[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
[Octave-bug-tracker] [bug #40057] exit criterion in pcg
From: |
Marco Caliari |
Subject: |
[Octave-bug-tracker] [bug #40057] exit criterion in pcg |
Date: |
Wed, 18 Sep 2013 09:40:47 +0000 |
User-agent: |
Mozilla/5.0 (X11; Linux i686; rv:13.0) Gecko/20100101 Firefox/13.0 |
URL:
<http://savannah.gnu.org/bugs/?40057>
Summary: exit criterion in pcg
Project: GNU Octave
Submitted by: caliari
Submitted on: Wed 18 Sep 2013 09:40:46 AM GMT
Category: Octave Function
Severity: 2 - Minor
Priority: 5 - Normal
Item Group: Matlab Compatibility
Status: None
Assigned to: None
Originator Name:
Originator Email:
Open/Closed: Open
Discussion Lock: Any
Release: 3.6.2
Operating System: GNU/Linux
_______________________________________________________
Details:
Dear all,
the exit criterion in pcg is
norm (B - A * X) <= TOL * norm (B - A * X0)
On the other hand, in C.T. Kelley (quoted as reference book), p.23, it is
used
norm (B - A * X) <= TOL * norm (B)
and the same in matlab (as I understand from the description of
[X,FLAG,RELRES] output). Of course, Octave criterion is a valid one and they
coincide for X0 = zeros (the default value). I think it should be changed for
compatibility. If one really needs the current behavior, he can give TOL *
norm (B - A * X0) / norm (B) as input tolerance.
Cheers,
Marco
_______________________________________________________
Reply to this item at:
<http://savannah.gnu.org/bugs/?40057>
_______________________________________________
Message sent via/by Savannah
http://savannah.gnu.org/
[Prev in Thread] |
Current Thread |
[Next in Thread] |
- [Octave-bug-tracker] [bug #40057] exit criterion in pcg,
Marco Caliari <=