[SciPy-User] Passing fmin_cg a gradient
Yury V. Zaytsev
yury at shurup.com
Sun Apr 17 10:11:31 EDT 2011
On Fri, 2011-04-15 at 17:26 -0400, Jason Rennie wrote:
>
> I'm actually on the look-out for a good open source CG or L-BFGS
> implementation. Thanks for the fmin_tnc mention---I'll check it out.
On Mon, 2011-04-11 at 20:52 -0400, Patrick Holvey wrote:
> I'm also open to alternate optimization calls if anyone has
> recommendations.
So far I have been very happy with NLopt which got to know thanks to the
advice someone posted on this list. It features a nice pure C
implementation of BFGS, which, I guess, for most problems would be a
good starting point and a number of other gradient-based methods.
Also, I have found through experimentation, that NLopt Downhill-Simplex
implementation is much more stable for the functions that I have to
optimize than fmin from SciPy.
Do check it out.
--
Sincerely yours,
Yury V. Zaytsev
More information about the SciPy-User
mailing list