Smarter algorithms ticket to a fast train


What makes European trains run on time?

No, not fascism. It's ARRIVAL—Algorithms for Robust and online Railway optimization: Improving the Validity and reliAbility of Large scale systems. Besides being the world's most tortured acronym, ARRIVAL is a new EU program aimed at using technology to manage transportation systems more efficiently. The idea is to create timetables that allow more trains to move faster through existing infrastructure, with fewer delays. Succeed at that, and you make mass transit‐an important element in sustainable development—more appealing to potential riders.

"The algorithms have not only cut waiting time between trains from four minutes to two on the Berlin underground network, but have also been used to draw up a new timetable for the Dutch national railway system, which handles 5,500 trains per day. In Switzerland, the system has been used to optimise a schedule so that additional trains may operate on high-risk sections of track, while trials at the Italian stations of Palermo and Genoa have reduced delays by 25%. "

A couple minutes might not seem like much, but averaged over thousands of trains, it can mean a very big difference. If you go further and multiply those minutes that would otherwise have been wasted by the number of passengers on those trains, you get thousands and thousands of hours that aren't being lost by people waiting around at stations.

There's actually a lot of similar work—based both on new, smarter timetables and on policy changes that allow for faster communication and better-protected right-of-ways—going on in the American Midwest. This approach is especially important here, because, without it, we're kind of stuck in a catch-22, where nobody wants to spend money on new infrastructure for an inter-city train system that sucks—but the system is likely to continue to suck without some kind of serious improvement.

Planet Green: Meet the Algorithm that Cut Delays in Half for German Trains

Image courtesy Flickr user jpmueller99, via cc