Discount Factor In Reinforcement Learning