Systems of problems

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
Bill Harris
Member
Posts: 31
Joined: Fri Mar 29, 2002 3:39 am

Systems of problems

Post by Bill Harris »

George Richardson writes:

> Bill asks if we cant model systems, apparently thinking that we can.
^^^^
There are too many of us "Bills" out here!. :-) That was another one.

> Im so attached to this notion of problem focus that I cant imagine you
> would not agree with it. (Have to admit to my myopia here.) So could you
> help me out by telling us what you have in mind when you suggest we can
> model a "system"?

Im firmly in the camp of "model to solve a problem", not "model a system,"
because everytime Ive tried to model a system, Ive failed. Most of the
time Ive tried to solve a problem, its worked.

I have seen a number of people who think you should model systems, though.
The usual rationale seems to be that one can create a comprehensive enough
model so that one can ask _any_ question of it and get a good answer. In
other words, one can create _one_ model and solve a wide range of problems
which occur in that system.

In a consulting style intervention, the hard part seems to be to get the
"client" to get involved in the process _before_ the modeling work is done
to help capture the decision making processes in a model. Those who want
to see a whole system approach seem to think the modeler can go off and
create this model without being involved, and the problem solving can be
done quickly and again without much involvement and thought.

So the rationale seem economic on two levels: economy in the number of
models created, and economy in the time the client/sponsor/whatever needs
to be involved. Ive just not seen it work that way.

My $0.02.

Bill
--
Bill Harris Year 2000 Program Office
mailto: billh@lsid.hp.com Hewlett-Packard Company
phone: (425) 335-2200 M/S 330
fax: (425) 335-2648 8600 Soper Hill Road
web: http://hpweb.lsid.hp.com:8080/~billh/ Everett, WA 98205-1298
cupertino: (408) 447-0452 (no voicemail)
DGPacker@aol.com
Junior Member
Posts: 4
Joined: Fri Mar 29, 2002 3:39 am

Systems of problems

Post by DGPacker@aol.com »

I cant resist this, which relates to modeling the "system" (rather than
issues). Theres a Stephen Wright segment in which he says "I spent all last
summer drawing a map of the United States--with a scale of 1 inch to 1 inch.
It was alot of fun, but a real problem to fold!"

Modeling the whole thing creates a model as complex and as incomprhensible as
the thing itself.

Regards, Dave Packer, STC.
From: DGPacker@aol.com
Bill Braun
Senior Member
Posts: 73
Joined: Fri Mar 29, 2002 3:39 am

Systems of problems

Post by Bill Braun »

In a number of recent threads many have cautioned against using SD to model
systems. Rather, its best application is when it is problem focused. I
believe that Akoff said that one aspect of applying systems thinking was to
think in terms of systems of problems (his way, I took it, of using
expansionistic thinking and recognizing large scale interdependencies). If
SD is best suited to exploring problems, and problems are considered from
an interlocking, interdependent systemic point of view, why is it not
reasonable to model large scale systems (admitting up front the probable
inverse relationship between complexity and [as noted in a current thread]
accuracy)?

Bill Braun
From: Bill Braun <medprac@hlthsys.com>
Jim Hines
Senior Member
Posts: 80
Joined: Fri Mar 29, 2002 3:39 am

Systems of problems

Post by Jim Hines »

Bill Braun wrote "...Akoff said that one aspect of applying systems thinking
was to think in terms of systems of problems...If SD is best suited to
exploring problems, and problems are considered from an interlocking,
interdependent systemic point of view, why is it not reasonable to model large
scale systems (admitting up front the probable inverse relationship between
complexity and..accuracy)?"

The problem with complexity is less with accuracy and much more with
understanding. (I think George Backus point was that predictive models will be
too complicated to understand easily, making policy creation more difficult).

In understanding a system you need some strategy of simplification -- any
human, biological or (most) physical systems are way to complicated to grasp in
their entirety. System dynamics has three main strategies for simplifying:
Concentrate on loops, view the world as stocks and flows, view the world as
aggregate actors, and (usually) view the world as deterministic . (In contrast
to, say, econometrics which simplifies inter alia by viewing the world as
linear. Or in contrast to pure stochastic approoaches which adopt the strategy
that the the world is probalistic, not causal).

But even with SDs strategic focus on loops, stocks, flows, etc, more
simplification is usually needed. Hence, the dictum to focus on a problem
instead of a system. Alternatively, you can focus on some other aspect of the
system. This usually requires a system dynamics "framework". The most famous
such framework is probably the project model.

What is happening here is that any systems approach (including SD) requires --
in addition to its usual simplification strategies -- an additional assumption
that some aspect of the system is separable from all other parts of the
system. The assumption that the causes of a problem are separable from all the
rest of the system seems to be a useful assumption in SD.

Regards,
Jim Hines
MIT and LeapTec
From: Jim Hines <
jimhines@interserv.com>
Bill Braun
Senior Member
Posts: 73
Joined: Fri Mar 29, 2002 3:39 am

Systems of problems

Post by Bill Braun »

Hi George,

>Id ask Bill, if someone asked you to model a city (e.g., Cleveland),
>would you know what to do? Id say I wouldnt, and I think you wouldnt
>either. So I conclude I cant model a system, without some definition of
>the dynamic problem focus of the modeling effort.

Good question. I wouldnt know where to start without a dynamic problem
focus as you say. The question also makes me think that Ive been using
the word system with a different meaning than you have. Although I used
the word system, I probably should have used the word subsystem. Another
question...if we describe a subsystem do we simultaneously and
automatically describe a dynamic problem focus?

>Im so attached to this notion of problem focus that I cant imagine you
>would not agree with it. (Have to admit to my myopia here.) So could you
>help me out by telling us what you have in mind when you suggest we can
>model a "system"?

What I had in mind when I posted the question was the healthcare system,
inclusive of all its stakeholders, and the dynamics of treating pathology
and improving health in a regional population of people. It acknowledges
dynamic factors external to the healthcare system per se (ones job,
physical environment, lifestyle, etc.). It excludes modeling any one
particular disease and its treatment.

I posed the question more as an exploratory one, not as a dissenting
opinion. I actually do find a pretty compelling argument in what you say.
It was the notion of Ackoffs systems of problems that seemed to make the
question a logical one. If problems can/should be considered from a
systems perspective, it seemed that eventually each of the dynamic problems
would eventually find their interdependent links.

By the way, Cleveland can be modeled by simply saying its a great town!

Thanks George,

Bill
From: Bill Braun <medprac@hlthsys.com>
Locked