عنوان مقاله [English]
Most of the multi-criteria decision methods (MCDM), including Analytic Hierarchy Process (AHP) and Analytic Network Process (ANP), use subjective judgments for obtaining the relative importance of the decision elements of a subject matter. This has been considered as one the important limitations of the MCDMs. The F’ANP model, a hybrid factor analysis and analytic network process (F’ANP) model was first introduced by Zebardast (2013) to overcome this inherent limitation of the MCDMs in the process of obtaining the relative importance of decision elements. The proposed F’ANP model uses factor analysis (FA) to extract the underlying dimensions of the phenomenon. These identified dimensions and their primary variables are then entered into a network model in Analytic Network Process (ANP). The ANP is used to calculate the relative importance of different variables of the subject matter, taking into consideration the results obtained from FA and the possible interdependence between variables of the individual dimensions of the phenomenon. The model was first used to measure the social vulnerability as a composite index. In this paper the F’ANP model is modified to overcome one of the inherent limitations of the ANP and thus the original F’ANP as well. This is the lengthy process of computation required by the ANP. In the original F’ANP model after using factor analysis to derive the components of the subject matter and in the ANP part, pair-wise comparison matrices are constructed to obtain the relative weights of the decision elements. In this paper we show that since F’ANP uses the objective measures derived from the factor analysis method to construct the pair-wise comparison matrices in ANP in order to obtain the relative importance/weights of the decision elements, this could be directly obtained by normalizing the original vectors without a need for constructing the pair-wise comparison matrices. This process results in a relative weight computation that is robust and does not need a process of controlling the possible inconsistencies in the decision process. This enormously shortens the computation process, thus making F’ANP an easier model to use. The application of the proposed model to a real world case study show that it is a robust approach for constructing a composite index such as social vulnerability index. Its application to assess the spatial distribution of social vulnerability to earthquakes in the 117 zones of Tehran Metropolis indicates that the southern parts of the city are most socially vulnerable to earthquake hazards. The proposed F’ANP model could be utilized to study any multi-dimensional phenomenon in urban and regional planning and in any other fields of study that are multi-dimensional in nature and for their assessment several indicators are used.