A finite branch-and-bound algorithm for nonconvex quadratic programming via semidefinite relaxations

Existing global optimization techniques for nonconvex quadratic programming (QP) branch by recursively partitioning the convex feasible set and thus generate an infinite number of branch-and-bound nodes. An open question of theoretical interest is how to develop a finite branch-and-bound algorithm for nonconvex QP. One idea, which guarantees a finite number of branching decisions, is to enforce the first-order Karush-Kuhn-Tucker (KKT) conditions through branching. In addition, such an approach naturally yields linear programming (LP) relaxations at each node. However, the LP relaxations are unbounded, a fact that precludes their use. In this paper, we propose and study semidefinite programming relaxations, which are bounded and hence suitable for use with finite KKT-branching. Computational results demonstrate the practical effectiveness of the method, with a particular highlight being that only a small number of nodes are required. This author was supported in part by NSF Grants CCR-0203426 and CCF-0545514.     

A simplicial branch-and-bound algorithm for solving quadratically constrained quadratic programs

We propose a branch-and-bound algorithm for solving nonconvex quadratically-constrained quadratic programs. The algorithm is novel in that branching is done by partitioning the feasible region into the Cartesian product of two-dimensional triangles and rectangles. Explicit formulae for the convex and concave envelopes of bilinear functions over triangles and rectangles are derived and shown to be second-order cone representable. The usefulness of these new relaxations is demonstrated both theoretically and computationally.     

A reformulation-convexification approach for solving nonconvex quadratic programming problems

In this paper, we consider the class of linearly constrained nonconvex quadratic programming problems, and present a new approach based on a novel Reformulation-Linearization/Convexification Technique. In this approach, a tight linear (or convex) programming relaxation, or outer-approximation to the convex envelope of the objective function over the constrained region, is constructed for the problem by generating new constraints through the process of employing suitable products of constraints and using variable redefinitions. Various such relaxations are considered and analyzed, including ones that retain some useful nonlinear relationships. Efficient solution techniques are then explored for solving these relaxations in order to derive lower and upper bounds on the problem, and appropriate branching/partitioning strategies are used in concert with these bounding techniques to derive a convergent algorithm. Computational results are presented on a set of test problems from the literature to demonstrate the efficiency of the approach. (One of these test problems had not previously been solved to optimality). It is shown that for many problems, the initial relaxation itself produces an optimal solution.     

Note on a sufficient condition for a local minimum of a class of nonconvex quadratic programs

A counterexample is given to show that a previously proposed sufficient condition for a local minimum of a class of nonconvex quadratic programs is not correct. This class of problems arises in combinatorial optimization. The problem with the original proof is pointed out. © 1998 The Mathematical Programming Society, Inc. Published by Elsevier Science B.V.     

Necessary and sufficient condition for local minima of a class of nonconvex quadratic programs

The author (1992, 1993) earlier studied the equivalence of a class of 0–1 quadratic programs and their relaxed problems. Thus, a class of combinatorial optimization problems can be solved by solving a class of nonconvex quadratic programs. In this paper, a necessary and sufficient condition for local minima of this class of nonconvex quadratic programs is given; this will be the foundation for study of algorithms.Research supported by Huo Yingdong Educational Foundation '93.     

On affine scaling algorithms for nonconvex quadratic programming

We investigate the use of interior algorithms, especially the affine-scaling algorithm, to solve nonconvex — indefinite or negative definite — quadratic programming (QP) problems. Although the nonconvex QP with a polytope constraint is a hard problem, we show that the problem with an ellipsoidal constraint is easy. When the hard QP is solved by successively solving the easy QP, the sequence of points monotonically converge to a feasible point satisfying both the first and the second order optimality conditions.Research supported in part by NSF Grant DDM-8922636 and the College Summer Grant, College of Business Administration, The University of Iowa.     

Global optimization of a class of nonconvex quadratically constrained quadratic programming problems

In this paper we study a class of nonconvex quadratically constrained quadratic programming problems generalized from relaxations of quadratic assignment problems. We show that each problem is polynomially solved. Strong duality holds if a redundant constraint is introduced. As an application, a new lower bound is proposed for the quadratic assignment problem.     

Hidden convexity in some nonconvex quadratically constrained quadratic programming

We consider the problem of minimizing an indefinite quadratic objective function subject to twosided indefinite quadratic constraints. Under a suitable simultaneous diagonalization assumption (which trivially holds for trust region type problems), we prove that the original problem is equivalent to a convex minimization problem with simple linear constraints. We then consider a special problem of minimizing a concave quadratic function subject to finitely many convex quadratic constraints, which is also shown to be equivalent to a minimax convex problem. In both cases we derive the explicit nonlinear transformations which allow for recovering the optimal solution of the nonconvex problems via their equivalent convex counterparts. Special cases and applications are also discussed. We outline interior-point polynomial-time algorithms for the solution of the equivalent convex programs. This author's work was partially supported by GIF, the German-Israeli Foundation for Scientific Research and Development and by the Binational Science Foundation. This author's work was partially supported by National Science Foundation Grants DMS-9201297 and DMS-9401871.     

A New Semidefinite Programming Bound for Indefinite Quadratic Forms Over a Simplex

The paper describes a method for computing a lower bound of the global minimum of an indefinite quadratic form over a simplex. The bound is derived by computing an underestimator of the convex envelope by solving a semidefinite program (SDP). This results in a convex quadratic program (QP). It is shown that the optimal value of the QP is a lower bound of the optimal value of the original problem. Since there exist fast (polynomial time) algorithms for solving SDP's and QP's the bound can be computed in reasonable time. Numerical experiments indicate that the relative error of the bound is about 10 percent for problems up to 20 variables, which is much better than a known SDP bound.     

