[Intelligence Seminar] Joint OM/OR Seminar - Marek Petrik

[Dana Houston]

====================================================
THIS ANNOUNCEMENT IS FOR OM/OR SEMINAR, WHICH MAY BE
OF INTEREST TO THE INTELLIGENCE SEMINAR AUDIENCE
====================================================

OR/OMM Faculty and PHDs

The following joint Operations Research and Operations Management and
Manufacturing Seminar has been posted to the Tepper School seminar
tracker.

The schedule can be found under Manufacturing. There are slots left to
fill, please take the time to complete the schedule.

Name:         Marek Petrik
University:   IBM T.J. Watson Research Center
Date:         Friday, November 16, 2012
Time:         10:30 a.m. - 12:00 p.m.
Location:     GSIA Faculty Conference Room 322

Title: Approximate Dynamic Programming By Minimizing Distributionally
Robust Bounds

Abstract: Approximate dynamic programming is a popular method for
solving large Markov decision processes. I will describe a new class of
approximate dynamic programming (ADP) methods---distributionally robust
ADP---that address the curse of dimensionality by minimizing a
pessimistic bound on the policy return. This approach turns ADP into an
optimization problem, for which we derive new mathematical program
formulations and analyze its properties. DRADP improves on the
theoretical guarantees of existing ADP methods---it guarantees
convergence and L1 norm-based error bounds. The empirical evaluation of
DRADP shows that the theoretical guarantees translate well into good
performance on benchmark problems.

-- 
Dana M. Houston
Language Technologies Institute
School of Computer Science
Carnegie Mellon University
5405 Gates Hillman Complex
5000 Forbes Avenue
Pittsburgh, PA 15213

T:  (412)268-4717
F:  (412)268-6298

-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mailman.srv.cs.cmu.edu/pipermail/intelligence-seminar-announce/attachments/20121105/f1c7bb6d/attachment-0001.html