Multi-Attribute Group Decision Making Method For Linguistic Distribution Assessments Considering Personalized Individual Semantics

Linguistic distribution assessments as a tool for dealing with uncertain language problems not only allows decision makers to express evaluative information using multiple linguistic terms,but the symbol proportion also reflects the importance of linguistic terms.At the same time,personalized individual semantics can quantify different decision-makers’understanding of the numerical meaning of linguistic terms.In this paper,we propose two methods to derive the personalized individual semantics of decision makers in the multiattribute decision making with linguistic distribution assessments as the linguistic form,and study the consensus problem of multi-attribute group decision making with linguistic distribution assessments based on personalized individual semantics,and establish a dynamic personalized feedback optimization mechanism based on minimum adjustment,considering the influence of personalized individual semantic functions on the consensus behavior of decision makers.In the first chapter,we introduce the research background and related concepts of linguistic distribution assessments,personalized individual semantics and multi-attribute group decision consensus.In the second chapter,two methods are proposed to determine the personalized individual semantic functions of decision makers for multi-attribute decision problems with preference information,under the condition that the subjective preference information is consistent with the objective decision information as much as possible.Method 1 is the product consistency method,which takes advantage of the uniqueness of the objective ranking vector corresponding to the fuzzy judgment matrix with product consistency to construct the objective model to derive the personalized individual semantic functions of the group experts.The second method is the eigenvector method,which uses the characteristic root property of the fuzzy judgment matrix to construct a model to derive a personalized individual semantic functions using the eigenvector as a subjective ranking vector about the alternatives.Finally,the method proposed in this paper is compared and analyzed with existing research methods and applied to the problem of rice evaluation.In the third chapter,the problem of consensus in group decision making for linguistic distribution assessments based on personalized individual semantics is studied.Firstly,a new distance metric is given for the linguistic distribution assessments considering personalized individual semantics,reflecting the cumulative change in the product of the neighboring symbol proportion and the corresponding semantics.Secondly,a boundary trust threshold is used to define acceptable and unacceptable levels,which reflect the similarity and dissimilarity between the evaluation information,and a new consensus measure is defined in conjunction with uninorm aggregation operators.Then,the influence of personalized individual semantic functions on the consensus behavior of experts is considered,and the dynamic personalized feedback mechanism under the minimum adjustment objective is constructed by combining the attitude consensus threshold and the personalized feedback parameters of experts.Finally,the proposed consensus method is applied to solve practical problems.