Re: Is anyone modeling human-human interaction with Swarm?

[gepr]

Greg Madey wrote:

> Check out Business Week, September 21, 1998, p. 80.

.FLIGHT SIMULATORS FOR MANAGEMENT;
                   Author: By John A. Byrne in New York;
                   Thor Sigvaldason is late. As part of a novel consulting 
cluster at
                   PricewaterhouseCoopers, he is supposed to be at an important
                   client meeting at 10 a.m. But he has overslept for the 
session
                   with...; Size: 14K;
                   Management; 09-21-1998;
                   Computer models may give execs previews of how decisions pan 
out

It's about Win Farrell's group (Emergent Solutions Group [ESG]).

There's a useful distinction to be made between this kind of modeling and
the kind I mentioned is occurring at PSL.  And that distinction lies in
the foci of the models.  The PSL focusses on "observability" and, given
a set of phenomena, what human cognitive constituent functions generate
that phenomena.  Hence, the modeling they do is inherently psychological
and reflexive.

The kind of simulation ESG does is focussed on (and correct me if I'm
wrong you ESG guys) statistical modeling of human behavior.  Hence it
takes as its primary data, demographics and trends.

Now, both of these types of modeling are "CAS modeling" in a nicely loose
sense [grin].  But, the PSL effort is more like trying to find the minimal
model for the micro that generates (with a kind of morphogenesis) the
relevant macro properties.  And the ESG effort takes a more coevolutionary
stance, considering the content of the agents (the psychology, if you will)
less important than the interactions between agents.

So, in Swarm terms, PSL is focussed on finding the right behaviors for
the agents and ESG is focussed on finding the right medium to facilitate
the global dynamic.

Ultimately, of course, both will do a little of the other type of modeling; but,
it's interesting to watch the two unfold and be applied in two seemingly
different domains.

Another interesting difference is what I would characterize as the
difference between something like formal methods (PSL) and estimation
(ESG).  In simulation, you have the option of which aspects of a system
you want to model explicitly and those you want to bundle up into
macro characterizations.  Since PSL is trying to find mathematical
ways to specify and explicate causal relationships, they are not allowed
to bundle up quite as many variables into a noise function as ESG might
be able to do.  On the other hand, ESG has a significantly more difficult
job of bounding and characterizing their noise functions.  Separation
and classification of families of variables to bundle is a pretty hard
job.

glen
p.s. I moved this to the swarm-modelling list where it belongs. [grin]
--
glen e. p. ropella      =><= Hail Eris!
the swarm corporation   W:(505) 995-0818
<address@hidden>        H:(505) 424-0448




                  ==================================
   Swarm-Modelling is for discussion of Simulation and Modelling techniques
   esp. using Swarm.  For list administration needs (esp. [un]subscribing),
   please send a message to <address@hidden> with "help" in the
   body of the message.
                  ==================================