Spatial Analysis

In order to complete my analysis, I used the Multi-Criteria Evaluation (MCE) Decision Support in IDRISI using the Decision Wizard under the GIS Analysis/Decision Support menu. First, I had to establish some constraints and some factors that need to be considered. They were as follows:

Constraints	Factors
Industrial Areas Transportation Areas	Commercial Areas Undeveloped Areas Elementary Schools Parks Water

My constraint files were indusbool and transbool and the factor files were comm-dist, parks-dist, undevel-dist, water-dist, and schools-dist. Using the FUZZY Factor Standardization method, I established the Function Type, Function Shape, and Control Points for each of the five factors. The table below lists values used:

	Commercial	Parks	Undeveloped	Water	Schools
Function Type	J-shape	J-shape	Linear	Sigmoidal	J-shape
Function Shape	monotonically decreasing	symmetric	monotonically decreasing	monotonically decreasing	symmetric
Control Points - a	-	0	-	-	0
b	-	50	-	-	300
c	1000	750	0	0	1500
d	6000	1500	1200	1500	5000

These values, however, did not produce good results due to large distance values. I changed the values for the "d" Control Points for Commercial, Undeveloped, Water, and Schools to end up with the following table:

	Commercial	Parks	Undeveloped	Water	Schools
Function Type	J-shape	J-shape	Linear	Sigmoidal	J-shape
Function Shape	monotonically decreasing	symmetric	monotonically decreasing	monotonically decreasing	symmetric
Control Points - a	-	0	-	-	0
b	-	50	-	-	300
c	1000	750	0	0	1500
d	3000	1500	1000	1000	3000

This produced better results. The images produced were called commfuzz, parksfuzz, undevelfuzz, waterfuzz, and parksfuzz. The next step in the Decision Wizard involved choosing a Factor Weighting Option. Using the Analytical Hierarchy Process (AHP) option. This created the following pairwise comparison file from the follwing matrix:

	commfuzz	parksfuzz	undevelfuzz	waterfuzz	schoolsfuzz
commfuzz	1
parksfuzz	5	1
undevelfuzz	5	1/5	1
waterfuzz	1	1/5	1/5	1
schoolsfuzz	3	1/3	1/3	1	1

The resulting eigenvector of weights (left) and decision support file (right) were:

From the eigenvectors of weights, the consistency ratio was acceptable allowing me to proceed and aggregate the data by Weighted Linear Combination (WLC). This enables the factors to be multiplied by their corresponding weights. The resulting image is an MCE WLC image called MCEWLCsuitable shown below:

In order to find the suitable places to live, I used the RECLASS module to classify unsuitable areas as values ranging from 0 to 175 and suitable areas as 175 to 255.

This produced too many areas. I finally ended up with a suitable image reclassed to portray unsuitable as 0 to 225 and suitable areas as 225 to 255. Here is the macro model I used:

The resulting final image using MCE and WLC is:

MOST SUITABLE PLACES TO LIVE WITHIN SURREY & LANGLEY!!!

Vitally Ufimtseff
GEOG 355 Term Project
November 2003

Where to Live Within Surrey and Langley, BC.