Hi
I would like to find some ways to easily build a report to be printed wit all the constants, auxiliaries, data, rates and levels of a model but in an order that permits easily the verification. any value being reported should be after any value that influences that value. Presently I build the report by looking at the sketch and starting with the values that are computed first and finding the next ones using the causal links. It is rather tedious and finding some ways that lists the variable in the order of calculation would ease the process, in particular when one has to rebuild the report as the model is extended progressively.
Regards.
JJ
ordering results on a report for verification of a model
-
bob@vensim.com
- Senior Member
- Posts: 1107
- Joined: Wed Mar 12, 2003 2:46 pm
Hi JJ,
There is, to some extent, a defined order of computation for auxiliary variables. This gets messed up if you use FIND ZERO or the SIMULTANEOUS function or if you have equations that require the unraveling of subscript relationships. It can also be different between the initial and active computations. Still, I can see that it would be useful to add an ordering option to the Document tool that would represent that sequence when possible.
You can, in the document tool, ask for a variable and its causes (and even the values for these) to be shown.
There is, to some extent, a defined order of computation for auxiliary variables. This gets messed up if you use FIND ZERO or the SIMULTANEOUS function or if you have equations that require the unraveling of subscript relationships. It can also be different between the initial and active computations. Still, I can see that it would be useful to add an ordering option to the Document tool that would represent that sequence when possible.
You can, in the document tool, ask for a variable and its causes (and even the values for these) to be shown.
ordering of computational order of elements of a report
Hi Bob
Thank you for your answer.
I avoid using find zero or simultaneous functions, as I do not control how they work, and I prefer to simulate while loops using subscripts. The values inside the loop are affected to a different value of subscript each time the loop progresses.
In this case it is still possible to order the calculations.
I will use my standard method, building the report, using the sketch and starting with elements with no influence firts and progressing towards the elements they influence.
Il would be convenient to have a report that builds automatically a report of all the elements correctly ordered.
It would too ease and promote the verification of how models calculate the values of the variables, which I am not sure that the majority of modelers do.
Best regards.
JJ
Thank you for your answer.
I avoid using find zero or simultaneous functions, as I do not control how they work, and I prefer to simulate while loops using subscripts. The values inside the loop are affected to a different value of subscript each time the loop progresses.
In this case it is still possible to order the calculations.
I will use my standard method, building the report, using the sketch and starting with elements with no influence firts and progressing towards the elements they influence.
Il would be convenient to have a report that builds automatically a report of all the elements correctly ordered.
It would too ease and promote the verification of how models calculate the values of the variables, which I am not sure that the majority of modelers do.
Best regards.
JJ