A role for sensitivity analysis in presenting the results from MCDA studies to decision makers
An important element of judgment in decision making is a quantitative appreciation of the uncertainties involved, together with an indication of the likely sources of the uncertainty. While uncertainty analysis is often seen in multicriteria decision analysis (MCDA) studies, sensitivity analysis (SA) is still largely absent or rudimentary. Desirable attributes of a sensitivity analysis in support to a MCDA would involve it being quantitative, computationally efficient and easy to read and understand. New quantitative SA methods, derived from Sobol' sensitivity indices and from the Fourier amplitude sensitivity test (FAST) appear adequate to the task for the capacity to break down quantitatively the variance of the model response, not only according to input factors, but also according to subgroups of factors. The approach suggested is computationally more efficient that factor-by-factor analysis.
Bibliographic Reference: Article: Journal of Multicriteria Decision Analysis
Record Number: 199711707 / Last updated on: 1998-01-20
Original language: en
Available languages: en