An Incentive Compatible Mechanism for Distributed Resource Planning

We study collaborative resource planning arises when resource managers need to coordinate their schedule with a group of internal or external customers. We design a “schedule selection game” where all participating agents state their scheduling preferences via a valuation scheme, and the mechanism selects the Þnal schedule based on the collective input. A majority of distributed coordination approaches in the literature assume that agents have the good will to coorperate and they always provide private information truthfully. Our contention is that without proper incentives, agents may not reveal this information trufully and they may not behave in a way that is aligned with overall system efficiency. We examine the issue of incentive compatibility, and we propose a direct revelation mechanism that implements an optimal schedule under dominant strategies. The mechanism is a modiÞcation of the well known Vickrey-Groves-Clarke mechanism, and we analyze its properties including budget balanceness, and individual rationality. We then illustrate the mechanism using realworld data obtained from an electronics manufacturer.