Structured-policy Q-learning: an LMI-based Design Strategy for Distributed Reinforcement Learning | IEEE Conference Publication | IEEE Xplore