Optimization afternoon

Monday 25 November 2013
LAAS-CNRS, Toulouse, France


This half-day workshop aims at reporting recent achievements in optimization.

Date and venue

The workshop takes place on Monday 25 November 2013 at LAAS-CNRS, 7 avenue du colonel Roche, 31077 Toulouse, in the main conference room (salle de conférences), on your right handside in the main entrance hall.


  • 14:00-14:45 - Michal Kocvara (Univ. Birmingham, UK).
    Title: Introducing PENLAB, an MATLAB code for nonlinear (and) semidefinite optimization.
    Abstract: I will introduce a new code PENLAB, an open Matlab implementation and extension of our older code PENNON. PENLAB can solve problems of nonconvex nonlinear optimization with standard (vector) variables and constraints, as well as matrix variables and constraints. I will demonstrate its functionality using several nonlinear semidefinite examples. This is a joint work with Jan Fiala (NAG) and Michael Stingl (Erlangen).

  • 15:00-15:45 - Miguel Anjos (Polytechnique Montreal, Canada).
    Title: Towards Efficient Higher-order Semi-definite Relaxations for Max-cut.
    Abstract: The basic semidefinite relaxation for max-cut can be tightened by adding facet-defining inequalities for the metric polytope, or more generally valid inequalities for the convex hull of incidence vectors of cuts, known as the cut polytope. We start with the basic semidefinite relaxation intersected with the metric polytope, and iteratively refine this relaxation using a new class of cuts called polytope cuts. We present theoretical insights as well as computational results. This is joint work with E. Adams, F. Rendl and A. Wiegele.

  • 16:00-16:45 - Fabien Caubet (Univ. Paul Sabatier Toulouse, France).
    Title: Shape optimization methods for the Inverse Obstacle Problem.
    Abstract: The aim of our work is to reconstruct an inclusion \omega immersed in a fluid flowing in a larger bounded domain \Omega via a boundary measurement on \partial\Omega. We study the inverse problem of reconstructing \omega thanks to the tools of shape optimization by minimizing a cost functional. We characterize the first and second order optimality conditions in order to make a numerical resolution. Thus, we explain why the inverse problem is ill-posed. Therefore we need some regularization methods to solve numerically this problem. Finally, we present some numerical simulations using a parametric method in order to confirm and complete our theoretical results.

    Last updated on 21 October 2013.