Improved hybrid indicator to identify key nodes in subway networks

In the subway network, key nodes are closely related to their robustness, which has practical significance for the development of future urban transportation systems. The conventional approach to identifying key nodes is constrained by its reliance on local (e.g., node degree, clustering coefficient) or global indicators (e.g., the number of intermediaries, proximity centrality). This narrow focus on isolated attributes overlooks the intricate interconnections and far-reaching impact of nodes within a network. This paper examines the constraints of prevailing key node identification techniques in metropolitan networks and puts forward a novel approach, designated BD, which integrates node location, connection, and information bridge functions to accurately identify key nodes and facilitate network optimization and emergency management. This new method combines local indicators and global indicators, which are the degree and betweenness of nodes, respectively. Then, in the experiment, subway networks in different cities were used as the subject of the study, which involved a comparison between the BD method and traditional algorithms for identifying key nodes. Experimental results demonstrate that the proposed method exhibits superior performance in the context of both static and dynamic attacks. It is capable of identifying and neutralizing key nodes within a subway network with greater efficacy than existing methodologies. This has a considerable impact on the network, offering valuable insights for enhancing security and risk assessment. This achievement not only improves network security, enabling urban planners to more reasonably design or transform subway networks, optimize line layouts, and enhance the carrying capacity of key nodes, but also promotes the development of intelligent transportation by real-time monitoring and analysis of these key nodes and network status, predicting potential risk points and taking preventive measures in advance.