Stochastic MADM Analysis based on Stochastic Dominance Degrees

The aim of this paper is to put forward an approach for stochastic multi-attribute decision making (MADM) problems with stochastic information, in which the attribute values are in the format of probability distributions. In the approach, we propose concepts and computation formulas of stochastic dominance degrees, which are utilized to measure the degrees of stochastic dominance for alternative pairs with respect to all attributes. Then concordance stochastic dominance degrees are used to aggregate different kinds of stochastic dominance degrees. Thus unified stochastic dominance degrees for all alternatives are calculated and a ranking of alternatives or selection of the most desirable alternative(s) is obtained. Finally, a numerical example is also used to illustrate the applicability and effectiveness of the proposed approach.