Artificial Intelligence

The project consists of the generation of a maze using a Ruled Based System and aims to solve it using a Reinforcement Learning algorithm.

For the Reinforcement Learning, Q-learning was used. It is a model-free Reinforcement Learning algorithm based on 4 elements:
• Agent: the player, which is represented as a red cube.
• States: all the situations that the agent encounters, in this project we find four states, Start, Correct Path, Hit and End.
• Actions: the transitions from one state to another, in this case the actions are Left, Right, Forward and Backward.
• Rewards: the rewards are given to the agent after every transition. The scoring system is based in a negative reward when the agent hits a wall and a positive reward when it reaches the end of the maze.