Flatland-rl : Real-World Baselines Challenge

[Adrian Egli]

>> https://www.flatland-association.org/home

Dynamic Train Rescheduling under Stochastic Perturbations with Multi-Agent Reinforcement Learning and Operations Research

In this competition we want to find out how to efficiently manage dense traffic on a complex railway network with unforeseen events.

This competition tackles a real-world problem faced by many transportation companies around the world: the vehicle rescheduling problem (VRSP). It aims to identify novel solutions and advance existing approaches for dynamic railway rescheduling to take a step towards real-time rescheduling demands of future railway networks. Your contribution to this challenge might influence how modern traffic management systems are implemented.

This competition is part of the AI4REALNET project (AI for REAL-world NETwork operation).