基于EMD提取专家语言评价信息的群决策方法  

The method of group decision making based on the EMD extracting experts linguistic assessment information

作　　者：周任军[1] 陈瑞先 陈跃辉[3] 陆佳政[3] 周胜瑜[3] 刘乐平[3] 

机构地区：[1]智能电网运行与控制湖南省重点实验室(长沙理工大学),长沙410004 [2]国网山东省电力公司烟台供电公司,烟台264000 [3]国家电网湖南省电力公司,长沙410007

出　　处：《系统工程理论与实践》2016年第3期743-749,共7页Systems Engineering-Theory & Practice

基　　金：国家自然科学基金(51277016;71371065;11171095)~~

摘　　要：针对传统语言群决策方法在集成专家评价信息时所遇到的专家权重求取主观性大、方法无标准的问题,提出一种基于经验模态分解(empirical mode decomposition,EMD)提取专家语言评价信息的群决策方法,利用EMD方法适于分解非线性、非平稳信号数列的特点,分解各专家对待决策事物做出的综合语言评价值,得到各专家所做评价值的客观趋势成分和主观随机成分,以客观趋势成分的均值,作为待决策事物的专家综合评价结果用于优选过程.相比于传统群决策理论中专家信息的集成方式,其由于不用求取专家权重而直接进行评价信息提取,因此能更大限度地排除人为主观因素的影响,使最终决策结果更加符合对客观事物的评价规律.Considering the defects of traditional language group decision-making method, such as the different subjective weight and judgment scale from each expert, it is proposed that a method of group decision-making based on the empirical mode decomposition extracting experts＇ linguistic assessment infor- mation. This method decomposes each expert＇s comprehensive linguistic assessment value into subjective random components and objective trend components by feature of EMD, suitable for decomposing nonlin- ear, non-stationary signal column. The mean value of objective trend components act as the last expert＇s comprehensive evaluation result. Compared to the traditional method, it extracts evaluate information directly without obtaining weight of expert, to eliminate the subjective factors influence to a greater extent. Then the decision-making consequence is more coincident with the evaluation laws of obiects.

关 键 词：语言评价信息 语言群决策方法 经验模态分解(EMD) 主观随机成分 客观趋势成分 

分 类 号：C934[经济管理—管理学]