Research And Implementation Of EinStein Würfelt Nicht! Computer Game Algorithm

Computer game,also known as machine game,is a great challenging branch of the field of artificial intelligence.The study of computer games dates early back to the first paper chess program designed by Alan Turing in the 1950 s.After nearly seven decades of research,experts and scholars in the field of global computer game have contribution to wealth theory and practical case studies of computer game,including relevant technology,method and strategy.Especially computer game,such as Chess,Go chess,has achieved amazing results.For example,the challenge of IBM “Deep Blue”and Chess King Kasparov in 1997,as well as the confrontation of World Go Championship Li Shishi and Google AlphaGo in 2016,have ended with the failure of the human chess player.Through these events,people had a new understanding of computer game and truly realized the force of in the field of computer and artificial intelligence.Compared to the splendor of foreign countries,domestic computer game research started later,but it also developed rapidly.For instance,starting from the first National Computer Game Competition in 2006,11 continuous games have been successfully held so far.There are more than 300 teams,more than 400 teachers and students,compete for 17 projects,and participate in nearly 40 colleges and universities.These activities have greatly promoted the development of computer game activities in China.This article selects EinStein Würfelt Nicht! as research object,which belong to computer game of incomplete information and unpopular among all the people.EinStein Würfelt Nicht! has features of a variety of initial game states and needs to be decided the scope of virtual move by dice.These characteristics not only increase the interest of the game process,but also enhance the research difficulty of game decision-making.The paper introduces around the famous UCT(Upper Confidence Bound Apply to Tree)algorithm in the field of computer game,puts forward the idea of ? ?improvement and its implementation method,and builds a new game strategy system based on this in order to improve the chess force.For this purpose,the full text primarily elaborated the basic principles of computer games and related technologies.Secondly,it briefly introduced the rules of EinStein Würfelt Nicht! and established a mathematical model.Then,it further analyzed the UCT algorithm and built applicationstrategies.Finally,EinStein Würfelt Nicht! computer game was developed,and participated in the National Computer Game Competition and verified the results.The paper carries out small innovations in the following areas:(1)Establish a basic strategy for offensive and defensive balance.Analyze particularly the characteristics of EinStein Würfelt Nicht! and establish mathematical model,from combination offensive and defensive and reduction decision-making influence by throwing dice,and by means of adjustment of offensive factors,defense factors and protection factors to optimize the valuation method and constructe quadratic valuation function.It greatly improves the accuracy of valuation of EinStein Würfelt Nicht! computer game.(2)Establish Hybrid Pptimization algorithm.Hybrid Optimization algorithm combination Monte Carlo algorithm based on stochastic simulation and Expected Search algorithm based on prediction is proposed,and quadratic valuation function is applied to the Hybrid Optimization algorithm to improve simulation search process and operational efficiency of machine learning method Monte Carlo.(3)Propose speed and global optimization method.Based on the powerful multi-core computing capabilities of PCs,the introduction of parallel computing methods has improved the operating speed of UCT algorithm.The current valuation factor(WINK)and sub-optimal node balancing factor(UCTK)are introduced as decision-making valuation method to partly overcome the optimization problem of move and promote the global optimization move of computer game.The author of this paper carried the KnighTeam-AN2 program developed with the above research results and followed computer game team to participate in the 2national competitions held in Shenyang Aerospace University in 2016 and in Chongqing Internet College and Chongqing University of Technology in 2017 and won the national championship.This directly proves the feasibility and effectiveness of the research results.