Mechanism Design for Task Procurement with Flexible Quality of Service
In this paper, we consider the problem where an agent wishes to complete a single computational task, but lacks the required resources. Instead, it must contract self-interested service providers, who are able to flexibly manipulate the quality of service they deliver, in order to maximise their own utility. We extend an existing model to allow for multiple such service providers to be contracted for the same task, and derive optimal task procurement mechanisms in the setting where the agent has full knowledge of the cost functions of these service providers (considering both simultaneous and sequential procurement). We then extend these results to the incomplete information setting where the agent must elicit cost information from the service providers, and we characterise a family of incentive-compatible and individually-rational mechanisms. We show empirically that sequential procurement always generates greater utility for the agent compared to simultaneous procurement, and that over a range of settings, contracting multiple providers is preferable to contracting just one.
Unable to display preview. Download preview PDF.
- 1.Foster, I., Jennings, N.R., Kesselman, C.: Brain Meets Brawn: Why Grid and Agents Need Each Other. In: Proc. of the 3rd Int. Conference on Autonomous Agents and Multiagent Systems, vol. 1, pp. 8–15 (2004)Google Scholar
- 2.Dash, R.K., Ramchurn, S.D., Jennings, N.R.: Trust-based mechanism design. In: Proc. of the 3rd Int. Conference on Autonomous Agents and Multi-Agent Systems, pp. 748–755 (2004)Google Scholar
- 6.Papakonstantinou, A., Rogers, A., Gerding, E.H., Jennings, N.R.: A truthful two-stage mechanism for eliciting probabilistic estimates with unknown costs. In: Proc. of the 18th European Conference on Artificial Intelligence, Patras, Greece, pp. 448–452 (2008)Google Scholar
- 7.Stein, S., Jennings, N.R., Payne, T.: Flexible service provisioning with advance agreements. In: Proc. of the 7th Int. Conference on Autonomous Agents and Multi-Agent Systems, pp. 249–256 (2008)Google Scholar
- 10.Press, W.H., Flannery, B., Teukolsky, S.A., Vetterling, W.T., et al.: Numerical recipes. Cambridge University Press, New York (1986)Google Scholar