Comparison Analysis of Fuzzy AHP and AHP in Multiple-Decision Making Problem Using Statistic Approach

Abstract

Analytic Hierarchy Process (AHP) is a robust approach for decision making under complex criteria. Decision makers express their opinions differently and arbitrarily, giving rise to uncertainty in the ranking of alternatives. Fuzzy AHP was then developed and applied under those circumstances to reduce the uncertainty. This paper generates a string of randomly simulated data to represent completely arbitrary triangle fuzzy number, and based on these data compare Fuzzy AHP with classical AHP in statistical manner. Then the paper conducts a series of SPSS linear regressions for this comparison with two critical factors: the pairwise comparison weight value of AHP and the fuzzy value range of Fuzzy AHP. The regression shows how these two factors affect the differences between the two approaches. Results indicate that the pairwise comparison weight value of AHP significantly influences the difference while the fuzzy value range does not. In general, Fuzzy AHP narrows the final weights between each criterion, but in some extreme situations, this conclusion does not exist any more.