%0 Journal Article
%T Learning of Soccer Player Agents Using a Policy Gradient Method : Coordination Between Kicker and Receiver During Free Kicks
%A Harukazu Igarashi
%A Koji Nakamura & Seiji Ishihara
%J International Journal of Artificial Intelligence and Expert Systems
%D 2011
%I Computer Science Journals
%X As an example of multi-agent learning in soccer games of the RoboCup 2D Soccer SimulationLeague, we dealt with a learning problem between a kicker and a receiver when a direct free kickis awarded just outside the opponent¡¯s penalty area. We propose how to use a heuristic functionto evaluate an advantageous target point for safely sending/receiving a pass and scoring. Theheuristics include an interaction term between a kicker and a receiver to intensify theircoordination. To calculate the interaction term, we let a kicker/receiver agent have areceiver¡¯s/kicker¡¯s action decision model to predict a receiver¡¯s/kicker¡¯s action. Parameters in theheuristic function can be learned by a kind of reinforcement learning called the policy gradientmethod. Our experiments show that if the two agents do not have the same type of heuristics, theinteraction term based on prediction of a teammate¡¯s decision model leads to learning a masterservantrelation between a kicker and a receiver, where a receiver is a master and a kicker is aservant.
%K RoboCup
%K Soccer Simulation
%K Multiagents
%K Policy-Gradient methods
%K Reinforcement Learning
%U http://cscjournals.org/csc/manuscript/Journals/IJAE/volume2/Issue1/IJAE-36.pdf