[SciPy-dev] Trouble with optimize.fmin_ncg

[dmitrey]

I'm not sure it's a good idea, because it requires having the file 
containing matrix + correct load by scipy.io
I think constructing matrices explicitly is better.
Also, I suspect the func <x,Ax>/<x,Bx> is non-convex even for B=I.
But if you want, I can add the test case (for fmin_ncg) after closing 
all other currently still open tickets, assigned to me.
Regards, D.

Alan G Isaac wrote:
> Many thanks to Nils for finding and localizing the bug,
> Dmitrey for helping to localize it,
> and Nathan for fixing it!
>
> It seems like a good idea to keep this around as
> a test case.  (A stripped down version.)
> Nils, any objections?
>
> Cheers,
> Alan Isaac
>
>
>
> _______________________________________________
> Scipy-dev mailing list
> Scipy-dev at scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-dev
>
>
>
>