RE: Criticisms and defense of ABM

[Konstantinos Alexandridis]

Paul,

        The theme of your proposed agent-based model, seams very
suitable for exploration using adaptive behavior agent attributes. The
literature is rich on models examining complex interactions between
individuals or groups of individuals on determining successful
strategies. Robert Axelrod published an excellent book - review of these
methodologies, in relation to Prisoner's Dilemma, adaptive behavioral
models as well as to the use of the genetic algorithm on determinining
successful strategies.
        Although in a primary, perhaps narrow view of agent-based
modeling, it might seam somewhat predictive; in reality though,
agent-based models project in the future the 'source' of the agent's
behavior. And that future projection is a more real-world process, since
it can overcome the usual barriers of error propagation and
statistical/econometric assumptions about the 'mean' behavior of the
sample data population.
        In contrast, the simulated results are able to reveal a variety
of complex features, interactions and systemic behavior of the
individual agents, not likely to appear in an averaged sample
population. It can, for example, reveal significant information about
the boundaries of the system's behavior, or about the limits that the
agent's can reach in their individual performances.
        My work with agent-based models is coping with land-use and
land-transformation processes that occur in a spatial/landscape setting,
including the human-behavioral component of the populations and
individuals occupying these spatial settings. Socio-economic behavior
and individual choices, are highly correlated with land use and land
cover changes through time and space. As in your case, decision-making
about the optimal choices cannot be assumed to be a matter of
rational-behavior in a traditional economic setting, but, rather, a
process where an agent (an individual) has to reach a decision from a
pool of available choices that they are subject to limited information
(and diverse access to it), hierarchical processes, and thus, they are
suboptimal, yet they can be viewed as optimal in the time-scale of the
individual. In other words, an individual on a given moment, and with
given (limited) information available, and with different access to
available choices, (and often subject to a series of other limitations
such as economic, social, behavioral, perceptional, age, or racial) has
to reach a decision. Although one could accept the rational-choice
assumption, must do so by recognizing the given limitations. There is no
economic instantiation of such a system, able to represent and
regenerate a whole population (and its data) in an individual level.
Agent-based models are able to perform this regeneration and to produce
likely-to-occur projections of reality.
        There are methods to test the validity of the simulated results,
especially, if the design and execution of the simulation can be
performed using available historical initial data (which can be used for
comparing the simulated with the historical data on sequential
time-steps). In addition, tests like updated versions of the Turing
test, or spectral analyses can be employed to assure the validity of the
simulated results.
        
Thank you,
With respect,

Kostas Alexandridis
Regional Environmental & Geospatial
Analysis Laboratory (REGAL)
Michigan State University
www.mabel.msu.edu


-----Original Message-----
From: address@hidden
[mailto:address@hidden On Behalf Of Paul Charteris
Sent: Tuesday, March 05, 2002 4:55 PM
To: address@hidden
Subject: Criticisms and defense of ABM


Dear Swarm Modelling Group

Yesterday I gave a rather informal seminar on the use of agent-based 
modeling to the animal genetics group at Colorado State University and 
(not unexpectedly) received a couple of criticisms - especially from my 
Ph.D. major advisor. Anyway, the two main criticisms that I attempted to

defend myself against were:

Criticism 1. One of my dissertation objectives is stated in this way: 
"To explore the utility of Agent-Based Models in describing the 
industry-level outcome of breeding decisions made by individual 
enterprises" As a note, these breeding technologies are animal genetic 
technologies that allow farmers to make more accurate genetic selection 
decisions.

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.

My view - Sometimes it is a valid objective to examine the methodology 
rather than just the outcomes of the methodology, especially to test if 
the methodology provides new insights into the system or opens up new 
hypotheses not previously testable.

Criticism 2 - All agent-based modeling is essentially prediction.

Their view - any outcomes of an agent-based model are a prediction of 
entities or processes in the system. Prediction can be thought of as any

declaration or estimate regarding the future.

My view - No the aim of agent-based modeling is frequently not a 
declaration or estimate of the future but frequently relates to 
observing the processes that lead to some future outcome. They argue 
that observing these processes is prediction in its own right. I 
counter-argued that we do not make a declaration or estimate of these 
processes at all- rather we simply parameterize the agents in their own 
model world and hope some "goodies" come out of the model. I guess the 
point is subtle, we (the modeler) do not try and anticipate the future -

we only parameterize the present and observe what possible futures may 
exist. Maybe this is an estimate of the future after all?

Any thoughts or comments on these points of view would be most welcome.

Cheers, Paul Charteris

---------------------------------------------------------
The Department of Animal Sciences
Fort Collins CO 80523-1171
USA
Ph. +1 970 491 5785
Fax. +1 970 491 5326
Personal homepage: http://ansci.colostate.edu/dep/people/grads/plc/
---------------------------------------------------------



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



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