Tridiagonal Operator::SOR

[John Maiden]

I tried using the SOR operator that comes with the
tridiagonal operator class, and get the result "NaN"
if I use a positive value for the tolerance. Any ideas
on what is going wrong? If I use a negative value,
then the SOR operator doesn't do much (since it's easy
to get out of the for loop). I know that the tridiagonal
operator is working well, since I use it in a
StandardFiniteDifferenceModel, and the result produced
after a rollback makes sense.


-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

[Luigi Ballabio]

On Thu, 2007-06-28 at 17:56 +0000, John Maiden wrote:
> I tried using the SOR operator that comes with the
> tridiagonal operator class, and get the result "NaN"
> if I use a positive value for the tolerance. Any ideas
> on what is going wrong?

Hmm. I'm not sure that SOR ever worked reliably. Nando, are you reading?
Any comments?

Luigi


--

Cogito ergo I'm right and you're wrong.
-- Blair Houghton



-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

[Ferdinando M. Ametrano-3]

Hi all

On 6/29/07, Luigi Ballabio <[hidden email]> wrote:
> Hmm. I'm not sure that SOR ever worked reliably. Nando, are you reading?
> Any comments?

it was few years ago. I remember it seemed to work OK, but I had
problems plugging in into the framework at that time.
I would not rely on it since it has not been really used... anyway it
could be a decent starting point if anyone want to fix/integrate it.

ciao -- Nando

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

[John Maiden]

I might work on it, but I'd need an idea as to why
it's producing "NaN" as an output for positive
tolerance. Any possible explanations?


Regards,

                 John Maiden


       
____________________________________________________________________________________
Get the Yahoo! toolbar and be alerted to new email wherever you're surfing.
http://new.toolbar.yahoo.com/toolbar/features/mail/index.php

-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users

[John Maiden]

I've figured out where the "NaN" is coming from. When having the tridiagonal
operator components act on the input array, the difference between each SOR
iteration gets progressively larger (and sometimes the numbers go negative). The
error amount becomes too large for the computer to handle, and thus spits out a
"NaN". Still trying to figure out how to make it work.


-------------------------------------------------------------------------
This SF.net email is sponsored by DB2 Express
Download DB2 Express C - the FREE version of DB2 express and take
control of your XML. No limits. Just data. Click to get it now.
http://sourceforge.net/powerbar/db2/
_______________________________________________
QuantLib-users mailing list
[hidden email]
https://lists.sourceforge.net/lists/listinfo/quantlib-users