REPLY The Death of System Dynamics? (SD6204)

[System Dynamics Mailing List]

Posted by  "Alan McLucas" <a.mclucas at adfa.edu.au>

At the risk of entering the impending fray like a prize fighter leading with
his chin, I declare my support of Richard's criticisms.  Before proceeding,
let me independently thank Cognitus Ltd for their long time continued
sponsorship and support. Cognitus Ltd is a name strongly associated with
both the System Dynamics Society and the System Dynamics discipline.  

Firstly, we members of the Society should feel uneasy that we have created a
situation where such criticisms, indeed, might be justified.  Of course, it
is essential that discussion of these criticisms limited to this closed
forum.

Secondly, it remains a serious concern (as I see it) that despite so much
having been written about feedback systems thinking and system dynamics (SD,
for brevity) that SD has struggled to achieve the strong influence over
strategy and policy formulation, and decision making that we enthusiastic SD
proponents might expect.  Richard's point that SD, its tools and techniques
are just too much trouble for most managers is unfortunately too true.  We
often distance ourselves from our organizational clients and the public and
I believe we can be fairly criticised for haring off at a tangent, pursuing
our own myopic goals.  On the positive side, there have been numerous
serious attempts to make SD more accessible and understandable.  Notable
amongst these are Peter Senge's 'The Fifth Discipline', John Sterman's
'Business Dynamics', Kim Warrens work on 'Strategy Dynamics', and the
30-year retrospective on 'Limits to Growth'.  But in a global sense the
extent of readership, and consequent impacts in terms of changing the public
views have been limited.   In a recent Government-initiated public forum on
environmental impact, I was stunned by the collective ignorance the work
Jorgen Randers and other "prominent" SD researchers, and of SD itself. To
engage clients (and the public) in a dialogue based upon a rigorous
understanding of SD is intellectually demanding.  Unfortunately, most people
cannot be bothered, if for no other reason than human cognitive capacity is
seriously outmatched by real-world complex dynamic problems. It therefore
comes as no surprise that in 'Business Dynamics' John Sterman suggests that
one should use expert modellers in SD interventions.  Unfortunately, not
everybody engaging in SD can be experts - we have all been novices at some
time.  Well meaning but inexperienced or naïve SD modellers can produce
incomplete, erroneous and misleading advice (and whilst I have considerable
anecdotal evidence for this effect, to pursue this in any detail would be a
distraction).  None of us can claim to have been right on every occasion,
and it is unfortunate that we cannot always agree amongst ourselves.
Further, I suspect that there have been too many occasions when we have
inadvertently given untimely or inappropriate advice or our clients have not
been able to interpret what we have told them - and we have to take full
responsibility for the latter.  Rouwette, et al in their 1999 study of group
model building projects found that the most problematic of nine areas of
investigated was in development of a 'common language'.  Within the Society
we must strive to achieve a common language and highly effective ways of
engaging both our clients and the public, thereby improving the credibility
of SD.  We must seek to understand what we can do to dramatically improve
the acceptance and implementation of SD-based strategies. 

Thirdly, exemplary practice in SD and the self discipline needed to produce
it are not always evident (most unfortunately).  Having reviewed many
conference papers and seen the products for which (some) consultants have
been paid, I have been appalled by instances of poor quality and lack of
rigour.  This is most evident in research papers which have been
subsequently presented, relatively unchanged, despite receiving adverse
reviews.  The Society has a duty to improve this aspect.  Here I must stress
that I believe that the majority of SD researchers, practitioners,
consultants, teachers and students do excellent work.  The greatest
contributions to the credibility of SD come from those exemplary practices
which come under public scrutiny.  So it is that threats to credibility
arise from examples of poor practice.  Unfortunately, the latter are
remembered longest.  We have a duty to be disciplined and strive continually
to achieve exemplary practice.

Fourthly, the SD discipline appears to be forging ahead oblivious to
hard-won knowledge of other systems disciplines.  In systems engineering the
need for top-down design and bottom-up construction is widely accepted, and
for very good reasons.  Optimisation of sub-systems does not lead to
system-level optimisation, and sub-systems cannot be considered in total
isolation.  We have forgotten how strongly the formative years of SD were
influenced by these principles.  Whilst these principles are strongly
embedded in control theory and electrical engineering, and SD is founded (in
part, at least) upon them, their significance appears to be unknown to many
SD modellers.  Jay Forrester makes the point the systemic problems we
address through use of SD can be significantly more complex in terms of
order and number of feedback loops than those which engineers face.
Consequently, our work is potentially more difficult. We need to strive
harder to develop our collective knowledge whilst never forgetting that SD
derives its strength from a solid understanding of systems principles.  

Regards,
Alan
 

Dr Alan McLucas
School of Information Technology and Electrical Engineering,
UNSW at ADFA,
Australian Defence Force Academy,
Northcott Drive,
CAMBPELL  ACT 2600
AUSTRALIA
Posted by  "Alan McLucas" <a.mclucas at adfa.edu.au>
posting date  Wed, 24 Jan 2007 15:26:54 +1000