关键词: 粒子群优化算法, 高斯扰动, 局部引导, 局部极值点, 社会认知

Abstract:

In order to solve the problem of low convergence rate and premature convergence of the particle swarm optimization (PSO), we propose an improved PSO algorithm which is based on local guidance and Gauss perturbation. Two measures on the particle velocity updating formula are proposed to improve the PSO. Firstly, social cognition is removed and the particles are only locally guided. Secondly, the Gauss perturbation term controlled by the global optimal particle is added. The combination of the two improvement measures can avoid the premature convergence problem and accelerate the convergence speed. Comparative experiments show that the improved PSO algorithm achieves better performance and stability than the classic PSO algorithm. Effect analysis experiments on the two improved measures and the comparative experiments on Gauss perturbation and social cognition further verify the effectiveness of the proposed algorithm.

Key words: particle swarm optimization algorithm (PSO);Gauss perturbation;local guidance;local extremum point;social cognition