Approaches To Conflict Evidence In D-S Evidence Theory And Its Applications

The D-S evidence theory, with its advantages in expression, measurement and combination of uncertainty, is thought to be quite important. The application of evidence reasoning is extended to the fields of probability and fuzzy set, because it has been improving itself and trying to take advantage of the strengths of other theories. It can not only use priori information as Beyes Reasoning does, but also deal with fuzzy evidence like languages. As to its applications, the expert system and data fusion can use the evidence theory at different levels, and the results can be quite satisfactory.With the development of the evidence theory, the prospects of its applications are even better. Issues like how to deal with conflict evidence and how to set mass function have become the focus in its theoretical and applied research. Some research work on the theoretical and applied issues of the D-S evidence theory has been done in this thesis. The main research work is as follows:Firstly, various concepts and their implications in the evidence theory and the latest developments in this field are introduced. The existing problems in the theoretical and applied research of the evidence theory are analyzed.Secondly, the conflict evidence issue and its treatment are studied, the existing classical methods of dealing with conflicts are analyzed, and some analyses and comparisons of these methods are done. Based on the research above, a new method of averaging the weighted evidences before combining them is put forward. The new method is then used to deal with some examples of conflict evidence to test its effectiveness in this aspect.Thirdly, the applications of the evidence theory in the comprehensive evaluation of the teaching quality are studied.Finally, the whole thesis is summarized and some issues worthy of studying and exploring are mentioned.