new port to be constructed, at the preliminary stage of planning. It is, therefore, aimed at providing a planning context for informed decision making by governments, shipping lines and port authorities, as well as potential investors and global port operators. The methodology has been applied for estimating demand for a proposed port development in Greece. Keywords container throughput, forecasting, logit model INTRODUCTION The shipping and port industry must respond to a new reality. The ongoing globalization processes, the current transport needs related to consumer and production patterns and the recent economic and technological changes, have created a new environment, full of challenges for all interested parties. The continuing growth of world trade, particularly from emerging economies, such as China, has resulted in increased demand for transport services, whilst the advantages of seaborne transport in terms of economic efficiency, security and safety, as well as environmental protection have become obvious. On the other hand, there is a strong emergence of global operators, as a result of policy measures that enhance competition, as ports tend to be divided into competing terminals and national monopolies are being decentralized, leading to new market opportunities for companies to operate further than the confines of their origin countries’ borders. Therefore, container ports constitute at present important nodal points in the entire global logistic chain of containerized freight, and there is a continuously increasing demand for them and their related services, resulting in low-risk investments in this sector. Short Sea Shipping, according to the European Commission, is the only transport mode able to keep up with the rapid economic growth of the EU, and is, therefore, regarded as the only freight mode that offers a realistic prospect for substantial modal shift from road in the future, as well as the one contributing to improving competitiveness, reducing environmental damage, and enhancing cohesion in an expanding EU [1]. More specifically, a similar initiative, the Motorways of the Sea, has as its main objective the improvement of the existing maritime links or the establishment of new, viable, regular and sufficiently frequent maritime links for the transport of goods, resulting in the concentration of transport flows on a selection of ports or port regions in a door-to-door logistics chain. The development of the ports and their hinterland connections not only contributes to the reduction of transport costs and the promotion of an environmentally friendly transport mode, but also meets the requirements of an intermodal transport system. The forecasting of container throughput to a new port is a challenging task. By providing cargo throughput forecasting and thus future shipping network requirements, port authorities or governments can set up master plans for port development, whilst shipping lines can improve their investment plans. This paper presents a methodology for carrying out an initial forecasting of future container traffic through a new port to be constructed. This is aimed at providing a planning context for informed decision making by governments, shipping lines and port authorities in the region in hand, as well as potential investors and global port operators. The proposed methodology consists of two different forecasting methods, the first based on the attraction of existing flows, with a scope to obtain an upper limit on the value of the future container throughput and to provide input to the second one, a probabilistic approach, based on the application 1 Dimitrios Tsamboulas, Professor, National Technical University of Athens, School of Civil Engineering, Department of Transportation Planning and Engineering. 5, Iroon Polytechniou Str., Zografou Campus, Zografou-Athens, GR-15773, Greece Tel.: +30-210-7721367, Fax: +30-210-7722404, E-mail: dtsamb@central.ntua.gr 2 Panayota Moraiti, Research Associate, National Technical University of Athens, School of Civil Engineering, Department of Transportation Planning and Engineering. 5, Iroon Polytechniou Str., Zografou Campus, Zografou-Athens, GR-15773, Greece Tel.: +30-210-7721265, Fax: +30-210-7722404, E-mail: pmoraiti@central.ntua.gr 313 Research in Transport and Logistics 2009 Proceedings Do u University of a Logit Model. The forecasting methodology has been applied for the specification of the demand for the proposed development of a new port in the south of Greece. BACKGROUND Various port authorities use different methods to forecast traffic of general cargo. A number of them rely on externally produced international trade forecasts and use shift and share methods by broad commodity groups, to estimate what part of the trade of the country will go through each port. Others use trend extrapolations or a mixture of trend analysis and shift and share [2]. In general, port cargo volume is directly related to the level of external trade and the macroeconomic conditions. Thus, regression analysis is often carried out, to relate variations in cargo throughput to variation of economic indicators [3]. In addition, Hong Kong, the busiest container port in the world, has been using a regression analysis approach to forecast port cargo throughput for its port planning and development over the decades [4]. Moon [5] has examined the validity of the port input-output model (PIM) as predictive tool to future planning for identifying the impact of the port industry and calculating the spreading effects of the port industry on the nation's economy. In another approach, aimed at developing an integrated model for forecasting both the number of ship visits and their characteristics, knowledge of future shipping trends for any port was considered, as the future number of ship visits and their characteristics may have implications for the physical facilities of a port. The proposed model identified the major economic determinants as the expected trade throughput, world shipping trends, standards of facilities and future plans of shipping companies/agents [6]. Finally, Zohil and Prijon, [7] in their forecasting of transhipment volumes in the Mediterranean ports, concluded that transhipment volumes depend on the diversion distance and the total traffic throughput. In more detail, the volume of cargo transhipped in a single port is an inverse linear function of the distance of the port from the main-line vessels’ route transiting the Mediterranean, and a linear function of the port container traffic volumes. METHODOLOGY The methodology proposed in this paper is applied on the proposed development of an international container port in the south of Greece, with a view to provide connection to other big ports in the Mediterranean, as well as the Far East (Singapore, China, Korea, Japan and Taiwan) and India. At this preliminary stage, a detailed development and operational plan is not necessary. Any feasibility study, however, would have to include a forecasting methodology. As described in the previous section, there are several ways to estimate the attracted traffic to the new container port. Consequently, in order to explore all possibilities, and take into consideration the fact that the port is not yet in operation, so that actual values could be obtained, it is proposed to employ two distinctive forecasting methods with a scope to derive a more realistic estimate, based on the outputs of each of these methods. Attraction from Existing Routes The first approach entails a simple record of existing flows and a crude assumption of how traffic will be attracted to the proposed port. This assumption is based on the “potential” traffic that can be diverted to the port from existing routes. Thus, this method presents the maximum traffic (upper limit) and can be used as a benchmarking value to provide input to the other methods. It assumes a common rate for transhipment costs, regardless of the port. Direct calling is more profitable than transhipment according to shipping costs, mainly by virtue of the extra feeder costs and container lift charges involved [8]. In other words, in case of intra Mediterranean or Mediterranean-Europe trade routes, wherever transhipment occurs at any ports, the associated costs are more than direct calling. Therefore, container cargo derived from the economic area around the south of Greece will not be transhipped at the New Container Port in order to be shipped to other European and Mediterranean ports. Possible container throughputs will be from ships calling at Gioia Tauro, Taranto, Port Said, Malta, and Piraeus for the trade routes between the western Mediterranean & Black Sea countries (such as Greece, Romania, Bulgaria, Ukraine, Russia, Turkey) and North American & Asian countries, Australia and South Africa. Based on the above assumptions, Table 1 presents the methodology for forecasting the future container throughputs of the New Container Port. 314 Research in Transport and Logistics 2009 Proceedings Do u University Based on the actual traffic data for the base year 2005 [9]-[10], the estimated traffic that could be attracted by the New Container Port is presented in Table 2. TABLE 1 Forecasting Methodology 1 Countries Possible Trade Routes & Container Throughputs Greece(GR) Total Container Throughputs of Greece - {(Trade Volume btw GR & EU) + (Trade Volume btw GR & Med.) + (Trade Volume btw GR & Asia & N. Am. using Ocean Going Direct Calling Trade Route-OGDCTR)} Romania(RO) Total Container Throughputs of Romania - {(Trade Volume btw RO & EU) + (Trade Volume btw RO & Med.) + (Trade Volume btw RO & Asia & N. Am. using OGDCTR)} Bulgaria(BG) Total Container Throughputs of Bulgaria - {(Trade Volume btw Bu & EU) + (Trade Volume btw Bu & Med.) + (Trade Volume btw Gr & Asia & N. Am. using OGDCTR)} Ukraine(UR) Total Container Throughputs of Ukraine - {(Trade Volume btw UR & EU) + (Trade Volume btw UR & Med.) + (Trade Volume btw UR & Asia & N. Am. using OGDCTR)} Black Sea Area, Total Container Throughputs of Russia - {(Trade Volume btw BSR & EU) + Russia(BSR) (Trade Volume btw BSR & Med.) + (Trade Volume btw BSR & Asia & N. Am. using OGDCTR)} Turkey(TK) Total Container Throughputs of Turkey - {(Trade Volume btw TK & EU) + (Trade Volume btw TK & Med.) + (Trade Volume btw TK & Asia & N. Am. using OGDCTR)} Cyprus(CY) Total Container Throughputs of Cyprus - {(Trade Volume btw CY & EU) + (Trade Volume btw CY & Med.) + (Trade Volume btw CY & Asia & N. Am.(incl. W. Africa) using OGDCTR)} TABLE 2 Estimated Container Throughputs of the New Container Port (TEUs) Countries A B C D A-B-C-D Greece 1,779,030 726,631 386,018 396,128 349,436 Bulgaria 110,085 46,502 48,726 NA 67,939 Romania 771,126 117,891 160,299 172,266 355,123 Ukraine 458,258 177,649 NA 35,577 252,174 Black Sea Area in 161,756 39,878 NA NA 121,878 Russia Turkey 3,170,357 1,800,791 NA 856,307 792,520 Total 6,450,612 2,909,342 595,043 1,460,278 1,939,070 Where: A (Total Container Throughputs), B (Total Volume with Europe), C (Trade Volume with North Africa (including Mediterranean)), D(Trade Volume with Asia & North America using OGDCTR), and E(Trade Volume with Asia & North America using OGDCTR) 315 Research in Transport and Logistics 2009 Proceedings Do u University Application of the Logit Model The second method involves the application of the Logit Model [11], which has been by far the most widely used in the field of transport mode and route choice studies. The formulation is developed in a utility context of a binary choice situation, in the form of (1): P(i) = expVi /( (1) where, x = a vector of attributes. With regards to the variables to be included in the model, Mangan [12] identified the following hierarchy of the key variables affecting the choice of mode/route: Cost/price/rate Speed Transit time reliability Characteristics of the goods Service Other key variables also assessed by Witlox [13] were reliability, flexibility and damage and loss. In transhipment, cargo moves through intermediate ports in its journey from origin to destination. These journeys are increasingly managed and designed to achieve the minimum point-to-point generalized transport cost, and not merely the minimum distance transport cost as before [14]. Hence, for the purpose of this study, it is assumed that the travel time and frequency of service, as well as reliability and quality of service will not differ significantly amongst the routes, and hence the decisive variable will be the travel cost. Also, since there is no possibility of introducing disaggregate data, the values for the variables introduced to the models are related to average values, derived from aggregate data Hence, (1) is employed to express the “probability of choosing the New Container Port”, as follows: P(New Container Port)= 1 / (1+ e U1) (2) where, U1= 1,73-0,018 C (3) The value of 1,73 represents the constant that reflects the relative attractiveness of the New Container Port. The value of 0.018 is a coefficient based on other similar studies [15]. Based on the above, travel costs are calculated for the routes and ports assumed in the first forecasting methodology. Since in this case, the critical value is the transhipment cost, an average value for the New Port is proposed, set at 84 US dollars per TEU [16]. Since the exact value is not known in advance, the forecasting methodology will be carried out for three distinct scenarios with respect to the transhipment cost of the proposed port, based on the transhipment values of 67$ (low), 84$ (medium) and 99$ (high). Based on the results of Table 2 in the previous section, and on the methodology described, Table 3 presents the predicted container throughput for the base year 2005, for the three scenarios of transhipment cost considered. 316 Research in Transport and Logistics 2009 Proceedings Do u University TABLE 3 Predicted Container Throughput to New Container Port 2005 (TEU) Countries Estimated Probability of choosing New Possible container throughput container Port to New Port for year 2005 throughput to New Port (max. Values) Transshipment $99 $84 $67 $99 $84 $67 cost at the New Port Greece 349436 0,2760 0,3328 0,4036 96441 116308 141034 Bulgaria 67939 0,2760 0,3328 0,4036 18750 22613 27420 Romania 355123 0,2760 0,3328 0,4036 98010 118201 143329 Ukraine 252174 0,2872 0,3454 0,4172 72430 87092 105213 Black Sea 121878 0,2872 0,3454 0,4172 35006 42092 50850 Area,Russia Turkey 792520 0,2872 0,3454 0,41722 227629 273707 330657 1939070 548267 660014 798503 Total In the case of the base year 2005, the predicted container throughput for the average transhipment cost of $84, is estimated at 660 014 TEU. It should be noted that this value is well below the upper limit that has been estimated and presented in the first column (1 939 070 TEU). Finally, in order to obtain values for any target year that the New Port would be operational, using linear extrapolation, an average growth factor could be applied to the predicted demand for 2005. In the absence of sufficient data, the model was neither calibrated nor verified. Nevertheless, Giannopoulos [17] used a similar, albeit more complex, Logit model in order to model the transhipment port choice situation in the same region, with the relative utility function defined as a composite measure of two utilities: one that relates to the distance between the origin and the destination port, and one that relates to the “level of service” (likely) to be offered by the destination port. The results indicated a future attraction of 650 000 containers for the same year, which is directly comparable to the container throughout predicted by the methodology proposed in this paper. CONCLUSIONS New trends in the economy, policy, and technological frameworks in the Mediterranean have lead to the development of new container shipping activities operating in the region, together with the evolution and growth of transhipment ports. In this context, this paper presented a simplified, novel and less data intensive methodology for forecasting future container traffic applied to a proposed New Container Port in the southern part of Greece. The methodology consisted of two approaches, the first being more general in nature, with the scope to obtain an upper limit on the future container throughput and to provide input for the second approach. In the latter, a port choice model was set up, based on the assumption that travel cost is the determining variable. Given its nature, the proposed methodology could be employed at the preliminary stage of planning for cases where the port in hand is not yet in operation, providing thus, a planning context for feasibility studies, and provide indicative information for stakeholders such as governments, shipping lines, port authorities, potential investors and global port operators. 317 Research in Transport and Logistics 2009 Proceedings Do u University REFERENCES [1] Baird, A.J., (2007), “The economics of Motorways of the Sea”, Maritime Policy Management, Vol. 34(2), pp 287- 310 [2] Dagenais, M.G. and Fernand M., (1987), “Forecasting Containerized Traffic for the Port of Montreal (1981-1995)”, Transport Research, Vol. 21A (1), pp 1-16 [3] Seabrooke, W., Eddie C.M. Hui, E.C.M. , Lam W.H.K., Wong G.K.C., (2003), “Forecasting cargo growth and regional role of the port of Hong Kong”, Cities, Vol. 20(1), pp. 51–64 [4] Lam, H.K.W., Ng, P.L.P, Seabrooke, W. and Hui, E.C.M., (2004), “Forecasts and Reliability Analysis of Port Cargo Throughput in Hong Kong”, Journal of Urban Planning, Vol. 130(3) , pp. 133-144 [5] Moon. S.H., (1995), “Port Economic Impact Model (PIM) and its Planning Applications”, Maritime Policy Management, Vol.22 (4), pp 363-387 [6] Tongzon, J., (1991),“A model for forecasting future supply of shipping services at Australian ports” Maritime Policy Management, Vol.18 (1), pp 55-68 [7] Zohil, J. and OHIL and Prijon, M., (1999), “The MED rule: the interdependence of container throughput and transhipment volumes in the Mediterranean ports”, Maritime Policy Management, Vol.26 (2), pp 175-193 [8] Baird, A.J., (2002), “The Economics of Container Transshipment in Northern Europe”, International Journal of Maritime Economics, Vol. 4(3), pp.249-280 [9] Containerisation International, (2005), Yearbook 2005, ISBN: 1859789854. [10] Eurostat, “Transport statistics - maritime”, www.eurostat.eu [11] Tsamboulas, D. Moraiti, P., (2007), “Estimating Freight to Assess the Viability of an International Intermodal Transportation Corridor”, CD-ROM Proceedings 86th Transportation Research Board Annual Meeting, Washington DC, U.S.A. [12] Mangan, J. Lalwani, C. and Gardner,B., (2002) “Modelling port/ferry choice in RoRo freight transportation”, International Journal of Transport Management no. 1 [13] Witlox, F. and Vandaele, E., (2005), “Determining the Monetary Value of Quality Attributes in Freight Transportation Using a Stated Preference Approach”, Transportation Planning and Technology, Vol.28 (2), pp 77-92 [14] Medda, F. and Carbonaro, G., (2007), “Growth of Container Seaborne Traffic in the Mediterranean Basin: Outlook and Policy Implications for Port Development”, Transport Reviews, Vol.27 (5), pp 573-587 [15] NEW.TON,, (2007), “Networking ports to promote intermodal transport and better access to hinterland”, Project Report, Work Package 2, Feasibility Study, INTERREG III, ARCHIMED Programme [16] Drewry Shipping Consultants Report, (2000), “Mediterranean container ports and shipping, Traffic growth versus terminal expansion - An Impossible Balancing Act?” [17] Giannopoulos, G., Aifadopoulou, G. and Torok, A., (2007), “A Port Choice Model for the Transhipment of Containers in Eastern Mediterranean”, CD-ROM Proceedings Transportation Research Board Annual Meeting, Washington DC, U.S.A. 318 Research in Transport and Logistics 2009 Proceedings Do u University SEA PORT HINTERLAND FLOWS AND OPENING HOURS: THE WAY FORWARD TO MAKE THEM MATCH BETTER Marjan BEELEN1, Hilde MEERSMAN2, Evy ONGHENA3, Eddy VAN DE VOORDE4 and Thierry VANELSLANDER5 Abstract Congestion at and around seaports has gained increasing media attention in the last years. This paper deals with the results of a research project on how to alleviate such congestion, and starts from the observation that the concentration of traffic is to a large extent due to the mismatch in opening hours at seaports and in the hinterland. The main research hypothesis states that a shift from peak hours to other moments during the hinterland working day, is the most viable option for shifting away traffic. Representative hinterland transport flows are selected upon statistical importance of flows and the type of goods. Three scenarios are considered: shifting to the late morning, the very early morning and the late working day. A shift to a moment within shipper opening hours is always the most feasible one. The largest benefit in all cases goes to the hinterland transport company. The largest cost usually goes to hinterland shippers. Keywords congestion, hinterland transport, opening hours, seaport, shift in time of day INTRODUCTION It is often said that seaports are the engine behind economic growth. Of all factors that determine the logistics position of a country or region, the capacity and efficiency of gateways, like seaports and airports, is probably the most important one. With increasing international trade, this statement is probably more true than ever before. All links and nodes within a chain should match as well as possible. Each bottleneck is a potential destructor, no matter how well the rest of the chain functions, especially to the extent that the bottleneck is located in a crucial node or link. Seaports are in the latter situation. To some part of the hinterland, they might be the only gateway. In that case, hinterland accessibility is at stake. In other cases, several seaports might serve a hinterland. In such situation, competition between seaports will come into play. Therefore, it is important to gain insight into the factors that impact on port efficiency and competitiveness. Past research, among others by [3] - [5], shows that a crucial competitive variable is potential time loss for ships, which can be incurred during the port entry or exit, or at berth. The latter may be caused by bad capacity management, lack of terminal capacity or congested hinterland connections. Most seaports worldwide have registered large growth figures, especially those that are active in the container business. Growth has occurred in tonnage as well as employment and value added. This growth was often referred to as evident, referring to the naturally advantageous location of the seaport and the available knowledge and training, which translate into high throughput respectively productivity. In the most recent years however, this evidence has disappeared. As far as containers are concerned for instance, the strongest threat is often no longer coming from neighboring seaports, but from ports in other port ranges, who have accumulated overcapacity. The competitive environment of seaports however has changed drastically over the last years, with competition getting fiercer among existing competitors, new competitors entering the market, and changing power of seaports at all in logistics chains. Ports have a hard time in keeping up with capacity expansion. Space needs to be found, and funds are needed for investing. But even when port space and funds are available, there is yet another problem. Other links in the chain need to be able to follow. Recent research by [6] – [7] indicates that especially the latter is an issue at many seaports, in Flanders like in a number of other European seaports. The ongoing and planned 1 Marjan Beelen, University of Antwerp, Faculty of Applied Economics, Department of Transport and Regional Economics, Antwerp, Belgium, marjan.beelen@ua.ac.be 2 Hilde Meersman, University of Antwerp, Faculty of Applied Economics, Department of Transport and Regional Economics, Antwerp, Belgium, hilde.meersman@ua.ac.be 3 Evy Onghena, University of Antwerp, Faculty of Applied Economics, Department of Transport and Regional Economics, Antwerp, Belgium, evy.onghena@ua.ac.be 4 Eddy Van de Voorde, University of Antwerp, Faculty of Applied Economics, Department of Transport and Regional Economics, Antwerp, Belgium, eddy.vandevoorde@ua.ac.be 5 Thierry Vanelslander, University of Antwerp, Faculty of Applied Economics, Department of Transport and Regional Economics, Antwerp, Belgium, thierry.vanelslander@ua.ac.be 319 Research in Transport and Logistics 2009 Proceedings Do u University expansion of maritime access and terminal capacity at many seaports will only worsen the hinterland situation, if no adequate measures are taken. This may imply that the new flows cannot be accommodated properly, but also that existing chains will experience worse service than before, impacting heavily on accessibility and competitiveness. Shipping lines will not undergo such deterioration of service, but will react by repositioning maritime loops, changing ports of call or call frequency. A core rule is that economies that materialize through scale increase and cost reduction on the maritime side, should not be nullified by time losses and cost increases on the hinterland side. Such shift will in turn make the port less attractive for shipping commodities, and will lead to a reduction of employment and value added generated at the port. Furthermore, also production and service companies in the port’s hinterland may relocate, again decreasing the level of employment and value added. A first way in which governments have reacted, is by trying to shift part of the traffic away from the road to other hinterland modes, if available at all. However, in most cases, those modes can only accommodate a small share of the traffic. A second answer was completing the hinterland network with a number of missing links. Congestion however is in most cases a peak hour phenomenon, putting high pressure on transport infrastructure and logistics systems at well-determined moments of the day, while leaving the infrastructure inefficiently used at other moments of the day. Moreover, adding new roads and/or missing links is expensive and will most likely not be a structural solution, as new traffic will be attracted that will fill up the extra capacity, in most cases even not only port-related. In that case, a third solution may be sought for, namely trying to make better overall use of infrastructure by shifting certain operations in time, to moments where overall traffic levels are lower. Shifting operations in time can be done in two ways: by extending opening hours at seaport terminals or in the hinterland, or by shifting the traffic to less congested moments during existing opening hours. Such shift can be enforced either by governments, like in Los Angeles [2], or by terminal operators, like in Felixstowe and Southampton [8].In order to select the best measure or optimal combination of measures, it is necessary to have insight in the exact identification and location of bottlenecks, and in the level of supplementary benefits and costs which either of the two options involves. This paper deals with the two issues for a selected set of cases in Flanders, in an attempt to find an answer to the main research question: whichever of the two is better, shifting within existing terminal opening hours, or shifting towards extended opening hours. The next section states the research hypothesis and the methodology used to test it. Section three deals with the context of hinterland flows and chain opening hours in Flanders. Section four elaborates on the flows that are selected for measurement and calculation, and the scenarios that are applied. Section five gives the quantified results of the calculations for the various cases and scenarios considered. The final section summarizes the main conclusions and provides a number of policy recommendations. RESEARCH HYPOTHESES AND METHODOLOGY As the paper revolves around one major research question, the main hypothesis is this one: shifting within existing terminal opening hours is more efficient than extending terminal opening hours. In order to test the hypothesis, first, a typology was made for classifying the different port-related hinterland flows. Second, statistical data were collected with respect to the importance of the various commodity groups and mode split of the Flemish ports, the geographical spread of port-related commodity flows around Antwerp, and a similar analysis based on the differentiation in time. Additionally, terminal-level truck calls were analyzed. The main input was a literature and data review on documents from the Flemish Community, the Flemish seaports, and terminal operators. Third, opening hours of the various actors involved in the landside of the logistics chains were assessed. Telephone and personal interviews were the main source of input. Fourth, based on the results from the previous steps, a selection of hinterland flows and shifting scenarios was made. Fifth, a model was developed, taking into account the results of previous measurement and calculation research. This model was used for calculating the supplementary costs incurred by all relevant actors in shifting traffic in time or in extending terminal opening hours, as well as the benefits that accrue from taking these measures. PORT HINTERLAND FLOWS AND CHAIN OPENING HOURS: A TYPOLOGY Traffic flows between the seaport and the hinterland can be summarized into the typology of figure 1, according to their type of origin, destination and transport mode. The typology applies to Flanders, but is representative for most countries. 