Reinforcement Learning in FlexSim v 23.1

[MarisolVera]

[ FlexSim 23.1.1 ]

Hello, I'm trying to follows the steps of reinforcement learning training ( https://docs.flexsim.com/en/22.1/ModelLogic/ReinforcementLearning/Training/Training.html)

On the version 22.1 I don't have issues, but when i try to follow the same steps on v 23.1 it doesn't work. On the "testing environment connection" step i got the next error when i try to run file flexsim_env.py :

I'm using the model provided for guide (ChangeoverTimesRL).

I also notice that you don't have documentation about this for v 23.1. Are you still supporting this feature ?

[Matthew_Gillespie]

The documentation for 23.1 has a slightly different path because it was split into two options:

Reinforcement Learning with OpenAI Gym

and

Reinforcement Learning with Bonsai (Note that Microsoft is ending support for Bonsai so we don't recommend starting to use it now)

[JordanLJohnson · Accepted solution]

Be sure you download and use the python scripts that correspond to the version of FlexSim you are using. In version 22.1, you need to download the scripts from the 22.1 version of the manual. For 23.1, you need to download the scripts from the 23.1 page in the manual:

https://docs.flexsim.com/en/23.1/ModelLogic/ReinforcementLearning/WorkingWithGym/GettingStarted/Gett...

Also be sure you are working with python version 3.8 or 3.9.

If that doesn't solve your issue, you'll need to post your scripts and model so we can investigate further.