search

mlii / mfrl

Mean Field Multi-Agent Reinforcement Learning
MIT License
369 stars 101 forks source link

how you define the markov state? #5

Closed lyers179 closed 6 years ago

lyers179 commented 6 years ago

I have read the paper Mean Field Multi-Agent Reinforcement Learning,agent's state is global or individual？would you mind explaining me the battle game'markov state.Thx！

KornbergFresnel commented 6 years ago

@lyers179 hi, the agent's state is individual as we described in the paper, and the state of agent in the battle game includes observation and feature vector. For more information, pls visit https://github.com/mlii/mfrl/issues/3 or read https://github.com/geek-ai/MAgent.

lyers179 commented 6 years ago

OK,thank you.