|
|
 |
|
Mardi 19 Février
| Heure: |
12:30 - 13:30 |
| Lieu: |
Salle B107, bâtiment B, Université de Villetaneuse |
| Résumé: |
Discrete optimization using semidefinite methods |
| Description: |
Angelika Wiegele Many real-world applications, although being non-linear, can be well
described by linearized models. Therefore, Linear Programming (LP)
became a widely studied and applied technique in many areas of science,
industry and economy. Semidefinite Programming (SDP) is an extension of
LP. A matrix-variable is optimized over the intersection of the cone of
positive semidefinite matrices with an affine space. It turned out, that
SDP can provide significantly stronger practical results than LP. Since
then SDP turned out to be practical in a lot of different areas, like
combinatorial optimization, control theory, engineering, and more
recently in polynomial optimization.
In this talk I will present some ideas how to model discrete
optimization problems using semidefinite programming in order to obtain
semidefinite relaxations. Some of this relaxations proved to be
succesful when using in a branch-and-bound framework. Furthermore, I
want to present a new idea of how to strengthen semidefinite relaxations.  |
|
|
