Characterizing System Dynamics

R.M.Mooy telecom.tno.nl
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by R.M.Mooy telecom.tno.nl » Thu Nov 17, 2005 11:40 am

Posted by <R.M.Mooy@telecom.tno.nl>
Hi all,

What are fundamental properties of the method of Systems Thinking and/or System Dynamics? I've composed a list of properties that may be defined within these methods. My purpose for this would be to assess whether a given method of study displays Systems Thinking properties.

Is the list complete? Is it correct? Is it useful?
Please comment!

- Rutger Mooy


Systems Thinking Properties

Core elements

1 Interrelationship of components
A system consists of a group of interacting, interrelated or interdependent components (elements)

2 Boundary definition
A system can be defined by a boundary which determines the system's scope and separates its elements from elements in the system's environment

3 Holism
The presence of the whole system, as opposed to the collection of individual system elements, is explicitly taken into account as object of study

4 Feedback between elements
Elements within a system mutually influence each other, and, indirectly, themselves

5 Causality of relationships
Relationships between elements within a system are fundamentally causal

6 Dynamics
Behaviour-over-time of the system, including historical patterns of behaviour, are important in examining the system and its implications

7 Emerging behaviour
A system exhibits behaviour which is detected only when all sub-systems are interacting properly

8 Control of a system
Elements of a system may be controlled by actors, either from within or from without the system


Other elements

9 Non-linearity
Relationships within a system are usually non-linear and should be treated as such

10 Balance of long term and short term
A system usually has different behaviour in the short term, when compared to the longer term

11 The presence of subsystems
Within a system, elements may be part of functional subsystems which may be controlled by different actors

12 Combination of qualitative and quantitative properties Elements and their relationships within a system may be studied both qualitatively and quantitatively


Elements I'm not sure about

13 Visual language
A system can be described by a visual language to provide easy access to the inherent complexity and to facilitate learning

14 Hierarchy of subsystems
Subsystems can be classified into a hierarchical system of subsystems

15 Transformation process
A system has a transformation process, in which input elements are transformed into output elements

16 Focus on problem-solving
An important goal of describing a system is to solve a problem, rather than to describe the system

17 Focus on facilitation of learning
Examining a system may facilitate (interactive) learning and application of the results

Posted by <R.M.Mooy@telecom.tno.nl>
posting date Wed, 16 Nov 2005 15:36:38 +0100

Finn Jackson finn.jackson tangle
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Finn Jackson finn.jackson tangle » Fri Nov 18, 2005 12:08 pm

Posted by ""Finn Jackson"" <finn.jackson@tangley.com>
To me there is just one fundamental belief:

1) there is cause and effect: A can cause B

The rest then follows. Because A can also cause C and D. And H, R and W can also cause B we need to stop and think about what we think the different interactions might be. If we then test and quantify those ideas against our experiences of the 'real' world, then we can create a deeper understanding. And that enables us to make better predictions of the future, and so take better decisions.


Finn Jackson
www.theeschercycle.com
Posted by ""Finn Jackson"" <finn.jackson@tangley.com>
posting date Thu, 17 Nov 2005 12:15:56 -0000

Bill Braun bbraun hlthsys.com
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Bill Braun bbraun hlthsys.com » Fri Nov 18, 2005 12:08 pm

Posted by Bill Braun <bbraun@hlthsys.com>
I suggest the addition of emergence, where observed behavior is not described by any of the rules that give rise to the behavior; ref. the ""boids"" metaphor.

Bill Braun
Posted by Bill Braun <bbraun@hlthsys.com>
posting date Thu, 17 Nov 2005 07:34:50 -0500

Jean-Jacques Laublé jean-jacques
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Jean-Jacques Laublé jean-jacques » Sat Nov 19, 2005 12:50 pm

Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr> I agree with Finn, but unfortunately finding the causes and the level of their implication is the principal difficulty of SD, and I would say of most of the problems of our world.

Give me the way to know the causes of most of things, and I will be pleased enough and do very well even without the help of SD!

Regards.
J.J. Laublé
Allocar Strabourg France
Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr> posting date Fri, 18 Nov 2005 15:08:55 +0100

wakeland pdx.edu
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by wakeland pdx.edu » Sat Nov 19, 2005 12:52 pm

Posted by wakeland@pdx.edu
I do not agree with posting SD5621 about adding emergence (as in ""boids"") to the list of SD characteristics.

I believe that the behavior in SD is almost the antithesis of emergence, since the behavior is calculated directly by solving the equations.

