Oleg Svidchenko (team captain) describes our approach to the RL solution for the Flatland challenge at NeurIPS 2020.
Team JBR_HSE represents the Agent Systems and Reinforcement Learning Lab at JetBrains Research as well as the National Research University Higher School of Economics. The team is based in Saint Petersburg, Russia.
Challenge website: https://www.aicrowd.com/challenges/neurips-2020-flatland-challenge
Source code: to be added.
Lab: https://research.jetbrains.org/groups/ml_human