Geodesic
Reduced Hessian methods as a perturbed Newton–Lagrange method
Vestnik rossijskih universitetov. Matematika, Tome 29 (2024) no. 145, pp. 51-64 Cet article a éte moissonné depuis la source Math-Net.Ru

Voir la notice de l'article

For an equality-constrained optimization problem, we consider the possibility to interpret sequential quadratic programming methods employing the Hessian of the Lagrangian reduced to the null space of the constraints’ Jacobian, as a perturbed Newton–Lagrange method. We demonstrate that such interpretation with required estimates on perturbations is possible for certain sequences generated by variants of these methods making use of second-order corrections. This allows to establish, from a general perspective, superlinear convergence of such sequences, the property generally missing for the main sequences of the methods in question.
Keywords: equality-constrained optimization problem, sequential quadratic programming, reduced Hessian of the Lagrangian, perturbed Newton–Lagrange method framework, second-order corrections, superlinear convergence 
@article{VTAMU_2024_29_145_a4,
     author = {A. A. Volkov and A. F. Izmailov and E. I. Uskov},
     title = {Reduced {Hessian} methods as a perturbed {Newton{\textendash}Lagrange} method},
     journal = {Vestnik rossijskih universitetov. Matematika},
     pages = {51--64},
     year = {2024},
     volume = {29},
     number = {145},
     language = {ru},
     url = {http://geodesic.mathdoc.fr/item/VTAMU_2024_29_145_a4/}
}
TY  - JOUR
AU  - A. A. Volkov
AU  - A. F. Izmailov
AU  - E. I. Uskov
TI  - Reduced Hessian methods as a perturbed Newton–Lagrange method
JO  - Vestnik rossijskih universitetov. Matematika
PY  - 2024
SP  - 51
EP  - 64
VL  - 29
IS  - 145
UR  - http://geodesic.mathdoc.fr/item/VTAMU_2024_29_145_a4/
LA  - ru
ID  - VTAMU_2024_29_145_a4
ER  - 
%0 Journal Article
%A A. A. Volkov
%A A. F. Izmailov
%A E. I. Uskov
%T Reduced Hessian methods as a perturbed Newton–Lagrange method
%J Vestnik rossijskih universitetov. Matematika
%D 2024
%P 51-64
%V 29
%N 145
%U http://geodesic.mathdoc.fr/item/VTAMU_2024_29_145_a4/
%G ru
%F VTAMU_2024_29_145_a4
A. A. Volkov; A. F. Izmailov; E. I. Uskov. Reduced Hessian methods as a perturbed Newton–Lagrange method. Vestnik rossijskih universitetov. Matematika, Tome 29 (2024) no. 145, pp. 51-64. http://geodesic.mathdoc.fr/item/VTAMU_2024_29_145_a4/

[1] J. Nocedal, S. J. Wright, Numerical Optimization, 2nd ed., Springer, New York, 2006 | MR | Zbl

[2] A. F. Izmailov, M. V. Solodov, Chislenniye Metody Optimizatzii, Fizmatlit Publ., Moscow, 2008 (In Russian) | MR

[3] A. F. Izmailov, M. V. Solodov, Newton-Type Methods for Optimization and Variational Problems, Springer Series in Operations Research and Financial Engineering, Springer, Cham, 2014 | DOI | MR | Zbl

[4] R. H. Byrd, “An example of irregular convergence in some constrained optimization methods that use the projected Hessian”, Mathematical Programming, 32 (1985), 232–237 | DOI | MR | Zbl

[5] Y. Yuan, “An only 2-step Q-superlinear convergence example for some algorithms that use reduced Hessian approximations”, Mathematical Programming, 32 (1985), 224–231 | DOI | MR | Zbl

[6] R. H. Byrd, “On the convergence of constrained optimization methods with accurate Hessian information on a subspace”, SIAM J. on Numerical Analysis, 27 (1990), 141–153 | DOI | MR | Zbl

[7] R. H. Byrd, J. Nocedal, “An analysis of reduced Hessian methods for constrained optimization”, Mathematical Programming, 49 (1990), 285–323 | DOI | MR

[8] J. Nocedal, M. L. Overton, “Projected Hessian updating algorithms for nonlinearly constrained optimization”, SIAM J. on Numerical Analysis, 22 (1985), 821–850 | DOI | MR | Zbl

[9] J. F. Bonnans, J. Ch. Gilbert, C. Lemaréchal, C. Sagastizábal, Numerical Optimization: Theoretical and Practical Aspects, 2nd ed., Springer, Berlin, 2006 | MR | Zbl

[10] T. F. Coleman, A. R. Conn, “On the Local Convergence of a Quasi-Newton Method for a Nonlinear Programming Problem”, SIAM Journal on Numerical Analysis, 21:4 (1984) | DOI | MR | Zbl