ECON 425/563 & CPSC 455/555
Lecture on December 2nd, 2008
Bill Walsh's December 2 lecture will touch on the following topics, in varying levels of
detail:
- Overview of CombineNet and real-world combinatorial auction applications
- Generalization of combinatorial auctions to "expressive commerce"
- Compact expression of real-world preferences using side
constraints
- Solving really large expressive auctions optimally
- Why the VCG auction is not practical in the real world
- Handling preference uncertainty
- Handling supply uncertainty
- Handling attributes
- Solving large NP-hard problems
The following papers provide good background reading; the first two are the most
important.
- A. T. Sandholm, "Expressive commerce and its application to sourcing:
How we conducted $35 billion of generalized combinatorial auctions"
http://www.cs.cmu.edu/~sandholm/Expressive%20commerce.aimag.scanned.pdf
(covers topics 1 - 3 at a high level)
- B. M. Rothkopf, "Thirteen reasons why the Vickrey-Clarke-Groves process
is not practical"
http://orforum.blog.informs.org/files/2007/04/rothkopf_article.pdf
(covers topic 4)
- C. T. Sandholm, "Optimal winner determination algorithms"
http://www.cs.cmu.edu/~sandholm/windetalgs.pdf
(covers topic 3 in a lot of detail)
- D. C. Boutilier, T. Sandholm, and R. Shields, "Eliciting bid taker non-price preferences in (combinatorial) auctions"
http://www.cs.cmu.edu/~sandholm/eliciting_bidtaker_prefs.aaai04.pdf
(covers topic 5)
- E. C. Boutilier, D.C. Parkes, T. Sandholm, W.E. Walsh, "Expressive banner ad auctions and model-based online optimization for clearing"
http://www.cs.cmu.edu/~sandholm/expressiveBannerAdAuctions.AAAI08.pdf
(covers topic 6)