Solving Maximum-Entropy Sampling Problems Using Factored Masks

Samuel Burer (samuel-bureruiowa.edu)
Jon Lee (jonleeus.ibm.com)

Abstract: We present a practical approach to Anstreicher and Lee's masked spectral bound for maximum-entropy sampling, and we describe favorable results that we have obtained with a Branch-&-Bound algorithm based on our approach. By representing masks in factored form, we are able to easily satisfy a semidefiniteness constraint. Moreover, this representation allows us to restrict the rank of the mask as a means for attempting to practically incorporate second-order information.

Keywords: design of experiments, entropy, eigenvalue optimization, branch and bound, semidefinite programming

Category 1: Convex and Nonsmooth Optimization

Category 2: Combinatorial Optimization

Category 3: Linear, Cone and Semidefinite Programming (Semi-definite Programming )

Citation: IBM Research Report RC23542, IBM T.J. Watson Research Center, February 2005.

Download: [PDF]

Entry Submitted: 03/02/2005
Entry Accepted: 03/02/2005
Entry Last Modified: 03/02/2005

Modify/Update this entry