An Improved Train Classification Procedure for the Hump Yard Lausanne Triage

In this paper we combine an integer programming approach and a computer simulation tool to successfully develop and verify an improved classification schedule for a real-world train classification instance. First, we derive an integer program for computing train classification schedules based on an earlier developed bitstring representation of such schedules. We show how to incorporate various practical restrictions in this model. Secondly, we apply the model to one day of traffic data of the Swiss classification yard Lausanne Triage. We incorporate all the operational and infrastructural restrictions of this yard instance in our integer program. Even with this high number of restrictions, we are able to compute a schedule that saves a full sorting step and one track compared to the currently applied procedure. We finally show this improved schedule is applicable in practice by a thorough computer simulation.