This is the mail archive of the
gsl-discuss@sources.redhat.com
mailing list for the GSL project.
Re: Bug in Levenbeg-Marquardt Algorithm
- From: Hans Ekkehard Plesser <hans dot plesser at itf dot nlh dot no>
- To: gsl-discuss at sources dot redhat dot com
- Date: Mon, 25 Feb 2002 16:29:51 +0100 (CET)
- Subject: Re: Bug in Levenbeg-Marquardt Algorithm
- Reply-to: hans dot plesser at itf dot nlh dot no
Hi!
I now had a look at the original MINPACK code (lmder.f). They set
par_lower to 0 for rank deficient r, so here is a patch that does the
same in lmder.c .
Best regards,
Hans
---------------------------------------------------------------------
Dr. Hans Ekkehard Plesser Tel. : +47 6494 8832
Physics Section / ITF Fax : +47 6494 8810
Agricultural University of Norway e-mail: hans.plesser@itf.nlh.no
N-1432 Ås, Norway WWW : arken.nlh.no/~itfhep
---------------------------------------------------------------------
270c270,277
< par_lower = fp / (delta * phider);
---
>
> /* w == zero if r rank-deficient,
> then set lower bound to zero form MINPACK, lmder.f
> Hans E. Plesser 2002-02-25 (hans.plesser@itf.nlh.no) */
> if ( wnorm > 0 )
> par_lower = fp / (delta * phider);
> else
> par_lower = 0.0;