You are on page 1of 18

Journal of China University of Science and Technology Vol.__-2011.

11

The purchasing selection of a decision making Uses AHP


AHP 使用於採購選擇的決策

Department of Business Administration,


China University of Science and Technology. Taipei, Taiwan, R.O.C.

Abstract

In this paper, we present a practical synthesis structure involving the modified Delphi
technique, the identified criteria method and the analytic hierarchy process (AHP).
Besides, we utilized the literatures and expert’s experience to develop an evaluative
method. It is an effective assessing technique to help the company’s purchasing
department, and it could directly affect the final demand. The product selection process
involved complex multi-criteria problems. These existing problems will be improved by
a systematic and logical approach based on the different product attributes. Here, the
AHP model may be very useful technique to arrive an optimum decision, and then
applied a real example to examine its feasibility to indicate how to satisfy the
purchasing needs. This method has the most contribution in the systematic and
rationality in the evaluation process that could reduce quite cost and time in the decision
making.

Keywords: Delphi technique, Purchasing decision making, Analytic hierarchy process, Identified criteria
method, Experts system.

摘要
本論文提出一種實用的綜合評選架構,包含德爾菲技巧、準則認定與層級分
析,我們閱讀相關文獻與專家經驗發展出的評估方法,這種有效的評估技巧能夠
協助採購部門減少成本與符合產品需求。AHP 評估過程包括有改善複雜的多數的準
則,藉系統與邏輯的推導,依據不同特性做合理比較,達成最佳的採購決策,是
一種非常實用的技巧,然後我們應用實際的例子,檢驗它的可行性,證實它能夠

1
滿足採購的需要。本文最大的貢獻是評估過程中,能夠減少相當時間成本,相當
合理的決策。
關鍵詞:德爾菲技巧,採購決策,層級分析,確認準則方法,專家系統。

1. Introduction

In recent years, the market environment has been undergone a revolutionary changes.
The changes were forcing such as increasing customer demands, functional advances,
and brand kinds. Many authors suggested the use of the analytic hierarchy process
(AHP) approach for vendor selection problems [1-3]. In relatively small and intensive
groups are becoming increasingly popular in both the private and public sector.
Purchasing on a more specific level, some studies on success factors have been carried
out [4]. The previous empirical study that explicitly deals with success factors for
managing purchasing groups [5] among 28 organizations that participate in purchasing
groups. However, the results of Hendrick are not fully consistent with the results of
existing studies on success factors for inter organizational cooperation [6]. In a large
number of cases, the advantages of cooperative purchasing can outweigh the costs of
cooperation, and the disadvantages such as anti-trust issues and disclosure of sensitive
information. In practice, small and intensive purchasing groups do not always popular
and premature endings of such groups occur [7]. Schotanus identify critical success
factors for managing small and intensive purchasing groups by comparing successful
and unsuccessful purchasing groups in a large-scale survey [8].
In addition, the product selection of the purchasing system is equally an important
problem involved many criteria, technical requirements, service specifications and unit
cost, etc. Although, we had read many literatures but we always felt that was a complex
problem. Thus, there is a need to develop a systematic selection process for more
simplification, in order to identify relevant criteria, and evaluate effective method for
purchasing process.
The approach will reduce cost and time in selection process to form the consensus
decision making. Nydick and Hill suggested the use of the AHP approach for vendor
selection problems. They suggested the AHP model mainly because of its inherent
capability by handling qualitative criteria in vendor selection problems. Furthermore,
the AHP model can be easily understood about the purchasing problems, and it can help

