This report is the result of work completed in line with the AIML440 (Directed Individual study) at Victoria University of Wellington. I am completing this voluntarily (albeit for credits) as the University's 400 level course selection didn't have enough Reinforcement Learning content in it. Therefore I undertook this project to give me an opportunity to really get involved with some Deep RL.

The project has a time commitment of one course which is the equivalent of 1/4 of my time for a 12 week semester.

My project is being completed under the supervision of Dr Aaron Chen. The project is split up into three parts. Firstly I will complete a review of the RL landscape. Then I will move onto evaluating some of the current baseline algorithms. Lastly will be me applying my ideas to try and augment and improve some of the existing algorithms. This last section is about half of the project and I expect to know more about what will do closer to the time.

More information on the project plan can be found in this PDF.

See the appendix section of the report for a more detailed breakdown of the project source code.