I do not have the Kempis reference, so I am shooting in the dark a little,
but here goes. What I hear are some aspects of the distinctions between
Continuous System modelling (i.e. Systems Dynamics) and Discrete Dynamic
Systems modelling. Altho some continuous modelling software has some
discrete capabilities, they are secondary to the conceptual framework
involved. To get at the Discrete Chaos in systems which I mentioned,
discrete causality must be explicitly modelled, continuous systems
modelling packages do not encourage this.
We regularly apply such discrete dynamic modelling to strategic planning
and decision making using a technique which we call "event-based financial
dynamics". This is where all of the revenues and expenditures of the
system under study are driven explicitly by the discrete transactions
which generate them, and Not by using all of the widly approximate and
aggregating assumptions in conventional financial modelling. At all times
we pay attention to accurate representation of dynamic variation and the
occurrence of stoppages and breakdowns. In this manner, we can obtain
verifiably good representations of current system behaviour, and have
strong confidence in the behaviour of alternative systems under study.
In making the financial and other strategic measurements conform to the
vagaries of the operational level behaviour of the system, practical
systems planning results.
With respect to the component systems referred to by Kampis... I am
assuming that this means systems with large numbers of similar components,
wherein the explicit modelling effort would be exhorbitant. There is a
technique called "meta-modelling", which tries to fit a multiple
regression curve to the generic behaviour of such components, and then to
use this regression sub-model in an abstract systems model. We have
experienced the need for this approach in some logistic and transportation
situations, and are gradually experimenting with using a sub-model
approach with better non-linear capability than regression, namely genetic
algorithms and possibly artificial neural networks.
I am frankly startled by Fred's observation about the incapacity of
systems dynamics to help with practical decision-making. With discrete
systems dynamics we have helped railroads in Australia for example, to
save millions of dollars in their strategic capital acquisitions.
Certainly, we have used systems dynamics principles as part of our "big
picture" modelling in such situations.
Hoping to learn more, Doug Seeley.
---------------------------------------------------------------------------------
-----------
Dr. Doug Seeley: compuserve 100433,133... Fax: +41 22 756 3759
InterDynamics Pty. Ltd. (Australia) in Geneva, Switzerland
"Integrity is not merely an ideal; it is the only reality."