320 Research in Transport and Logistics 2009 Proceedings Do u University Seaport 3. ROAD Hinterland Container 1. ROAD Distribution centre terminal 2. ROAD ROAD 5. ROAD / INLAND Bulk terminal Shipper / NAVIGATION / RAIL Production unit6. INLAND NAVIGATION / (ROAD) ROAD Industrial Inland container production unit terminal / ROC 4. ROAD 7. RAIL ROAD Inland rail 8. RAIL terminal Ro/ro terminal FIGURE. 1 Port Hinterland flow Typology. Following types of flows can be distinguished. Figures between brackets refer to the corresponding flows in figure 1. This typology is used as a structuring basis for the further research in this paper. From seaport container terminals to shipper or logistics service provider distribution centres in the hinterland and vice versa. An example of this kind of chain is the import of non-European products like computers, DVD players, etc. for large supermarket chains. (1) From seaport container terminals directly to hinterland shippers/production units and vice versa. An example is car components transported from seaports to car manufacturers outside the port. (2) From container terminals in seaports to industrial companies in the same seaport. An example is car components transported from terminals to car manufacturers in the same port. (3) From seaport ro/ro terminals directly to hinterland shippers/production units and vice versa. An example is paper pulp transported from a processing unit to foreign customers. (4) From seaport bulk terminals to hinterland shippers/production units and vice versa. An example is iron ore transported from a bulk terminal to a steel production plant. (5) From seaport container terminals to inland container terminals and vice versa. The bulk of these flows are handled by inland navigation. Only for urgent transport, road is used. Pre- and post-haulage between inland terminal and hinterland customers/shippers are done through road. An example is food shipped by using one of the inland terminals. (6) From seaport container terminals via rail to inland rail terminals and vice versa. An example is electronic components shipped through one of the rail terminals. Pre- and post-haulage between inland terminal and hinterland customers/shippers are done through road. (7) From seaport ro/ro terminals via rail to inland rail terminals and vice versa. An example is car traffic entering the country and using an inland rail terminal. With respect to chain opening hours, the disequilibrium between opening hours on the port side and the hinterland side is apparent from table 1. In all Flemish ports, it appears that the maritime side is open 24/24 and 7/7. The port landside in most cases shuts down at night, but the opening hours are still relatively long, especially in comparison to the hours that prevail in the hinterland. There, a distinction is to be made between small, medium-size and large shippers. Larger shippers are often open 24/24 as well. Medium-size shippers usually have opening hours that correspond to those at the port landside. Small shippers usually apply the regular office hours, and therefore are most limited. 321 Seaport Research in Transport and Logistics 2009 Proceedings Do u University TABLE 1 Comparison of opening hours in the chain (2007) Maritime Customs / Phyto-sanitary Land Truck Rail Barge Antwerp Containers 7/7, 24/24 8h – 12h/12.30h + 6h - 21.30h 6h - 22h 6h - 21.30h 12.30h/13h –16.30h Bulk 7/7, 24/24 8h – 12h/12.30h + 8h – 15.30h - 7/7, 24/24 12.30h/13h –16.30h Break bulk 7/7, 24/24 8h – 12h/12.30h + 8h – 15h 8h – 15h 8h – 15h 12.30h/13h –16.30h Zeebruges Containers 7/7, 24/24 6h – 22h / 8h – 12.30h + 6h – 21.30h - - 13h – 16.30h Ro/ro 7/7, 24/24 6h – 22h / 8h – 12.30h + 6h – 22h - - 13h – 16.30h Inland terminal 6h/8h – 6h/8h – 18h/22h 6h/8h – 18h/22h 18h/22h Shipper 24/24, 8h – 24/24, 8h – 17h 24/24, 8h – 17h 17h CASES SELECTED FOR MEASUREMENT AND CALCULATION Based on the importance of the commodity categories, the size of the freight flows and the mode split, this paper focuses on container terminals in the Port of Antwerp and on ro/ro and container terminals in the Port of Zeebruges. Of all modes, only pure road sections or combinations involving road among others are considered. Five specific cases are selected, involving actual flows by road between various types of companies and various terminals, and considering the eventual involvement of actors like shipping companies, customs, agents, phyto-sanitary services, storage companies and forwarders. A first case study concerns the export of containers by a medium-size shipper via the Port of Antwerp, as shown in figure 2. Red boxes indicate activities controlled by the shipping companies, green ones activities under the shipper’s control, and black boxes show activities controlled by the forwarder. On average 25 containers take this route. The shipping company is also in charge of road transport (carrier haulage), and outsources that to a road transport company. Containers on chassis can be picked on and off at the shipper’s premises 24/24 (from Monday 6am till Saturday 5pm). The shipper hás a proper container lift to pick on and off containers. It is manned from 5am till 7pm. Alternatively, the chassis can also be left and changed for a different one. AGENT CUSTOMS CUSTOMS SHIPPING TERMINAL ROAD TRANSPORT SHIPPER COMPANY OPERATOR COMPANY PHYTO-SANITARY PHYTO-SANITARY SERVICES SERVICES FIGURE. 2 Case 1: Container export through the Port of Antwerp. A second case study, shown in figure 3, deals with export of containers by a small shipper via Antwerp, with intermediation by a forwarder. One container per week is transported this way. Containers can be picked up each working day from 7am till 4pm, but a clear peak is noticeable between 7am and 8am. 322 Research in Transport and Logistics 2009 Proceedings Do u University AGENT CUSTOMS CUSTOMS SHIPPING TERMINAL ROAD TRANSPORT FORWARDER ROAD SHIPPER COMPANY OPERATOR COMPANY TRANSPORT COMPANY PHYTO- SANITARY PHYTO-SANITARY SERVICES SERVICES FIGURE. 3 Case 2: Container export through the Port of Antwerp via forwarder. A third case study treats an importer using the Port of Zeebruges, shown in figure 4. Upon arrival in Zeebruges, containers are railed to the Port of Antwerp. From Antwerp, the containers are shipped to the shipper’s warehouses by a fixed road transport operator. On a weekly basis, 115 containers are involved. From the distribution centres, the further dispatching to outlet stores is done by road transport again. The distribution centres operate in two shifts and are open from 6am till 10pm, and on Fridays till 8pm. AGENT FORWARDER CUSTOMS CUSTOMS SHIPPING TERMINAL OPERATOR RAIL OPERATOR RAIL TERMINAL ROAD DISTRIBUTION COMPANY TRANSPORT CENTRE COMPANY PHYTO- PHYTO-SANITARY SANITARY SERVICES SERVICES FIGURE. 4 Case 3: Container import through the Port of Zeebruges and via rail through the Port of Antwerp. Fourth, import containers via Zeebruges are modeled, as in figure 5. 25 containers a week are involved. Containers can be delivered at the shipper’s premises in two ways: by leaving the chassis, or by picking off. The latter is only possible between 5am and 9pm. AGENT CUSTOMS CUSTOMS SHIPPING TERMINAL ROAD TRANSPORT SHIPPER COMPANY OPERATOR COMPANY PHYTO-SANITARY PHYTO-SANITARY SERVICES SERVICES FIGURE. 5 Case 4: Container import through the Port of Zeebruges. The fifth and final case concerns import of ro/ro through the Port of Zeebruges, represented in figure 6. Loading and unloading on the terminal are possible 24/24. Delivery at the shipper is theoretically possible between 8am and 6pm, but in practice there are limitations. 323 Research in Transport and Logistics 2009 Proceedings Do u University AGENT CUSTOMS CUSTOMS SHIPPING TERMINAL ROAD TRANSPORT SHIPPER COMPANY OPERATOR COMPANY PHYTO-SANITARY PHYTO-SANITARY SERVICES SERVICES FIGURE. 6 Case 5: Ro/ro import through the Port of Zeebruges. Per case, different scenarios are developed and prepared for measurement in the next section. The scenarios depend on the moment to which the shift is made. A crucial variable is shipper size: for large shippers, there is no difference between night and day deliveries, as their services are open anyway. Small shippers on the contrary have strongly limited opening hours. The timing of the shift also impacts on the maritime side of port terminals: due to limitations in opening hours, supplementary costs need to be made for night deliveries. DIRECT COSTS AND BENEFITS OF SHIFTS IN TIME For calculating costs and benefits for the various cases and scenarios described higher, following assumptions are made. The standard split-up of transport costs between time and kilometre costs, as described by [1] is used. External effects are not taken into account. The benefits of gaining time in transport equal the costs that are avoided. Not taken into account are the extra trips that can be made. Working hours are considered as night labour by applying the Belgian law: if for five consecutive nights, more than five hours are worked 8pm and 6am. If these conditions are fulfilled, a surplus wage applies. The period between 7am and 9am is considered the morning traffic peak, whereas the period between 4pm and 7pm is considered the evening peak. Terminal dwell times, waiting times at the gates and travel times are average values, not taking into account exceptional events. For night operations, an extra full shift is assumed. Due to the special regulation for customs night tariffs, no surplus value is assumed. For case one, the starting situation is shown in table 2. TABLE 2 Case one starting situation. Action Duration Tijdstip Transport company arrives at shipper and puts off container 7.00am Full containers is picked up 30min 7.30am Transport company drives from shipper to container terminal in 50min 8.20am Antwerp, during morning peak Processing of the container at the terminal (including gate waiting 90min 9.50am times) Three alternative scenarios are considered, with two options each time: using a container elevator, or picking up and off entire chassis. An early morning (3am) start A late morning (10am) start 324 Research in Transport and Logistics 2009 Proceedings Do u University A evening (7pm) start In the first alternative, as compared to the starting situation, all actors incur supplementary costs, whereas only the transport company incurs a relatively small benefit. In the second alternative scenario, no extra costs are to be made, but the transport company incurs a small disbenefit due to longer gate waiting times. In alternative scenario three, the shipper incurs a surplus cost, whereas the transport company benefits from a small time gain. Therefore, in case 1, none of the alternatives means a real improvement to the current situation. The general results for any possible alternative are summarized in table 3, per time block. Green boxes refer to no surplus cost, red ones to a relatively high surplus cost, and brown ones to a very high surplus cost. It turns out that especially the nights present prohibitive surplus costs to most actors. TABLE 3 Case one surplus costs compared to starting situation. scenario with container elevator shipper road transport operator terminal customs scenario picking up and off entire chassis shipper road transport operator terminal customs For case two, the starting situation is the same, with that difference that travel times towards the port may are different because of a different starting location and distance towards the port. Only the early morning and late morning scenarios are now considered as alternatives. The evening scenario would be comparable to the late morning scenario, because the shipper is a big company, with extended opening hours. In the late morning scenario, the transport company earns a lower benefit than in the early morning scenario, but, at the same time, the surplus costs to all actors together are lower. Therefore, alternative two would be preferable in this case. In case three, like in case two, only two alternative scenarios are considered, for the same reason. For these two alternatives, the benefit is equal and accrues entirely to the transport company. The surplus cost however is a lot higher under scenario one, and affects all actors. Therefore, just like in case two, scenario two is the better one. Exactly the same conclusion as in case three applies to cases four and five, although the benefit is higher in case four, due to the time gain obtained at the terminal. The overall results in terms of surplus costs for the different cases are summarized in table 4. Minor differences are observed between the cases, so that the overall lesson is that the heaviest surplus cost, to all parties concerned, is found for trips between 10pm and 6am, with some deviations according to the specific case considered. CONCLUSIONS AND POLICY RECOMMENDATIONS The analysis in this paper allows narrowing down congestion problems in port-bound hinterland traffic to a limited time, geography and commodity type scope, and allows drawing a number of conclusions on the causes of the problem and the obstacles that prevent solutions from materializing. The time scope is clear: problems are strongest during morning and evening peak, and the calculations indicate that of the tested scenarios, shifting to other moments in the day overall gives the best result. Geography is important however, as the calculations show that not every trip origin results in the same benefits and surplus costs. Only containers were considered. It is probable that other commodity types might benefit from similar initiatives, although there again, benefits and surplus costs may be different. 325 00:00 00:00 01:00 01:00 02:00 02:00 03:00 03:00 04:00 04:00 05:00 05:00 06:00 06:00 07:00 07:00 08:00 08:00 09:00 09:00 10:00 10:00 11:00 11:00 12:00 12:00 13:00 13:00 14:00 14:00 15:00 15:00 16:00 16:00 17:00 17:00 18:00 18:00 19:00 19:00 20:00 20:00 21:00 21:00 22:00 22:00 23:00 23:00 Research in Transport and Logistics 2009 Proceedings Do u University Although the analysis focuses on Flanders, the conclusions are generalizable to a large extent to other countries with seaports, most of which experience similar problems. This type of initiative therefore certainly opens opportunities, and could be considered by any government of countries where similar inefficiencies occur. More in-depth research is however needed, into the best alternative trip moment that could be suggested. Subsequently, a way of convincing actors of the benefits they obtain has to be found. A calculation tool like the one developed here could be helpful. TABLE 4 Case two-five surplus costs compared to starting situation. case two verlader hinterlandvervoerder terminal douane case three verlader hinterlandvervoerder terminal douane f verlader hinterlandvervoerder terminal douane verlader hinterlandvervoerder terminal douane [1] Blauwens, G., De Baere, P. and Van de Voorde, E., 2006, "Transport Economics (2nd ed.)", De Boeck, Antwerp [2] Giuliano, G., O’Brien, T. and Maggadino, J., 2005, "Evaluation of Terminal Gate Appointment System at Los Angeles and Long Beach Ports. ", Final Report Project 04-06, METRANS Transportation Center [3] Heaver T., Meersman, H. and Van de Voorde, E., 2005, "Co-operation and competition in international container transport: strategies for ports", in Leggate, H., McConville, J. and Morvillo, A. (ed.), International Maritime Transport: Perspectives, Routledge, London, 145-159 [4] Janssens, S., Meersman, H. and Van de Voorde, E., 2003, "Port throughput and international trade: have port authorities any degrees of freedom left?", in Loyen, R., Buyst, E. and Devos, G. (eds), Struggling for Leadership: Antwerp-Rotterdam Port Competition between 1870-2000, Physica-Verlag (a Springer-Verlag Company), Heidelberg - New York, 91-113 [5] Meersman, H. and Van de Voorde, E., 2002, "Port management, operation and competition. A focus on north European continent", in Grammenos, C.T. (ed.), The Handbook of Maritime Economics and Business, Lloyd's of London Press, London/Hong Kong, 765-781 [6] Meersman, H., Monteiro, F., Pauwels, T., Van de Voorde, E. and Vanelslander, T., 2006, Social Marginal Cost Calculation for Ports, Report for the GRACE Project - WP 2.1 & 2.3, European Commission [7] Meersman, Van de Voorde, E. and Vanelslander, T., 2007, "Port congestion problems: some evidence from european and US ports", Proceedings of the First Intermodal Conference, University of the Aegean [8] Port of Felixstowe, 2008, Vehicle Booking System, http://www.portoffelixstowe.co.uk/vbs/ 326 0:00 0:00 00:00 00:00 1:00 1:00 01:00 01:00 2:00 2:00 02:00 02:00 3:00 3:00 03:00 03:00 4:00 4:00 04:00 04:00 5:00 5:00 05:00 05:00 6:00 6:00 06:00 06:00 7:00 7:00 07:00 07:00 8:00 8:00 08:00 08:00 9:00 9:00 09:00 09:00 10:00 10:00 10:00 10:00 11:00 11:00 11:00 11:00 12:00 12:00 12:00 12:00 13:00 13:00 13:00 13:00 14:00 14:00 14:00 14:00 15:00 15:00 15:00 15:00 16:00 16:00 16:00 16:00 17:00 17:00 17:00 17:00 18:00 18:00 18:00 18:00 19:00 19:00 19:00 19:00 20:00 20:00 20:00 20:00 21:00 21:00 21:00 21:00 22:00 22:00 22:00 22:00 23:00 23:00 23:00 23:00 Research in Transport and Logistics 2009 Proceedings Do u University INTERNATIONAL ROAD FREIGHT TRANSPORT IN GERMANY AND THE NETHERLANDS DRIVER COSTS ANALYSIS AND FRENCH PERSPECTIVES Laurent GUIHERY1 Abstract -: These last few years French international road freight transport has been undergoing a loss of influence within Europe while traffic has increased and great manoeuvres are taking place since the opening of the European Union towards East. Some of the French transporters are then focusing back on the French market showing a worrying loss in competitiveness. On the contrary, German and Dutch companies are increasing their shares in the French market and have reorganized themselves within Europe to face eastern Europe competition: follow-up on customers delocalizing in the East, networking, hyperproductivity, markets segmentation between high quality transport in the West, specific markets and low cost segment in eastern Germany or/and Poland, intensive geographical closeness to a great harbour (Rotterdam)… What should we learn from German and Dutch experiences to be used towards a renewal of French road transportation in the international field? On the basis of a comparison of our neighbours’ driving costs and road freight transport structure, our contribution - a synthesis of two recent studies ordered by the Comité National Routier (CNR, studies free to be downloaded by www.cnr.fr) - will first propose a cooperation with German or Dutch companies in order to propose a winner-winner model based on exchange of competencies : North Africa (Morocco for instance) and Southern Europe for French partners (specialization Storage - Logistics) and transport business model and opening towards the East for the German and Dutch partners. An interesting proposition would be to get close to Moroccan operators in order to rebalance the international freight road transport in Europe between a dominant Centre-East and a developing Centre-Mediterranean. Keywords International Road Freight Transport, Germany, the Netherlands, Driver Costs The year 2007 and 2008 have been in France a full one in terms of studies, debates and propositions, in order to understand, accompany and relaunch international road freight transport in France, in particular as far as its engagement towards international markets is concerned, one that has been undergoing a downfall these last few years. Within the framework of the actions undertaken by the Conseil d’Analyse Stratégique (www.strategie.gouv.fr), many reports have been published (9 reports at all that is to say 875 pages !) : they offer a full analysis of road freight transport in France but also internationally. The synthesis presented by Claude Abraham and his team – “For an ongoing regulation of road freight transport2“ – and the reports of the different working groups3 have been going through a wide consultation of all actors and show the need to compare the French situation with that of our European partners4 in order to learn a lesson for the French transporters. Within the framework of market intelligence missions and prospective analysis of the sector led by the Comité National Routier (CNR) some investigations have been undertaken in 2007-2008 on road freight transport and driving costs in Germany and Holland by the author, directed by Alexis Giret (CNR). This article proposes a synthesis of those two studies. METHODOLOGY AND FRENCH SITUATION The study’s methodology relies on a series of meetings/discussion led among road freight transport professionals (around 10 companies by country in more than 3 different regions for every country) but also 1 Laurent Guihéry, Maître de Conférences / Associate Prof. in Economics, Laboratoire d’Economie des Transports (LET-ISH), Université Lumière Lyon 2, 14 avenue Berthelot, 69363 LYON CEDEX 07, Téléphone : 04-72-72-64-03, Télécopie : 04-72-72-64-48, Laurent.guihery@let.ish-lyon.cnrs.fr 2 http://www.strategie.gouv.fr/article.php3?id_article=838 3 « Développement, compétitivité, et emploi » ; chairman : M. Maurice Bernadet « L’acceptabilité sociale des poids lourds » ; chairman : M. Jean-Noël Chapulut « Les relations et les évolutions sociales ; chairman : M. Georges Dobias « Transport routier de marchandises et gaz à effet de serre » ; chairman : M. Michel Savy 4 Germany, as France does, has a good knowledge of the sector through the mission of market analysis of the B.A.G. (Bundesamt für Güterverkehr, Cologne). As expressed in a meeting by the Dutch Ministry of Transport, The Netherlands have a strong liberal culture and then few interests of markets analysis of privates companies, for example in the transport sector. Some statistics of the sector for example are managed by NEA, a private company. TLN (Transport Logistic Nethelrands, the professional association of Dutch Road Haulage) is a key actor too, for statistics for instance. 