nonsymm eigenvalues...

nonsymm eigenvalues...

[Robert G. Brown]

Hi, I'm just starting to write a little code snippet to find eigenvalues
and vectors for a non-symmetric, real matrix and am not finding the
nonsymm functions or include files as described in the current online
manual.  My current version is 1.8-1.1 (FC 6).  Were these functions not
available in that release?  Would upgrading help or is the problem
elsewhere (as in I'm not looking at the right place in e.g.
gsl/gsl_eigen.h)?

   rgb

-- 
Robert G. Brown
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone(cell): 1-919-280-8443
Web: http://www.phy.duke.edu/~rgb
Lulu Bookstore: http://stores.lulu.com/store.php?fAcctID=877977

Re: nonsymm eigenvalues...

[Robert G. Brown]

On Wed, 14 Nov 2007, Robert G. Brown wrote:

> Hi, I'm just starting to write a little code snippet to find eigenvalues
> and vectors for a non-symmetric, real matrix and am not finding the
> nonsymm functions or include files as described in the current online
> manual.  My current version is 1.8-1.1 (FC 6).  Were these functions not
> available in that release?  Would upgrading help or is the problem
> elsewhere (as in I'm not looking at the right place in e.g.
> gsl/gsl_eigen.h)?

Also, what is likely to happen if I call the symmetric routines for a
nonsymmetric matrix?  They seem to "work" and still produce eigenvalues
and eigenvectors.  Are those garbage?

Sorry to ask what are probably elementary questions, but the last time I
actually took a course on linear algebra or associated computational
methods was something like 1976, and I don't use it much in my research.
Or haven't until now...;-)

   rgb

>
>  rgb
>
>

-- 
Robert G. Brown
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone(cell): 1-919-280-8443
Web: http://www.phy.duke.edu/~rgb
Lulu Bookstore: http://stores.lulu.com/store.php?fAcctID=877977

Re: nonsymm eigenvalues...

[Brian Gough]

At Wed, 14 Nov 2007 06:58:48 -0500 (EST),
Robert G. Brown wrote:
> Also, what is likely to happen if I call the symmetric routines for a
> nonsymmetric matrix?  They seem to "work" and still produce eigenvalues
> and eigenvectors.  Are those garbage?

It just uses the lower triangular part and assumes symmetry, anything
in the upper triangular part is ignored.

-- 
Brian Gough

Re: nonsymm eigenvalues...

[Robert G. Brown]

On Wed, 14 Nov 2007, Brian Gough wrote:

> At Wed, 14 Nov 2007 06:58:48 -0500 (EST),
> Robert G. Brown wrote:
>> Also, what is likely to happen if I call the symmetric routines for a
>> nonsymmetric matrix?  They seem to "work" and still produce eigenvalues
>> and eigenvectors.  Are those garbage?
>
> It just uses the lower triangular part and assumes symmetry, anything
> in the upper triangular part is ignored.

Garbage indeed.  Thanks.

   rgb

-- 
Robert G. Brown
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone(cell): 1-919-280-8443
Web: http://www.phy.duke.edu/~rgb
Lulu Bookstore: http://stores.lulu.com/store.php?fAcctID=877977

Re: nonsymm eigenvalues...

[Brian Gough]

At Wed, 14 Nov 2007 06:50:52 -0500 (EST),
Robert G. Brown wrote:
> 
> Hi, I'm just starting to write a little code snippet to find eigenvalues
> and vectors for a non-symmetric, real matrix and am not finding the
> nonsymm functions or include files as described in the current online
> manual.  My current version is 1.8-1.1 (FC 6).  Were these functions not
> available in that release?  Would upgrading help or is the problem
> elsewhere (as in I'm not looking at the right place in e.g.
> gsl/gsl_eigen.h)?
> 

You need 1.9 or later.  The NEWS file has a complete list of when
features were added.

-- 
Brian Gough

Re: nonsymm eigenvalues...

[Robert G. Brown]

On Wed, 14 Nov 2007, Brian Gough wrote:

> At Wed, 14 Nov 2007 06:50:52 -0500 (EST),
> Robert G. Brown wrote:
>>
>> Hi, I'm just starting to write a little code snippet to find eigenvalues
>> and vectors for a non-symmetric, real matrix and am not finding the
>> nonsymm functions or include files as described in the current online
>> manual.  My current version is 1.8-1.1 (FC 6).  Were these functions not
>> available in that release?  Would upgrading help or is the problem
>> elsewhere (as in I'm not looking at the right place in e.g.
>> gsl/gsl_eigen.h)?
>>
>
> You need 1.9 or later.  The NEWS file has a complete list of when
> features were added.

NEWS in the sources?  Is there a link in the current manual or on the
gnu page?  I looked around for one and didn't find it, and of course I
was asking to find out if I needed to upgrade because I hadn't yet...;-)

I'll do so now, thanks.

   rgb

-- 
Robert G. Brown
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone(cell): 1-919-280-8443
Web: http://www.phy.duke.edu/~rgb
Lulu Bookstore: http://stores.lulu.com/store.php?fAcctID=877977

Re: nonsymm eigenvalues...

[Brian Gough]

At Wed, 14 Nov 2007 11:18:50 -0500 (EST),
Robert G. Brown wrote:
> NEWS in the sources?  Is there a link in the current manual or on the
> gnu page?  I looked around for one and didn't find it, and of course I
> was asking to find out if I needed to upgrade because I hadn't yet...;-)

I've added a link to the NEWS file on the web page, there wasn't one before.

Question about Lagrange multiplier

[Gyenese János]

Dear All,

I would like to build a library of the constrained maximization based on
Lagrange multiplier and using this multiplier a minimize library. (I have to
use Lagrange because the second library use the result of the first, the
first is an entropy maximization and the second is the cross entropy
minimize.)

I think about using GSL.

Are there anybody have experience to build the above mentioned libraries?

Any help appreciate.


Regards,

Janos

Re: Question about Lagrange multiplier

[Brian Gough]

At Sat, 17 Nov 2007 16:50:52 +0100,
Gyenese János wrote:
> I would like to build a library of the constrained maximization based on
> Lagrange multiplier and using this multiplier a minimize library. (I have to
> use Lagrange because the second library use the result of the first, the
> first is an entropy maximization and the second is the cross entropy
> minimize.)
> 
> I think about using GSL.
> 
> Are there anybody have experience to build the above mentioned libraries?

Hello,
Have you checked OOL? http://ool.sourceforge.net/
There is a list of other packages on http://www.gnu.org/software/gsl/

-- 
Brian Gough