Multi-Objective Exploration for Proximal Policy Optimization