Particle swarm optimization based on an improved diversity mechanism

Particle swarm optimization (PSO) algorithm has shown fast and good search abilities in many unimodal and simple multimodal problems. However, PSO as well as other evolutionary algorithms (EAs) also suffers from the problem of premature convergence in solving some complex multimodal problems. The main reason is that the diversity of swarm decreases very quickly. In this paper, we propose a new PSO variant (DPSO) based on an improved diversity mechanism. Experimental verifications on 13 famous benchmark functions show that the proposed approach achieves better results than standard PSO on the majority of test problems.