基于改进遗传算法的RBF网络的截球策略  

Intercepting Ball Strategy of Improved Genetic Algorithm and RBF Network

作　　者：廖本先[1] 杨宜民[1] 项凡[1] 

机构地区：[1]广东工业大学自动化学院,广东广州510090

出　　处：《控制工程》2009年第S2期98-99,102,共3页Control Engineering of China

摘　　要：在RoboCup2D仿真足球比赛中,球员的底层动作设计的好坏直接影响着比赛的输赢,特别是一些关键的动作,比如截球、传球、射门等动作。鉴于此,提出一种新的截球方法,即改进的遗传算法和RBF神经网络相结合的截球方法。该方法是用改进的遗传算法优化RBF神经网络的结构参数,通过优化,提高了网络的全局搜索效率。实验表明,经过改进遗传算法优化的RBF神经网络的截球成功率比单一的RBF神经网络截球成功率高很多。In the RoboCup2D simulation soccer game,the design of the player bottom action is very important and it directly impacts on winning or losing the game,especially some fundamental action,such as intercepting ball,passing ball,shooting ball.In view of the fact,based on the improved genetic algorithm and RBF neural network in intercepting ball,a new method of intercepting ball is proposed.Namely,the improved genetic algorithm is to optimize the structure parameter of RBF neural network.It enhances the network generalization capability and improves global search efficiency of network.The experiment shows that the success rates of intercepting ball with RBF neural network optimized by improved genetic algorithm is much more than the success rates of intercepting ball with RBF neural network.

关 键 词：截球 RBF神经网络 遗传算法 ROBOCUP 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]