After create the script is time to configure it from the Inspector in the Unity IDE. We need to associate the necessary elements like the target, and modify the MaxStep variable. In this variable we indicate the steps that the agent must do before restart the environment calling the OnEpisodeBegin function.
We, also, add the component Behaviour Parameter, it’s necessary to configure the communication between our agent and the Machine Learning Agent.
At the end we add a new componet: Decision Requester . It’s a really simple to configure component, but totally essential, without him our agent would remain frozen. We can indicate a number of steps required to make a decision. In the video we indicated 5, it means that the Machine Learning engine will do at least 5 steps before to send a new action to the agent.