Beta Policy

Reimplement the paper: Improving Stochastic Policy Gradients in Continuous Control with Deep Reinforcement Learning using the Beta Distribution on MountainCarContinuous-v0

Note that the model is not very stable, if you find any tricks that can make it stable, please keep me informed.

Reference: https://github.com/dennybritz/reinforcement-learning/blob/master/PolicyGradient/Continuous%20MountainCar%20Actor%20Critic%20Solution.ipynb