Method for Identifying Key Stations Based on Importance Evaluation Matrix
Gang Zhang, Tao Zhao, Yunming Wang, Xuehang ShanIntroduction:
As the mainstay of urban transportation, the accurate identification of key stations in the metro–bus two-layer coupled network is essential for ensuring network anti-interference capability, improving operational reliability, and enhancing urban traffic efficiency. Most existing methods focus on a single traffic layer or a single-index approach. These often ignore the inter-layer coupling characteristics between metro and bus systems, leading to low identification accuracy. To address this issue, this study proposes a key station identification method based on an importance evaluation matrix
Methods:
The Space-L method is adopted to construct a two-layer coupled composite transportation network model. Based on the importance evaluation matrix, intra-layer passenger flow, network efficiency, and average degree are taken as core indicators. A coupling strength factor—defined by internal degree (intra-layer connections) and external degree (inter-layer transfer connections)—is further introduced. On the basis of the importance evaluation matrix and coupling strength modeling, an improved key station identification method is proposed.
Results:
Simulation results based on Chengdu’s metro–bus composite network show that compared with traditional methods (e.g., Degree Centrality, Betweenness Centrality) that ignore inter-layer coupling or rely on single indicators, the proposed method can effectively identify crosslayer key stations with significantly improved accuracy. Specifically, the optimal internal degree contribution ratio a = 0.3 (within the range (0.1, 0.5)) maximizes the method’s recognition precision, and the failure of stations identified by this method results in more pronounced degradation of the network’s largest connected subgraph.
Discussion:
The proposed method demonstrates promising applicability in metro–bus composite transportation networks. Compared with traditional approaches that ignore inter-layer coupling or rely on single indicators, it better captures both the structural and functional roles of stations by jointly considering coupling strength, node efficiency, average degree, and passenger flow. This improves the identification of cross-layer key stations and provides a practical basis for network robustness evaluation and reliability-oriented transportation planning.
Conclusion:
The proposed method provides a scientific basis for the reliability optimization of urban transportation networks, emergency station layout, and transfer resource allocation. Additionally, its computational efficiency meets the requirements of large-scale network analysis, supporting the enhancement of urban transit system resilience.