Comptes Rendus
Sensitivity of solutions computed through the Asymptotic Numerical Method
Comptes Rendus. Mécanique, Volume 336 (2008) no. 10, pp. 788-793.

The Asymptotic Numerical Method (ANM) allows one to compute solution branches of sufficiently smooth non-linear PDE problems using truncated Taylor expansions. The Diamant approach of the ANM has been proposed for hiding definitively the differentiation aspects to the user. In this Note, this significant improvement in terms of genericity is exploited to compute the sensitivity of ANM solutions with respect to modelling parameters. The differentiation in the parameters is discussed at both the equation and code level to highlight the Automatic Differentiation (AD) purposes. A numerical example proves the interest of such techniques for a generic and efficient implementation of sensitivity computations.

La Méthode Asymptotique Numérique (MAN) permet de calculer des branches de solutions de problèmes d'EDP suffisamment réguliers à l'aide de séries de Taylor tronquées. L'approche Diamant de la MAN a été proposée pour cacher définitivement les aspects différentiation à l'utilisateur. Dans cette Note, cette amélioration significative en terme de généricité est exploitée pour calculer la sensibilité des solutions MAN par rapport aux paramètres de modélisation. La différentiation en les paramètres est discutée au niveau des équations et du code pour souligner les aspects Différentiation Automatique (DA). Un exemple numérique prouve l'intérêt de ces techniques pour l'implémentation générique et efficace de calculs de sensibilité.

Received:
Accepted:
Published online:
DOI: 10.1016/j.crme.2008.09.003
Keywords: Continuation, ANM, Diamant, Sensitivity analysis
Mot clés : Continuation, MAN, Diamant, Analyse de sensibilité

Isabelle Charpentier 1

1 Laboratoire de physique et mécanique des matériaux, UMR 7554, Île du Saulcy, 57045 Metz cedex 1, France
@article{CRMECA_2008__336_10_788_0,
     author = {Isabelle Charpentier},
     title = {Sensitivity of solutions computed through the {Asymptotic} {Numerical} {Method}},
     journal = {Comptes Rendus. M\'ecanique},
     pages = {788--793},
     publisher = {Elsevier},
     volume = {336},
     number = {10},
     year = {2008},
     doi = {10.1016/j.crme.2008.09.003},
     language = {en},
}
TY  - JOUR
AU  - Isabelle Charpentier
TI  - Sensitivity of solutions computed through the Asymptotic Numerical Method
JO  - Comptes Rendus. Mécanique
PY  - 2008
SP  - 788
EP  - 793
VL  - 336
IS  - 10
PB  - Elsevier
DO  - 10.1016/j.crme.2008.09.003
LA  - en
ID  - CRMECA_2008__336_10_788_0
ER  - 
%0 Journal Article
%A Isabelle Charpentier
%T Sensitivity of solutions computed through the Asymptotic Numerical Method
%J Comptes Rendus. Mécanique
%D 2008
%P 788-793
%V 336
%N 10
%I Elsevier
%R 10.1016/j.crme.2008.09.003
%G en
%F CRMECA_2008__336_10_788_0
Isabelle Charpentier. Sensitivity of solutions computed through the Asymptotic Numerical Method. Comptes Rendus. Mécanique, Volume 336 (2008) no. 10, pp. 788-793. doi : 10.1016/j.crme.2008.09.003. https://comptes-rendus.academie-sciences.fr/mecanique/articles/10.1016/j.crme.2008.09.003/

[1] J.M.T. Thompson; A. Walker The nonlinear perturbation analysis of discrete structural systems, International Journal of Solids and Structures, Volume 4 (1968), pp. 757-768

[2] B. Cochelin; N. Damil; M. Potier-Ferry Asymptotic numerical method and Padé approximants for non-linear elastic structures, International Journal for Numerical Methods in Engineering, Volume 37 (1994), pp. 1187-1213

[3] B. Cochelin; N. Damil; M. Potier-Ferry Méthode asymptotique numérique, Hermes Science Publications, 2007 (p. 298)

[4] A. Griewank Evaluating Derivatives: Principles and Techniques of Algorithmic Differentiation, Frontiers in Appl. Math., vol. 11, SIAM, Philadelphia, 2000 (p. 369)

[5] I. Charpentier; M. Potier-Ferry Différentiation automatique de la Méthode Asymptotique Numérique Typée : l'approche Diamant, Comptes Rendus Mecanique, Volume 336 (2008), pp. 336-340

[6] I. Charpentier; A. Lejeune; M. Potier-Ferry The diamant approach for an efficient automatic differentiation of the asymptotic numerical method (C.H. Bischof; H.M. Bücker; P. Hovland; U. Naumann; J. Utke, eds.), Advances in Automatic Differentiation, Lecture Notes in Computational Science and Engineering, vol. 64, 2008, pp. 139-149

[7] Y. Koutsawa, I. Charpentier, E.M. Daya, M. Cherkaoui, A generic approach for the solution of nonlinear residual equations. Part I: The Diamant toolbox, Computer Methods in Applied Mechanics and Engineering, | DOI

[8] L. Hascoët, V. Pascual, TAPENADE 2.1 user's guide, Rapport technique de l'INRIA RT-0300, 2004

[9] E. Doedel; H. Keller; J.P. Kernevez Numerical analysis and control of bifurcation problems (I). Bifurcation in finite dimensions, International Journal of Bifurcation and Chaos, Volume 1 (1991), pp. 493-520

[10] J. Morgenstern How to compute fast a function and all its derivatives, a variation on the theorem of Baur–Strassen, SIGACT News, Volume 16 (1985), pp. 60-62

[11] I. Charpentier; M. Ghémirès Efficient adjoint derivatives: Application to the atmospheric model MESO-NH, Optimization Methods and Software, Volume 13 (2000), pp. 35-63

Cited by Sources:

Comments - Policy