Policy Evaluation In Reinforcement Learning