Research On Uncertainty Multi-Attribute Decision Making Method Based On Evidential Reasoning With Certitude Degree

Decision science is closely related to system science and management science, it is the interdiscipline of natural science and social science. The uncertainty of decision making problem has been highlighted in the face of increasingly complicated decision environment. Under the uncertainty environment, how to analyze the subjective judgment of decision maker using quantitative tool is a worthy problem of research. As evidential reasoning is an uncertainty reasoning approach which aims at information fusion, it provides a new solution of uncertainty multi-attribute decision making problem. The concrete research ideas are: decision information and knowledge will be collected as evidences; based on the collected evidences, the fusion of information is given using uncertainty reasoning approach; then the priorities of alternatives can be obtained according to combination information.In the viewpoint of the application requirements, we introduce the certitude structure into the evidential reasoning. On the one hand, it conforms to human beings' language cognizance mode. On the other hand, it predigests the knowledge representation and knowledge reasoning algorithm. In the light of the research results existed in the fuzzy set theory, interval-valued certitude structure extends from certitude structure. This interval-valued certitude structure represents a variety of uncertain quantitative information and qualitative knowledge. With further research, uncertainty multi-attribute decision making methods and their applications are proposed.This dissertation includes five main research contents as follows:Firstly, considering that most of the decision information is uncertainty, the natural language expression method and computer language expression method is different, the certitude structure and certitude rule base are proposed. Then, two algorithms of evidential reasoning with certitude structure for information fusion, and their properties and relations are given. Further, the certitude rule base inference method using evidential reasoning with certitude degree is presented.Secondly, considering interval of certainty degree is easier to understand and acquired than exact values, certitude structure is promoted as interval-valued certitude structure. Then the interval-valued certitude structure transformation method is given. The random variable, fuzzy number, linguistic variable, incomplete information and so on can be transformed into interval certitude structure using the transformation method. Further, the evidential reasoning with interval-valued certitude degree and interval-valued certitude rule base inference method using evidential reasoning with interval-valued certitude degree are presented. These methods have extensive applicability, as the hybrid data with randomness, fuzziness and incompleteness can be modeled using interval-valued certitude rule base inference method.Thirdly, considering the decision making environment and decision making information are uncertainty, and the cognition of decision maker is not completely rational, the uncertainty multi-attribute decision making method with certitude structure is proposed. Certitude structure is used to describe attribute value, evidential reasoning with certitude degree is used to fuse decision information. According to the research results existed in the prospect theory and considering the uncertainty of states, the single-state multi-attribute decision making with certitude structure and multi-state multi-attribute decision making with certitude structure are proposed.Fourthly, the single-state multi-attribute decision making with interval-valued certitude structure and multi-state multi-attribute decision making with interval-valued certitude structure are proposed based on the evidential reasoning with interval-valued certitude degree and uncertainty multi-attribute decision making with certitude structure. The prospect theory is generalized from real number field to interval number field based on fuzzy set theory. In this case, the interval value function and interval weight function are presented, the using extent of prospect theory is expanded.Finally, concerning the issue that the radio interference has become increasingly complex and serious, as well as the difficulty of radio management is increasing, the aviation radio interference investigation intelligent decision support system is built based on the expertise, evidential reasoning with interval certitude degree and uncertainty multi-attribute decision making with interval certitude structure. This system is used to given the monitoring schemes and ranking of monitoring schemes according to the characterization of interference and monitoring result using evidential reasoning with interval certitude degree and uncertainty multi-attribute decision making with interval certitude structure. It is a support tool for finding out the interference source and interference reason accurately and quickly. In general, the aviation radio interference investigation intelligent decision support system has some practical value as the efficiency and effectiveness of radio management are improved.