Impact of Information Feedback in Continuous Combinatorial Auctions: An Experimental Study of Economic Performance

Let I be the set of distinct items to be sold in a combinatorial auction, and let N = | I |. The terms auction set and auction size refer to I and N, respectively. Bidders can place bids on any item set, which refers to any non-empty subset of I. A bid b can be represented by the tuple b = (S, v, id). Here S denotes the item set the bid was placed on (∅ ⊂ S ⊆ I), also called the span of the bid; v denotes the value of the bid (v > 0), for example, the monetary amount specified in the bid; and id denotes the bidder who submitted this particular bid. Given bid b, S(b), v(b), and id(b) are used to refer to the span, value, and bidder of the bid, respectively. We also use the notion of auction states. In particular, auction state k (where k = 0, 1, 2, ...) refers to the auction after the first k bids are submitted. The bid set is denoted as Bk, that is, Bk = {b1,...,bk}. Auction state 0 refers to the auction before any bids are made, i.e., B0 = ∅. Obviously, Bk ⊆ Bl, for any k and l such that k ≤ l.