摘要
提出计算机博弈五大构成要素并基于分层递进思想设计了博弈智能体的分层结构,以久棋为案例,构建了久棋博弈智能体的知识库,重点提出了规则库、开局库的构造方法。针对藏族久棋规则的特殊性,面向久棋特有的3个阶段,基于量化思想,建立了分段评估体系。最后,以此方法开发久棋博弈程序,参加了全国计算机博弈大赛并获得好成绩,验证了方法的可行性和有效性。
The five major components of the computer game were put forward and the hierarchical structure of the game agent based on the hierarchical and progressive ideas was designed.Taking Jiu Chess as a case,the knowledge base of Jiu Chess game agent was constructed,and the construction methods of rule base and opening base were mainly put forward.A segmented evaluation system was established aiming at the particularity of the Tibetan Jiu Chess rules and facing the three unique stages of Jiu Chess based on quantitative thinking.In the end,this method was used to develop the Jiu Chess game program,which participated in the national computer game competition and obtained a good result,thus the feasibility and effectiveness of the method were verified.
作者
张小川
刘溜
陈龙
涂飞
ZHANG Xiaochuan;LIU Liu;CHEN Long;TU Fei(School of Artificial Intelligence,Chongqing University of Technology,Chongqing 401135,China)
出处
《重庆理工大学学报（自然科学）》
CAS
北大核心
2021年第12期119-126,共8页
Journal of Chongqing University of Technology：Natural Science
基金
国家自然科学青年基金项目(61702063)。
关键词
藏族久棋
知识库
博弈智能体
棋型匹配
分段评估方法
Tibetan Jiu Chess
knowledge base
game agent
type matching
segmented evaluation method
