The Methods Research And The Sensitivity Analysis In Interval Multiple Attribute Decision Making

The multiple attribute decision making refers to the choice of the most superior alternative or the ranking of alternatives when there are many attributes are considered. It is one important component of the modern decision science. Its theories and methods are widely applied to many areas such as engineering, technology, economics, management, military and so on. In recent years, the uncertainty multiple attribute decision making question have aroused many people's attention. And interval multiple attribute decision making is one of them.Because of the objective situation's complex polytropy, the uncertainty of subjective thinking, survey errors and so on, the decision information is often given in the form of interval numbers.Therefore,the study on interval multiple attribute decision making has important theoretic significance and higher practical application value.This paper mainly presents the analysis and the research on the theory of interval multiple attribute decision making methods.The main content of this paper is summarized as follows:(1)The present research of interval multiple attribute decision making question is summarized .The question of the existing decision methods and the necessity of further studies are analyzed, thus the main content of this paper is established.(2)An example is given to explan that a rank reversal problem exists in the method based on the concept of vector projection for interval multiple attribute decision making.This paper presents a modified method and proves the modified method is rank preserving.(3)The concept of the attribute importance is defined by referring to the idea of attribute importance in rought sets and a new method to definite attribute weight is proposed using the attribute importance.(4)The sensitivity to the weight and the attribute values are analyzed separately in the TOPSIS method. The necessary and sufficient conditions of weights and attributes variation result in ranking change are proposed.