Cliff Walking Exercise: Sutton's Reinforcement Learning 🤖

My implementation of Q-learning and SARSA algorithms for a simple grid-world environment.

The code involves visualization utility functions for visualizing reward convergence, agent paths for SARSA and Q-learning together with heat maps of the agent's action/value function.

Contents: ⭐

• cliff_walking.py: Q-learning, SARSA, Visualization Functions
• cliff_walking_report.pdf: Analysis on the Q-learning and SARSA algorithms

References:

• Sutton's Reinforcement Book