This repo implements the state-of-the-art methods for deep RL in a networked multi-agent system, with observability and communication of each agent limited to its neighborhood. For fair comparison, all methods are applied to A2C agents. Under construction ...
Available IA2C algorithms:
- Fingerprint: Foerster, Jakob, et al. "Stabilising experience replay for deep multi-agent reinforcement learning." arXiv preprint arXiv:1702.08887, 2017.
- Policy inferring: Lowe, Ryan, et al. "Multi-agent actor-critic for mixed cooperative-competitive environments." Advances in Neural Information Processing Systems, 2017.
Available MA2C algorithms:
- Self-other modeling: Raileanu, Roberta, et al. "Modeling Others using Oneself in Multi-Agent Reinforcement Learning." arXiv preprint arXiv:1802.09640, 2018.
- IC3: Sukhbaatar, Sainbayar, et al. "Learning multiagent communication with backpropagation." Advances in Neural Information Processing Systems, 2016.
- Neural message passing: Gilmer, Justin, et al. "Neural message passing for quantum chemistry." arXiv preprint arXiv:1704.01212, 2017.
- Python3
- Tensorflow
- SUMO