Jan Kronqvist: Mixed-integer optimization: algorithms and strong relaxations
Tid: Ti 2021-09-28 kl 10.15
Plats: KTH, 3721, Lindstedtsvägen 25 and Zoom (meeting ID: 659 3743 5667)
Medverkande: Jan Kronqvist (KTH)
The presentation focuses on how to solve optimization problems containing both continuous and integer variables. The integer restrictions can originate from disjunctions, logical relations, or general integer properties of some variables. Some algorithms for solving so-called convex MINLP problems are described in the presentation, and we briefly analyze some of their properties. In the presentation, we will go through some techniques for generating tight polyhedral outer approximations of sets defined by a convex set intersected by an integer lattice. A new approach for modeling disjunctive terms is briefly presented along with an application in robust verification of ReLU-based neural networks. The presentation is intended to give a brief overview of several topics in mixed-integer optimization and some of the presenter’s research interests.