David C. Parkes
The generalized Vickrey auction (GVA) is a strategy-proof combinatorial auction, in which truthful bidding is the optimal strategy for an agent. In this paper we address a fundamental problem with the GVA, which is that it requires agents to compute and reveal their values for all combinations of items. This can be very difficult for bounded-rational agents with limited or costly computation. We propose an experimental design for an iterative combinatorial auction. We have a theoretical proof that the the auction implements the outcome of the Vickrey auction in special cases, and initial experimental results support our conjecture that the auction implements the outcome of the Vickrey auction in all cases. The auction has better information properties than the sealedbid GVA: in each round agents must only bid for the set of bundles that maximize their utility given current ask prices, which does not require agents to compute their exact values for every bundle.