mlii / mfrl

Mean Field Multi-Agent Reinforcement Learning
MIT License
377 stars 100 forks source link

Why the number of agents has not decreased in battle game? #30

Open lzh-awesome opened 2 years ago

lzh-awesome commented 2 years ago

hi,i am trying to reproduce the battle game (red army use MFQ and blue army use DQN.) i trained the model and find both algorithums has a negative reward.and the agents number of two side has not decreased in the battle which is confused.can u help me?

1 JPG