When is it Best to Best-Respond? NOAM NISAN School of Computer Science and Engineering, The Hebrew University of Jerusalem, and Google Tel Aviv. and MICHAEL SCHAPIRA Computer Science Department, Princeton University and GREGORY VALIANT Computer Science Department, UC Berkeley and AVIV ZOHAR Microsoft Research, Silicon Valley Often, in both computerized settings and economics settings, the prescribed behavior for participants is to repeatedly best respond to each others actions. We aim to understand when such myopic local rationality is also globally rational , i.e., when is it best for a player, given that the others are repeatedly best-responding, to also repeatedly best-respond? Categories and Subject Descriptors: J.4 [Social and Behavioral Sciences]: Economics General Terms: Algorithms, Economics, Theory Additional Key Words and Phrases: Mechanism Design, Best-Response Dynamics 1. MOTIVATION The basic object of study in game theory and economics is the equilibrium. However, an equilibrium is a static concept which often abstracts away the question How is an equilibrium reached? . A truly satisfactory answer to this question is one where each player follows a simple and natural rule of behavior yet, mysteriously, the resulting dynamics reach an equilibrium point. The (arguably) most compelling examples of such
/lp/association-for-computing-machinery/when-is-it-best-to-best-respond-xWdk8KZkAj