On the Global Convergence of Policy Gradient in Average Reward Markov Decision Processes

Open in new window