Exploring significant edges of public transport network under targeted attacks

Abstract

Edges of public transport network (PTN) play important roles in transporting passengers of cities, especially in metropolises. Understanding the significant edges of PTN can provide insights for managers to increase the efficiency of networks. In this work, we construct a new measure called community bridge (CB), which is based on the number of nodes in communities. We propose an edge removal process to test network efficiency (NE), average transfer times (ATT) and correlation coefficient (CC). For comparison, edge measures of degree product (DP), edge betweenness (EB), edge overlap (EO), closeness centrality index (CCI) are introduced for the process. The results show that there are only 12.6% edges are CBs in the network. Removing edges according to CBs can decrease NE more effectively compared with other indicators. However, it cannot increase the ATTs effectively. It is found that removing edges according to CCI is the most effective way to increase ATT under the outranked 15% edges removed. In addition, the results indicate the CC has very different changes according to the introduced removal strategies.