Exam 1
Instructions
You will work in your current teams.
After class, each team will be sent an email with their specific alias name and
exam_env.py.
You must keep the teams alias anonymous!
Not doing so will result in a penalty of \(10\%\) of the total grade of Exam 1 for you and your group.
- From Monday (October 20th) to Friday (October 24th), each team is in charge of coding one
main.pyandmodel.pyfor the Dynamic GridWorld environment.
Only Classical Reinforcement Learning algorithms that are model-free will be considered:
- On-Policy Monte Carlo
- Off-Policy Monte Carlo
- SARSA
- Q-Learning
- Double Q-Learning
- n-step Bootstrapping
- Please keep the following hyperparameters intact:
SEED = 123
SIZE = 10
EPISODES = 100
FPS = 0Do not alter exam_env.py!
Doing so might result in skewed results for your team.
Submission
Each team is responsible for submitting one ALIAS_main.py and ALIAS_model.py as a DAY#_ALIAS.zip.
Remember to put the corresponding alias name for your team in ALIAS_main.py.
ALIAS = 'your alias goes here'Submit your work on the same email!
Remember the email you and your team members will get after class with your alias name? Submit your DAY#_ALIAS.zip by simply replying to that email chain.
Please make sure that your code runs properly!
Any submission that is not following any of the steps above or does not run in ALIAS_main.py will not be accepted.
Leaderboard
Each day a leaderboard will be updated here to display the results of each team.