Abstract: This paper analyzes the feedback of the basic ant based clustering algorithm，and presents an improved ant based clustering algorithm.The improved algorithm improves four strategies for basic ant clustering algorithm，the strategy of ants pick up the object，lay down the object and the strategy of adaptive change to the parameter α，and discrete object handling strategy.The experimental results show that the improved algorithm has a better time performance and clustering result.

Key words: ant colony algorithm, clustering analysis, swarm intelligence

摘要： 分析了现有蚁群聚类算法的特点与不足，并在此基础上提出了一种改进的蚁群聚类算法。改进算法分别从蚂蚁捡起对象、放下对象的策略、参数α的自适应改变策略及游离对象的处理策略四个不同方面对现有蚁群聚类算法进行改进。仿真实验结果表明，改进算法可以获得更好的聚类效果和时间性能。

关键词: 蚁群算法, 聚类分析, 群体智能