Robustness Analysis And Key Node Identification Of Transportation Networks Based On Complex Network Theory

The progress of science and technology has made the trend of urbanization in major regions of our country more and more obvious,and a series of traffic problems have followed.If the traffic problem is not solved,it will inevitably affect the life of urban residents and the operation efficiency of social and economic development.In order to alleviate the urban traffic problem,it is necessary to analyze the overall stability of the urban traffic network.Complex network is a discipline that has emerged in recent years.The urban traffic network is essentially a large complex network.Using the relevant knowledge of complex network to perform robust analysis of the traffic network can identify weak links in traffic and improve the stability of traffic operation.At the same time,the robust optimization of the transportation network can improve the utilization efficiency of limited resources and provide a new perspective for the planning of urban transportation networks.This article first collected the station and route data of Chongqing Public Transport Network,Chengdu Public Transport Network and Shenzhen Public Transport Network,and established three network topology models based on Python software.Using connectivity,relative size of the largest connected subgraph and network efficiency as robustness indicators,the robustness of the network is analyzed by attacking the network by means of deliberate and random attacks.The results show that in the deliberate attack mode,the robustness of the three networks is poor,while in the random attack mode,the networks have strong robustness,and the Shenzhen Public Transport Network has the best robustness.Subsequently,the network was robustly optimized by means of high-degree addition,highbetween addition,low-degree addition,low-between addition,and random addition.The results show that the network robustness can be improved most obviously by adding edges with low degree and low betweenness,and the higher the edge rate,the better the effect.Finally,the recognition effects of the five key node recognition methods of degree recognition,neighbor degree recognition,KSD recognition,DKS recognition and DKSN recognition in the three networks are compared.The results show that KSD recognition can obtain the best key node recognition effect under dynamic recognition.