基于粗集理论的多属性与多准则决策分析综述  

A survey on multi-attribute and multi-criteria decision analysis based on rough sets theory

作　　者：安利平[1] 仝凌云[2] 

机构地区：[1]南开大学商学院,天津300071 [2]河北工业大学管理学院,天津300130

出　　处：《云南大学学报（自然科学版）》2009年第S1期177-181,共5页Journal of Yunnan University(Natural Sciences Edition)

基　　金：国家自然科学基金资助项目(70601013)

摘　　要：多属性与多准则决策分析(MA&MCDA)的方法涉及多种学科,包括统计学、人工智能和运筹学等.粗集理论被证明是MA&MCDA的有利工具.从二元关系的角度对基于粗集理论的各种MA&MCDA方法进行了回顾与总结.指出了在该领域需要进一步研究的方向,即,基于粗集理论的MA&MCDA的框架体系研究,包括建立概念间的联系、建模过程研究、提出评价偏好模型的指标及其优化策略等;基于粗集理论的建模方法研究,这需要综合运用多种智能技术,采用问题机理分析与数据分析、归纳学习与分析学习技术相结合的研究方法;基于粗集理论的群决策分析研究,从而解决多个决策者决策策略互不相容的问题.Methodologies for addressing multi-attribute and multi-criteria decision analysis (MA&MCDA) problems have been developed from a variety of research disciplines,including statistics,artificial intelligence and operations research.Rough set theory (RST) has proved to be a useful tool for analysis of a large class of MA&MCDA problems.The objective of this article is to review different forms of MA&MCDA models based on RST from the perspective of binary relations.And,we point out some rough set based research directions in this field as follows:First,the research framework for MA&MCDA based on RST is to be established,including concepts and their relations,modeling process,indices for evaluating the model,and strategies for optimizing the indices.Second,a suite of methodology of modeling MA&MCDA problems is to be put forward.In the modeling process,several artificial intelligence technologies is to be utilized; problem mechanism analysis is combined with data analysis; and induction learning is combined with deduction learning.Lastly,in order to detect the concordant and discordant parts of the decision makers,group decision analysis based on RST is to be researched.

关 键 词：多属性决策 多准则决策 粗集 二元关系 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]