Global minimization of difference of quadratic and convex functions over box or binary constraints

In this paper, we present necessary as well as sufficient conditions for a given feasible point to be a global minimizer of the difference of quadratic and convex functions subject to bounds on the variables. We show that the necessary conditions become necessary and sufficient for global minimizers in the case of a weighted sum of squares minimization problems. We obtain sufficient conditions for global optimality by first constructing quadratic underestimators and then by characterizing global minimizers of the underestimators. We also derive global optimality conditions for the minimization of the difference of quadratic and convex functions over binary constraints. We discuss several numerical examples to illustrate the significance of the optimality conditions.

[1]  P. Pardalos,et al.  Optimization in computational chemistry and molecular biology : local and global approaches , 2000 .

[2]  Marc Teboulle,et al.  Global Optimality Conditions for Quadratic Optimization Problems with Binary Constraints , 2000, SIAM J. Optim..

[3]  Hoai An Le Thi,et al.  A continuous approch for globally solving linearly constrained quadratic , 2001 .

[4]  Arkadi Nemirovski,et al.  Lectures on modern convex optimization - analysis, algorithms, and engineering applications , 2001, MPS-SIAM series on optimization.

[5]  Eranda Çela,et al.  The quadratic assignment problem : theory and algorithms , 1999 .

[6]  Panos M. Pardalos,et al.  Computational aspects of a branch and bound algorithm for quadratic zero-one programming , 1990, Computing.

[7]  Roummel F. Marcia,et al.  Iterative Convex Quadratic Approximation for Global Optimization in Protein Docking , 2005, Comput. Optim. Appl..

[8]  Michael Jünger,et al.  Quadratic 0/1 optimization and a decomposition approach for the placement of electronic circuits , 1994, Math. Program..

[9]  Christodoulos A. Floudas,et al.  Computational Experience with a New Class of Convex Underestimators: Box-constrained NLP Problems , 2004, J. Glob. Optim..

[10]  Zhi-You Wu,et al.  Conditions for Global Optimality of Quadratic Minimization Problems with LMI Constraints , 2007, Asia Pac. J. Oper. Res..

[11]  Panos M. Pardalos,et al.  Optimization in Computational Chemistry and Molecular Biology , 2000 .

[12]  M. Pinar,et al.  Sufficient Global Optimality Conditions for Bivalent Quadratic Optimization , 2004 .

[13]  Jean-Baptiste Hiriart-Urruty,et al.  Global Optimality Conditions in Maximizing a Convex Quadratic Function under Convex Quadratic Constraints , 2001, J. Glob. Optim..

[14]  J. Ben Rosen,et al.  Nonconvex Piecewise-Quadratic Underestimation for Global Minimization , 2006, J. Glob. Optim..

[15]  N. Q. Huy,et al.  Sufficient global optimality conditions for multi-extremal smooth minimisation problems with bounds and linear matrix inequality constraints , 2006, The ANZIAM Journal.

[16]  Geir Dahl,et al.  A note on diagonally dominant matrices , 2000 .

[17]  P. Pardalos,et al.  Handbook of global optimization , 1995 .

[18]  Zhi-You Wu,et al.  Non-convex quadratic minimization problems with quadratic constraints: global optimality conditions , 2007, Math. Program..

[19]  R. Horst,et al.  Necessary and Sufficient Global Optimality Conditions for Convex Maximization Revisited , 1998 .

[20]  Zhi-You Wu,et al.  Sufficient Global Optimality Conditions for Non-convex Quadratic Minimization Problems With Box Constraints , 2006, J. Glob. Optim..

[21]  Christodoulos A. Floudas,et al.  A New Class of Improved Convex Underestimators for Twice Continuously Differentiable Constrained NLPs , 2004, J. Glob. Optim..