System Dynamics definition

[JerzyW]

Why did you redefine the hallmark of System Dynamics?
From: Stock and Flow diagrams
To: Level and Rate diagrams

[Administrator]

We didn't redefine anything. I don't think I've ever heard anyone referring to a "level and rate diagram".

A stock is the same ss a level, and a flow is the same as a rate. Sometimes the term "level" makes more sense than "stock" (eg, water in a reservoir), but they are still the same thing.

[JerzyW]

The Vensim help files are telling different story. Stock and Flow are not even mentioned in the Glossary of Terms but Levels and Rates are everywhere. Stock and Flow have much broader meaning then Levels and Rates and they are correct.
Jay Forrester must be turning in his grave at the thought of what is being perpetrated to the System Dynamics.

[Administrator]

Jay Forrester must be turning in his grave at the thought of what is being perpetrated to the System Dynamics.

Some of the staff at Ventana were close friends with Jay and worked alongside him right up until he passed away. They are still close to his family. Statements like this can easily be considered an insult.

And yes, stock and flow diagrams are mentioned in the help system and are easy to find.
https://www.vensim.com/documentation/usr05.html

[LAUJJL]

Hi 'everybody'

I have never been bothered by stocks or levels or flow or rates. Anyhow, whatever word you use, it will be understood differently by people.
Using two different names for the same thing is not necessary a bad idea. For some people a rate is more suggesting than a flow, same for stocks and levels. Such a controversy will not help the field.

Jay Forrester was more concerned by the general lack of model's quality generating a poor impact than by rates or flows.
A more interesting question to ask is: since Jay left us, did the general quality progress?

Regards to ‘everybody’.

JJ

[JerzyW]

I know Jay Forrester as well, from videos and books. My impression of him is that he was very strict about System Dynamics implementation.

For example: "flow" is always function of time: liters/hour, but "rate" could be anything: people/square_meters.

Anyhow, I misread the whole thing so the best course of action is to deleted this chat from the server.

Regards,
JerzyW

[LAUJJL]

Hi

It is not necessary to delete anything. All opinions are welcome.

"flow" is always function of time: liters/hour, but "rate" could be anything: people/square meters.

Flow is not necessarily a function of time and can be a constant. Flow must be divided by some time value. It is the value of something during a time period; the fact of dividing by a time period does not make it function of time.

About the “rate”. The rate must be too, like the flows divided by a time period and represent a unit during a time period. Then rate could be anything: people/square meters is wrong. If you use this rate with a stock, it will generate a unit error, unless you divide the rate by some time period inside the stock equation, something that is not recommended by good practice.

About Jay being strict about implementation, you should explain how Jay was strict about implementation.

Can you give me an example of Jay having ever implemented a model with success?
Jay worked with G.E. but never mentioned it in ‘Industrial Dynamics’. It looks like G.E. never implemented Jay’s model, or the implementation did not produce the expected results.

About Urban dynamics, the ex-Boston’s Mayor, had not the opportunity to implement anything, and the members of the Club of Rome had not the authority to implement the world model, and I doubt that the newer versions are more successful.

I do not know if Jay ever implemented any model successfully.

But it does not make his teaching bad. And he is not alone in the SD field. Implementation seems to be very rare in SD. You never see any implementation reported during the SDS conferences.

Best regards.

JJ

[JerzyW]

Hi,

I have never met Jay personally but I have soft spot in my heart for inventors like him. I can imagine that at the end of his days he realized what a "monster" called System Dynamics he created. "Monster" that on the surface appears so simple that is being introduced at elementary schools but in practice the development of correct SD model requires advanced knowledge of Control Theory with hands-on experience at the engineering level.

Yes, he did not achieved his goal of developing World Model but I bet that Artificial Intelligence based on Large Language Models will fail to achieve this goal as well.

I can sign off on the last paragraph in your email (you nailed it):

"But it does not make his teaching bad. And he is not alone in the SD field. Implementation seems to be very rare in SD. You never see any implementation reported during the SDS conferences."

Give it a last shot (if you don't mind) and let's close this thread.

Regards,
JerzyW

[LAUJJL]

Hi

I think that advanced knowledge like control theory is more harmful than helpful because it may lead to think that the more math the better. I think that the less math the better.

JJ

[LAUJJL]

The less math and more life experience.

[rdudley]

Just for clarity. Forester used the term 'Level' in much of his work. See Industrial Dynamics (1961) and Principles of Systems (1971).

IMG_4651.JPG (1002.92 KiB) Viewed 33 times