Parallel Reinforcement Learning With Minimal Communication Overhead for IoT Environments | IEEE Journals & Magazine | IEEE Xplore