This is an implementation of an AI that plays the board game Quoridor, designed in 1997 by Mirko Marchesi. Quoridor is a 2 player turn based board game akin to the likes of Chess or Go. For details on how the game works and how I approached the AI implementation feel free to read my paper on it: Quoridor, a board game AI
For those with little time to spare, here comes the short version.
This is the initial state of the board, being both pawns the players. Each player needs to reach the row at the other side of the board. Not necessarily the tile in the center of the row.
To do so, there are 2 possible moves to choose from each turn, either move the pawn orthogonally in the board (never diagonally) or place a wall blocking the path between two tiles. It is against the rule to place a wall in such a manner that a player has no possible path to reach the destination row.
Check here a game played by me (blue) against the AI (green). The top left console is the input system, a bit too raw and uncomfortable to use, but it gets the job done for testing purposes.
If you managed to follow on what is going on in the game, you might have noticed that I lost the game with a fatal mistake at 3:45. Up to that point it was going pretty well, but the machine never forgives!