Addressing the valuation problem in multi-round combinatorial auctions

Multi-round combinatorial auctions have been infrequently employed for essentially two reasons, the computational complexity of the Winner Determination Problem, and the need to provide information feedback to bidders to enable them to bid meaningfully. Here we take a fresh look at the second issue. In business environments, a bidder’s valuation of a package can change in response to signals from other bidders. In the absence of feedback, bidders are forced to expend time and resources to determine these package valuations as accurately as possible. We propose a multi-round combinatorial auction scheme with inter-round feedback called RevalSlot that helps a bidder to converge to an appropriate valuation for a package. This is likely to prove attractive to bidders and thereby increase bidder participation. We also propose a simplified version of the scheme called RevalBundle that requires fewer solutions of the Winner Determination Problem.