Disruption Management on Shunting Yards with Tabu Search and Simulated Annealing
Summary
Trains are cleaned and inspected regularly to keep passengers happy, and costs low by preventing accidents. These services are done within a shunting yard based on a schedule created by the Dutch Railways. In practice, the execution of a schedule often deviates because a train arrives too late, or arrives with carriages in an unexpected order. If these disruptions break the schedule in any way, it will have to be corrected manually. We propose a local search technique to aid the adjustment of the schedule, while also aiming for the resulting schedule to be similar to the original. We introduce a score for similarity based on comparing the machinist schedules.