[SciPy-user] Integrating

Rob Clewley rob.clewley@gmail....
Mon Jan 14 10:03:32 CST 2008


Hi Tommy,

I am not sure how strict your requirement is for 5th order exactly,
but our PyDSTool package (pydstool.sourceforge.net) has interfaces
both to the 5th order Radau integrator (a good integrator for stiff
systems and those with algebraic constraints) and the Dormand-Prince
8th order multi-stage (8-5-3) integrator. Both codes were developed in
the 80s by Hairer & Wanner, renowned experts who have written a couple
of Springer books about these. Radau is in Fortran and Dopri is in C.

Our package lets you write your vector field definitions in
pseudo-code from inside a python script, which gets automatically
converted to C code for linking to the integrator object files. This
results in very fast integration! Many python interfaces to C
integrators only support python vector field definitions, creating an
additional overhead in the frequent callback to those functions. BTW
we also support C-level accurate detection of discrete events like
zero crossings.

Feel free to contact me if you have further questions.

-Rob

On 14/01/2008, Tommy Grav <tgrav@mac.com> wrote:
>
> On Jan 14, 2008, at 10:22 AM, Marian Jakubik wrote:
>
> > Hi Tommy,
> >
> > I am an astronomer (researcher) and for the job you mentioned I use
> > the
> > SWIFT package (written by Levison et. al). I think you know this
> > package so my question is - "are you working on some Python-code for
> > numerical integration of the small-bodies orbits?" I am very
> > interested
> > in it...
>
> I have used the SWIFT package yes. For the problem I have before me it
> is a little overkill :). I need to run a lot of small integrations
> with a lot of
> input/output of orbits and plotting. Since code development is much
> faster
> in python I am mostly my coding in that these days, as it is
> significantly
> easier to use in handling input/output of text files. Since the
> integrations
> are a a little slow in pure python to be efficient I was looking for an
> interface to some C/C++ integrator to speed that part up :)
>
> I am working on some astronomy python code. Mostly orbital mechanics
> with orbital determination and integration. The code is in a very early
> stage of development and has a tendency of undergoing major changes
> in the underlaying class structure as I add new functionality.
>
> > Please contact me (on my email) and we could debate about this
> > topic from the "astro-view", if you like...
>
> I have sent my reply to the list as well, in the hope that some other
> orbital mechanics will be interested and chime in.
>
> Cheers
>    Tommy
>
> +-----------------------------------------------------------------------
> ------------------------------------------+
>    Associate Research Scientist                 Dept. of Physics and Astronomy
>    Johns Hopkins University                     Bloomberg 243
>    tgrav@pha.jhu.edu                                    3400 N. Charles St.
>    (410) 516-7683                                       Baltimore, MD21218
> +-----------------------------------------------------------------------
> ------------------------------------------+
>
>
>
>
> _______________________________________________
> SciPy-user mailing list
> SciPy-user@scipy.org
> http://projects.scipy.org/mailman/listinfo/scipy-user
>


-- 
Robert H. Clewley, Ph. D.
Assistant Professor
Department of Mathematics and Statistics
Georgia State University
720 COE, 30 Pryor St
Atlanta, GA 30303, USA

tel: 404-413-6420   fax: 404-651-2246
http://www.mathstat.gsu.edu/~matrhc
http://brainsbehavior.gsu.edu/


More information about the SciPy-user mailing list