摘要
在动态对抗性环境下竞争的多智能体系统里,能够对对手行为进行及时而准确的预测,从而采取相应对策,是战胜对手的必要条件。本文提出了一种预测对手行为的新方法,并在此基础上实现了多智能体的合作。最后以RoboCup仿真比赛作为实验平台证明了该方法的可行性和有效性。
It is necessary to forecast the opponent＇s action effectively and exactly and then adopt corresponding strateges for MAS under dynamic environments to win in a competition. This paper brings forward a new method of opponent＇s action forecast and realizes the multi-agent cooperation using this method. Finally, it is proved that the method is feasible and effective in RoboCup simulating competition system.
出处
《计算机与现代化》
2006年第5期25-27,共3页
Computer and Modernization
