R help - Apr 2011 - Krylov subspace computations of matrix exponentials

I use the very nice expm functions available from the expm and Matrix
packages. My understanding is that for large sparse matrices the
currently best methods available are Krylov subspace methods, but
they are as far as I can tell not implemented in either of the packages
mentioned, nor in any other R package I have found.

Does anybody know if Krylov subspace methods are available from
any R package?

If not, is there anybody working on this or planning to do so?

If not, I might be interested in making such an implementation, because
I find myself in the need of fast methods for very large sparse matrices.

A third question. There exists a Fortran implementation called Expokit
with an ad hoc license stating that

"Permission to use, copy, modify, and distribute EXPOKIT and its
supporting documentation for non-commercial purposes, is hereby
granted without fee, provided that this permission message and
copyright notice appear in all copies."

If I choose to use Expokit and just write an interface to R, would
there be any problems with having such a copyright notice in the package?

Thanks, Niels

-- 
Niels Richard Hansen                     Web:   www.math.ku.dk/~richard	
Associate Professor                      Email: Niels.R.Hansen at math.ku.dk
Department of Mathematical Sciences             nielsrichardhansen at gmail.com
University of Copenhagen                 Skype: nielsrichardhansen.dk	
Universitetsparken 5                     Phone: +1 510 502 8161	
2100 Copenhagen ?
Denmark

Look at ExpoKit (http://www.maths.uq.edu.au/expokit/)
It implement Krylov methods and is really fast.
Unfortunately no R wrapper that I know yet.

If you plan to implement it, I can provide my testing code.
Ciao!
                     mario

On 14-Apr-11 10:16, Niels Richard Hansen wrote:
> I use the very nice expm functions available from the expm and Matrix
> packages. My understanding is that for large sparse matrices the
> currently best methods available are Krylov subspace methods, but
> they are as far as I can tell not implemented in either of the packages
> mentioned, nor in any other R package I have found.
>
> Does anybody know if Krylov subspace methods are available from
> any R package?
>
> If not, is there anybody working on this or planning to do so?
>
> If not, I might be interested in making such an implementation, because
> I find myself in the need of fast methods for very large sparse matrices.
>
> A third question. There exists a Fortran implementation called Expokit
> with an ad hoc license stating that
>
> "Permission to use, copy, modify, and distribute EXPOKIT and its
> supporting documentation for non-commercial purposes, is hereby
> granted without fee, provided that this permission message and
> copyright notice appear in all copies."
>
> If I choose to use Expokit and just write an interface to R, would
> there be any problems with having such a copyright notice in the package?
>
> Thanks, Niels
>
-- 
Ing. Mario Valle
Data Analysis and Visualization Group            | http://www.cscs.ch/~mvalle
Swiss National Supercomputing Centre (CSCS)      | Tel:  +41 (91) 610.82.60
v. Cantonale Galleria 2, 6928 Manno, Switzerland | Fax:  +41 (91) 610.82.82

On Apr 14, 2011, at 10:16 , Niels Richard Hansen wrote:
> 
> A third question. There exists a Fortran implementation called Expokit
> with an ad hoc license stating that
> 
> "Permission to use, copy, modify, and distribute EXPOKIT and its
> supporting documentation for non-commercial purposes, is hereby
> granted without fee, provided that this permission message and
> copyright notice appear in all copies."
> 
> If I choose to use Expokit and just write an interface to R, would
> there be any problems with having such a copyright notice in the package?
> 
Probably. GPL and "non-commercial" are incompatible, so you can't
distribute anything that mixes the two licenses. R itself has chosen to allow
non-GPL packages, although there is a very fine line between a package as a
"plug-in" vs. a "derived work".

-- 
Peter Dalgaard
Center for Statistics, Copenhagen Business School
Solbjerg Plads 3, 2000 Frederiksberg, Denmark
Phone: (+45)38153501
Email: pd.mes at cbs.dk  Priv: PDalgd at gmail.com