Winning Strategy of The Online Tic-tac-toe Game Using Artificial Neural Network

Tic-tac-toe is a brain teaser game. The board of the game includes 24 different positions. Each of players has 12 homochromatic taw. At every turn, a taw is placed onto a position. Based on the map of the board, The game is led to winning or losing. In this game, if 3 homochromatic taw were in a consecutive positions, the winning is approved. In this paper, an artificial neural network is proposed for the winning strategy of the computer opponent. For this purpose, a web based application using asp .net technology, C# programming language and SQL Server DBMS is develpoed. Then, a variety of different online users played the game. All of the players’ moves are gathered and saved. Also, a multi-layer, feed-forward artificial neural network is used. The network inputs are 24 pairs of status and chance of each position. The network output is the selected position for computer opponet. The network is trained by gathered move patterns. Results show acceptable accuracy, precision and performance of the proposed approach

کلیدواژه ها:

Tic-tac-toe Game ، Neural Network ، Winning Strategy ، Online Game