I'm in the early stages of getting a large model working.
The model fails before all variables have been calculated, because of the order in which things are computed.
A value that is a denominator is not calculated before the variable in which it is used, and so I get a floating point error. Is there any way to influence the order of computation?
thanks
Computational order error
-
- Junior Member
- Posts: 5
- Joined: Wed Jun 05, 2019 3:39 pm
- Vensim version: DSS
-
- Super Administrator
- Posts: 4590
- Joined: Wed Mar 05, 2003 3:10 am
Re: Computational order error
Can you send the model? It should not be possible for something like this to happen (I don't recall it ever happening).
Advice to posters seeking help (it really helps us to help you)
http://www.ventanasystems.co.uk/forum/v ... f=2&t=4391
Units are important!
http://www.bbc.co.uk/news/magazine-27509559
http://www.ventanasystems.co.uk/forum/v ... f=2&t=4391
Units are important!
http://www.bbc.co.uk/news/magazine-27509559
-
- Junior Member
- Posts: 5
- Joined: Wed Jun 05, 2019 3:39 pm
- Vensim version: DSS
Re: Computational order error
Thanks for the response.
My mistake. Have managed to amend several equations that were failing, by looking at the limited data that was available before the failure. The model now calculates every variable.
Perhaps would be useful to put into the help somewhere? Maybe I'm the only one who has thought this (!), but it looked like the computation order was contributing to the floating point error.
My mistake. Have managed to amend several equations that were failing, by looking at the limited data that was available before the failure. The model now calculates every variable.
Perhaps would be useful to put into the help somewhere? Maybe I'm the only one who has thought this (!), but it looked like the computation order was contributing to the floating point error.