Optimization Issues in KL-Constrained Approximate Policy Iteration

Many reinforcement learning algorithms can be seen