On Incentives and Updating in Agent Based Models

Abstract

This paper introduces the concept of incentive based asynchronous updating in which the order of updating is determined by incentives. Previously, asynchronous updating has been shown to yield greater stability than synchromous updating for a variety of dynamical systems. However, in those models the order of updating is random. When incentives determine the ordering, the dynamics and end states change. For a conformity model on a two dimensional cellular automata, incentive based asynchronous updating yields greater linear disparity. Fot the game of life, it results in much greater sensitivity to initial conditions.