Emergence is usually associated with agent based models (such at the Boids model), where the simple rules executed by the agents at the micro level look nothing like the behavior we observe at the macro level, and we are hard-pressed to see **any** connection between these simple rules and the observed behavior.

Wayne Wakeland
Portland State University
wakeland@pdx.edu
Posted by wakeland@pdx.edu
posting date Fri, 18 Nov 2005 08:20:28 -0800

Paul Newton plnwtn gmail.com
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Paul Newton plnwtn gmail.com » Sat Nov 19, 2005 12:54 pm

Posted by Paul Newton <plnwtn@gmail.com>
Donella Meadows' and Jennifer Robinson's book ""The Electronic Oracle"" (1985 or thereabouts), which compares the modeling paradigms of econometrics, optimization, input-output analysis, and system dynamics, has a chapter describing the characteristics of each of these paradigms. I think the system dynamics section in this chapter is one of the best short characterizations of system dynamics I've seen anywhere.

Paul Newton
Tukwila, WA USA (near Seattle)

Posted by Paul Newton <plnwtn@gmail.com>
posting date Fri, 18 Nov 2005 12:18:33 -0600

Bill Harris bill_harris facilita
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Bill Harris bill_harris facilita » Sat Nov 19, 2005 12:58 pm

Posted by Bill Harris <bill_harris@facilitatedsystems.com>
""R.M.Mooy telecom.tno.nl"" <system-dynamics@VENSIM.COM> writes:


>> What are fundamental properties of the method of Systems Thinking
>> and/or System Dynamics? I've composed a list of properties that may
>> be defined within these methods. My purpose for this would be to
>> assess whether a given method of study displays Systems Thinking
>> properties.


Rutger,

I think we had a similar discussion a few years ago. Many of your comments seem true of many systemic approaches (or systems approaches). I think we agreed the two key, unique attributes of SD were accumulation and feedback.

If you look for that earlier discussion, you may find it as part of a thread on discrete vs. continuous simulation. I recall John Sterman and I were both part of that thread; I don't recall the other participants' names.

Bill
- --
Bill Harris http://facilitatedsystems.com/weblog/
Facilitated Systems Everett, WA 98208 USA
http://facilitatedsystems.com/
Posted by Bill Harris <bill_harris@facilitatedsystems.com>
posting date Fri, 18 Nov 2005 08:57:13 -0800

Bill Braun bbraun hlthsys.com
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Bill Braun bbraun hlthsys.com » Mon Nov 21, 2005 11:57 am

Posted by Bill Braun <bbraun@hlthsys.com>
At 07:11 AM 11/19/2005 -0500, wakeland pdx.edu wrote:

Posted by wakeland@pdx.edu
>I do not agree with posting SD5621 about adding emergence (as in
>""boids"")
> >to the list of SD characteristics.
>
>I believe that the behavior in SD is almost the antithesis of
>emergence, since the behavior is calculated directly by solving the
>equations.

Wayne questions the addition of emergence to the characteristics of SD.

As a preliminary point, the original question inquired into ST and/or SD. That said, what of his comment on the relevance of emergence to SD? And, does the following response describe an accurate understanding of some SD basics?

The equations represent policies, do they not? In a social system, is there any one policy that describes the behavior of the whole? It is not their interaction that explains the behavior or performance of the system as a whole?

Further, my understanding is that all the equations that occur between one or more stocks and a flow are all part of the policy, disaggregated for clarity. I happen to be working with Urban Dynamics at present; I cannot identify any policy taken separately that determines the behavior of the model as a whole.

While the ""boids"" simulation does illustrate the behavior of the agents, the rules are still policies. From a Structure-Paterns-Events pyramid point of view, is there a distinction between the ""boid"" rules/policies and, for example, the rules/policies governing A/R and A/P, and the resulting Cash on Hand (all other inflow and oputflow policies being equal).

Perhaps I am using the term ""emergence"" incorrectly and Wayne's comments are on the mark. And, perhaps, Jim Thompson is once again wondering what I'm serving for breakfast.

Bill Braun
Posted by Bill Braun <bbraun@hlthsys.com>
posting date Sat, 19 Nov 2005 15:12:53 -0500

Jim Hines Jim VentanaSystems.com
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Jim Hines Jim VentanaSystems.com » Wed Nov 23, 2005 11:56 am

Posted by ""Jim Hines"" <Jim@VentanaSystems.com>
For me, the only defining characteristic of SD work is an interest in how structure creates behavior.

