Research On Multi-Attribute Group Decision Making Approaches With Multi-Granular Hesitant Fuzzy Linguistic Information

Multi-attribute group decision making(MAGDM)is an important branch in the field of modern management science,which exists widely in politics,economy,culture and other fields of human society.Due to the complexity of the world,the uncertainty of the decision making environment and the vagueness of experts’ thinking,experts usually tend to provide qualitative assessments over alternatives using natural languages.For some complex decision making problems with high uncertainty,experts may hesitate between some linguistic terms due to the lack of experience,ability and knowledge when providing their linguistic assessments.In order to better characterize the fuzziness and hesitancy of experts’ linguistic assessments,Rodriguez et al.introduced the concept of hesitant fuzzy linguistic term set(HFLTS).Although great progresses have been made about models and approaches for MAGDM with HFLTSs in recent years,there are still some problems that need to be tackled:(1)In group decision making problems,due to the differences in knowledge and culture backgrounds,experts may use multi-granular linguistic term sets to provide his/her hesitant fuzzy linguistic assessments.Moreover,in some situations,experts may also provide HFLTSs defined on unbalanced linguistic term sets.(2)Consensus reaching process plays an important role in MAGDM.However,most existing consensus models about MAGDM based on HFLTSs do not take the weight vectors of experts and attributes,adjustment cost and interpretability of adjustment advice into account.(3)The existing alternative ranking methods usually need to transform multi-granular balanced HFLTSs into linguistic information defined on the same linguistic term set which may result in information loss.Moreover,experts’ psychological behaviors,such as reference dependence and loss aversion have not been fully considered for MAGDM with HFLTSs.(4)The existing studies about large-scale group decision making(LGDM)cannot deal with the situations when different types of multigranular HFLTSs are provided by experts,and usually neglect the interpretability of the decision results.All these problems bring new challenges for MAGDM with HFLTSs.To solve the above problems,the aim of this thesis is to develop some novel models and approaches for MAGDM with multi-granular HFLTSs.The main work of this thesis is summarized as follows.(1)Methods for MAGDM with multi-granular balanced HFLTSs.First,for MAGDM problems with multi-granular balanced HFLTSs,the fuzzy envelope method is first used to unify multi-granular HFLTSs provided by experts and then the individual consensus levels and the group consensus level are measured by considering the weight vectors of experts and attributes.On this basis,an optimization model which aims to minimize the overall adjustment amounts of experts’ assessments is established and a consensus reaching algorithm is further devised.Second,an approach is developed to calculate the individual weighted rank frequency matrix and collective weighted rank frequency matrix for all alternatives and an assignment model is then established to derive the collective ranking of alternatives.Finally,a numerical example for healthcare waste treatment technology selection and some comparative analyses are provided to demonstrate the feasibility and effectiveness of the proposed method.(2)Methods for MAGDM with multi-granular unbalanced HFLTSs.First,the multi-granular unbalanced hesitant fuzzy linguistic weighted averaging operator is developed to fuse multigranular unbalanced HFLTSs.Afterwards,for MAGDM problems with multi-granular unbalanced HFLTSs,the consensus measures are defined on the element level,the alternative level and the decision matrix level and a linguistic distribution-based feedback adjustment mechanism is then designed,based on which an interactive consensus reaching algorithm is proposed.Second,by considering different relationships between two HFLTSs,it is proposed some new formulae to calculate the gain and loss for unbalanced HFLTSs.On this basis,a TODIM-based alternative ranking method is developed for MAGDM problems with multi-granular unbalanced HFLTSs,which takes the experts’ psychological behaviors,such as reference dependence and loss aversion into account.Finally,a numerical example for old town reconstruction scheme selection is provided.Detailed simulation analysis and comparative analysis are utilized to demonstrate the feasibility and effectiveness of the proposed method.Moreover,the characteristics of the proposed method are discussed.(3)Methods for multi-attribute LGDM with multi-granular HFLTSs.First,two algorithms are developed to transform an unbalanced HFLTS into a balanced linguistic distribution and to transform a balanced linguistic distribution into an unbalanced linguistic distribution,respectively.On this basis,for multi-attribute LGDM problems with different types of multi-granular HFLTSs,some linguistic distribution-based methods are proposed to unify experts’ linguistic assessments and cluster experts.Subsequently,a minimum adjustment consensus optimization model with accurate constraints is established to help expert clusters reach consensus.Furthermore,it is proposed a method to aggregate experts’ decision matrices and select the optimal alternative.To provide interpretable results to experts,a retranslation model for alternative collective assessments based on accurate constraints is developed.Finally,the feasibility and effectiveness of the proposed method are demonstrated by using an example for the selection of subway lines.To summarize,this study enriches the theories and methods about computing with words and group decision making under hesitant fuzzy linguistic environment.The proposed decision models and methods can be applied to deal with practical decision making problems,such as supplier selection,enterprise investment selection and healthcare waste treatment technology selection.