Sensitivity analysis on fuzzy analytic hierarchy process using elementary effects method /
Marie Analiz April Arceņo Limpoco.
- 2011
- 101 leaves.
Thesis (BS Applied Mathematics) -- University of the Philippines Mindanao, 2011
The study was done to evaluate the sensitivity of Buckley's fuzzy AHP model using the elementary effects methods developed by Morris (1991) and improved by Campolongo (2007). To cater the fuzzy nature of the input factors of Buckley's model, the elementary effects method was modified. Two possible sources of variation in the output were considered: the influence of individual factors and of groups of factors. Different cases, consisting of a varied number of criteria and alternatives, were considered. Sensitivity measures were computed for individual factors and for groups of factors. Kruskal Wallis test was performed on the sensitivity measures to determine if there is at least one factor or group that is significantly influential on the output. Nemenyi's multiple comparison test was used as a posthoc test. The individual factors and the groups of factors were classified into influential and non-influential. For almost all cases, group 2, consisting of pairwise comparisons among alternatives under criterion 1, was the influential group. Also among the individual factors, those factors belonging to group 2 were influential in most cases. This implies the critical role of the individual factors belonging to group 2 on the rankings of the alternatives based on their importance. Furthermore, since the results of the sensitivity analysis on Buckley's fuzzy AHP model is a general one, the assignment of which criterion will be considered as criterion 1 is also crucial in the output.