Talk:System dynamics/Archive 1

Archive 1

Archive 2

This archive page contains old discussion that is no longer relevant to the current version of the page.

First draft

hi we just started this "system dynamics" wikipedia page today 30/11/2002. please help make it better (sorry for saying the obvious...)

we're at a dynamics workshop in Gdansk, Poland...

need more references

It's a good start. One thing that would help immensely is to add some references for the interested reader. Are there any textbooks you would recommend? The reference to the article on "Simile" seems to be of very little value.

Strange

Some notes that might help to improve the page:

• It is not clear what makes systems dynamics differ from classical cybernetics, scientific modeling, systems theory, and other approaches to model complex systems. Feadback loops are, e.g., at the core of cybernetics.
• Maybe the only difference (innovation) is the application domain?
• Feadback loops do not necessarily imply nonlinearity!!!

Note by 141.35.14.148 on 2 Dec 2005

My feeling is that this is really a bunch of people using the term system dynamics to describe population dynamics and try to popularise it using non-free software packages (but with the best of educational intentions). i was invited to a course on the topic, and the convenors didn't seem to realise that in the internet era, we as a class could form a self-sustaining dynamical system...

i don't see how a positive feedback loop can not imply nonlinearity (if there's nothing else in the system). ${\frac {dx}{dt}}=Ax$ has the solution $x=e^{At}$ which is definitely not linear.

Boud 14:55, 7 August 2006 (UTC)

But ${\frac {dx}{dt}}=Ax$ is linear in the sense that the differentials can be expressed as a linear combination of the variables, eg. in the general case ${\frac {dx}{dt}}=A(x)$, where $A(x)$ might be any arbitrary old expression. I think that's the sense implied above. —Preceding unsigned comment added by 85.228.100.47 (talk) 17:27, 19 August 2008 (UTC)

You make a good point about the similarities between system dynamics and cybernetics. In fact, Jay Forrester was very well versed in control system theory, and noted the similarities with management policies. Hi work on industrial dynamics was the application of control system theory to problems in the domain of management. His second book was Urban dynamics, which has some similarity with population dynamics.

Since then, system dynamics has found appliation in many other fields.

I think the article can benefit from some examples in other fields. For example, the California management review of Summer 2001 (Vol 43 no 4) contains some example in other fields, including fire-fighting; process improvement programs and service and quality drift.

Apdevries 10:40, 1 October 2006 (UTC)

Sierrasedge (talk) 13:01, 8 September 2008 (UTC)

The example given is very instructive, although, the initial conditions need to be explicitly stated. I was able to reproduce the results very closely using Excel and although it is not free either the point here is that simulation software is not necessarily needed to implement system dynamics based models. I set the initial conditions at time = 0, to the following:

        Potential Adopters = 1000
        Adopters = New Adopters = 0
        Innovators (Early Adopters) = 30
        Imitators = 0

The published simulation results seemed to have plotted the results incorrectly, causing the graph of Imitators and Innovators to appear to start at 30 at time = 0, when in actuality their intersection at 30 occurs a couple of timesteps later.

It would be nice if another example could be added and a comparison to real world reference data provided to open the discussion on calibration of the model. I will try to supply an example modeling defects in a generic software development lifecycle.

Non-free is pretentious

Unless all software referenced on Wikipedia specifies its license, then I think noting "Non-free" is insulting, and often misleading, epecially if the software is of zero-cost. Don't get me wrong, I use a great deal of "free software," but for a general publication such as an encyclopedia, it's incredibly slanted to judge such matters in a limited context, when no licensing is called for.

I propose, the licenses be only in the fact tables of the software entries, unless it is specificially relevant to the grandparent article. Politics aside, it's poor organization to specify trivia about a link outside of a link.

Comment by 24.130.226.184 on 20 August 2006

I agree, and have culled the list of software. Apdevries 10:19, 1 October 2006 (UTC)

I must recall that wikipedia isn't just another encyclopedia, its "The Free Encyclopedia" as you can read at the upper left corner of your screen, so I think that readers of a Free Encyclopedia would like to know which of the software listed is free and which no. Even so, as 24.130.226.184 says I don't think it should be added here, please see [Comparison of wiki software] and lets start a new page that compares functionality an licences for each System Dynamics software, say [Comparison of System Dynamics Software] ?

Apdevries, I really like the new looks of the page, thanks, but why did you took most of the software cited? the only thing you should have done is take the licence statement. Some of the software previously cited is actively used in the industry in order to simulate engineering processes, it is cataloged as control systems but at the end they all apply System Dynamics theory, so shouldn't this be mentioned to in the article?

I'll wait your response to reload the missing software, and possibly start the comparison page. Best regards WiKimik 03:23, 19 October 2006 (UTC)

Equations need to be reviewed

The current equation for Potential Adopters is such that Potential Adopters = - Adopters hence the probability that contact has not yet adopted is infinite. I presume that the correct formulae would be Potential Adopters = Initial # For Potential Adopters + the integral. —The preceding unsigned comment was added by Ahmedbax (talk • contribs) 23:46, 18 January 2007 (UTC).