|
|
Ublas : |
From: Fred (hakonbh_at_[hidden])
Date: 2006-10-24 21:32:23
Hi all,
In truth, I'm trying to solve linear and nonlinear sparse systems as fast as
Thanks again,
Fred
On 10/24/06, James N. Knight <nate_at_[hidden]> wrote:
possible because my program repeat this operation many times.
I got a good time using the algorithm of example 2 in:
http://www.bauv.unibw-muenchen.de/~winkler/ublas/examples/
but it isn't sufficient yet.
It's possible that many increments people write aren't add to the official
ublas version, so I asked for help. That's it.
>
> You can read about Nonlinear cg here: http://tinyurl.com/yd37bt
>
> This thread seems a bit off topic, but before it dies I wanted to say that
> I'm working on a generic nonlinear optimization library that can (and
> currently does)
> use ublas for the linear algebra representation and algorithms. The code
> is generic
> with respect to the function being optimized, the linear algebra
> facilities, and the
> space being optimized over. If the original poster or anyone else is
> interested please
> email me offlist.
>
> James
>
> Gunter Winkler wrote:
> > Hi Fred and Karl,
> >
> > Karl Meerbergen schrieb:
> >> I vaguely recall that conjugate gradients is an optimization method.
> But I do
> >> not recall the details for nonlinear problems. Convergence is only
> guaranteed
> >> for specific math properties, as this is also the case for linear
> systems.
> >>
> >
> > Yes. Linear CG is a method to find (the unique) vector x that minimizes
> > a (scalar) quadratic function
> >
> > f(x) := 1/2 <x, Ax> - <b, x> -> min
> >
> > ( <a,b> is any inner product, such that <x, Ax> > 0 for all x<>0 )
> >
> > Although I must admit that I never heard of "the nonlinear CG". There
> > are lots of gradient based methods for nonlinear minimization.
> >
> > Fred, can you explain your method?
> >
> >
> > mfg
> > Gunter
> >
> > _______________________________________________
> > ublas mailing list
> > ublas_at_[hidden]
> > http://lists.boost.org/mailman/listinfo.cgi/ublas
> _______________________________________________
> ublas mailing list
> ublas_at_[hidden]
> http://lists.boost.org/mailman/listinfo.cgi/ublas
>