Agent Behavior Strategy Based on Fuzzy Evaluation and Inference in RoboCupSoccer

10.6138/JIT.2010.11.3.09

Jian-Huai Cai；De-Lin Luo；Mao-Qing Li；Shun-Xiang Wu

RoboCupSoccer ； Fuzzy evaluation ； Fuzzy inference ； Heterogeneous players ； Ball controlling strategy

網際網路技術學刊

11卷3期（2010 / 05 / 01）

371 - 378

英文

In RoboCupSoccer simulation 2D game, the decision layer is the core of a player agent, which is composed of a series of behavior strategies, for example positioning, ball controlling and using heterogeneous players. In order to solve the uncertainly problem caused by the randomly generated parameter of the heterogeneous players, fuzzy evaluation is used to obtain the evaluation results of all heterogeneous players on various roles. A fuzzy inference system is constructed to dynamically select equal heterogeneous players for various roles. All passing routes are estimated and the best dribbling angle is chosen by correspondingly fuzzy inference system. Additionally considering shooting decision, the ball controlling strategy of a player agent is finally formed. The simulation experiments and competition results show that the behavior strategies are feasible and effective.