Mathematics VS? Agent Based Modeling

[Stephen Guerin]

Let me restate my previous comment and repost the quote. I suspect the ABM
vs. EBM debate is ill-posed. I favor Bonabeau's characterization of ABM more
than Parunak's...

"ABM is a mindset more than a technology. The ABM mindset consists of
describing a system from the perspective of its constituent units. A number
of researchers think that the alternative to ABM is traditional differential
equation modeling; this is wrong, as a set of differential equations, each
describing the dynamics of one of the system's constituent units, is an
agent-based model. A synonym of ABM would be microscopic modeling, and an
alternative would be macroscopic modeling."

Robert writes:
> So my contention is that while ABMs may be great for advisory, open-loop
> systems they just aren't robust enough to form the core of a closed-loop
> system. How would you feel if the automatic control systems in your car
> (plane, nuclear reactor etc.) where all written in NetLogo? ;)

More comfortable than if it was written as a VBA add-in to Excel ;-)

I know you're kidding, but I'm not sure what software platforms have to do
with determining the appropriate modeling perspective for understanding the
problem at hand.

Also, there is a distinction to be made between Agent-Based Modeling and
Agent-based systems. The routers on the internet coordinating to deliver
this email is an example of a distributed closed-loop agent control system.
(granted, they're doing it very poorly today, Roger. We need a new hosting
company...). The Internet is an agent-based system, not an ABM. If I was
proposing a new protocol I would want to simulate it first with an
abstraction of reality. That would be an ABM.

-S

[Robert Holmes]

I agree Steve, I think the Bonabeau statement is more accurate. But it does
seem a rather  broad definition: it could be applied to any part of
scientific reductionism from Democritus onwards. Take for instance
thermodynamics vs. statistical mechanics. Both describing the same phenomena
just from opposite directions (macros vs micro). So by this defintion ABMs
have been a subject of real hard-core research since the nineteenth century.

Does this matter? I know it shouldn't but it still manages to bug me. ABMs
are sold as a New Thing, an exciting new forefront of scientific discovery
that has emerged within the last couple of decades. Yet by one of the more
coherent definition of ABMs I've seen (Bonabeau's) they clearly aren't. So
although the working definition of an ABM as "something that is written in
NetLogo, Swarm or Repast" just seems intuitively a bad thing, at least it
has the timing right.

Maybe this is all pointing towards the conclusion that there isn't a
meaningful discipline of agent based modelling... maybe it's all just some
social construct created by a bunch of academics with too much time and
processing power on their hands...

Robert (who seems to be in contentious mood today)

> -----Original Message-----
> From: Stephen Guerin [mailto:[hidden email]]
> Sent: Friday, March 11, 2005 1:06 PM
> To: The Friday Morning Applied Complexity Coffee Group
> Subject: RE: [FRIAM] Mathematics VS? Agent Based Modeling
>
> Let me restate my previous comment and repost the quote. I
> suspect the ABM vs. EBM debate is ill-posed. I favor
> Bonabeau's characterization of ABM more than Parunak's...
>
> "ABM is a mindset more than a technology. The ABM mindset
> consists of describing a system from the perspective of its
> constituent units. A number of researchers think that the
> alternative to ABM is traditional differential equation
> modeling; this is wrong, as a set of differential equations,
> each describing the dynamics of one of the system's
> constituent units, is an agent-based model. A synonym of ABM
> would be microscopic modeling, and an alternative would be
> macroscopic modeling."
>
> Robert writes:
> > So my contention is that while ABMs may be great for advisory,
> > open-loop systems they just aren't robust enough to form
> the core of a
> > closed-loop system. How would you feel if the automatic control
> > systems in your car (plane, nuclear reactor etc.) where all
> written in
> > NetLogo? ;)
>
> More comfortable than if it was written as a VBA add-in to Excel ;-)
>
> I know you're kidding, but I'm not sure what software
> platforms have to do with determining the appropriate
> modeling perspective for understanding the problem at hand.
>
> Also, there is a distinction to be made between Agent-Based
> Modeling and Agent-based systems. The routers on the internet
> coordinating to deliver this email is an example of a
> distributed closed-loop agent control system.
> (granted, they're doing it very poorly today, Roger. We need
> a new hosting company...). The Internet is an agent-based
> system, not an ABM. If I was proposing a new protocol I would
> want to simulate it first with an abstraction of reality.
> That would be an ABM.
>
> -S
>
>
> ============================================================
> FRIAM Applied Complexity Group listserv
> Meets Fridays 9AM @ Jane's Cafe
> Lecture schedule, archives, unsubscribe, etc.:
> http://www.friam.org
>
>
>