As the team Electric Sheep we participated in the Flatland Challenge hosted by the SBB.
The Flatland Challenge is a competition to foster progress in multi-agent reinforcement learning for any re-scheduling problem (RSP). The challenge addresses a real-world problem faced by many transportation and logistics companies around the world (such as the Swiss Federal Railways, SBB. Different tasks related to RSP on a simplified 2D multi-agent railway simulation must be solved. Your contribution may shape the way modern traffic management systems (TMS) are implemented not only in railway but also in other areas of transportation and logistics. This will be the first of a series of challenges related to re-scheduling and complex transportation systems. -- SBB
Our team (Nien-Chun Yin, Wenjie Cai, Marcel Fernandez Rosas) achieved place 17 in year 2019 and place 15 in year 2020 on the worldwide leaderboard. During the challenge we learnt to solve a problem with reinforcement learning methods. We have acquired the theoretical foundations with books, online courses and whiteboard sessions.
We successfully implemented following algorithmns:
- Q-learning
- Deep Q Networks
- Double Deep Q Networks
- Dueling Deep Q Networks
- Prioritized Experience Replay
The code can be found in this repository. Keep exploring!