Abstract
Some recent algorithms for nonsmooth optimization require solutions to certain piecewise quadratic programming subproblems. Two types of subproblems are considered in this paper. The first type seeks the minimization of a continuously differentiable and strictly convex piecewise quadratic function subject to linear equality constraints. We prove that a nonsmooth version of Newton’s method is globally and finitely convergent in this case. The second type involves the minimization of a possibly nonconvex and nondifferentiable piecewise quadratic function over a Euclidean ball. Characterizations of the global minimizer are studied under various conditions. The results extend a classical result on the trust region problem.
Similar content being viewed by others
References
C. Chen and O.L. Mangasarian, A class of smoothing functions for nonlinear and mixed complementarity problems,Computational Optimization and Applications 5 (1996) 97–138.
F.H. Clarke,Optimization and Nonsmooth Analysis (John Wiley and Sons, New York, 1983).
J.E. Dennis Jr., S.-B.B. Li and R.A. Tapia, A unified approach to global convergence of trust region methods for nonsmooth optimization,Mathematical Programming 68 (1995) 319–346.
M.R. Garey, D.S. Johnson and L. Stockmeyer, Some simplified NP-complete problems,Theoretical Computer Science 1 (1976) 237–268.
D.M. Gay, Computing optimal locally constrained steps,SIAM Journal on Scientific Statistical Computing 2 (1981) 186–197.
J. Kreimer and R.Y. Rubinstein, Nondifferentiable optimization via smooth approximation: General analytical approach,Annals of Operations Research 39 (1992) 97–119.
O.L. Mangasarian, Least-norm linear programming solution as an unconstrained minimization problem,Mathematical Analysis and its Applications 92 (1983) 240–251.
J.M. Martínez and S.A. Santos, A trust region strategy for minimization on arbitrary domains,Mathematical Programming 68 (1995) 267–302.
J.J. Moré, Generalizations of the trust region problem, Preprint MCS-P349-0193. Argonne National Lab. (Argonne, IL. 1993).
J.J. Moré and D.C. Sorensen, Computing a trust region step,SIAM Journal on Scientific Statistical Computing 4 (1983) 553–572.
J.-S. Pang, S.-P. Han and N. Rangaraj, Minimization of locally Lipschitzian functions,SIAM Journal on Optimization 1 (1991) 57–82.
J.-S. Pang and L. Qi, Nonsmooth equations: motivation and algorithms,SIAM Journal on Optimization 3 (1993) 443–465.
M.C. Pinar and S. Zenios, On smoothing exact penalty functions for convex constrained optimization,SIAM Journal on Optimization 4 (1994) 486–511.
L. Qi, Convergence analysis for some algorithms for solving nonsmooth equations,Mathematics of Operations Research 18 (1993) 227–244.
L. Qi and J. Sun, A nonsmooth version of Newton’s method,Mathematical Programming 58 (1993) 353–367.
L. Qi and J. Sun, A trust region method for minimization of locally Lipschitzian functions,Mathematical Programming 66 (1994) 25–43.
R.T. Rockafellar, Monotone operators and the proximal point algorithm,SIAM Journal on Control and Optimization 14 (1976) 877–898.
R.T. Rockafellar, Proximal subgradients, marginal values, and augmented Lagrangians in nonconvex optimization,Mathematics of Operations Research 6 (1981) 427–437.
R.T. Rockafellar,The Theory of Subgradients and its Applications to Problems of Optimization, Convex and Nonconvex Functions (Heldermann, Berlin, 1981).
R.T. Rockafellar and J. Sun, A finite simplex-active-set method for monotropic piecewise quadratic programming. in: D. Du and J. Sun, eds.,Advances in Optimization and Approximation (Kluwer Academic, Dordrecht, 1994) 275–292.
S. Scholtes, Introduction to piecewise differentiable equations. Preprint No. 53/1994, Institut für Statistik und Mathematische Wirtschaftstheorie. University of Karlsruhe (Karlsruhe, 1994).
D.C. Sorensen, Newton’s method with a model trust region modification,SIAM Journal on Numerical Analysis 19 (1982) 409–426.
R.J. Stern and H. Wolkowicz, Indefinite trust region subproblems and nonsymmetric eigenvalue perturbations,SIAM Journal on Optimization 5 (1995) 286–313.
J. Sun and H. Kuo, Applying a Newton method to strictly convex separable network quadratic programs. Technical Report. Department of Decision Sciences, National University of Singapore (1994).
J. Sun and L. Qi, An interior point algorithm of O(√m|ln ε|) iterations forC 1-convex programming,Mathematical Programming 57 (1992) 239–257.
Y. Ye, On affine scaling algorithms for nonconvex quadratic programming,Mathematical Programming 56 (1992) 285–300.
Author information
Authors and Affiliations
Additional information
Partially supported by National University of Singapore under grant 930033.
Rights and permissions
About this article
Cite this article
Sun, J. On piecewise quadratic Newton and trust region problems. Mathematical Programming 76, 451–467 (1997). https://doi.org/10.1007/BF02614393
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF02614393