Tetris AI app that can train a Tetris bot that can outperform the vast majority of Competitive Tetris Players using a Genetic Learning Algorithm Model built in C++ that can evaluate the best possible next moves, taking into account at most 4 (at depth=2) future pieces at once.
[ APP FEATURES ]
-> You start off with you own personal bot that has a randomized evaulation system
-> There is also a pre-built bot that has a set evaulation system that I have personally obtained by training the same bot that you start off with for a long period of time.
-> You can choose to battle against either your personal bot or the pre-built bot, or you can choose to have the two AI bots challenge each other
-> You can train your own bot using different training enviornments that can alter the behaviour that your personal bot will learn
-> Modes: 1 vs 1, highest score wins, last to survive
-> There is a settings menu to alter certain game options
[ VERSION UPDATES ]
1.0 -> Naive O(2^n) approach, still works for 4(2 on display) depth 1.1 (in progress) -> Will implement MCTS to optimize further, and will improve the ui, with images, better sprites, and audio.
[ REQUIRMENTS ]
-> MinGW 13.1.0 installed and set in your global PATH variables
Download from : [ https://winlibs.com/ ]
-> All else included in installation folder
[ SCREENSHOTS ]
Menu:
Battle (Player vs Personal Bot)
Training (1 vs 1):
Settings:
Example Selection Screen (Training):
