Parameter analysis of particle swarm optimization algorithm
XU Yu-ru · Harbin Gongcheng Daxue Xuebao/Journal of Harbin Engineering University · 2007
Particle swarm optimization(PSO) algorithm is a swarm-intelligence-based stochastic global optimization technique originating from artificial life and evolutionary computation.Though the algorithm has been shown to perform well,the researchers haven't adequately explained how it works.In this paper,the swarm optimization was considered as the evolution of a dynamical system.The convergence property of PSO was analyzed by using the linear discrete time system method.The convergence conditions for simplified PSO algorithm were derived.Since the parameters were very important to the performance and efficiency of PSO,the selection of parameters was systematically discussed through several benchmark functions,and some instructional suggestions were also presented.