REPLY Discrete event software (SD6708)

[SDMAIL Bill Harris]

Posted by  Bill Harris <bill_harris at facilitatedsystems.com>

"SDMAIL Bill Braun" <bbraun at hlthsys.com> writes:
> I am involved in a project where I may need to contrast SD and discrete 
> event simulation. I want to avoid significant investment until the 
> future path becomes clear.

Bill B.,

I think there's a thread about this from a few years back.  I recall it,
because I recall making the point that SD and discrete event form a
false dichotomy: SD is all about feedback and accumulation, which can be
done in continuous or discrete time and with quantized or non-quantized
levels.  (I really remember that because John Sterman followed up with
"Bill Harris is right!", which I /so/ wanted to post on my Web site  :-)  .)

> > I would appreciate suggestions/recommendations on open source and/or low 
> > cost packages. The software should be GUI based (like Vensim, Powersim, 
> > iThink, etc.).

I presume what you want is what normally is called discrete event
simulation: no inherent feedback (it may be harder to find a tool that
allows no feedback) and quantized time and perhaps levels, right?

CPNTools (http://wiki.daimi.au.dk/cpntools/cpntools.wiki) is a (the?)
GUIfied tool for Colored Petri Nets.  I'm not sure what the status of
its graphical output capabilities is; I haven't really used it in a
while.  You'll have to decide if a CPN fits your needs, and you do need
to program bits of a model (guards, arc expressions, declarations, etc.)
in a subset of ML.

SimPy (http://simpy.sourceforge.net/) seems to be a (the?) major
contender in that world today.  It's on my list to learn; I've just
played around with it a bit so far.  It has a GUI but perhaps not in the
sense you mean: you can program a GUI to provide a user interface for
others, but you still program the model in Python.

I sense that either would be robust enough to do production work, as
long as they have the capabilities you seek.  Both are no-cost solutions.

Having said all that, perhaps I should have started by asking what you
really need to do.  That might focus the answers better; mine were just
quick, off-the-cuff responses.

I hope that helps; I'll be curious what others say.

Bill
- -- 
Bill Harris 
Posted by  Bill Harris <bill_harris at facilitatedsystems.com>
posting date  Fri, 04 Jan 2008 20:21:38 -0800