A Cooperative Evolutionary Algorithm Based on Dual Environment selection strategies

Abstract

Constrained multi-objective optimization problems (CMOPs) involves both conflicting objective function and various constraints. Numerous constrained multi-objective evolutionary algorithm (CMOEAs) have been proposed to deal with CMOPs. Accordingly, this paper adopts a co-evolutionary framework for constrained multi-objective optimization. The main population is responsible for the original constrained multi-objective optimization. The help population search the unconstrained derived from the original CMOP. Dual population coevolution is helpful to promote the convergence of the algorithm. To further improve the performance of the algorithm, this article proposes dual environment selection strategies. The two populations adopt different environment selection strategies to improve the diversity of the individuals. The proposed algorithm is compared with four advanced constrained multi-objective evolutionary algorithms on 24 bench-mark CMOPs. The experimental results demonstrate the competitiveness of the proposed algorithm.