previous post in the SDS forum

Use this forum to discuss any issues relating to Systems Dynamics and Systems Thinking.
Post Reply
LAUJJL
Senior Member
Posts: 1421
Joined: Fri May 23, 2003 10:09 am
Vensim version: DSS

previous post in the SDS forum

Post by LAUJJL »

Hi everybody

I post here a previous post in the SDS forum, from where I had no answer.

My question was probably not well formulated. I find the two questions fundamental.

One is about two options: build a simplistic and eventually unrealistic depending on the parameter values used or other circumstances or building a more realistic but much more complex and costly model. Which solution to chose and in which circumstances? Or maybe both at the same time or one after the other?

The other is about the boundaries of a model. Is it yes or no and in which circumstances necessary to include in a model, the decision system that governs the decision taken about using or not the policies and insights of the model and how the model will be studied and analyzed. In fact describing how the client takes his decision, works etc..

I know the question is not simple but it seems to me to explain most of the model's failures that I have seen these last years and the reason of the relatively poor use of SD.

In the recent post of Roberta Spencer there were recommendations from Tom Fiddaman about a good conference paper. It gives nearly infinite opportunities to improve one's own models even if not published.
I have noticed two of them in 'writing a good system dynamics paper II'

<If all of your graphs show constant values, linear increases (ramps), or exponentials, my eyes glaze over, unless you can make a <compelling case that your model world is really that simple, or that people fail to appreciate the implications of those behaviors.

In fact the question is: is it absolutely necessary that the model be realistic to be useful something that is maybe meant with 'or that people fail to appreciate the implications of those behaviors'.

I join a simplistic model without inflation and economic growth that is obviously unrealistic about the national debt of a country (the parameters are for France starting in 2015). It is unrealistic with some parameters and after some years, because one can hardly imagine anybody lending money to such a country in this case. The debt should too influence the interest rate which would make the behavior still worse in the case of a growth of the debt. If one adds a lookup effect of debt on interest rate the debt is literally exploding and becomes unsustainable after already 25 years.

But it is still interesting because it gives an idea about how unsustainable can become a debt with the time if nothing is done about it and that something bad should happen in that case even if nobody knows exactly what.

Another question is about:

<Discuss what lies beyond the boundary of your model. What did you leave out and why? How does this limit the applicability of the <results?

The objective of a model unless used for teaching, is to help understand and hopefully improve the behavior of a system. But to do that, if the model can generate useful insights, the decision preconised must be decided and implemented. One should then logically include in the model the way that decisions described in it are taken in the real world unless one takes the risk to see after having studied for one year all kinds of things, Mister X an influential decision maker leaving a committee meeting at 4 o'clock because his plane takes of at 6, ruining one year of effort. Something that I have seen many times even if not exactly like the situation described precedently.

The question is: where should be put the boundaries. Is it the responsibility of the modeler at least to mention the boundaries and the influences they can have on the final results of the modeling job; eventully proposing different offers, with more or less wide boundaries?

I join two models that can be read either with the Vensim PLE or the Vensim reader.

It is possible to transform an unrealistic behaviour into a more realistic one, applying a specific policy. For instance the behaviour becomes realistic and sustainable if the primary deficit ramps from 27.4 to a benefit of 80 during 30 years. The total debt becomes stable.

The stabilizing policy is in the stabilized.vdf run joined and compressed into a zip archive file to make it possible to upolad it.
I do not understand the extensions restrictions to upload a file. I tried too to change the extensions name which seems to be no more possible. Life could be easier!

Best regards.

JJ
Attachments
stabilize.zip
(2.45 KiB) Downloaded 409 times
exponential growth.mdl
(2.13 KiB) Downloaded 409 times
exponential growth.vpm
(15 KiB) Downloaded 406 times
Post Reply