计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (20): 41-44.DOI: 10.3778/j.issn.1002-8331.2008.20.012

• 理论研究 • 上一篇    下一篇

基于聚类分析的微粒群算法

郝武伟,曾建潮   

  1. 太原科技大学 系统仿真与计算机应用研究所,太原 030024
  • 收稿日期:2007-10-24 修回日期:2007-12-29 出版日期:2008-07-11 发布日期:2008-07-11
  • 通讯作者: 郝武伟

Particle swarm optimization algorithm based on cluster analysis

HAO Wu-wei,ZENG Jian-chao   

  1. Division of System Simulation and Computer Application,Taiyuan University of Science and Technology,Taiyuan 030024,China
  • Received:2007-10-24 Revised:2007-12-29 Online:2008-07-11 Published:2008-07-11
  • Contact: HAO Wu-wei

摘要: 在对基本PSO算法进行分析的基础上,针对PSO算法中的早熟收敛问题,提出了一种基于聚类分析的PSO算法(CPSO)。CPSO算法保证了微粒种群的多样性,使微粒能够有效地进行全局搜索。并证明了它依概率收敛于全局最优解。最后以典型的基准优化问题进行了仿真实验,验证了CPSO的有效性。

关键词: 微粒群算法, 全局优化, 收敛性, 聚类分析

Abstract: A new PSO algorithm based on the cluster analysis(CPSO) is proposed for the problem of the premature convergence by the analysis of the standard PSO.The CPSO is guaranteed that the particles are diversiform,and can make particles explore the global optimization more efficiently.The CPSO is guaranteed to converge to the global optimization solution with probability one.Finally,several examples are simulated to show that CPSO is more efficient than the standard PSO.

Key words: Particle Swarm Optimization(PSO), global optimization, convergence, cluster analysis