Jim Hines
jim@ventanasystems.com
Posted by ""Jim Hines"" <Jim@VentanaSystems.com>
posting date Tue, 22 Nov 2005 20:01:59 -0500

Martin Schaffernicht martin utal
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Martin Schaffernicht martin utal » Fri Nov 25, 2005 1:22 pm

Posted by Martin Schaffernicht <martin@utalca.cl>
Hi,

I've always read and heard that ""how structure creates behavior"" is what SD is about. Or, a little more general: ""how things change over time"".

In the first definition, ""structure"" appears as something different from behavior. But surely it also belongs to the ""things [that] change over time"".

How is ""structure"" generated according to the SD worldview?

""Saludos"",
Martin Schaffernicht
University of talca - Chile
Posted by Martin Schaffernicht <martin@utalca.cl>
posting date Thu, 24 Nov 2005 17:12:21 +0100

Richard Turnock richardturnock c
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Richard Turnock richardturnock c » Sat Nov 26, 2005 1:07 pm

Posted by ""Richard Turnock"" <richardturnock@comcast.net> Characterizing System Dynamics

R.M.Mooy reported: ""My purpose for this would be to assess whether a given method of study displays Systems Thinking [or System Dynamics] properties.""

Let's compare the characteristics of System Dynamics to a standard that has the same principles used in the physical sciences. Berlinkski (1) describes just such a standard. To use differential equations you have to justify the initial conditions and causality connections. In addition, Berlinkski concludes that when differential equations meet three requirements of a ""well-posed problem"" then they are a model for what science should be. (solutions exist, they are unique, and they vary continuously as the initial conditions change.)

Using words like ""emergence"" implies there is either a gap in the evidence for the causality chain or the system (of differential equations) is so complex we need a computer to do the calculations. So here is my suggestion.

A given method of study displays System Dynamics properties when:

1. Differential equations are used.
2. The initial conditions are based on evidence.
3. Causality is based on evidence.
4. The differential equations meet three requirements of a ""well-posed
problem"".

But suppose I'm trying to develop a theory without having sufficient evidence. Maybe there are gaps in the evidence for the causality chain or incomplete evidence about the initial conditions. I can propose a theory using an SD model and then follow a process of experiments or fact finding to complete the justification of the initial conditions and causality. As long as I acknowledge the gaps in evidence (ethics) and satisfy the four properties listed above, then I'm following a method of study called System Dynamics.

1. ""On the Origins of the Mind"" by David Berlinkski. Original source: Commentary, November 2004. ""The Best American Science Writing 2005"", page 227.


Richard Turnock
Beaverton OR
Posted by ""Richard Turnock"" <richardturnock@comcast.net> posting date Fri, 25 Nov 2005 11:24:58 -0800

Jean-Jacques Laublé jean-jacques
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Jean-Jacques Laublé jean-jacques » Sun Nov 27, 2005 2:34 pm

Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr> Hi everybody

Jim Hines writes
< the only defining characteristic of SD work is an interest in how < structure creates behavior.

What does creates mean?
Does it mean
1.. Has an influence
2.. Has a determinant influence
3.. is the only influence on the behaviour

What does structure mean?
A simplistic example

A business sells widgets and the number of widgets sold in a period of time is the demand for widgets constrained by the production capacity.

What is the structure in this case?
Is it only the fact that the widgets sold depend on the demand and the production capacity? Is the level of the demand of the production capacity included in the structure?

If I am interested in the behaviour of the turnover of the business, knowing that it depends on the demand and on production capacity is not enough, especially if the future demand is unknown.

The difficulties of all my models were not to find the causal structure of the model but the values of the parameters and the behaviour of the exogenous data needed by the model. And depending on these values, the behaviour of the results could change drastically, whatever the structure.

Regards.
J.J. Laublé Allocar
Strasbourg France.
Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr> posting date Sat, 26 Nov 2005 13:45:26 +0100

Justin Lyon justin1028 yahoo.com
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Justin Lyon justin1028 yahoo.com » Mon Nov 28, 2005 11:56 am

Posted by Justin Lyon <justin1028@yahoo.com>
J.J.

I tend to think of structure in a system dynamics
simulation as a hierarchy of objects that provide
insight into the interrelationships of the various
objects.

System Dynamics 'Structure' is a hierarchy of objects
assembled in the following process:

A. You define the boundary of the system.
1. You identify the feedback loops.
a. You identify the stocks
b. You identify the flows
1. You articulate the policies as
flow equations

Note that two variables, stocks and flows, capture ALL
the aspects of the system under study. And, the flows
capture all the information on the policies of the
system under study.

These processes are the building blocks of
simulations.

Everything in the universe is a process. EVERYTHING.

Simulation Science is commercially very valuable
stuff. Think billions of dollars not millions.

