A Causal Inference Approach to Measure the Vulnerability of Urban Metro Systems

摘要: Transit operators and passengers need vulnerability measures to understand the level of service degradation under disruptions. This paper contributes literature with a novel causal inference approach for estimating station-level in metro systems. The empirical analysis is based on large-scale data historical incidents population-level passenger demand, thus obviates assumptions made by previous studies human behaviour disruption scenarios. We develop three metrics impact disruptions travel demand average speed. unbiased estimates are obtained adopting propensity score matching method, which adjusts confounding biases caused non-random occurrence An application proposed framework London Underground indicates that station depends location, topology, other characteristics. find 2013 central stations more vulnerable terms loss. However, loss speed reveals from outer suffer longer individual delays due lack alternative routes.