Request of SD advice - Ph.D Thesis results

[Nub_User]

I want to make an academic inquiry regarding SD.

Currently, I'm finalizing the research for my Ph.D., where I employ SD as a modeling tool for a complex system. Essentially, without delving into specific or thematic details, in my thesis, I develop, based on literature review and expert interviews, a simulation model whose results I analyze according to the relevant state of the art. There, I examine if the obtained results align with the main current theories and utilize the model.

The consult I want to make as experts in SD are:

1. What details about modeling through SD would you expect to find in a thesis of this type? (please exclude academic requirements such as methodological details, references, citations, source usage, etc.)
I refer to what details you would expect to find (causal model, Forrester model, list of variables, formal mathematical equations used, equations used in simulation software, model testing, validation, others) and where you would expect them to be presented (in the main body, appendices, annexes, others) and in what form (graphics, tables, writing, others).

2. What information about the model can I not forget/omit under any circumstances about the model?

3. What kind of discussions about the results would you expect to read given the SD model?

4. Any recommendations that you believe could be pertinent or valuable.

I want to clarify that I have read some Ph.D. theses that employ SD as a modeling tool for various topics, and they are all different in terms of what they present and how they present it regarding the developed model.

Thank you very much in advance.

[Administrator]

There are two things I do not like to see when reading papers.

1. Models that have unit errors or missing units. I immediately assume the model is full of errors. If the model cannot past this basic check, then it suggests to me that the author has not taken the time to make sure things are correct elsewhere.

2. Equations in the paper that do not match what is in the model. For example, a model might have "cumulative costs = INTEG (monthly costs, 0)", but in the text it's "CC = sum (MC)". I hope that makes sense.

The big one for me is units. I see it regularly where someone is looking at results from a model that has errors in it.

[LAUJJL]

Hi

I would add: units with 'use strict testing' on, in the model settings.

Regards.

JJ

[Nub_User]

Hi

I would add: units with 'use strict testing' on, in the model settings.

Regards.

JJ

Thank you very much for your advices! Please could you explain to me about the "use strict testing"?

[LAUJJL]

Hi
Sometimes but not always, Vensim if strict unit is not set, does not consider that adding a dimensionless variableor constant to a variable with dimension is not and error, guessing probably that the constant or variable without a dimension has the same dimension than the variable with a dimension.
This unit error can hide an equation error.
i explained this in the discuss forum of the SDS and in the Vensim forum at:
Re: two models explaining how to correct some errors found in the world3 model
You can download the model that I have joined.
In that model the first page is an ageing chain. The first rate equation called maturation is dimensionally correct but not the other ones. If the strict testing is not set, Vensim will not detect any error, if it is set, the two rates that calculate the maturation from adults to older people and older to very old people will show an error. This error exists too in the world3 model. In fact this error is only significative in extreme cases
And does not probably change the overall results of the models.
But it shows from my point of view that the people who changed the world2 to the world3, did not care very much about avoiding possible errors. The errors were caused from changing the time step from one year to a fraction of year.
I have sometimes found some errors in my models that would not have been found with the strict unit not set.
By the way hundreds of SD people have read my explanation in the SD discuss forum and about 80 people have downloaded the models explaining how to correct the errors. It has not triggered the slightest comment or question. I have never found any explanation to that fact. But whatever the explanation, the facts being there, I am no more member of the SDS, having a completely different SD conception.
Regards.
JJ

[Nub_User]

Thank you very much for the detailed explanation and the model @LAUJJL. I understand the point.