A Trust Based Interactive Group Consensus Method For Multi-attribute Group Decision Making

The rapid development of Web 2.0 has made the information exchange between people more and more convenient,the interaction activities also increased.The trust relationship and decision-making opinions in groups spread through social networks,and the networked trust relationship in groups is more significant.When multiple users participate in group decision making(management decisions,etc.)based on professional knowledge or common goals,group decision behaviors exhibit social network characteristics.Group decision-making based on social network analysis methods has been widely concerned by the academic community.In a social network,there is a certain trust relationship between multiple users.This kind of trust relationship not only affects individuals,but also influences group decision-making behavior.Group consensus is a research hotspot for multi-attribute group decision making.Due to the differences in knowledge fields and professional backgrounds,multi-attribute group decision-making experts have certain inconsistencies in decision-making opinions.In group decisions,inconsistencies should exist within a reasonable range.Excessive differences may lead to unreliable decision-making results and expert dissatisfaction.Finding the optimal integration strategy within the reasonable consensus of the group is the goal of multi-attribute group decision-making.The traditional group interaction consensus mechanism does not consider the trust relationship between experts,so it is unreasonable to force consensus experts to accept the opinions of each expert on average.Therefore,how to generate reasonable consensus recommendation and reach group consensus by constructing trust network is the key problem that this study tries to solve.On the basis of analyzing the characteristics of trust network in multi-attribute group decision problem,the trust network between experts is constructed,and a more reasonable trust based interactive group consensus method for multi-attribute group decision making method is proposed.In order to verify the rationality and superiority of the method,the case study of the cold chain logistics cloud platform selection runs through the whole research process.The visual consensus process and the mathematical proof to illustrate the rationality of the proposed method are given in the group interaction consensus process,and finally the decision result is obtained.The main research results of this paper are as follows:(1)Expression of trust.This paper analyzes and contrasts the traditional fuzzy expression of trust and introduces the meaning of trust scores.Based on this meaning,the concept of knowledge coverage is proposed.The knowledge coverage includes two parts: incomplete and contradictory fuzzy information.Its influence on the results of trust propagation cannot be ignored.Based on the correlation coefficient,the ranking method of trust scores is proposed.In addition,the concept of trust network based on trust scoring is proposed.(2)A trust network construction method based on trust propagation is proposed.In a trusted network,due to the lack of trust information,it usually appears as an incomplete network.In order to estimate the missing trust information,this study proposes a single path trust propagation operator,which considers the experts' knowledge coverage and analyzes its expected proposition.In order to aggregrate multi-path trust information,a trust aggragation method based on penalty for the length of trust path is proposed.Thereby constructing a complete trust network and providing a basis for the trust relationship of the group consensus recommendation process.(3)Trust-based interactive group consensus mechanisms are proposed.Traditional consensus recommendation methods force experts to averagly accept the recommendation of consistency experts.This study proposes two trust-based consensus recommendation methods.The first method indicates accepts the recommendation of some trust relationship experts based on the objective trust relationship,the method effectively reduces the cost of consensus interaction.Another approach to recommend for inconsistency element in decision matrix is based on a complete trust network obtained by trust propagation and the degree of consistency is objective trust.The subjective and objective trust ratios are adjusted by subjective and objective adjustment parameters,so that experts can accept consensus recommendations by the inducing of subjective and objective trust.(4)A selection process based on Einstein operator is proposed.Aiming at resolving the decision information aggregation problem expressed by,the Trust Score Einstein weighted average(TFEWA)operator and the Trust Score Einstein ordered weighted average(TFEOWA)are proposed and their expected propositions are analyzed.Thereby construct an optimal selection method of MAGDM in this study.The trust-based interactive group consensus method for multi-attribute group decision-making method proposed in this study enriches the theory of extension of group consensus and multi-attribute group decision.What's more,it provides methods to support group decision-making problems such as management decision-making and consumer behavior decision-making under the development of social media.