MazeLearnerExercise

Overview

In this exercise, you will write an agent much like the previous Maze Solver, in this case however, rather than programming a direct search agent, you will be writing a learning agent that will develop a model of the environment through some means, and then use this model to reliably solve the maze.

Instructions

Locate a list of temporal difference learning algorithms. We suggest looking at the AI Textbook AI: A modern approach. After this, in the same way as the Maze Solver Exercise, implement one of these algorithms and compare to the random baseline, Q-Learning, and SARSA implementations already completed.