Evaluation of the mean-variance mapping optimization for solving multimodal problems

Abstract

Based on swarm intelligence principles and an enhanced mapping scheme, the extension of the original single-particle mean-variance mapping optimization (MVMO) to its swarm variant (MVMOS) is investigated in this paper. Numerical experiments and comparisons with other heuristic optimization methods, which were conducted on several composition test functions, demonstrate the feasibility and effectiveness of MVMOS when solving multimodal optimization problems. Sensitivity analysis of the algorithm parameters highlights its robust performance.