基于相似度的蚁群聚类算法  

Ant Colony Clustering Algorithm Based on Similarity

作　　者：沈兴鑫 杨余旺[1] 肖高权 徐益民 陈响洲 SHEN Xingxin;YANG Yuwang;XIAO Gaoquan;XU Yimin;CHEN Xiangzhou(School of Computer Science and Engineering,Nanjing University of Science and Technology,Nanjing 210094;Hunan Yunjian Group Co.,Ltd.,Huaihua 419500)

机构地区：[1]南京理工大学计算机科学与工程学院,南京210094 [2]湖南云箭集团有限公司,怀化419500

出　　处：《计算机与数字工程》2021年第6期1052-1057,共6页Computer & Digital Engineering

摘　　要：针对于蚁群聚类算法在搬运数据项过程中随机选择移动位置时,由于无效移动导致的算法收敛速度缓慢等缺陷,论文提出了一种基于相似度的蚁群聚类算法。通过设计相似度矩阵,基于相似移动机制将蚂蚁随机移动方式优化为按照相似度矩阵规则实施目的性的关联。实验选取Iis、Wine、Haberman和Balance-scale四种经典数据集,相较于现有的LF算法及GACC算法,结果表明在蚂蚁空载率都为90%的条件下,论文提出的SMACC算法的迭代次数明显降低,均体现出较优的聚类速率。This paper proposes an ant colony clustering algorithm based on similarity when the ant colony clustering algorithm randomly selects the moving position during the process of moving data items and the slow convergence speed of the algorithm due to invalid movement.By designing the similarity matrix,the ant random movement method is optimized based on the similarity move⁃ment mechanism to implement the purpose association according to the similarity matrix rule.Four classic data sets of Iis,Wine,Haberman and Balance-scale are selected in the experiment.Compared with the existing LF algorithm and GACC algorithm,the re⁃sults show that the SMACC algorithm proposed in this paper is under the condition that the ant no-load rate is 90%.The number of iterations is significantly reduced,and it both shows better clustering rates.

关 键 词：蚁群聚类 相似度矩阵 相似移动 高速率 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]