A hybrid architecture for hierarchical reinforcement learning | IEEE Conference Publication | IEEE Xplore