In business, we create objects that are assembled into representations of workflow processes with mapping to data warehouses for real-time information access.

In nature, like in a biochemical process, we can
create a simulation model that captures with some
degree of fidelity the processes that we wish to
better understand (like production of an amino acid in bacteria).

Stocks, such as a stock of customers, are
accumulations of many autonomous humans (agents). The
inflow of new customers captures all the information
on the policies (such as marketing) that drive the
accumulation of customers. The outflow of customers
captures all the information on the policies (such as
poor customer service) that drive the depletion of
customers.

Steve Guerin at Redfish refers to these as 'aggregated super-agents.' Dr. Suarez at Trinity University refers to them as 'upper-level agents.' Others refer to them as 'stocks.' Others refer to them as 'levels.' I call them 'bathtubs.'

Behavior over time EMERGES from structure. I use
EMERGENCE as a concept because all too often the
behavior of a system is counter-intuitive to what we
actually observe; it arises from nothing, yet is
clearly visible. Some explain, where possible,
behavior over time as resulting from the interaction
of multiple feedback loops and the concomitant
switching loop dominance.

That is, I think that the ABM folks (at Santa Fe,
Trinity and other spots around the world) and the SD
folks (at MIT, LBS, WPI, LSE and other spots around
the world) have their own particular jargon which, in
my not so humble opinion, all means something similar
– life is not reducible to a simple set of
deterministic and solvable equations.

That is, I argue that the concept of 'emergence' is
related or perhaps even the same as the SD concepts of 'behavior over time' as a result of structure?

Of the interesting systems for study in the world,
99.99% are not solvable by humans without the aid of
computers. And, many of those 99.99% of non-linear
systems are not 'solvable' with mathematics known to
humans.

This should really irritate a lot of people because
everyone keeps trying to find some Newtonian-like super-equation for biology that will 'answer' all our questions, but no one will ever find it. It is a will-o-wisp.

We can only solve the problems of the business,
nations and the world by the close and intimate
interaction between humans and computers. This
interaction must be governed by an appreciation of
reality as understood through the lens of Simulation
Science, that is, complexity science, chaos theory,
nonlinear science – choose the jargon that you like.

That is, this must always be kept in mind, ALL models
are false, some are useful and some are dangerous.

There are NO point predictions. No ULTIMATE answers.
No TRUTH.

There is only foresight into plausible and probable
future scenarios that may or may not happen.

More and more I'm thinking that system dynamics,
aggregated super agents, upper level agents, etc. ad
nausea, are the 'physics of biology;' the
'meta-engineering' of life through the wonderful, yet
blind and simple algorithmic process of evolution by
natural selection of agents that replicate.

Why is this important? Well, Simulation Science is
incredibly well-suited to studying human-designed
systems like global businesses, surfacing mental
models and enabling humans to re-design and manipulate
these systems to achieve goals -- like making more
money.

Before the 21st century, physicists and mathematicians
were the 'gods' of science. Think of Newton, Leibniz,
Kelvin, Fermi, Einstein. The 21st century 'gods' of
'science' are the biologists, computer scientists,
artists and engineers – people who are well-versed in
the power and manipulation of emergence (behavior over
time) in everyday life. Think of Darwin, Andronov,
Forrester, Kaufman

-Justin
Posted by Justin Lyon <justin1028@yahoo.com>
posting date Sun, 27 Nov 2005 10:11:50 -0800 (PST)

Jean-Jacques Laublé jean-jacques
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Jean-Jacques Laublé jean-jacques » Tue Nov 29, 2005 12:12 pm

Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr> Hi Justin I have a more pragmatic definition of system dynamics: It is a set of equations calculating the values of different variables Xi i = 1. n at different period of time:
Xi(t) = Fi (Xj(t),exogenk(t),paraml) j included in 1 .. N excepted i exogenk k = 1.m are data that take a different values depending on the period of time paraml are the r parameters l = 1..r

Some variables have particular equations;
Xi(t) = Xi(t-1) + Fi(Xj(t),exogenk(t),paraml) one generally calls them stocks or levels. in that case the Fi(Xj(t),exogenk(t)) is called a net flow or net rate. The exogenk data represent the outside world.

It is a mathematical description, and I have always thought that SD should be a branch of applied mathematics.

What I think is called the structure is the pure qualitative description of the set of equations. It is the set of equations with no description of the Fi and no values for the parameters.

The problem is what sort of assumption on can make about the behaviour of the variables if one knows only the structure?

Not to mention that one needs too the exogenous data if it is the case.

About the role of simulation
You say:
Everything in the universe is a process. EVERYTHING.

