SD Reviews of A New Kind of Science by Stephen

This forum contains all archives from the SD Mailing list (go to http://www.systemdynamics.org/forum/ for more information). This is here as a read-only resource, please post any SD related questions to the SD Discussion forum.
Locked
Jay Forrest
Junior Member
Posts: 12
Joined: Fri Mar 29, 2002 3:39 am

SD Reviews of A New Kind of Science by Stephen

Post by Jay Forrest »

Greetings Peter!

Unfortunately your link to slashdot seems to be dead so I cant respond
directly to what you read, but it seems to have given you an erroneous
interpretation of Wolframs book.

Wolfram clearly does not suggest that the relatively simple cellular
automata (or rules based systems) he espouses can "faithfully replicate
reality." He also clearly denies the ease and ability of interpreting
results from rule based systems and automata.

On page 4, "simple programs seem to capture the essential mechanisms
responsible for all sorts of important phenomena phenomena"..."seemed to
complex to allow a simple explanation." He clearly stops short of saying
they can capture reality. He sounds more like Jay Forrester suggesting that
a simple SD model can capture the bulk of the behavior of a complex system.

On page 31, "it becomes almost impossbile to predict -- even approximately
-- what the cellular automaton will do."

On page 40, "For normally we start from whatever behvior we want to get,
then try to design a system that will produce it."..."unless we can foresee
how a system will behave we cannot be sure the system will do what we
want." He goes on to suggest that rule based systems and CAs are equivalent
to computer programs and observes on page 46 "it is usually very difficult
to fresee what even a simple program will do."

I have not yet finished the book but I see no evidence of his suggesting
that one can look at the output of a CA model and interpret how the pattern
of behavior arose or explain how the behavior arose. My experience with
agent based modeling suggests that low level rules create a layer of direct
and obvious behaviors along with cross impacts which are not. As depth and
richness of the model rises those emergent behaviors react to beget new
emergent behaviors. Trying to interpret what causes behavior at the surface
is like trying to peel an onion of layered cross impacts and influences --
much like analyzing the real world -- except that one never knows for sure
that the rules of an agent model are complete or accurate relative to
reality. And a lot of emergent behavior -- and particularly departures in
behavior -- can arise or not from how exceptions to the rules are handled
and from missing rules.

On page 6, Wolfram says "the very results of the book show that there will
inevitably be fundamental limits to the application of scientific methods."

Regards!
Jay Forrest

P.O. Box 701488
San Antonio, TX 78270
Tel: 210.355.0429
E-mail: jay@jayforrest.com
Locked