Researches On The Consistency Measures And Priority Weights Of Probabilistic Linguistic Preference Relations

Probabilistic linguistic term set(PLTS)is an extension of hesitant fuzzy linguistic term set(HFLTS).It is a new tool to describe uncertain linguistic information.By PLTS,the preference information is more in line with the uncertainty characteristics of human thinking,more accurately reflects the preferences of decision makers,and is more suitable for dealing with more complex decision-making problems.Therefore,this paper studies the eonsistency measure of probabilistie linguistic preference(PLPR).The main results are as follows:When the probability values are same,hesitant fuzzy linguistic preference relation(HFLPR)as the special form of PLPR can be obtained,then study the consistency measure of HFLPR.In this paper,a new definition of the consistency of HFLPR is proposed.Aiming at the problems in the existing consistency measure method for HFLPR,a consistency measure method based on hesitant goal programming model is proposed.For the consistent HFLPR,priority weights from HFLPR for ranking the alternatives can be directly derived.Using the consistency index of Linguistic preference relation(LPR)to further measure whether HFLPR satisfies satisfactory consistency.For the inconsistent preference relation,the model optimization method and the iterative optimization method are proposed.In order to illustrate the effectiveness of the proposed methods,this paper gives several examples including consistency,satisfactory consistency and consistency improvement,and compares with the existing research content of HFLPR.When the sum of the probability values is less than 1,this paper studies the consistency measure of preference relation in the general form of PLPR.Normalized probabilistic linguistic term set(NPLTS)is a probabilistic language that normalizes PLTS with incomplete probability information into interval form.Based on NPLTS,NPLPR is proposed,and the definition of expected consistency of PLPR is proposed.The target programming model is applied to test its consistency.For the consistent PLPR,the interval weights are obtained by the paired linear programming model as the final ranking weights,and the probabilistic linguistic geometric consistency index(PLGCI)is proposed to judge the satisfactory consistency of PLPR.In order to deal with the inconsistency,this paper proposes to improve the consistency by adjusting the probability information and linguistic terms in the preference relation.Combined with these two methods,the inconsistency correction algorithm of PLPR is proposed.And the effectiveness of the proposed methods is verified by several examples.Apply the methods proposed in this paper to the emergency decision support of fire accidents in petrochemical plants,and use HFLPR and PLPR to evaluate the emergency plans and give the results analysis.