This repository contains the first lab for the Deep Learning – Reinforcement Learning module. The focus is on implementing and comparing two key approaches in reinforcement learning:
Model-Based (Markov Decision Process: Policy Iteration / Value Iteration)
Model-Free (Q-Learning with GridWorld)
The lab helps develop practical coding skills, understand convergence behavior, and analyze execution time differences between these methods.
Upload and run the Policy Iteration and GridWorld Q-Learning notebooks.
Complete missing code sections.
Experiment with larger grid sizes and observe performance changes.
Insert screenshots of completed code and results.
Modify MDP notebook to compare execution time and convergence with Q-Learning.
Explain differences between both approaches.
Provide screenshots of results.
Successful implementation of Policy Iteration and Q-Learning.
Performance comparison between Model-Based and Model-Free approaches.
Observations on scalability with larger grid sizes.