A Solution Approach for Agricultural Service Center Location Problem Using TOPSIS, DEA and SAW Techniques - Technology Management for Sustainable Production and Logistics - page 29

Fig. 2 The procedure of MADM method
First action in MADM is specifying attributes and also sub-attributes if it is
necessary. In this phase, Delphi method can be used to help. The sub-attribute defi-
nition is to facilitate the data gathering. The data must be gathered or generated
from databases, surveys and Geographical Information System (GIS) and etc. Next
step is the giving score to the attributes for each alternative (candidate location).
In this step, according to attribute data type, quantitative or qualitative, decision
matrix method or AHP can be selected. In the current study, because of several
technical reasons, village points are proposed as candidate location in each region.
Quantitative date more than qualitative data will lead to reliable judgment about
candidate locations. So in this study qualitative data is preferred and decision
matrix method is selected for giving score to the attributes of candidate locations.
Some decision methods such as SAW and TOPSIS use the additive values of
attributes to analysis. Since the type of data is different for all attributes, data nor-
malization is required. Some other ranking methods, such as DEA, are not sensi-
tive to the unit of data, but in this study normalized data is used for all analysis.
In this research fuzzy and linear normalization methods are used. The data order
in the linear normalization method is retained, while in the fuzzy normalization
method the order is disturbed. Totally in this chapter seven MADM methods are
applied and because of the linear normalization method ability, in this research lin-
ear method is used in five methods.