Stochastic evolution and equilibrium selection in n-person asymmetric coordination games

给出一个有多个均衡的博弈，均衡选择问题问：如果有的话，哪一个均衡是最“稳定”的？为了回答这个问题，均衡选择模型通过指定和研究玩家玩游戏的策略调整过程来挑选出一个均衡。利用随机进化方法，研究了多人博弈中的均衡选择问题。与已知的两人博弈结果形成鲜明对比的是，我们发现选择结果对策略调整过程非常敏感。特别是，自适应博弈或多种群随机匹配下的选择结果可能与单种群随机匹配下的选择结果有根本不同。

Given a game with multiple equilibria, the equilibrium selection problem asks : Which equilibrium is, if any, most "stable"? To answer the question, an equilibrium selection model singles out an equilibrium by specifying and investigating the strategy adjustment process through which the players play the game. Employing the stochastic evolution approach, we examine the equilibrium selection problem in many-person games. In a sharp contrast to the previously known results for two-person games, we find that the selection outcome is very sensitive to the strategy adjustment process. In particular, the selection outcome under the adaptive play or the multi-population random matching may be radically different from that under the single population random matching.