Posted by Justin Lyon <
justin1028@yahoo.com>
Francisco and others,
Please allow me to clarify and at the risk of starting
a dialog (or even worse, a flame war) that many will
have already discussed in the past . . .
Simulation Science, in my mental model, is the
application of 'linear science' **AND** 'nonlinear
science' to solving problems in the real world.
'Linear science' is located within the larger set of
'nonlinear science'.
I hate the phrase 'nonlinear science' as to me it's
like calling Zoology the 'study of nonelephant
animals.' (thanks go to the mathematician Stanislaw
Ulam for this simile).
So, system dynamics is a discipline within the larger
set of Simulation Science.
Simulation Science is the use of computers and the
application of nonlinear techniques (and 'linear'
techniques when appropriate) for understanding our
complex world so that people can solve real world
problems more creatively and efficiently.
Techniques used within Simulation Science might come
from a variety of fields, e.g., regression analysis, agent-based modelling, discrete-event simulations, system dynamics, fractal mathematics, dynamical systems, chaos, etc. ad nausea depending on the problem being studied.
That is, people need to understand the real-life
problem and then choose the appropriate scientific
discipline to analyze it and develop policy. Maybe
it's an SD model, maybe a discrete-event model or
maybe some combination of a variety of techniques.
Just make sure the scientific technique you are using
embraces the facts of complex adaptive systems.
Greater than 90% of all business problems in the world currently being analyzed by the vast majority of consultants, economists, marketers, etc., are almost exclusively analyzed using techniques that paint patently false simulacrums of reality.
Many of these traditional simulacrums (often in the
form of highly complex spreadsheet models and implicit
mental models) include assumptions that violate the
most basic of physical laws.
They are more of a hindrance to decision making than a
benefit. I argue that reductionism analysis is more
often used for improving perceived certainty and
reducing perceived risk (leading to the all to often
heard lament, ""we did a bunch of analysis before it
all went wrong, so its not our fault, rather something extrinsic to us caused our failure).
Nothing wrong with doing what you’ve always done other
than the technology for applying simulation science is available (after decades of refinement) in the form of fast computers, excellent modelling software, solid management science, etc., so that humans can build evermore robust simulacrums for improving their decision making.
Like a good book that has been relentlessly edited
over time, simulation models become simpler over time
(unlike traditional analysis which gets more complex
over time). Over time, the user interface gets more
complex and robust, but the underlying simulation
models (objects) should get simpler.
Or, the more time I have, the simpler I can make the
simulation objects. The complexity of the model arises
when you connect a bunch of very simple objects
(molecules) together into a larger simulation. Sort of
like the complexity that arises when you connect a
bunch of atoms together into molecules into cells into
organs into humans.
It was *nearly* impossible *before computers* to truly understand the dynamics of businesses, because businesses are complex adaptive systems and it is IMPOSSIBLE for any human being, no matter how clever, to solve high-order, nonlinear, dynamic systems other than in the most 'gut-feel' sort of way by using the most powerful computer on the planet (our brains).
Nothing wrong with that either -- I just think it's
easier for people to follow their leaders if they make
their mental models explicit in a computer simulation
instead of requiring faith in the hidden mental
models.
Simulation Science insights have diffused slowly over
the past fifty years from the natural sciences to the
social sciences.
A 2002 citation search compared 5,400 social science
journals against the 100 natural science disciplines
covered by INSPEC (>4,000 journals) and Web of Science
( >5,700 journals) indexes. It shows keywords comput*
and simulat* peak at around 18,500 in natural science,
whereas they peak at 250 in economics and around 125
in sociology. For the keyword nonlinear citations peak
at 18,000 in natural science, at roughly 180 in
economics, and near 40 in sociology.
In the words of the researchers who conducted this
study, ""How can it be that sciences founded on the
mathematical linear determinism of classical physics
have moved more quickly toward the use of nonlinear
computer models than economics and sociology—where
those doing the science are no different from social
actors—who are Brownian Motion?""
My answer: It takes time.
So, one of the key challenges I'm facing as I
promulgate simulation science insights outside the
academic world, is . .
How do I make the CEO realize that the reliability and
accuracy of his decision-making can be improved by
using simulation science?
I believe that this involves educating managers that
it is time to set aside their historical (and
erroneous) preconceptions and prejudices of how to
manage a business in the 21st century.
Few executives today realize that the complexity of
business is genuinely beyond them. And that's not an
easy sell.
Humans, thanks to computers and only with the help of computers, can deepen our understanding of the world we live in. And, that’s an even harder sell.
-Justin
Posted by Justin Lyon <
justin1028@yahoo.com>
posting date Sun, 4 Sep 2005 04:57:25 -0700 (PDT)