A collections of solutions for openai gym environments, using mainly reinforcement learning techniques.
As of 12/10/2018 11:30PM MDT, currently #1 on the leaderboard
A Sarsa(λ) Agent with eligibility trace and linear function approximation.
Generating feature vectors was done with the tiles3 library.
The bulk of the environment code is from the RandomAgent example on the OpenAI github.
Modified to run the agent as necessary.
Max Episode = 1000
Required Avg score over 100 consecutive episodes = -110.00
Achieved Avg score over 100 consecutive episodes = -109.41
Number of Episodes for solution = 341
