Hi,
I have a question about muscle recruitment options.
AnyMuscleRecruitmentSolver has 4 options for using either
Simplex or Quadratic promming. It refers to an “OOSOL library”, have you
explained this a little more in detail in any of your tutorial sections.
If not, then could you give me a reference which would help me understand
these option better?
thanks
prabhav.
[Non-text portions of this message have been removed]
Hi Prabhav
There is a tutorial about these options, please see
http://www.anybodytech.com/691.0.html
In the reference manual on page 269 (AnyBodyStudy) i have found the text
below concerning the InverseDynamicAnalysis
Inverse Dynamics: The inverse dynamics problem is solved by an optimization
approach that minimizes
some muscle performance criterion with respect to the unknown forces of the
system.
The equations of motion, i.e., the dynamic equilibrium equations, enter the
problem as constraints, as well as the condition that the muscles cannot
push.
In AnyBodyTM the basic performance criterion is the maximal muscle activity,
where the muscle activity is defined as the muscle force divided by the
current muscle strength. The problem is solved iteratively eliminating
uniguely determined muscles and redundant equations after each step of the
iteration. The optimization in each step can be handled with different
solvers according to the RecruitmentSolver attribute. Basically these
solvers can be divided into linear programming (Simplex) and quadratic
programming (QP) algorithms. The numerical library OOSol, which is used by
AnyBodyTM contains both a Simplex and a QP algorithm. Furthermore, a simplex
algorithm based on algorithms in the book Numerical Recipes in C++ by Press
et. al. The attribute RecruitmentSolver allows you to select either of these
three algoritms with value names composed of the library name (OOSol or NR)
and the algorithm type (simplex or QP). The MinMaxSimplex is the default
simplex solver, which currently is MinMaxOOSolSimplex. Both types of
algoritms feature a penalty setting that can assist in making the problem
better conditioned. In addition these penalties can be used to change the
appearance of the problem: A large penalty in the linear formulation of the
problem makes the problem converge towards a linear problem, i.e.,
minimization of the sum of muscle activities. A large penalty in the
quadratic formulation makes it a minimization of the sum of squares of
muscle activities. Solution in terms of muscle forces can be obtained using
a model where all muscle models have constant strength equal to one.
Please ask again if you have further questions.
Best regards
Søren
From: anyscript@yahoogroups.com [mailto:anyscript@yahoogroups.com] On Behalf
Of Prabhav Saraswat
Sent: 14 November 2007 22:14
To: anyscript@yahoogroups.com
Subject: [AnyScript] OOSOL library??
Hi,
I have a question about muscle recruitment options.
AnyMuscleRecruitmentSolver has 4 options for using either
Simplex or Quadratic promming. It refers to an “OOSOL library”, have you
explained this a little more in detail in any of your tutorial sections.
If not, then could you give me a reference which would help me understand
these option better?
thanks
prabhav.
[Non-text portions of this message have been removed]
[Non-text portions of this message have been removed]