This supposes that causality is ruling the universe.
Nobody ever demonstrated this.
It is the better approximation found so far of the way the world is working, but what about the freedom of individualities? Not to mention people who believe in God or something like. This involves automatically some finality somewhere. Is the world absolutely deterministic and is there no finality somewhere?

Determinism needs always to have a preceding event.
So in determinism, there is no beginning of time.
This is why the big bang theory has a lot of difficulty staying deterministic when it gets near the singularity point, the 'beginning of the universe"". It needs some finality to explain the beginning as there is no preceding event. It needs some finality to get started.

There are some theories that explain that the beginning is a singularity point and has no real existence etc. but it still has to be demonstrated. I do not say that determinism is not ruling the universe, but that there is no proof of it. The definition of emergence is not clear for me. It is one of the two or both the next definitions.

Something that becomes visible and that was not visible before. Or Something that is slightly visible with weak effects, but that can once be much stronger.

In the first definition, the explanation of the non visibility of the emergence may be caused by some positive feed back loops becoming dominant. Why not.

But it can be caused by many other things, and why not by individual freedom? This is anyhow a very philosophical question and I do not know it is the place for such a discussion. Regards. J.J. Laublé Allocar Strasbourg France.Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr> posting date Mon, 28 Nov 2005 16:15:42 +0100

Don.Mango ge.com
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Don.Mango ge.com » Wed Nov 30, 2005 11:58 am

Posted by Don.Mango@ge.com
Dear SD Practitioners

I wish to encourage you to define your science in broader terms than equations and applied maths. I am an actuary by training and am learning SD because it provides tools and language to describe phenomenon which are typically beyond the scope of single-dimensional or static or closed form models. I would encourage you to avoid the ""applied maths"" description and instead focus on SD as a means for solutions to complex risk situations in organized activity of any kind. There is a big opportunity for SD to play a key role in the evolution of Enterprise Risk Management (ERM). The Sterman example of the auto leasing problem is state-of-the-art ERM. Most firms are not prepared to conduct studies like that because they do not know that understandings and models such as that are even possible.

I encourage you to apply your own science to your science. Why hasn't SD caught on more? Why isn't it taught in more business schools? Why aren't SD terms used more in the mainstream business press, which is increasingly technical in nature? I would venture (based on my own experience trying to learn and apply it) that the learning curve is high, and the practitioners tend to caught up in jargon-heavy discussions. The same critiques are made of the actuarial profession BTW (we have a lot in common).

I am merely encouraging you all to focus some of your time on making SD more accessible, and touting its strengths in the organizational solutions space.


Thank you

Don Mango, FCAS, MAAA
Director of R&D, GE Insurance Solutions
Director, Casualty Actuarial Society (2005-2007)
Posted by Don.Mango@ge.com
posting date Tue, 29 Nov 2005 07:43:18 -0600

Kim Warren Kim strategydynamics.
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Kim Warren Kim strategydynamics. » Thu Dec 01, 2005 11:59 am

Posted by ""Kim Warren"" <Kim@strategydynamics.com>
I agree with your comments Don.

I have a short section on why SD may not have caught on in 'Why has Feedback Systems Thinking Struggled to Influence Strategy and Policy?', Systems Research and Behavioral Science, 21, 2004, pp.1-17.This poses the question of whether we have either [a] excited people about the simple qualitative feedback view of their world, but not left them with anything that they can reliably and practically use on their own to make real and important decisions, or scared the living daylights out of them by implying that they can't hope to understand and solve their problems unless they undertake big modeling exercises.

Kim Warren
Posted by ""Kim Warren"" <Kim@strategydynamics.com>
posting date Wed, 30 Nov 2005 13:54:44 -0000

John Gunkler jgunkler sprintmail
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by John Gunkler jgunkler sprintmail » Thu Dec 01, 2005 12:00 pm

Posted by ""John Gunkler"" <jgunkler@sprintmail.com>
For fundamental questions about system dynamics, I always find it enlightening to return to the original source -- Jay Forrester. His ""Principles of Systems"" gives very clear, direct answers to most of the questions being asked here.

For example, to find out ""what is structure?"" turn to page 4-3 of that book, where you'll find ...

""Feedback Loop--Structural Element of Systems
Within the system boundary, the basic building block is the feedback loop. The feedback loop is a path coupling decision [defined nicely on the next page], action, level (or condition) of the system, and information, with the path returning to the decision point.""

And, on page 4-5:

""Principle 4.2-2. Feedback loop--the structural element of systems.
The feedback loop is the basic structural element in systems. Dynamic behavior is generated by feedback. The more complex systems are assemblies of interacting feedback loops.""

