Page 1 of 1

Static vs. dynamic submodels

Posted: Mon Jun 02, 2003 12:54 pm
by Salima
When you model a problem with SD, is it preferable to start modeling the submodel that is "static" or rather start with the "dynamic " submodel.

I mean by "static" the submodel where the equations and relationships between variables are clear mathematically (example : financial variables where the equations are already defined ).

By "dynamic" I mean the submodel that needs some investigations on the nature of relationships that may be complex.

Thanks in advance, Salima

Posted: Wed Jun 04, 2003 2:33 pm
by Administrator
Hello,

I am not entirely sure what you mean by submodel.

Can you give a real example?

Tony.

Posted: Fri Jun 06, 2003 11:41 am
by Lee Jones
If I understand you fully, then the "sub model", as you put it, with fully understandable relationships can, in my opinion, wait while the interesting, underlying dynamics are much more important to get right.

That said, many financial outputs (where relationships are known and agreed) can provide useful calibration comparisons, so including these as the dynamic (interesting) part of the model develops could be useful.

I would suggest concentration on the "dynamic sub model" should be paramount while known "static?" relationships can be added easily at times where their presence would be benefitial in the understanding, reporting or calibration.

Posted: Sun Jun 08, 2003 10:01 pm
by Salima
Thanks Tony and Lee for your answers.

I meant by submodel, a sector or subsystem of the whole model.

A "static" submodel would be, for example, an accounting/bookings submodel whom purpose is not to model behavior, but to report transactions.

These transactions are caused by variables that do not belong to the accounting/bookings submodel, but rather are caused by some other
"dynamic" submodel.

I understand your suggestion Lee. I asked the question because I wanted to know which path is faster.

Thanks a lot, Salima