The Research On Inference System Based On D-S Evidence Theory

The D-S evidence theory, which provides a method of efficiently solving uncertain information reasoning and integration of the multiple data source in practical application, can synthesize independent evidence and gives consistent results, in addition, it also can dealing with synthesis problems of fuzzy and uncertain information in order to achieve complementary information. Compared with other reasoning methods, this evidence theory is much closer to the decision process of human-beings.With the theory of evidence being used in data fusion, target recognition, decision-making, predicting and artificial intelligence, how to construct the network of reasoning in practical application according to the data and the rule and how to combine conflicting evidence becomes the focus of study of the theory of evidence. This thesis has made the following contributions:(1)The latest theory and application advances are summarized. The essence of each concept is analyzed.The problems in the theory of evidence are also described and analyzed.(2)How to apply the theory of evidence to the field of decision-making is detailed described. A model based on the theory of evidence is presented according to the DSS. A decision method which based on DS evidence theory has been proposed, and its effectiveness and feasibility has also been proved according to the experimental results.(3)The problem of conflicting evidence is studied which normally exists in the application of the theory of evidence. Several classical improved-methods are analyzed and compared. On the base of these methods, a new method of combine evidence is presented. The example proves that the solution satisfies the requirements and the method is in line with the actual field.