Mathematics of Cooperative Reinforcement Learning

Mathematics of Cooperative Reinforcement Learning

Author		: Bayu Erfianto; Bambang Riyanto Trilaksono; Hilwadi Hindersah; Intan Muchtadi
Published on	: International Journal of Electrical and Computer Engineering (IJECE)

Abstract

In this paper, we propose a cooperative reinforcement learning and provide a mathematical analysis of proposed model. Our cooperative multi-agent reinforcement learning is intended to create adaptive and intelligent system to be applied in mobile agent robot. This cooperative learning algorithm works by considering local subproblem of an agent and its neighbors to update Q-value. We provide the convergence proofs to see the iteration in proposed Q-learning algorithm converges to an optimal Q-value.