Pytorch implementations of the multi-agent reinforcement learning algorithms, including QMIX, VDN, COMA, MADDPG, MATD3, FACMAC and MASoftQ, which are the state of the art MARL algorithms. We trained these algorithms on MPE, the Multi Particle Environments in PettingZoo. Then they are trained for path planning of swarm of mobile robots.
- QMIX: Monotonic Value Function Factorisation for Deep Multi-Agent Reinforcement Learning
- Value-Decomposition Networks For Cooperative Multi-Agent Learning
- FACMAC: Factored Multi-Agent Centralised Policy Gradients
- Counterfactual Multi-Agent Policy Gradients
- Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments
- Reducing Overestimation Bias in Multi-Agent Domains Using Double Centralized Critics
- Multiagent Soft Q-Learning
Use pip install -r requirements.txt
to install the requirements.
ros2 run marl_planner main.py