Evidence Theory Fusion Method Based on Variable Granulation Rough Set

Abstract

Variable granulation rough set and D-S evidence theory are two kinds of complementary evidence fusion method. In this paper, we discuss the two theories in detail, and proposing an evidence fusion method based on variable granulation rough set according to the contact between the two theories. This method transforms the multi-source evidence fusion problem to the data analysis problem of multi-source information system. We use granulation space to represent evidence space, and use granulation distance to represent evidence distance, so that we can judge if there is a conflict between the evidences or not. And then the corresponding fusion rules are used to set the parameter β and carry on the evidence fusion respectively. At the end of the paper, an example of calculation shows that the fusion method is effectively.