Dispatching trains during seriously disrupted traffic situations

Train services are planned in detail, defining several months in advance the train order and timing at crossings, junctions and platform tracks. A robust timetable is able to deal with minor perturbations even though in case of large delays or blocking of some tracks multiple timetable modifications are required in order to recover the feasibility of operations. Due to the interaction between trains, such exceptional situations result in consecutive delays to other trains in the network, making the railway system very vulnerable to disruptions. This paper investigates disruption handling strategies for large and busy railway networks. We consider seriously disturbed traffic conditions on double track railway lines where some block sections of one track are unavailable for traffic, e.g., due to a temporary track blockage, and disrupted train services are rerouted in other areas while still with the same origin and destination. Centralized approaches are used to solve the whole scheduling problem. Distributed approaches are also presented to manage effectively larger networks, in which a coordinator sets ad-hoc constraints between areas and delegates the scheduling decisions to local schedulers. Computational experiments on a large Dutch railway network, actually controlled by ten dispatchers, assess the performance of the centralized and distributed procedures, showing that both the procedures face increasing difficulty in finding feasible schedules in a short computation time for increasing time horizons of traffic prediction