Template-Type: ReDIF-Paper 1.0
Title: Updating Strategies Through Observed Play - Optimization Under Bounded Rationality
Author-Name: R. Cressman
Author-Name: K.H. Schlag
Author-Postal: 
Author-Phone: 
Author-Homepage: 
Classification-JEL: C72, C79
Keywords: Multi-Armed Bandit, improving, undominated behavioral rule, play-wise imitating, 
 replicator dynamic, monotone dynamics
Abstract: Individuals repeatedly face a multi-decision task with unknown payoff
 distributions. They have minimal memory and update their strategy by
 observing previous play (and not strategy) of someone else. We select
 behavior rules that increase average payoffs as often as possible in a large
 population where all use the same rule. Here imitation generalizes to a
 pasting procedure. When decisions within the task are unrelated, individuals
 eventually learn the efficient strategy but the underlying dynamic is not
 monotone. However, when choices influence which decisions are subsequently
 faced in the task, play may not be efficient in the long run as it
 approaches a Nash equilibrium of the agent normal form.
Series: Sonderforschungsbereich 303, University of Bonn, Germany
Length: 
Creation-Date: 1998-04
Revision-Date: 
File-URL: http://www.wiwi.uni-bonn.de/bgsepapers/bonsfb/bonsfb432.pdf
File-Format: application/pdf
File-Size: 5500000 bytes
File-URL: http://www.wiwi.uni-bonn.de/bgsepapers/bonsfb/bonsfb432.ps
File-Format: application/postscript
File-Size: 8421376 bytes
Handle: RePEc:bon:bonsfb:432