Page 1 of 1

Vensim constant behavior

Posted: Tue Apr 17, 2018 10:43 am
by anca91
Hello everyone, sorry if this topic could be banal. I have the following model and I try to verify his behavior. Unfortunately, when I start the simulation I get a constant trend, when I expect a decrease in the number of uninfected and an increase in the number of those infected. You could kindly help me. :oops:

Re: Vensim constant behavior

Posted: Tue Apr 17, 2018 10:56 am
by Administrator
Try changing
rate of plant infections = Plant Infected/LATENCY TIME
to
rate of plant infections = Plant Not Infected/LATENCY TIME

At t = 0, the number of plants infected is zero. So rate of plant infections is also zero.

Re: Vensim constant behavior

Posted: Tue Apr 17, 2018 12:47 pm
by anca91
thanks for the reply, I tried to do what you say. but I get this answer: :cry:

ERROR: Floating point error computing - rate of plant infections - at time = 34.000000.
Trying to save the results anyway

Re: Vensim constant behavior

Posted: Tue Apr 17, 2018 12:54 pm
by Administrator
Have a look at a table of the values for rate of plant infections and see what is happening.

Re: Vensim constant behavior

Posted: Tue Apr 17, 2018 1:06 pm
by anca91
I can not find a meaning

Re: Vensim constant behavior

Posted: Tue Apr 17, 2018 1:27 pm
by Administrator
One problem is that your model is running for 100 years with a time step of one year. This needs to change, the time step should be 1/4 of the smallest delay in your model (I think this is 0.08333, the "latency time").

What is the "latency time"? I know it's value is one month, but what actually is it?

Re: Vensim constant behavior

Posted: Tue Apr 17, 2018 1:41 pm
by anca91
It is the time it takes to understand how much time there is between when a plant shows the disease and when it takes the infection. :roll:

Re: Vensim constant behavior

Posted: Tue Apr 17, 2018 11:35 pm
by tomfid
The model will not generally be stable unless the time step is shorter than the shortest delay. Powers of 2 are generally better, so perhaps .03125 would be a good start (you can set this via the dropdown menu at Model>Settings>Time Bounds.