Publication: On the Use of a Differentiated Finite Element Package for Sensitivity Analysis
Introduction
Applications
Tools
Research Groups
Workshops
Publications
   List Publications
   Advanced Search
   Info
   Add Publications
My Account
About
Impress

On the Use of a Differentiated Finite Element Package for Sensitivity Analysis

- Part of a collection -
 

Area
Computational Fluid Dynamics

Author(s)
C. H. Bischof , H. M. Bücker , B. Lang , A. Rasch , J. W. Risch

Published in
Computational Science -- ICCS 2001, Proceedings of the International Conference on Computational Science, San Francisco, USA, May 28--30, 2001. Part I

Editor(s)
V. N. Alexandrov, J. J. Dongarra, B. A. Juliano, R. S. Renner, C. J. K. Tan

Year
2001

Publisher
Springer

Abstract
Derivatives are ubiquitous in various areas of computational science including sensitivity analysis and parameter optimization of computer models. Among the various methods for obtaining derivatives, automatic differentiation (ad) combines freedom from approximation errors, high performance, and the ability to handle arbitrarily complex codes arising from large-scale scientific investigations. In this note, we show how ad technology can aid in the sensitivity analysis of a computer model by considering a classic fluid flow experiment as an example. To this end, the software tool ADIFOR implementing the ad technology for functions written in Fortran 77 was applied to the large finite element package SEPRAN. Differentiated versions of SEPRAN enable sensitivity analysis for a wide range of applications, not only from computational fluid dynamics.

AD Tools
ADIFOR

Related Applications
- Differentiation of the SEPRAN Package

BibTeX
@INPROCEEDINGS{
         Bischof2001OtU,
       author = "C. H. Bischof and H. M. B{\"u}cker and B. Lang and A. Rasch and J. W.
         Risch",
       title = "On the Use of a Differentiated Finite Element Package for Sensitivity Analysis",
       booktitle = "Computational Science -- ICCS~2001, Proceedings of the International Conference on
         Computational Science, San Francisco, USA, May~28--30, 2001. Part~I",
       editor = "V. N. Alexandrov and J. J. Dongarra and B. A. Juliano and R. S. Renner and C. J. K.
         Tan",
       volume = "2073",
       series = "Lecture Notes in Computer Science",
       pages = "795--801",
       address = "Berlin",
       publisher = "Springer",
       abstract = "Derivatives are ubiquitous in various areas of computational science including
         sensitivity analysis and parameter optimization of computer models. Among the various methods for
         obtaining derivatives, automatic differentiation~(AD) combines freedom from approximation errors,
         high performance, and the ability to handle arbitrarily complex codes arising from large-scale
         scientific investigations. In this note, we show how AD technology can aid in the sensitivity
         analysis of a computer model by considering a classic fluid flow experiment as an example. To this
         end, the software tool ADIFOR implementing the AD technology for functions written in Fortran~77 was
         applied to the large finite element package SEPRAN.\@ Differentiated versions of SEPRAN enable
         sensitivity analysis for a wide range of applications, not only from computational fluid dynamics.",
       ad_area = "Computational Fluid Dynamics",
       ad_tools = "ADIFOR",
       year = "2001"
}


back
  

Contact:
autodiff.org
Username:
Password:
(lost password)