Convexification and Global Optimization of Nonlinear Programs
We develop a framework of convex extensions that allows us to
construct convex envelopes of many nonlinear functions. In particular,
we construct relaxation schemes for fractional terms and for polynomial
terms over a general hypercube and compare them with the relaxations
used in the literature. We develop polyhedral relaxations for nonconvex
programs and automatically exploit convexity in constructing relaxations
without explicitly identifying it. We discuss modeling features that can
enhance performance of global optimization algorithms. Finally, we
present computational results on a variety of application problems.
Chemnitz Workshop
Last modified: Wed Nov 3 08:46:51 CEST 2004