基于多粒度理论的不完备决策规则获取  

Incomplete decision rule acquisition based on multigranulation theory

作　　者：王丽娟[1,2] 杨习贝 杨静宇[1] 吴陈 

机构地区：[1]南京理工大学计算机科学与工程学院,江苏南京210094 [2]江苏科技大学计算机科学与工程学院

出　　处：《南京理工大学学报》2013年第1期12-18,共7页Journal of Nanjing University of Science and Technology

基　　金：国家自然科学基金(61100116);中国博士后科学基金(20100481149);江苏省自然科学基金(BK2011492);江苏省高校自然科学基金(11KJB520004)

摘　　要：将多粒度理论引入不完备决策系统中,针对其不完备性,结合非对称相似关系提出基于相似关系的不完备多粒度粗糙集模型,并讨论该模型的性质。随后重点讨论基于相似关系的不完备多粒度粗糙集模型中属性约简和决策规则获取的问题,将多粒度理论和近似分布约简相结合,提出了不完备多粒度近似分布约简,获得所有的最简确定决策规则和最简可能决策规则。最后通过实例分析,发现使用多粒度近似分布约简与规则获取方法比原有的单粒度近似分布约简与决策获取方法更合理有效。Multigranulation theory is introduced into the incomplete decision systems.According to the incompleteness,the similarity relation-based incomplete multigranulation rough set model is proposed combining with the non-symmetric similarity relation,and its properties are discussed.The investigation focuses on the attribute reductions and decision rule acquisitions in the new model.Multigranulation theory and approximate distribution reduction are combined to solve the two problems,and the concept of the incomplete multigranulation approximate distribution reduction is presented.All the simplified certain decision rules and the simplified possible decision rules are obtained.A numerical example is employed to substantiate the conceptual arguments.The research results show that the methods of multigranulation approximate distribution reduct and rule acquisition are more reasonable and effective than the original methods in single granulation.

关 键 词：多粒度理论 不完备决策系统 近似分布约简 决策规则获取 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]