Multi-agent reinforcement learning based on policies of global objective

1 .INTRODUCTIONBecause an agent’s rewardis a function of all agents’joint action, when applying RL1]to multi-agent do-mains ,some fundamental change should be made .Byadopting single agentQlearning2]to Markovgames,several algorithms have been proposed,suchas Littman’s mini maxQ-learning( mini max-Q)3],Hu et al’s NashQ-learning(Nash-Q)4 ,5], Claus etal’s cooperative multi-agentQ-learning6], Bowlinget al’s multi-agent learningQ-learning using a vari-able learning rate7 ~9],…

Multi-agent reinforcement learning based on policies of global objective

In general-sum games, taking all agent's collective rationality into account, we define agents' global objective,and propose a novel multi-agent reinforcement learning(RL) algorithm based on global policy. In each learning step, all agents commit to select the global policy to achieve the global goal. We prove this learning algorithm converges given certain restrictions on stage games of learned Q values, and show that it has quite lower computation time complexity than already developed multi-agent learning algorithms for general-sum games. An example is analyzed to show the algorithm' s merits.