The increasing popularity of adopting auctions is largely due to its efficiency of allocating goods. However,in the face of uncertainties on services, the winner determination solution is often not robust enoughto ensure a reliable outcome. This paper aims to design a more robust auction by introducing redundancyinto the selected solution. More specifically, we construct an algorithm and a mechanism for incentivizingtruth-telling in public procurement problems with uncertainties. Our contributions are the developmentof a framework for studying such procurement problems, proving that minimizing cost in this frameworkis NP-complete, developing a quick algorithm that minimizes this cost, and providing a novel multi-stagemechanism that has desirable properties such as efficient, truthful in dominant strategies, and post-executionindividually rational. We show experimentally that our approach significantly outperforms the current practicein many settings.

hdl.handle.net/1765/81918
24th Benelux Conference on Artificial Intelligence, BNAIC 2012
Erasmus University Rotterdam

Zhang, Y, & Verwer, S. (2012). Mechanism for robust procurements. Presented at the 24th Benelux Conference on Artificial Intelligence, BNAIC 2012. Retrieved from http://hdl.handle.net/1765/81918