Publication: Adjoint gradients compared to gradients from algorithmic differentiation in instataneous control of the Navier-Stokes equations
Introduction
Applications
Tools
Research Groups
Workshops
Publications
   List Publications
   Advanced Search
   Info
   Add Publications
My Account
About

Adjoint gradients compared to gradients from algorithmic differentiation in instataneous control of the Navier-Stokes equations

- Article in a journal -
 

Area
Computational Fluid Dynamics

Author(s)
Michael Hinze , Thomas Slawig

Published in
Optimization Methods & Software

Institution
Institute of Mathematics, Technische Universität Berlin

Year
2003

Abstract
The authors first used TAMC and then TAF. The model uses an iterative solver, for which, after inserting 5 TAF flow directives, TAF can generate a very efficient adjoint. The TAF generated adjoint is slightly faster than its hand coded counterpart.

AD Tools
TAF, TAMC

AD Theory and Techniques
Iteration, Self Adjoint Solvers

BibTeX
@ARTICLE{
         Hinze2003Agc,
       author = "Michael Hinze and Thomas Slawig",
       title = "Adjoint gradients compared to gradients from algorithmic differentiation in
         instataneous control of the {N}avier-{S}tokes equations",
       journal = "Optimization Methods \& Software",
       institution = "Institute of Mathematics, Technische Universit{\"a}t Berlin",
       type = "Technical Report",
       ad_tools = "TAF, TAMC",
       ad_area = "Computational Fluid Dynamics",
       ad_theotech = "Iteration, Self Adjoint Solvers",
       url = "http://www.math.tu-berlin.de/preprints/abstracts/Report-735-2002.rdf.html",
       abstract = "{The authors first used TAMC and then TAF. The model uses an iterative solver, for
         which, after inserting 5 TAF flow directives, TAF can generate a very efficient adjoint. The TAF
         generated adjoint is slightly faster than its hand coded counterpart.}",
       YEAR = "2003",
       VOLUME = "18",
       NUMBER = "3",
       PAGES = "299--315"
}


back
  

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