The Interplay between Analytics and Computation in the Study of Congestion Externalities: The Case of the El Farol Problem

In this paper I study the El Farol problem, a deterministic, boundedly rational, multi-agent model of a resource subject to congestion externalities that was initially studied computationally by Arthur (1994). I represent the interaction as a game, compute the set of Nash equilibria in mixed strategies of this game, and show analytically how the method of inductive inference employed by the agents in Arthur's computer simulation leads the empirical distribution of aggregate attendance to be like those in the set of Nash equilibria of the game. This set contains only completely mixed strategy profiles, which explains why aggregate attendance appears random in the computer simulation even though its set-up is completely deterministic. Copyright 2004 Blackwell Publishing Inc..