This AI learns to solve 4 handmade mazes from scratch, in scratch! As the AI goes around, it uses the Q learning formula to calculate values to tell the AI which moves to make later on down the track. The values are stored in a Q table and when the epsilon equals to 0, the AI will finish training and use the table to determine which move to make. All you need to do is press that green flag and the simulation will start!
Thanks to ChatGPT and EdChat for my numbers and thanks to the 2 or 3 tutorials that I used to learn Q learning! This project took a long time and I made it from scratch so it would be appreciate if the hit the heart and star buttons! If you want more AI, you can go into the code and find the comment to change to number of AI! Play on turbo mode for a faster experience (shift click the green flag)!