Dual support method for solving convex quadratic programs

In this article, we present a new dual method for solving convex (but not strictly convex) quadratic programs (QPs). Our method is the generalization of the dual support method, developed by Gabasov and co-workers in 1981, for solving convex QPs. It proceeds in two phases: the first is to construct the initial support, called coordinator support, for the problem and the second is to achieve the optimality of the problem. Results of numerical experiments are given comparing our approach with the active-set method.     

A relaxation method for nonconvex quadratically constrained quadratic programs

We present an algorithm for finding approximate global solutions to quadratically constrained quadratic programming problems. The method is based on outer approximation (linearization) and branch and bound with linear programming subproblems. When the feasible set is non-convex, the infinite process can be terminated with an approximate (possibly infeasible) optimal solution. We provide error bounds that can be used to ensure stopping within a prespecified feasibility tolerance. A numerical example illustrates the procedure. Computational experiments with an implementation of the procedure are reported on bilinearly constrained test problems with up to sixteen decision variables and eight constraints.This research was supported in part by National Science Foundation Grant DDM-91-14489.     

Lagrangian decomposition and mixed-integer quadratic programming reformulations for probabilistically constrained quadratic programs

Probabilistically constrained quadratic programming (PCQP) problems arise naturally from many real-world applications and have posed a great challenge in front of the optimization society for years due to the nonconvex and discrete nature of its feasible set. We consider in this paper a special case of PCQP where the random vector has a finite discrete distribution. We first derive second-order cone programming (SOCP) relaxation and semidefinite programming (SDP) relaxation for the problem via a new Lagrangian decomposition scheme. We then give a mixed integer quadratic programming (MIQP) reformulation of the PCQP and show that the continuous relaxation of the MIQP is exactly the SOCP relaxation. This new MIQP reformulation is more efficient than the standard MIQP reformulation in the sense that its continuous relaxation is tighter than or at least as tight as that of the standard MIQP. We report preliminary computational results to demonstrate the tightness of the new convex relaxations and the effectiveness of the new MIQP reformulation.     

Decomposition methods for solving a class of nonconvex programming problems dealing with bilinear and quadratic functions

We develop convergent decomposition branch and bound algorithms for solving a class of bilinear programming problems. As an application of the proposed method, we apply it to quadratic programs with a few negative eigenvalues, and to a class of mixed integer programming problems.This paper was completed during the stay of the first author at LMI-INSA Rouen, CNRS URA 1378, France.     

On the convergence of two branch-and-bound algorithms for nonconvex programming problems

This note presents a new convergence property for each of two branch-and-bound algorithms for nonconvex programming problems (Falk-Soland algorithms and Horst algorithms). For each algorithm, it has been shown previously that, under certain conditions, whenever the algorithm generates an infinite sequence of points, at least one accumulation point of this sequence is a global minimum. We show here that, for each algorithm, in fact, under these conditions, every accumulation point of such a sequence is a global minimum.The author would like to thank Professor R. M. Soland for his helpful comments concerning this paper.     

The semi-continuous quadratic mixture design problem: Description and branch-and-bound approach

The semi-continuous quadratic mixture design problem (SCQMDP) is described as a problem with linear, quadratic and semi-continuity constraints. Moreover, a linear cost objective and an integer valued objective are introduced. The goal is to deal with the SCQMD problem from a branch-and-bound perspective generating robust solutions. Therefore, an algorithm is outlined which identifies instances where decision makers tighten requirements such that no ε-robust solution exists. The algorithm is tested on several cases derived from industry.     

The equivalence of optimal perspective formulation and Shor's SDP for quadratic programs with indicator variables

In this paper, we compare the strength of the optimal perspective relaxation and Shor's SDP relaxation for convex quadratic optimization problems with indicator variables and prove that they are equivalent.     

Robust global optimization with polynomials

We consider the optimization problems max_z∈Ω min_x∈K p(z, x) and min_{x ∈ K} max_{z ∈ Ω} p(z, x) where the criterion p is a polynomial, linear in the variables z, the set Ω can be described by LMIs, and K is a basic closed semi-algebraic set. The first problem is a robust analogue of the generic SDP problem max_{z ∈ Ω} p(z), whereas the second problem is a robust analogue of the generic problem min_{x ∈ K} p(x) of minimizing a polynomial over a semi-algebraic set. We show that the optimal values of both robust optimization problems can be approximated as closely as desired, by solving a hierarchy of SDP relaxations. We also relate and compare the SDP relaxations associated with the max-min and the min-max robust optimization problems.     

Some Fundamental Properties of Successive Convex Relaxation Methods on LCP and Related Problems

General successive convex relaxation methods (SRCMs) can be used to compute the convex hull of any compact set, in an Euclidean space, described by a system of quadratic inequalities and a compact convex set. Linear complementarity problems (LCPs) make an interesting and rich class of structured nonconvex optimization problems. In this paper, we study a few of the specialized lift-and-project methods and some of the possible ways of applying the general SCRMs to LCPs and related problems.     

Monotone projected gradient methods for large-scale box-constrained quadratic programming

Inspired by the success of the projected Barzilai-Borwein (PBB) method for large-scale box-constrained quadratic programming, we propose and analyze the monotone projected gradient methods in this paper. We show by experiments and analyses that for the new methods, it is generally a bad option to compute steplengths based on the negative gradients. Thus in our algorithms, some continuous or discontinuous projected gradients are used instead to compute the steplengths. Numerical experiments on a wide variety of test problems are presented, indicating that the new methods usually outperform the PBB method.