An approach to aggregating interval weights for hierarchical multiple criteria decision making

Priority weights derived from uncertain preference relations are often characterized by interval weights. A key stage in multiple criteria decision making (MCDM) with a hierarchical structure is to aggregate local interval weights into global interval weights. In this paper, a pair of linear programming models is devised to maximize the lower and upper bounds of the aggregated interval value where criteria weights are independently treated as decision variables for each alternative. By integrating a number of optimization models, a linear program is developed to obtain a unified criteria weight vector. The criteria weights are subsequently incorporated to aggregate local interval weighs into global interval weights for final alternative ranking. A numerical example and analysis is presented to show the feasibility and effectiveness of the proposed models.