Working version of DRAGG-Gym

This is the initial working version of DRAGG-Gym. Using the RL agent contained in nonoise10.zip the RL agent improves the score of the aggregator agent over the baseline and TOU cases.

This version requires some general cleaning before it should be officially released. To do before officially released: decouple the DRAGG and DRAGG-Gym packages, pull out DRAGG-Gym parameters such as the reward function and state variables to make modifications easier.