2012
DOI: 10.1080/10556788.2011.577775
|View full text |Cite
|
Sign up to set email alerts
|

On higher-order differentiation in nonlinear mechanics

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
21
0

Year Published

2012
2012
2021
2021

Publication Types

Select...
7
1

Relationship

0
8

Authors

Journals

citations
Cited by 18 publications
(21 citation statements)
references
References 47 publications
0
21
0
Order By: Relevance
“…The solutions for the under-determined system (4) define one-dimensional continua, hereafter called branches of solutions. Continuation methods and related software, see [26,24] enable to compute solution branches. Most of them rely on first order predictor-corrector principles described in [27,28,29,30] and approximate the solution branches as a collection of points satisfying (4).…”
Section: Mechanism Analysis Using Higher-order Continuation Methodsmentioning
confidence: 99%
See 1 more Smart Citation
“…The solutions for the under-determined system (4) define one-dimensional continua, hereafter called branches of solutions. Continuation methods and related software, see [26,24] enable to compute solution branches. Most of them rely on first order predictor-corrector principles described in [27,28,29,30] and approximate the solution branches as a collection of points satisfying (4).…”
Section: Mechanism Analysis Using Higher-order Continuation Methodsmentioning
confidence: 99%
“…A second main contribution is the ability that offers the method to determine workspace and kinematic properties of mechanisms that can be employed using only basic knowledge of a mechanism. It builds on the higher-order continuation method [23,24] for the continuous computation of joint angular positions along any parameterized end-effector path. The results are presented in the form of so-called bifurcation diagrams displaying joint positions as functions of the end-effector position.…”
mentioning
confidence: 99%
“…In a nutshell, AD is a set of techniques designed to augment the source code of a software with derivative computations (Griewank & Walther, 2008;Naumann, 2011). It implements the chain rule in either a source to 260 source AD tool like Tapenade (Hascoet & Pascual, 2013) or an AD overloading library (Griewank et al, 1996;Charpentier, 2012) to be linked with the computer code. The AD tool (Hascoet & Pascual, 2013) is here applied to our MHFEM code.…”
Section: Inverse Modelingmentioning
confidence: 99%
“…Nevertheless after taking into account these uncertainties, we see that the trace phase (not the initiation or termination of the phase) has an important impact on the runtime of each of the operations and that it clearly forms a bottleneck of the computation 8 . We attribute this to the fact that tracing tracks the sequence of operations dynamically, resulting in a lot of dynamic allocations, indirections, cache misses, pipeline stalls and far jumps through the binary executable.…”
Section: A Comparison To Adol-cmentioning
confidence: 99%
“…A list of packages that allow for the differentiation of C++ codes is provided on the AutoDiff site. Beyond that, the differentiation of linear solvers or linear transformations, fixed-point methods [7], nonlinear solvers [8,9] and special functions of mathematical physics [10] requires a careful study to be accurate and efficient. Although general developments were proposed and automated for most of these issues, they are not systematically included in the existing AD tools.…”
Section: Introduction and Overviewmentioning
confidence: 99%