2
Journal of China University of Science and Technology Vol.__-2011.11

to improve the decision making process. The hierarchical structure was used in
formulating model to visualize the problem systematically in relevant criteria and sub-
criteria.
We try to take a practical case for buying a folding bike, so that how to make the
purchasing decision making by above the systematic and logical approach, and how to
enumerate the success criteria and sub-criteria then to set up an effective selection
method. Here, we describes the collection of the critical factors; the first group into two
chief criteria such as cost and quality, the next make up four sub-criteria such as brand,
price, style, and specification. The success factors selection in Brand factors include
reputation, guarantee, service, operation facility, configuration, network information,
advertisement, reliability, and availability. In Price factors consist of unit price,
accessories value, functional items, maintenance expense and technology development
expense. In Style factors involve style design, weight, high, color, handlebar, frame,
folding size and saddle. In Specification factors compose of material, shifter, shock
absorber, tire, brake system, fork and chain wheel, etc. Above preliminary collection
were extracted 35 factors from the expert’s different viewpoints. In the AHP
construction, a conceptual foundation involves four phases; structuring the decision
problem, measurement technique, data collection, determination of normalized weights
then synthesis finding solution in the assessment approaches. In addition, the Delphi
technique was used to collect more large number of the critical factors, and then the
survey questionnaire was used to extract appropriate number of the success factors, so
that to form a hierarchy level for assessing priority weights. Final to determine the
highest weight as the optimum product.

2. Modified Delphi technique

The Delphi technique is proposed from earlier response. Delbecq suggested the
experts who are anonymous and do not meet each other in person [9]. Dijk point out the
Delphi technique is a suitable communication technique on the subjective base on
norms and opinions for social research [10]. In addition, the decision-making group
probably should not be too large; a minimum of 5 to a maximum of about 50, a work
group of five to nine members [11]. Later on, Delphi technique is a conventionally
adopted qualitative forecasting method [12], which involves the systematic application

3
and collation of experts on a particular issue, then through a set of carefully designed
sequential questionnaires, and interspersed with summarized information and feedback
of opinions derived from earlier Delbecq’s response. Chaw applied the Delphi to select
procurement system for construction project [13]. Hsu and Chen adopted the modified
Delphi technique to develop and implement a selection model for chain store [14].
Recently elicit the knowledge of ventilation management to provide a baseline for
evaluating the performance of an expert system for neonatal ventilation [15]. Therefore,
we will adopt the modified Delphi technique based on results of literature review and
interviews with experts to select the probably criteria.

2.1. Sample and data collection

This study establishes and demonstrates the effectiveness of the product selection
model. The results coming from the different viewpoints of the experts are described
separately. In this case, we asked 16 participants took part in the first round to use the
modified Delphi method. However, the participants who were unable to attend another
second round using the focus group’s meeting, but they able to view the summary of the
group’s first round responses. We help purchaser to choose five companies such as
Dahon, Shimano, KHS, Merida and Doppelganger CO., LTD. These were all the large
company and well-known manufacturing in Taiwan for selling various bikes to over the
world. We wish to select a better folding bike. But, we have not an objective selection
method. Therefore, we followed the expert system to group a team including
manufacturer, user, technician and sales experts. They could offer many critical factors
from their experiences. Notice, the questionnaire was send using e-mail; the evaluation
and selection were defined; the finally factors was extracted; and the results were
collected after passing second round of using the modified Delphi technique. Result, the
decision making had selected five brands, two strategy issues, four success criteria and
more 30 sub-criteria (see Fig. 1).

2.2. Identified the criteria and sub-criteria

4
Journal of China University of Science and Technology Vol.__-2011.11

Dickson [16] had identified the different criteria for vendor selection including
quality, delivery, performance history, warranties, price, technical capability and
financial position. Many authors identified a number of criteria with respect to financial,
technical and operational aspects that are applicable to select a decision making system
[17-18]. In this paper, we are grouped into two chief criteria with cost and quality, four
sub-criteria with brand, price, style, and specification. The brand factor includes
reputation, guarantee, service, operation facility, configuration, network information,
advertisement, reliability, and availability factors. The price factor consists of unit price,
accessories value, functional items, maintenance expense and technology development
expense. The style factor involves the style design, weight, high, color, handlebar,
frame, folding size and saddle. The specification factor composes of material, shifter,
shock absorber, tire, brake system, fork and chain wheel, etc. Above preliminary
collection of 35 factors was extracted from the expert’s different viewpoints.
One of the contribution manners will try to identify the chief criteria and sub-criteria,
we enumerated the large number of critical factors; it was a difficult to solve the
complex selection problems facing so many factors. To overcome this problem; First we
shall cut-off lower 2.30 values to reduce the number of factors as a fewer desirable.
However, a survey questionnaire approach is organized into a simple evaluation model.
In general, asking all participants rated each factor using the three-point scale of “not
important”, “somewhat important” and “very important” with the values of 1, 2 and 3,
the criteria value are arranged in descending order of their personal viewpoint. We
would draw out a suitable quantity of the success factors from the largely factors (see
Table. 1).

