DFAC Framework: Factorizing the Value Function via
       Quantile Mixture for Multi-Agent Distributional Q-Learning

                 Wei-Fang Sun 1 2 3 Cheng-Kuang Lee 2 Chun-Yi Lee 1

            Abstract                optimize the overall rewards in each episode. Nevertheless,
                                each agent’s policy may not converge owing to two main
   In fully cooperative multi-agent reinforcement
                                difficulties: (1) non-stationary environments ca ...