THE EMERGENCE OF CORDINATION RULES IN AN AGENT-BASED MODEL: A CONTRIBUTION TO THE EVOLUTIONARY APPROACH TO ECONOMIC CHANGE

This article aims to illustrate how artificial intelligence and cognitive sciences may complement economic approaches to learning and coordination. This article proposes a model of individual rationality which microeconomic assumptions are based on the principles of induction, classification and generalisation of knowledge. Building on the "El Farol" model first introduced by Arthur [1994], we use a particular class of classifier systems to study how a population of artificial learning agents develop adapted decision rules in a variety of contexts. These rules allow agent to coordinate their actions by using only minimal information about their environment.