A generalization of jeffrey’s rule in the interval-valued dempster-shafer framework

Jeffrey’s rule of conditioning is an effective tool to update the current information under the given information. However, traditional Jeffrey’s rule can only process information under the framework of probability theory. So we generalize it based on Dempster-Shafer evidence theory which was seen as a generalization of probability in this paper. In this generalization, interval-valued prior and conditional probability which satisfies weaker conditions is joined with the basis of original Jeffrey’s rule and Dempster-Shafer evidence theory. And we then achieve the update of information by building an optimization model under interval prior and conditional probability. We achieve comparison of interval-valued belief degree with the basis of TOPSIS. One of the main advantages of this generalization is its ability to handle information with wider imperfections. Finally, we demonstrate the application of our generalization on an example of multi-criteria decision-making.