327 Research in Transport and Logistics 2009 Proceedings Do u University among professional organizations, trade unions, the university community and public institutions (Ministry of Transport). Statistics have been obtained among appropriate organizations, public ones in Germany (BAG - Federal Office for Freight Transport, or professionals road haulage associations (BGL, DLSV in Germany and TLN in Holland) or private/semi private ones in Holland working for the Dutch Transport Cabinet (NEA, NIWO). Costs assessments are a synthesis of crossed-information (BGL information), average values and estimates for Germany and panel results (NEA information) for Holland. The starting point of the interest of professionals and French public institutions in the study of international road freight transport in Germany and Holland originates in the worrying observation in France of a decrease of French companies’ market share in international road freight transport. In 2007, the French international road freight transport has decreased by 3,2 % in t.km (27,2 billion t.km) following a stabilisation noticed in 2006. In tonnes, the decline is 1,4 % by 2007/2006. Cabotage of French companies is also declining by around 24 % in 2007 ! This lack of competiveness is part of a long trend : while, between 2000 and 2004, international exchanges between France and its 15 European Union’s partners have increased by 17%, French transport companies has undergone a drop by 17% [40]. This drop can be explained partly by the decreasing share of French transporters in French international exchanges. French transporters have lost 3 points in market shares every year, dropping from 35,9% in 2000 to 25,5% in 2004. The increase in imports, for which French transporters have found it difficult to position themselves, is acting against French transporters. Despite a moderate increase in exchanges with Germany (+9%) which explains more than a third of the decrease of French transporters [40], French transporters’ market share dropped by 12 points and German transporters are reinforcing their position as well as third parties (among which Dutch transporters) who have seen their activities progress by 50% since 2000 [40]. This situation is to be seen on the field specifically in Germany and can be partly explained with the introduction in 2005 in Germany of a pricing policy for the use of motorway infrastructure which has induced an increase in transport costs for French transporters and more particularly when return is an “empty” one. This factor has thus been well identified by the German authorities of the sector in its annual report : since the introduction of toll on motorways (2005), the annual report 2005 of the Federal Freight Road Transport Office [2, p. 13] is expressing that French companies are less present in Germany which gives German transporters more development opportunities5. The screening of German and Dutch long distance freight road transport markets is very interesting in terms of their engagement at the heart of Europe in a hard competition with the European Union’s New Member States. It is thus interesting to see how in both countries a whole sector, well organized and structured (in terms of legislation and socially), has reacted to this external impact. Commercially speaking these newcomers show a hyperdynamism, as we will see. Focused on its internal market, the road freight transport in France seemed protected from “big manoeuvres” developing today in the East and South (Maghreb) but this fact is no longer true : French road freight transport is now confronted with a double compression from both East (Germany, The Netherlands, East Europe) and South (Spain) on the international market but also as far as national transportation is concerned when limits on “cabotage” will be abandoned. Moreover, this compression is far from being a static one and thus, it is in the framework of both an eastern and southern compression that the French freight road transport should be studied. This situation could legitimately lead us to question the existence of French long distance road freight transport in the long term. Which strategic direction should we impulse to take advantage of the Europeanization of road freight transport on the European continent? - Network strategies like suggested by German middle-sized companies - Concentration and search for a critical mass around big road/rail/sea intermodal groups - Withdrawal towards market niches - Specialization in more profitable high quality and standards transports 5 « Eine positive Entwicklung der Beförderungsentgelte sahen deutsche Transportunternehmen im ausgehenden Verkehr nach Frankreich, da es auf dieser Relation seit der Mauteinführung in Deutschland zu einer Laderaumverknappung gekommen war. Französische Transportunternehmen engagierten sich seitdem in geringerem Maß im Verkehr mit Deutschland. Problematisch waren für in diesem Bereich tätige deutsche Unternehmen die deutlich niedrigeren Beförderungsentgelte für Rückladungen aus Frankreich und die Tatsache, dass französische Auftraggeber die Zahlung der deutschen Maut in vielen Fällen grundsätzlich ablehnen. » [2] 328 Research in Transport and Logistics 2009 Proceedings Do u University - Closeness to a requalified and modern harbour (Holland) or what we recommend: first a merging with Dutch and German companies in a European winner-winner perspective, taking into account the necessity of protecting competition, and, in a second best solution, a quick setting up of subsidiaries in Morocco or in Southern Spain in order to segment one’s global transport supply (low cost/niches/high-rank) and thus rebalance the Center-East stream with a South-East powers… A Centre- Mediterranean answer to a freight road transport critical mass moving towards the Centre-East. AT THE CENTRE OF EUROPE: A GERMAN ECONOMY FULLY EXPORTING AND A HARBOUR FOR EUROPE, ROTTERDAM Leaving aside the current crisis linked with petrol prices (july 2008) and financial market (september-october 2008), the road freight transport sector is not feeling to bad at the centre of Europe and enables Holland and Germany to position themselves as leaders in this field (table 1) for three fundamental reasons : a real dynamism of the German economy towards exports these last few months, the large outsourcing of subsidies and plants in East Europe linked with the new process of “Bazar Economy” described by H.-W. Sinn which imply a lot of transport between all the plants and subsidiaries all over Europe and the world, and a very profitable industrial woldwide positioning (success of a the famous structure of middle size companies likely to exports widely) and, for Holland, predominance of a great harbour for European inward and outward flows, this harbour being very closely connected with the German industrial structure. In the field of freight transport, the completion in 2006 of the freight railroads “Betuwelinie” between Rotterdam and Germany is considered essential. TABLE 1 : ROAD FREIGHT TRANSPORT IN FRANCE, GERMANY AND THE NETHERLANDS IN 2004/2005 France Germany The Netherlands in… (2005) (2004) (2005) T.km road freight transport 205 384 84 Billion t.km on national territory - from national companies 177 267 32 Billion t.km National companies in 28 71 52 Billion t.km international t.km T.km freight rail transport 41 86 5 Billion t.km T.km watervay 9 64 42 Billion t.km Number of companies 36 000 48 500 12 000 Trucks 185 500 345 500 77 500 (more than 3.5 t.) Source : Alexis Giret and Laurent Guihéry, Synthesis Study CNR Europe, march 2007 For Germany, road freight transport is leading in Europe (table 1) which is easy to understand with the very central location of Germany in Europe between West and East. The reason for this are multiple but the main factor is the dynamic structure of middle size companies (“Mittelstandunternehmen”) typical of German industrial and service oriented sector. These companies employ 72 % of the 22 millions employees and are making 50 % of investment. German road freight transport is twice the French sector (see table 1) and is growing quite well the last years (lack of drivers for the shipping of presents for Christmas 2006) like in Holland. Transport for third parties is important but some companies are thinking of “re-nationalizing” transport operation (own account) to keep 100 % reliability, punctuality and full efficiency of transport operations, which is in Germany presented as a label : “transport made in Germany”. This label is very 329 Research in Transport and Logistics 2009 Proceedings Do u University important for shippers working with just-in-time model of production. Sales of below 12 tons vehicles are booming because the new motorway toll implemented in 2005 do not take these vehicles into consideration. German transport companies are organized in both way: a high proportion of family and small companies organized sometimes outside the system of collective labour, middle size companies organized on a European network, and large operators like Schenker. Like in the Netherlands, the structures of companies in Germany is based on a dual system following the model of “insider”(protected / collective labour agreement) / “outsider model” (small companies, high flexibility and reactivity, lack of job security). Costs in Germany are higher than those corresponding in France but this is changing rapidly : trucks are 10 % more expensive than in France, gasoline more expensive (+6,6 %, data CNR), indirect taxes (“droits d’accises”) higher up to 20 %, higher cost of insurance, toll pricing on motorway of around 0,20 cents euro / km since the 1 January 2005 (increased in January 2009) but this new costs was transferred on the shippers. Truck involvement is around 130 000 km/year more than 120 000 km reported in France (CNR). Wages are ruled by a labor agreement negotiated at the Länder level which implies disparities between East and West Länder. It can be noticed that “low cost” transport companies are existing in East Germany facing directly East European competition. East German drivers can then be paid 30 % less than in the West when labour agreements are applied which is not the case every time. The daily rate in Thüringen is 6,14 € one hour without length of service. West German transport operators are then mainly focused on “high quality transport” with high profitability and are likely to outsource transport operations in their East European or East German subsidiaries. Comparing productivities and wages of drivers, there are few differences between Germany and France … but a German driver seems to drive 22 % more than a French one by using a more “company friendly way of driving” (few time of disposal, switching mainly between time of rest and time of driving). Following strictly European regulations and working for a flat rate imply a maximal management of driving time which is one of the lessons to draw of the competitive advantages of German and Dutch drivers. Concerning Netherlands, road freight transport is growing rapidly in NL with 7% growth each year in average. The Dutch Ministry of Transport is expecting a growth of 20 % - 30 % by 2020. 12 000 companies are involved with 80 % specialized on international transport (70 % transport for others). The Dutch are very specialized in international transport with more t.km in international transport than in national transport ! The international road freight transport in Holland is then twice the French one : Dutch operators carry 57 % of their exports like France for 15 years but Dutch operators are also leading in imports (52 % !), which is surprising ! More than Germany, the Netherlands are THE reference in terms of hyper-productivity in road freight transport : linked with the framework of the “polder model” (liberal way of life, entrepreneurs spirit, consensus, negotiation, compromises, flexi-security social system, dual system following the model of “insider”(protected / collective labour agreement) / “outsider model” (small companies, high flexibility and reactivity, lack of job security)) and far away from public interventionism, the labor relations in Dutch road freight transport are managed by a system of collective and autonomous labour agreement (see TLN and the 112 pages of their very precise labour convention 2007). The daily rate is 9€84 while beginning to work till 12€54 after six years of work in the same company (value October 2006) : 40 hours a week ; +30 % if overtime ; + 50 % if worked on Saturday ; +100 % if work on Sunday. Operating costs are higher : buying a truck seems o be a little bit more expensive than in France, insurance are 50 % higher, gasoline and maintaining / repairing are around 5 % less expensive. As expressed in the table 2, driving cost are widely higher and are one of the higher in Europe and this positive point for employers are compatible with a leading position of Dutch operators in Europe . To sum up, it seems that the Netherlands have set up a winner – winner model : - Wages (50 % more than a French driver in average) - Competitiveness (despite an hourly cost 8 % higher than a French driver specialized in international road freight transport). - Large turnover and jobs available: lack of drivers is a reality. The success of the Dutch transport operator in international operation is then based on a high volume of work (trucks are then driving till 150 000 or 180 000 km/year and a high flexibility of the labour organisation of drivers, which lead to a competitive advantages of Dutch operators on the European market. Drivers are, in Germany and in The Netherlands, maybe more implicated in the success of the company by choosing “driving 330 Research in Transport and Logistics 2009 Proceedings Do u University time” position or “rest time” position more than “disponibility time” on the tachograph. The yearly working time is around 213 days like in France. But differences are obvious in the organisation of companies looking for productivity gains, management of tours and efficient allocation of drivers : normal day of work in this sector can reach 11 hours a day, weekly working time near 55 hours (reaching 60 hours in certain cases),... On this point, due to the central position of Rotterdam Harbour in European Transport facilities, the European Commission has given an additional delay since march 2010 (5 years more) to transfer the European Guideline 2002/15 (limit of working time of freight road driver to 48 hours a week in average). HYPER PRODUCTIVITY OF DUTCH AND GERMAN COMPANIES In both Germany and The Netherlands, transport operators are hyper productive and put gains of productivity at the centre of their business management. For middle size companies, they develop networks of companies for achieving a better critical weight and a better visibility. They get more concentred too. Table 2 is expressing this hyper productivity of German and Dutch operators by showing a compared analysis of driver costs in four European countries. Methodology of the study was introduced in introduction of this paper. For Poland and France, data are coming from the CNR. East Germany can appear as a “low cost region” in Western Europe facing directly East European competitors but are full part of the famous German label of “Transport made in Germany”. Western German companies are then outsourcing easily unprofitable or complicated transport operations to their East German subsidiaries for the benefit of both shippers and transport companies, both in West (keep competitiveness) and in East (having some work in a difficult economic framework). TABLE 2 : MAIN RESULTS OF THE STUDY ON DRIVING COSTS IN GERMANY AND THE NETHERLANDS, IN COMPARISON WITH FRANCE AND POLAND (2006) Estimations Unit France : Germany : Germany : Germany : Netherlands Poland study CNR West West East (Source : NEA data (estimation 2006 Länder Länder Länder international to CNR ; rapid (average) (maximum) (minimum) France) and own growth) adjustments ; Wages (with € / 2173 2734 2967 1718 3223 from 820 to 1,360 overtime and bonus) month Employers charges % 36 (Fillon 25 25 25 36 22.57 support deducted) Travelling expenses €/day 38-40 20 20 20 40 international from 25 to 40 in average by day Average with -7 national 50 % international Total yearly cost € / year 44 173 45 463 48 960 30 425 55 132 20 000 Weekly working Hours / 49,6 56,5 56,5 56,5 55 56,5 time week Number of working Week / 42 42 42 44 43 45 to 46 weeks by year year Yearly working time Hours / 2100 2373 2373 2486 2343 2500 year Yearly driving time Hours / 1554 1890 1890 1980 1917 2015 year Cost of one hour of € / hour 21 19,2 20,6 12,2 23,5 8 work Cost of one driving € / hour 28,4 24,1 25,9 15,4 28,8 10 hour Base 100 France on 100 85 91 54 101 35 the driving hour Source : Studies CNR Europe, data CNR, own calculations and cross comparisons, firms and univ. interviews 331 Research in Transport and Logistics 2009 Proceedings Do u University WHY LOOKING TO ROAD FREIGHT TRANSPORT IN GERMANY AND THE NETHERLANDS: WHICH RECOMMENDATIONS? Germany and The Netherlands are leading in international road freight transport in Europe today and are facing a strong competition from East European countries. They are pioneers in reaction and strategies developed facing new member’s states and French operators can then learn from them. Objectives are, like the history of the European Union Integration process, to balance benefits and risks among European nations in a long term process of mutual convergence. If we consider that the competency of international road freight transport is a key function of each nation states and then cannot disappear in France (no total specialization on the European level in the transport sector is likely to emerge - this assumption has to be investigated -, then we can propose some recommendations for French transport companies and public authorities. Road freight transport in Germany and Holland is dynamic and successful since the joining of the New Member States in the European Union (opportunities), even if the competition is becoming stronger and profitability low (or negative like in Holland in average). Companies in both countries are facing a lack of drivers. “Cabotage” is, in a short future, challenging the future of French road freight transport, especially with Romania and Bulgaria. Ukraine, Turkey and Russia are also in a middle term perspective also formidable partner. Facing this new European environment, German and Dutch companies have developed interesting strategies: - Following of industrial companies in the outsourcing of production process in East Europe (especially Germany) ; Transport and logistic subsidiaries set up in East Germany (driving costs between 20 and 30 % less expensive) or in East Europe with merging with local operators (driving cost inferior to 40 % following a study of the BGL and the German Transport Ministry). - Productivity gains if possible - In Germany: decrease of average personal costs, maybe driving costs but difficult to show evidence on this point. Germany is the only country in Europe that has experienced an average decrease in wages the last 5 years (decrease of real wages of 0,8 % the last 8 years (2000-2008, Source : foundation Hans Böckler, Les Echos, 18.09.2008 ). More and more workers (maybe drivers?) in Germany are not integrated in collective labour conventions, which have an impact on competitiveness. This trend is considered as “competitive disinflation” which has an impact on restoring cost-competitiveness of Germany the last years. This model is less transferable to France for socio-historical raisons (strong trade unions, lack of confidence between social partners,…). - Networking of medium size operators to increase the critical mass and get more visibility; strategy to be developed with French operators; outsourcing of non profitable transport to small operators or East European partners, like in Germany. - Better connection with ports (example of Rotterdam and European Distribution Center) ; balance development with other European ports to be investigated, especially with congestion issues in Rotterdam Harbour. - Specialisation in high quality transport as the German operators do : reliability, punctuality, services (logistic, packaging,…), know how ; “niche” market ; label quality transport (like the famous “transport made in Germany”). Concerning the French transport operators, the development of European International Road Freight Transport implies a rapid and strategic reaction : if we consider and accept the process of European integration process as a complicated balance and trade-off between nation states in a winner – winner game – this was the case for the last 50 years of European integration, we cannot accept both in Germany and in The Netherlands the disappearance of the French international road freight transport. Solutions has to be found in the merging of French operators with German and Dutch operators in a winner – winner model : giving access to south European or north African markets for central and north European operators and increase of critical weight for accessing East European markets for both French and Dutch / German operators. In a second best solution, if merging is unlikely to appear, French operators would have interests to “move South”, by setting up 332 Research in Transport and Logistics 2009 Proceedings Do u University subsidiaries in North Africa (Morocco). If this solution is likely to boost, on a short term perspective, the competitiveness of French international road freight companies, the impact of such strategies on the current European integration process (political integration) is difficult to assess. BIBLIOGRAPHY: [1] Bundesamt für Güterverkehr (BAG), Geschäftsbericht 2005, Cologne [2] Bundesamt für Güterverkehr (BAG), Markbeobachtung Güterverkehr : Sonderbericht zum Strukturwandel im Güterkraftverkehrsgewerbe, 2005 [3] Bundesamt für Güterverkehr (BAG), Struktur der Unternehmen des gewerblichen Güterkraftverkehrs und des Werkverkehrs“, November 2004 [4] Bundesverband Güterkraftverkehr Logistik und Entsorgung e.V. (BGL), Jahresbericht, October 2005 [5] CNR, base de données des TRM européens du SESP, 2006 [6] Deutscher Speditions- und Logistikverband e.V. (DSLV), Lohntarifvertrag 2006 für die gewerblichen Arbeitnehmer des privaten Transport- und Verkehrsgewerbes in Hessen (“Tarifs”), 2006 [7] Guihéry L, « Le transport routier de marchandises en Allemagne », l'Officiel des Transporteurs, Actualités internationales, nr. 2401, April 2007 [8] Guihéry L, « Transport routier de marchandises et coûts de personnel de conduite en Allemagne », in Georges Dobias, Patrice Dupuy, Christine Raynard, "Pour une régulation durable du transport routier de marchandises « Les relations et les évolutions sociales », Conseil d'Analyse Stratégique, p. 85-88, Paris, april 2008 [9] Guihéry L., « Transport routier de marchandises et coûts de personnel de conduite aux Pays-Bas », Etudes CNR Europe, Comité National Routier, Paris, 2008, 46 pages. ; Synthesis : "Le transport routier de marchandises aux Pays- Bas", in Georges Dobias, Patrice Dupuy, Christine Raynard, "Pour une régulation durable du transport routier de marchandises « Les relations et les évolutions sociales », Conseil d'Analyse Stratégique, p. 89-94, Paris, april 2008 [10] Herry M., « Transportpreise und Transportkosten der verschiedenen Verkehrsträger im Güterverkehr. 2001 [11]Information zur Tarifpolitik, WSI-Tarifarchiv, april 2006, Nr. 61 [12] Klaus P. , “Go East-Logistik”, Transparents de la Conférence, September 2004 [13] Klaus P., Kille C., « Die Top 100 des Logistik 2006 », 4ème Edition, Deutscher Verkehr Verlag Hamburg, 2006 [14] Lafontaine F., Malaguzzi Valeri L., « The Deregulation of International Trucking in the European Union : Form and Effect, working paper, April 2005 [15] MINEFI, « Fiche de synthèse Pays-Bas », Mission économique aux Pays-Bas, July 2006 [16] MINEFI, « Fiche de synthèse Pays-Bas », Mission économique aux Pays-Bas, March 2007 [17] Ministère de l’Equipement du Transport et du Logement, “Réglementation sociale européenne dans les transports routiers, DTT, 20005 [18] Ministère de l’Equipement, du Transport et du Logement, SESP, « La transport routier de marchandises en Europe en 2004 : forte croissance du pavillon espagnol » 2006 [19] Ministry of Transport Netherlands, “Freight Transport in NL”, 2004 and 2007. [20] NEA Transport Research and Training, Cost comparison and costs developments in the European Road Haulage Sector, 2005 [21] NEA, “Cost Comparison and Cost Developments in the European Road Haulage Sector”, 2006 [22] NEA, “The main Features 2005-2015 : a strategic vision on European Transport Flows”, August 2005 [23] NEA, many studies on transport sectors and costs, 2006 [24]NIWO, Jaarverslag 2006 [25] OCDE, Note de synthèse Pays-Bas, ECO/CPE(2007)7/EO81/2, mai 2007 [26] Prognos, « Regulations in the transport market and personnel costs of driving staff in Germany, January 2003 333 Research in Transport and Logistics 2009 Proceedings Do u University [27] Revue de l’OFCE, « France : le coût d’Outre-Rhin : perspectives 2006-2007 pour l’économie française », April 2006, Nr. 97 [28] Revue de l’OFCE, fiche pays Allemagne, n°97, September 2006 [29] Schulz W., „Industrieökonomik und Transportsektor. Marktdynamik und Marktanpassungen im Güterverkehr“, Kölner Wissenschaftsverlag, 2004 [30] Schulz W., « Application of systems Dynamics to Empirical Industrial Organization – The Effects of the New Toll Systems », Jahrbuch für Wirtschaftswissenschaften, Band 56, 2005, Heft 2 [31] Sociétal, « Le modèle nordique », April 2006, n°52 [32] Speech of the States Secretary of Transport, M. Hennerkes, BMVBS, February 2007 [33] TLN, “Transport in cijfers”, Edition 2005 [34] Warning A., Diplomarbeit Universität Karlsruhe, „die Wettbewerbsfähigkeit von Logistik- und Transportunternehmen nach der EU-Osterweiterung“, 2006 [35] “Public policy intervention in freight transport costs: effects on printed media logistics in the Netherlands”, Hens Runhaar, Rob van der Heijde, Transport Policy 12 (2005) 35–46, 2005 [36] “Innovative Behaviour and productivity in Dutch logistics industries”, Lourens Broersma, Jeroen Segers, University of Groningen, Working paper, 33 p. [37]”Continuous poor profitability in the container trucking industry : is there a way out ?”, Rob Konings, Delft University of Technology, Working Paper [38] “Public policy in freight transport costs : effects on printed media logistics in the Netherlands’, Hens Runhaar, Rob van der Heijden, Transport Policy, 12(2005), p. 35-46 [39] Study on Road Cabotage in the freight transport market”, Final report, Framework Contract TREN/A1/56-2004, Lot 2: Economic assistance activities, European Commission, DG TREN [40] Ministère de l’Equipement, du Transport et du logement, SESP, « La transport routier de marchandises en Europe en 2004 : forte croissance du pavillon espagnol » , 2006 334 Chapter 12 Transport and Land Use Research in Transport and Logistics 2009 Proceedings Do u University LAND RENT AND NEW TRANSPORT INFRASTRUCTURE: HOW TO MANAGE THIS RELATIONSHIP? Elena SCOPEL1 Abstract This paper discusses the relationship between land rent and new transport infrastructure, and analyses the problems of social structure and distribution of income that can arise. The building of a new transport infrastructure, increasing the accessibility of the area and the level of mobility, can result in the increase in the value of land in that area. This is also tightly linked to building constraints on the area, because they raise the scarcity of land that is an element of land rent. So, if accessibility increases and if there are building constraints, also land rent will rise. This land rent can be transferred into the prices of houses and buildings, and in turn of goods and services, which are in the ownership of individuals and firms. This raises the issue of social distribution: while the transport infrastructure is paid for by all taxpayers, one important benefit, land rent, is captured only by a limited number of owners (it can be also the total of owners). Is it possible to avoid this and how? There are, at least, two ways. First, the capture of land rent by general financing options. This practice, common in USA, takes part of the benefits arising from the building of infrastructure to finance it. Second, the removal or reduction of building constraints on the land involved. Rent, among other factors, is a function of the scarcity of land and building constraints help establish scarcity of land. If new building land becomes available, total rent decreases. In this paper we examine the connection between the building of a transport infrastructure and the land rent that it creates, and we try to explore two solutions to manage the issues of social distribution that may arise. Keywords building constraints, fiscal instruments, transport infrastructure, land rent, social distribution, rent skimming. INTRODUCTION: LAND RENT The relationship between land rent and transport infrastructure is tightly established. Land rent is due to two conditions: first the scarcity of the resource “land” and second the difficulty to substitute this resource with another one. So, it is possible to explain land rent as income that a land owner receives from his land thanks to these features of scarcity and uniqueness. This latter feature is a function also of its accessibility and so, of the infrastructure level in that area. This relationship is clearer when a new transport infrastructure is built: it increases the mobility level of the area and so its accessibility. This aspect enhances the real estate price (accessibility is a component to establish the real estate price) that raise the level of land rent. The new rent can be transferred on to the price of house and building and, in turn, of goods and services. The consequences of these relationships can raise problems of social structure and distribution of income: while the transport infrastructure is paid for by all taxpayers, one important benefit, land rent, is captured only by a limited number of owners (it can be also the total of owners). Is it possible to avoid this and how? There are, at least, two ways. First, the capture of land rent by an impact fee; and second, the removal or reduction of building constraints on the land involved. RENT SKIMMING Land rent is linked to transport infrastructure: where once transport infrastructure is built, the level of rent of the building in the nearly area is raised, without any effort for the land owners. Transport infrastructure is a public service for the community, and it seems incorrect that one of its benefits, the accessibility, is used to increase the value of some private goods or services. How is it possible to distribute the benefits to the community? What can the public administrators do to recover part of land value increase linked to a new transport infrastructure? One solution is the application of some type of financing options. 