I interpret this to mean that ""structure"" (as applied to feedback systems, which are the only province of SD) therefore means the system's feedback loops and how they interact with each other within the boundary of the system under study.

You don't need to know much more than this to understand what we mean by ""structure."" You'll note that there are only two kinds of feedback loops (called ""positive"" and ""negative."") Also, feedback loops may also be characterized by their ""order"" -- a count of the number of level (condition, or state) variables within the loop. This leads to simple descriptions of structure, such as ""a second-order negative feedback loop.""

It is this simplicity which is the genius of the SD approach, in my humble opinion. Attempts to ""muddy the waters so that they may look deep"" are very much misguided.

Here's another gem from ""Principles of Systems"" that some seem to have forgot (from page 4-2):

""Principle 4.1-1. Closed Boundary.
In concept a feedback system is a closed system. Its dynamic behavior arises within its internal structure. Any interaction which is essential to the behavior mode being investigated must be included inside the system boundary.""

Forrester also declares: ""Formulating a model of a system should start from the question 'Where is the boundary, that encompasses the smallest number of components, within which the dynamic behavior under study is generated?'""

This means, then, that having exogenous data that ""represent the outside world"" is distinctly NOT a characteristic of system dynamics. Those who want to focus on exogenous inputs as causative agents in their models are wanting to do something else, not SD. Exogenous inputs can create transitory effects on a system, and while it can be useful to study these effects, the stable behavior of the system is not determined by exogenous inputs but, rather, by the system's internal structure (i.e., feedback loops and how they are connected to each other.)

As to ""... what sort of assumption on[e] can make about the behaviour of the variables if one knows only the structure?"", I believe the SD answer is, simply, ""One can completely describe the qualitative behavior of the variables. There is no 'assumption' required."" As to the quantitative behavior of the variables of interest, one must parameterize the model to get to that.


John Gunkler
Posted by ""John Gunkler"" <jgunkler@sprintmail.com>
posting date Wed, 30 Nov 2005 10:54:05 -0600

Carl Betterton carlb uga.edu
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Carl Betterton carlb uga.edu » Fri Dec 02, 2005 11:55 am

Posted by Carl Betterton <carlb@uga.edu>
As John Gunkler suggests, if the boundary is defined and structure is fixed, how can there be any emergence? The concept of emergence, so far as I have seen, relies upon contextual factors that are ""outside"" the system boundary.

Posted by Carl Betterton <carlb@uga.edu>
posting date Thu, 01 Dec 2005 16:21:24 -0500

Jean-Jacques Laublé jean-jacques
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Jean-Jacques Laublé jean-jacques » Fri Dec 02, 2005 11:56 am

Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr> Hi Don I leave the exact definition of SD to the head designer of the field Jay Forrester, who knows better than me. As a user there is a strong difference between the tools, highly based on mathematics, offered and the way to apply them. The way to apply SD considered as a mathematical tool, depends on a host of elements that are far away from the scientific mathematical definition of SD. It is up to anybody to include or not the way to use SD, in the SD definition or not. The advantage of the mathematical SD definition is that it has at least the benefit of the rigor, while keeping the same rigor with the way to use it is to my opinion not precise at all at the actual state of knowledge of the field. This high difference between the two parts of the field, makes the application of the method uneasy. One has on one side a very mathematical method and on the other side, a huge imprecision on where and when to apply it to make it useful. Lots of other methods have not the same difficulty.

I prefer the mathematical definition because the problem and the tools to solve it are different. Identifying a problem with a definitive tool is highly dangerous in the perspective of the user who is more interested by the resolution of his problem than by the tools used. Identifying the problem and the tools, limits too the different approaches possible and deserves the possibility for the SD field to move towards more useful practices or even definition.

As to the direction the field tends to follow, I think it is more on the mathematical side. Suffice to look at the evolution of the software. They emphasize more the analysis of a model, or increasing the range of problems being solved than the ease of constructing a useful model, forgetting that before a model is analysed it must first be built. The reason of this evolution is to my opinion the influence of the experts of the field that know by experience how to build a useful model, and need more power to operate their current models and broaden the range of the field. The experts having automatically a big influence on the development, it is their influence that is predominant. The second reason maybe is that the development of the mathematical capabilities are less prone to be copied easily but I am not sure of the fact. One other reason is that they all want to add value to their software, meaning more complexity. Another reason is maybe the difficulty to improve methods of using SD and eventually to protect them. The developers of the software know certainly better than me about this question. I prefer when solving problems to follow the advice of Francis of Assisi: 'Do the necessary, then the possible and suddenly you do the impossible.' The field to my opinion would largely benefit from an effort towards making the SD as a mathematical tool, easier to use and I mean the software tools should largely participate to this effort. Unfortunately the tendency is always towards making things more complex then they are because it values intellectual speculation versus pragmatic application. The possibilities of development are numerous: First of all, the conditions that must meet a problem to benefit from the application of SD should be clearly defined. This is the first and more important effort that should be undertaken. Second, helping people to define precisely a problem that can be solved with the tools, expertise and time available, and can deliver a sufficient useful outcome is critical too. If that was already improved, that would be fantastic.

