Three-way Multi-attribute Behavior Decision-making Research On Multi-scale Information Systems

In the real-life decision-making environment and practice,the psychological behavior of individuals can have some influence on the decision-making process,traditional multi-attribute decision-making approaches based on utility decisions have some deviations from actual decisions.Behavioral decision theory improves on traditional decision theory and has wide applicability.Regret theory,as one of the important components of behavioral decision theory,has received extensive attention in theory and applications.In addition,the traditional multi-attribute decision-making approach requires decision-makers to discriminate decision objects into two categories of acceptance or rejection,which is too strict from a practical point of view,thus making it difficult to reflect the risk attitude and psychological behavior of decision-makers.In this paper,the proposed generalized three-way decision model for multi-scale information systems combines three-way decision thoery and regret theory,which can improve the shortcomings of these two aspects.The details are as follow two aspects:(1)A generalized three-way multi-attribute decision model based on regret theory is established on a multi-scale information system.First,the multi-scale evaluation information is converted into numerical evaluation information using the conversion function,and then the multi-scale information system is integrated according to the scale characteristics,regret theory is used to build a regret-rejoicing index on the integrated multi-scale information system to construct a outranking relation between decision objects;Secondly,the objects in the outranking relation are classified into three different domains by the concepts of clustering and neighborhood,the ranking of the decision objects in each domain can also be calculated by using the posting coefficient.Finally,we use cases in the database to simulate the experiment to verify the decision effect of the model.Comparative analysis and experimental analysis also show the effectiveness,superiority and stability of the proposed model.(2)A generalized three-way multi-attribute decision-making method based on regret theory is developed for incomplete multi-scale decision information systems with interval fuzzy numbers.Firstly,for the incomplete multi-scale information system,we select an incomplete optimal sub-system,and then the multi-scale evaluation information is converted into an interval fuzzy number by using a linguistic terms set.Secondly,based on the probability distribution of evaluation values and trade-off factors,we propose a target-dependent approximation estimation method for the incomplete interval fuzzy sub-system.Then,the regret-rejoicing preferences between objects are obtained.Finally,a tri-partition method based on a max-bipartition is established,and the ranking method is also obtained by a preference index.Further,comparative and parametric analyses show the good effectiveness and stability of our proposed method.