Problems with Sample Yeast Culturing Model?

[egrdavie]

I am using a sample Vensim model of yeast growth and decline in my graduate course on systems modelling. The model used to work quite well, but without any changes to the model except for an update to Vensim PLE version 6.1, the behaviour now seems very sensitive to small changes in the temperature. For example, with a normal culturing temperature of 29 degrees, and a "dormancy temperature" of 10 degrees, I would expect to find yeast growing almost normally at 22 degrees. Instead, the yeast count grows only very slowly, reaching a maximum of 1600 cells at 300 min. Under normal conditions, the cell count grows and then collapses, with a maximum cell count of ~60000.

Can you tell me why the model is so sensitive now? (I can change the lookup tables to fix the problem, but I am wondering specifically whether anything changed in Vensim from version 5.9 to 6.1 that might have affected the model behaviour.)

[Administrator]

No, nothing has changed.

Can you post the model so we can take a look?

[tomfid]

To debug, I think you need to locate what changes first. An easy way to do this is:

- right-click the Runs Compare tool and select Levels and Auxiliaries (in addition to the constants and data, which should already be selected)
- load a v6.x and v5.9 run that differ
- click the Runs Compare tool, and start causal tracing the first few differences that show up

It's also convenient (if you haven't already done it) to change the settings of the table tool to show Causes, so that you can easily see a variable and its inputs for debugging.

[egrdavie]

My apologies for the delay in responding -- I had not checked for responses since posting my question.

I followed tomfid's suggestions, and it seems the model was always quite sensitive (I found a simulation file from 2011, and results were the same as the current results). Sorry for wasting your time, and thanks for answering so quickly.

[tomfid]

If you have the time and inclination, I'd be interested to hear about your experiments with the model. Have you identified room for improvement?