Construction of test problems in quadratic bivalent programming

Abstract

A method of constructing test problems for constrained bivalent quadratic programming is presented. For any feasible integer point for a given domain, the method generates quadratic functions whose minimum over the given domain occurs at the selected point. Certain properties of unconstrained quadratic zero-one programs that determine the difficulty of the test problems are also discussed. In addition, a standardized random test problem generator for unconstrained quadratic zero-one programming is given.