Will we finally have on time trains?

Train traffic has been always a source of complaints. I mean, trains that are too few. Or when they might have a decent frequency, they might be late. Or when they are on time, they might be too short. Or even, when they are long enough, on time, and every 5 minutes (like most of the cases now in the Netherlands, fairly speaking), then the problem might be that the wireless does not work.

But apart from this psychological aspects, can train traffic possibly be on time, sooner or later? This requires setting up optimized train traffic control, helping the dispatcher in the control room.


That was in a nutshell, the goal of the EU research project Ontime, that just finished a few weeks ago.

I have been involved in that project as an scheduling specialist – I had the luck to see most of my work done for my PhD thesis useful for something. My experience in the project basically involved the definition, development, implementation and setup of automated systems that are able to change (online and during operations) times, sequences and routes of running trains, in order to minimize delays.
The mathematical details of how such a problem can be solved, and solved quickly and effectively, are hidden in a bunch of research papers. Basically scheduling trains is a very difficult kind of scheduling problem, that can be thought of as a job shop problem. Such problems are relatively common in industrial production environments, transportation networks, supply chains, and they belong to the NP-hard complexity class. The mathematical formulation and the algorithms to solve the problem were the subject of two PhD Theses (by Andrea D’Ariano and Francesco Corman, myself), successfully defended at TUDelft in 2008 and 2010.

One key objective was trying the algorithms in some kind of validated, independent, realistic environment. That is something that I always put in the nice section of the many papers i wrote, the one called “future research”. In fact, it turned out to be a nice and complex challenge on his own, requiring two years from Egidio Quaglietta, the post-doc researcher who was basically doing the job, at the Department of Transport and Planning, with the help of myself.

A small video shows a quick overview of how such a closed loop structure would work, showcasing the results on a final test case of the project.
YouTube Preview Image

Trying things in reality would have been even nicer, but even more complicated! And possibly that would take some more years, but the possibility to have better train traffic is now demonstrated. More details on the Project page:


Be Sociable, Share!
This entry was posted in Uncategorized and tagged , , , , . Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *