Why Imitate, and If So, How? A Boundedly Rational Approach to Multi-armed Bandits Export

@article{Schlag1998Why, abstract = {Individuals in a finite population repeatedly choose among actions yielding 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 memorythat increase expected payoffs regardless of the distributionunderlying their realizations. It is shown that the rule thatoutperforms all others is that which imitates the action ofan observed individual (whose realized outcome is better thanself) with a probability proportional to the difference in theserealizations. When each individual uses this best rule,the aggregate population behavior is approximated by thereplicator dynamic. Journal of Economic Literature ClassificationNumbers: C72, C79, D83.Copyright 1998 Academic Press.}, author = {Schlag, K. H.}, citeulike-article-id = {5402135}, citeulike-linkout-0 = {http://www.ingentaconnect.com/content/ap/et/1998/00000078/00000001/art02347}, issn = {0022-0531}, journal = {Journal of Economic Theory}, keywords = {bandit, social-learning}, month = {January}, pages = {130--156}, posted-at = {2009-08-10 00:10:36}, priority = {2}, publisher = {Academic Press}, title = {Why Imitate, and If So, How? A Boundedly Rational Approach to Multi-armed Bandits}, url = {http://www.ingentaconnect.com/content/ap/et/1998/00000078/00000001/art02347}, year = {1998} }

TY - JOUR ID - Schlag1998Why L3 - citeulike-article-id:5402135 N2 - Individuals in a finite population repeatedly choose among actions yielding 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 memorythat increase expected payoffs regardless of the distributionunderlying their realizations. It is shown that the rule thatoutperforms all others is that which imitates the action ofan observed individual (whose realized outcome is better thanself) with a probability proportional to the difference in theserealizations. When each individual uses this best rule,the aggregate population behavior is approximated by thereplicator dynamic. Journal of Economic Literature ClassificationNumbers: C72, C79, D83.Copyright 1998 Academic Press. JF - Journal of Economic Theory SN - 0022-0531 EP - 156 TI - Why Imitate, and If So, How? A Boundedly Rational Approach to Multi-armed Bandits PB - Academic Press SP - 130 KW - bandit KW - social-learning AU - Schlag, KH PY - 1998/01// UR - http://www.ingentaconnect.com/content/ap/et/1998/00000078/00000001/art02347 ER -