Re: models in the wild (was Re: [Swarm-Modelling] Re: [Swarm-Support] Repast vs. Swarm)

[Rick Riolo]

re glenn's and others comments about ABM as another "tool"--
right, different modeling approaches are appropriate or not
depending on the modeling goals, system being modeled, etc.

re glen's comment:
> The direction to move into is to create computational devices that you
> experiment with/on.  You have to treat these _devices_ like they are
> material objects to be used in an experiment.  And that means abandoning
> the idea of mimicking things.

I may not be understanding what you are saying, but if i am, i
would put it a bit differently:

*One* direction to go is the direction you talk about, i.e.,
treating the ABS as a system:
> I believe that agent-based _systems_ (not "models") are an ideal way  
> to approach this type of "models gone wild" (sorry, I couldn't resist
> ;-) development of _method_.  It's much more natural and intuitive to
> think of ABSes as "analogs", with an existence apart from any
> referent, than it is to think of systems of equations that way.

But I think *another* reasonable direction to go is to treat ABM's
as *models*, just as diffyQs or other formalisms (or informalisms)
are models, i.e., a simplified version of some "real" system,
which (more or less) generates some of the same dynamics,patterns,etc,
of interest in the real system, which needs "verifcation" and 
"validation" as appropriate given the modeling goals, and given
the possibilities (given the complexity of the system being modeled...
see Steve Bankes on "Deep Uncertainty").

A thought provoking article in the ALife literature that
covers many of these issues is:

  Di Paolo, Nobe and Bullock.  Simulation Models as Opaque Thought 
  Experiments (Postscript)  Artificial Life VII: The Seventh International 
  Conference on the Simulation and Synthesis of Living Systems, Reed 
  College, Portland, Oregon, USA, 1-6 August, 2000.

But as i say, maybe i'm missing glen's point.

- r

-- 
Rick Riolo                           address@hidden
Center for the Study of Complex Systems (CSCS)
4477 Randall Lab                
University of Michigan         Ann Arbor MI 48109-1120
Phone: 734 763 3323                  Fax: 734 763 9267
http://cscs.umich.edu/~rlr

On Thu, 29 Jan 2004 address@hidden wrote:

> Date: Thu, 29 Jan 2004 15:23:06 -0800
> From: address@hidden
> Reply-To: address@hidden
> To: address@hidden
> Subject: models in the wild (was Re: [Swarm-Modelling] Re:
>     [Swarm-Support] Repast vs. Swarm)
> 
> Steve Railsback writes:
>  > One key to making ABS more acceptable as science is developing common, 
>  > widely used modeling concepts, terms, frameworks, and tools. To me one 
>  > of the coolest things about Swarm is that the project started off with 
>  > the goal of meeting these needs; and has succeeded to some extent.
> 
> Exactly.  Modeling and simulation are as useful as any other tool.  As
> Marcus pointed out, advocating the use of ABM (especially as "the
> only" or "the best" way to do something) is about as silly as
> advocating the use of the beaker in chemistry. [grin] It's a tool that
> you either use or don't.  If you use it, have a good reason.  If you
> don't use it, have a good reason.
> 
> The real issues lie in the _methods_.  How do you use simulation?  How
> do you use your models?  Are your models well specified?  (In the same
> way your experimental devices either have spec sheets that are
> believable or they don't.)
> 
> But, as I've said before, the best way to move into discussions about
> methods and materials is to get away from the concept of "simulation".
> Simulation is all about being able to mimic some phenomenon without
> acurately implementing the machine that generates that phenomenon.
> That requirement is akin to the Turing test for intelligent machines
> and presents us with all sorts of time-wasting arguments.
> 
> The direction to move into is to create computational devices that you
> experiment with/on.  You have to treat these _devices_ like they are
> material objects to be used in an experiment.  And that means abandoning
> the idea of mimicking things.
> 
> (Note: There is a broader sense of the word "simulation" that still
> applies here.  It is that of man- and hardware-in-the-loop simulation
> or those activities that might be more appropriately termed
> "exercises" or role-playing.  Those "simulations" push the boundaries
> towards that of building analogs -- chemical, physical, electrical,
> etc. -- such that experiments can be performed on multiple analogs and
> the outcomes compared.  I'm suggesting we move further toward this
> sense of the word "simulation," which is less focused on mimicry and
> more focused on the methods involved in carrying out experiments over
> those systems.)
> 
> I believe that agent-based _systems_ (not "models") are an ideal way
> to approach this type of "models gone wild" (sorry, I couldn't resist
> ;-) development of _method_.  It's much more natural and intuitive to
> think of ABSes as "analogs", with an existence apart from any
> referent, than it is to think of systems of equations that way.
> 
>