COLLECTIVE BEHAVIOR OF EL FAROL ATTENDEES

Abstract

Arthur's paradigm of the El Farol bar for modeling bounded rationality and inductive behavior is employed. Manipulating the memory horizon available to the agents and the selection criteria they utilize for prediction algorithms, one can maneuver the mean attendance away from the externally provided threshold. We observe that a transition occurs in the attendance distribution at a critical memory, beyond which a larger part of the crowd becomes more comfortable. Agents' confidence in their algorithms and the delayed feedback of attendance data increase the overall collectivity of the system behavior. It is possible to manipulate the time evolution of the attendance either externally, by providing past data with delay, or internally, if agents postpone algorithm modification upon failure.