REPLY The Death of System Dynamics? (SD6212)

[System Dynamics Mailing List]

Posted by  Jean-Jacques Laublé <jean-jacques.lauble at wanadoo.fr>

Hi John

About your agreement, you say that we must be disciplined.
I agree with you that SD as a methodology needs a very high level of discipline.
Unfortunately until somebody entering the profession knows what sort of
discipline is required, will take several years.
He will learn to be disciplined by experience and after having made a lot of
bad models.

I then think that one must help people to be disciplined.
Actually the profession is going to much towards adding more power to the
tools than making them easier to use. SD for me is a powerful technique that needs year
of practice to be mastered. The problem is not adding more power but adding
more control.
Control on how the problem to be modelled is chosen so that it matches the
resources needed (time, money, expertise), if it is or not susceptible to
profit from the SD methodology, if the stake is high enough to justify the
amount of resources involved.
The control must be kept from the beginning of the definition up to the end
of the study.
It must help the modeller avoid the following traps (not precise enough
definition, setting bad boundaries, bad horizon, bad time step, bad level of
aggregation, and most important finishing with a too complex model,
difficult to analyze).
The control must permit to keep the level of complexity sufficient low, or
help analyze complex models (loops dominance etc.) and help simplify models
(suppress not useful material).
So the profession should change its goals if it wants more people using SD.

About the possibility to change one method against another, I think that if
people find that it can help them, they are ready to make a lot of efforts.
The competition in the business world is so hard, that anything that may
give a competitive advantage will be used, at the condition that the competitive
advantage is proven and available.

About the disagreement:
There are truly a lot of subjects to be studied at the condition that the
method does not focus on why things are supposed to have happened in the past but proposes
credible policies for the future.
Is SD represented at Davos? What could the profession show?
The last world model that has thousands of feed back loops?
The model is easy to use, but the problem is that it requires a blind belief
in its reliability.
Regards.
Jean-Jacques Laublé Allocar
Strasbourg France 
Posted by  Jean-Jacques Laublé <jean-jacques.lauble at wanadoo.fr>
posting date  Fri, 26 Jan 2007 10:55:01 +0100