An Evidence-Based Nonlinear Programming Model for MADA under Fuzzy Uncertainties

Abstract

Multiple-attribute decision analysis (MADA) problem consists of both quantitative data and qualitative information with subjective judgments of ambiguity. The evidential reasoning (ER) approach is well-suited for dealing with such MADA problems and can generate comprehensive distributed assessments for different alternatives. In the ER approach, many types of uncertainties could be considered such as imprecision, vagueness and fuzziness. In this paper, both triangular fuzzy weights and fuzzy utilities are proposed in the ER approach, which may be incurred in a lot of circumstances such as group decision making situation. Several pairs of ER based non-linear programming algorithms for MADA are conducted to generate the global fuzzy belief degrees of the assessed alternative. A case study based on R&D project assessment in a car manufacturer is implemented to illustrate the effectiveness of the proposed approach.