Hybrid model for multiagent reinforcement learning

In This work we propose a new method for reducing space and computational requirements of multiagent reinforcement learning based on Markov games. The proposed method estimates value functions by using two Q-value tables or function approximators. We formulate the method for symmetric and asymmetric multiagent reinforcement learning and also discuss some numerical approximation techniques. Additionally, we present a brief literature survey of multiagent reinforcement learning and test the proposed method with a simple example application.