计算机工程与应用 ›› 2008, Vol. 44 ›› Issue (35): 129-131.DOI: 10.3778/j.issn.1002-8331.2008.35.039

• 数据库、信号与信息处理 • 上一篇    下一篇

基于捕食-被捕食粒子群优化的模糊聚类

王 琳,罗 可,罗永红   

  1. 长沙理工大学 计算机与通信工程学院,长沙 410076
  • 收稿日期:2008-07-07 修回日期:2008-10-17 出版日期:2008-12-11 发布日期:2008-12-11
  • 通讯作者: 王 琳

Fuzzy clustering based on predator prey particle swarm optimization

WANG Lin,LUO Ke,LUO Yong-hong   

  1. School of Computer and Communication Engineering,Changsha University of Science and Technology,Changsha 410076,China
  • Received:2008-07-07 Revised:2008-10-17 Online:2008-12-11 Published:2008-12-11
  • Contact: WANG Lin

摘要: 粒子群优化聚类算法具有参数简单,收敛快等优势,但也有局部极值问题。为解决此问题,提出一种基于捕食-被捕食的粒子群优化模糊聚类算法且聚类中心采用密度函数初始化。捕食者追逐被捕食者中心,加速收敛,而被捕食者逃离捕食者,促进多样性,以防局部极值出现。实验测试数据表明,算法具有防止局部极值、收敛快、全局寻优能力强等性能优势,能够比较好客观地反映现实世界。

关键词: 捕食-被捕食, 粒子群优化, 模糊聚类, 密度函数, 局部极值

Abstract: PSO clustering algorithm is known to have simple parameters and fast convergence,but there are also local optimal problems.To solve the problem,a fuzzy clustering based on predator prey PSO algorithm is presented,which is using density function to initialize cluster centre.Predators chase preys centre,to accelerate convergence,and the prey escape predators,to promote diversity and to prevent the local optimal there.The experimental test data show that this method is limited to prevent the extreme,fast convergence,global optimization capabilities,and other performance advantages,better able to objectively reflect the real world.

Key words: predator prey, Particle Swarm Optimization(PSO), Fuzzy C-Mean clustering algorithm(FCM), density function, local optimal