PyTorch-Counterfactual-Multi-Agent-Policy-Gradients-COMA

This is PyTorch implementation of paper

Foerster, Jakob N., et al. "Counterfactual multi-agent policy gradients." Thirty-Second AAAI Conference on Artificial Intelligence. 2018.

in the multiagent environment "findgoals" https://github.com/Bigpig4396/Multi-Agent-Reinforcement-Learning-Environment The discription of environment is in 'FindGoals.pdf'

You have to install opencv-python and pytorch to run the code. run 'COMA2.py' you will get the training curve like