Delay management with re-routing of passengers and capacity constraints

Delay management focuses on wait-depart decisions, in which the main question is whether trains should wait for a delayed train or should depart on time. In the traditional delay management models passengers are assumed to always take their originally planned route. In the latest literature however re-routing of passengers is incorporated in the delay management model. Passengers are now assumed to be rational: they always take the shortest route to their destination, also in case of a delay. The delay management model with re-routing (DMwRR) obtained significant improvements to the passengers‟ traveling times. However, in this model, capacity constraints are not taken into account. Therefore in this thesis, we will present three new models with different capacity constraints. The first model includes “platform constraints”, which make sure that trains can only arrive at a station at a certain platform if that platform is empty. The second model includes “not overhaul” constraints, which make sure that trains can only overhaul each other when the next station is a „big‟ station (a station with more than four tracks). The third and last model includes both “platform” constraints and “not overhaul” constraints. At the end experiments based on real-world data from the NS (Nederlandse Spoorwegen, Dutch railways) show that adding capacity constraints to the DMwRR-model makes the model even more complex than it already was. For small real-world instances it turned out that the constraints do not influence the optimal solution. The large real-world instances turned out to be unsolvable due to the high solving times.