Research On The Computer Game Key Technology And System Design Based On Checkers Complete Information

Computer game is one of the most important and challenging research directions in the field of Artificial Intelligence(AI).From 2016 to 2017,AlphaGo has defeated the professional Go players and the champion of the world successively.The Computer game has drawn wide attention from all over the world.For a complete information game,this paper takes Checkers(100)as the object to design and implement Checkers computer game system.It focuses on the key technologies of Checkers.The new evaluation and the search algorithm are put forward.That improves the efficiency of evaluation and search,and enhances the power of Checkers.The main research work in this paper is as follows:(1)An evaluation algorithm of the value matrix of chess pieces based on expert knowledge is put forward.For traditional evaluation methods,it is difficult to take active measures in the face of the changing and complicated game.In response to these deficiencies,and the chess pieces distribution of the whole game is classified and processed according to the expert knowledge,and the different chess pieces sorted out are calculated and compared by different methods firstly.Then,the value matrix of chess pieces is used to make corresponding score points to the evaluation.Finally,the values of each type of different pieces are weighted,and a new chess piece evaluation function is obtained.Experiment shows that the promotion effect of the evaluation is obvious.(2)The sorting strategy of UCT algorithm(UCB applied to Tree)based on Monte Carlo search tree(MCTS)is proposed.There are no rules to choose when selecting the drop point,which causes many excellent optional points are easily overlooked.A new method is proposed,that is,before the UCB(Upper Confidence Bound)algorithm selects the maximum node,sorting all the optional points according to the sorting strategy firstly.Then UCB selects the nodes according to the results,and according to the selected nodes to continue to expend finally.By comparing this method with traditional algorithm,the experimental results show that the above proposed method improves game search efficiency,improves the chess power,and achieves the expected results.(3)By studying the principle and method of computer checkers game,the basic data structure of checkers chessboard elements are designed and modeled.The programming language is used to realize the front-end game platform and back-end search engine of checkers computer games.The system effectively combines the traditional mode method and the modern evaluation search technology.This system can complete the basic requirements of the Checkers,for example,playing chess,eating chess pieces,becoming king and so on.The functions of human-human game,human-machine game and machine-machine are realized.In short,it has a certain degree of Artificial Intelligence.