Font Size: a A A

Research And Implementation Of Einstein W(u|")rfelt Nicht! Computer Game System

Posted on:2017-03-11Degree:MasterType:Thesis
Country:ChinaCandidate:Y GuangFull Text:PDF
GTID:2308330485464136Subject:Computer application technology
Abstract/Summary:PDF Full Text Request
Computer game, also known as computer games, is a great challenge to the field of computer research and development prospects, its artificial intelligence as an extremely important issue, known as artificial intelligence "flies," said. For the study board games and computer games during the study of a standard and extremely important issue. Many important existing theories and methods, such as various types of search algorithms and intelligent method, initially proposed are present in order to solve some of the problems in the chess game. For the study before the computer game can be traced back half a century, in the past 50-60 years, scholars of the painstaking research of artificial intelligence computer game, including Othello, chess, checkers, backgammon, Go, etc., we made a lot of impressive results.1997 IBM’s "Deep Blue" beat chess champion Garry Kasparov is attracting worldwide attention and vibration, let the world for the first time realized that the ice machine intelligence beyond human chess game. Study compared to foreign and domestic research for this area started late, but with several years of game and domestic competitions conducted, attracting more and more students and scholars involved in the game, the domestic game Computer ushered in the golden period, which greatly promoted the development of domestic computer game research.Conventional machine game model is more complete information game, chess process that fully understand the status of the other competitors features and benefits of the policy approach, the existing computer games such as research studies mostly carrier, such as backgammon, chess etc.. Games such data structures and design depend on the situation described accurately state the status of policy information, search engines also have ready-made templates for reference. Object of this study EinStein wurfelt nicht! game belongs to incomplete information game, for the game in the process of understanding the information is incomplete, and the situation before data structure to describe the design state, you need to state other opponents characteristic information such as an estimate, let the game information "pseudo-full" and then the related design, the same, the need for the traditional game search algorithm can be modified to adapt to the game with incomplete information.In this paper, EinStein wurfelt nicht! game platform for research support, research some key technical aspects of computer games, including the basic concepts, study and analyze a number of search engines, designed EinStein wurfelt nicht! game machine system. In the checkerboard data structure design process, opponents of state information were to predict and quantify the estimate to complete information, evaluating and attack simultaneously from two directions defense got the offensive and defensive evaluation function; the same time, Research on parameter optimization method of evaluation function, to ensure accuracy and reasonableness evaluation function design; in the search algorithm, based on the traditional method of complete information game-Minimax algorithm, reference desired search algorithm, in great artificially added between the layer and the layer of tiny dice layer, and the application of the above evaluation function simultaneously added to the algorithm permutation table optimization technology is designed to adapt to EinStein wurfelt nicht! chess offensive and defensive expectations algorithm to solve the traditional Game algorithm can not build a game tree problem and improve the level of the game.From the search efficiency levels and game two angles experiments can be obtained:the traditional method with respect to the desired application of the valuation function in the use of offensive and defensive expectations algorithm, although the decline in the efficiency of the search, but in the game there are levels good break, the game proved significantly improved algorithm applies the desired offensive and defensive evaluation function within a certain depth range game and game efficiency levels achieved a good balance, both have a high level of the game, but also make full use of the game time. Meanwhile, after the addition of permutation table optimization, search efficiency and the level of the game have some improvement, innovation of this paper is as follows:(1) research and design applicable EinStein wurfelt nicht! game with incomplete information data structure, in-depth study constitutes the valuation function for the valuation of the traditional one-sided evaluation function exists the problem of inaccurate, two from defense to attack taking into account the direction and EinStein wtirfelt nicht! combined win at chess two ways to design accurate and efficient evaluation function better offensive and defensive; while analysis of a number of optimization parameters used in the valuation methods and strategies. In the design of the valuation function, the use of aggressive factors and threats factor to adjust the path of the law, effectively improve the level of the game.(2) specifically on the use of climbing method with manual adjustment of the valuation of both offensive and defensive parameter optimization, the algorithm will be applied to the valuation after optimization to greatly improve the level of the game.(3) based on the traditional game Minimax algorithm, the desired reference search algorithm, and improved design suitable for both offensive and defensive EinStein wurfelt nicht! desired algorithm, adding a layer in the original layer dice Minimax layer, throw dice to simulate the process, the process of solving the game problem of incomplete information, and applications (1) the valuation function to the algorithm, permutation table is added to optimize the algorithm to reduce the number of nodes in the search process, improve search algorithm efficiency.These methods have been successfully applied in EWN System which is called "Flowing Water" and "Fire" program. And The "Flowing Water" program won the first prize in "CDUTCup", the Dots and Boxes project of the National Undergraduate Computer Game Competition in 2014. It proved that these methods are workable and practical.
Keywords/Search Tags:computer game, EinStein würfelt nicht!, evaluation, ODEMA, Minimax algorithm, except algorithm
PDF Full Text Request
Related items