Abstract:

The multiattribute decision making result is influenced directly by how to select weights. Sensitivity analysis is an efficient method for processing weight uncertainty. A new kind of sensitivity analysis model on the weight in ELECTREⅢ is constructed. In this model, the weights that change least and make certain alternative rank first are obtained by a classical linear programming method. By this way the question how the rank order changes after the weights over a stable interval is answered. The analysis result proves that this method is valuable for helping decision makers to adjust the weights and reach a reasonable result in practice.