Fast convergence in evolutionary models: A. Lyapunov approach

被引：6

作者：

Ellison, Glenn ^{[1
]}

Fudenberg, Drew ^{[2
]}

Imhof, Lorens A. ^{[3
]}

机构：

[1] MIT, Cambridge, MA 02139 USA

[2] Harvard Univ, Cambridge, MA 02138 USA

[3] Univ Bonn, Bonn, Germany

来源：

JOURNAL OF ECONOMIC THEORY | 2016年 / 161卷

基金：

美国国家科学基金会;

关键词：

Hitting time; Learning model; Local interaction; Lyapunov function; Markov chain; Recency; FICTITIOUS PLAY; EQUILIBRIUM SELECTION; STOCHASTIC STABILITY; GLOBAL CONVERGENCE; FINITE POPULATIONS; RESPONSE DYNAMICS; MIXED EQUILIBRIA; GAMES; INNOVATION; SPEED;

D O I：

10.1016/j.jet.2015.10.008

中图分类号：

F [经济];

学科分类号：

02 ;

摘要：

Evolutionary models in which N players are repeatedly matched to play a game have "fast convergence" to a set A if the models both reach A quickly and leave A slowly, where "quickly" and "slowly" refer to whether the expected hitting and exit times remain bounded when N tends to infinity. We provide simple and general Lyapunov criteria which are sufficient for reaching quickly and leaving slowly. We use these criteria to determine aspects of learning models that promote fast convergence. (C) 2015 Elsevier Inc. All rights reserved.

引用

页码：1 / 36

页数：36

共 42 条

[1]

[Anonymous], 1960, A First Course in Stochastic Process

[2]

[Anonymous], 1996, Graduate Texts in Mathematics

[3] Deterministic approximation of stochastic evolution in games [J].