Geodesic
Convergence of primal-dual solutions for the nonconvex log-barrier method without LICQ
Kybernetika, Tome 40 (2004) no. 5, pp. 571-584 Cet article a éte moissonné depuis la source Czech Digital Mathematics Library

Voir la notice de l'article

This paper characterizes completely the behavior of the logarithmic barrier method under a standard second order condition, strict (multivalued) complementarity and MFCQ at a local minimizer. We present direct proofs, based on certain key estimates and few well–known facts on linear and parametric programming, in order to verify existence and Lipschitzian convergence of local primal-dual solutions without applying additionally technical tools arising from Newton–techniques.
This paper characterizes completely the behavior of the logarithmic barrier method under a standard second order condition, strict (multivalued) complementarity and MFCQ at a local minimizer. We present direct proofs, based on certain key estimates and few well–known facts on linear and parametric programming, in order to verify existence and Lipschitzian convergence of local primal-dual solutions without applying additionally technical tools arising from Newton–techniques.
Classification : 49K40, 49M37, 65K10, 90C30
Keywords: log-barrier method; Mangasarian–Fromovitz constraint qualification; convergence ofprimal-dual solutions; locally linearized problems; Lipschitz estimates 
@article{KYB_2004_40_5_a3,
     author = {Grossmann, Christian and Klatte, Diethard and Kummer, Bernd},
     title = {Convergence of primal-dual solutions for the nonconvex log-barrier method without {LICQ}},
     journal = {Kybernetika},
     pages = {571--584},
     year = {2004},
     volume = {40},
     number = {5},
     mrnumber = {2120997},
     zbl = {1249.90252},
     language = {en},
     url = {http://geodesic.mathdoc.fr/item/KYB_2004_40_5_a3/}
}
TY  - JOUR
AU  - Grossmann, Christian
AU  - Klatte, Diethard
AU  - Kummer, Bernd
TI  - Convergence of primal-dual solutions for the nonconvex log-barrier method without LICQ
JO  - Kybernetika
PY  - 2004
SP  - 571
EP  - 584
VL  - 40
IS  - 5
UR  - http://geodesic.mathdoc.fr/item/KYB_2004_40_5_a3/
LA  - en
ID  - KYB_2004_40_5_a3
ER  - 
%0 Journal Article
%A Grossmann, Christian
%A Klatte, Diethard
%A Kummer, Bernd
%T Convergence of primal-dual solutions for the nonconvex log-barrier method without LICQ
%J Kybernetika
%D 2004
%P 571-584
%V 40
%N 5
%U http://geodesic.mathdoc.fr/item/KYB_2004_40_5_a3/
%G en
%F KYB_2004_40_5_a3
Grossmann, Christian; Klatte, Diethard; Kummer, Bernd. Convergence of primal-dual solutions for the nonconvex log-barrier method without LICQ. Kybernetika, Tome 40 (2004) no. 5, pp. 571-584. http://geodesic.mathdoc.fr/item/KYB_2004_40_5_a3/

[1] Adler I., Monteiro R. D. C.: Limiting behavior of the affine scaling continuous trajectories for linear programming problems. Math. Programming 50 (1991), 29–51 | DOI | MR | Zbl

[2] Bank B., Guddat J., Klatte D., Kummer, B., Tammer K.: Non-linear Parametric Optimization. Akademie–Verlag, Berlin 1982 | Zbl

[3] El-Bakry A. S., Tapia R. A., Zhang Y.: On the convergence rate of Newton interior-point methods in the absence of strict complementarity. Comput. Optim. Appl. 6 (1996), 157–167 | DOI | MR | Zbl

[4] Fiacco A. V., McCormick G. P.: Nonlinear Programming: Sequential Unconstrained Minimization Techniques. Wiley, New York 1968 | MR | Zbl

[5] Forsgren A., Gill P. E., Wright M. H.: Interior methods for nonlinear optimization. SIAM Rev. 44 (2002), 525–597 | DOI | MR | Zbl

[6] Grossmann C.: Convergence analysis for penalty/barrier path-following in linearly constrained convex programming. Discuss. Math.: Differential Inclusions, Control and Optimization 20 (2000), 7–26 | MR

[7] Grossmann C., Kaplan A. A.: Strafmethoden und modifizierte Lagrange Funktionen in der nichtlinearen Optimierung. Teubner, Leipzig 1979 | MR | Zbl

[8] Grossmann C., Schöniger G.: Sensitivität und Anwendbarkeit von Straf–Barriere–Methoden. Z. Angew. Math. Mech. 57 (1977), 419–430 | DOI | Zbl

[9] Guddat J., Guerra, F., Nowack D.: On the role of the Mangasarian-Fromovitz constraint qualification for penalty-, exact penalty-, and Lagrange multiplier methods. In: Mathematical programming with data perturbations (A. Fiacco, ed.), (Lecture Notes in Pure and Appl. Mathematics 195). Marcel Dekker, Dordrecht 1997, pp. 159–183 | MR

[10] Güler O.: Limiting behavior of weighted central paths in linear programming. Math. Programming 65A (1994), 347–363 | DOI | MR | Zbl

[11] Klatte D., Kummer B.: Nonsmooth Equations in Optimization – Regularity, Calculus, Methods and Applications. Kluwer, Dordrecht 2002 | MR | Zbl

[12] Kummer B.: On solvability and regularity of a parametrized version of optimality conditions. Z. Oper. Res.: Math. Meth. Oper. Res. 45 (1995), 215–230 | MR | Zbl

[13] Kummer B.: Parametrizations of Kojima’s system and relations to penalty and barrier functions. Math. Programming, Ser. B 76 (1997), 579–592 | DOI | MR | Zbl

[14] Nožička F., Guddat J., Hollatz, H., Bank B.: Theorie der linearen parametrischen Optimierung. Akademie–Verlag, Berlin 1974 | Zbl

[15] Owen G.: Spieltheorie. Springer–Verlag, Berlin 1971 (Translation) | MR | Zbl

[16] Ralph D., Wright S. J.: Superlinear convergence of an interior point method despite dependent constraints. Math. Oper. Res. 25 (2000), 179–194 | DOI | MR | Zbl

[17] Robinson S. M.: Generalized equations and their solutions. Part II: Applications to nonlinear programming. Math. Programming Stud. 19 (1982), 200–221 | DOI | MR | Zbl

[18] Wright S. J.: Superlinear convergence of a stabilized SQP method to a degenerate solution. Comput. Optim. Appl. 11 (1998), 253–275 | DOI | MR | Zbl

[19] Wright S. J.: On the convergence of the Newton/log-barrier method. Math. Programming 90A (2001), 71–100 | DOI | MR | Zbl

[20] Wright S. J.: Effects of finite-precision arithmetic on interior-point methods for nonlinear programming. SIAM J. Optim. 12 (2001), 36–78 | DOI | MR | Zbl

[21] Wright S. J., Orban D.: Properties of the Log-barrier Function on Degenerate Nonlinear Programs. Preprint ANL/MCS-P772-0799, Mathematics and Computer Science Division, Argonne National Laboratory, Argonne 1999, revised May 2001 | MR