一种新的文化算法框架下知识融合策略  

New Knowledge Integration Strategy Based on Culture Algorithm Framework

作　　者：陈思华[1] 陶长琪[1] 

机构地区：[1]江西财经大学信息管理学院,江西南昌300013

出　　处：《小型微型计算机系统》2009年第10期2030-2033,共4页Journal of Chinese Computer Systems

基　　金：国家自然科学基金项目(70663002;70671003)资助;国家社会科学基金项目(07BJY140)资助;教育部新世纪优秀人才支持计划项目(NCET-07-0382)资助

摘　　要：对不同主体,多种来源的知识进行融合,得到一个高效的知识库是提高决策质量并增强组织核心竞争力的关键.提出一种在文化算法框架下的知识融合策略,将知识统一进行编码,通过群体空间和信任空间两个阶段的进化,并建立两个空间之间的通讯机制,生成一个有效、精练的知识库.实验表明,相比传统遗传算法,此模型可以更准确地就将知识分类,并能减少知识冗余,剔除矛盾的知识.Integrating knowledge from different subjects and sources to obtain an effective knowledge base is the key to improve decision quality and enhance organizational core competency. We put forward a knowledge integration strategy under the framework of culture algorithm. It encodes the knowledge uniformly and through the evolvement of the two phases of population space and belief space the communication protocol is established among the two spaces. Then an effective and concise knowledge base is automatically produced. The experiment shows that comparing with traditional genetic algorithm, the model can classify the knowledge more precisely, reduce redundant knowledge and remove contradictory knowledge.

关 键 词：文化算法 群体空间 信任空间 遗传算法 知识融合 

分 类 号：TP182[自动化与计算机技术—控制理论与控制工程]