REPLY Models, Problems and Systems (SD7015)

[SDMAIL Jay Forrest]

Posted by  "Jay Forrest" <systems at jayforrest.com>

Thanks for bringing this up, Kim! 

I sympathize with your question for though the nature of system dynamics
includes consideration of boundaries, there is a certain tension between the
concepts of modeling the problem and modeling the system. It would seem to
me that in an ideal world the boundary examination related to modeling the
problem SHOULD expose the pertinent factors outside the problem details that
deserve consideration. In practice, I have to question that the boundary
examination is routinely adequate and would suggest that the ideal path
involves some balance between modeling the problem and modeling the system.

System modeling tends to be difficult for the resulting models are: 1) too
large, 2) increasingly involve extraneous variables and inputs beyond
control, and 3) likely to increasingly involve interpretation rather than
observable, clearly pertinent issues, as perhaps three of the most
significant barriers. And, as Jack suggests, the meaning and goal for the
model grows fuzzier as one models the system.

A successful, clean, problem-oriented model is a wonderful thing and can be
highly influential for its obvious pertinence and relative clarity. But it
seems critical to consider at some point the larger systems with which the
problem oriented model fits and functions. At risk of being overly
simplistic, imagine modeling a brake problem on a car that has a cracked
block or a leaking radiator. Failure to recognize the implications of other
areas are likely to result in model results that are locally valid but
trivial or meaningless to the entity in question. 

If the goal is to provide insights that have maximum value, I must conclude
that some balance between problem and system oreintation are necessary and
feel it is appropriate to ask the possibly painful question, "Is it possible
that overly narrow modeling has contributed to the failure of SD to achieve
the impact on business, industry, and government that we all (I think) feel
it is capable of?" The same question could also be asked of system-wide
models.  Again, it seems clear that each problem demands some different
balance between focus on problem and whole system.

Thanks!
Jay Forrest    
Posted by  "Jay Forrest" <systems at jayforrest.com>
posting date  Thu, 8 May 2008 09:49:37 -0500