地铁网络节点的聚类分析与关键节点识别  

Cluster Analysis and Key Node Identification of Metro Network Nodes

作　　者：薛锋[1,2] 刘泳博 施政 户佐安[1,2] 何传磊 XUE Feng;LIU Yongbo;SHI Zheng;HU Zuoan;HE Chuanlei(School of Transportation and Logistics, Southwest Jiao tong University, Chengdu 611756, China;National Engineering Laboratory of Integrated Transportation Big Data Application Technology, Southwest Jiao tong University, Chengdu 611756, China;Graduate school of Tangshan, Southwest Jiao tong University, Tangshan 063000, China;China Railway First Survey and Design Institute Group Co. Ltd., Xi’an 710043, China)

机构地区：[1]西南交通大学交通运输与物流学院,成都611756 [2]西南交通大学综合交通大数据应用技术国家工程实验室,成都611756 [3]西南交通大学唐山研究生院,唐山063000 [4]中铁第一勘察设计院集团有限公司,西安710043

出　　处：《武汉理工大学学报（交通科学与工程版）》2021年第5期811-816,共6页Journal of Wuhan University of Technology（Transportation Science & Engineering）

基　　金：国家重点研发计划项目(2017YFB1200702);四川省科技计划项目(2019YJ0211);综合交通大数据应用技术国家工程实验室开放基金项目(CTBDAT201902)。

摘　　要：为克服传统评价模型在寻找网络关键节点时主观因素的干扰,使用聚类思想来进行网络关键节点识别.结合聚类分析的思想,提出了识别网络关键节点的数学模型.将蚁群聚类算法和遗传算法进行融合,对成都市地铁网络的136个节点进行聚类分析,完成了关键节点的识别.采用网络效率和最大连通子图来衡量网络鲁棒性,观察关键节点类被随机攻击后网络指标的变化趋势.结果表明:改进后的蚁群聚类算法聚类性能和效率有了较大提升,并且准确识别出了26个关键节点;关键节点类被随机攻击后,网络效率和最大连通子图下降80%以上,远高于其他节点类,验证了通过聚类进行关键节点识别的可行性.In order to overcome the interference of subjective factors when traditional evaluation models are looking for key nodes of the network,this paper attempted to use clustering to identify key nodes of the network.Combined with the idea of cluster analysis,a mathematical model for identifying key nodes of the network was proposed.The ant colony clustering algorithm and genetic algorithm were fused,and 136 nodes of Chengdu subway network were clustered,and the key nodes were identified.The network efficiency and maximum connectivity subgraph were used to measure the robustness of the network,and the changing trend of the network indicators was observed after the key node classes were randomly attacked.The results show that the improved ant colony clustering algorithm has greatly improved the clustering performance and efficiency,and accurately identified 26 key nodes.After the key node classes are randomly attacked,the network efficiency and maximum connected subgraph decrease by more than 80%,which is much higher than other node classes,which verifies the feasibility of identifying key nodes by clustering.

关 键 词：城市交通 地铁网络 关键节点识别 聚类分析 蚁群聚类算法 

分 类 号：U231[交通运输工程—道路与铁道工程]