Identification of critical stations in multilayer railway networks based on evidence theory

Multilayer networks have received considerable attention because a network with one layer cannot be used to describe a complex system with multiple relationships inside it. In this paper, we encapsulate the high-speed railway network into dynamic multilayer infrastructures based on the information of train operations and the railway network topology constrained by geographic factors. In our multilayer railway network, each station defines a node and each train route defines a layer. In order to identify key stations in a multilayer network, we first propose some performance metrics to evaluate the importance of stations. Then, an information fusion method based on the Dempster-Shafer evidence theory is introduced. Through this method, the importance of each node in different layers can be fused, and the results of fusion are used to determine the importance of the stations in the whole multilayer network. Finally, some numerical experiments are discussed to illustrate the usefulness of the proposed method. The results developed in this study could provide guidance for improving traffic managements under emergencies.