On the importance of adopting a multi-centrality approach to detecting the vital nodes of urban road networks

Abstract

Transportation networks are prone to various types of disturbances, ranging from regular rush hour congestion to occasional closures due to construction zones, accidents, etc. It is impossible to avoid all disruptions, but detecting the critical points of networks (i.e., nodes that noticeably affect the connectedness of the network when closed) would help urban transportation managers prioritize preventive actions. Centrality measures are used to quantify the importance of network nodes. In this study, we calculated various centrality measures for six existing urban road networks and evaluated their importance to the connectivity and functionality of the networks via a percolation method. Along with well-established centrality measures such as betweenness, communicability, and the clustering coefficient, we evaluated the collective influence (CI) and the enhanced collective influence (ECI) indices in a transportation context. We found that nodes with high values of CI or ECI are not the ones with high values of betweenness, communicability and the clustering coefficient. Nevertheless, failures of nodes with high values of CI, ECI or betweenness centrality most significantly affect the connectivity and functionality of urban road networks. We identified three distinct sets of vital nodes in the networks we analyzed. Hence, we conclude that a set of centrality measures should be used to detect vital topological nodes of urban networks rather than just one centrality measure. Moreover, we investigated employing various aspects of CI and ECI to reveal the critical nodes of urban road networks.