Examining Pareto optimality in analytic hierarchy process on real Data: An application in public transport service development

Szabolcs Duleba,Sarbast Moslem
DOI: https://doi.org/10.1016/j.eswa.2018.08.049
IF: 8.5
2019-02-01
Expert Systems with Applications
Abstract:Recent findings in operations research and decision theory show that the eigenvectors derived from pairwise comparison matrices are not Pareto optimal in all cases. Thus, looking at the traditional eigenvector method of Saaty, a better approximation may exist for characterizing the decision makers’ opinion. The Pareto optimal modified vector may lead to better decisions for better expressing the real intention of the evaluator. This paper aims to examine Pareto-optimality on real data pairwise comparison matrices gained from a public transport Analytic Hierarchy Process (AHP) survey. Moreover, detecting the impact of weight score modification on the whole AHP structure and thus the significance of Pareto test is also in the scope of this study. Further, a detailed description of the general process of Pareto optimal AHP is also included. The application has been conducted in Mersin, Turkey with the purpose of determining public preference on the importance of developing supply quality elements in local bus transportation service.
computer science, artificial intelligence,engineering, electrical & electronic,operations research & management science
What problem does this paper attempt to address?