A preference-based interpretation of AHP

The relationship between MAUT and AHP for three-level hierarchic structures is demonstrated based on a common framework interpreting multiattribute decision making techniques. A theorem showing the conditions under which two multiattribute decision making techniques result in a consistent preference structure is proved. It can be justified based on this theorem that the rank reversal problem in AHP resulting from the addition or deletion of alternatives is caused by multiplying inappropriate criteria priorities with local priorities for alternatives. A scaling method, AHP', is introduced synthesizing judgments from pairwise comparisons within and among criteria into value scales in MAUT.