1 Elena Scopel. DiAP - Politecnico di Milano, Via Bonardi 3, 20133 Milano (Italy). Tel.: (+39)02.2399.5424. E-mail: elena.scopel@mail.polimi.it. 337 Research in Transport and Logistics 2009 Proceedings Do u University Land use policy The aim of land use policies is the regulation of the use and transformation of land, and they involve all planning tools to do it. A section of these policies are the financing policies: they are dedicated to recover the new rent generated for the community and to finance infrastructures or public services. These policies can be applied because the land owners have not paid “production costs” to have the increase of land value, but it has been given via a public action (the building of transport infrastructure). There are different broad ways to skim this rent that depend on the type of partnerships. Below is reported a proposed list. Public - private partnership The increase of new land rent is, as it already said, tightly linked to the building of new transport infrastructure and to the relative property development. High levels of mobility, due to new infrastructure, establish benefits to whoever lives in that area and, in the case of property development, to the promoters of the project. Indeed, it seems lawful to apply a system of fees to skim the rent on the subjects, the property promoters, which enjoy the most advantages due to the location of the project. Often this practice is a public-private partnership between local administrators and property developers. It is sometimes considered that those developers who obtain planning permission should be required to pay for the betterment generated by a planning system which limits land supply and thus concentrates land value. This value may then be captured by the lucky developer with planning permission. In view of this, it is legitimate to seek planning gain for local communities as an ad hoc local approach to collecting betterment. Extracting such gain from developers is also attractive to local authorities and the infrastructure providers as a way of obtaining contributions to general and specific funds over and above those needed to cope with the specific impacts of project (Healey P., Purdue M., Ennis F., 1995). There is a risk that this proliferation of exactions based on property development is in large part due to today’s crisis in public finance. Many sources of funding for public infrastructure and services have not kept pace with costs. It is possible that local authority use these tools to find resources to fund new public infrastructures or services, in order to accommodate growth (Deakin E., 1996). Despite this fact, the fee is a good instrument to skim the rent due to new property development, and it is the simplest way to capture the economic benefits concerning the building of new transport infrastructure. Tools There are many ways to apply this practice, and depend on the legal authority granted to local jurisdictions by the various states. Generally there are, at least, four principal tools to skim the rent: Proffers; Impact fees; Benefit assessment; Dedicated taxes. Proffers are a premise to have, from public administrator, a building permission. They involve some actions that property developer must to do to have the permit. These actions are in relation to the dimension of impact of the project, and they can consist of a new road, new public facilities (school, hospital, etc.), etc. Impact fees were pioneered in United State in 1960. They are fees, paid from land owners, to cover the costs of services and infrastructures in proportion to the project. These fees are used to cover the social costs to the new project and they are considered a reimbursement, from property developer, to compensate the negative impacts of the project. Although development impact fees intend to transfer the burden of infrastructure provision to the developer, some evidence suggests that the cost of infrastructure gets shifted to new residents of the community and that a new homebuyer ultimately absorbs the cost. It has also been suggested that the existing community pays a portion of the cost through inflated prices on existing housing and land. 338 Research in Transport and Logistics 2009 Proceedings Do u University Development impact fees have emerged as away to pass the cost of new infrastructure to the development community. The fees can be used to pay for new roads, extending water and sewer lines, and schools, among other things. The city put in the impact fees to pass this additional cost on to the new homeowners and businesses that wanted to develop in this area. They are directly tied to planning in that they are used to help finance a local capital improvement program that itself implements overall community planning objectives. Impact fees are used in areas linked to new property development where new infrastructures are built. There are two other types of impact fees: linkage fees. These fees finance social infrastructure to meet new social needs linked to new property development. mitigation fees. This fees concern environment impacts due to new property development, and they are used to refund the negative impact on community. Impact fees are subject at some criticisms. The first blame these fees for damaging the local economy, because some property developers can move their projects to places where there are not any fees. The second considers impact fees to increase the price of houses. It is possible that property developers add to house prices the part of fees that they should pay themselves. Indeed, the fees are transferred to new owners that pay, for the same house, more costs. Often these criticisms are considered unwarranted and unrealistic, especially when impact fees are used to finance infrastructure necessary to community. Today, impact fees have become a popularly used method. In USA about 60% of all cities with over 25,000 residents along with 40% of metropolitan counties use impact fees on new developments for public services or infrastructure. In some cities or states such as Florida, 90% of communities use impact fees. Twenty six states have implemented the use of impact fees in the western portion of the country, along the Atlantic coast, and within the Great Lakes region. Benefit Assessment Districts (BAD) are considered a tax that local administrators charge property developers for the benefits introduced with a new project. The fees concern the benefits of new infrastructure and they are proportional to the services offered. They are commonly characterized as geographical areas (Business Improvement District BID) within which fees or taxes are collected to fund capital investments or special services that clearly benefit properties within the district. The distinctive feature of special assessment districts is the very close and visible tie between the facility constructed or maintained and those who benefit from and pay for it. Benefit Assessment Districts are attractive for several reasons. They shift the burden of infrastructure finance from the general public to properties receiving direct benefit, while avoiding the short-term time horizon of purely private infrastructure provision. BAD is often used to cover the costs of a new public transport lines in an area, through the fees on beneficiary (Newport Partners LLD, Davidsonville MD, 2007). One important experience of BAD is been applied in Los Angeles to finance the first part of red line of the tube. The costs of the whole project is respected in 1,4 million dollars and, in the 1985, the “Southern California Rapid Transit District” (RTD) has established a BAD to obtain and refund a funding of 130 euro million of that (about the 9%). The BAD is been funded to the owners of offices, shops, hotels and other business’ activities localized in the areas around the main station of the tube. The value of fee is depended to the different planning destination, to the distance of the building to the station and to the cost of the infrastructure in that area. Dedicated taxes are particular fees applied to some activities or people to achieve one share purpose. The essential argument is based on the principle that who receive a service should pay for it. In fact, on the theory, people will be more willing to pay if their money is dedicated to programs they want or need. It is important that the public service (school, health, police, etc.) remain to the public responsibility, and only a extraordinary services must to be finance with a dedicated taxes. These taxes are use also to assure a minimum level of support and continuity of funding for specific projects or services. Dedicated taxes have also some criticism. Main risk is about the possibility that public administrations use always these taxes to finance a project and this may become a only way. Then it can increase the difficulty of 339 Research in Transport and Logistics 2009 Proceedings Do u University adapting budgets to changing economic and social conditions because part of budget is fixed and assigned for some years to a determinate project. One infrastructure funded with a dedicated tax is the Bay Area Rapid Transit District (BART) of San Francisco. All these tools are often used to compensate a negative impact, rather than to decrease the impact itself. This can be a problem when this fee, linked to new property development, is overlapped with other tools for the mitigation of the negative impacts of the project. BUILDING CONSTRAINTS Zoning is a system of land-use regulation. It is a practice of designating permitted uses of land based on mapped zones which separate one set of land uses from another. One task of planners is to set clearly in which portion of territory is possible to build. This implies building constraints to manage the territory and to narrow the building; constraints are able to render land “scarce” and, as a consequence, to create differential rent. Camagni notes that land is a scarce resource and it is hardly to expend. For this reason it is appoint to have a potential extra-remuneration (Camagni R., 2000, pag. 184). In this setting, zoning has an influence on the land rent. Land rent is linked to the role of the public administrator, that can establish residential zoning (through the city plan) and the location of new transport infrastructure (which improve the accessibility), and it is influenced by the attractiveness of the area (through urbanization works, bigger services and urban quality). If, for example, the public authority decides to locate an infrastructural project in a specific area (due to a decrease of the building constraints in that area), there will be a benefits2 to that territory, and so an increase in land rent (Camagni R., Capello R., 2005). Why can the decrease of building constraints resolve the problem of social distribution? As it already said, building constraints help establish scarcity of land, and this characteristic increases land rent. This rent can be transferred into the prices of houses and buildings, which became more expensive. In this way there is price discrimination, because some people cannot pay for the same building or services. Moreover, building constraints are more binding in a central location (where there are economies of scale), and so the higher prices are localized in a central location. So, people with a lower willingness to pay for those goods (houses) must to move themselves to where the price (and land rent) is lower, and the city is expanding in space toward marginal areas: there is an increase in transport costs and a sprawl phenomenon. In these conditions, elements of the community are discriminated and there is an incorrect income distribution. If all activities try to concentrate in central locations, where both accessibility and economies of scale are maximum, only the price of land (= rent) counter-weights this tendency. But the price of land, as far as total saturation (skyscrapers of Shanghai scale) is not achieved, depends on its man-made scarcity, and so, on building constraints. If there is a removal or reduction of building constraints on the land involved and if new building land becomes available, total rent decreases and: the land owners do not enjoy economic benefits due to additional rent, only to have land in a particular location (for example, near new transport infrastructure); anyone can build in the same way and there is not different land; there is not sprawl induced by the scarcity of land; there are also efficiency gains, as by definition there are economic benefits stemming from any reduction of constraints; there is a transfer of benefits from land owners to all community; it is possible to change a planning system that could prefer some subjects rather than others, sometimes with a blurry mechanism of building constraints assignment. Reducing building constraints will generate a reduction both of transport costs and externalities (and more efficient non-subsidized public transport), and distribution benefits (less rent). 2 For example larger accessibility in the city, larger urban quality, larger environmental quality. 340 Research in Transport and Logistics 2009 Proceedings Do u University The tie between land rent and building constraints is summarized as follow: planning, through the definition of building constraints on the territory and the location of new transport infrastructure or new public transport system, can influence the increase of rent and the economic benefits concerning this additional rent. This supplementary rent will be high in a central area, thanks the scarcity and uniqueness of “its” land; rent weighs upon the social distribution of income, favouring land owners and giving to them benefits without any effort. If building constraints, assigned to different land with an urban plan, can favour conditions to rise land rent (already strictly linked to economies of scale and economies of agglomeration), then it is possible to hypothesize that without building constraints determined a priori on a plan, the specific rent will be equal to zero. In this situation, the additional economic benefits, owing to the constraint, could disappear. Instead of these, it will be competition that could distribute the benefit to all community. If it hypothesized, in a last case, an abstract city collapsed in one point, then it will be possible to notice that, without constraints, there could be not land rent and all the construction will be built in a place with higher economies of scale. In theory, this is the best solution to have a better social distribution. In reality, it is very difficult to apply this, because the entire community would be built near the area with most economies of scale (the limit case is an oversize skyscraper). Realistic contexts are different and they are obviously within a standard urban circular scheme and it is indeed likely that building constraints tend to be relaxed in more external and less dense areas. There are no doubts that policy recommendations have to favour the eliminating building constraints as close as possible to the central area (where there are the most economies of scale and the land is more attractive), because when building constraints are lowered there are positive impacts on income distribution. CONCLUSION The tools to skim the rent and the will to remove building constraints follow the direction to reduce the additional rent for a better social and income distribution; each of these have different characteristics but both pursue one same aim3. Therefore, is it possible to join some features of fiscal instruments and planning measures in an integrated system? In this way, it establishes only one articulate structure with one shared purpose. How can it be organized? Observing that it is not possible to remove completely building constraints and that rent will always be created, the solution is a mechanism that gives rise to rent but “under control”. In other words, with a specific distribution of building constraints targeted, it gives the possibility to use benefits concerning high accessibility (that lead to benefits on public transport system and environment) but, at the same time, through the fiscal system, it is possible to distribute these benefits to the community. The accessibility is influenced by the location of new transport infrastructure or new public transport on territory, and it is, as already said, one function of land rent. It has got special potential that, if used in better way, can exploit the benefits of the economies of scale in central area and so can reduce private mobility (in favour of the environment). Paradoxically, it should start a relation in which if accessibility increases, also the level of building on the territory increases in a proportional way, within the limits of the zoning in force. In this way, can become a further instrument to incentivise the development of areas with high levels of accessibility (Franceschini S., 2008). The solution could be including two steps: building constraints should be calculated to incentivise the proximity at points, on territory, with higher accessibility. In this way, development tends to evolve near central locations (where the most economies of scale are located) across public transport structure, and so it can disincentive private mobility and the sprawl. Planning should correlate possibilities and costs of urban expansion with the accessibility of the territory; 3 Fiscal instruments and removal of building constraints are equivalent for distribution issues, but they are very different for efficiency issues: while the fees concern the economic benefits of the new transport infrastructure but they keep the scarcity of land, the removal of building constraints eliminate also the scarcity of land. The first solution takes care of effects of land rent concerning new transport infrastructure, while the second one aims to prevent his formation. 341 Research in Transport and Logistics 2009 Proceedings Do u University benefits generated with a central location (additional rent) must be captured with fiscal instruments and, later on, spread to the whole community. The results of these actions (to apply in a unique system) are a concentration of new activities, building and services near the new node of transport, that: provide demand for transport where there is supply. In this way, there will be a (partial) reduction4 of private mobility and a betterment of the environmental5; through increasing accessibility, land rent arises and it is possible to design a fiscal system to capture the benefits deriving from the increase in property value. This solution is possible only in an urban and dense location, where public transport system is efficiency and competitive with network of road transport. With public transport system, the accessibility, and the rent that is generated, is concentrated in some points, which can be controlled. In extra-urban location, where the activities and the buildings are spread in the territory, public transport system is not useful at the necessity of movement, because it is not flexible in the territory (it has got fix timetable and fix route). Therefore the network of road transport, that is “capillary”, is winning. The accessibility is spread and so the rent is less and most distributed. In this case the fiscal instruments cannot use, because there are not essential requirements (high rent in a specific location) and lack some definite subjects to apply the mechanism (community is scatter in the territory): the system cannot controlled. The mix of instruments in use today follow different directions and the fiscal field and planning field often do not converse. This system does not ensure a rational planning on territory, because it does not: incentivise a neat organization of territory; exploit the capability of accessibility; use the benefits of transport infrastructure. Moreover, fiscal instruments do not give a correct signal and, often, their aims are obfuscated to the necessity of local authority to have a supplement voice of budget. The correct signal of the measures is an essential condition to the fiscal system. This paper wants to be a little step to increase awareness towards these issues and to start the well-known integration between transport and planning. Indeed, from many years a mixing of this know-how has been discussed theoretically, but there are not significant results in practice. Often it is difficult discover a concrete and coherent application to do that. Today, the scientific debate is in agreement on the theoretical dynamic, but it is missing the instruments to apply the integration and guidelines that address public authority towards attainment of this policies. The aim of this paper is to focus attention on these issues in an urban area and try to give a “key” to read the conflict in a different, but concrete, way. BIBLIOGRAPHICAL REFERENCES [1] Arnolfi S., Curti F. (2000), Forme alternative di finanziamento, in Karrer F. e Monardo B., a cura di, Territori e città in movimento. Strategie infrastrutturali e strumenti finanziari per lo spazio della mobilità collettiva, Alinea, Firenze [2] Ballabio F., (2000), Attualità del rapporto tra urbanistica e rendita fondiaria urbana, Tesi di Laurea in Architettura, Politecnico di Milano, Milano [3] Camagni R., (1993), Principi di economia urbana e territoriale, Carocci, Roma [4] Camagni R., Capello R., (2005), “Una valutazione dei benefici collettivi di un grande progetto urbano attraverso un indicatore sintetico: la rendita urbana”, Scienze Regionali, vol. 4, n. 2, pp. 51-92 [5] Crane R. (2000), “The influence of urban form on travel: an interpretive review”, Journal of Planning Literature, vol. 15, n. 1, pp. 3-23 4 The level of reduction is in function of the type of transport infrastructure: if it is a public service the reduction is total (the user don’t catch a private transport to move himself). 5 The betterments for the environment are linked with less movement to arrive until the new transport infrastructure. If the infrastructure is a public service, the betterment involve all the movement. 