3. AHP methodology

A decision method should decompose the complex multi-criteria into a hierarchy


construction [19]. The AHP approach is a measurement theory that prioritizes the
hierarchy and consistency of judgmental data provided by a group of decision makers.
The AHP model incorporates the evaluations of all experts’ participants into a final
decision, without used their subjective criteria, which was used by pair-wise
comparisons of the alternatives [20]. Sinuany applied the AHP approach to determine

5
the location of a hospital in a rural region [21]. In general, the AHP approach is a
scoring method that was designed to change a complex decision problem into a simple
hierarchy process, and then formed a priority weight within each level of the hierarchy
by carrying out simple pair-wise comparisons, the relative importance of decision
criteria, attributes and alternatives [22]. For a conceptual foundation, the AHP modeling
process involves four phases; structuring the decision problem, measurement technique,
data collection, determination of normalized weights and synthesis finding solution to
the problem [23]. While we used this four-phase to formulate an AHP model for an
assessment process that could be applied to the product selection. Thus, the AHP has
been successfully applied to a various problems with the calculation procedure as
follows:

Establishment of pair-wise comparison matrix A; Let C 1, C2, …, Cn denote the set of


elements, while aij represents a quantified judgment on a pair of elements Ci , and
C j . The relative importance of two elements was rated using a scale with the values 1,
3, 5, 7, and 9, where 1 refers to “equally important”, 3 denotes “slightly more
important”, 5 defines “strongly more important”, 7 represents “demonstrably more
important”, and 9 denotes “absolutely more important”. These scales yield a

n×n matrix A as follows:

C1, C2, …, Cn where aij  1 and aij  1 aij , i, j  1,2  , n. In matrix A, the problem
becomes one of assigning to the n elements C1 , C 2 ,  , C n a set of numerical weights
W1 , W2 ,  , Wn that was reflected the recorded judgments.

C1  1 a12  a1n 
C 2 1 a12 1  a2 n 
A   aij    
    
 
C n 1 a1n 1 a 2 n  1 

Saaty suggested that the largest eigen-value max would be

n Wj
max   aij 
j 1 Wi

6
Journal of China University of Science and Technology Vol.__-2011.11

And then Saaty proposed utilizing consistency index (CI) and consistency ratio (CR) to
verify the consistency of the comparison matrix. CI and CR are defined as follow:

CI   max  n  /  n  1 ,

CR  CI / RI ,

Here, the RI represents the average consistency index over numerous random elements
of same order reciprocal matrices. If CR≦0.1 indicated that the matrix reached
consistency.

3.1. Application of the AHP model to a specific product selection


problem

As customer loving the folding bike he wants to buy it in general market, but he was
unable to make an available assessing technique. We could help him to consider a
decision problem to use an AHP model for buying the better folding bike, and must be
demonstrated how the model is reasonable assessment. We successfully applied to the
purchasing problems that the calculation procedure was illustrated as follows: First
process would collect the related data by specific experts in the public; the next process
we need to consider the strategic factors by expert system; the final defined the criteria
and sub-criteria. Here we formed a priority weight within each level of the hierarchy by
pair-wise comparisons, the relative importance of decision criteria, attributes, and
alternatives to formulate the AHP model (see fig. 2). Notice, if we find that the priority
weights are equal in some level, while we can delete the level to reduce the number of
the matrices. In level 2, the strategic issues can not delete because possessed different
priority weights. In buying a folding bike, it is not involved in research or design
activities, but it is a new applicable method to qualify the bike’s various attributes. Now,
we have had five kinds of the potential bikes, and try to select optimal one depending on
existing qualified information. We apply the proposed AHP model to demonstrate how
it can be used and how the results can be compared with all criteria. Although, it is not a
formal AHP application any reference to literatures, but we used them in the selection
product as a better assessment technique. We find AHP model is suitable for a product
selection having two main contribution: (1) AHP model is an effective tool for dealing

7
with a large number of tangible and intangible factors; (2) with a user-friendly feature,
AHP approach allows the decision maker not only visible preferences but also identified
inconsistent judgments, we could correct them to get an optimum selection during the
decision process.

