It is important to make sure that the Pathmind Helper is correctly installed and set up before exporting your model. Performing a run with random actions will allow you to make sure the Pathmind Helper is correctly configured while also demonstrating how a model performs without a reinforcement learning policy.

Open the Pathmind Helper properties and double check the following items:

  1. Make sure that Pathmind is enabled.
  2. Turn on debug mode.
  3. Set "Mode" to "Use Random Actions".

Now run the simulation. Once it's started, open the Developer Panel. If the Pathmind Helper is correctly installed and set up, you will see data printed.

You may inspect this data to confirm that Pathmind is receiving the correct information.

Pathmind:
Step: 15
Observations:
Array: [0.0, 0.0, 0.0, 1.0, 0.0]
Total Observations: 5
Reward Variables:
Array: [1.0, 0.0, 0.0, 0.0, 0.0, 0.0]
Total Rewards: 6
Last Action: [3, 2]
Done: false

Once the run is completed, observe how the model performed. The results are likely to be poor since random actions are being used, but can serve as an interesting baseline for comparison once a policy is in place.

Did this answer your question?