(α,β)集对限制优势粗糙集及决策模型  

On(α,β)Set-pair Limited Dominance Rough Set and the Decision-making Model

作　　者：翟育明[1] 蔡红[2] 郭斌[3] 

机构地区：[1]上海应用技术学院经济管理学院,上海201418 [2]复旦大学经济学院,上海200433 [3]江西师范大学软件学院,南昌330027

出　　处：《系统管理学报》2014年第3期437-443,共7页Journal of Systems & Management

基　　金：上海市教委科研创新项目(14YS119);上海应用技术学院引进人才项目(YJ2013-13);国家自然科学基金青年项目(71301100)

摘　　要：针对不完备信息系统中的偏好多属性决策问题,提出了一种基于(α,β)集对限制优势粗糙集的决策分析模型。引入集对优度和关键属性构建一种集对限制优势关系,构造了基于(α,β)集对限制优势关系的粗糙集模型;在此基础上,研究了粗糙近似算子的性质、知识的属性约简和决策规则的提取方法,并采用计算机仿真的方法分析其分类误判率。(α,β)集对限制优势关系比原有的集对优势关系更加符合实际,分类的整体性能更优。比较研究的结果表明:(α,β)集对限制优势关系在保持较高的分类精度和质量的基础上,其分类误判率显著优于集对优势关系和限制相似优势关系。In order to deal with multi-attribute decision-making problems with incomplete information, we present a decision model based on （α,β） set-pair limited dominance rough set. We define the set-pair dominance degree and the key attributes to build a （α,β） set-pair limited dominance relation and construct a rough set model based on this relation. We then study the property of the rough approximation operators, knowledge and attribute reduction of the rough set model, and use computer simulation to analyze its classification error rate; From the analysis, we know （α,β） set-pair limited dominance relation is more realistic than the original set-pair dominance relations, and the overall performance of the classification of this relation is better. An example is discussed to validate the theory, The results of comparative research show that （α,β） set limited dominance relationship maintains a high classification accuracy and quality, and its classification error rate is better than set-pair dominance relation and limited similar dominance relation.

关 键 词：限制相似优势关系 集对优势关系 集对限制优势关系 集对优度 关键属性 

分 类 号：O225[理学—运筹学与控制论]