Topics covered in this module-

1. Components of Reinforcement learning and Environment dynamics.

2. Custom Environment creation using OpenAI gymnasium API.

3. Registering and Running of Environment.

4. Analysis and code of Grid walk Environments ( Bandit-Walk and Random-Walk Enironments. )

5. Theroies of methods based on Markov's decision process and Bellman's Equation.

6. Algorithms to explore an environment -

     * Greedy Approach ( Pure Exploitation )
     * Pure Exploration
     * Epsilon Greedy Approach
     * Decaying Epsilon Approach
     * Softmax Exploration Strategy
     * UCB Strategy
    

7. Probabilistic Prediction methods -

      * Monte-Carlo FV
      * Monte-Carlo EV
      * Temporal Difference 
      * n - Step TD
      * TD Lambda 
     

8. Control Algorithms -

      * Monte-Carlo FV Control
      * Monte-Carlo EV Control
      * SARSA
      * Q- Learning
      * Double Q-Learning 
      * SARSA lambDa ( Accumulating and Replacing Traces )
      * Q lambDA ( Accumulating and Replacing Traces )
      * DYNA - Q ( Model Based )
      * Trajectory Sampling ( Model Based )
     

9. Deep Reinforcement Learning -

      * Neural-Fitted Q ( NFQ )
      * DQN , DDQN , D3QN
      * PER-D3QN
      * Reinforce
      * VPG
      * DDPG
      * TD3
      * PPO
     

10. Plots and reports based on above methods.

Notes-

1. The assignment contains all the implementation and plots.
2. It has two solutions - the one which I wrote and the one provided by tutors.
3. All the codes in this repo may or may not be correct. Please verify once before using.