Combinatorial Auctions and Knapsack Problems

Revision with unchanged content. With the rapid diffusion of the internet, different kinds of market designs for e-commerce have emerged. Among these, combinatorial auctions (CAs) have gained increased interest as they permit bidders to express their valuations more easily. For applying a CA in real world scenarios, fast algorithms are needed to compute the optimal allocation of the offered goods. Although much research has dealt with this so-called Winner Determination Problem (WDP), only recently has its equivalence to the well-studied multi-dimensional knapsack problem (MDKP) been noticed. Therefore, there is a lack of research which compares approaches for solving MDKP and WDP problems. With this work, the author, Jella Pfeiffer, provides the missing integrative step, aiming at a more intense understanding and a mutual inspiration of both research areas. She examines structural differences of test instances from both domains, compares the performance of different algorithms, and alludes to a successful search behavior of non-exact but fast algorithms. The book addresses readers interested in electronic market design and algorithms solving complex combinatorial optimization problems.