Researchand Implementation Of Search Technology In Connect Six Game

Computer game is the drosophila in the field of artificial intelligence. It’s the main research field in the artificial intelligence. The nature of the computer game is programming to realize the game between computer and people or computer and computer. It makes the computer simulate the mind of people play chess. For example, the algorithm of AlphaGo development by google. Currently, many researchers have studied computer game, and they have obtained a lot of outcome. It has greatly promoted game theory and artificial intelligence.This article, the connect6 has been research and optimization, include search technology and algorithm, evaluation function. In search technology, a threat search techno liogy has been designed as Defensive strategy. In search algorithm, a road based local search evaluation and "type of chess" based global search have been design and integrated into the Alpha-Beta algorithm. In evaluation function, a road based and "type of chess" based evaluation function have been designed, and use the genetic algorithms to self-learning offline for evaluation parameters. It is based on part of "chess-type" to genetic training. Through the experiment contrasts, it proves the correctness of the research methods. These technologies are applied to the program of the "Surprise Move". Through the university computer games championship it prove the feasibility and effectiveness of the search technology.The main contributions of this article are listed as following:(1) Based on threat search technology, an anti-threat search technology is designed. This search technology applies to search best move position in early stage. Through simulation to predict whether the other party can be continuously threat to win. It’s improving the accuracy of the defense. The experiment shows that search efficiency and the game level ware greatly improved by use threat search technology.(2) Apply "road" and "type of chess" based combinations search technology to Alpha-Beta algorithm. Because of the easy and quick characteristic of the road scanning, the efficiency of the valuation can be improved. "Type of chess" Based global scanning mode to evaluate the leaf node. In addition, change the "type of chess" based global search instead of local search. It’s can improve the accuracy of evaluate and game level. The experiment shows that the new combinations search technology can keep higher search efficiency and improve the game level of computer, with the "type of chess" based global search.(3) Apply adaptive genetic algorithm to train the evaluation function. It can make the parameter and evaluate more accuracy and adapt to vary different environment The experiment shows that the chess type value of evolution after using adaptive evolutionary genetic algorithm parameters can improve the level of game.The connect6 computer program "Surprise Move" adopts method (1) and (2). In the more intense competition situation, the "Surprise Move" achieved better results. The method (3) has been proved by experiment, and it will apply to university computer games championship in next year. We are hoping to get a better race results.