Local Optimization and Adversarial Consideration: Agent Skills in Everest 2002 RoboCup Simulation Team

msra(2008)

引用 23|浏览2
暂无评分
摘要
Skilled individual agents are firm basis of a strong soccer team. Agent skills are short-term goals in the soccer simulation game. The skills available to Everest 2002 agents include kicking, dribbling, forwarding, ball interception and tackling. These intermediate sub goals are implemented by a combination of local optimization which hopes to determine the optimal primitive action from a local perspective and adversarial consideration which takes into account opponents and limitations imposed by simulation environment. Everest 2002 RoboCup simulation teams, building on 11 skilled agents and an on-line coach, became the 2nd place winner of
更多
查看译文
关键词
robocup2002 simulation league. keywords: loac,robocup.,skill,agent
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络
Chat Paper
正在生成论文摘要