Hi,
I've got a question regarding the length of my vensim simulation.
I would like to stop the model-simulation at the time when a variable used in the model reaches a specific value.
Is that possible? Is there an equation for a FINAL TIME which depends on a variable used in the model?
In case this doesn't work, would it be possible to type in a similar equation for the end time of the time axis?
Thanks for any answers, I appreciate it!
Cheers,
Seb
FINAL TIME dependent on a variable?
-
- Senior Member
- Posts: 1107
- Joined: Wed Mar 12, 2003 2:46 pm
If I wanted to have a variable lenthf of simulation, I would prefer to test the possibility of setting a final time sufficient large and simulating up to that final time, and taking into account only the results up to the real length of time wanted that can be named differently. Making the final time vary may have subtle implications, for example if you optimize, or performing sensibility analysis on cumulated pay offs. It is always better to my opinion to choose simpler methods when possible.
SD is sufficiently complicated by itself!
JJ
SD is sufficiently complicated by itself!
JJ
-
- Senior Member
- Posts: 1107
- Joined: Wed Mar 12, 2003 2:46 pm
In practice I agree with JJ - just set a reasonable time and look at that. The difference in the amount of time it takes to simulate is almost always trivial and varying lengths can, as JJ points out mess up sensitivity and optimization.
That said, Richardson and Pugh in their classic Introduction the System Dynamics with Dynamo, do use a variable final time, so you would at least be in good company.
That said, Richardson and Pugh in their classic Introduction the System Dynamics with Dynamo, do use a variable final time, so you would at least be in good company.