Evaluating the Vulnerability of the Sydney Train Network by Comparing Access-based and Network Centrality Metrics

Abstract

Operational incidents are a significant cause of unreliability on rail transit networks. These incidents cause major delays in services, impact passenger travel time, and have knock-on effects that interrupt other public transport services. Consequently, the vulnerability of the rail transit network is a crucial concern for managers and operators. This paper employs network vulnerability analysis to characterize individual critical stations in a railway network. The concepts of graph theory and person-weighted access are implemented to identify the critical nodes in the Sydney train and metro network, and the results are compared. In the first method, weighted and unweighted centrality measures are computed to find the most critical station. In particular, eigenvector centrality is used to identify the critical nodes by scoring all nodes in the network using the first eigenvector of the graph adjacency matrix. In the second approach, stations are ranked by the reduction of access before and after an incident. Finding of this study may have implications not only for the train operators and managers but also for the transit network planners to enhance the resilience of the public transport network.