Incentive-compatible online auctions for digital goods

  • Authors:

  • Ziv Bar-Yossef;Kirsten Hildrum;Felix Wu

  • Affiliations:

  • -;-;U.C. Berkeley, Berkeley, CA

  • Venue:
  • SODA '02 Proceedings of the thirteenth annual ACM-SIAM symposium on Discrete algorithms
  • Year:
  • 2002

Quantified Score

Hi-index 0.00

Visualization

Abstract

Goldberg et al. [6] recently began the study of incentive-compatible auctions for digital goods, that is, goods which are available in unlimited supply. Many digital goods, however, such as books, music, and software, are sold continuously, rather than in a single round, as is the case for traditional auctions. Hence, it is important to consider what happens in the online version of such auctions. We define a model for online auctions for digital goods, and within this model, we examine auctions in which bidders have an incentive to bid their true valuations, that is, incentive-compatible auctions. Since the best offline auctions achieve revenue comparable to the revenue of the optimal fixed pricing scheme, we use the latter as our benchmark. We show that deterministic auctions perform poorly relative to this benchmark, but we give a randomized auction which is within a factor O(exp√log log h) of the benchmark, where h is the ratio between the highest and lowest bids. As part of this result, we also give a new offline auction, which improves upon the previously best auction in a certain class of auctions for digital goods. We also give lower bounds for both randomized and deterministic online auctions for digital goods.