Abstract.
The classical trust-region method for unconstrained minimization can be augmented with a line search that finds a point that satisfies the Wolfe conditions. One can use this new method to define an algorithm that simultaneously satisfies the quasi-Newton condition at each iteration and maintains a positive-definite approximation to the Hessian of the objective function. This new algorithm has strong global convergence properties and is robust and efficient in practice.
Similar content being viewed by others
References
Bongartz, I., Conn, A.R., Gould, N.I.M., Toint, Ph.L., CUTE: Constrained and unconstrained testing environment. Département de Mathématique Facultés Universitaires de Namur, 1993
Conn, A.R., Nicholas, I.M. Gould, Toint, P.L.: Trust-Region Methods. Society for Industrial and Applied Mathematics (SIAM), Philadelphia PA, 2000
Fletcher, R.: An Algorithm for Solving Linearly Constrained Optimization Problems. Math. Program. 2, 133–165 (1972)
Fletcher, R.: Practical Methods of Optimization. John Wiley and Sons 1987
Friedlander, A., Martínez, J. M., Santos, S. A.: A new trust region algorithm for bound constrained minimization. Appl. Math. Optim. 30, 235–266 (1994)
Gill, P.E., Goluband, G.H., Murray, W., Saunders, M. A.: Methods for modifying matrix factorizations. Math. Comp. 28, 505–535 (1974)
Michael Gertz, E.: Combination trust-region line-search methods for unconstrained optimization. University of California San Diego 1999
Hallabi, M.El: Globally convergent multi-level inexact hybrid algorithm for equality constrained optimization Département Informatique et Optimisation Institut National des Postes et Télécommunications 1999 Technical Report RT11-98 (revised) Rabat Morocco
Gill, P.E., Hammarling, S.J., Murray, W., Saunders, M.A., Wright, M.H.: Department of Operations Research Stanford University Stanford CA User’s guide for LSSOL (Version 1.0): A Fortran package for constrained linear least-squares and convex quadratic programming Report SOL 86-1 1986
Gill, Philip E., Wright, M.H.: Department of Mathematics University of California San Diego. Course Notes for Numerical Nonlinear Optimization 2001
Moré, Jorge J., Sorensen, D.C.: Computing a Trust Region Step. SIAM J. Sci. Stat. Comput. 4 (3), 553–572 sep(1983)
Nocedal, Jorge, Yuan, Ya Xiang : Combining Trust-Region and Line-Search Techniques. Optimization Technology Center mar OTC 98/04 1998
Powell, M.J.D.: A New Algorithm for Unconstrained Optimization. J. B. Rosen and O. L. Mangasarian and K. Ritter Nonlinear Programming Academic Press 1970
Powell, M.J.D.: A new algorithm for unconstrained optimization A.E.R.E. Harwell 1970 T.P. 382
Powell, M.J.D.: Convergence Properties of a Class of Minimization Algorithms. O. L. Mangasarian and R. R. Meyer and S. M. Robinson Nonlinear Programming 2 Academic Press 1975, pp. 1–27
Stoer, J.: On the numerical solution of constrained least squares problems. SIAM J. Numer. Anal. 8, 382–411 (1971)
Steihaug, T.: The Conjugate Gradient Method and Trust Regions in Large Scale Optimization. SIAM J. Numer. Anal. jun 20 (3), 626–637 1983
Thomas, S.W.: Sequential estimation techniques for quasi-Newton algorithms. Cornell University 1975
Toint, Ph.L.: Towards an efficient sparsity exploiting Newton method for minimization. Sparse Matrices and Their Uses Academic Press New York 1982 I. S. Duff 57–87
Wilson, P.B.: A Simplicial Algorithm for Concave Programming. Harvard University 1963
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Michael Gertz, E. A quasi-Newton trust-region method. Math. Program., Ser. A 100, 447–470 (2004). https://doi.org/10.1007/s10107-004-0511-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10107-004-0511-1