342 Research in Transport and Logistics 2009 Proceedings Do u University [6] Curti F. (1999, a cura di), Urbanistica e fiscalità locale. Orientamenti di riforma e buone pratiche in Italia e all’estero, Maggioli, Rimini [7] Deakin E. (1996), La negoziazione dello sviluppo immobiliare in fase di recessione economica, in Curti F. e Gibelli M., a cura di, Pianificazione strategica e gestione dello sviluppo urbano, Alinea, Firenze [8] Echenique M., (2007), “Mobility and income: the relationship between income and mobility”, Environment and Planning, vol. 39, pp. 1783-1789 [9] Franceschini S., (2008), Una nuova concezione degli oneri urbanistici e degli indici di edificabilità per l’integrazione territorio e trasporti, X riunione scientifica SIET, Sassari [10] Gaffney M., (1972), “Land rent, taxation and public policy: taxation and the functions of urban land rent”, American Journal of Economics and Sociology, vol. 31, n. 3, pp 241-257 [11] Geerlings H., Stead D., (2003), “The integration of land use planning, transport and environment in European policy and research”, Transport Policy, n.10, pp. 187-196 [12] Healey P., Purdue M., Ennis F. (1995), Negotiating development: rationales and practice for development obligations and planning gain, E and FN Spon, London [13] Krugman P., Venables A., (1995), “Globalization and the inequality of nations”, The Quarterly Journal of Economics, vol 110, vol. 4, pp.857-880, The MIT Press [14] Levine J., Inam A. (2004), “The market for transportation-land use integration: do developers want smarter growth than regulation allow?”, Transportation, vol. 31, pp. 409-427 [15] Newport Partners LLD, Davidsonville MD, (2007), Impact fees and housing affordability, a guidebook for practitioners, U.S. Department of Housing and Urban Development, Washington, DC [16] Ponti M., (2007), Land rent and transport policy, Politecnico di Milano, Milano [17] RICS Policy Unit (2002), Land value and public transport. Summary of findings, London:ODPM/RICS [18] Ryan S. (1999), “Property value and transportation Facilities: Findings The Transportation-Land Use Connection”, Journal of Planning Literature, vol. 13, pp. 412-427 [19] Van Wee B., Maat K., (2004), Land use and transport: a review and discussion of Dutch research, Delft University of Tecnology, Delft, The Nederlands 343 Research in Transport and Logistics 2009 Proceedings Do u University EFFECTS OF PAVEMENT CHARACTERISTICS ON THE TRAFFIC NOISE LEVELS Aybike ONGEL1 and John HARVEY2 Abstract In recent decades, noise pollution has become a major concern around the world due to industrialization and increased motorization causing a tread to human well-being. Road traffic is the most prevalent source of noise emissions by transportation. Tire/pavement noise is a major contributor to traffic noise at highway speeds. Tire/pavement noise is affected by different pavement properties. A study conducted in California measured the noise levels of different pavement types and the pavement characteristics affecting noise levels as measured by California On-Board Sound Intensity Method. Data was collected on dense graded asphalt concrete mixes (DGAC), conventional open graded mixes (OGAC), open graded rubberized asphalt concrete mixes (RAC-O), and gap graded rubberized asphalt concrete mixes (RAC-G). A total of 72 field pavement sections were included in the study, all of which were less than 8 years old at the time of the measurements. This paper evaluates the effects of pavement characteristics including the air void content, gradation properties, IRI, texture, pavement surface condition, and age on third-octave band frequency noise levels and identifies the pavement characteristics that would be more annoying to human ear. Keywords Flexible pavements, noise, pavement surface characteristics, tire/pavement noise INTRODUCTION Noise pollution caused by industrialization and increased motorization is a growing concern to the public due to its effect on human well-being. It may have negative effects on health, productivity, and economics. Health consequences of noise pollution include hearing impairment, sleep disturbance, and cardiovascular effects while productivity consequences include interference with social behavior, performance loss, interference with speech communication, and annoyance [1]-[3]. The economic consequences of noise include property value loss in areas subject to noise, lower work performance of those affected by noise [4], and medical costs of improving the state of health of those affected by noise [5]. The adverse effects of noise on health, productivity, and economy have forced highway agencies to abate traffic noise levels. Quieter pavements have become an attractive option for minimizing the impacts of traffic noise levels in neighborhoods adjacent to highways. A noise reducing surface is defined as “a road surface which, when interacting with a rolling tire, influences vehicle noise in such a way as to cause at least 3dB(A) lower vehicle noise than obtained on conventional and most common road surfaces” [6]. Literature shows that open graded asphalt mixes can reduce the tire/pavement noise, and hence traffic noise compared to dense graded mixes [7]-[9]. Tire/pavement noise is affected by pavement surface characteristics such as texture, roughness, air void content, thickness, and age. Different pavement characteristics affect different frequency levels of the tire/pavement noise. The frequency content of sound is important since it was shown that human ear is more sensitive to sound in the frequency range between 1000 Hz and 3000 Hz [10] and annoyance increased as the high frequency component of the noise increased even though overall noise level stayed the same [11]. In the last decade, open graded mixes have been placed in California and other states, in part to benefit from their noise reducing properties. There is a need to better understand the long-term acoustic properties of pavements as well as the noise reduction provided by open graded mixes. The purpose of this study is to determine the noise levels of different types of asphalt pavement mixes at different ages and to identify the effects of pavement characteristics on the noise levels as well as to identify the pavement characteristics that would be more annoying to human ear. 1 Aybike Ongel, Istanbul Kultur University, Faculty of Engineering, Civil Engineering Department, Atakoy, 34156, Istanbul, Turkey, a.ongel@iku.edu.tr 2 John Harvey, University of California, Davis, Faculty of Engineering, Civil and Environmental Engineering Department, Davis, CA, 95616, USA, jtharvey@ucdavis.edu 345 Research in Transport and Logistics 2009 Proceedings Do u University METHODOLOGY Site Selection This study presents the analysis of data collected over two years from 72 field pavement sections in California. The experimental design is a full factorial including four different asphalt pavement surface types, three different age categories, two traffic types, and two rainfall regions. There are some replicates in the factorial. The four mix types include open graded asphalt concrete with conventional and polymer-modified binders (OGAC), open graded asphalt concrete with rubberized binder (RAC-O), rubberized gap graded asphalt concrete (RAC-G), and dense graded asphalt concrete with conventional and polymer-modified binders (DGAC). Age categories include less than a year old, one to four years old, and four to eight years old. Traffic type, based on California Department of Transportation (Caltrans) 2004 annual average daily traffic (AADT) data for highways and freeways (Caltrans 2004), was categorized as “high” if the AADT (two-way) is greater than 32,000 vehicles per day and was categorized as “low” otherwise. Rainfall is based on annual average rainfall in California from 1960-1990 obtained from CDIM software (PaveSys, 2004). The rainfall was categorized as “high” if average annual rainfall is greater than 620 mm (24.4 inches) and was categorized as “low” otherwise. Data Collection In this study tire/pavement noise was measured by On-Board Sound Intensity (OBSI) method. In OBSI measurements two locations of the sound intensity probe are used: one is at the leading edge and the other at the trailing edge of the tire/pavement contact patch. The probe consists of two 25 mm phase-matched microphones spaced 16 mm apart and preamplifiers in a side-by-side configuration. A foam windscreen is placed over the microphones to reduce the wind noise. Signals from the two microphones are input to a two- channel real time analyzer. OBSI measurements are taken at 97 km/h. When that is not possible, an alternative speed of 58 km/h is used. The OBSI results measurements at 58 km/h were adjusted to their equivalent values at 97 km/h based on a field testing correlation study. Three replicate measurements are collected at each probe location, which are the results of consecutive passes with the instrumented vehicle on the 150 m sections selected for this study. Air and pavement temperatures are also recorded during OBSI measurements. Measurements were conducted using a Goodyear Aquatread III tire and Dodge Stratus car. The OBSI results are expressed in terms of A-weighted sound intensity levels, dB (A). In addition to the OBSI measurements, data on the pavement characteristics were collected. Microtexture was measured using the British Pendulum Tester and the results expressed in terms of British Pendulum Numbers (BPN). Macrotexture was measured using a high sampling frequency laser profilometer on the instrumented vehicle used for the sound intensity measurements. Macrotexture results are reported in terms of Mean Profile Depth (MPD) and Root Mean Square (RMS). Roughness was measured with the inertial laser profiler and reported as International Roughness Index (IRI). Pavement condition surveys were conducted using the Caltrans Condition Survey Manual (version year 2000) on the 150 m segments. A total of twelve cores were also collected, 6 in the wheelpath and 6 between the wheelpath, at 25 m intervals from the selected pavement sections to determine the air void content and aggregate gradation. Air-void contents were calculated using the bulk specific gravity value obtained from CoreLokTM measurements and the theoretical maximum specific gravity value obtained according to ASTM D2041. After the asphalt from the core samples was burned off in an ignition oven, the aggregate gradation was obtained by sieve analysis according to ASTM C136 and ASTM C117. Thickness of the cores were also measured and recorded in the laboratory. DATA ANALYSIS Analysis of Overall Sound Intensity Levels In this study, the effects of the following pavement variables were investigated: BPN, MPD, RMS, nominal maximum aggregate size (NMAS), coefficient of uniformity (Cu), fineness modulus, air void content, permeability, mix type, IRI, surface thickness, pavement age, rubber inclusion, and pavement distresses on the sound intensity levels. Figure 1 shows the variation of sound intensity levels of the OGAC, RAC-O, RAC-G and DGAC mixes in different age categories for the two years of data collection as Phase 1 and 2. It can be seen that sound 346 Research in Transport and Logistics 2009 Proceedings Do u University intensity values generally increase with age. It can also be seen that DGAC mixes have the highest sound intensity levels. RAC-G mixes are the quietest among all the mix types that are less than a year old, however they lose their noise reducing properties with time. Overall, open-graded mixes may reduce noise levels, on average 2 dB (A) compared to dense- and gap-graded mixes. The effects of pavement characteristics on the OBSI levels were investigated using correlation analysis. OBSI was found to be positively correlated with IRI, age, Cu, surface layer thickness, and presence of transverse cracking and raveling and negatively correlated with mix type, air void content, and fineness modulus at 0.05 significance level. OBSI levels increase with increasing IRI, age, Cu, surface layer thickness, and presence of transverse cracking and raveling and decrease with increasing air void content and fineness modulus. Open graded mixes have lower noise levels compared to dense and gap graded mixes. 109 108 107 106 105 104 103 102 101 100 Phase ID 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 1 2 Age Category <1 1-4 >4 <1 1-4 >4 <1 1-4 >4 <1 1-4 >4 Mix Type DGAC OGAC RAC-G RAC-O FIGURE 1. Sound Intensity Levels for Different Mix Types at Different Ages Figure 2 shows the sound intensity levels versus the air void content for different mix types. It can be seen that the sound intensity levels go down as the air void content increases for dense and gap graded mixes, while the noise levels of open graded mixes are insensitive to air void content for open graded mixes. 107 Mix Type RAC-G&DGAC OGAC&RAC-O 106 105 104 103 102 101 100 5.0 7.5 10.0 12.5 15.0 17.5 20.0 22.5 Air Void Content (%) FIGURE 2. Scatter Plot of A-weighted Sound Intensity Levels versus Air Void Content for Different Mix Types 347 A-Weighted Sound Intensity, dB(A) A-Weighted Sound Intensity, dB(A) Research in Transport and Logistics 2009 Proceedings Do u University Figure 3 shows the sound intensity levels versus the surface layer thickness for different mix types. Noise levels go down as the thickness increases for open graded mixes increase while the noise levels increase with increasing thicknesses for dense and gap graded mixes. Although the trend between the noise levels and thickness for open graded mixes is due to the four points, shown in the circle; it suggests that when the thickness is above 50 mm, thickness may have a noise reduction effect for open graded mixes. 107 Mix Type DGAC&RAC-G OGAC&RAC-O 106 105 104 103 102 101 100 0 20 40 60 80 100 120 Surface Layer Thickness (mm) FIGURE 3. Scatter Plot of A-weighted Sound Intensity Levels versus Surface Layer Thickness for Different Pavement Types Frequency Analysis of Noise Levels The effects of pavement characteristics on the noise levels at different frequencies were investigated using regression analyses. Separate single variable regressions were conducted for each frequency level, from 500 Hz to 5,000 Hz. shows the variables affecting the sound intensity levels at 5 % significance level for each frequency and their sign. The variables are shown in order from the one with the highest coefficient of determination to lowest coefficient of determination. A positive (+) sign indicates that increasing the value of the independent variable increases noise at the given frequency. For the categorical variables, a positive sign indicates that the variable coded as “1” increases the noise levels compared to the variable coded as “0.” It can be seen that the 500 Hz and 630 Hz band frequencies are mainly affected by texture variables (MPD and RMS) while 800 Hz and 1000 Hz band frequencies are mainly affected by mix type (open graded mixes versus dense and gap graded mixes). For frequencies above 1000 Hz, air void content has the biggest effect on the noise levels. Increasing texture also reduces the noise levels at frequencies above 1,600 Hz. Since the human ear is more sensitive to high frequency noise and more annoyed with increasing noise levels at higher frequencies, open graded mixes which have higher air void content and texture may be perceived as quieter although the overall A-weighted noise levels are not significantly different than those of dense-graded mixes. 348 A-Weighted Sound Intensity ,dB(A) TABLE 1 Pavement Characteristics Affecting the Noise Levels at Different Frequencies Significant Variables 500 Hz 630 Hz 800 Hz 1,000 Hz 1,250 Hz 1,600 Hz 2,000 Hz 2,500 Hz 3,150 Hz 4,000 Hz 5,000 Hz MPD (+) MPD (+) Mix type (-) Mix type (-) AV (-) AV (-) AV (-) AV (-) AV (-) AV (-) AV (-) RMS (+) RMS (+) FM (-) AV (-) Mix type (-) Cu (+) Mix type (-) Mix type RMS (-) RMS (-) RMS (-) (-) AV (+) AV (+) Cu (+) Cu (+) Cu (+) Mix type (-) Cu (+) RMS (-) MPD (-) MPD (-) FM (-) FM (+) IRI (+) Air-void FM (-) Age(+) FM (-) FM (-) Cu (+) FM (-) FM (-) MPD (-) Content (-) Mix type Surface IRI (+) Surface FM (-) Surface RMS (-) FM (-) Cu (+) Cu (+) Cu (+) (+) Thickness Thickness (+) Thickness (-) (+) Cu (-) Cu (-) Rubber NMAS (+) NMAS (+) RMS (-) Surface MPD (-) Mix type (-) Mix type Mix type (-) Inclusion (-) Thickness (-) (+) IRI (+) Mix type IRI (+) Presence of NMAS (+) MPD (-) Surface Surface Surface Surface (+) Transverse Thickness Thickness Thickness (+) Thickness (+) Cracking (+) (+) (+) Surface FM (+) IRI (+) MPD (-) NMAS (+) Transverse Transverse Transverse Transverse Thickness Cracking Cracking (+) Cracking (+) Cracking (+) (-) (+) Fatigue Age(+) Transverse NMAS (+) Age(+) Cracking (+) Cracking (+) Surface Transverse Fatigue Thickness Cracking (+) Cracking (+) (+) Fatigue Cracking (+) Notes: Surface type and presence of fatigue and transverse cracking are categorical variables and are coded as “0” or “1” in the regression analysis. Research in Transport and Logistics 2009 Proceedings Do u University SUMMARY AND CONCLUSIONS This study evaluated the effects of air void content, age, surface distresses, roughness, texture, and surface layer thickness on the overall sound intensity levels as well as frequency content of sound intensity levels. The pavements studied include open graded mixes, conventional and rubberized; gap graded rubberized mixes; and dense graded mixes that are less than 8 years old. The variation of sound intensity levels at different ages for different mixes was presented. The pavement characteristics affecting the overall noise levels were identified using correlation analysis while the pavement characteristics affecting the frequency content of sound intensity levels were investigated using regression analysis. The results showed that open graded and gap graded mixes have lower noise levels than dense graded mixes. Gap graded mixes are the quietest mixes among all the mixes that are less than a year old while open graded mixes are the quietest among mixes older than one year. It was found that increasing air void content and fineness modulus reduce the noise levels while increasing roughness (IRI) and the gradation’s coefficient of uniformity increase the noise levels. The trend between pavement surface thickness and air void content for open graded mixes suggested that increasing thickness may reduce the noise levels for thicknesses above 50 mm. It was also shown that increasing air void reduces the noise levels of dense and gap graded mixes while open graded mixes are quite insensitive to air void content changes. As expected, the noise levels increase with age and with the presence of raveling and transverse cracking on the pavement surface. Increasing air void content and texture reduce the noise levels at higher frequencies. It is known that human ear is more sensitive to high frequency noise, therefore open-graded mixes which have higher air void content and surface texture may be perceived as quieter even though the overall A-weighted noise levels are not significantly different than those of dense-graded mixes. REFERENCES [1] Jakovljevic B, Belojevic G, Paunovic K, Stojanov V, 2006, “Road traffic noise and sleep disturbances in an urban population: Cross-sectional study”, Croatian Medical Journal , 47 (1), 125-133 [2] Ohrstrom E, Skanberg A, Svensson H, Gidlof-Gunnarsson A, 2006, “Effects of road traffic noise and the benefit of access to quietness” , Journal of Sound and Vibration , 295 (1-2), 40-59. [3] World Health Organization Regional Office for Europe, 2005, “Noise and health” , Available online at http://www.euro.who.int/noise/ [4] Berglund B, Lindvall T, and Schwela DH, 2000, “Guidelines for community noise” Geneva, World Health Organization, Available online at http://www.who.int/docstore/peh/noise/guidelines2.html [5] Swiss Agency for the Environment, Forests, and Landscape (SAEFL). Monetization of the Health Impact due to Traffic Noise. Berne, Environmental Documentation No.166 Noise, 2003 [6] Sandberg, U and Ejsmont J.A, 2002, “Tyre/Road Noise Reference Book”, Informex, Kisa, Sweden. [7] Sandberg, U, August 2005, “State-of-the-art of Low-noise Pavements” , Presentations from SILVIA Final Seminar, Brussels, Belgium, (Feb. 10, 2006). [8] Colwill D.M, Bowskill, G.J, Nichols, J.C, and Daines M.E, 1993 , “Porous Asphalt Trials in the United Kingdom”. Transportation Research Record 1427, Transportation Research Board, National Research Council, Washington, D.C., 13-21 [9] Camomilla G, Malgarini M, and Gervasio S, 1990 , “Sound Absorption and Winter Performance of Porous Asphalt Pavement”. Transportation Research Record 1265, Transportation Research Board, National Research Council, Washington, D.C., 1-8 [10] Bray J, Cragg P, Macknigh A, Mills R, 1998, “Lecture notes on human physiology”, 4th edition, Blackwell Science, Boston, USA. [11] Ishiyama T, Hashimato T, 2000, “The impact of sound quality on annoyance caused by road traffic noise: an influence on frequency spectra on annoyance”, JSAE Review, 21, 225-230 350 Research in Transport and Logistics 2009 Proceedings Do u University FUZZY MEDICAL WASTE DISPOSAL FACILITY LOCATION PROBLEM 1, Müjde EROL GENEVOIS 2 and H. Ziya ULUKAN 3 Abstract This paper provides a two step solution procedures for medical waste disposal facility location problem in Istanbul. Medical waste must be disposed without damaging environment and human health, complying with new regulations. Our aim is to provide an alternative method to the medical facility location problem in Istanbul, able to handle the fuzziness of the real world. People attempt to minimize the undesirable effects introduced by the new facility by maximizing its minimum Euclidean distance with respect to all demand points and also to minimize the total transportation costs. In our study, we propose the following two step solution procedure where we obtain all efficient solutions at the first step and then we choose the best solution at the second step. We use the Ishii’s model to obtain a list of candidate points. Then, we made a selection between these candidate points by solving the fuzzy goal programming problem. Keywords Fuzzy facility location, Fuzzy goal programming, Maximin, Medical Waste INTRODUCTION Recently, the controlled collection and safe disposal of medical waste has become an important field of environment protection in Turkey as in whole world. The 50% of environmental pollution has occurred in last 35 years. As a result, the scientists work on eliminating all harmful factors on air, earth, water and human life. The safe disposal and the regain of waste are some of the important fields that they work on. Even if the medical waste consist a small part of waste amount, considering the threat they constitute on human life, their safe disposal is vital. The infectious, pathological, cutter-driller wastes emanating from medical institutions are known as medical wastes. Throwing away without control or common disposal with household waste poses a serious threat for all livings. That is why medical waste must be disposed without damaging the human mental and physical health, the animal health, the flora, the water and the welfare of the society, complying with new regulations. The aim of this study is to provide an alternative method to solve the medical facility location problem in Istanbul, able to handle the fuzziness of transportation costs and people preferences. Nuclear plant, oil refining plant, waste disposal plant are plants undesirable close to residential area. People attempt to minimize the undesirable effects introduced by the new facility by maximizing its minimum Euclidean distance with respect to all demand points and also to minimize the total transportation costs. These are semi-obnoxious facilities and to locate them, facility planners determined two objectives [1]. The first aims to maximize the minimum distance from the new facility to the demand points; this is the maximin problem. The second aim is to minimize the total distance from the facility to the demand points in order to minimize the transportation costs. In our study, we propose the following two-step solution procedure where we obtain all efficient solutions at the first step and then we choose the best solution at the second step. Considering the fuzzy nature of the people attitude towards the location of this kind of facilities we will try to find the site of the facility which maximizes the minimal satisfaction degree among all demand points and maximizes the preference of the site by using a method based on the Ishii’s where the attitudes of people are expressed by a trapezoidal membership function [2]. The function represents the satisfaction degree of demand points with respect to the distance from these points to the facility site. We reformulate the Ishii’s model to obtain a list of candidate points. Then, we made a selection between these candidate points by solving the fuzzy obnoxious facility location problem, using an algorithm based on the Chen’s [3]. 1 l, Turkey, ykop@gsu.edu.tr 2 Müjde Erol Genevois, Galatasaray University, Faculty of Engineering and Technology, Industrial Engineering Department, Ortaköy, Istanbul, Turkey, merol@gsu.edu.tr 3 H. Ziya Ulukan, Galatasaray University, Faculty of Engineering and Technology, Industrial Engineering Department, Ortaköy, Istanbul, Turkey, zulukan@gsu.edu.tr 351 Research in Transport and Logistics 2009 Proceedings Do u University DISPOSAL OF MEDICAL WASTE There are many proven techniques for the safe disposal of medical waste all around the world. Nowadays, new contagious diseases that appeared in diverse countries induced all medical authorities to take severe precautions. In 2003, with the outbreak of severe acute respiratory syndrome (SARS) the authorities take more serious steps in managing medical waste. The procedures for handling, treatment and disposal of this waste were required to comply with the most stringent standards [4]-[5]. The health authority in Taiwan had to handle the current status of waste production for further management planning. The quantity of waste generated by hospitals varied by changes in local legislation according to the studies [6]-[7]-[8]-[9]. In Turkey, the medical waste production increases due to the economic and social changes and the population increase. The amount of waste is too huge to dispose in the dumping ground and requires an integrated management concept involving collection, transportation and disposal. By 2005 the number of hospital in Turkey has become 1198 and due to development of medical technologies and hygienic precautions total waste amount has been increasing. According to MEF’s researches, total waste amount in Turkey is 238.26 tons per day and 86 968 tons per year. 23 000 tons of this amount is collected in Istanbul by ISTAC (sub institution of Istanbul Metropolitan Municipality) and district municipalities. The mortal contagious diseases that appeared recently in Turkey as Avian Influenza, as well-known name bird flu, Crimean-congo haemorrhagic fever (CCHF) and the holocausts related to these diseases, furthermore the increase of other contagious diseases as aids, hepatitis and tuberculosis accentuate the vitality of the collection and the safe disposal of medical waste. The regulations of medical waste control inure on 20.07.2005 in Turkey. The regulations are based on the 8th article of the environment code where it is denoted that inappropriate collection, transportation and disposal are forbidden. In the 35th article of regulations there is a legal decision on the medical waste disposal facility location: The distance of disposal facility can’t be closer than 1000 meters to the residential area. METHODOLOGY Fuzzy Facility Location with Preference of Candidate Sites Facilities as power plants, chemical plants, dumping grounds, airports are undesired close to residential area but they must also be at a reasonable distance easy to reach in order to minimize transportation costs. By using different decision making methods we must compromise two objectives. People don’t want to live near a dumping ground but want also to get rid of the waste they produce as fast as possible. So we can not place the facility to an inaccessible distance. The first step of this study is bas attitudes of the demand points (people, clients) towards the location of new facility are expressed by a trapezoidal membership function [2]. We reformulate Ishii’s problem to obtain a list of candidate points. Ishii categorizes people attitudes in three categories. In our study we have two categories of attitude: 1. People don’t want the facility near. 2. People don’t want high cots of transportation. As it is mentioned in the medical waste regulations the burning and storage facility can not be close more than 1000 meters. Anyway people prefer the facility to be more than 5000 meters far from their residential area. a = 1 0 (di a) 1 b = 5 (d di ai ) (a d b)b a i 1 (b di ) (1) a b km FIGURE. 1 Membership function related to distances. 