Decision support for multi-unit combinatorial bundle auctions

Single item auctions are by far the most common auction format, but they are not always efficient. Combinatorial auctions are beneficial, when complementarities exist between the items to be auctioned. There are, however, several problems with the implementation of combinatorial auctions. Firstly, they are computationally challenging. Secondly, in combinatorial auctions it is difficult for bidders to know what kind of bids to place, since the winning bids complement each other. We consider a progressive electronic procurement situation with a monopsonistic buyer. We propose a decision support tool for iterative e-auctions generating suggestions for bids that would be among the current winners of the auction. We have tested the mechanism and report on the results.