An Approach to Equilibrium Selection
An Approach to Equilibrium Selection
复制标题
平衡选择的方法
DOI：
10.1006/jeth.1995.1015
复制
发表时间：
1995
影响因子：
1.6
通讯作者：
Kiminori Matsuyama
中科院分区：
文献类型：
--
作者：
A. Matsui;Kiminori Matsuyama
我们通过将其嵌入由匿名代理连续体玩的动态随机匹配游戏中来考虑具有两个严格纳什均衡的 2x2 双矩阵（对称和非对称）游戏中的均衡选择。与进化博弈文献不同，我们假设参与者是理性的，寻求最大化预期贴现收益；但他们在选择行动时却只能做出短期承诺。以这种方式对摩擦进行建模会产生动态平衡，其稳态对应于原始博弈的纳什结果。我们的选择是基于稳态的微分稳定性特性。研究表明，当摩擦力变得任意小时，严格的纳什结果变得独特且具有全球性，当且仅当它满足 Harsanyi 和 Selton（1988）的风险主导标准概念。因此，除了文献中给出的方法之外，我们的方法还为风险主导提供了另一种支持。
We consider equilibrium selection in 2x2 bimatrix (both symmetric and asymmetric) games with two strict Nash equilibria by embedding it in a dynamic random matching game played by a continuum of anonymous agents. Unlike in the evolutionary game literature, we assume that the players are rational, seeking to maximize the expected discounted payoffs; but they are instead restricted to make a short run commitment when choosing actions. Modeling the friction this way yields the equilibrium dynamics, whose stationary states correspond to the Nash outcomes of the original game. Our selection is based on differential stability properties of the stationary states. It is shown that, as friction becomes arbitrarily small, a strict Nash outcome becomes uniquely absorbing and globally accessible if and only if it satisfies the Harsanyi and Selton (1988) notion of risk-dominance criterion. Our approach thus supplies another support for risk-dominance in addition to those given in literature.