Namespace List

The namespaces specified in this document are:

NamespaceAssembly
Net.Kniaz.Math.AHP.TestsNet.Kniaz.AHP
Net.Kniaz.Math.AHPNet.Kniaz.AHP

Namespace : Net.Kniaz.Math.AHP.Tests



Net.Kniaz.Math.AHP.Tests Type List

Classes

TypeSummary
TestAHP Summary description for TestProrities.


Net.Kniaz.Math.AHP.Tests Classes

TestAHP Class

Summary

public class TestAHP

Summary description for TestProrities.

Constructor Members

NameAccessSummary
TestAHP()publicInitializes a new instance of the class.

Method Members

NameAccessSummary
TestChoiceMatrix() : Voidpublic
TestConsistencyRatio() : Voidpublic
TestExpansionUtility() : Voidpublic
TestFeatureSelection() : Voidpublic AHP model for the computer system with 3 criteria - scalability, maintanablity, reliability and 19 features f0 to f18. Features have been rated against ech other and feature f12 is most preferable with 19%
TestPriorities() : Voidpublic
TestVacationSpotSelection() : Voidpublic Vacation Hierarchy Model by Saaty et all Criteria: Activities, Nightlife, Siteseeing, Cost Choices: Orlando, San Fran, New York; Inputs are Activities Orlando SF NY Orlando 1 4 3 SF 1 2 NY 1 That is Orlando is more preffered than SF and NY. Also SF is more prefferred than NY

Namespace : Net.Kniaz.Math.AHP



Net.Kniaz.Math.AHP Type List

Classes

TypeSummary
AHPModel This class implements a decision model of the Analytic Hierarchy Process by Saaty. (T.L. Saaty "The Analytic Hierarchy Process", McGraw-Hill New York, 1980) The client of the class must provide decision making criteria that are scored against each other via a pairwise comparison (using the 1-9 scale proposed by Saaty) and the choices tht are scored for every criteria also using binary comparison process. Client should use the AddCriteria method for the criteria and AddCriterionRatedChoices for choices scored agains the criteria. Order in which choices are added to the model must match the ordering of criteria in the matrix. It is assumed that all matrixes submittd to the model are symmetrical, inverse. for example: say we are trying to decide between Orlando, SF and NY as vacation places (see the unit test for a full example). We have 4 criteria to based our selection on: Activities, Nightlife, Siteseeing, Cost We start build our crteria matrix by comparing activities with remaining 3 criteria and using the scale from 1 (equally preferable to 9: extremely more preferable). Then we score nightlife against siteseeing and Cost and finally we compart Siteseeing ang cost. There is no reason to compare Cost to Activities again because it will be just a reverse from the comparison made in the first choice. As a result we get the following matrix: Activities NightLife SiteSeeing Cost Activites 1 3 4 5 Nigthlife 1 0.5 0.33333 SiteSeeing 1 3 Cost 1 The left lower triangle of the matrix should be left to 0 because the class will transpose and take inverse of the upper right corner. After setting up criteria and choices the client should call the calculate model method. final result (choices scored against each criteria) are contained in the model result matrix. Each column of the result matrix represents scored of choices against each criteria This assembly uses the GeneralMatrix package for the matrix algebra.
Constants Random Indices values (Consistency Indices for randomly selected values in the priority matrix) taken from Saaty.
PrioritiesSelector This class implements the Saaty's method for estimating eigenvalues of the priorities matrix.


Net.Kniaz.Math.AHP Classes

AHPModel Class

Summary

public class AHPModel

This class implements a decision model of the Analytic Hierarchy Process by Saaty. (T.L. Saaty "The Analytic Hierarchy Process", McGraw-Hill New York, 1980) The client of the class must provide decision making criteria that are scored against each other via a pairwise comparison (using the 1-9 scale proposed by Saaty) and the choices tht are scored for every criteria also using binary comparison process. Client should use the AddCriteria method for the criteria and AddCriterionRatedChoices for choices scored agains the criteria. Order in which choices are added to the model must match the ordering of criteria in the matrix. It is assumed that all matrixes submittd to the model are symmetrical, inverse. for example: say we are trying to decide between Orlando, SF and NY as vacation places (see the unit test for a full example). We have 4 criteria to based our selection on: Activities, Nightlife, Siteseeing, Cost We start build our crteria matrix by comparing activities with remaining 3 criteria and using the scale from 1 (equally preferable to 9: extremely more preferable). Then we score nightlife against siteseeing and Cost and finally we compart Siteseeing ang cost. There is no reason to compare Cost to Activities again because it will be just a reverse from the comparison made in the first choice. As a result we get the following matrix: Activities NightLife SiteSeeing Cost Activites 1 3 4 5 Nigthlife 1 0.5 0.33333 SiteSeeing 1 3 Cost 1 The left lower triangle of the matrix should be left to 0 because the class will transpose and take inverse of the upper right corner. After setting up criteria and choices the client should call the calculate model method. final result (choices scored against each criteria) are contained in the model result matrix. Each column of the result matrix represents scored of choices against each criteria This assembly uses the GeneralMatrix package for the matrix algebra.

Constructor Members

NameAccessSummary
AHPModel()public Parametrized Constructor

Property Members

NameAccessSummary
CalculatedChoices : GeneralMatrixpublic
CalculatedCriteria : GeneralMatrixpublic Criteria priorities as calculated by the model. A matrix of n by 1
ChoiceMatrix : GeneralMatrixpublic Model choices scored in pairwise comparison for each criteria A matrix of n by n*m
ConsistencyRatio : GeneralMatrixpublic Matrix of consistency ratios for the model size of n+1 where n is number of criteria First row represents consistency ratio for the
ModelResult : GeneralMatrixpublic Result
RatedCriteria : GeneralMatrixpublic Raw criteria scored after the pairwise comparison A matrix of n by n

Method Members

NameAccessSummary
AddCriteria() : Voidpublic
AddCriteria() : Voidpublic
AddCriterionRatedChoices() : Voidpublic
AddCriterionRatedChoices() : Voidpublic
CalculateModel() : Voidpublic
ExpandUtility() : GeneralMatrixpublic (

Constants Class

Summary

public class Constants

Random Indices values (Consistency Indices for randomly selected values in the priority matrix) taken from Saaty.

Constructor Members

NameAccessSummary
Constants()publicInitializes a new instance of the class.

Field Members

NameAccessSummary
randomIndices : Double[]public

Method Members

NameAccessSummary

PrioritiesSelector Class

Summary

public class PrioritiesSelector

This class implements the Saaty's method for estimating eigenvalues of the priorities matrix.

Constructor Members

NameAccessSummary
PrioritiesSelector()publicInitializes a new instance of the class.

Property Members

NameAccessSummary
CalculatedMatrix : GeneralMatrixpublic Final matrix of n by 1 containing results of the selection
Consistency : Doublepublic Consistency of the selection
ConsistencyRatio : Doublepublic Consistency ratio of CI/RI
LambdaMax : Doublepublic Estimate of the max eigenvalue

Method Members

NameAccessSummary
ComputePriorities() : Voidpublic Calculates priorities using the Saaty method
PCalc() : Voidpublic Average values of the priority matrix over sum of columns. set values of sum of averaged rows into a new matrix