Study on Convergent Fuzzy Particle Swarm Optimization and Performance Analysis

Abstract

Fuzzy particle swarm optimization (FPSO) has shown its great searching ability and high computing precision, while it can not assure the algorithm is convergent. In this paper, a new kind of FPSO is proposed, called convergent fuzzy particle swarm optimization (CFPSO), employing the convergent gene. It differs from normal FPSO in that a convergent gene is introduced in the velocity equation. And it differs from convergent particle swarm optimization (CPSO) in that it employs the fuzzy membership function in the velocity equation. The CFPSO performance is evaluated with four popular benchmark functions, compared with FPSO, CPSO and standard PSO. The experimental results show the model is well built and better performance can be gained with the new optimization algorithm.