Research And Implementation Of A Chinese Chess Computer Game System With Self Learning

Computer game is an important branch of artificial intelligence research. Many methods and theories originate from it. It also brings remarkable social and scientific influence. The research of game playing is very popular. With the deeper study of research, computer Chinese chess becomes one of the most active parts of computer game research area recently.After summarizing related researches on computer Chinese chess, Some key problems are studied and discussed in this dissertation. Based on above work, an integrated computer chess-playing system are designed and developed. The whole work mainly focuses on the following three aspects:1. Introduce the key component parts of a computer Chinese chess system which involve how to express the chessboard, store the board position in computer and generate legal moves. Then the pruning algorithm and the evaluation function of chessboard position are discussed chiefly.2. Make a study on the optimization of search algorithm and static evaluation function. In the dissertation, four kinds of methods for optimizing pruning algorithm are summarized. And according to the practice, a proper search algorithm is designed and implemented. Then some advantages and disadvantages of static estimation method for game situation are analyzed.3. In order to solve the problem that linear evaluation function depend on the chess knowledge too much and can't evaluate game situation exactly. A method that the BP network combining the optimized temporal difference reinforcement-learning algorithm replaces linear evaluation function is proposed. At last, a platform system is made to implement this method and discuss some relative question.The Chinese chess programs carried out by this dissertation achieves the initial goal via optimizing and enhancing the evaluation, which increases the level of playing chess, is worth for research and use.