Hybrid leader-follower and fuzzy-Petri-net traffic rate control and supervision in network systems

Abstract

This paper addresses networking and traffic control problems in network systems along with the potential for introducing soft-computing applications at supervisory control level. The incentive Stackelberg strategy control concept was introduced to the networking model that comprises subsidiary systems of users and of network. A linear strategy is proposed for the elastic traffic problem, and results are demonstrated and illustrated via Kelly's (1997) example. The presented method is extended to the non-elastic traffic problem and results are also illustrated via a slightly modified Kelly example. Ways of employing combined analytical and soft-computing approaches and techniques in a hybrid framework have been investigated. A fuzzy-Petri-net supervisor is proposed to implement the strategic decision and control layer. Insofar obtained results are rather encouraging, and this research continues further towards possible implementation as well as some open theoretical issues