Lane-Merging Using Policy-based Reinforcement Learning and Post-Optimization | IEEE Conference Publication | IEEE Xplore