摘要
针对蚁群聚类算法在蚂蚁之间缺少信息交互导致误识别率高和蚂蚁单独移动带来的资源浪费的问题,本文将三支决策思想融入蚁群聚类算法来改进其性能.通过蚂蚁激活机制、微簇生成机制将待聚类的蚂蚁“一分为三”,区分为已激活的单个蚂蚁、未激活的单个蚂蚁和蚂蚁微簇.再使用精英蚂蚁机制、二次验证机制,遴选相似度达到一定阈值的微簇,形成聚类正域,并赋予精英蚂蚁更高的优先级和固定的平面位置,最后利用不同正域间的信息熵值为属性加权,引导边界域中蚂蚁向着更相似且优先级更高的蚂蚁方向移动.实验结果表明,本文所提出的算法不仅提升了蚁群聚类的质量,还具有良好的时间效率.
In order to address the high misidentification rate and resource wastage caused by individualant movement due to poor information interaction among ants in ant colony clustering algorithms.This paper integrated three-way decision idea into ant colony clustering progress.The ant activation and micro-cluster formation mechanisms are used to categorize ants into activated,non-activated and micro-clustered ants.The elite ant mechanism and secondary verification mechanism are then employed to select micro-clusters with similarity above a threshold to form positive clustering regions.And then elite ants are assigned higher priorities and fixed flat positions.Finally,the information entropy values between different positive regions as attribute weighting are used to guide the movements of boundary ants towards more similar and prioritized ants.Experimental results show that the proposed algorithm not only enhance the quality of ant colony clustering but also demonstrates good time efficiency.
作者
王梦绚
万仁霞
苗夺谦
赵杰
WANG Mengxuan;WAN Renxia;MIAO Duoqian;ZHAO Jie(Faculty of Computer Science and Engineering,Northern University for Nationalities,Yinchuan 750021,China;Faculty of Electronic and Information Engineering,Tongji University,Shanghai 201804,China)
出处
《昆明理工大学学报（自然科学版）》
北大核心
2024年第1期83-97,共15页
Journal of Kunming University of Science and Technology(Natural Science)
基金
国家自然科学基金地区项目(62066001)
宁夏科技领军人才项目(2022GKLRLX08)
宁夏自然科学基金项目(2021AAC03203)。
关键词
蚁群聚类算法
三支决策
精英蚂蚁
微簇
属性加权
ant colony clustering
three-way decision
elite ants
micro-clusters
attribute weighting
