Group Decision-Making Method From A Reliability Perspective And Its Application

In many complex real-world decision-making problems,it is quite difficult for decision-makers to get the accurate values due to the complex decision-making systems and numerous factors,the problems often have uncertainty.Then many uncertain information forms,such as interval number,fuzzy number,grey number,hesitant fuzzy number,cloud model and linguistic terms,have been widely applied.Among these forms,linguistic terms are not only closer to human's thinking habit,but also simple and flexible to express,thus,it is very conmmon in practice.To enhance the decision-making accuracy and comprehensiveness,group decision making gradually becomes a common decision mode instead of individual decision making.However,experts often have different opinions due to the differences in their status,background,knowledge and experience,so how to effectively solve group confilicts becomes a key part of relevant research.Many existing studies consider the expert difference from perspectives of similarity,consensus and consistency,which ignore the internal property of experts.Therefore,we mainly study the dynamic linguistic group decision-making problems from a reliability perspective.In the study,both group conflicts and interval property are considered to study the reliability from multiple aspects.In the study,the reliability and aggregation of linguistic group decision making are studied from simple issues to complex ones,making the study logic.The main contents of the thesis are presented as follows.(1)A linguistic group decision-making problem is studied from perspertive of the reliability of decision-makers.To the problem on reliability of self-selection,which is expressed in terms of linguistic variables,a method to measure reliablity is constructed based on linguistic information degree of greyness.According to group consensus,linguistic information degree of greyness and prior subjective information,a method to calculate comprehensive weights of experts is put forward.On the basis of aggregated linguistic information degree of greyness,the method to measure and analyze the reliability of decision making is established.Then the sorting algorithm cosidering linguistic information degree of greyness is put forward as well.(2)A probabilistic linguistic multi-attribute group decision-making problem is studied from a reliability perspective.First,two sets of information(i.e.,probabilistic linguistic terms on alternative grades,and linguistic terms on familiarity)are given by experts as measures of expert reliability and attribute reliability.A granulation optimization model is then constructed to generate the utility of grades while maximizing group similarity,and a model that redistributes the probabilities of grades is established for the situation of incomplete information.A method to measure the expert reliability is proposed by considering group similarity and familiarity,we further provide another granulation optimization model to generate the degree of familiarity by maximizing the reliability.Additionally,through using the evidential reasoning rule,the aggregated weights of experts and attributes are obtained employing an optimization method that is used to retrieve attribute information.Then the ranking order of laternatives can be obtained as well.(3)a large-scale group decision-making problem is studied from perspetives of classification,reliability and aggregation methods based on the hesitant fuzzy linguistic terms.Specifically,a method to measure the similarity between two hesitant linguistic sets is proposed by considering both hesitancy and distance.Moreover,the preliminary expert cluters are generated using the netting method based on a similarity matrix,and then an acceptable level is set to help realize a second clutering.Through a cllutering accuracy index,the final classes are obtained.Furthermore,two aggregating frames are constructed respectively for the information within one cluter and between the cluters.The probabilistic linguistic groups are obtained through combing the information within one cluter.On that basis,the degree of reliability of each cluter is calculated and the cluter weights are determined based on the cluter reliability and the percentage of expert number in one cluter to the total number to aggregate the information among cluters.Additionally,the expected values of the alternatives are calculated to make a selection.(4)A dynamic hesitant fuzzy linguistic group decision-making problem is studied from perspective of information reliability.First,an approach is applied to transform the dynamic hesitant fuzzy linguistic term sets into a set of proportional linguistic pairs to eliminate the time dimension.Second,information reliability of experts is measured from multiple dimensions,i.e.group similarity and degree of certainty,and an optimization method is employed to maximize the group similarity.Third,a measurement of the reliability of an alternative on each attribute is provided based on the degree of stability that is also obtained through the idea of information certainty.Then based on the combination rule which considers both weight and reliability of experts and attributes,the ranking order is obtained according to the aggregated values.(5)A dynamic large-scale group decision-making problem using hesitant fuzzy linguistic term sets(HFLTSs)is studied from perspectives of clustering,reliability and aggregation.First,considering the characteristics of HFLTSs,an expert similarity measurement is proposed by considering both expectation distance and hesitancy similarity between two HFLTSs.Second,multiple reference points are considered to make a selection,i.e.positive ideal reference point and development reference point.Then the reference similarity is measured by combing these two points.Third,an optimization model using the multi-attribute weight and multi-stage weight is established with the aim of maximizing the reference similarity of alternatives.Fourth,a fuzzy clustering method is applied to realize the expert clustering based on the expert similarity as well as the the proposed clustering validity index.In addition,a detailed aggregation process contains two steps,the first one is the aggragation process within one cluster which is considered of no difference(i.e.the experts have no difference),the second one is the aggregation process between clusters based on the cluster weight and reference reliability(i.e.the clusters have the difference).Finally,the ranking order is obtained according to the aggregated reference similarity.(6)A suppliers' selection and evaluation problem for steel products is studied on the background of low carbon economy.After analyzing the existing researches on low carbon economy,the suppliers' selection and evaluation with the consideration of low carbon,a relevant indicator system is built for the problem.Then giving the muti-stage hesitant fuzzy linguistic terms,the proposed method in the study is applied to evaluate the comprehensive performance of four steel product's suppliers with low carbon being included,with the aim to verify the rationality and feasibility.