考虑个性化语义和一致性的群体决策共识模型  

A consensus-reaching model for group decision making that considers personalized individual semantics and consistency

作　　者：张震[1] 李卓琳 ZHANG Zhen;LI Zhuoin(Institute of Systems Engineering,Dalian University of Technology,Dalian 116024,China)

机构地区：[1]大连理工大学系统工程研究所,辽宁大连116024

出　　处：《管理工程学报》2023年第6期157-168,共12页Journal of Industrial Engineering and Engineering Management

基　　金：国家自然科学基金资助项目(71971039、71731003);大连市科技创新基金项目(2018RQ69)。

摘　　要：本文针对专家给出语言偏好关系的群体决策问题,考虑不同专家语言术语的个性化语义和个体一致性,提出了一种群体决策共识模型。首先,基于一致性驱动的方法论确定专家语言术语的数值标度,将专家的个体语言偏好关系转化为模糊偏好关系,并通过计算个体模糊偏好关系与群体模糊偏好关系的相似度对群体共识度进行度量。为辅助不满足个体共识水平的专家调整其语言偏好关系,建立了考虑一致性控制的混合0-1线性规划模型来确定其语言偏好关系元素的调整范围。在此基础上,提出了考虑个性化语义和一致性控制的共识达成算法。最后,以在线教育平台选择问题为例说明了所提算法的可行性,并设计仿真实验验证了所提算法的有效性。实验结果表明,本文所提算法可以在保证个体一致性不被破坏的前提下提升群体共识水平。Group decision making(GDM)refers to the selection process of the best alternative among several alternative options based on preference information of multiple experts.This is widely practiced in areas of economic,political and cultural fields in the society.Given the high uncertainty surrounding the decision-making environment and the ambiguity of human cognition,experts usually tend to use linguistic terms to express their preferences among alternatives during GDM.This type of problem is called a linguistic GDM problem.As a common way of expressing preference information,linguistic preference relations allow experts to use linguistic terms to provide their preference information among alternatives through pairwise comparisons in linguistic GDM.For this type of GDM problem,experts may provide inconsistent linguistic preference relations due to the lack of knowledge and ability,leading to poor decisions.In addition,different experts hold different preferences which makes it difficult for a group to reach consensus,which reduces experts′degree of satisfaction with decisions achieved.Therefore,in linguistic GDM it is necessary to simultaneously consider individual consistency and group consensus of experts′linguistic preference relations to achieve decision results that are accepted by most experts.Furthermore,when dealing with linguistic GDM problems,linguistic information is usually transformed into numerical information using computing with words models.However,it should be noted that words mean different things for different people.For instance,when reviewing a paper,two experts may both evaluate a paper as“good”,however,one may consider“good”to be equivalent to a score of 85,while the other considers“good”to be equivalent to a score of 80.Recently,significant research attention has been directed towards understanding individual cognitive differences among experts in relation to linguistic GDM problems and how to model personalized individual semantics(PISs)of experts′linguistic information.Fo

关 键 词：群体决策 语言决策 共识达成过程 一致性 个性化语义 

分 类 号：C934[经济管理—管理学]