[Intelligence Seminar] April 23, 3:30pm: , Presentation by Sasa Pekec

Dana Houston dhouston at cs.cmu.edu
Wed Apr 17 15:23:07 EDT 2013

APRIL 23 AT 3:30PM, IN GHC 4405

SPEAKER: SASA PEKEC (Duke University)
Host: Tuomas Sandholm
For meetings, contact Charlotte Yano (yano at cs.cmu.edu 
<mailto:yano at cs.cmu.edu>)


Multiple identical items are being sold to unit-demand buyers in a
network. Buyers are willing to pay a premium if they obtain the item
exclusively in their neighborhood, i.e., if no neighbor obtains the
item. Each buyer has a private value for the item, as well as a
private value for obtaining the item exclusively among neighbors. The
seller limited to offering posted prices should inflate the price in
order to maximize revenues and capture the value buyers put on the
allocative externalities. However, we show that these posted price
mechanisms are suboptimal by solving the revenue maximizing mechanism
design problem for several special information structures. We show
that this mechanism design problem could be computationally hard even
when the corresponding perfect information problem is trivial to
solve. Furthermore, optimal seller revenues are non-monotonic in
buyers' valuations because information rents that buyers capture
depend not just on the private values, but also on the underlying
network topology. Finally, we present an ascending auction
implementation of the optimal mechanism on the complete graph with
private values that admit a generic one-dimensional representation.
Our results provide insights on when and where exclusive allocations
should be considered, and how to price exclusivity.


Sasa Pekec is an Associate Professor in the Decision Sciences Area at
the Fuqua School of Business. He holds a Ph.D. in Applied Mathematics
from Rutgers University. He joined Duke in 1998 and is currently
teaching the core statistics course in the daytime MBA and the Cross
Continent MBA programs. Professor Pekec research is interdisciplinary
and revolves around decision-making in complex competitive
environments, and includes work on multiple object auction design,
subset choice and bundle valuations, preference elicitation, and
information aggregation. He has published articles in Management
Science and Operations Research, as well as in top academic journals
in other fields such as economics, mathematics, and psychology.
His work on combinatorial auctions had been widely cited and had
influenced design of a new generation of now standard procurement
auction procedures in a variety of industries. Professor Pekec's
consulting experience includes banking, internet, pharmaceutical,
retail, and telecommunications industries. He serves on the
Supervisory Board of Atlantic Grupa, one of the leading FMCG companies
in SE Europe. Professor Pekec is a member of the Council of Economic
Advisors to the President of Croatia.


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.srv.cs.cmu.edu/pipermail/intelligence-seminar-announce/attachments/20130417/4531d7ec/attachment-0001.html>

More information about the intelligence-seminar-announce mailing list