Reinforcement Learning Project

Apply reinforcement learning methods to solve a problem of your choice. You may choose any
problem that engages your interest, but its state/action space must be large enough so that tabular
reinforcement learning methods cannot be used to solve it effectively.
Examples:
• playing Atari arcade games,
• teaching a four-legged creature how to walk,
• teaching a simulated bipedal robot how to navigate an obstacle course,
• driving a simulated car around a racetrack,
• playing Super Mario Bros.,
• landing a simulated spacecraft on the moon, and
• swinging-up a simulated double pendulum.