scalable multi-agent reinforcement learning
Forked from https://github.com/livey/scalable_maddpg.Rewrite the environment.
The prey training was not successful.