/MADDPG-from-starry-sky6688

Pytorch implementation of the MARL algorithm, MADDPG, which correspondings to the paper "Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments".

Primary LanguagePython

MADDPG

This is a pytorch implementation of MADDPG on Multi-Agent Particle Environment(MPE), the corresponding paper of MADDPG is Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments.

Requirements

Quick Start

$ python main.py --scenario-name=simple_tag --evaluate-episodes=10

Directly run the main.py, then the algrithm will be tested on scenario 'simple_tag' for 10 episodes, using the pretrained model.

Note

  • We have train the agent on scenario 'simple_tag', but the model we provide is not the best because we don't want to waste time on training, you can keep training it for better performence.

  • There are 4 agents in simple_tag, including 3 predators and 1 prey. we use MADDPG to train predators to catch the prey. The prey's action can be controlled by you, in our case we set it random.

  • The default setting of Multi-Agent Particle Environment(MPE) is sparse reward, you can change it to dense reward by replacing 'shape=False' to 'shape=True' in file multiagent-particle-envs/multiagent/scenarios/simple_tag.py/.