Study of Game Strategy Emergence by using Neural Networks

In artificial intelligence systems, various machine learning algorithms are used as learning algorithms. The most used artificial intelligence approaches are symbolic rule-based systems and subsymbolic neural networks. The main objective of this work is to study the game strategy emergence by using subsymbolic approach neural networks. From the viewpoint of artificial intelligence, games in general are interesting. The games are often complex even if their definitions, rules and goals are simple. In this work we are concerned about the Battleship game. The Battleship game is a representative of games with incomplete information. We will design and implement solutions based on using subsymbolic artificial intelligence approach to solve the Battleship game. We will use machine learning techniques as the supervised learning and the reinforcement learning for this purpose. We will compare machine learning techniques used by using simulation results and statistical data of human player.