search

quantumiracle / nash-dqn

Official code of Nash-DQN for paper: Nash-DQN algorithm for two-player zero-sum Markov games, details see our paper: A Deep Reinforcement Learning Approach for Finding Non-Exploitable Strategies in Two-Player Atari Games. Zihan Ding, Dijia Su, Qinghua Liu, Chi Jin
16 stars 2 forks source link

Can this method solve the non-zero sum game? #4

Closed hwz9612 closed 1 year ago

hwz9612 commented 1 year ago

Hi, Can this method solve the non-zero sum game? Thanks.

quantumiracle commented 1 year ago

Hi,

Yes. It can be extended to general-sum games for solving Nash equilibrium (NE) or coarse correlated equilibrium (CCE), by plugging in NE or CCE solvers for general-sum, in place of default solver for zero-sum.

But the performance of these general-sum solvers is not guaranteed. Some exploration and changes may be required here.