Train rescheduling with consideration of passenger reservicing during severe disruptions

Train dispatching is critical for ensuring the punctuality and reliability in passenger transportation, especially in railway systems with a seat reservation mechanism, such as that in China. In the presence of a disruption, e.g., a 4-h segment blockage, passengers holding reservations on disrupted trains face significant travel plan alterations. As the flexibility for these stranded passengers to self-shift to alternative travel times and/or modes is often limited with pre-booked tickets and infrequent services, they are very likely to wait for their reserved trains for a long time, causing negative consequences on passenger satisfaction. In this study, we propose a method to address this rescheduling problem from the perspective of passenger reservicing. Measures of adding extra stops and inserting extra trains are conditionally used with the limitation of an effective time window (ETW), alongside retiming, reordering, and local-rerouting, to offer alternative travel options for stranded passengers, enabling them to reach their destinations with minimal delays. To accurately capture the limitation of ETW, we reconstructed the space-time network by introducing an alternative arc set, followed by formulating an ILP model. This NP-hard problem is then solved using a Lagrangian relaxation solution method, which integrates a decomposition approach, a label correcting algorithm, and a heuristic algorithm. We also conduct computational experiments on the Chinese high-speed railway network to validate the feasibility, effectiveness, and efficiency of our proposed method, and perform experimental analysis to help dispatchers to make more informed decisions in daily operations.