Deep Reinforcement Learning scripts working with ROS+Gazebo+"Gym like wrappers"

First algorithm will be DDPG working on an "inverted pendulum" environment.