This repo provides solutions for problems 0 and 1. The solution for problem 0 is simply a pdf file that can be found in the appropriate folder.
With regards to problem 1, I created a dueling deep q-learning agent in a separate file that can be found in p1's folder. I also slightly modified the test_agent.py
script in order to make the state processing part easier for the agent. The solution may be tested by simply running python3 test_agent.py
. The specifcs of the algorithm may be found in the form of comments in test_agent.py
and dqn_agent.py
.
This project is licensed under the MIT License - see the LICENSE.md file for details