A game theoretical perspective on incentivizing collaboration in system design

Abstract

Proponents of value-driven design hypothesize that multidisciplinary design optimization architectures can serve as templates for coordinating domain experts in the system design process. However, such architectures can rely on domain model sharing among the domain experts. In this paper, we do not take for granted that domain experts would provide these models. Therefore, we use game theory to formulate a mathematical model of agents' decisions to collaborate or not and analyze how we can formulate a piece rate incentive to motivate agents to collaborate. A piece rate incentive structure provides a marginal increase in reward with a marginal increase in some performance figure of merit. We provide a lower bound condition for the marginal increase in reward term in the piece rate structure that theoretically motivates an agent to collaborate. However, this lower bound is only reasonable if the agent believes that unilaterally collaborating produces an increase in performance.