Performance Assessment of Natural Survivor Method-Based Metaheuristic Optimizers in Global Optimization and Engineering Design Problems

Abstract

This study presents the comparative performance analysis of Natural Survivor Method (NSM)-based algorithms in solving the IEEE CEC 2022 test suite benchmark problems and four real-world engineering design problems. Three different variants (Case1, Case2, Case3) of the NSM-TLABC, NSM-SFS and NSM-LSHADE-SPACMA algorithms were used in the study. The data obtained from the experimental studies were statistically analyzed using Friedman and Wilcoxon signed-rank tests. Based on the Friedman test results, NSM-LSHADE-SPACMA_Case2 showed the best performance with an average Friedman score of 3.96. The Wilcoxon signed-rank test showed that NSM-LSHADE-SPACMA_Case2 outperformed its competitors in 13 out of 16 experiments, achieving a success rate of 81.25%. NSM-LSHADE-SPACMA_Case2, which was found to be the most powerful of the NSM-based algorithms, is used to solve cantilever beam design, tension/compression spring design, pressure vessel design and gear train design problems. The optimization results are also compared with eight state-of-the-art metaheuristics, including Rime Optimization Algorithm (RIME), Nonlinear Marine Predator Algorithm (NMPA), Northern Goshawk Optimization (NGO), Kepler Optimization Algorithm (KOA), Honey Badger Algorithm (HBA), Artificial Gorilla Troops Optimizer (GTO), Exponential Distribution Optimization (EDO) and Hunger Games Search (HGS). Given that all results are together, it is seen that NSM-LSHADE-SPACMA_Case2 algorithm consistently produced the best results for the global and engineering design problems studied.