Methodology:

Multi-Criteria Evaluation (MCE) is an efficient and helpful analysis tool to evaluate certain problems such as suitability to place certain features in certain places. MCE works by having two types of criteria: constraints and factors. Constraints are expressed in Boolean map form where a limit is set to remove areas that are not of interest. For example, a value of 1 is assigned to areas that are in INCLUDED in consideration for analysis and a value of 0 is given to areas that will be EXCLUDED from consideration for analysis. Factors are the type of criteria that will help define to what degree each factor included in consideration will be suitable compared to others.

Constraints:

The constraints for this project were restricted to landuse types for the city of Surrey. The new police station must be located within the city of Surrey. A value of 0 was given to the following landuse areas, which are excluded from consideration:

  • Transportation, Communication and Utilities

  • Industrial

  • Institutional

  • Commercial

  • Agricultural

  • Lakes and Water Bodies

  • Recreation and Protected natural areas
These areas were given a value of 1 because they are areas that are included in consideration for the new location of the police station:
  • Open and Undeveloped

  • Residential - Single Family

  • Residential - Rural

  • Residential - Townhouse and Low-rise Apartments

  • Residential - High-rise Apartments


Factors:

Landuse: The New police station should be further away from excluded landuse types (landuse with value=0) but within the considered landuse types (landuse with value=1)
Average Income: The new police station should be suitable to locate in areas of low income because crime and low income have always been considered tied with one another.
Unemployment Rate: The new station should be in areas where unemployment rates are high because high crimes rates are also associated with high unemployment rates.
Total Number of Crimes (Surrey Districts):  Example
 The new station should be in areas where total number of crimes is high.
Police Station: The new station should not be near the existing police station in Surrey.
Schools: The new station should be away from schools because it would cause panics and would be disrupting for people attending surrounding schools.
Bus Routes: The new station should not be close to existing bus routes as it will be inconvenient for police to navigate through traffic, and it would be inconvenient for locals that are commuting by bus.
Water: The new station should be far away from waterways because the fear of natural factors such as flooding and other issues.
Skytrain Station: The new station should not be near existing skytrain station, because there will be a large influx of people moving in and out of the skytrain station, although there are only a few skytrain stations in the city of Surrey.

Fuzzy:

The Decision Wizard function within the Decision Support in IDRIS was used to create the fuzzy sets. The purpose of the fuzzy set is to standardize all the factors that will be used in the multi-criteria evaluation, to a byte level range from 0-255.

Cartographic Model for Fuzzy  <--- interactive (requires Flash plug-in to view). Please follow the cartographic model link to view the complete fuzzy process in interactive mode. May take time to load...

Factors to Standardize Membership Function Shape Membership Function Type Min Max Control Points Fuzzy Raster
avginc_rcl1 Monotonically decreasing Sigmoidal 0 32767 c=100   d=32767 avginc_fuzzy
unemp_rcl2 Monotonically decreasing Sigmoidal 0 67 c=0       d=67 unemp_fuzzy
crime_rcl Monotonically increasing Sigmoidal 0 5 a=0       b=5 crimefuzzy
police_dist Monotonically increasing Sigmoidal 0 16187.4 a=100   b=16187 policefuzzy
school_dist Monotonically increasing Sigmoidal 0 5448.32 a=100   b=3500 schoolfuzzy
bus_dist Monotonically increasing Sigmoidal 0 6440 a=100   b=4500 busfuzzy
water_dist Monotonically increasing Sigmoidal 0 11453.1 a=100   b=3000 waterfuzzy
skystation_dist Monotonically increasing Sigmoidal 0 23265.5 a=100   b=20000 skystationfuzzy


This is a visual example of a sigmoidal curve. A Sigmoidal Membership function was used to show and explain concepts such as, the further away the new police station is from the existing police station, bus routes, skytrain station, or schools the more suitable. Also to show that it is more suitable to locate close to areas with higher total crimes or high unemployment rates.


For the Landuse raster file, the decision wizard was not used to standardize the data into a new raster fuzzy image. The landuse file was reclassed to standardize the data to byte level range of (0-255). It was reclassed the following way:

0 0 2
255 3 3
150 4 4
0 5 5
200 6 6
0 7 8
100 9 9
150 10 10
0 11 12
-9999


The landuse raster file was reclassed this way because as mentioned above, the areas that are included for considerations are areas with a value of 1 in the landuse constraints. The areas that are considered are reclassed given a value from 0-255 to show the importance and suitability for each of these landuse factors. The following table shows the corresponding landuse code, landuse name, the reason as to why it is suitable to build in these areas, and the (value of importance) for each factor:


Landuse Code Landuse Name Reason to Locate in these Area Values given in Reclass
3 Open and Undeveloped best to locate new police station in open areas 255
4 Residential - Single Family medium range suitability to locate in these areas 150
6 Residential - Rural more Suitable to build in areas with rural residential 200
9 Residential - Townhouse and Low-rise Apartments less suitable to locate in these areas 100
10 Residential - High-rise Apartments medium range suitability to locate in these areas 150


The following images are example of some of the map results after fuzzy sets was applied to them to standardize them to value of 0-255:

   




Peter Chan . Copyright © 2006.