Parameterization of equilibrium assessment in Bayesian game with its application to belief computation
Parameterization of equilibrium assessment in Bayesian game with its application to belief computation
复制标题
DOI：
10.1109/cdc.2016.7798812
复制
发表时间：
2016-12
期刊：
影响因子：
--
通讯作者：
K. Kogiso;Takashi Suzuki
中科院分区：
文献类型：
--
作者：
K. Kogiso;Takashi Suzuki
本文证明了对于一类贝叶斯对策，存在一个均衡评价、一对贝叶斯纳什均衡和一个相应的信念的参数表示。参数化可以从确定性状态空间模型的稳态分析中得到，该模型从平衡评估转移到平衡评估。所得到的参数化以分数形式表示，其比例不随平衡评估的转变而改变。由于比率恒定，参数化使使用给定或测量的BNE作为博弈结果的信念计算成为可能。数值算例验证了参数化的有效性和实用性。
In this paper, the authors show that for a certain kind of Bayesian games, there exists parametric representation in terms of an equilibrium assessment, a pair of a Bayesian Nash equilibrium and a corresponding belief. The parameterization can be derived from a steady-state analysis of a deterministic state-space model that transfers from the equilibrium assessment to one. The resulting parameterization is expressed in fractional form and its ratio is unchanged by transitions of the equilibrium assessment. Owing to the constant ratio, the parameterization enables belief computation using a given or measured BNE as a game outcome. A numerical example confirms the validity and usefulness of the parameterization.