A Behavioral Model for Mechanism Design: Individual Evolutionary Learning

We describe a new and different behavioral model for individuals playing in repeated games, called Individual Evolutionary Learning (IEL). It is based on a flexible learning process, does not require calibration, and can handle very large strategy sets. It can be used as a computer testbed to study the probable performance of a wide range of mechanisms over a wide range of environments prior to testing them in a laboratory or using them in practice. We illustrate the behavior of IEL and the utility of the testbed approach by analyzing an open question in mechanism design the dynamics of Groves-Ledyard mechanisms. Contrary to standard theories, the prediction from the IEL behavioral model is that the average time to convergence varies smoothly and is U-shaped in the mechanism’s free parameter. We validate the results from the testbed with data from economic experiments with humans.