MRRL Research

Single-robot Reinforcement Learning

We first provide several implements of state-of-the-art single-robot reinforcement learning baselines which includes:

• DQN

• Doublee-q learning

• Actor-critic

• DDPG

• SAC

• TD3

• PPO

Multi-robot Reinforcement Learning

Besides, we also provide several implements of state-of-the-art multi-robot reinforcement learning baselines which includes:

• Independent Q-learning

• IPPO

• MAPPO

• MADDPG

• QMIX

• VDN

• COMA

• MAAC

• HRL

• Shapley PPO

• Others