RL-on-OpenAI-Gym

In this repository we aim to:

1.Use CliffWalking-v0 from OpenAI gym:

• Create two agents to find the optimal policy using Policy Iteration and Value Iteration.
• Test-run and visualizing learning.

2.Use Taxi-v3 from OpenAI gym:

• Prepare and train your agent using i) On-Policy Monte Carlo and ii) Off-Policy Monte-Carlo using Important Sampling.
• Prepare and train two more agents using i) Q-Learning and ii) SARSA.

---

File Structure:

1. requirements_doc.pdf gives more detailed explanation of the requirements and the scope of this repository.
2. mc_qlearn_sarsa.ipynb aims to implement
   1. On-Policy Monte Carlo
   2. Off-Policy Monte Carlo+Importance Sampling
   3. Q-Learning
   4. SARSA
3. policy_iter_value_iter.ipynb aims to implement
   1. Policy Iteration
   2. Value Iteration

---