The spherical constraint in Boolean quadratic programs

Jérôme Malick

Abstract: We propose a new approach to bound Boolean quadratic optimization problems. The idea is to re-express the Boolean constraints as one “spherical” constraint, whose dualization amounts to semidefinite least-squares problems. Studying this dualization provides an alternative interpretation of the sdp relaxation. It also reveals a new class of non-convex problems with no duality gap.