Re: [Swarm-Support] development priorities (was Re: Membership in Swarm Developmen Group)

[glen e. p. ropella]

Scott Christley wrote:
> Another viewpoint is that we know the behavior of the collective for
> this is what we can observe, so we want to know what individual
> behaviors can give rise to that collective action.  This isn't so
> obviously characterized as a simulation task, more one of model
> selection, but if viewed from the meta-level of a set of parameterized
> models (versus a single model with a set of parameters) then simulation
> becomes an exploration of meta-models of individual behaviors.  This is
> even more intractable which just gives even more reason to run on big
> computers!  :-)

Exactly!  In fact, one of the early ideas that we never pursued was to
integrate generic search tools for characterizing and searching behavior
spaces.  JJ Merelo developed a GA tool for that specific type of search
and Rob Keweley implemented something a little more generic; but, those
were two particulars in a class of search related methods.

This is exactly the type of original requirement that we should consider
for a next generation ABM tool.

-- 
glen e. p. ropella, 971-219-3846, http://tempusdictum.com
I have the heart of a child. I keep it in a jar on my shelf. -- Robert Bloch