Train Shunt Planning using Genetic Algorithms

Because demand for train transportation is very high during rush hours, train operators need a lot of rolling stock to supply this demand. Outside the rush hours, especially during the night, this results in a surplus of rolling stock which need to be parked at a shunt yard. Because of limit track capacity and the diversity in train types, scheduling this process is a hard and time consuming task. We have developed a tool that is able to generate good shunt schedules to support local shunt planners. With the use of a genetic algorithm, we were able to extend the basic problem formulation with exible shunt times for each shunt activity in the schedule. Our model is successfully tested on a series of theoretical cases and two real-world cases from Netherlands Railways.