In my case, the model that I found the more useful, was built in four hours, and is a pure half financial and half actuary problem, that is not easy at all to solve without using a model. I tried to understand why. I thing it is related to the fact that the structure, and the parameters of the model are stable or easy to forecast over the years, and does not use any outside data (John Gunkler will be glad to know that) as is often the fact for actuary and financial problems.

I think that SD could be widely used in this field.
This represents only my opinion, being only a simple user, and I do not pretend that it represents even a fraction of the average user ideas on the subject. The best would be to investigate further on. Regards. J.J. Laublé Allocar Strasbourg France. Posted by =?iso-8859-1?Q?Jean-Jacques_Laubl=E9?= <jean-jacques.lauble@wanadoo.fr> posting date Fri, 2 Dec 2005 10:48:17 +0100

the invigilator the_invigilator
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by the invigilator the_invigilator » Fri Dec 02, 2005 11:57 am

Posted by ""the invigilator"" <the_invigilator@hotmail.com>
i would concur with Don's comments. i think that SD has been selling itself short.

rather than following the worn path of the pursuit of publications, i think SD should take a more revolutionary route. SD should seriously consider marketing itself as the 'universal modelling science' and claim that it is the *only thing* that can unite all the other fields of study.

while this may seem a tad arrogant, i can't honestly see how else SD can find itself an ongoing role. SD is like a formula 1 racing car used to go to the shops for bread and milk. it seems like a fish out of water, none too practical and idles like a dog.

i say put SD on the race track and show the rest of the world what we can do. assert our claim to the throne. invite the psychologists, the sociologists, the anthropologists to play by our rules. for too long these different perspectives on the human condition have been isolated and thereby ineffective

SD can unite them. We are not a science in ourselves, we are the framework that unites the sciences. we can give them the direction, relevance and credibility they have been lacking...and the power of concerted effort.

we have to stop thinking like engineers defending our niche on the underbelly of academia, thinking we can't possibly have the right to power. i mean...we're engineers for god's sake...we don't even tell our wives what to do.

well, we have the power at our fingertips, and its about time we started taking the responsibility that comes with it. no one else can or is going to, and lord knows the whole sorry lot of us are running out of time while the SD community remains hesitant to raise its voice above a querulous whisper...

invig
Posted by ""the invigilator"" <the_invigilator@hotmail.com> posting date Fri, 2 Dec 2005 06:00:22 +1100

Justin Lyon justin1028 yahoo.com
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Justin Lyon justin1028 yahoo.com » Fri Dec 02, 2005 12:03 pm

Posted by Justin Lyon <justin1028@yahoo.com>
Don,

I agree with your comments about jargon-heavy
discussions.

One of the primary reasons SD has not caught on is the
SD community itself.

It's a classic case of compensating feedback.

As we move thru the tipping point for simulation
science in 2006/2007, we can expect to see a lot of
resistance within parts of the community regarding the
wider adoption of SD and other simulation techniques.

Similar compensating feedback occured within the
defence and science communities when the Internet
tipped in 1993/1994.

-Justin
Posted by Justin Lyon <justin1028@yahoo.com>
posting date Thu, 1 Dec 2005 05:14:00 -0800 (PST)

Bill Braun bbraun hlthsys.com
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Bill Braun bbraun hlthsys.com » Sat Dec 03, 2005 12:57 pm

Posted by ""Bill Braun"" <bbraun@hlthsys.com>
I'll ask again if my understanding of the word emergence is faulty. I think it is the behavior that results from the interaction of policies. In any organization, no one policy - or even a subset of policies, randomly or intentionally selected - will completely describe behavior. While the policies themselves are a defined structure (and fixed, as Carl notes) and the boundary of the model is what it is (likewise as Carl notes), why is it that (given a complete list of policies) we cannot intuit the system's behavior.

Likewise, if the aggregate behavior of a system is not an emergent property of its policies, should we not be able to observe the system behavior and know the underlying policies? Why would modeling be necessary?

