Ranking with Statistical Variance Procedure Based Analytic Hierarchy Process

This study introduces an objective multicriteria ranking method based on the Analytic Hierarchy Process (AHP). Different multicriteria decision analysis methods generate different solutions for the same ranking problem because of their varying mathematical models. In AHP, decision makers construct positive comparison matrices from their preferences by using a scale of 1-9. However, even a simple ranking problem requires numerous comparison matrices while subjective judgments lead to inconsistent rankings. As a simplified version of the AHP, the Statistical Variance Procedure (SVP) based AHP (SVP-AHP) extracts the ranking of alternatives from a multicriteria dataset without referring to costly survey processes. SVP-AHP uses pairwise comparison matrices, the powerful tool of AHP, and it does not need to measure consistency. For an objective ranking of alternatives, SVP-AHP embeds vector normalization and SVP into the AHP. SVP determines criteria weights while pairwise comparison matrices for alternatives are constructed using the normalized observations. In SVP-AHP, it is sufficient to know only criteria and alternative values, unlike AHP, where the model requires decision makers' judgments. Results of the AHP and SVP-AHP for the example in this study point out that SVP-AHP is an efficient ranking method because of its computational efficiency and objectivity