Action-driven Markov Decision Process and the Application in RoboCup

Abstract

For solving a special kind of Agent planning problems,this paper proposes the concept of Action-Driven Markov Decision Process and analyzes its theory model.Besides,this paper proposes the algorithms for solving Action-Driven Markov Decision Process,which are used for the proximal dribble problem in 2D Competition of RoboCup Soccer Simulation League.The empirical result shows that the new algorithm is much better than the old algorithm of our team in robot′s dribble performance.The new algorithm is also used in WrightEagle 2D Soccer Simulation Team,which shows a good performance in the RoboCup Competitions.