Re: Criticisms and defense of ABM

[M Lang / S Railsback]

Paul Charteris wrote:
> 
> Dear Swarm Modelling Group

Hey, welcome to agent-based modeling! It's like a fraternity, there's
this hazing process you have to go through before everybody laughs, says
they were just joking, and let's you in.

(we wish)
 
> Criticism 1. One of my dissertation objectives is stated in this way:
> Their view - The aim of assessing whether a methodology is appropriate
> to a problem is not a suitable research objective in its own right, i.e.
> the study should not undertake simulation for the sake of determining if
> simulation works.

Seems like rephrasing your objective could be a useful way around this
opinion. You could point to the fairly substantial literature pointing
to ABM as a tool for understanding complex systems, and say that one of
your objectives is to apply this approach to your problem. In
particular, you can pose ABM as one of the only approaches for
understanding the links between system dynamics and individual-level
behaviors; so it seems natural to apply ABM to understanding the
relation between industry-level trends and decisions of individual
breeders. 

Or a popular approach is comparing an ABM with a more conventional
model, so you are now (a) comparing two modeling approaches to your
problem and/or (b) attacking your problem from two different directions.
I couldn't guess if that would make things better or worse with your
committee. 
 
> Criticism 2 - All agent-based modeling is essentially prediction.

Aha, perhaps another example of the problem of "How do you get results?"
question?, which Paul Johnson & I plan to discuss at SwarmFest.  Again,
you may be well served by talking about how you can use ABM to develop
an understanding of the mechanisms by which industry dynamics and
individual decisions. The "pattern-oriented" analysis approach is very
useful if you have some historic patterns to use as a basis- use these
patterns as a way to test, in your ABM, models of how industry dynamics
emerge from individual decision. I have a paper outlining this approach,
and an appplication (concerning fish behavior) now in press at a fancy
ecology journal. (See http://math.humboldt.edu/~simsys/Products.html and
let me know if you want a copy of anything you can't download.) Leigh
Tesfatsion mentions lots of examples in economics in her paper
"Agent-based computational economics: growing economies from the bottom
up" available at: http://www.econ.iastate.edu/tesfatsi/ace.htm . For fun
ask your advisor to read Auyang (1998 "Foundations of complex-system
theories in economics, evolutionary biology, and statistical physics",
Cambridge University Press) (and be very impressed if they actually
can/do read it), which discusses "synthetic microanalysis"- a process
amenable to ABMs for understanding how complex system behavior is linked
to individual behaviors.

But what's wrong with "prediction"? Isn't that what models are for, in
part? Haven't countless grad students gotten their PhDs by writing
models and using them in a prediction mode to propose solutions to some
problem (my wife did)? 

It sounds like your committee maybe doesn't see a lot of modeling
projects?

Steve
-- 
address@hidden
Lang, Railsback & Assoc.
250 California Ave., Arcata CA 95521
707-822-0453; Fax 822-1868


                  ==================================
   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.
                  ==================================