352 Research in Transport and Logistics 2009 Proceedings Do u University E(x pi ) N (y qi ) (x pi , y qi ) W (x pi ) N (y qi ) (x pi , y q )di (X ) i E(x pi ) S(y qi ) (x pi , y qi ) W (x pi ) S(y qi ) (x pi , y qi ) (2) N: North = 1; S: South = -1; E: East = 1; W: West = -1 Let be n demand points in the rectangular area. These points are represented as pi , qi , i 1,2,..., n . The facility that we will build is represented as X x, y and X (x, y) pA x pÜ , qA y qÜ [2]. Defin (di ) . Then, we model our problem as [2]: P :Maximize s.t. di (x) a b a 0 1 pL x pU q y q (3) L U But the way that we will pursue to solve this problem is to partition the rectangular area to small rectangles by tracing the horizontal and vertical lines passing by each demand points. For each line intersection point called x jm we will solve the sub problem below. The optimums of each sub problem can be handled by another decision making tool in order to select the best location for our facility. Q :Maximize s.t. (b a) a E(x pi ) N (y qi ) (b a) a W (x pi ) N (y qi ) (b a) a E(x pi ) S(y qi ) (b a) a W (x pi ) S(y qi ) 0 1 x X i 1,2,...,n (4) Selection of the Best Candidate Site with Fuzzy Goal Programming A starting point for the GP model can be found by restating the LP model, its assumptions and modeling notation. Each constraint that makes up an LP model is separate function, called a functional. These functionals are viewed as individual objectives or goals to be attained. Charnes and Cooper (1961) referring to these functionals as goals, suggested that goal attainment is achieved by minimizing their absolute deviation. They illustrated how that deviation could be minimized by placing the variables representing deviation directly in the objective function of the model. This allows multiple goals to be expressed in a model that will permit a solution to be found. A generally accepted statement of this type of GP model was presented in Charnes and Cooper (1977), [10]. To deal with uncertainty, many attempts have been made but the most fruitful was the theory of Zadeh. Since the single objective fuzzy linear programming (FLP) study made by Zimmermann in 1976 and multi objective fuzzy linear programming in 1978, the fuzzy theory has been applied to many decision making problems. One of these applications is the fuzzy goal programming (FGP) study of Narsimhan in 1980 with imprecise aspiration levels of fuzzy goals. There are many studies involving different kinds of FGP method to deal with uncertain data about a certain parameter (fuzzy alternatives, fuzzy objective functions, fuzzy deviation functions etc.). The study of Hannan in 1981 represents a fuzzy logic based method where decision maker satisfaction in goal attaining are represented by piecewise linear functions [12]. In 1991, nonlinear membership functions are used in FGP by Yang, Ignizio and Kim. Tiwari et al. (1987) proposed a method 353 Research in Transport and Logistics 2009 Proceedings Do u University similar to lexicographic GP where the problem is decomposed into n sub problems. Here, n is the number of preemptive priority levels [13]. To solve our FGP model we use the algorithm proposed by Huey-Kuo Chen, 1994, which is a modified version of the method developed by Tiwari, Dharmar and Rao in 1993, [3]. Our solving algorithm uses symmetrically triangular membership functions of fuzzy goals. The membership function of this FGP is [3]: 1 if (AX )i gi 0 if (AX )i gi AX g i i if gi (AX ) g i AX i i g gi (AX )i if gi (AX )i gi g i (5) g g 0 if gi (AX )i The aim of the method proposed by Tiwari et al. is to find the maximum membership value by enumerating all possible combinations which cover the entire feasible region. We consider two subsets which are left and right hand sides intersecting at the point having the highest membership value equal to 1. Therefore, there will be 2m sub problems taking into account of different combinations of membership functions of fuzzy goals [3]: AX g gi AX i max min i i max min x 0 i and x 0 i i i gi AX gi i 1,2,..., m gi AX g i 1,2,...,mi i (6)-(7) i As the sub problems are linear with a single objective function, the FGP method has the advantage that a commercially available software as LINDO may be used for solving it. APPLICATION This two-step model is applied to the medical waste facility selection for the most populated region of showing the application region. The distances and coordinates are provided by Google Earth (Figure 2). FIGURE. 2 Selection of a medical waste disposal facility in Istanbul. In rectangular area there are four demand points and all lines passing through form 25 points which means that we will solve the sub problem (4) 25 times. The optimum membership values and optimum candidate closer to each other. So, the most reasonable 354 Research in Transport and Logistics 2009 Proceedings Do u University .600 in terms of an observation made on the map and we will make our selection by solving a fuzzy goal programming problem for facility location. TABLE 1 SUB PROBLEMS OPTIMUM OPTIMUM POINT VALUE SUB PROBLEM POINT VALUE 1 - - 14 (6.350; 7.110) 0.422 2 (1.790; 3.450) 0.110 15 (2.675; 0.000) 0.549 3 (6.350; 0.000) 0.140 16 (0.000; 0.565) 0.565 4 (5.330; 7.110) 0.167 17 (3.230; 7.110) 0.567 5 (5.080; 2.160) 0.217 18 (6.350; 3.645) 0.579 6 (5.080; 3.645) 0.261 19 (1.640; 5.645) 0.621 7 (5.330; 3.645) 0.324 20 (6.350; 3.045) 0.631 8 (4.445; 2.160) 0.376 21 (0.000; 3.100) 0.645 9 (5.080; 2.795) 0.376 22 (0.000; 0.000) 0,700 10 (5.080; 2.795) 0.376 23 (1.640; 7.110) 0.965 11 (3.230; 0.000) 0.410 24 (0.000; 5.520) 1.000 12 (6.350; 2.160) 0.410 25 (0.080; 5.690) 1.000 13 (2.675; 0.540) 0.413 from the new facility to the demand points (minisum). As we have 2 goals, we have to solve 22 subproblems. Our seven candidate waste disposal facilities are highlight in the Table 1 and the medical waste amounts of each districts (demand points) are in the Table 2. These amounts will help us to determine the weights of each district that we will use in the minisum goal. Table 3 contains the distance between all demand points and candidate sites. TABLE 2 Waste amounts of each district and weights (Istanbul Metropolitan Municipality) DISTRICTS AMOUNT WEIGHT (KG/YEAR) SISLI 87 600 0.362 64 800 0.268 KAGITHANE 64 800 0.268 SARIYER 24 600 0.102 TABLE 3 Distances of candidate sites to each demand point SISLI BESIKTAS KAGITHANE SARIYER 1 6.680 5.870 3.800 10.770 2 3.580 7.890 2.580 7.670 3 3.525 3.525 5.595 3.915 4 5.300 10.310 5.000 5.250 5 3.785 8.795 3.485 3.485 6 5.390 10.400 5.090 5.000 7 5.250 10.260 4.950 4.860 Our objectives are: 7 G1 : max j z j where z j 0,1 j 1 z 7 4 j 0,1 G2 : min widij z j where dij are dis tan ces j 1 i 1 wi are weights (8) We solve our fuzzy goal programming problem using Chen’s algorithm. We use LINDO as tool and we found that we can construct our medical waste disposal facility to the candidate site 6: (0.080; 5.690). This point is in a green area at 355 Research in Transport and Logistics 2009 Proceedings Do u University transportation cots. So we can conclude that our study is meaningful for this kind of problems. But there can be also other logistical constraints. In that case this model may be weak. CONCLUSION This paper deals with the problem of determining the undesirable facility location site in Istanbul. This kind of problems own naturally two objectives one is to maximize the minimum distance from the new facility to the demand points and the second is to minimize the total distance from the facility to the demand points in order to minimize the transportation costs. We tried first to find the site of the facility which maximizes the minimal satisfaction degree among all demand points and maximizes the preference of the site by using a method based on the Ishii’s. We used fuzzy membership function to represent the attitudes of demand points related to the distances between demand points and new facility location. By maximizing the membership value we try to find the optimum point. There are two drawbacks of this kind of solution procedure. The first is that this model suggests that residential areas are points discrete. This problem can be handled by revising the intervals of membership function. Second is that more than more than one sub-problem can have the same optimum solution. To handle this problem we use the optimum points having nearly same values bigger than a fixed value to solve a fuzzy goal programming where we add the minisum objective. After that procedure we obtain a unique solution optimizing the two objectives. The apparent advantage of our solution procedure is that all problems are converted to linear problems and we solve them easily using LINDO software. ACKNOWLEDGMENT This research has been financially supported by Galatasaray University Research Fund. REFERENCES [1] Melachrinoudis, E., Xanthopulos, Z., 2002. “A Semi-obnoxious single location in Euclidean space”. Computers&Operations Research, 30, 2191-2209. [2] Ishii, H., Lee, Y. L., Yeh, K. Y. 2007. “Fuzzy facility location problem with preference of candidate sites”. Fuzzy Sets and Systems, 158, 1922-1930. [3] Chen, H-K., 1994. “A note on a fuzzy goal programming algorithm by Tiwari, Dharmar and Rao”. Fuzzy Sets and Systems, 62, 287-290. [4] Chiang, C.F., Sung, F.C., Chang, F.H., Tsai, C.T., 2006. “Hospital waste generation during the SARS outbreak in Taiwan”. Infection Control and Hospital Epidemiology, 27, 519-522. [5] Cheng, Y.W., Sung, F.C., Yang Y., Lo Y.H., Chung Y.T., Li K.-C. 2008. “Medical waste production at hospitals and associated factors”. Waste Management, in press. [6] Askarian, M., Vakili, M., Kabir G. 2004. “Results of a hospital waste survey in private hospitals in Fars Province, Iran”. Waste Management, 24, 347-352. [7] Fisher, S., 2005.“Healthcare waste management in the UK: the challenges facing healthcare waste producers in light of changes in legislation and increased pressures to manage waste more efficiently”.Waste Management, 25, 572-574. [8] Mohee, R., 2005. “Medical wastes characterization in healthcare institutions in Mauritius”. Waste Management, 25, 575-581. [9] Bdour, A., Altrabsheh, B., Hadadin, N., Al-Shareif, M., 2007. “Assessment of medical wastes management practice: a case study of Northern Part of Jordan”. Waste Management, 27, 746-759. [10] Keeney R.L., Raiffa H., 2003. “Decisions with Multiple Objectives Preferences and Value Tradeoffs”. Cambridge University Press. [11] Dubois, D., Prade, H., 2000. “Fundamentals of Fuzzy Sets”, Kluwer Academic Publishers. [12] Aköz O., Petrovic D., 2006. “A fuzzy goal programming method with imprecise goal hierarchy”. European Journal of Operational Research, 181, 1427-1433. [13] Kumar M., Vrat P., Shankar R., 2004. “A fuzzy goal programming approach for vendor selection problem in a supply chain”. Computers & Ind. Eng., 46, 69-85. 356 Chapter 13 Transport Infrastructure and Investment Appraisal Research in Transport and Logistics 2009 Proceedings Do u University AGENTS’ BEHAVIOUR IN FINANCING ITALIAN TRANSPORT INFRASTRUCTURES Paolo BERIA1 Abstract The aim of the paper is to critically describe the behaviour of the main Italian transport infrastructural agents (national roads and rails, highway, airports concessions) in the field of infrastructure financing. All the actors are trying to avoid the effects of regulation and, at the same time, to reinforce their dominant position and power. The thesis of the paper is that the funding and building of new infrastructures is often the pillar of this strategy, since this is the field where the legal framework is weaker. A short review of the present normative is followed by the description of the behaviour of the main agents. Recent planning documents will be commented, pointing out the tendency of the regulated monopolists to find out new strategies to maximise their objectives. Among these, the proposal of new infrastructures is the field where monopolists are more active and where regulation should be more effective. Keywords regulation, investment, infrastructure, airports, roads, highways, railroads, price cap. INTRODUCTION: INVESTMENTS AND REGULATION The aim of the paper is to critically describe the behaviour of the main Italian transport infrastructural agents in the field of infrastructure financing. All the actors are trying, according to the structure of respective markets, to avoid or minimise the effects of regulation and, at the same time, to reinforce their position and power against the regulator. The thesis of the paper is that the funding and building of new infrastructures is often one of the pillars of this strategy, since this is the field where the regulatory framework is weaker. The topic of the relationship between the regulation of transport markets and the investments in new infrastructures is seldom discussed. The well-known approaches of privatisation, price-cap, yardstick competition, tenders, etc. do not explicitly consider the investments as part of the regulatory problem or do it in a way that reveals as non satisfactory in the practice. In particular, the price-cap, one of the most applied methods for the regulation of non contestable sectors and of natural monopolies, does not distinguish in the majority of applications, the type of investment. As discussed before in Beria and Ponti (2008), in fact, a distinction between “endogenous” and “exogenous” investments is necessary to avoid the tendency to overinvestment or gold-plating practices. The “endogenous” investments are those that reduce the costs (introducing a new technology, for example) or raise the demand (building a new section) at a level that makes the investment self-profitable for the regulated actor, also at a given tariff. In other words, for these investments an explicit remuneration is not necessary because self- financing. To the contrary, those investments that are not able to raise revenues or lower costs at a level that make them profitable, should be exogenously financed through a tariff increase or public lump sum subsidies. This, of course, only in case they result profitable in socio-economic terms. The lack of distinction between those two categories and the remuneration through tariff increases of any kind of investment, may generate extra-profits. This gives an incentive to overinvestment, at the expenses of the users that pay unnecessarily high tariffs. The already mentioned paper demonstrated the wide occurrence of this practice in Italian regulated sectors. The present paper aims to investigate a further aspect of the problem of investments done by a regulated (or a weakly-regulated) concessionaire or agency. This aspect deals with the use of investments to increase the contracting power of the regulated, up to the substitution of the public planner with a private-like (or even private) planner, and to grant long term monopoly and rents. In some cases this behaviour lays in the insufficiency of the regulatory normative; in other cases it is the outcome of an opaque application of it or of the capture of the regulator by the agent. In this second case it is not always possible to bring proofs supporting facts that are actually circumstantiated impressions. 1 Paolo Beria. DiAP - Politecnico di Milano, Via Bonardi 3, 20133 Milano (Italy). Tel.: (+39)02.2399.5424. E-mail: paolo.beria@polimi.it. 359 Research in Transport and Logistics 2009 Proceedings Do u University The paper starts with a short review of the present Italian normative, followed by the description of the behaviour of the main agents and modes, namely the highway concessionaires, the national road agency, the national railways and the major airport concessionaires. Some conclusions will be drawn at the end. BRIEF NORMATIVE OVERVIEW The Italian normative framework is supposed to regulate all the transport natural monopolies through fares, both in terms of operations and financing. However, the normative is not homogeneous among the modes and largely insufficient to promote efficiency (Ponti et al., 2007). Highway and airport fares are regulated via price-cap. In both cases the regulation applied is unsatisfactory due to the lack of actual pressure to efficiency and for the favourable investment remuneration (Beria and Ponti, 2008, Ragazzi, 2008). Rail track access tolls are calculated with a formula defined in a decree. There is no price-cap or any other form of efficiency promoting mechanism. Rail services fares are defined by the state or by the regions (for those considered as social services) or are de facto left to the incumbent operator choice. Road system is free of charge and no regulation on subsidies is present. Since the national regulatory framework is so contradictory, it is generally not completely effective. Major rents are still present, also in form of inefficiency. Concerning the financing, some effects can be recognised in the past, like “gold plating” and cost overruns. GOLD PLATING AND PROFITS IN THE HIGHWAY SECTOR Italian highway system consists of approximately 6.500 km, the majority of which is franchised. Part of this network, 600 km, has characteristics of highway, but is built and owned by the National Road agency ANAS (see below). The concessionaires are 25, usually privately owned or mixed public-private. The main concessionaire is the private company group “Autostrade”, which owns the 61% of the network and serves the 67% of the traffic. It was formerly owned by the state and was privatised in 1999. Another group, the “Gruppo Gavio” owns the 20% of the network. The rest of the network is owned by local authorities or by minor private shareholders (Ragazzi, 2008). The privatisations and the extension of concessions Until the end of the Nineties, all the concessions were automatically renewed at the same conditions and without any tender. The renewal was formally justified with the promise of further investments, which needed to be adequately remunerated. In 1999, the Government decided to privatise the main concession, “Autostrade” in order to accomplish at European request to dismantle the former owner, IRI, a public company. However, the hidden rationale of such privatisation was to earn as much as possible in a period of financial constraints by expanding the value of the stocks (Boitani, 2004; Coco and Ponti, 2006). The maximisation of the company value was done by introducing some distortions in the price-cap mechanism adopted since 1996 for the yearly update of the tolls. These distortions regard both the lack of an effective incentive to efficiency maximisation and profit minimisation, and the issue of investments. According to the decree named “Costa–Ciampi” (20 October 1998, n. 283), following the criterion of past costs recovery, the renewal periods should have been very limited for all the concessions (1 to 7 years, with only one exception of 18 years) (Ragazzi, 2008). This short concession was a threat for the survival of the present concessionaires. In few years all of them should have transferred the infrastructure back to the regulator, due to the fact that all their investments were completely repaid and widely remunerated. However, the still public “Autostrade” obtained easily the renewal of the concession for another 40 years up to 2038, just before the privatisation and without any tender (in 1997, until 2038). This fact opened the way, between 1999 and 2000, for many other concessionaires to conveniently extend their concession period, despite the European rules forbidding automatic renewals without tendering. The mix of weak regulation, extension of the concession and approval of huge investment plans, made ex-post Autostrade (but also the other concessions) extremely profitable for the investor and increased noteworthy its value2. The counterpart is that such a favourable regulation cannot, by definition, promote efficiency and favour the users. 2 Boitani (2004) quotes some documents of that time evidencing the fact that the expected profits of Autostrade were largely superior to the official ones, due to the particularly favourable regulatory framework. 360 Research in Transport and Logistics 2009 Proceedings Do u University Price-cap and new investments The main justification to the extension of the concessions to Autostrade and to the other networks under expiration, laid in the field of new investments. In fact, all the concessionaires succeeded in having a further concession period in order to repay new investment plans. These investments, once their social utility was demonstrated, could have been properly financed by fixing an end of concession value to be paid by the new concessionaire after the expiration of the present one. This would have avoided of having longer concession periods before tenders. However, this did not happen and concessions were extended for 6 to 27 years (with two shorter exceptions only). Moreover, only few concessions would have expired around 20103, all the rest around 2020 or later, up to 20384. The extension was calculated as the sum of the extension due to the past unrecovered investments, plus the extension due to recover the future, declared, investments. The new investments, with the exclusion of Autostrade, consisted in approx. 4.500M€. The figure is slightly lower than the initial proposal made by the concessionaires of approx. 5.500M€ (Ragazzi, 2008). As already mentioned, the remuneration is done by applying, since 2000, a new regulatory regime ruled by a price cap formula. However, the model applied is very far from the theoretical one (Coco and Ponti, 2006) especially concerning investments. The principle determining fares is that of uniformity across parts of the network and tariffs are, in fact, defined to fully recover the investment costs. No criterion is applied in order to efficiently allocate the demand (Boitani, 2004). The effect seems to be that these investments are not only not risky, due to the favourable price-cap, but also that they are assessed by the concessionaire and not by the regulator, whose control on methodology and unit costs is not transparent. Moreover, as Italian norms require, the assessment refers to single, specific, investments and not to a plan. Finally, there is no distinction between endogenous investments, i.e. those that can be directly financed through the direct increase of demand or lower of operating costs, and exogenous investments, i.e. those that must be paid by public funds or by other users because not self-profitable (Beria and Ponti, 2008). The outcome of this normative framework (together with the feeble assessment process used in Italy, especially at the planning level) is to stimulate overinvestment: the more investments (promised), the longest is the concession period and the higher can be the tolls. The important modifications introduced with CIPE decree 1/2007 were not sufficient to limit the power of the concessionaires, but only to stop for one year (2007) the increase of tolls5 (Ragazzi, 2008). The decree has been issued because of many cases of investments declared, included and pain in tolls, but actually never built. The resistance of all the concessionaires to this decree has been strong. Another correction of the past, problematic, framework deals with the quantification of the value of the investment. Before 2007, the amount of the investment was updated at the end on the basis of the declared ex- post costs. This did not give any incentive to correctly forecast the investment cost and opened the way to large and incontrollable overruns. Since 2007 the investment remunerated in price cap formula is the one declared ex-ante by the promoter – concessionaire. Recently (2008), the formula is, once again, under modification, this time in favour of concessionaires. In the future, the price-cap formula will fix also the “X” parameter to a predetermined share of the inflation rate, neutralising also any driver to efficiency. Regulation or incentive to gold-plating? The given description stresses that the Italian regulatory framework concerning highways investments is critical. It is not only a matter of weakness and normative inadequacy. Rather, some contents of the regulation are distorting the sector and inducing opportunistic behaviours of the agents. Their resistance has been until now capable of reducing the effectiveness of the regulation. Some problems, limiting the analysis to the field of the investment choices only, can be stressed. a. The way initial tariffs are set, i.e. the “historical” ones, determines an unregulated monopolistic rent once the past investments are repaid. 3 Centropadane (Piacenza-Brescia) in 2011, Cisa (Parma-La Spezia) in 2010, Padova- Mestre in 2009. 4 Torino – Savona. 5 For example, Autostrade per l’Italia had an increase of only 0,17% instead of 2,93% if calculated with the previous rules. This means that the phenomenon of declared but not realised investments is crucial. 361 Research in Transport and Logistics 2009 Proceedings Do u University b. New investments are automatically remunerated in the tariff, without applying any efficiency criterion and in a scarcely transparent way. Moreover, these increases are spread on the whole network even if benefits (e.g. a new section) are limited to some users only (Boitani, 2004). c. The extension of the concession and the extra-remuneration of new investments during the concession period mean that new investments are paid twice (Ragazzi, 2008). d. Present contracts remunerate any kind of investment. However, only exogenous ones should be remunerated (Beria and Ponti, 2008). e. Every new investment is proposed by the concessionaire (Ponti and Boitani, 2005), which actually acts as a planner. The plan is only formally approved by ANAS, which is supposed to be the planner/regulator. The distinction between investments promoted by the concessionaire (to be paid by new demand or cost reduction) and by the authority/planner (to be paid in tariff or by a tender) is necessary, but absent. The outcomes of this regulatory framework are important. Privatisation and regulation aimed at maximising the value of the public companies to be privatised and not to obtain efficiency and fair tolls. The consequence is that now the concessionaires are doing, with no exceptions, extraordinary extra-profits (Ragazzi, 2008), increasing their already strong position in front of the state. Also the transparency of such profits is scarce. Furthermore, no productive and allocative efficiency is promoted. A perverse incentive to overinvestment is given, known as Averch-Johnson effect, both in quantity (new and unnecessary investments) and in unitary prices (no control on cost overruns and inefficient engineering situations, defined as gold plating) (Gomez Ibanez, 2003). Obtaining the approval of monumental and unjustified investment plans, even if longer than the concession period, is easy. This behaviour, even if not linked to a longer concession period, gives immediately extra-profits (Ponti and Boitani, 2005), thanks to the price-cap parameter. Moreover, the socio-economic rationale of these investments is weakly verified by the authority. ANAS, in fact, approved investments plans without any priority verification, stressing that investments are not publicly funded. NATIONAL ROADS: A STATE AGENT BECOMING A POWERFUL PLANNER The Italian national road system consists of approx. 21.500 km, including some 4 lanes roads and approx. 1.200 km of highways or roads with characteristics of highways. All its network, except part of the highways, is completely free. The whole national system is managed by one single concessionaire, named ANAS6 and 100% owned by the Treasury. The concession will expire in 20307. ANAS is now a stock company, even if fully state owned, but its financial autonomy is still very scarce and depends mainly on earmarked state transfers. In the past the “strategy” of the agency was quite simple and typical of the operative branches of ministries. Its task was to build, maintain and operate the national network, under state transfer and monopoly conditions. However, due to historical reasons, ANAS is not only the network manager. It is also in charge of an authority-like function for franchised highways. Under this function, it manages the bids and regulates the fares of the concessions. However, ANAS is now in charge also of some expired toll concessions and aims to introduce tolls on many trunk roads: this double function is generating a conflict of interests and threatening its independence as authority. The industrial plan 2007-2052 and a new powerful actor The analysis of ANAS case became recently extremely significant, after the recent issue of an ambitious industrial plan. The plan, transmitted to the Treasury in June 2007, is still under approval in 20088. The agent is dependent, both on operation and investments, on the actual state transfers, allocated by law every year. Since transfers come always from the state or from local budget, ANAS concur to national debt, constrained by European agreements. The largest part of national transfers is, up to now, for new investments, that sum up to more than 1 billion Euros per year, with few exceptions9. 