Highlights • We model the real-time timetable rescheduling problem due to a complete blockage. • Decisions on where to wait and when to depart should be taken for all trains. • This model is tested on instances of the Chinese high-speed railway line. • Disruptions at different locations and with different durations are considered. • In short computation time, good solutions are found.

Abstract This paper focuses on real-time rescheduling of railway traffic on a high speed railway line in case of a complete blockage of the railway infrastructure. Due to the disruption, all tracks in a railway segment are out of order for a certain period of time. In the situation that we consider, trains that are blocked by the disruption do not return to their origin by taking over train services in the opposite direction, but wait inside the stations until the disruption is over. Thus the main decisions to be taken are the following: in which stations do trains have to wait, in which order do they have to leave when the disruption is over, and which trains have to be canceled? A Mixed Integer Programming model is formulated to minimize the total weighted train delay and the number of canceled trains, while adhering to headway and station capacity constraints. Most instances can be solved in a single optimization run, but for the most complex instances we propose a two-stage optimization approach to improve the computational efficiency. The model is tested on real-world instances of the Beijing–Shanghai high speed railway line. The results show that the model is promising for reducing the effect of a disruption on passenger service, especially in comparison with a heuristic method used in practice.