Robust travel time predictions are of prime importance in managing any transportation infrastructure, and particularly in rail networks where they have major impacts both on traffic regulation and passenger satisfaction. We aim at predicting the travel time of trains on rail sections at the scale of an entire rail network in real-time, by estimating trains' delays relative to a theoretical circulation plan. Predicting the evolution of a given train's delay is a uniquely hard problem, distinct from mainstream road traffic forecasting problems, since it involves several hard-to-model phenomena: train spacing, station congestion and heterogeneous rolling stock among others. We first offer empirical evidence of the previously unexplored phenomenon of delay propagation at the scale of a railway network, leading to delays being amplified by interactions between trains and the network's physical limitations. We then contribute a novel technique using the transformer architecture and pre-trained embeddings to make real-time massively parallel predictions for train delays at the scale of the whole rail network (over 3000 trains at peak hours, making predictions at an average horizon of 70 minutes). Our approach yields very positive results on real-world data when compared to currently-used and experimental prediction techniques.
Transformers \`a Grande Vitesse
2021
10 pages including 1 page of appendices, 5 figures. Presented at
IAROR RailBelgrade 2023 and published in Journal of Rail Transport P&M
Journal of Rail Transport Planning & Management, Volume 29, 2024,
100418, ISSN 2210-9706
Preprint
Electronic Resource
English
IuD Bahn | 2006
|IuD Bahn | 1996
|Dossier grande vitesse - Eurailspeed 2005. La grande vitesse fait salon à Milan
Online Contents | 2005
TIBKAT | 1995
|IuD Bahn | 1995
|