We identify and address algorithmic and game-theoretic issues arising from welfare maximization in the well-studied two-stage stochastic optimization framework. In contrast, prior work in algorithmic mechanism design has focused almost exclusively on optimization problems without uncertainty. We show both positive results, by demonstrating a mechanism that implements the social welfare maximizer in sequential ex post equilibrium, and also negative results, by showing the impossibility of dominant-strategy implementation. In this letter, we describe the relationship between mechanism design and stochastic optimization, and highlight our key technical results. An extended abstract will appear in WINE 2007, and a journal version is under preparation.
/lp/association-for-computing-machinery/mechanism-design-for-stochastic-optimization-problems-h0FLwkMIU5