Aggregating Strategy for Online Auctions
We consider the online auction problem in which an auctioneer is selling an identical item each time when a new bidder arrives. It is known that results from online prediction can be applied and achieve a constant competitive ratio with respect to the best fixed price profit. These algorithms work on a predetermined set of price levels. We take into account the property that the rewards for the price levels are not independent and cast the problem as a more refined model of online prediction. We then use Vovk’s Aggregating Strategy to derive a new algorithm. We give a general form of competitive ratio in terms of the price levels. The optimality of the Aggregating Strategy gives an evidence that our algorithm performs at least as well as the previously proposed ones.
Unable to display preview. Download preview PDF.
- 1.Blum, A., Kumar, V., Rudra, A., Wu, F.: Online Learning in Online Auctions. In: Proc. 14th SODA (2003)Google Scholar
- 2.Blum, A., Hartline, J.D.: Near-Optimal Online Auctions. In: Proc. 16th SODA, pp. 1156–1163 (2005)Google Scholar
- 3.Bar-Yossef, Z., Hildrum, K., Wu, F.: Incentive-Compatible Online Auctions for Digital Goods. In: Proc. SODA 2002, pp. 964–970 (2002)Google Scholar