NS to work with algorithms AI Labs
Presentation of research during RAIL Lab event
NS (Dutch Railways) is developing smart software for planning trains at shunting yards. It is based on the research conducted by PhD candidate Roel van den Broek for the AI & Mobility Lab. He presented his findings during the kickoff event of the RAIL Lab, that took place in November. The software is expected to be put into operation next year.
NS manages a large number of trains. Outside of peak hours, train sets need to be parked, cleaned, inspected, and coupled before they can be deployed again during the next rush hours. Due to limited space at maintenance locations and the anticipated increase in train passengers in the future, NS faces a complex planning problem. Van den Broek developed an algorithm that calculates an option for a station like Enkhuizen within a few seconds, allowing these tasks to be planned quickly and efficiently without trains interfering with each other. The algorithm also works for larger stations like Amsterdam and Utrecht, but it requires significantly more computing time—more than ten hours.
RAIL Lab
The RAIL Lab, part of the AI & Mobility Lab, is a collaboration between the University of Utrecht, NS, ProRail, and TU Delft. Together, they are developing AI technologies to improve the capacity of train transport for a robust railway network. Thanks to these algorithms, logistical operations and capacity planning can be carried out quickly, safely, and reliably. During the kickoff event, all involved students and researchers presented their research. Read more about the projects of the RAIL Lab.