Re: Combining ABM and CFD in Swarm?

[Tom Wainwright]

I also am working on a similar problem, but am not far enough along to
help--I'm looking for advice as well.  Perhaps we're ready for an aquatic
modeling interest group for Swarm.

I am developing a marine food web model for the California Current System,
which will eventually have food webs embedded in a 3D ocean.  (My "agents"
are abstract trophic pools rather than individual organisms, so it is not
quite analogous to the particles in water problems you are working on.) 
All the physics will be external to the Swarm model, and will be read in as
part of the environment each time step.  Like Steve, I am using a
space-time grid that is coarser than that needed for the physics, so there
is some averaging that needs to be done.  Swarm currently has no spatial
structures appropriate for this model, so I was assuming I'd have to write
those myself after I get a 2D-grid version working.

To be honest, I recently came back to Swarm reluctantly after abandoning it
two years ago.  I come from a strong scientific/statistical programming
background, and have a preference for strongly-typed, compiled languages,
and have recently been using "design-by-contract" (DBC) extensively.  I
couldn't bring myself to work in Objective-C, Java was far too slow, and
the method-passing structure of Swarm appeared antithetical to safe,
verifiable programming practices.  Recently, JIT compilers, gcj, and the
development of some DBC techniques for Java  made me look again, and
realize I was spending too much time writing libraries that exist already
in Swarm.  There remains the problem of verifying correctness in a
message-passing environment, but that can be addressed with the message
passing largely restricted to probes and the GUI.  My other main problem is
that as my model grows, I will need parallel code for a Beowulf-class
cluster, and I'm betting that Swarm/Java will come through with support for
that soon enough.  I am hoping that Swarm is up to the task of large
real-world problems.

M Lang / S Railsback wrote:
> 
> Uffe Høgsbro Thygesen wrote:
> >
... 
> > My question is: Do you know of anyone who has combined Swarm with a CFD
> > package? Or extended Swarm with specialised features for CFD?
> >
> > Any pointers are much appreciated. Cheers,
> >
> > Uffe
> 
> I admire your ambition very much!
> 
> We have built several fish models in Swarm ...

-- 
Tom Wainwright
NOAA/NMFS/NWFSC
2030 S Marine Science Dr
Newport, OR 97365 USA
address@hidden

                  ==================================
   Swarm-Support is for discussion of the technical details of the day
   to day usage of Swarm.  For list administration needs (esp.
   [un]subscribing), please send a message to <address@hidden>
   with "help" in the body of the message.