摘要
爱恩斯坦棋是一种双人不完全信息博弈棋类,目前多用于计算机博弈领域的研究,其较为灵活且有随机性,可使用多种搜索算法对AI进行加强。为了提高AI在搜索时的效率及增强对战时的智能程度,采取了一种基于MCTS搜索并配合动态混合局面评估函数的算法,在评估函数中使用了棋子动态估值的方法,并对残局进行了一定的优化。经过多次对弈得出,应用改进算法后的新程序在与其他传统搜索算法程序对战时,显著提高了对弈时的胜率和搜索的效率,有效验证了该算法的实用性。
Einstein chess is a two-player game with incomplete information,which is mainly used in the field of computer game research at present.It is flexible and random,and its AI can be strengthened by using a variety of search algorithms.In order to improve the search efficiency of AI and enhance its intelligence in competition,an algorithm based on MCTS search and dynamic mixed situation evaluation function is adopted.In the evaluation function,the method of dynamic evaluation of chess pieces is used,and the residual game is optimized to some extent.After many games,the new program with the improved algorithm can significantly improve the winning rate and search efficiency when playing chess with other traditional search algorithms,which effectively verifies the practicability of the algorithm.
作者
宋英健
侯荣旭
孙嘉荣
史广阔
SONG Ying-jian;HOU Rong-xu;SUN Jia-rong;SHI Guang-kuo(School of Information,Shenyang Institute of Engineering,Shenyang 110136,Liaoning Province)
出处
《沈阳工程学院学报（自然科学版）》
2022年第3期72-76,共5页
Journal of Shenyang Institute of Engineering:Natural Science
