A new resilience evaluation method of the railway system based on multi-source heterogeneous data

In railway systems, the evaluation of resilience plays a crucial role in risk management. However, the comprehensive quantitative evaluation of resilience often encounters challenges due to heterogeneous data from multiple sources or incomplete datasets, resulting in uncertainty. Furthermore, traditional probabilistic models are inadequate for characterizing the resilience of railway system, necessitating a new framework with targeted modeling and an appropriate index system. This paper proposes a novel distribution-based method to address the uncertainty arising from undesirable datasets: Firstly, methods for transforming multi-source heterogeneous data are constructed. Subsequently, the traditional Bayesian network is enhanced by introducing a fuzzy Bayesian network and proposing new fuzzy inference methods. Finally, new indexes are introduced that align with the characteristics of resilience and the uncertainty-based method. The functional curve-based index synthesis calculation method is established. In the case study, we take the Taiyuan-Jiaozuo Railway as an example to apply the evaluation framework. The results highlight the efficiency and adaptability of the proposed method when dealing with multi-source heterogeneous data and suitable for railway systems while considering its internal dependencies. Additionally, the diverse range of indexes caters to different decision-making goals and provides valuable insights.