3.2. Sensitivity analysis

We initially seek out about more ten kinds of the brands in the network; we deleted
some brands for poor quality or too high price or ugly style by Delphi technique. We
identified the most dominant to draw out the five brands as the final candidate
alternatives. Notice, we wouldn’t require an excessive number of pair-wise comparisons
and an exceeding judgment capabilities for AHP applying in selection system. These
five brands were screened using an AHP-based stratification scheme analogous
followed Erkut and Moran [24]. Therefore, the five brands were established for further
consideration to appear on the bottom level shown in fig. 2; (1) Dahon Espresso 21, (2)
Shimano Deore LX9, (3) KHS F20-R 18, (4) Merida VW-Blumotion, (5) Doppelganger
FX04.
As explained earlier, we recorded 35 critical factors in the spreadsheet, and then we
can use the Delphi technique to obtain the consensus ratings more 2.30 values to
transfer them into 12 sub-criteria, 4 success criteria and 2 strategic issues. Furthermore,
the priority weights of the criteria are determined as the local weights; the composite
priority weights of the sub-criteria are determined for the different product attributes;
the sub-criteria weights multiplied by the local criteria weights were the global priority
weights, we can find each strategic issue weight by multiplying the global priority
weight getting the assessment rating, and adding the resulting values up. Or, as
suggested [25], we would find the global priority weights of assessment systems by the
Pair-wise comparison with judgment matrices, and use them to select the highest weight
as a best product.
In this case study, the purchasing evaluator assigned the ratings to each criterion with
respect to each sub-criterion. Example, the evaluator qualified the cost weight is more
important than the quality weight as 1.25:1 and we wish to compute that thought the
n×n matrix A to get the cost priority weight 0.556 and the quality weight 0.444 (see
Table 2). Here, the priority weights of each criterion rating is already determined, then

8
Journal of China University of Science and Technology Vol.__-2011.11

we use the priority weights to contrast each sub-criterion on a spreadsheet, and it is


formatted the n×n matrix A to structure and determine the global priority weights by
composite local priority weights (see Table 3). Notice, these global priority weights
need to be normalized based on composite success factors. Above was applied the
success factors to the folding bike selected data switching from network (see Table 4).

3.3. Selecting the optimal product to satisfy a customer

In this section, we make the attributes score for success factors. For example, the
alternative reputation has superior and good both, we make an interval between 1 and 0;
the superior score 0.113 and the good score 0.005 computed by the n×n matrix A.
However, the alternative reputation’s score (RPT) 0.113 multiplied by the global weight
(GW) 0.064 are the Dahon E21 WG 0.007. And then plus all attributes’ GW equals total
values (TV) 0.268. Final makes the n×n matrix A of the 5 alternative to get the Dahon
E21 represents renormalized scores (RS) 0.268 (see Table 5).
We find the Merida VW-B has the highest priority weight. However, it was selected
as the best one to satisfy customer. Interestingly, the evaluation team used the pre-
existing the selection process, this actual assessment process was modified the best
solution approach using the proposed AHP model. And the decision making was
considered to gain the highly satisfactory and able to be accepted under the comparative
result. This approach was proved the regularity value selection.
This result also shows that both the pre-existing product selection process and the AHP
successful decision. Using the AHP approach, when the criteria and sub-criteria are
clearly defined, and the selection problem is structured as the systematically matrix.
Enable decision-makers examine the strengths and weaknesses of the product attributes,
to compare with their relative criteria. It is easier the evaluation method to arrive at a
consensus decision. We also used the proposed AHP model in another product selection
problem, and same can arrived at a selection decision that was considered to be
successful [27]. The AHP model involved selecting a new platform to replace the
existing complex data problem. Again, we used the general model and rated the product
attributes given each sub-criteria, and determined the corresponding global priority
weights. Based on these priority weights we would select the best one. We can conclude
that the proposed AHP model can help facilitating the decision making and significantly

