Very-Large-Scale Generalized Combinatorial Multi-Attribute Auctions: Lessons from Conducting $60 Billion of Sourcing
Sourcing is the process by which companies acquire goods and services for their operations. Drawing from personal experiences of designing and fielding over 800 sourcing auctions worth over $60 billion, this chapter examines issues that arise in very-large-scale generalized combinatorial auctions. It discusses how combinatorial and multi-attribute auctions can be hybridized. It addresses preference and constraint expression languages for the bidders and the bid taker, as well as techniques for effectively using them. It presents scalable optimization techniques for the market clearing (a.k.a. winner determination) problem. It also considers other issues that this study uncovered as well as the significant efficiency gains and other benefits that followed.