6 Also Provinces and Municipalities build and manage their own road networks. 7 Unless renewed until 2050 as requested by the agency itself. 8 The first comments of the technicians of the Ministry is extremely “cautious” (Ministero Economia, 2008). 9 The year 2006 was an anomaly, with “only” 100M€ spent. 362 Research in Transport and Logistics 2009 Proceedings Do u University ANAS dependency on end-of-budget state transfers and the stochastic character of them across years, limits the actual responsibility of the managers towards an efficient behaviour. It is not surprising that the budget showed very negative results often in the past years, despite the large transfers. In 2007 a new industrial plan has been issued, covering a very long period until 2052. It is proposing radical modifications on the status, the role and the autonomy of ANAS. The plan has not been approved yet at the moment of writing, but it will likely be. The key issues of the plan are listed below. All of them are heavily affecting the way Italian roads will be financed in the future (Beria and Ponti, 2008): introduction of shadow tolls and real tolls on some trunk roads, now free. State subsidies remain for the “non commercial” network; acquisition of the network of some highway concessions under expiration; some new roads will be financed by ANAS budget (obviously larger than now) instead of direct state transfers. Direct funding remains for the new “non commercial” roads, i.e. the large majority. All these measures have the primary purpose to “deconsolidate” ANAS from the Italian public debt. Also, the public ownership, the sovereign guarantee, the status of legal monopoly and the lack of any regulation makes ANAS financial rating high and very attractive for borrowing loans and for private investors. The plan demonstrates the huge importance of investments in defining the strategy of the actor. ANAS is mainly aiming at becoming an independent subject, independently promoting development plans, increasing revenues, introducing tolls and increasing debt expositions. However, the ambiguous status of public company and of authority, in contradiction with the new role of market agent, was never put into discussion. Many comments can be drawn on the most questionable aspects of the proposed evolution, especially related to the investment topic: Huge increase of the dimension of ANAS in terms of assets, including the whole trunk road network, all the highway concession under expiration and the whole new network that they will decide to build. Also the functions are increasing in number and extent: not only regulator and agent, but also concessionaire of many main roads. There is no clear discussion on the most efficient dimension of the road monopolist, which is now matching with national borders. Moreover, it is not demonstrated that further enlarging the dimension of the concession with new roads and functions goes in the right direction. No toll regulation is explicitly proposed. ANAS seems to become both the regulated actor and the regulator of itself. The main argument used to support the evolution is related to investments, in particular the fact that the new rules allows to skip every state budget constraint and that new investments will come “at no cost”. The argument is fictitious. The source of funds for the financing of the new network is supposed to come mainly from the toll revenues. However, part of these revenues will continue to come from the state budget (with another name) and part from users that now have free roads. The projections included in the plan go from 15 millions of real tolls out of 185 millions of total toll revenues in 2012, to approx 1000 millions out of 2800 millions in 2052 (plus the transfers for the investments in the non commercial network). So, the largest part of revenues is still from treasury budget, differently from what stated. But this is not the main point. Three further aspects must be underlined, stressing the importance of new investments regulation in ANAS strategy: The plan is based on the existence and perpetuation of the unregulated monopoly, which can dispose of the totality of the national network. Cross subsidisation is the rule, with profitable roads financing the ancillary network. Moreover, the monopoly is not regulated and the ambiguity of ANAS as regulator and concessionaire at the same time is not solved at all. The investment plan is the pretext for the increase of the dimension of the “agency”. The national dimension is never discussed. The agent prefers to be the largest possible to avoid competition (also in comparative terms) and to maximise the contracting power. The selection of new investments (in other words, the planning of national network) will move from ministry and public agency to a private-like company, whose goal is profit. The list of projects is no more supposed to come from central planning, at least for the main network. The risk of capture and gold plating is present, both rising unitary costs due to lack of regulation, and allocating resources out of an efficiency-based framework (Beria and Ponti, 2008). 363 Research in Transport and Logistics 2009 Proceedings Do u University NEW RAILWAYS AND OVERCAPACITY The Italian rail network consists of approx. 16.000 km in charge of the national railways (Ferrovie dello Stato S.p.A.), plus some 3.500km of conceded local railways. The main operator, FS, was formerly a vertically integrated public company in charge of the network and with the legal monopoly on services. After the introduction of liberalisation principles, the group has been divided into a service company (Trenitalia) and a network manager (RFI), plus some smaller specific societies10. However, all the companies are still grouped under one single public holding, Ferrovie dello Stato S.p.A.. The financial soundness of FS S.p.A. is scarce. The budget of the company, also including state subsidies, is often performing null or negative operative results (Beria and Ponti, 2008). The liberalisation process in Italy is aligned with Europe, with some specificities. National railways have been unbundled, but tracks and services are still under the same state owned holding. Tracks are a legal monopoly managed by one single company and its tolls are defined via a decree. No subsidy-cap mechanisms exist. No discussion on the minimal efficient dimension of the network took place and the concession is one for the whole country. On goods sector, there is a penetration of new-entrants for approx. 10% in terms of tonn·km. For passengers, Trenitalia is still the monopolist both for long distance and regional services, even if both markets are legally opened (some new rail companies are nearly ready to enter in the market of high speed and some regions performed bids, always won by Trenitalia or by consortia including it). FS group is absent in logistics. Rail investments with no remuneration Historically and presently, all new investments, especially those involving infrastructures, are fully state financed. In general, investment costs are transferred from public purse to RFI, that spends it. These subsidies are ruled by a contract, called “Contratto di Programma”, that includes the list of new projects that RFI will build and the cost for it. The new infrastructures (or new equipments), excluding high speed lines, are never supposed to remunerate the investment, but only the running costs including maintenance. The amortization is not due for the new infrastructures financed by state transfers, except for part of the high speed network. The Italian high speed network is under construction since the 90s11 and will consist until 2010 in the line Turin – Milan – Rome – Salerno, plus some doublings between Milan and Venice. According to public declarations, the main line was initially supposed to cover the 60% of total investments with revenues from fares. Ten years later the line, not yet completed, is costing much more than expected and only 5,1 billions€ out of 24 are financed by equity, all the rest by public transfers (Ponti and Beria, 2007). The average available capacity is described in the plan as “wide”. It is likely that no capacity constraints will exist after the construction of the main HS network Turin-Salerno. Only some terminals or urban sections suffer of congestion and must limit the introduction of new trains. Much more problematic in limiting the services is the lack of availability of appropriate rolling stock. The ratio underlining the investments is contradictory. The HS network under construction is positively seen by FS, because it is the only one capable to host profitable services. The investment cost has been paid for a very small share by the group, but the revenues constitute the largest part of total traffic revenues. Moreover, the HS service is seen as the only one attractive for the users and is thus necessary to improve the image of the company. To the contrary, the largest part of the conventional network (and probably the majority of the HS extensions) is requested by the political bodies, with the exclusion of some urban doublings that are necessary to solve bottlenecks. FS and Trenitalia will not benefit of the extra capacity provided by doublings and new secondary lines, because the present capacity is still largely sufficient. Moreover, the investments had, historically, also the function of hiding the lack of liquid resources. The transfers to FS for the investments constituted until 2000 a source of available money. However, the present definition of such transfers does not allow the “surround” anymore. Finally, it is worth noting that the large capacity now existing on the main commercial network can also be a threat for the public company. The Italian long distance market is substantially opened and potentially very 10 In charge, for example, of commercial revitalisation of major stations (“Grandistazioni”) or of estates development (“FS Real Estate”). 11 With the exclusion of the Florence – Rome direct line dating back to the 80s. 364 Research in Transport and Logistics 2009 Proceedings Do u University attractive, despite the cross-subsidies benefiting FS. Some newcomers are entering in the market and picking the most profitable services only, thanks to the fact that the available capacity is large. AN AIRPORT IN EVERY CITY The majority of Italian airports are managed by concessionaires, with the exception of some minor airports managed directly by the Civil Aviation authority. Often airports are franchised by public companies owned by local authorities. Only few private airports exist, especially among the larger ones. In particular, Rome and Naples airport systems have been privatised and the discussion about the Milan one is frequently at stake, even if not yet decided. Concessions are 40 years long. Regulatory mechanisms have been modified in 2007 (Beria and Ponti, 2008). Both in 2000 and 2007 regulatory frameworks, a point of weakness raises in the issue of investments remuneration. Exactly as for any other regulated sector, investments that improve efficiency and give temporary extra profits to the concessionaires must not be remunerated. The concessionaire will decide independently whether it is worth doing that investment or not, and pay for it. Only when an investment gives a net social surplus, but not a benefit for the concessionaire, it must be publicly financed or paid by users through an increase of tariff. A problem arises when any investment financially feasible at a given increase of tariff is proposed and then approved. This seems to be the situation of Italian airports regulation, where ENAC, that is called to approve investments and consequent tariff increase, do not perform any kind of public evaluation. The procedure usually followed should be the approval of a contract (“contratto di programma”) that may include various new investments, without an explicit estimation of their effect on demand, efficiency and profits12. In these cases there is no distinction between financially feasible investments (decreasing costs or rising demand) to be paid by airports and socially desirable ones: all investments are paid by an automatic increase of tariffs or by public transfers. This fact is giving to the concessionaires a strong incentive to overinvestment. The outcome is that all Italian airports suffer of overcapacity, except Linate and Fiumicino (Sebastiani, 2004). Then, every expansion or investment is not unlocking new demand, before constrained, because no unsatisfied demand exists. A second incentive to develop extra-capacity comes also from the local authorities, that tend to build and operate as much airports as possible13. This is due both to consensus practices (the airport is seen as a necessary tool for regional development), but also because of the expectations of large rents in a period of expanding air market. CONCLUSIONS The present regulatory framework in Italy gives to the agencies many tools to enforce their positions, to increase profits or to minimise efforts. Aside to the “usual” tools described by literature (informative rents in contracts, capture power, dominant positions in the market, cross subsidies, etc.), this paper adds the issue of infrastructural investments. We stress that the incumbent agencies are intensively using the tool of investments and the related informative rents on actual costs, to avoid the effects of regulation, because investments field is weakly regulated. The strategy of the agents is to take control also of the planning procedures and maximise the investments, in order to soften or avoid the regulatory pressure. For the regulated sectors, such as airports or highways, the investments are used to maximise profits thanks to a wrong toll regulation, usually focusing on services and fares. For the less regulated sectors, such as railways, the maximisation of investments has the traditional meaning of increasing the contracting power by expanding the assets. ANAS, being a “newcomer” among the concessionaires, is a fundamental example of the aggressive strategic behaviours in this issue. Their strategy is heavily focusing on the relationship between infrastructural investments and market regulation, exploiting the weaknesses of the normative framework and the absence of an independent transport regulator. The evolution ANAS proposes for itself is to move from being a purely operative agent, becoming a planner and private-like investor. The argument of skipping national debt constraints is used to make attractive the industrial plan to the political decision maker, capturing it. The 12 However, for reasons that cannot be discussed here, no contracts have been signed recently. 13 Italy there are more than 100 airports, some of them really small, and nearly all of them are performing investments and expanding capacity. 365 Research in Transport and Logistics 2009 Proceedings Do u University counterpart of this, is the unlimited expansion of the dimension and power of the former agent that becomes a huge monopolist concessionaire. The Italian highways explicitly benefited of an ineffective and favourable regulation, in order to maximise their financial value. The favourable conditions are based on the issue of the investments, that extend the concession on the basis of development plans proposed by the concessionaire and never assessed in terms of costs and utility by the regulator. The fact that the “regulator” (ANAS) is also a “potential” concessionaire is clearly introducing a conflict of interests: why ANAS should apply a stringent regulation and limit investments, if it plans to apply tolls and maximise their own investments? The statement included in the plan that the two roles will not conflict is not realistic at all. Similar is the case of airport concessions, where the regulator ENAC is not limiting the investments, since they are wanted and paid by local authorities or paid by higher tariffs. The capture of it may be supposed. The airport companies, on their side, do not have any incentive to minimise investments, since the more facilities they build, the higher are the tariffs and the profits. Railway sector is not lucrative and not regulated. For these reasons the use of investment is still limited to the expansion of the assets and to the tentative of surviving the competition with air and private car. The general policy indication, discussed somewhere else (Beria and Ponti, 2008) is to subtract the topic of investments remuneration to the concessionaires, with the exclusion of the endogenous ones. BIBLIOGRAPHICAL REFERENCES [1] ANAS (2006), Bilancio 2006, ANAS, Roma (Italy). [2] Ponti M., Beria P., Erba S. (2007), Una politica per i trasporti italiani, Laterza, Roma/Bari (Italy). [3] Beria P., Ponti M. (2008), The role of regulation in financing transport infrastructures in Italy, SIET Conference 2008, Sassari (Italy). File available at: http://www.sietitalia.org/siet2008/PaperB/B2.pdf [4] Boitani A. (2004), La regolazione delle autostrade italiane tra Piano Generale dei Trasporti e privatizzazioni, paper presented at the workshop “Highways”, Bergamo (Italy), November, 27 2004. [5] Coco G. and Ponti M. (2006), Riflessioni per una riforma della regolazione nel settore autostradale in De Vincenti C. e Vigneri A. (2006, edited by), Le virtù della concorrenza, Il Mulino, Bologna, 185-228. [6] Ministero dell’Economia e delle Finanze (2008), Politica economica e finanziaria. Glossario di due anni, Ministero dell’Economia e delle Finanze, Roma. [7] Ponti M., Beria P. (2007) “La rotaia arrugginita e il vagone del futuro”, Il Mulino, No. 6, p. 1028-1041. [8] Ponti M., Beria P., Erba S. (2007), Una politica per i trasporti italiani, Laterza, Roma/Bari (Italy). [9] Ponti M., Boitani A. (2005), “Autostrade, galline dalle uova d'oro e polli da spennare”, laVoce.info, 18/11/2005 [10] Ragazzi G. (2008, forthcoming), I signori delle autostrade. [11] Ragazzi G. and Rothengatter W. (2005, edited by), Procurement and Financing of Motorways in Europe, proceedings of the conference “Highways: Costs and Regulation in Europe”, Bergamo, 26/11/2004, Elsevier, Amsterdam (NL). [12] Sebastiani M. (2004), “Perché Alitalia resta a terra”, laVoce.info, 01/04/2004. 366 Research in Transport and Logistics 2009 Proceedings Do u University FREE TRADE AGREEMENTS IN THE MEDITERRANEAN REGION: A BOX-COX ANALYSIS Matthew KARLAFTIS1, Konstantinos KEPAPTSOGLOU 2 and Dimitrios TSAMBOULAS 3 Abstract Free Trade Agreements (FTA) have been tools widely implemented for enhancing trade between countries. As a result, various models have been developed in the past, in an effort to explain the effects of these agreements in trade. The objective of this study is to develop a model suitable for exploring FTA effects in trade flows in the Mediterranean region. For that purpose, a model is developed for analyzing trade flows based on the family of gravity models; the model parameters are estimated by using a Box – Cox transformation, in an effort to overcome the assumptions of the linear model typically used in similar cases. Results of the derived models are discussed and indicate that transportation costs have a dominant effect in trade; agreement effects are of a lower magnitude and can be negative in cases of specific commodities, whose volume may significantly be affected by other, external factors, as well as the mode used for transporting them. Keywords: Free Trade Agreements, Mediterranean, Gravity Model, Box-Cox Transformation INTRODUCTION Free trade agreements (FTAs) are forms of trade pacts between countries on a bilateral or multilateral basis; these agreements eliminate tariffs, quotas and other barriers, for a number of goods (if not all), traded between involved partners. The aim of FTAs is to increase trade between countries as a result of the relaxation or removal of existing institutional and economic barriers between them. FTAs have been a tool widely implemented for enhancing trade between countries, with examples such as the North American Free Trade Agreement between the US, Canada and Mexico, the Free Trade Agreement of the Americas (FTAA), the ASEAN FTA between eastern Asian countries and the Euro-Mediterranean Free Trade Agreement (EMFTA). In particular, EMFTA has been EU’s instrument for establishing FTAs since the 1995 Barcelona Declaration set the principles for establishing FTAs between the EU and other Mediterranean countries in the years to follow [1], [2], [3]; non-EU Mediterranean countries interested in the agreement are expected to adopt it by no latter that 2010. The interest of the research community in the effects of FTAs to trade has been considerable. For example, all Mediterranean countries will be part of the EMFTA agreement in 2010 and therefore analysis of EMFTA’s effects so far is invaluable for examining the agreement’s policy and potential. As a result, a large number of existing studies have attempted to model and analyze the effects of FTAs in trade flows and trade in general. In this context, the objective of this paper is to develop a model suitable for exploring EMFTA effects in trade flows in the Mediterranean region. The paper is structured as follows: the next section reviews modeling approaches for analyzing FTA effects. Based on that review, the proposed modeling approach is described in the subsequent sections: the problem, data and model are described and results are discussed. The paper concludes with a summary of the study’s findings. BACKGROUND Trade flow modeling in general has been widely discussed in the last three decades; most efforts for expressing and analyzing trade flows have concentrated (a) on simulation models, aiming at replicating that phenomenon and its impacts, and (b) on econometric models, attempting to make predictions based on past, actual performance [4]. Simulation models capture the underlying structure of trade flows (comprising of activities such as production, consumption and transportation) [5], respond to inputs and estimate potential impacts to trade flows. Commonly, input-output (I/O) and generalized equilibrium models have been 1 Matthew G. Karlaftis, Assistant Professor, School of Civil Engineering, National Technical University of Athens, Greece, mgk@central.ntua.gr 2 Konstantinos Kepaptsoglou, PhD Candidate, School of Civil Engineering, National Technical University of Athens, Greece, mgk@central.ntua.gr 3 Dimitrios Tsamboulas, Professor, School of Civil Engineering, National Technical University of Athens, Greece, dtsamb@central.ntua.gr 367 Research in Transport and Logistics 2009 Proceedings Do u University exploited for simulating trade flows with the later having the advantages of calculating trade flows endogenously and incorporating transportation costs in the model [5]. On the other hand, econometric approaches for modeling trade flows have been largely based on the gravity model specification; since its introduction by Timbergen [6] and Linneman [7] the gravity model has demonstrated a considerable empirical robustness in describing trade flows [8], [9]. As reported by Porojan [9], in the last decade, gravity models have been employed in numerous studies for analyzing and assessing trade flows. Indeed, the literature reveals a considerable number of publications, either offering modeling improvements and refinements or attempting to explain policy impacts to trade. As noted earlier, regardless of the method used, FTA effects in trade have been widely examined in the literature, often with controversial conclusions. Among econometric based studies of the last decade, Endoh [10] found that the Latin American Free Trade Agreement (LAFTA) has exhibited neither trade creation nor trade diversion on trade with Japan. On the other hand, Fukao et al. [11] provided evidence of some trade diversion as a result of the North American Free Trade Agreement (NAFTA). Roberts [12] investigated the potential of a FTA between China and ASEAN countries; he concluded that neither trade creation or diversion were expected. Eger [13] indicated that while FTAs are not expected to have a short-term impact on trade volumes, a considerable long-run trade creation is anticipated; he reported a 15% long-term increase for NAFTA members. An analysis of trade agreements in African countries (COMESA, ECCAS, ECOWAS) by Musila [14] did not find any considerable impacts in trade diversion and creation; similar conclusions were drawn for the COMESA by Rojid [15] and for the AGADIR agreement by Peridy [16], as a result of the lack of trade complementarity between AGADIR member countries. Tang [33] examined the effects of the NAFTA, ANZCER and ASEAN FTAs and reported that trade within member countries has increased, ANZCER FTA has resulted in trade diversion from non-members and ASEAN FTA has led to a trade increase with non-members (something that has not been observed for NAFTA). Peridy [17] investigated trade effects of the Euro-Mediterranean FTA (EMFTA); he reported that the FTA resulted in an increase of Mediterranean countries’ exports to the EU by 20%-27%, indicating trade creation and accounted for the large EU share of Mediterranean exports. Carrere [18] concluded that “…regional trade agreements generate a significant increase in trade between members, often at the expense of the rest of the world”. Lee and Park [19] proposed new FTAs for East Asia; they noted that trade facilitation enhances trade creation between FTA members and reduces trade diversion among them and indicated that their proposed FTAs would be beneficial compared to existing conditions. Kalijaran [20] reported that Australia is expected to have more gains of its potential exports because of the IOC-ARC agreement. As for simulation approaches, in a mid-1990s study, Breston et al. [21] simulated a potential trade agreement between EU and Russia and concluded its infeasibility if no other CIS countries were included in the same agreement. Venditto [1], examined the practicability of EMFTA and reported that its benefits were uncertain, expected to be obtained in the long run and did really modify existing relationships between countries. On the contrary, another two studies on EMFTA effects by Augier and Gasiorek [22], [23] indicated a substantial impact on non-EU member countries, especially in production and welfare. McQueen [2] investigated the FTAs between EU and developing countries; he reported significant potential gains for EU but less clear benefits for the partner countries, as a result of limited product coverage and rules of origin. Francois et al. [24] arrived at the same conclusions with McQueen [2]. The FTA of the Americas (FTAA) was analyzed by Brown et al. [25] using a CGE model. The authors found that (a) FTAA increases the economic welfare of members (with the largest share to the USA) and (b) FTAA is trade diverting for non-members. Finally, Siriwardana [26] investigated the FTA between Australia and the USA and reported growth in exports between the