The real-time railway rescheduling problem is a crucial challenge for human operators since many factors have to be considered during decision making, from the positions and velocities of the vehicles to the different regulations of the individual railway companies. Thanks to that, human operators cannot be expected to provide optimal decisions in a particular situation. Based on the recent successes of multi-agent deep reinforcement learning in challenging control problems, it seems like a suitable choice for such a domain. Consequently, this paper proposes a multi-agent deep reinforcement learning-based approach with different state representational choices to solve the real-time railway rescheduling problem. Furthermore, comparing different methods, the proposed learning-based approaches outperform their competitions, such as the Monte Carlo tree search algorithm, which is utilized as a model-based planner, and also other learning-based methods that utilize different abstractions. The results show that the proposed representation has more significant generalization potential and provides superior performance.
1 September 2021 in Periodica Polytechnica Transportation Engineering
Rail transportation helps to reach the global climate targets because it is characterized by low emission. The passenger and freight volumes on the railway increase yearly in line with EU targets. However, delays of passenger and freight trains decrease the punctuality and the reliability of the railway sector and the development of the infrastructure is not enough to increase the average speed of trains. Delays mean cost to the passengers, railway operators, infrastructure managers, and all railway undertakers. Therefore, the reason for the most significant optimization target is to minimize delays. In this paper, a possible solution has been described to solve the real-time railway traffic management problems by applying a mixed-integer linear programming approach. For validation of the research result, one simplified case study has been presented. Based on the result, the presented solution can provide effective support to dispatchers in solving real-time traffic management problems.
12 December 2024 in Journal of Rail Transport Planning & Management