A Novel Behavioral Strategy for RoboCode Platform Based on Deep Q-Learning

Figure 6

Percentage of actions selected during model training in combat with the Robot “SittingDuck.”