Page 1 of 1

Sensitivity distributions

Posted: Wed Sep 09, 2009 10:13 pm
by mikdale
Hi all

I'm having a bit of a strange problem.

I have run some sensitivity testing on my model using normal distributions, got it all running very well.

I have since come back and changed a couple of the variable distributions and the sensitivity confidence bounds are now MUCH larger than before. 'That's strange!' I thought and set about trying to figure out what was causing it by systematically setting the standard deviation of each of the variable distributions to zero.

I have now set all to zero (so each variable should only take on the mean value, right?) but the confidence bounds are still as large as ever (i.e. bigger than they were initially)

This seems a bit weird to me. Does anyone have any idea what might be going on? Am I being a bit dense?

Cheers

Mik

Posted: Wed Sep 09, 2009 10:44 pm
by mikdale
Hi again

Don't worry, it seems to be sorted now, not sure what was going on though - bit weird

Cheers

Mik

Posted: Thu Sep 10, 2009 10:40 am
by bob@vensim.com
A bit of a mystery. The parameters being varied are all saved, so one thing to do is to make a bar graph for each of those to see if it offers any clues as to what is happening.