COMPUTATIONAL SOCIAL SCIENCE

Department of Computational Social Science Seminar Abstract

Friday, February 17, 3:00 p.m.
Center for Social Complexity Suite
Third Floor, Research Hall

Title: Learning to Consume: Individual vs. Social Learning

Winslow Farrell & Nathan Palmer
CSS PhD Students
George Mason University

Abstract: Allen and Carroll (2001) show an exciting result: with enough time, it is possible for consumers using simple trial-and-error learning to get very close to an optimal consumption rule. Their negative result was that the amount of time necessary to achieve this was astronomical and unattainable under any plausible scenarios. We extend Allen and Carroll’s model to incorporate social learning, and shown that with this addition, we can retain the ability of consumers to find an optimal rule while pushing the amount of time necessary to reliably find such a rule arbitrarily close to the lowest possible bound, clearly within the “lifetime” of a single generation of agents. In addition, we identified an alternative trial-and-error exploration rule which decreases the amount of time needed to reliably find a near-optimal rule by a full order of magnitude. This alternative exploration rule also opens the door for social learning processes incorporating heterogeneous S0 values, potentially eliminating the need for an entire set of costly parameter sweeps. In addition, the model is constructed in both Python and NetLogo to facilitate a docking experiment. Successes and difficulties in that experience are discussed.

Reference:
Allen, Todd, and Christopher Carroll (2001): “Individual Learning About Consumption,” Macroeconomic Dynamics 5(2), 255-271.