9
reducing the selection time. Also, we hope the successful method would be used facing
other problem in the future.
As this actual case, we had conducted the whole assessment process, understand and
verify the AHP model, and we believe it was in worth trust. All evaluators who assigned
pair-wise comparison judgments appear to be satisfied. Also, the managers of the
concerned departments were happy with the application of the proposed AHP model. To
overcome the selection problems for assessing pair-wise comparison judgments, the
evaluators were first trained on AHP principles and assessment techniques. The
questionnaires were mailed to obtain the pair-wise comparisons from experts. Gaining
the support and commitment to evaluation team from technical and management
personnel would also encourage the continued application of the proposed model.

4. Conclusion

This study is developed an evaluation method to select the critical factors by Delphi
technique, and conducted a survey sequential questionnaire to select the success factors.
In general, the imposing issue of the impact on the management implication is
obviously offered an evaluative method by the AHP analysis. A detailed sensitivity
analysis has already been performed to identify the list of the variations by Delphi
technique. Furthermore, we identified the strategic issues, criteria and sub-criteria, and
alternatives by a survey sequential questionnaire. Finally formulated an AHP based
model. This is a new method of the hybrid evaluation about the objective survey,
effective evaluation, scientific and systematical assessing technique for applicable any
selection problem. In this case, we find the global priority weights that can easily
compute using Excel Software from determined final renormalized scores. First, the
synthesis values of four criteria weights are showed brand (0.529), price (0.471), style
(0.643) and specification (0.357). Next, an assessment can compute the constructing
sub-criteria into the global weight. On the basis of data, the Merida VW-B was ranked
as highest (0.213) from five alternatives.
It should be noted, the data collection would increase the number of criteria and sub-
criteria, as well as the number of alternatives in the selection structures. In fact, we can
increase the collective data and determine the priority weights to prevent the evaluator
subjective views. One of the reasons, as several case studies in the literature using the

10
Journal of China University of Science and Technology Vol.__-2011.11

AHP indicates the use of three to seven evaluators. We suggested short-listing the
number of criteria to apply the Excel Software to reduce the complex computation
process. Also, as shown here, the number of success factors can be grouped to minimize
the number of criteria and sub-criteria used the Delphi technique. In this method, we can
conduct sensitivity analysis and determine the reasonable number of assessment criteria.
The proposed AHP model is applied to this case, the decisions had reached customer
anticipative goal. However, the AHP model has the capable of handling multiple
conflicting views such as the minimization of cost, the maximization of market
opportunities. This method enables decision makers to examine the strengths and
weaknesses by comparing with respect to criteria and sub-criteria. Moreover, AHP
model can significantly reduce the evaluation time and computation process. In
addition, the results can be transferred to a spreadsheet to arrive at an optimum
selection.
From the results of the case studies, it can be concluded that improving the decision
making process is desirable. The hybrid AHP model developed in this paper can be used
as a basis model for purchasing systems. In addition, the attributes rating of assessment
product based on the separately product’s catalog. Similarly, any relative to the
purchasing problems can be included in the routine evaluation fashion.

Reference

[1] Narasimahn R. An analytical approach to supplier selection. Journal of purchasing


and Materials Management 1983;19(4):27-32.
[2] Partovi FY., Burton J., Banerjee A., Application of analytic hierarchy process in
operations management. International Journal of Operations and production
Management 1989;10(3):5-19.
[3] Nydick RL., Hill RP., Using the analytic hierarchy process to structure the supplier
selection procedure. Journal of purchasing and Materials Management
1992;25(2):31-6.
[4] Kogut, B., 1988. Joint ventures: theoretical and empirical perspectives. Strategic
Management Journal 9 (4), 319-332.
[5] Hendrick, T.E., 1997. In: Purchasing Consortiums: Horizontal Alliances Among
Firms Buying Common Goods and Services: What? Who? Why? How?. Center for

