摘要
国际跳棋的博弈系统中,UCT算法是一个比较新颖的算法,其效果得到了广泛认可。但是UCT算法的缺点也同样明显,终局下的处理明显不如稳定性更强的Alpha-Beta算法。为避免UCT算法的不稳定性,本文采取限制搜索深度和增加扩展条件的方法对UCT算法进行改进,改进后的UCT算法胜率远超过改进前的UCT算法胜率,使得博弈水平得到了极大的提升。
Against draughts game system,UCT algorithm is a relatively new algorithm,and its effect has been widely recognized.However,the shortcomings of UCT algorithm are also obvious,which is obviously inferior to Alpha-Beta algorithm with stronger stability in the final game,In order to avoid the instability of UCT algorithm,this paper adopts the method of limiting the search depth and increasing the extension conditions to improve UCT algorithm,and compares with the improved UCT algorithm,it is concluded that the improved UCT algorithm win rate is much higher than the improved UCT algorithm,which greatly improves the game level.
作者
张家铭
王静文
李媛
ZHANG Jiaming;WANG Jingwen;LI Yuan(School of Science,Shenyang University of Technology,Shenyang 110870,China)
出处
《智能计算机与应用》
2022年第1期128-131,共4页
Intelligent Computer and Applications
作者简介
张家铭(2001-),男,本科生,主要研究方向:计算机博弈;通讯作者:王静文(1965-),男,学士,工程师,主要研究方向:人工智能、信息安全,Email:wangjingwen007@126.com;李媛(1976-),女,博士,教授,主要研究方向:人工智能和随机过程。
