Robust planning for making a train schedule

In the Netherlands, the process of creating a new train schedule begins more than a year in advance. Although this has contributed to having one of the most punctual while also one of the most heavily used rail networks in Europe, it has also become increasingly more difficult to create high quality plans. One of the challenges when making these plans is the continuous disruptions that necessitate various changes. This is particularly prevalent in freight transportation, where disruptions occur on a large scale.

In this research, we aim to explore whether we can design schedules in such a way that, firstly, disruptions can be better accommodated within the planning. Secondly, we want to determine if certain details can be finalized much later, reducing the amount of work that needs to be redone repeatedly.

We aim to solve this by applying algorithms that schedule trains at a global level. Specifically for freight trains, we can for example configure the algorithm in such a way that it does choose the paths between major stations for a train, but does not yet precisely determine the route within the station itself. This will be done in such a way that the planning respects the maximum capacity at each station.

The results of this research will be of particular interest to infrastructure managers, who will need less time to create good schedules. It is also relevant for carriers, as they will have more certainty and clarity about where and when they can operate their trains

Researcher

Casper Loman (Department of Information and Computing Science)

Academic supervisors

Dr. Han Hoogeveen, dr. ir. Marjan van den Akker, dr. Roel van den Broek, Floris Visser

Grant funding agency and (co-)funding non-academic partners

NS, ProRail and NWO