Week 10: Final Project Proposal

Inspiration/Motivation:

I was inspired by last weeks discussion on reinforcement q-learning and wanted to look more into this concept.  I see a potential for many applications of q-learning and want to take time applying it to my own scenario to understand the process.  My plan is to use the example we saw last week in class by Aidan Nelson as a baseline, and further his example.  Hopefully, I will be able to make the project interactive, so that users can manipulate the q-table.  Ultimately, this will be an exploration for myself to learn about this type of machine learning while also providing an opportunity for others to see an agent learn using reinforcement q learning in various scenarios. 

Source Material/Code:

  1. Aidan Nelson’s code

  2. Reinforcement Learning Tutorial

  3. My starter code

Implementation options:

  • grid/ automated maze

  • automated frogger

  • other ideas?

Current one sentence description:

This project will create a maze field that the user can manipulate and an automated agent will learn using reinforcement q-learning.