基于蚁群优化K-medoids的变电站特性聚类研究  

Study of Substation Characteristics Clustering Based on Ant Colony Optimization K-medoids

作　　者：刘建华[1] 孟颖[2] 谭智[1] 

机构地区：[1]长沙理工大学电气与信息工程学院,长沙410114 [2]长沙理工大学计算机与通信工程学院,长沙410114

出　　处：《电气技术》2012年第2期6-10,共5页Electrical Engineering

基　　金：国家自然科学基金（71071025）

摘　　要：为建立合适的变电站负荷模型,将聚类方法引入到变电站负荷特性分析,提出了一种基于蚁群优化K-medoids的综合聚类算法。该综合算法是K-medoids算法对蚁群的历史最优位置进行聚类分析,蚁群算法全局搜索能力强,克服了K-medoids算法易陷入局部最优的缺点,提高了聚类的准确率。最后通过变电站特性聚类实例,验证了综合算法在变电站特性聚类的可行性和有效性。To establish a proper substation load model, this paper applies clustering method into load characteristics analysis, and proposes a synthetic clustering algorithm of based on ant colony optimization K-medoids. The synthesis algorithm is makes clustering analysis for the history optimal position of ant colony using K-medoids algorithm, ant colony optimization algorithm has higher global search capability, overcoming the K-medoids algorithm easily trapped into local optimal shortcomings, improved clustering accuracy. Finally, it analyzes transformersubstation the characteristics clustering examples and results verify the feasibility and effectiveness of the proposed synthetic clustering method of substation characteristics.

关 键 词：蚁群优化K-medoids 蚁群算法 K-medoids算法 负荷特性 聚类分析 

分 类 号：TM63[电气工程—电力系统及自动化] TP311.13[自动化与计算机技术—计算机软件与理论]