On convex relaxations for quadratically constrained quadratic programming

We consider convex relaxations for the problem of minimizing a (possibly nonconvex) quadratic objective subject to linear and (possibly nonconvex) quadratic constraints. Let F denote the feasible region for the linear constraints. We first show that replacing the quadratic objective and constraint functions with their convex lower envelopes on F is dominated by an alternative methodology based on convexifying the range of the quadratic form (1;x)(1,x') for x in F. We next show that the use of "alphaBB" underestimators as computable estimates of convex lower envelopes is dominated by a relaxation of the convex hull of the quadratic form that imposes semidefiniteness and linear constraints on diagonal terms. Finally, we show that the use of a large class of "D.C." underestimators is dominated by a relaxation that combines semidefiniteness with RLT constraints.

Citation

Department of Management Sciences, University of Iowa, July 2010.

Article

Download

View On convex relaxations for quadratically constrained quadratic programming