The Design Of Japanese Mahjong Assistant Decision System Based On Multi-agent Game Analysis

With the rapid development of information technology and computer technology,artificial intelligence technology has developed rapidly in many aspects,such as finance and trade,medicine,diagnosis,heavy industry,transportation,telecommunication,online and telephone services,legal services,scientific discovery,toys and games,and music.Among them,the ability to think and act like a person is the ultimate goal of learning anthropomorphic behavior pattern,which plays a crucial role in the performance of artificial intelligence.Games have always been a great way to test and train artificial intelligence,which in some areas has surpassed human players.In the past few years,AlphaGo has defeated lee-Sedol and ke jie,becoming the first artificial intelligence program to defeat a human professional go player and the first go world champion.As one of the important applications in the field of artificial intelligence,the introduction of artificial intelligence technology in games plays an important role in promoting research.As a typical form of artificial intelligence technology,multi-Agent technology in intelligent robot in recent years,traffic control,expert system for flexible manufacturing,coordination,etc.has a broad application,by introducing multi-Agent decision fusion techniques,such as by using the methods of game theory,can be based on different scene situation analysis and countermeasures of intelligent evolution,obtain the best decision.As a popular game in east Asia,Japanese mahjong has perfect rules and platforms,as well as fierce antagonism,which requires high decision-making of artificial intelligence.Based on the analysis and study of multi-Agent,this thesis chooses Japanese mahjong as the research object to study the game behavior of multi-Agent,and puts forward the method of multi-Agent decision fusion.This method can effectively improve the decision accuracy of Agents,Firstly,this thesis introduces the basic theory of game theory and multi-Agent,and the mathematical theory and method of game theory to study the phenomena with the nature of struggle or competition.Game theory considers the predicted and actual behavior of individuals in a game and studies their optimal strategies.Multi-Agent system is a collection of multiple Agents.Its goal is to build large and complex systems into small systems that communicate and coordinate with each other and are easy to manageSecond,the rules of the game for the Japanese mahjong interpretation,on the basis of the modeling for four people of non-cooperative game under the incomplete information problem,the phases of the decision-making model and the practical strategy of Agent model and computer program implementation,and puts forward a recursive tree search based on the Japanese mahjong algorithm of automatic play,with multiple Agent model of different strategies,each Agent to simulate human players in the game a strategy.Finally,the different Agent in Japan mahjong standardized platform "chicken" on test,through respective analysis and statistical analysis to verify the advantages and disadvantages of different strategies,and then puts forward the decision fusion method,and fusion of different Agent decision in testing new Agent after fusion,compared with the performance of the single Agent,finally formed a complete game aided decision making software.The results show that this method can effectively improve the performance of multi-Agent cooperation.