基于可拓学和HowNet的策略生成系统研究进展  

Strategy-generating system based on Extenics and HowNet

作　　者：杨春燕[1] 李卫华[2] 汤龙[1] 汪中飞 陶星[2] 

机构地区：[1]广东工业大学可拓学与创新方法研究所,广东广州510006 [2]广东工业大学计算机学院,广东广州510006

出　　处：《智能系统学报》2015年第6期823-830,共8页CAAI Transactions on Intelligent Systems

基　　金：国家自然科学基金资助项目(61273306);广东省科技计划资助项目(2012B061000012)

摘　　要：在已有可拓策略生成系统框架基础上,利用可拓信息-知识-策略形式化表示体系、HowNet的知识系统描述语言(KDML)和Agent的智能引导,建立了基于可拓学和HowNet的策略生成系统的基本流程和系统框架结构,增强了可拓策略生成系统解决矛盾问题的能力,改善了知识资源缺乏问题,提高了其问题模型建立的准确性和知识处理能力。初步的应用显示了该研究的优越性。把可拓学和HowNet这2个中国原创的理论和应用工具相结合研究策略生成系统,将使这项有别于传统研究思路的基础研究取得重要突破,产生具有完全自主知识产权的成果。这也是为将来能实现矛盾问题的智能化处理作基础性的工作。Due to the lack of a sufficient knowledge base, traditional extension strategy-generating systems （ESGS） are often not sufficiently robust to generate effective strategies for solving given contradictory problems. To overcome this drawback, this paper proposes the use of the HowNet knowledge base to help ESGS to generate strategies. We established the basic flow chart and system frame structure using the formalized system of extension information knowledge-strategy, the knowledge database description language （KDML） of HowNet, and the intelligence guid- ance by Agent. Preliminary applications have shown the superiority of this concept. Combining Extenics with How- Net, both original Chinese theories and application tools, respectively, will establish a research foundation that will necessarily diverge from traditional research ideas toward the likelihood of significant breakthrough. This research will produce original results characterized by independent intellectual property rights, which will make possible the realization of the intelligent processing of contradictory problems in the future.

关 键 词：可拓学 HOWNET 策略生成 可拓策略生成系统 矛盾问题 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]