A Study on Multi-objective Chaotic Evolution Algorithms Using Multiple Chaotic Systems

Abstract

We investigate the optimization performance of multi-objective chaotic evolution (MOCE) algorithm with implementations using different chaotic systems. A comparison experiment of MOCE algorithms with four chaotic systems are employed in MOCE to analyse whether chaotic systems will affect the optimization performance of the MOCE algorithms. We analyze and discuss the performance of the MOCE algorithms implemented using different chaotic systems. Four chaotic systems are introduced in this work, i.e., the logistic map, the Hénon map, the tent map, and the Gauss map, respectively. The number of Pareto solution and the diversity of Pareto solution are two evaluation metrics to evaluate the performance of the multi-objective optimization algorithm. We apply the statistical tests to analyse and investigate the number of Pareto solution and their diversity. The evaluation results indicate that the MOCE with the logistic map has the best optimization performance in both the number of Pareto solution and their diversity. The statistical significance demonstrates that chaotic systems have a great influence on the optimization performance of MOCE algorithms.