摘要
级别不低于方法是一类解决多属性决策问题的重要方法。相对于其他方法力图建立可行方案集上的完全序,级别不低于方法所构建的是一种较弱的二元关系——级别不低于关系,它所要求的条件较弱,而得到的结果更为可靠,是近年来研究最为活跃、应用十分广泛的方法之一。本文将级别不低于方法划分为ELECTRE、PROMETHEE和PCCA等三大类分别进行了归纳和总结;介绍了近年来级别不低于方法的新发展,对级别不低于方法的公理化,不确定信息下的级别不低于方法以及多属性决策的敏感性分析和鲁棒性分析等问题给出了一些有意义的讨论。特别是对鲁棒性的定义和分析方法提出了一条切实可行的思路。
Outranking methods are one type of important methods for solving multiple attributes decision making problems. Compared to other methods who engage in constructing complete orders in feasible alternative set, outranking methods present one more relax binary relation outranking relation. This relation produces more reliable result, and conducts one kind of the most active and extensive methods, though it dosen＇t need too strong conditions. In this paper, outranking methods are divided into three types and summarized separately, such as ELECTRE, PROMETHEE and PCCA etc. Some new development on outranking methods in recent years are introduced, in which, axiomatic approach of outranking methods, outranking methods with uncertainty information, sensitivity analysis and robustness analysis on multiple attributes decision making are discussed significatively. In particular, a practical way is provided in the robustness definition and analysis method.
出处
《管理工程学报》
CSSCI
2008年第3期41-45,56,共6页
Journal of Industrial Engineering and Engineering Management
基金
国防“十五”预先研究基金资助项目(413040201)
关键词
级别不低于方法
综述
多属性决策
公理化方法
鲁棒性
模糊
outranking method
summary
multiple attributes decision making
axiomatic approach
robustness
fuzzy
