Fairness Aware Reinforcement Learning via Proximal Policy Optimization