This paper presents iAuctionMaker as a novel tool that serves as a decision support for e-sourcing professionals on their pursuing of auction optimisation. Given a set of items to auction, iAuctionMaker helps an auctioneer determine how to separate items into promising bundles that are likely to produce better outcomes than the bundle of items as a whole. Promising bundles are those that satisfy certain properties believed to be present in competitive sourcing scenarios. These properties are defined by esourcing professionals and capture their experience and knowledge in the domain. iAuctionMaker models this knowledge as constraints to be satisfied by any bundle, and implements an optimisation algorithm to find the bundles that maximize satisfaction. Experimental results are shown to demonstrate the applicability of the approach. Case studies are presented to demonstrate that iAuctionMaker improves current e-sourcing practices and provides an alternative to combinatorial scenarios whos...
Antonio Reyes-Moro, Juan A. Rodríguez-Aguil