The meanings of trade-offs in multiattribute evaluation methods: a comparison

Multiattribute evaluation methods are being used without full understanding of the meanings of the weights which characterize the relative importance of the attributes. The weights in the methods do not carry the same meaning, even though the questions asked in the procedures are similar. The weights imply the trade-offs among attributes with respect to certain transformed measures, and the decisionmaker is presumed to be able to specify the appropriate trade-off judgments based on preferences. Five multiattribute evaluation methods are compared deductively based on a common framework in which the meanings of weights are clarified: multiattribute utility theory (MUT), weighting and rating, the analytic hierarchy process, concordance analysis, and computation of equivalent alternatives. For each method, the evaluation procedures are described in mathematical terms and weights are derived and interpreted by means of the preference structure represented by MUT. Transformations between the methods make clear the meanings of weights and make possible iterative procedures for conducting empirical comparisons of the methods.