Optimization-driven Hierarchical Deep Reinforcement Learning for Hybrid Relaying Communications | IEEE Conference Publication | IEEE Xplore