two countries. FTA IN THE MEDITERRANEAN REGION The EU has been pursuing trade agreements with non-member Mediterranean countries since the mid 1950s; early agreements between the former European Economic Community and Morocco, Algeria and Tunisia were established in the treaty of Rome, in 1957 [3]. Few years later (1960), the EFTA agreement was initiated, ensuring free trade between EC members. Only in the early 1990s though, the EU commenced a comprehensive trading policy in the Mediterranean area which resulted in the Barcelona Declaration, in 1995; that declaration set the principles for developing FTAs between the EU and other Mediterranean countries in the years to follow [1], [2], [3]. The instrument introduced for achieving these objectives was the Euro- Mediterranean Free Trade Agreement (EMFTA), which is expected to be fully adopted by the EU and other 368 Research in Transport and Logistics 2009 Proceedings Do u University interested Mediterranean countries by 2010. An update of the status of EMFTA agreements up until the early 2000’s is given by McQueen [2]; most south Mediterranean countries have adopted the EFMFTA agreement by the early 2000s. Mediterranean countries should enter the EMFTA agreement by 2010 and for that reason the analysis of the already established FTAs and their effects to trade flows is invaluable for examining the actual impact of the EMFTA policy so far, along with its potentials for the years to come. Based on the review results, gravity models are widely used for examining FTA effects; in this study such a specification is used for the case of the Mediterranean region, to capture the effects of the EMFTA agreement in trade between potential EMFTA partners. MODEL SPECIFICATION While being largely empirical, gravity models have been successfully used in the past for describing trade flows. Their concept is based on Newtonian physics; trade between two partners is affected by their sizes and proximity [9]. In particular, a flow of goods Fij between two areas i and j is expressed as a function of the characteristics of the origin (Oi) and the destination (Dj) and some measure of impedance between them (Rij): Fij Oi D j Rij (1) Equation (1) can be modeled as a linear function by taking its logs: F ' 2ij LogFij , 0, (2) where X: Vector containing the logs of explanatory variables : Vector of parameters to be estimated : error term Typically, explanatory variables include proxies for the sizes of the economies of the two partners (for example GDP, population etc), and their proximity (distance, transportation cost and other barriers to trade). Based on the state-of-the-art review, the modeling needs of the study and data availability, the specification used analyzed in this paper is of the following form: log Fijc a0 a1 logTEic a2 logTI jc a3 logTFc a4 GAFTA a5 AGADIR a6 EU a7 EMFTA a 8 logTC (3)ij where i: Origin j: Destination c: Commodity (c=1,2,3,4,5) Fijc: Flow of commodity c between origin i and destination j TEic: Total exports of commodity c from origin i TIjc: Total imports of commodity c by destination j TFc: Tariffs for commodity c TCij: Transportation cost between i and j GAFTA: Dummy variable indicating a GAFTA agreement between i and j AGADIR: Dummy variable indicating an AGADIR agreement between i and j EU: Dummy variable indicating an EU/EEA agreement between i and j EMFTA: Dummy variable indicating an EMFTA agreement between i and j a0,…,a8: Model coefficients As mentioned earlier, commodities are grouped in five categories, namely 1 to 5. Total exports and imports for each commodity act as mass variables, indicating the supply and demand for that commodity and therefore the attraction between the origin and the destination of the flow. Transportation costs, tariffs and existence of trade agreements on the other hand quantify the friction anticipated to be experienced for the flow of commodities between i and j. It has to be noted that while the different gravity models found in the literature exhibit a variety of explanatory variables, this paper focuses on selecting those factors that can capture the free trade agreement effects between two countries. 369 Research in Transport and Logistics 2009 Proceedings Do u University DATASET DEVELOPMENT A necessary step towards developing the model was to construct an appropriate dataset for the Mediterranean region. A dataset including data for all 48 EMFTA participants (EU member states and other partners) was set up for the years 1992-2006 (the EMFTA process was initiated in 1995). Participating countries were aggregated into 21 zones, shown in Figure 1 (blue boundaries); aggregation was necessary to reduce estimation of transportation costs and therefore modeling efforts into an acceptable level. The SITC3 commodity classification was adopted. Again, in order to reduce modeling efforts without compromising forecasting reliability, the ten Standard International Trade Classification (SITC3) 1-digit classes have been grouped into five macro-classes, presented in Table 1. Aggregation of classes aimed at producing homogeneous macro-classes and balancing the number of records among those classes. FIGURE 1 EMFTA zones considered in the study TABLE 1 SITC aggregation into macro-classes SITC3 1-digit code Product Description Macroclasses 0 Food and live animals 1 2 Crude material ex food/fuel 2 5 Chemicals / Products n.e.s. 2 3 Mineral fuel / lubricants 3 6 Manufactured goods 4 7 Machinery / Transp equipment 4 8 Miscellaneous manuf arts 4 Trade flows between zones are the model’s dependent variables, expressed in US $, for a single year and a given commodity group. The COMTRADE database was exploited for that purpose (http://comtrade.un.org/). COMTRADE provides both import and export data for each pair of countries. Import and export values are usually different for a number of reasons, the main being the fact that exports are usually recorded fob (free- on-board) while imports are recorded as cif (cost, insurance and freight). Moreover, imports are usually recorded more accurately than exports because since only the former generate tariff revenues. For that reason, 370 Research in Transport and Logistics 2009 Proceedings Do u University in the dataset analyzed, import values are used for trade flows. The geographical and commodity classification coverage provided by the COMTRADE database is satisfactory for the European Countries and sufficient for the remaining. In the gravity model mass variables are expected to capture the impact of the magnitude of origin and destination zones in explaining the related trade flows. GDP values are normally used for this purpose and therefore, for each of the 48 countries within the study area, such information has been collected with the main sources being the EUROSTAT, the Arab Monetary Fund and national statistics bureaus. All GDP data are expressed in US $. Moreover, for modeling purposes, it is worth exploring to what extent a more proper choice of the mass variables can increase the model goodness-of-fit. In detail, it seems natural to adopt the total import and the total export trade of a zone as mass variable respectively of trade destination and origin. Therefore such variables have been also included into the estimation database. The basic impedance variable adopted is the generalized cost for transporting goods between zones. A typical proxy for transportation costs used in relevant studies is distance between zone centroids; however in this study detailed transportation costs and tariffs are explicitly used. Tariffs were extracted from the TRAINS database (www.unctad.org/Trains); the database included three types of tariffs: MFN (most favored nations): those are the nominal tariffs applied by World Trade Organization member states to other countries, unless preferential agreements are in force; PRF (preferential rates): usually lower than MFN tariffs, represent the tariffs nominally applied among countries with preferential agreements in force; AHS (effectively applied tariffs): when available, they denote the tariffs effectively applied to trade between two countries. Tariffs can be expressed as percentage of the value of the imported goods or calculated by complex rules, for instance an increasing-by-step tariff (i.e. zero tariffs for trade up to a certain volume threshold and then tariffs increasing with trade volume itself). Furthermore, non ad-valorem tariffs can be applied as well, for instance related to the quantity of the good imported. The WITS TRAIN database allows overcoming this issue by considering an ad-valorem equivalent (AVE) tariff, which turns each type of non ad-valorem tariff in a corresponding ad-valorem equivalent. For the dataset of the study, the AVE tariffs for the lowest between MFN, PRF and AHS have been taken into account and tariffs were expressed as percentages of the traded value. With respect to commodity aggregation, tariffs were weighted for each commodity subgroup, using the corresponding subgroup trade volume as weight. A detailed transportation cost process was adopted; this is described in detail in [27]. For each pair of geographical entities, i.e. NUTS2 zones for European countries and national level for the remaining countries, travel times and costs are available for road, rail (traditional and combined) and shipping (Ro-Ro and containerized) modes. Zones for estimating the transportation costs are indicated in Figure 1 with green boundaries. In brief, based on available data and past research results, detailed transportation cost functions were developed, for calculating transportation costs of shortest paths between entity centroids. These were later weighted with each entity’s GDP in order to provide transportation costs between demand zones. A detailed description of the transportation costs is provided in [28]. Free trade agreements in the study area were captured with the use of dummy variables; the agreements explicitly considered in the model were the European Union membership and the EFTA and EMFTA agreements already in existence in the Mediterranean. Regarding the EFTA agreement, for years from 1992 to 1994 it has been set equal to one for the following group of countries: Austria, Finland, Iceland, Norway, Sweden and Switzerland. Some of these countries joined EU in 1995 and therefore, since that year EMFTA dummy has been set in force only between Iceland, Norway and Switzerland. In conclusion, the structure of both aggregated and disaggregated database consists of the following fields: reporting zone (i.e. the destination of the flow), partner zone (i.e. the origin of the flow), year, commodity class, trade value ($000), tariffs (weighted average by commodity), dummy EFTA agreement, dummy EU/EEA agreement, dummy EMFTA agreement, reporting zone total GDP, partner zone total GDP, reporting zone total imports by commodity, partner zone total exports by commodity, transportation costs. MODEL ESTIMATION TECHNIQUE As previously discussed, the goal of the present paper is to examine the relationship between trade flows, attraction and impedance variables . To date, such investigations have been performed in either of two ways; 371 Research in Transport and Logistics 2009 Proceedings Do u University with the use of linear regression, or with the use of neural networks. While such investigations have yielded results that are quite satisfactory, they suffer in that when neural networks are used only predictions of the dependent variable (i.e. total trade value between two regions) can be made; that is, there can be no meaningful inference regarding the underlying relationship between trade flow and exogenous factors (independent variables) and, as such, policy implications cannot be extracted. Further, when linear regression is used, the implicit hypothesis is made that the underlying function is linear-in-the-parameters; this is a hypothesis that can be, neither theoretically nor empirically, justified. To overcome these apparent shortcomings, the present paper develops a Box-Cox regression model. In essence, the Box-Cox methodology is used as a method of generalizing the linear model; the basic transformation is [29]: x 1 x (4) In a regression model, this transformation could be done conditionally, where for a given value of , the model becomes K y a k xk (5) k 1 This transformation leads to a linear regression that can be estimated via least squares [30]. Basic model theory allows for each regressor to be transformed by a different value of , but this level of generality becomes excessively cumbersome; as a result, is assumed to be the same for all variables in the model. If in Eq. (5) is considered an unknown parameter, then the regression becomes nonlinear-in-the-parameters. And, although no transformation will reduce it to linearity, nonlinear least squares are straightforward. As a general rule-of-thumb, is typically between –2 and 2. Interestingly, when equals zero and using L’ x 1 Hospital’s rule, lim ln x . To obtain the values of a, , from Eq. (5), right-hand side derivatives of 0 this equation (with respect to these parameters are needed); these are given as h . 1, a h . xk , (6) k h . K x Kk 1 k k xk ln xk xk k 1 k 1 Of course, it needs to be noted that, from a practical perspective, the parameters of the nonlinear model are not equal to the slopes with respect to the variables, as is the case in regression. As such, the slopes (or equivalently the marginal effects and elasticities) can be given as [31]: x 1 ln y a (7) d ln y x d ln x The extension of Eq. (5) to include a transformation of the dependent variable (y) as well as a transformation of the independent variables is rather straightforward and can be given as K y a k xk k 1 or (8) y x Again, this form includes the linear and log-linear functional forms as limiting cases. Allowing to differ from increases the flexibility of the model. Nevertheless, for clarity of presentation, it is assumed 372 Research in Transport and Logistics 2009 Proceedings Do u University here that these two values are equal. Further, assuming that ~ N 0, 2 , then the log-likelihood for a sample of n observations can be written n n n 1 n 2 ln L ln 2 ln 2 1 ln yi y x (9) 2 2 i 1 2 2 i i i 1 The maximum likelihood estimator of 2 will be the average squared residual, making the last term of Eq. (5) n 2 , collapsing the log-likelihood into: n n ln L 1 ln yi ln 2 1 ln ˆ 2 (10) i 2 2 Model Results The dependent variable considered in this estimation includes the total value of the trade between pairs of regions for four commodity categories; commodities are categorized as (a) food and live animals, (b) crude material (except fuel) and chemical products, (c) mineral fuel and lubricants and (d) various manufactured goods. In this way, it is possible to examine numerous parameters (such as total import and export trade between the regions, transport cost and tariffs and the effect of trade agreement on total values) that might have a significant effect on various trade flow categories. Results are presented in Table 2: TABLE 2 Box-Cox Gravity Model Parameters Commodity Class Model Specification 1b 2c 3d 4e Explanatory variables Coefficient Coefficient Coefficient Coefficient(t-statistic) (t-statistic) (t-statistic) (t-statistic) Total Commodity 0.557 0.509 1.819 3.755 Imports (21.93) (22.651) (7.578) (5.988) Total Commodity 0.661 0.741 1.628 4.38 Exports (28.557) (36.433) (8.080) (6.116) Tariffs -0.522 -0.625 -0.500 -0.29 (-1.817) (-1.705) (-1.975) (-1.637) Transport Costs -1.353 -1.475 -3.918 -1.685 (-5.208) (-5.638) (-6.645) (-3.752) EU/EEA agreement ns -1.068 (-4.628) ns ns EMFTA agreement -1.331 -0.813 0.4422 0.597 (-4.312) (-3.865) (1.912) (2.671) f 1.895 1.789 1.684 1.895 (21.41) (21.125) (22.270) (20.515) f 2.01 1.14 (3.17) (3.3650 = = 2 0.51 0.32 0.61 0.35 a Estimated using Box – Cox transformation d Mineral fuel and lubricants b Food and live animals e Various manufactured goods c Crude material (except fuel) and chemical products f Parameters from Eq. (4) Based on these results, three important observations regarding the estimated models can be made: It can be observed that for most models statistically significant results are obtained; that is, there are independent variables whose influence on the dependent variable is important and are, by and large, consistent with the general directions of previous findings in the literature. In total, four separate models were estimated, each corresponding to a different category of goods; Box- Cox transformations were required for both the dependent and independent in all four models. However, for the first two models, the and parameters are different, while for models three and four, both the dependent and independent variables were transformed by the same Box-Cox parameter. It must be noted 373 Research in Transport and Logistics 2009 Proceedings Do u University that both parameters ( and ) were allowed to be different from one only when a likelihood ratio test suggested this to be statistically acceptable. All models offer a rather good fit to the data. 2, the nonlinear equivalent of R2 in linear regression, takes on values ranging from 0.32-0.61 which is considered to be good fit for nonlinear models. According to Table 2 results, total commodity exports and imports have a strong positive correlation with trade flows for all commodity categories; this is expected since these variables are direct representations of the demand and supply for each commodity type. Therefore, their selection instead of proxies such as the GDP is justifiable. Tariffs seem to have an impact on trade flows of commodity categories, particularly on categoriew 1,2 and 3. Their negative sign is expected since tariffs pose a barrier to trade; however in all cases their impact is low. A possible explanation is the existence of FTAs; large part of the dataset covers a period and set of countries already participating in such agreements or custom unions (the EU). As a result, such barriers are gradually eliminated and play a lesser significant role in trade flows. Transport cost impacts are negative and high in most cases; similarly to other studies examining the EMFTA agreement [16], [32], transportation costs have a considerable impact in flows. The EMFTA agreement has an impact to all five commodity categories; interestingly that impact in the cases of commodity categories 1 and 2 is negative. This is probably due to the particular nature of the goods within those commodities, grouping mainly crude material and mineral fuel: it is well known that the trade volume of such commodities strongly depends on external factors highly variable in small time periods. Therefore, the possibility that the EMFTA agreements were put in force in a period of decreasing trend in trade implies that the model captures this trend through a negative sign of EMFTA agreements for those commodities. From one side, this would suggest introducing for those commodities further explanatory variables within the model specification. However, those commodities make mainly use of very specific transport modes (e.g. pipelines) and therefore they can be considered as marginal with respect to the demand segments impacting on transport services to be analyzed. Moreover, with respect to category 1, restrictions on food and live animals posed by the EU even through the EMFTA agreement, still pose barriers to trade for these commodities. CONCLUSIONS The objective of this paper was to review modeling efforts for analyzing effects of Free Trade Agreements to trade flows and to develop an econometric model for analyzing the effects of such agreements to trade in the Mediterranean region. An econometric gravity model specification was developed for that purpose and the model parameters were estimated using a Box – Cox approach. Results indicated that transportation costs have a dominant effect in trade; agreement effects are of a lower magnitude and can be negative in cases of specific commodities, whose volume may significantly be affected by other, external factors, as well as the mode used for transporting them. The model is capable of providing useful insights to decision makers regarding the characteristics and details of FTAs in the Mediterranean region. ACKNOWLEDGEMENTS This work is part of the Interreg IIIb Archimed project “FREEMED – New Mobility Scenarios in the FREE trade zone in the MEDiterranean basin”, co-funded by the European Commission and National funds REFERENCES [1] Venditto, B., 1998, The Euromediterranean Free Trade Area: A New Form of Regional Cooperation. Proceedings of the Encounters of the Mediterranean Economists, Split, Croatia. [2] McQueen M., 2002, The EU’s Free-trade Agreements with Developing Countries: A Case of Wishful Thinking? The World Economy, 25(9), 1369-1385, 2002. [3] Siousiouras P., 2003. The Euro-Mediterranean Free Trade Zone: Prospects and Possibilities. Mediterranean Quarterly, 14(3), 112-121. [4] The World Bank, 2005. Global Economic Prospects 2005, 57-76. [5] Transek AB, CERUM, INRO, TOI, NEA, 2001. Ideas for a new Swedish Freight Model. Report, Norway. [6] Timbergen, J., 1962. Shaping the World Economy, Twentieth Century Fund, New York. [7] Linneman, H., 1966 The Gravity Equation in International Trade: Some Microeconomic Foundations and Empirical Evidence. The Review of Economics and Statistics, 67(3), 474-481. 374 Research in Transport and Logistics 2009 Proceedings Do u University [8] Bergstrand, J.H., 1985. The gravity equation in international trade: some microeconomic foundations and empirical evidence, Review of Economics and Statistics, 67. 474–481. [9] Porojan A., 2001. Trade Flows and Spatial Effects: The Gravity Model Revisited. Open economies review, 12, 265-280. [10] Endoh M., 1999 Trade creation and trade diversion in the EEC, the LAFTA and the CMEA: 1960-1994. Applied Economics, 31, 207-216. [11] Fukao K., Okubo T., Stern R.M., 2003. An econometric analysis of trade diversion under NAFTA. North American Journal of Economics and Finance. 14, 2-24. [12] Roberts B.A., 2004. A gravity study of the proposed China-Asean free trade area. The International Trade Journal, 18(4), 335-353. [13] Egger P., 2002. An Econometric View on the Estimation of Gravity Models and the Calculation of Trade Potentials, The World Economy, 25(2), 297–312. [14] Musila J.W. , 2005.The Intensity of Trade Creation and Trade Diversion in COMESA, ECCAS and ECOWAS: A Comparative Analysis”. Journal of African Economies, Vol. 14, No. 1, 117-141. [15] Rojid S., 2006. COMESA trade potential: a gravity approach”. Applied Economics Letters, 13, 947-951. [16] Peridy N., 2005. The trade effects of the Euro-Mediterranean partnership: what are the lessons for ASEAN countries? Journal of Asian Economics, 16, 125-139. [17] Peridy N., 2005 Toward a Pan-Arab free trade area: assessing trade potential effects of the Agadir agreement. The Developing Economies, XLIII-3: 329-45. [18] Carrere C., 2006. Revisiting the effects of regional trade agreements on trade flows with proper specification of the gravity model. European Economic Review, 50, 223-247. [19] Lee H. and Park I., 2007. In Search of Optimised Regional Trade Agreements and Applications to East Asia. The World Economy, 783-806. [20] Kalirajan K., 1999. Stochastic varying coefficients gravity model: an application in trade analysis. Journal of Applied Statistics, Vol. 26, No. 2, 185-193,. [21] Breston P., Tourdyeva N. and Whalley J., 1997. The Potential Trade Effects of an FTA between the EU and Russia. Weltwirtschaftliches Archiv, Vol. 133 (2). [22] Augier P. and Gasiorek M., 2003. The welfare implications of trade liberalization between the Southern Mediterranean and the EU. Applied Economics, 35, 1171-1190. [23] Augier P., Gasiorek M., Lai Tong C., 2005. Rules of Origin-The impact of rules of origin on trade flows. Economic Policy, 567-624. [24] Francois J.F., McQueen M. and Wignaraja G., 2005. European Union-Developing Country and Analysis. World Development, Vol. 33, No. 10, 1545-1565. [25] Brown D.K., Kiyota K., Stern R.M., 2005. Computational analysis of the Free Trade Area of the Americas (FTAA). North American Journal of Economics & Finance, 16, 153-185. [26] Siriwardana M., 2007. The Australia-United States Free Trade Agreement: An economic evaluation. North American Journal of Economics & Finance, 18, 117-133. [27] Regione Campania, NTUA, Regione Calabria, CERTH, ADEP, 2007. Decision Support System Implementation, Deliverable D4, Project FREEMED – Interreg III B Archimed, EU. [28] Regione Campania, NTUA, Regione Calabria, CERTH, ADEP, 2007. Transportation Supply Model, Deliverable D2, Project FREEMED – Interreg III B Archimed, EU. [29] Box, G. and Cox, D., 1964. An analysis of transformations. Journal of the Royal Statistical Society Series B, 211- 264. [30] Seaks, T. and Layson, K., 1983. Box-Cox estimation with standard econometric problems. Review of Economics and Statistics, 65, 160-164. [31] Spitzer, J., 1982. A fast and efficient method for estimation of parameters in models with the Box-Cox transformation. Journal of the American Statistical Association, 77, 760-766. [32] Kandogan Y., 2005. Evidence for the Natural Trade Partners Theory from the Euro-Mediterranean Region. University of Michigan-Flint School of Management. Working Paper Series, No.2005-01. [33] Tang D., 2005. Effects of the Regional Trading Arrangements on Trade: Evidence from the NAFTA, ANZCER and ASEAN Countries, 1989-2000. The Journal of International Trade & Economic Development, 14(2), 241- 265, 2005. 375 Research in Transport and Logistics 2009 Proceedings Do u University AUTHOR INDEX , Berrin 37 B. 277 AKIN, 43, 95 KARLAFTIS Matthew 367 AKTA , Emel 37 KEPAPTSOGLOU Konstantinos 367 ALKAYA Ali Fuat 167 KOP 351 ALVANCHI Amin 211 KOPSACHEILI, Angeliki 21 173, 219 , Ivan 29 AZIMI Reza 211 KU Burcu 285 BABALIK-SUTCLIFFE, Ela 89 KUMRU Mesut 203 BASKAN Ozgur 231, 241 Orhan 225 , 67 LYNCE Ana-Rita 291 BAYRAKTAR Demet 219 149, 189 BEELEN Marjan 319 MAUNDER, David 59 141 MEERSMAN Hilde 319 BERIA Paolo 351 Romeo 155 BOZKAYA, 173 Feliciana 299 , n 73 MORAITI Panayota 313 53 MORVILLO, Alfonso 79 CEBI, Selcuk 108 Gholam Reza 277 CELIK, Metin 105, 108 Senay 268 CEYLAN Halim 231, 241 ONGEL Aybike 345 CEYLAN Huseyin 231, 241 ONGHENA Evy 319 CICEK, Kadir 105 , 67 , Emir 29, 155 , 37 CLAUSEN Uwe 251 , 133 , 73 IR, Vildan 67 Dilay 219 POLYDOROPOULOU Amalia 179 , Mehtap 95 POSACI, Evren 43 , Nihan 125 REIS Vasco 149 Selmin 219 RONGVIRIYAPANICH Terdsak 307 DAVOUDPOUR Hamid 277 ROTHENGATTER, Werner 3 DE KERVENOAEL, Ronan 173 SCOPEL Elena 337 DE MARTINO, Marcella 79 Bahar 285 , Branislav 29, 155 SILLA Anne 291 - , Vesna 29, 155 SILVA Jorge 189 DUMAN Ekrem 167 , Maja 29 133 SRIMUANG Kawee 307 ERENSAL Yasemin Claire 161 Mehmet 257 EROL GENEVOIS 141, 351 TOSUN 225 UYSAL Fahriye 225 TSAMBOULAS Dimitrios 21, 313, 367 GROSSO Monica 291, 299 TSIRIMPA Athena 179 GSCHWENDER Antonio 197 TUZKAYA 268 GUIHERY Laurent 327 ULUKAN H. Ziya 351 HALDENBILEN Soner 231, 241 , 37, 117 HARVEY John 345 VAGGELAS Georgios K. 291 IDDINK Ute 251 VAN DE VOORDE Eddy 319 JARA- Sergio R. 197 VANELSLANDER Thierry 319 155 YANIK Seda 257 KABAK 117 Y , G.Nilay 125 379