Abstract: This paper starts from basic model that imitates the aggregation of dead bodies by ants. With introducing new concepts: similarity factor and dissimilarity factor, it redefines the expression of the accepting fraction, and makes a more pure and rapid clustering result. The result of experiment shows the new algorithm can obviously improve the clustering quality.

Key words: Ant colony algorithm, Clustering, Swarm algorithm, LF algorithm, Similarity factor, Dissimilarity factor

摘要： 从模仿蚂蚁堆积的尸体的基本模型出发，引入了相似因子和相异因子的概念，通过重新定义接受分数的计算公式，能够使聚类结果更纯，聚类速度更快。实验结果表明新算法明显改善了聚类质量。

关键词: 蚁群算法, 聚类, 集群算法, LF算法, 相似因子, 相异因子