11
Advanced Purchasing Studies, Tempe.
[6] Hoffimann, W., Schlosser, R., 2001. Succss factors of strategic alliances in small
and medium-sized enterprises, an empirical study. Long Range Planning 34 (3),
357-381.
[7] Schotanus, F., 2007. Horizontal cooperative purchasing. Ph.D. Dissertation,
University of Twente, Enschede, The Netherlands.
[8] Schotanus, F., Jan Telgen, J., Luitzen B,. 2010. Critical success factors for
managing purchasing groups. Journal of Purchasing & Supply Management 16,
51-60.
[9] Delbecq, A.L., Van de Van, A.H., (1975). Gustafson DH. Group Techniques for
Program Planning: a Guide to Nominal Group and Delphi Process. Illinois: scott,
foresman, Glenview.
[10] Dijk, J.V., (1986). Method in applied social research: Special characteristic and
quality standards. Qual. Quant. 20(4), 357-370.
[11] Robbins, S.P., (1994). Management. Prentice Hall, New Jersey.
[12] Anderson, D.R., Sweeney, D.J., Willuams, T.A., (2001). Quantitative Methods for
Business. South-Western College, Cincinnati.
[13] Chaw, P.C., (2001). Application of Delphi method in selection of procurement
systems for construction projects. Constr. Manage. Econ. 19(7), 699-718.
[14] Hsu, P.E., Chen, B.Y., (2007). Developing and implementing a selection model for
bedding chain retail store franchisee using Delphi and fuzzy AHP. Qual. Quant. 41,
275-290.
[15] Kenneth T., Gordon B., Simon N., Steve S., Peter D., Keith B., Jonathan D.,
Knowledge elicitation for Validation of a neonatal ventilation expert system
utilizing modified Delphi and focus group techniques. Int. J. Human-Computer
Studies 2010;68:344-354.
[16] Dickson GW., An analysis of vendor selection systems and decisions. Journal of
purchasing 1966;2:5-17.
[17] Beck MP., Lin BW., Selection of automated office systems: A case study. OMAGA
1981;9(2):169-76
[18] Zviran MA., Comprehensive methodology for computer family selection. Journal
Systems Software 1993;22:17-26.
[19] Saaty TL., The analytic hierarchy process. New York: McGraw Hill;1980.
[20] Saaty TL., How to mark a decision: the analytic hierarchy process. European

12
Journal of China University of Science and Technology Vol.__-2011.11

Journal of Operational Research 1990;48:9-26.


[21] Sinuany SZ., Mehrez A., Tal AG., Binyamin S., The location of a hospital in a rural
region: the case of the Negev. Elsevier Science 1995;3:255-66.
[22] Vargas LG., An overview of the analytic hierarchy process and its applications. Eur
J Oper Res 1990;48:2-8.
[23] Tummala VMR,. Wan YW., Analytic hierarchy process (AHP) in practice: a survey
of applications and recent developments. Journal of Mathematical Modelling and
Scientific Computing 1994;3(1):1-38.
[24] Erkut E., Moran SR., Locating obnoxious facilities in the public sector: an
application of the analytic hierarchy process to municipal landfill siting decisions.
Socio-Econ Planning Sci 1991;25(2):89-102.
[25] Liberatore MJ., An extension of the analytic hierarchy process for industrial R&D
project selection and resource allocation. IEEE Transactions on Engineering
Management 1987;EM-34(1):12-8.
[26] Liberatore MJ., Nydick RL,. Sanchez PM., The evaluation of research papers (or
how to get an academic committee to agree on something). Interfaces
1992;22(2):92-100.
[27] Tam CY., An application of the analytic hierarchy process in vendor selection of a
telecommunications system. MSc Engineering and Engineering Management, city
University of Hong Kong, Kowloon, Hong Kong, 1996.
[28] Saaty TL., Vargas LG.. Decision making in economic, political, social, and
technological environments with the analytic hierarchy process. Pittsburgh: RWS
Publications, 1994.

