Chapter 5: Optimisation, Planning and Decision Support

Title: D-optimal designs approach for AHP incomplete comparisons

Author(s): Ioannis Karamitsos, Khalid Al-Arfaj, Charalampos Apostolopoulos

Address: Department of Computing School of Engineering, University of Sunderland, Sunderland, SR1 3SD, UK | Rayadh Investment Company, King Fahad Road, Riyadh, Kingdom of Saudi Arabia | School of Engineering and Mathematical Sciences, City University London, Northampton Square, London, EC1V 0HB, UK

Reference: Software, Knowledge, Information Management and Applications (SKIMA 2013) pp. 213 - 222

Abstract/Summary: Analytic hierarchy process (AHP) is an established method for multiple criteria decision-making. The main characteristic is pairwise factors comparison and final classification based on a relative scale. Even though, AHP is powerful enough, many AHP experiments are not complete for several reasons making the decision problem a rather complicated process, time consuming or even low validity. For this reason, it is considered very important to select and know which comparisons to take into consideration in order to avoid repetitions and minimize result bias. In this paper, concerning the AHP incomplete comparisons, we examined a solution by using a D-optimal criterion with linear modeling. In order to enhance the statistical analysis overall, a detailed example is presented which uses D-optimal design for further evaluation.

Order a copy of this article