2018

Heure:	12:30 - 13:30
Lieu:	Salle B107, bâtiment B, Université de Villetaneuse
Résumé:	Decomposition methods for quadratic programming
Description:	Emiliano Traversi The purpose of this talk is to present two decomposition methods for quadratic problems. First, we propose a methodological analysis on a family of reformulations combining Dantzig-Wolfe decomposition and Quadratic Convex Reformulation principles for binary quadratic problems. As a representative case study, we apply them to a cardinality constrained quadratic knapsack problem. Secondly, we analyze a simplicial decomposition like algorithmic framework that handles convex quadratic programs in an effective way. In particular, we propose two tailored strategies for solving the master problem and we describe a few techniques for speeding up the solution of the pricing problem. We report extensive numerical experiments on both real-world and generic quadratic programs.