The cartpole system control implementation and testing was done in simulation, using the environments available through OpenGymAI as the main testbed frameworks.  OpenGymAI is an open source python platform dedicated to testing various reinforcement learning techniques to a variety of problems; the cart-pole was  readily available testing environment [5], although minor modifications were made to make it more acceptable for our purposes.

The acrobot system control implementation and testing was done in a similar environmental setup as the cartpole system.  The acrobot is also an environment available through OpenGymAI.  The standard layout of the problem involves applying a unit torque of +1, -1, or 0 to the joint of directions in an attempt to swing the pole up.  This simulation relies on dynamics derived by Dr. R Sutton in Reinforcement Learning, an Introduction[].  The objective for the controllers were to swing the first link up to vertical by applying torque to the hing link.