Policy Gradient Methods to Actor Critic Model