EECS498-008-Room-Gridworld

Implementation of Gridworld as MDP and simulation environment.

Gridworld

Algorithms

Many introductory algorithms will be implemented for the above gridworld

Value Iteration

Implementation and description here.

Policy Iteration

Implementation and description here.

On-policy Monte Carlo

Implementation and description here.

Off-policy Monte Carlo

Potentially buggy implementation and description here.

SARSA

• SARSA 0

• SARSA n-step

• Expected SARSA

Q-learning

Dyna Q

To Run an algorithm

In your terminal run the command. Where the file is the file you want to run minus the .py extension.

python -m Algorithms.path.to.file