A question about Shapley Q-value

hsvgbkhgbv / SQDDPG

This is a framework for the research on multi-agent reinforcement learning and the implementation of the experiments in the paper titled by ''Shapley Q-value: A Local Reward Approach to Solve Global Reward Games''.

109 stars 42 forks source link

A question about Shapley Q-value #26

Open LHDLLL opened 2 months ago

LHDLLL commented 2 months ago

Deer author,

Can l use code 'value = self.value(state_, action_pol)' calculate to the credit assignment or Shapley Q-value of each agent? And the sample size need to be set 1.

Thank you.

hsvgbkhgbv commented 2 months ago

Yes, you can.

LHDLLL commented 2 months ago

Thank you very much！