Dipartimento di Ingegneria informatica, automatica e gestionale

In this paper we consider Quadratic Programming (QP) problems with general linear constraints. We show, through a computational investigation, that a careful selection of a suitable reformulation of such problems, together with the related relaxation, and an intensive application of bound tightening are simple but very effective ingredients in order to make a standard branch and bound approach very competitive and in some cases able to outperform even well known commercial solvers.