Abstract

We show that a new sufficient condition for the convergence with q-order two of the inexact Newton iterates may be obtained by considering the normwise backward error of the approximate steps and a result on perturbed Newton methods. This condition is in fact equivalent to the characterization given by Dembo, Eisenstat and Steihaug.

Authors

Emil Cătinaş
(Tiberiu Popoviciu Institute of Numerical Analysis, Romanian Academy)

Keywords

nonlinear system of equations in Rn; inexact Newton method; residual; local convergence; forcing term; q-convergence order.

Cite this paper as:

E. Cătinaş, A note on the quadratic convergence of the inexact Newton methods, Rev. Anal. Numér. Théor. Approx., 29 (2000) no. 2, pp. 129-133.

PDF

Scanned paper.

Latex-pdf version of the paper.

About this paper

Print ISSN

1222-9024

Online ISSN

2457-8126

MR

MR

Online ISSN

2457-8126

Google Scholar citations

References

Paper in html form

References

[1] E. Catinas, Inexact  perturbed  Newton  methods,  and  some  applications  for  a  class  of Krylov solvers, J. Optim. Theory Appl.,108 (2001) no. 3, pp. 543–570.

[2] E. Catinas, On  the  high  convergence  orders  of  the  Newton-GMBACK  methods, Rev. Anal. Num ́er. Th ́eor. Approx., 28 (1999) no. 2, pp. 125–132.

[3] E. Catinas, Newton and Newton-Krylov methods for solving nonlinear systems in Rn, Ph.D. thesis, ”Babes-Bolyai” University of Cluj–Napoca, Romania, 1999.

[4] E. Catinas, The  relationship  between  three  practical  models  of  Newton  methods  with high convergence orders , submitted.

[5] E.  Catinas, Inexact  perturbed  Newton  methods  for  nonlinear  systems  with  singular Jacobians, submitted.

[6] E. Catinas, The high convergence orders of the successive approximations, submitted.

[7] E.  Catinas, Finite  difference  approximations  in  the  Newton-Krylov  methods,  manuscript.

[8] D. Cores and R.A. Tapia, Perturbation Lemma for the Newton Method with Application  to  the  SQP  Newton  Method,  J. Optim. Theory Appl.,  97 (1998),  pp. 271–280.

[9] R. S. Dembo, S. C. Eisenstat and T. Steihaug, Inexact Newton methods, SIAM J. Numer. Anal., 19 (1982), pp. 400–408.

[10] J. E. Dennis, Jr. and R. B. Schnabel, Numerical Methods for Unconstrained Optimization and Nonlinear Equations, Prentice–Hall Series in Computational Mathematics, Englewood Cliffs, NJ, 1983.

[11] N. J. Higham, Accuracy  and  Stability  of  Numerical  Algorithms, SIAM, Philadelphia, PA, 1996.

[12] St.  Maruster, Numerical  Methods  for  Solving  Nonlinear  Equations,  Ed.  Tehnica, Bucharest, 1981 (in Romanian).

[13] J. M. Ortega and W. C. Rheinboldt, Iterative Solution of Nonlinear Equations in Several Variables, Academic Press, New York, 1970. 5 Quadratic convergence of the inexact Newton methods 133

[14] I.  Pavaloiu, Introduction  to  the  Approximation  of  the  Solutions  of  Equations,  Ed. Dacia, Cluj-Napoca, 1976 (in Romanian).

[15] F. A. Potra, On Q-order  and R-order  of  convergence, J. Optim. Theory Appl., 63 (1989), pp. 415–431.

[16] W. C. Rheinboldt, Methods for Solving Systems of Nonlinear Equations, Second ed., SIAM, Philadelphia, 1998.

[17] J. L. Rigal and J. Gaches, On the compatibility of a given solution with the data of a linear system, J. ACM, 14 (1967), pp. 543–548.

[18] H. F. Walker, An approach to continuation using Krylov subspace methods, Research Report  1/97/89,  Dept.  of  Math.,  Utah  State  University,  appeared  in  ComputationalScience in the 21st Century, M.–O. Bristeau, G. Etgen, W. Fitzgibbon, J. L. Lions, J.

Periaux and M. F. Wheeler, eds., John Wiley and Sons, Ltd., 1997, pp. 72–82

Related Posts

Menu