WHY IMITATE, AND IF SO, HOW - A BOUNDEDLY RATIONAL APPROACH TO MULTIARMED BANDITS

Individuals in a finite population repeatedly choose among actions yie lding uncertain payoffs. Between choices, each individual observes the action and realized outcome of one other individual. We restrict our search to learning rules with limited memory that increase expected pa yoffs regardless of the distribution underlying their realizations. It is shown that the rule that outperforms all others is that which imit ates the action of an observed individual (whose realized outcome is b etter than self) with a probability proportional to the difference in these realizations. When each individual uses this best rule, the aggr egate population behavior is approximated by the replicator dynamic. ( C) 1998 Academic Press.