Russell Ackoff Dies
SteveM
Moved the center of OR from Case Western to Penn. Ahead of his time.
One of the first guys to forecast over-modeling as being a significant
negative for the practice of Operations Research.
SteveM
Paul
Mike Trick had an interesting (I thought) blog post on Ackoff's
passing: http://mat.tepper.cmu.edu/blog/?p=928.
/Paul
SteveM
> SteveM wrote:
> >http://www.philly.com/inquirer/obituaries/20091103_Dr__Russell_L__Ack...
>
> > Moved the center of OR from Case Western to Penn. Ahead of his time.
> > One of the first guys to forecast over-modeling as being a significant
> > negative for the practice of Operations Research.
>
> > SteveM
>
> Mike Trick had an interesting (I thought) blog post on Ackoff's
> passing:http://mat.tepper.cmu.edu/blog/?p=928.
>
> /Paul
Thanks Paul.
OR is fundamentally a consulting discipline. A consultant point of
view is usually required to understand the true nature of the
underlying problem. That's where Ackoff got it right. The need to
understand the "mess" in all it's complexity. But like it or not OR
attracts "engineers" who like to model. And they often prefer to
leave out the messy things that they can't measure. The product of
that mindset is a trash heap of OR recommendations that were rejected
because they only half baked business reality. And an OR academic
superstructure that is fading into irrelevance.
I'm still bemused by the Decision Analysis people waxing poetic about
their arcane methods that management teams barely understand and
rarely use. Handing out Ph.D.'s in DA is nuts because it only
perpetuates complexity with real little utility (no pun.) I know, I
know - But obtuse modeling can win you a DA Society award! Well that
and a buck thirty five will get you Metro ride here in DC.
Sorry for sardonic. But I'm more for Ackoff than against him....
RIP
SteveM
Paul
>
> OR is fundamentally a consulting discipline. A consultant point of
> view is usually required to understand the true nature of the
> underlying problem. That's where Ackoff got it right. The need to
> understand the "mess" in all it's complexity. But like it or not OR
> attracts "engineers" who like to model. And they often prefer to
> leave out the messy things that they can't measure. The product of
> that mindset is a trash heap of OR recommendations that were rejected
> because they only half baked business reality. And an OR academic
> superstructure that is fading into irrelevance.
Ackoff was in a sense before my time -- I was a "pure" math major all
the way through my PhD and didn't get into management science/OR until
I started teaching -- and I've never worked in OR in the "real world",
so I don't know enough about Ackoff's position to know if I would
support it or not. I'm also not averse to applying sophisticated
models to real problems, as long as they're actually attacking the
right problem and the assumptions aren't made for mathematical
convenience (the latter has been known to set me at odds with some of
my colleagues from time to time -- and I think that they believe a
mathematician asking about the realism of assumptions is some sort of
heresy).
At the recent INFORMS meeting, there was an interesting session about
the teaching of modeling to MBAs, which started with an exercise for
the audience (the audience acting as consultants to a non-profit) and
ended with a second exercise where two of the moderators did the
"consulting" and the audience watched the process. In between we
discussed the failings of MBA candidates when confronting a rather
general, data-poor problem (what Ackoff might call a "mess", albeit a
small-scale one). What fascinated me was that both the audience (of
OR professionals, if you count academics as professionals) and the
moderators engaged in the sort of "I have a hammer, ergo this must be
a nail" nonsense for which we would cheerfully hold the MBAs
accountable (assuming, of course, that their hammer was not our type
of hammer).
So I'm somewhat sympathetic to your position, although I don't know
that the problem lies in the use of advanced models so much as in the
force-fitting of the problem to the model, rather than the other way
around.
/Paul
SteveM
Paul,
Agree with you 100% I'm totally sympatico to your position. I LOVE
to model. I still get a rush when I see an math program that I've
formulated execute properly. I just understand modeling's limitations
in many business contexts.
I'm with you about the hammer and nail thing too. The MBA's are even
more dangerous, because giving them something like a simulation add-in
is like giving a teenage boy whiskey and car keys. It may be fun, but
wreckage is just around the corner. At least the OR guys can keep the
model on the road of mathematical reality.
Steve