A Revised Evidential Reasoning Model Based on Interval Information

The basic theory of the evidential reasoning (ER) approach is firstly introduced. Since many uncertainties and ambiguities are included in a variety of real decision making problems, accurate number is not competent for the description of the assessment value any more. Some ER models in the circumstance of interval belief degrees or interval weights are then proposed by some people. The analytical ER algorithm which is not appropriate is used in these models. In this paper, the utility of each evaluation grade is extended from crisp value to interval value, and a revised ER programming model based on interval belief degree and interval utility is constructed. In this model, the recursive ER algorithm is used to substitute for the analytical ER algorithm. The effectiveness of this revised ER programming model is also illustrated in the end through a case study.