Auctions have been proposed as a way to provide economic incentives for primary users to dynamically allocate unused spectrum to other users in need of it. Previously proposed schemes do not take into account the fact that users’ power constraints might prevent them from transmitting their bid prices to the auctioneer with high precision, and that these transmitted bid prices must travel through a noisy channel. These schemes also have very high overheads which cannot be accommodated in wireless standards. We propose auction schemes where a central clearing authority auctions spectrum to users who bid for it, while taking into account quantization of prices, feedback overheads and noise in the channel explicitly. Our schemes are closely related to channel output feedback problems and specifically to the technique of posterior matching. We consider several scenarios where the objective of the clearing authority is to award spectrum to the bidders who value spectrum the most. We prove theoretically that this objective is asymptotically attained by our proposed scheme when the bidders are non-strategic with constant bids. We propose separate schemes to make strategic users reveal their private values truthfully, to auction multiple sub-channels among strategic users, and to track slowly time-varying bid prices. Our simulations confirm the optimality of our schemes for constant bid prices, and also illustrate the effectiveness of our tracking algorithm for slowly time-varying bids.


Secondary spectrum markets, auctions, posterior matching

Date of this Version