Deep Reinforcement Learning (CS-866)

Department of Computer Science
University of The Punjab

Assignment 1

Study the gridworld_value_iteration.ipynb notebook that trains an agent to navigate a simple Grid World using Value Iteration, and perform the following modifications.

  1. (40 marks) Make a new notebook called GridWorld_3H.ipynb. Train an agent for a GridWorld with 3 holes. Visualize V, the policy, and the Q-table. Save GIF of the agent in action.
  2. (60 marks) Make a new notebook called GridWorld_RandomStart.ipynb. Train an agent that starts at a random location. Visualize V, the policy, and the Q-table. Save GIF of the agent in action.
Submit notebooks and your visualizations in the Google Classroom.