基于认知程度的双论域粗糙集理论及在动态决策中的应用研究

项目名称： 基于认知程度的双论域粗糙集理论及在动态决策中的应用研究

项目编号： No.71301100

项目类型： 青年科学基金项目

立项/批准年度： 2014

项目学科： 管理科学

项目作者： 阎瑞霞

作者单位： 上海工程技术大学

项目金额： 23万元

中文摘要： 针对不确定性决策过程中决策者会随决策环境的变化调整决策方案的现象，本课题提出基于认知程度的双论域粗糙集理论以提高决策的时效性：(1)利用构造性与公理化方法构造基于认知程度的双论域粗糙集近似算子并研究其基本性质，提出基于认知程度的双论域粗糙集的属性重要度、匹配度、粒度和粗糙熵等概念，进一步研究基于认知程度的双论域粗糙集属性约简、规则提取和不确定性度量等理论和方法，发展双论域粗糙集理论体系，为制定动态决策方案提供理论指导。(2)设计基于认知程度的双论域粗糙集近似算子、属性约简的增量求解算法和决策规则的动态提取方法，为提高决策速度提供方法支持。(3)结合本体理论获取决策领域的基本知识和描述语言，利用基于认知程度的双论域粗糙集模型推理、获取动态决策规则，构建基于认知程度的双论域粗糙集动态决策模型。(4)将该模型应用于轨道交通应急管理中，对轨道交通突发事件的变化及环境变化进行动态分析，提供应急方案。

中文关键词： 知识库；双论域粗糙集；认知精度；认知范围；知识约简

英文摘要： In uncertainty decision-making process, decision alternatives should be adjusted to adapt environment. Rough set over dual-universes based on cognitive degree (RSDU-CD for short)is built from cognitive degree to confirm the timeliness and effectiveness of dynamic decision-making. (1) Approximation operators of rough set over dual-universes based on cognitive degree are proposed by constructive and axiomatic method. Properties of RSDU-CD are researched. New concepts of RSDU-CD, such as attribute significance, matching degree, granularity, rough entropy and so on, are presented. Properties, uncertainty measure, attribute reduction and rule extraction are researched further to develop rough set over dual-universes theory system, which provides a theoretic guide for dynamic decision-making. (2) Incremental algorithm of approximation operators, attribute reduction algorithm and dynamic rule extraction methodology are designed based on cognitive degree, which provide support to dynamic decision-making for timeliness. (3) General knowledge and describe languages of decision problem can be obtained based on the ontology theory. Then, rough set over dual-universes based on cognitive degree is utilized to reason and generate decision rules automatically. A dynamic decision-making model is built, which can analysis dynamic

英文关键词： knowledge base；rough ste over dual-universes；cognitive accuracy；cognitive scope；reduction