Evidential reasoning with discrete belief structures

In real-world applications, apart from being precise-valued or interval-valued, belief structure measurements can also be discrete-valued. However, problems relating to the combination of discrete-valued belief structures have not been resolved. Therefore, in the research presented in this paper, we explore the counterintuitive behavior associated with the combination of discrete evidence and extend the concept of evidential reasoning (ER) to evidential reasoning with a discrete structure in order to serve as the theoretical basis and as technical support for the fusion of discrete information. This method offers an approach to the normalization of discrete evidence, provides a means of objectively determining the weight of discrete evidence, and optimizes the combination of discrete evidence based on evidential reasoning. The results of various examples show that the method not only offers an effective solution to the combination of non-conflicting discrete evidence, but it also overcomes the counterintuitive results of combining internally or externally conflicting evidence.