Learning of Agent Behavior Based on Hierarchical Modular Reinforcement Learning

Abstract

Reinforcement learning is a promising approach to realize intelligent agent such as autonomous mobile robots. This paper presents hierarchical modular reinforcement learning method that Q-Learning is combined with modular structured Profit-Sharing reinforcement learning algorithm hierarchically based on task decomposition in multi-agent environment. By using the learning method, this research aims at learning the behavior of the agent in the pursuit problem, i.e. distributed AI problem in the multi-agent environment. Through numerical experiments, we found that the proposed method has good convergence property of learning compared with the conventional algorithms.