Fatal error when adding policies as an int

[lvanopstal]

Hello,

Currently I am coupling a Vensim model to the EMA workbench to do some analyses for my master thesis. I am running into an error and was wondering if this is the right place to ask what might be going wrong. After having defined uncertainties and outcomes and running the "evaluator.perform_experiments(1000, policies=5)" I get the following error (see attached picture).

EMAError: some fatal error has occurred while running the experiments, not all runs have completed. expected 5000, got 0

I tried diving into the code snippet where the error gets thrown but couldn't figure out what I am doing wrong here. What could the problem be?

Thanks in advance!

[quaquel]

Impossible to answer based on the code shown. The error is a catchall for fatal errors. I need to see the complete code to judge this.

Also, have you checked https://emaworkbench.readthedocs.io/en/latest/best_practices.html and made sure to apply this?

[lvanopstal]

I am currently running the code through jupyter notebooks. Even without the multiprocessing I get this error. The project directory looks similar as proposed in the link you provided. I hope this clarifies some things.

[quaquel]

A few remarks

1. why are you sampling 5 policies while you have not defined any levers? Most likely, this is the root of your error.
2. the r before the strings of wd and model_file is unnecessary.
3. parallel processing should work even in jupyter, but I would still strongly advise against performing experiments in a notebook.

[lvanopstal]

Thanks for your reply. I looked into the mentioned points and this resolved the problem. Thank you!