The principal aim of this project is to develop an AI of a the 2-players game GO. The major issue with designing an AI of a GO player is that it is very difficult to implement an heuristic to describe and to evaluate a position of a goban (the board of the game) because of the simplicity of the rules and the relative huge depth of the game.
This is why recently some people try to use the methods of Monte-Carlo in order to construct the min-max graphs of the positions of the game. The principal idea is to simulate a large number of complete games and use the probability in order to assign a value of each node of the tree of the positions in function of the number of game won.
The aim of the project is to implement these methods of Monte-Carlo to improve results found nowadays.