Résumé: The search process in population-based metaheuristic algorithms (MAs) can be classified into two primary behaviours: diversification and intensification. In diversification behaviour, the search space will be explored considerably based on randomization. Whereas intensification alludes to the search for a promising region locally. The success of MAs relies on the balance between two search behaviours. Nonetheless, it is strenuous to get the right balance between these behaviours due to the scholastic nature of MAs. Chaotic maps are proven an excellent tool to enhance both behaviours. This work incorporates the Logistic chaotic map into the recently proposed population-based MA called Electromagnetic field optimization (EFO). This suggested algorithm is named chaotic EFO (CEFO). An improved diversification step with chaos in EFO is presented to efficiently control the global search and convergence to the global best solution. CEFO is tested on different case studies, 40 unconstrained CEC 2014 and CEC 2019 benchmark functions, seven real-world nonlinear systems and three mechanical engineering design frameworks. All experiments are compared with other recent and improved algorithms in the literature to show the performance and effectiveness of the proposed algorithm. Two nonparametric statistical tests, the Wilcoxon rank-sum and the Friedman test, are performed on CEFO and other compared algorithms to determine the significance of the results and show the efficiency of CEFO over other algorithms.

Langue: Anglais