bguan's lunar lander model #3 using PPO trained for 1M timesteps

bguan's lunar lander model #2 using PPO trained for 500K timesteps

bguan's lunar lander model using PPO trained for 500K timesteps

initial commit