Identifying critical risk factors in railway operations based on directed weighted complex networks and combinatorial weighting-TOPSIS

The increasingly complex nature of modern railway operations has led to frequent accidents. Identifying the root causes of these accidents is essential for effective risk control. However, traditional risk identification methods are no longer adequate for handling such complexity. This paper introduces a novel approach for identifying key risk factors and performing a dynamic analysis of railway operation accidents based on a directed weighted complex network. First, risk factors and causal relationships were extracted from 168 railway accident reports based on accident chain theory to construct a Directed Weighted Railway Operation Accidents Complex Network (ROACN). Second, to identify key nodes within the network, a comprehensive node importance evaluation method based on combination weighting-TOPSIS was developed. Finally, a dynamic simulation analysis of network robustness was conducted under seven different targeted attack strategies. The results demonstrate that the ROACN exhibits scale-free network characteristics. Furthermore, attacking the ROACN based on the proposed comprehensive node importance index revealed that the network possesses excellent robustness, indicating that the proposed method reflects key node importance more accurately than traditional evaluation indicators. Consequently, implementing targeted interventions for these key risk nodes can effectively mitigate the propagation of risks and reduce the occurrence of railway accidents.