博弈论中重复可允许(Iterated Admissibilty)算法对于快速约简博弈模型、寻找合理置信的纳什均衡具有重要意义,但该算法的认知基础存在悖论。本文构建一个完备的博弈认知逻辑系统EL_G,利用该系统语言描述博弈相关概念和性质,使得我们可以基于EL_G逻辑刻画重复可允许算法,从而达到为该算法提供合理的认知基础,解决算法背后的认知悖论的目的。
The algorithm of Iterated Admissibility is significant for the rapid reduction of game models and seeking reasonable Nash equilibria in Game Theory. But there is a paradox in the Epistemic foundation of Iterated Admissibility algorithm. In this paper, we build a complete game epistemic logic systems ELG, and describe some concepts and properities related to a game in the logic language, so that we can characterize the Iterated Admissibility algorithm based on ELG. Thus, we can provide a reasonable epistemic foundation for the algorithm, and in a sense resolve the epistemic paradox for the algorithm.