| 求解聚类问题的混合PSO算法设计 |
| |
| 引用本文： | 杨勋,王江晴. 求解聚类问题的混合PSO算法设计[J]. 微电子学与计算机, 2007, 24(10): 43-45,49 |
| |
| 作者姓名： | 杨勋 王江晴 |
| |
| 作者单位： | 中南民族大学,计算机科学学院,湖北,武汉,430074 |
| |
| 摘 要： | 提出了一种基于混沌理论的求解聚类问题的混合PSO算法,该算法结合相关算法的优势采用混沌扰动生成初始化聚类种子,以保证解的分布性能;采用PSO算法进行初步聚类,以改善算法的全局搜索性能;然后采用K-均值算法进行局部精练,以加快算法的收敛速度。将粒子分为静态和动态两种,精心设计了动态粒子的属性,用碟型数据和IRIS数据仿真实验表明,该算法是有效的,能很好地解决聚类问题。
|
| 关 键 词： | 聚类 微粒群 混沌 演化计算 |
| 文章编号： | 1000-7180（2007）10-0043-03 |
| 修稿时间： | 2007-05-08 |
A Hybrid Particle Swarm Optimization Algorithm to Solve Clustering Problem |
| |
| YANG Xun,WANG Jiang-qing. A Hybrid Particle Swarm Optimization Algorithm to Solve Clustering Problem[J]. Microelectronics & Computer, 2007, 24(10): 43-45,49 |
| |
| Authors: | YANG Xun WANG Jiang-qing |
| |
| Affiliation: | College of Computer Science, South-Central University for Nationalities, Wuhan 430074, China |
| |
| Abstract: | This article presents a hybird Particle Swarm Optimization(PSO) algorithm to solve clustering problem. It bases on chaos theory, combines PSO with K-means algorithm. Firstly, uses the method of chaotic stir to create initial seed, secondly design PSO algorithm to cluster in advance, the last introduces K-means algorithm to quicken convergence's steps. Meanwhile the particles is labeled two kinds. Two benchmark cases are tested and show that the method can efficiently cluster the candidates. |
| |
| Keywords: | clustering particle swarm optimization chaos evolutionary compute |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
