Journal of Convex Analysis 25 (2018), No. 2, 623--641
Copyright Heldermann Verlag 2018
Minimization of Quadratic Functions on Convex Sets without Asymptotes
Juan-Enrique Martinez-Legaz
Dep. d'Economia i d'Història Econòmica, Universitat Autónoma de Barcelona, 08193 Bellaterra, Spain
JuanEnrique.Martinez.Legaz@uab.es
Dominikus Noll
Institut de Mathématiques, Université de Toulouse, 118 route de Narbonne, 31062 Toulouse, France
dominikus.noll@math.univ-toulouse.fr
Wilfredo Sosa
Programa de Pôs-Graduação em Economia, Universidade Católica de Brasilia, Brazil
sosa@ucb.br
The classical Frank and Wolfe theorem states that a quadratic function which is bounded below on a convex polyhedron P attains its infimum on P. We investigate whether more general classes of convex sets F can be identified which have this Frank-and-Wolfe property. We show that the intrinsic characterizations of Frank-and-Wolfe sets hinge on asymptotic properties of these sets.
Keywords: Quadratic optimization problem, asymptotes, conic asymptotes, Motzkin decomposition, Frank and Wolfe theorem, complementarity problem.
MSC: 90C20, 90C26
[ Fulltext-pdf (155 KB)] for subscribers only.