BPMSG |
Ishizaka, A., Balkenborg, D., & Kaplan, T. (2011). Influence of aggregation and measurement scale on ranking a compromise alternative in AHP. Journal of the Operational Research Society, 62(4), 700–710. Added by: Klaus D. Goepel (10 Jun 2019 22:26:16 Asia/Singapore) Last edited by: Klaus-admin (11 Jun 2019 09:17:18 Asia/Singapore) |
Resource type: Journal Article BibTeX citation key: Ishizaka2011b Email resource to friend View all bibliographic details |
Categories: AHP/ANP Keywords: compromise, decision analysis, scale, weighted product model (WPM), weighted sum model (WSM) Creators: Balkenborg, Ishizaka, Kaplan Collection: Journal of the Operational Research Society |
Views: 2/553 Views index: % Popularity index: 37% |
Abstract |
Analytic Hierarchy Process (AHP) is one of the most popular multi-attribute decision aid methods. However, within AHP, there are several competing preference measurement scales and aggregation techniques. In this paper, we compare these possibilities using a decision problem with an inherent trade-off between two criteria. A decision-maker has to choose among three alternatives: two extremes and one compromise. Six different measurement scales described previously in the literature and the new proposed logarithmic scale are considered for applying the additive and the multiplicative aggregation techniques. The results are compared with the standard consumer choice theory. We find that with the geometric and power scales a compromise is never selected when aggregation is additive and rarely when aggregation is multiplicative, while the logarithmic scale used with the multiplicative aggregation most often selects the compromise that is desirable by consumer choice theory.
|
Notes |
Under 3.5 the authors show an example, where the compromise solution is alternative candidate B, but AHP calculates candidate B as worst alternative. I cannot reproduce the result (A=33.9%, B=32.5%, C=33.6%). AHP-OS gives (A=33.5%, B=32.3%, C=33.2%). Uncertainty analysis shows that all alternatives are overlapping within the error margins. Sensitivity analysis shows further that the solution is not robust. Using WPM aggregation instead of WAM gives candidate B the preference (A=18.8%, B=32.2%, C=18.0%).
Added by: Klaus-admin Last edited by: Klaus-admin |