A Study On Multiple Attribute Decision Making Based On D Numbers Theory Under Uncertain Environment

Multiple attribute decision making(MADM)has been a hot issue that widely used in military,industrial engineering and economics.D numbers theory,as the extension of Dempster-Shafer theory of evidence,has been used in MADM widely in recently years.However,some deficiencies have been proved in D numbers theory by recent research articles.In this paper,the study on MADM is conducted under uncertain environment based on D numbers theory which prefects the theory as well as dealing with the optimal path selection problems together with adaptive amoeba algorithm.The whole paper has been divided into five chapters.The first chapter introduces the application background and research status of Dempster-Shafer theory of evidence,D numbers theory and MADM,the main work done in this thesis and some preliminary knowledge are also introduced in this chapter.Chapter two analyzes the deficiency in D numbers' combination rule.Two data-driven combination rules based on the assessment grades and the credibility are proposed respectively.The comparisons of the numerical simulation results to the previous articles have proved the effectiveness of the proposed method when dealing with MADM.In the third chapter,the problem of the D numbers' integration representation is studied.A modified D numbers' integration representation is given and a real example of assessing the performance of four motorcycles is analyzed,the results show the modified method has improved recognizable when making decisions.In the forth chapter,some real applications about optimal path selection problem based on D numbers theory are discussed.Some mathematical models which combine D numbers theory with adaptive amoeba algorithm are established to handle the road information of conflicting or incomplete weights and the restricted passing road,the numerical simulations show the proposed method can find out the optimal path effectively and lower complexity is involved in the proposed method.In chapter five,some conclusions on the studies in this paper as well as the prospect of the research in the future are discussed.