Dynamic railway timetable rescheduling for multiple connected disruptions

Unexpected disruptions occur in the railways on a daily basis, which are typically handled manually by experienced traffic controllers with support of predefined contingency plans. When several simultaneously, it is rather hard for to make rescheduling decisions, because (1) plans corresponding these may conflict each other and (2) no plan considering combined effects multiple available. This paper proposes Mixed Integer Linear Programming (MILP) model reschedule timetable case that at different geographic locations but have overlapping periods pairwise connected least one train line. The dispatching measures retiming, reordering, cancelling, adding stops flexible short-turning formulated MILP also considers rolling stock circulations terminal stations platform capacity. We develop two approaches dynamic environment: sequential approach approach. In approach, single-disruption applied handle new disruption last solution as reference. multiple-disruption every time an extra occurs all ongoing disruptions. A rolling-horizon method has been developed long more efficient way. tested real-life instances subnetwork Dutch 38 10 lines operating half-hourly direction. few cases, did not find feasible solutions, while obtained solutions considered cases. Besides, was able less cancelled services and/or delays than For disruptions, proposed generate high-quality acceptable time.