Robert
Weismantel
Title:
Convex relaxations in mixed integer nonlinear programming
Abstract:
The core of any global optimization algorithm for MINLP
is to determine convex or polyhedral relaxations for the
graph of a nonlinear function.
We prove results about a polyhedral approach for the discrete case and
present
existing and several novel results about convex envelopes in the
general
setting.
This is joint work with D. Michaels and M. Jach