摘要
蒙特卡罗树搜索(MCTS)在许多完备的信息双人游戏中获得成功。本文给出了UCT(Upper Confidence Bound Apply to Tree)算法结合了UCB公式和蒙特卡洛树搜索算法,同时与局面评估相结合,根据点格棋长链和环的特点对算法进行了优化。有利于更快更准地找到当前局面的最优解。
Monte Carlo tree search(MCTS)has been successful in many perfect information games.In this paper,UCT(Upper Confidence Bound Apply to Tree)algorithm is proposed,which combines UCB formula and Monte Carlo tree search algorithm.Meanwhile combined with situation assessment,the selection of nodes for evaluation by UCB algorithm is conducive to find the optimal solution of the current situation faster and more accurately.
作者
朱良双
王静文
李媛
ZHU Liangshuang;WANG Jingwen;LI Yuan(School of Science,Shenyang University of Technology,Shenyang 110870,China)
出处
《智能计算机与应用》
2021年第2期129-131,共3页
Intelligent Computer and Applications
关键词
UCT算法
估值函数
点格棋
UCT algorithm
evaluation function
Dots and Boxes
