Equilibrium's Action Bound in Extensive Form Games with Many Actions
national conference on artificial intelligence(2013)
摘要
Recently, there has been great progress in computing optimal strategies in large extensive games.But there are still many games where computing an equilibrium remains intractable, mostly due to the size of the game tree.In many of them, such as no-limit poker, huge game tree size is caused by too many actions available in informationsets. We show that it's possible to limit number of used actions in Nash equilibirum.We bound this number by a private information and show how to compute this Nash equilibirum.
更多查看译文
关键词
game theory,extensive form games
AI 理解论文
溯源树
样例
生成溯源树,研究论文发展脉络