Research on application of multi-agent Q-learning algorithm in multiAUV coordination
Yipeng Huang · Applied science and technology · 2008
Several major reinforcement learning algorithms for single agent and multi-agents,that are developed in recent years,are investigated,compared and analyzed.A multi-agent Q-learning algorithm is proposed and used in multi-AUV coordination.This algorithm combines the thoughts or expressions of Nash-Q,CE-Q and WoLF-PHC algorithms,so that the program is simple,the computation is easy,and the convergence rate is high.Experiment results of multiAUV coordination and control prove the effectiveness of this algorithm.