This is a reproduction of the MADDPG realization of OpenAI. Similarly, this code is to be run in conjunction with environments from a revised version of Multi-Agent Particle Environments (MPE)

• To install, cd into the root directory and type pip install -e .

• Known dependencies: Python (3.8.8), OpenAI gym (0.18.0), pytorch (1.8.0), numpy (1.19.2)

1. Similar document structure to the OpenAI's implementation
2. Similar performance under pure CPU mode
3. Removed operations for benchmarking, saving, or loading files.

1. The implementation with GPU (Quadro RTX 6000) runs slower than that with pure CPU [TODO].

• original MADDPG paper

@article{lowe2017multi,
  title={Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments},
  author={Lowe, Ryan and Wu, Yi and Tamar, Aviv and Harb, Jean and Abbeel, Pieter and Mordatch, Igor},
  journal={Neural Information Processing Systems (NIPS)},
  year={2017}
}

• OpenAI's implementation with tensorflow