[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [Help-gsl] Comments/questions about BFGS algorithm implemented for t
Re: [Help-gsl] Comments/questions about BFGS algorithm implemented for the GSL
Tue, 18 Apr 2006 17:32:37 -0400
> I suspect there is a substantial inefficiency problem with the GSL line-
> search implementation (since the BFGS update part is completely
> straightforward). I am not any sort of expert in this field though so I
> cannot pin down where the problem is occurring, but a superficial google
> search reveals a number of complaints about the GSL minimization efficiency.
I don't fully understand the line search code that's called by the
multidimensional minimizers, but it looks like it's not using the
information from the derivative. I hope I'm wrong about this, though.
Meanwhile, you could look into other free optimizers, including the
optimization code in the R project (http://www.r-project.org/, which
supposedly uses L-BFGS-B as well) or the TAO toolkit