Multi-Agent Deep Reinforcement Learning-Based Power Control and Resource Allocation for D2D Communications | IEEE Journals & Magazine | IEEE Xplore