Approximate articulation of preference and priority derivation

Abstract

This paper presents an approach aimed at expanding the capabilities of encoding and processing of preference articulation in the Analytic Hierachy Process (AHP). Assessment of preference in the AHP is done by asking pairwise comparison questions about strength of preference between subjects of comparison. If preference can be articulated by a single value taken from the 1–9 preference scale, then one proceeds to fill a comparison matrix and derives the priority vector given as the principal eigenvector of that matrix. While asking the decision maker to provide pairwise preference statements is easy to implement in most cases, this process may be met with some resistance in other cases. This resistance on the part of the decision maker does not necessarily stem from his reluctance to apply the AHP, or any other analytical approach, but rather reflects in most cases his own uncertainty as to the correct level of inetensity to be assigned to a particular preference question. In these cases one may be motivated to consider approaches where the decision maker is allowed to state preference as a range of scale values. This paper explores the question of approximate articulation of preference values encountered when the decision maker prefers to state his strength of preference as a range of scale value rather than a single ‘precise’ value.