基于层次聚类的共享单车维修点规划模型  

Planning model of shared bike maintenance points based on hierarchical clustering

作　　者：毛昊迪 汤鲲 MAO Haodi;TANG Kun(Wuhan Research Institute of Posts and Telecommunications,Wuhan 430000,China;Fiberhome Telecommunication Technologies Co.,Ltd.,Nanjing 210000,China)

机构地区：[1]武汉邮电科学研究院,湖北武汉430000 [2]南京烽火天地通信科技有限公司,江苏南京210000

出　　处：《电子设计工程》2022年第21期20-23,28,共5页Electronic Design Engineering

摘　　要：该文针对共享单车的固定维修点设置问题,提出了使用蚁群算法优化的层次凝聚聚类方法,该聚类方法使用蚁群算法的信息素机制和簇间距离来进行聚类,通过引入随机性来解决层次聚类的局部最优问题,使用层次聚类的树形结构来应对维修点的数量变化,将每个簇的聚类中心作为维修点位置。该算法在西雅图2015年自行车租车站点位置的真实数据集和加州大学欧文分校的数据集上的仿真实验结果表明,相对于传统聚类算法,该算法可以找出更好的维修点位置。Based on shared problems setting up the bicycle fixed maintenance station,puts forward the use of ant colony algorithm to optimize the level of the condensed clustering,the method using ant colony clustering algorithm of pheromone mechanism and the distance between the cluster to cluster,by introducing randomness to hierarchical clustering of local optimal solution,using hierarchical clustering tree structure to deal with the number of maintenance point of change. The cluster center of each cluster is used as the repair point location.Simulation experiments of this algorithm on real data sets of Seattle bike rental station locations in 2015 and data sets of University of California,Irvine show that this algorithm can find better repair point locations than traditional clustering algorithms.

关 键 词：层次凝聚聚类 蚁群优化 聚类中心 共享单车 

分 类 号：TN91[电子电信—通信与信息系统]