Reward Function Shaping of an Autonomous Reinforcement Learning Agent