ıstrial 
ct to land 
ial estates 
uch more 
' decision 
on?" with 
n diverse 
methods, 
evidence, 
h-west of 
inally, we 
different 
could be 
10dels for 
itroducing 
dels. 
tions that 
n. These 
's in study 
sing some 
91). 
es that are 
analytical 
International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, Vol XXXV, Part B5. Istanbul 2004 
  
functions. Each processing method has one or some outputs 
that depends on processing method type, entering raw data 
structure (Vector or Raster) and requiring data for entering 
analytical functions. There are different kinds of processing 
methods in GIS environment (for example: data combination, 
vector to raster conversion, distance map production, 
reclassification, etc). 
2.3 Integration Models 
Various models were used for real world events simulation in 
GIS environment (Aronof, 1989). Integration model is one of 
them that used for site selection by integrating related spatial 
data and effective criteria. There are various integration models 
that categorized by kind of functions and their executive routine 
(Knowledge driven or Data driven). - 
Knowledge driven: Experts experience and scientists science 
are used for executing models. 
Data driven: Models are executed based on existent solutions 
and dependency value computation. 
We'll describe some useful models that have been used in our 
application and mention our reasons for selecting them. These 
models consist of: Boolean Operation, Indexing Overlay, Fuzzy 
Logic and Genetic Algorithm. 
2.3.1 Boolean Operation : In this model, input maps are 
integrated by using logical operators such as AND, OR, XOR 
and NOT (Bonham Carter and G.F., 1991). Although, Boolean 
Operation is an easy and fast model to run, there are some 
problems in its execution routine. In this model, all input factor 
maps (processed data) have same weights and appropriate sites 
can’t be separated based on their priority. In our research for 
industrial estate site selection, this model isn't used because of 
mentioned problems and weighted nature of efficient criteria 
and conditions in our application. 
2.3.2 Index Overlay Model: Factor maps are integrated 
using following Equation : 
5S. 
code NE. (1) 
>; Wi 
where Wi- The weight of ith factor map 
Si- The ith spatial class weight of jth factor map 
S= The spatial unit value in output map 
Comparing Index Overlay with Boolean Models executive 
routines, it is identified that Index Overlay model has more 
flexibility and ability for priority indication on spatial units of 
factor maps (Bonham Carter and G.F., 1991). With respect to 
mentioned characteristics, this model is useful for comparing 
and evaluating integration models in industrial estate site 
selection process. 
2.3.3 Fuzzy Logic Model: In industrial estate site selection 
application, for each factor map we can define classes and 
spatial units as a subset that their membership values in 
industrial estates convenient location are mapped between 0 and 
1. There are some fuzzy operators such as Fuzzy AND, Fuzzy 
OR, Fuzzy Product, Fuzzy Sum and Fuzzy y that are used for 
factor maps integration (An and Moon, 1991). 
Fuzzy AND : This operator operates like common statements 
in classic sets theory as Equation 2. 
H combination = MIN (u P Hu T Hh y) (2) 
where 
Keombination= Each spatial unit value in output map 
HA Rc..- Spatial units membership values 
This is used when there have been two or more factors or 
evidences together that can help solving the problem. In our 
application, this operation isn't used because of its weakness on 
participating all effective factors and lack of specific evidences 
for industrial estates location. 
Fuzzy OR : This operator operates like union statements in 
classic sets theory as Equation 3. 
M Combinatia = MAX ME) (3) 
where combination“ Each unit value in output map 
Ua pc... = Spatial units membership values 
This operator is used when there are sufficient positive factors 
and evidences in study area. In our evaluation, because of lack 
of positive factors for industrial estates site selection, Fuzzy OR 
operator isn't used. 
Fuzzy Product : Fuzzy Product Operator produces input factor 
maps membership values and emplaces results on output map 
(Equation 4). Therefore, it has decreasing affects on results and 
is used when input factor maps debilitate each other. 
n 
HV. Combination ES ; [1 ju (4) 
lI = i 
where combination” Each unit value in output map 
u;- The weight of ith factor map 
In industrial estates site selection, There is not weakener factor 
maps then fuzzy product isn't used in our evaluation. 
Fuzzy Sum : This operator is defined as following Equation. 
n 
H combination = b C Holl H, )) (5) 
i=] 
Where combination“ Each unit value in output map 
Wi= The weight of ith factor map 
This operator is used when input factor maps have increasing 
effects in each other. In our application, because of increasing 
effects of Accessing factors and Infrastructure parameters, this 
operator was used and compared with other integration 
operators. 
Fuzzy y : This operator is a general form of Fuzzy Product and 
Sum operators (Equation 6). 
ö -à 
M combination = (Fuzzy Sum) % (FuzzyProduet) (6) 
According to Equation 6, y can change from 0 to 1. This 
operator is used when there have been increasing and 
decreasing effects on interactions of evidences 
(Zimmermannand and Zayo, 1980). This operator is used for