A Study of Fundamental Scale for Pairwise Comparisons in Analytical Hierarchy Process
碩士 === 國防管理學院 === 國防決策科學研究所 === 91 === The Analytical Hierarchy Process Model was designed by T.L. Saaty as a decision making aid. In order to cluster the decision elements for alternatives, the AHP used pairwise comparison and transferred from oral expression to numerical values-the Fundamental Sca...
Main Authors: | , |
---|---|
Other Authors: | |
Format: | Others |
Language: | zh-TW |
Published: |
2003
|
Online Access: | http://ndltd.ncl.edu.tw/handle/04969225674995602746 |
id |
ndltd-TW-091NDMC1683002 |
---|---|
record_format |
oai_dc |
spelling |
ndltd-TW-091NDMC16830022016-06-22T04:20:05Z http://ndltd.ncl.edu.tw/handle/04969225674995602746 A Study of Fundamental Scale for Pairwise Comparisons in Analytical Hierarchy Process 分析層級程序法成對比較基礎標度系統之研究 Guey-Rong Shu 許桂溶 碩士 國防管理學院 國防決策科學研究所 91 The Analytical Hierarchy Process Model was designed by T.L. Saaty as a decision making aid. In order to cluster the decision elements for alternatives, the AHP used pairwise comparison and transferred from oral expression to numerical values-the Fundamental Scale (1-9) to determine the priority. The translation from oral to numerical scale (1-9) which designed by Saaty had some issues about the Systematic Bias. The purpose of this paper is to examine this kind of problem and by using the Accuracy Scale to avoid the bias in consistent test and assessment of the alternatives. The results which carried out by the research are described below: (1) According to the questionnaire, the mean values of the Accuracy Scale are 『1.000, 1.976, 3.384, 5.353, 9.820』 and 『1.000, 1.813, 3.078, 5.098, 9.190』 under the cognitive relative importance which is in the scale with reference and the scale without reference. It is different from the scale (1-9) which designed by Saaty. (2) Comparing to six examples in the Saaty[27], we found that the Consistency Ratio by using the Accuracy Scale is less than using the Saaty’s Scale. This result showed that using Accuracy Scale can achieve better consistency. (3) Random Analysis (a) The value of which is equal to the number of 1.000 will decrease when the matrix expand in 100 random sample sets. The alternatives priority order using Saaty’s Scale and Accuracy Scale are different, the number of difference increases as matrix expansion. Among the optimal alternatives identified by using the Accuracy Scale, there are about 5 to 14 samples sets having different optimal using Saaty’s Scale. (b) To Compare with the Saaty’s Scale, the number of the Consistency Ratio using the Accuracy Scale is less than using Saaty’s Scale increases as the matrix expansion. Ching-Pu Chen 陳勁甫 2003 學位論文 ; thesis 87 zh-TW |
collection |
NDLTD |
language |
zh-TW |
format |
Others
|
sources |
NDLTD |
description |
碩士 === 國防管理學院 === 國防決策科學研究所 === 91 === The Analytical Hierarchy Process Model was designed by T.L. Saaty as a decision making aid. In order to cluster the decision elements for alternatives, the AHP used pairwise comparison and transferred from oral expression to numerical values-the Fundamental Scale (1-9) to determine the priority. The translation from oral to numerical scale (1-9) which designed by Saaty had some issues about the Systematic Bias. The purpose of this paper is to examine this kind of problem and by using the Accuracy Scale to avoid the bias in consistent test and assessment of the alternatives.
The results which carried out by the research are described below:
(1) According to the questionnaire, the mean values of the
Accuracy Scale are 『1.000, 1.976, 3.384, 5.353, 9.820』
and 『1.000, 1.813, 3.078, 5.098, 9.190』 under the
cognitive relative importance which is in the scale with
reference and the scale without reference. It is different
from the scale (1-9) which designed by Saaty.
(2) Comparing to six examples in the Saaty[27], we found that
the Consistency Ratio by using the Accuracy Scale is less
than using the Saaty’s Scale. This result showed that
using Accuracy Scale can achieve better consistency.
(3) Random Analysis
(a) The value of which is equal to the number of 1.000 will
decrease when the matrix expand in 100 random sample sets.
The alternatives priority order using Saaty’s Scale and
Accuracy Scale are different, the number of difference
increases as matrix expansion. Among the optimal
alternatives identified by using the Accuracy Scale, there
are about 5 to 14 samples sets having different optimal
using Saaty’s Scale.
(b) To Compare with the Saaty’s Scale, the number of the
Consistency Ratio using the Accuracy Scale is less than
using Saaty’s Scale increases as the matrix expansion.
|
author2 |
Ching-Pu Chen |
author_facet |
Ching-Pu Chen Guey-Rong Shu 許桂溶 |
author |
Guey-Rong Shu 許桂溶 |
spellingShingle |
Guey-Rong Shu 許桂溶 A Study of Fundamental Scale for Pairwise Comparisons in Analytical Hierarchy Process |
author_sort |
Guey-Rong Shu |
title |
A Study of Fundamental Scale for Pairwise Comparisons in Analytical Hierarchy Process |
title_short |
A Study of Fundamental Scale for Pairwise Comparisons in Analytical Hierarchy Process |
title_full |
A Study of Fundamental Scale for Pairwise Comparisons in Analytical Hierarchy Process |
title_fullStr |
A Study of Fundamental Scale for Pairwise Comparisons in Analytical Hierarchy Process |
title_full_unstemmed |
A Study of Fundamental Scale for Pairwise Comparisons in Analytical Hierarchy Process |
title_sort |
study of fundamental scale for pairwise comparisons in analytical hierarchy process |
publishDate |
2003 |
url |
http://ndltd.ncl.edu.tw/handle/04969225674995602746 |
work_keys_str_mv |
AT gueyrongshu astudyoffundamentalscaleforpairwisecomparisonsinanalyticalhierarchyprocess AT xǔguìróng astudyoffundamentalscaleforpairwisecomparisonsinanalyticalhierarchyprocess AT gueyrongshu fēnxīcéngjíchéngxùfǎchéngduìbǐjiàojīchǔbiāodùxìtǒngzhīyánjiū AT xǔguìróng fēnxīcéngjíchéngxùfǎchéngduìbǐjiàojīchǔbiāodùxìtǒngzhīyánjiū AT gueyrongshu studyoffundamentalscaleforpairwisecomparisonsinanalyticalhierarchyprocess AT xǔguìróng studyoffundamentalscaleforpairwisecomparisonsinanalyticalhierarchyprocess |
_version_ |
1718317077554528256 |