A dynamic incentive mechanism for security in networks of interdependent agents

Abstract

We study a dynamic mechanism design problem for a network of interdependent strategic agents with coupled dynamics. In contrast to the existing results for static settings, we present a dynamic mechanism that is incentive compatible, individually rational, budget balanced, and social welfare maximizing. We utilize the correlation among agents’ states over time, and determine a set of inference signals for all agents that enable us to design a set of incentive payments that internalize the effect of each agent on the overall network dynamic status, and thus, align each agent’s objective with the social objective.