final_project_checkin.md

Final Project Check-in

Name: Kevin Loftis

Finalized Research Question:

Can build a game/environment called Runaway (see README for details) and train a Deep Q-Network agent to play it?

Environment:

Custom built for this project. The environment class is called Game in src/game.py.

Random Agent:

Random Agent is in src/agent.py. The class name for ranomd Agent is Random_Agent. The agent is the green slime, and the minion is the floating eyeball.

RL Agent:

The RL agent lives in src/agent.py. The class name for the RL agent is DQN_Agent.

Total rewards over time.

To finish the project I need to train my agent for longer.