This is an variant of 2048 game. In this project, the AI use TD(lambda) Learning and Alpha-Beta search to implement two AI. One is a solver to maximize the tile, and the other is an evil preventing high score tiles from generating.
First lauch the platform (see platform/readme.md)
git clone https://github.com/tonylearn09/2584AI.git
cd 2584AI/
make
-
To train the AI:
./train_game [Numbers of iteration] -t -
To run the AI:
./ai_client [account] [password]