My understanding of emergence is that is is completely endogenous. In a video (whose setting I cannot recall) Dr. Forrester notes that it is rarely events that happen exogenously that have adverse impacts on organizations; rather, it is the failure of endogenous policies to effectively respond to exogenous events that spells out its fate.

The emergence of which I speak is the endogenous behavior of a system, not its exogenous environment.

Bill Braun
Posted by ""Bill Braun"" <bbraun@hlthsys.com>
posting date Fri, 2 Dec 2005 10:55:02 -0500 (EST)

Francisco Perez francisco77pp ya
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Francisco Perez francisco77pp ya » Sat Dec 03, 2005 12:57 pm

Posted by Francisco Perez <francisco77pp@yahoo.com>
Hi.

I would like to consider Invigilator comments about SD as Universal Modelling Science. I think that if that's the reference to put SD on the highspeed lane, a fast crash is for sure. Trying to put the SD Community in that track is like saying the we have no ears and eyes to accept that SD is just one strong modelling approach/technique/tool among the existing and potential techniques/approaches.

I think SD will evolve as science and approach as long as the SD community accepts in a wider concept that SD is not the only way to model our perceptions on whatever subjects of the real or vritual world. In fact, SD will gain more involvement from non-SDists as long as we apply synergy (conceptually and technically) between SD and other modelling tecniques such as Agent based modelling, OR Modelling (LP, NLP, QLP), etc. We have to take advantage that SD is the closest approach to common management and communitiy language since its foundations rest on Systems Thinking.

I would like to see more papers and discussions about real life practice on combining SD with other methods and techniques, but I mean in a very high percentage of the papers (say more than 40%).

Another important thing, under my perspective, is the difficulty fo SD modeling to communicate to the final stakeholders what the real benefit in financial terms will be. For example. ""from this model you will gain such a marvelous insight that your decision framework / policies will help you improve/optimise business overall efficiency by xxx %, ie. yy MM USD"". That is a matter we all have to work on. Building frameworks to quantify potential benefits from SD modeling (or hibrid SD+X modeling).


Regards
Francisco Perez
Posted by Francisco Perez <francisco77pp@yahoo.com>
posting date Fri, 2 Dec 2005 08:33:38 -0800 (PST)

Bill Harris bill_harris facilita
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Bill Harris bill_harris facilita » Sat Dec 03, 2005 12:57 pm

Posted by Bill Harris <bill_harris@facilitatedsystems.com>
""Kim Warren Kim strategydynamics.com"" <system-dynamics@VENSIM.COM> writes:


>> This poses
>> the question of whether we have either [a] excited people about the
>> simple qualitative feedback view of their world, but not left them
>> with anything that they can reliably and practically use on their own
>> to make real and important decisions, or scared the living
>> daylights out of them by implying that they can't hope to understand
>> and solve their problems unless they undertake big modeling
>> exercises.


It seems possible the ""or"" could be replaced by an ""and."" :-)

Bill
- --
Bill Harris http://facilitatedsystems.com/weblog/
Facilitated Systems Everett, WA 98208 USA
Posted by Bill Harris <bill_harris@facilitatedsystems.com>
posting date Fri, 02 Dec 2005 19:36:16 -0800

Colm Toolan subscriptions toolan
Junior Member
Posts: 14
Joined: Fri Mar 29, 2002 3:39 am

Characterizing System Dynamics

Post by Colm Toolan subscriptions toolan » Sat Dec 03, 2005 12:57 pm

Posted by ""Colm Toolan"" <subscriptions@toolan.de>

Invig wrote >> SD should seriously consider marketing itself as the 'universal modelling science' and claim that it is the *only thing* that can unite all the other fields of study.

Slightly off-topic, but there is an entertaining story which centres on SD as the answer to all of the worlds problems in Andreas Eschbach's novel ""Eine Billion Dollar"" (which translates as ""A **Trillion** Dollars"") published in German by Bastei Lübbe, 2003.

The story centres on a hero, who inherits a trillion dollars, with which he is supposed to ""save the world"", using his own defition for that term. In the process of seeking a solution, his advisor (who unfortunately turns out to be the ""baddie"") makes extensive use of - and explains - SD concepts and models. It certainly provides a good introduction to SD for the lay reader in the context of entertaining SF-reading.

Maybe more widespread adoption by popular writers would do more to increase the awareness?

Regards,

Colm Toolan, Business Architect
Posted by ""Colm Toolan"" <subscriptions@toolan.de>
posting date Fri, 2 Dec 2005 13:03:30 +0100

Locked