Hi
first of all season's greetings to everybody....
then my question: is there any way in Vensim PLE plus to perform a Monte Carlo analysis with a randomized FINAL TIME? I trid to insert the FINAL TIME variable among the others during the Sensitivity Simulation Setup as RANDOM UNIFORM 8 12 with a current value of 12 but when the MC is run I get a lot of errors of this kind
ERROR: Length or Saveper changing during sensitivity-some results lost..
so how much can I be confident about the resuluts I obtain in the sensitivity graphs? Is there any other way to proceed?
Best regards and best wishes
Lorenzo at work on Xmas morning
Random FINAL TIME
-
- Senior Member
- Posts: 102
- Joined: Wed Jul 30, 2014 7:09 am
- Vensim version: PLE
Re: Random FINAL TIME
The issue is that storage for results is allocated at the time of the first (ordinary) run, so if FINAL TIME is longer than that in a sensitivity instance, they get truncated.
As a workaround, you could set things up such that an ordinary simulation is longer than your sensitivity range. You'd still get the warning, but all the data would be there. Extracting the results would be a little tricky though.
As a workaround, you could set things up such that an ordinary simulation is longer than your sensitivity range. You'd still get the warning, but all the data would be there. Extracting the results would be a little tricky though.
/*
Advice to posters (it really helps us to help you)
http://www.ventanasystems.co.uk/forum/v ... f=2&t=4391
Blog: http://blog.metasd.com
Model library: http://models.metasd.com
Bookmarks: http://delicious.com/tomfid/SystemDynamics
*/
Advice to posters (it really helps us to help you)
http://www.ventanasystems.co.uk/forum/v ... f=2&t=4391
Blog: http://blog.metasd.com
Model library: http://models.metasd.com
Bookmarks: http://delicious.com/tomfid/SystemDynamics
*/
-
- Senior Member
- Posts: 102
- Joined: Wed Jul 30, 2014 7:09 am
- Vensim version: PLE
Re: Random FINAL TIME
Hitomfid wrote:The issue is that storage for results is allocated at the time of the first (ordinary) run, so if FINAL TIME is longer than that in a sensitivity instance, they get truncated.
As a workaround, you could set things up such that an ordinary simulation is longer than your sensitivity range. You'd still get the warning, but all the data would be there. Extracting the results would be a little tricky though.
thank you a lot... so I fix (for instance) statically FINAL TIME at x (in the model with the equation editorl) and then during the Monte Carlo set up phase define it as (again for instance) RANDOM WEIBULL(0,x,2,0,1). If this is the procedure I don't see why collecting the rusults should be tricky: I run the simulation and the either visualize the sensitivity graphs for the interesting variables or I export the results of the simulations as a tab delimited file...
Agian thank you a lot, best wishes
Lorenzo