This is the mail archive of the
gsl-discuss@sourceware.cygnus.com
mailing list for the GSL project.
Re: multidimensional optimization
Brian Gough wrote:
>
> It might be important for the user to follow the progress of the 1d
> minimisations, so perhaps the algorithm should be split twice. A
> separate inner loop for the 1-d minimisation could be used like this,
>
> do {
> get gradient
> do {
> minimise in 1d
> } while (!converged_1d)
> ...
> } while (!converged)
>
> I'm don't think that the user should call the existing gsl_min
> functions --- we would provide a suitable minimisation function (maybe
> a wrapper around an existing gsl_min function) that will take care of
> the vector arithmetic so the 1d minimisation is as simple as the outer
> loop.
Right. I've already written a wrapper that allows to consider a R^n function,
a starting point and a direction as a R function. I will provide a way to
control the 1d minimisation and a way to automatize everything (with a
suitable way to specify parameters for the internal 1d minimisation).
Fabrice Rossi