Efficiency levels in sequential auctions with dynamic arrivals

摘要

In an environment with dynamic arrivals of players who wish to purchase only one of multiple identical objects for which they have a private value, we analyze a sequential auction mechanism with an activity rule. If the players play undominated strategies then we are able to bound the efficiency loss compared to an optimal mechanism that maximizes the total welfare. We have no assumptions on the underlying distribution from which the players' arrival times and valuations for the object are drawn. Moreover we have no assumption of a common prior on this distribution.