 |
 |
    |
On any given day, millions of auctions for thousands of items run simultaneously at online marketplaces such as eBay,
uBid, Yahoo, and Amazon Auctions. A comparative buyer looking to obtain an item of a given type is therefore faced
with a large spectrum of opportunities. Accordingly, she must process a large amount of data in order to decide in
which auctions to bid, when and by how much. To address this issue, we have developed a bidding agent that can
participate in multiple auctions with the goal of winning exactly one of them at the lowest price, before a deadline,
and with a given eagerness. The behaviour of a bidding agent is based on a prediction method and a planning algorithm.
The prediction method estimates the likelihood of winning an auction with a given bid. The planning algorithm determines
where and how much to bid, in such a way as to ensure that the probability of winning one among all the alternative
auctions is above the eagerness. A series of experiments based on real datasets has demonstrated that the use of
probabilistic bidding agents increase the individual payoff of the buyers that use them, as well as the collective
welfare of the market. |