Research On Methods Of Linguistic Z-num Ber Multi-attribute Decision-Making Based On Cloud Model

Z-number is a further fuzzy concept on the basis of random number and fuzzy number,so it has both fuzziness and randomness.At present,the research on the fuzziness of Z-number has been relatively mature,but the research on its randomness still needs to be improved.Aiming at the most widely used linguistic Z-number,in this thesis,two methods of transforming linguistic Z-number into normal cloud model which has strong ability to express randomness are proposed,and the methods are applied to the multi-attribute decision-making problems in linguistic Z-number environment,The main works are as follows:(1)Based on the impact of decision makers' risk preferences and semantic coverage of linguistic terms on linguistic evaluation scales,two new types of linguistic scale functions are proposed.The transformation method of linguistic term to normal cloud model based on linguistic scale functions is defined,and on this basis,the indirect synthesis model of linguistic Z-number transformed into normal cloud is completed.In order to discuss the dominance relationship between the two cloud models,the dominance degree measurement between the two cloud models is defined.In this thesis,the traditional PROMETHEE method is applied to the multi-attribute decision-making problems in composite cloud environment,and the feasibility and effectiveness of the model are demonstrated through an example analysis.(2)A multi-objective optimization model is proposed to improve the underlying probabilistic solving model of discrete Z-number.The traditional backward cloud generator is improved into a probabilistic backward cloud generator,and based on it,the direct transformation of linguistic Z-number into normal cloud model is completed.For the multi-attribute decision-making problems in generated cloud environment,the traditional TODIM decisionmaking method is improved by replacing the distance measure with the dominance measure.Finally,the rationality of the model is proved by an example of human resource evaluation.(3)Reference alternatives with overall evaluation are introduced,by exploring the relationship between the attributes of the reference alternatives and the overall evaluation,the attribute set that is most relevant to the overall evaluation and has the minimum redundancy between the attributes is selected.And the attribute weight is assigned according to the correlation between the selected attributes and the overall evaluation of reference alternatives.In group decision-making problems,an expert weight model is established based on the correlation between expert evaluation matrices.Finally,combined with the improved TODIM method proposed above,the effectiveness of the proposed method is demonstrated through a case of multi-attribute group decision-making in linguistic Z-number environment.