An Optimal Control-Based Distributed Reinforcement Learning Framework for a Class of Non-Convex Objective Functionals of the Multi-Agent Network | IEEE Journals & Magazine | IEEE Xplore