MADDPG-DDPG

use tensorflow to implement the paper: Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments

use ddpg to training agents use maddpg to training agents

tensorflow==1.4.1

you should install multiagent-particle-envs first

you can run test_three_agent_maddpg.py and test_three_agent_ddpg.py to see the result