Misuses of Clustering and Absolute Scales of Measurement in the Analytic Hierarchy process

Kirti Peniwati
PPM Graduate School of Management

Publication date: Aug, 1994

Journal: The 3rd International Symposium on the Analytic Hierarchy Process

Publisher: Creative Decisions Foundation

Abstract: The grouping of elements, called clustering, is a fundamental operation in the Analytic Hierarchy Process. What elements are chosen and how they are clustered and located in a hierarchy have a profound effect on the outcome and the conclusions derived from the outcome. A valid AHP hierarchy must represent a problem well, and also ensure that all AHP axioms are satisfied. Examples are given to show how violations of the AHP axioms can lead to wrong priorities, infeasible interpretations, and erroneous conclusions. I also show that .a mechanical approach to priority setting can mislead one to believe that the theory needs to be modified to accommodate one's commitment tc an arithmetic procedure rather than the other way around.

Keywords: AHP, MCDM, Clustering, Absolute Measurement, Relative Measurement