It's really a system with only one enter, scenario, and only one output, action (or behavior) a. You can find neither a different reinforcement enter nor an advice input in the surroundings. The backpropagated worth (secondary reinforcement) would be the emotion toward the consequence circums