✨ Welcome to Flatland 3! Get started with the challenge here

:steam_locomotive: Welcome to The Flatland 3 Challenge!

:fast_forward: This is the 3rd consecutive year of the Flatland Challenge! :muscle: Participate in the challenge and push the boundaries of Multi-agent RL on trains!

:sparkles: What’s New in Flatland 3?

  • In the previous Flatland environments, the trains departed and arrived at anytime - as the goal was to make the trains reach their destinations as fast as possible.
  • However, things are quite different in the real world! Timing and punctuality are crucial to railways, and trains have specific schedules with pre-decided departure and arrival times.
  • We are excited to introduce this real-life concept into the Flatland 3 environment!

Trains now have a time window within which they are expected to start and reach their destination.

:rewind: Previous Challenges



  • The next year, Flatland challenge was accepted to the NeurIPS 2020 Competitions track and saw phenomenal participation from researchers and enthusiasts from all around the world!
  • Check out some stellar community contributions here and here
  • Watch the winners’ talks at NeurIPS here!


  • Flatland relaunched as a part of the AMLD conference in 2021 :muscle:
  • Check out the PPO Baseline here | And DQN Baseline here

:memo: How To Make Your First Submission

  • With our starting instructions, you should be able to make your first submission to the Flatland 3 Challenge in under 10 minutes!! :sparkles:
  • Just follow the instructions here and if you face any issues, let us know in the reply to this post right below! :raised_hands:

Participate Now!

All the best!
Team AIcrowd

:speech_balloon: Chat with the community members
:busts_in_silhouette:Find your teammates here!
:muscle: Leaderboard

1 Like