Opposition-based particle swarm optimization with adaptive mutation strategy

Abstract

To solve the problem of premature convergence in traditional particle swarm optimization (PSO), an opposition-based particle swarm optimization with adaptive mutation strategy (AMOPSO) is proposed in this paper. In all the variants of PSO, the generalized opposition-based PSO (GOPSO), which introduces the generalized opposition-based learning (GOBL), is a prominent one. However, GOPSO may increase probability of being trapped into local optimum. Thus we introduce two complementary strategies to improve the performance of GOPSO: (1) a kind of adaptive mutation selection strategy (AMS) is used to strengthen its exploratory ability, and (2) an adaptive nonlinear inertia weight (ANIW) is introduced to enhance its exploitative ability. The rational principles are as follows: (1) AMS aims to perform local search around the global optimal particle in current population by adaptive disturbed mutation, so it can be beneficial to improve its exploratory ability and accelerate its convergence speed; (2) because it makes the PSO become rigid to keep fixed constant for the inertia weight, ANIW is used to adaptively tune the inertia weight to balance the contradiction between exploration and exploitation during its iteration process. Compared with several opposition-based PSOs on 14 benchmark functions, the experimental results show that the performance of the proposed AMOPSO algorithm is better or competitive to compared algorithms referred in this paper.