2048 has caught the interest of gamers and researchers alike as a challenging puzzle game. This study seeks to evaluate the effectiveness of AI algorithms designed for playing 2048. By testing and comparing different approaches, the research aims to identify successful strategies and analyze their performance. The evaluation will focus on key factors such as achieving high tile scores and computational efficiency. The main algorithms investigated in this study include Minimax, Expectimax, and Monte Carlo Tree Search. The ultimate goal of this assessment is to gain deeper insights and assess how well these strategies work in the context of playing 2048.
The code provided in this repository is for CS686 my final project. All the results of the evaluation of algorithms are in this notebook