Dans ce papier, nous examinons l'influence des dérivées premières et secondes approximées sur l'algorithme de filtre de type région de confiance développé pour résoudre des problèmes d'optimisa-tion non-linéaire sans contraintes et proposé par Gould, Sainvitu et Toint dans [12]. Des résultats numériques effectués sur un ensemble de problèmes de petite taille provenant de la collection CUTEr décrivent l'effet de l'utilisation de dérivées approximées sur la robustesse et l'effi-cacité de la méthode de filtre de type région de confiance.
In this paper, we examine the influence of approximate first and/or second derivatives on the filter-trust-region algorithm designed for solving unconstrained nonlinear optimization problems and proposed by Gould, Sainvitu and Toint in [12]. Numerical experiments carried out on small-scaled unconstrained problems from the CUTEr collection describe the effect of the use of approximate derivatives on the robustness and the efficiency of the filter-trust-region method.
Keywords: unconstrained optimization, filter techniques, trust-region algorithms, approximate derivatives, numerical results
Mots-clés : optimisation sans contraintes, méthode de filtre, algorithme de type de région de confiance, dérivées approximées, résultats numériques
@article{RO_2009__43_3_309_0, author = {Sainvitu, Caroline}, title = {How much do approximate derivatives hurt filter methods ?}, journal = {RAIRO - Operations Research - Recherche Op\'erationnelle}, pages = {309--329}, publisher = {EDP-Sciences}, volume = {43}, number = {3}, year = {2009}, doi = {10.1051/ro/2009016}, mrnumber = {2567991}, zbl = {1175.65074}, language = {en}, url = {http://www.numdam.org/articles/10.1051/ro/2009016/} }
TY - JOUR AU - Sainvitu, Caroline TI - How much do approximate derivatives hurt filter methods ? JO - RAIRO - Operations Research - Recherche Opérationnelle PY - 2009 SP - 309 EP - 329 VL - 43 IS - 3 PB - EDP-Sciences UR - http://www.numdam.org/articles/10.1051/ro/2009016/ DO - 10.1051/ro/2009016 LA - en ID - RO_2009__43_3_309_0 ER -
%0 Journal Article %A Sainvitu, Caroline %T How much do approximate derivatives hurt filter methods ? %J RAIRO - Operations Research - Recherche Opérationnelle %D 2009 %P 309-329 %V 43 %N 3 %I EDP-Sciences %U http://www.numdam.org/articles/10.1051/ro/2009016/ %R 10.1051/ro/2009016 %G en %F RO_2009__43_3_309_0
Sainvitu, Caroline. How much do approximate derivatives hurt filter methods ?. RAIRO - Operations Research - Recherche Opérationnelle, Tome 43 (2009) no. 3, pp. 309-329. doi : 10.1051/ro/2009016. http://www.numdam.org/articles/10.1051/ro/2009016/
[1] A limited memory algorithm for bound constrained optimization. SIAM J. Sci. Comput. 16 (1995) 1190-1208. | MR | Zbl
, , and ,[2] Convergence of quasi-Newton matrices generated by the Symmetric Rank One update. Math. Program. 50 (1991) 177-196. | MR | Zbl
, and ,[3] Trust-Region Methods. MPS-SIAM Series on Optimization 1, SIAM, Philadelphia, USA (2000). | MR | Zbl
, and ,[4] Recent progress in unconstrained nonlinear optimization without derivatives. Math. Program. Ser. B 79 (1997) 397-414. | MR | Zbl
, and ,[5] Benchmarking optimization software with performance profiles. Math. Program. 91 (2002) 201-213. | MR | Zbl
and ,[6] Nonlinear programming without a penalty function. Math. Program. 91 (2002) 239-269. | MR | Zbl
and ,[7] A multidimensional filter algorithm for nonlinear equations and nonlinear least-squares. SIAM J. Optim. 15 (2005) 17-38. | MR | Zbl
, and ,[8] Solving the trust-region subproblem using the Lanczos method. SIAM J. Optim. 9 (1999) 504-525. | MR | Zbl
, , and ,[9] Sensitivity of trust-region algorithms on their parameters. 4OR, Quarterly Journal of Operations Research 3 (2005) 227-241. | MR | Zbl
, , and ,[10] CUTEr, a constrained and unconstrained testing environment, revisited ACM Trans. Math. Software 29 (2003) 373-394. | MR | Zbl
, and ,[11] GALAHAD - a library of thread-safe Fortran 90 packages for large-scale nonlinear optimization. ACM Trans. Math. Software 29 (2003) 353-372. | MR | Zbl
, and ,[12] A Filter-Trust-Region Method for Unconstrained Optimization. SIAM J. Optim. 16 (2005) 341-357. | MR | Zbl
, and ,[13] FILTRANE, a Fortran 95 Filter-Trust-Region Package for Solving Systems of Nonlinear Equalities, Nonlinear Inequalities and Nonlinear Least-Squares Problems. Technical report 03/15, Rutherford Appleton Laboratory, Chilton, Oxfordshire, UK (2003).
and ,[14] On the limited memory BFGS method for large scale optimization. Math. Program. Ser. B 45 (1989) 503-528. | MR | Zbl
and ,[15] Matrix conditionning and nonlinear optimization. Math. Program. 14 (1978) 149-160. | MR | Zbl
and ,Cité par Sources :