Title: A New Paradigm for Robust Optimization: Using Persistence as a Theory of Evidence

Speaker: Tod Morrison, University of Colorado, Denver

Date/Time: Thursday, February 17, 2011, 9:30am MST, 8:30am PST       

Location: CSRI Building/Room 90 (Sandia NM), 915/W133 in CA

Brief Abstract: Robust optimization is a name given to a collection of mathematical programming techniques for proactively addressing uncertainties in mathematical optimization problems. We add to this collection a new paradigm for combinatorial optimization by which the persistence of decisions is treated as evidence of robustness. We begin by introducing necessary terminology and developing the conceptual foundations for our notion of persistence. We then review the necessary elements of Dempster-Shafer Theory before presenting the main thesis: by treating persistence as evidence, we can select solutions based on their evidential robustness. Finally, we demonstrate the potential of this new approach with a discussion of some applications.

The speaker:  Tod Morrison recently completed his PhD in Applied Mathematics and Operations Research studying with Harvey Greenberg.

CSRI POC: Cindy Phillips, 505-845-7296



©2005 Sandia Corporation | Privacy and Security | Maintained by Bernadette Watts and Deanna Ceballos