In your Pathmind reward function editor, you can dynamically pick and choose which observations to use for training. This is helpful for rapidly testing various observations without the need to go back and forth between AnyLogic and Pathmind.
Step 1 - Configure individual variables or arrays of observations in Pathmind Helper.
Step 2 - In the Pathmind reward function editor, you will see a section called "Observations". This allows you to check and uncheck specific observations to use for training.
Step 3 - Click "Train Policy" to start training.