First Round
16
participants

Second Round Second Round


Focus group Modified Delphi

Fig. 1. Design of the study

13
Table 1 the survey questionnaire for drawing success factors

Modify factors Identify factors Sub-criteria Chief-criteria


Cost*** (2.90) Cost***(2.90) Cost***(2.90) Cost***(2.90)
Brand**(2.75) Brand**(2.75) Brand**(2.75)
Reputation*, Guarantee*, Reputation*,
Service*, Operation facility, Guarantee*, Service*,
Configuration, Network
information, Advertisement, Price**(2.85)
Reliability, Availability. Unit cost*, Accessories Price**(2.85)
Price**(2.85) cost*(2.40),
Unit cost*(2.60), Accessories
cost*(2.40), Function item,
Maintenance cost,
Technology development.
Quality***(2.90) Quality***(2.90) Quality***(2.90) Quality***(2.9
Style**(2.70) Style**(2.70) Style**(2.70) 0)
Style design*(2.55), Style design*(2.55),
Weight*(2.50), High, Color, Weight*(2.50),
Handlebar, Frame, Folding
size, Saddle. Specification**(2.65) Specification**(2.65)
Specification**(2.65) Material*(2.55),
Material*(2.55), Shifters*(2.35), Shock
Shifters*(2.35), Shock absorber*(2.50),
absorber*(2.50), Tire*(2.60), Tire*(2.60), Brake
Brake system*(2.65), Fork, system*(2.65),
Chain wheel
35 18 6 2

14
Journal of China University of Science and Technology Vol.__-2011.11

Level 1: Selection of purchasing folding bike


Goal Overall weighted index (OWI)

Level 2:
Strategic Issues Cost Quality (QLT)

Level 3: Brand Price Style Specification


Criteria (BR) (PRC) (STL) (SPF)

Reputation Unit cost Style design Frame material


(RPT) (UC) (SD) (FM)
Shifters (SFT)

Guarantee Accessories cost Weight (WT) Shock ABS


(GRT) (AC) (SABS)
Tire (TIR)

Service Brake system


Level 4: (SVC) (BS)
Sub-criteria

Level 5: Dahon Shimano KHS F20-R Merida VW- Doppelganger


Alternative Espresso 21 Deore LX9 18 Blumotion FX04

Fig. 2 AHP model for product selection of a purchasing system

Table 2
Pair-wise comparison judgment matrices of purchasing selection problem
Goal Cost Quality Priority
Cost 1 1.25 0.556
Quality 0.8 1 0.444
λ=2.000, CI=0.000, CR=0.000
Cost Brand Price Priority
Brand 1 1.125 0.529
Price 0.889 1 0.471
λ=2.000, CI=0.000, CR=0.000
Quality Style Specification Priority
Style 1 1.800 0.643
Specification 0.556 1 0.357

15
λ=2.000, CI=0.000, CR=0.000
Brand Reputation Guarantee Service Priority
Reputation 1 0.571 0.556 0.219
Guarantee 2.000 1 0.875 0.387
Service 1.600 1.143 1 0.393
λ=3.015, CI=0.007, CR=0.013
Price Unit cost Accessories cost Priority
Unit cost 1 1.500 0.600
Accessories cost 0.667 1 0.400
λ=2.000, CI=0.000, CR=0.000
Style Style design Weight Priority
Style design 1 0.571 0.348
Weight 2.000 1 0.652
λ=2.069, CI=0.069, CR=0.000
Specification FM SFT SABS TR BS Priority
Frame material 1 0.778 1.400 1.400 0.875 0.206
Shifters 1.286 1 1.800 1.500 1.125 0.255
Shock ABS 0.714 0.556 1 0.833 0.625 0.142
Tire 0.857 0.667 1.200 1 0.750 0.170
Brake system 1.143 0.889 1.600 1.333 1 0.227
λ=5.040 ,CI=0.010, CR=0.009

Table 3
Composite priority weights for critical success factors
Strategic Local Criteria Local Sub-criteria Local Global
issues weights weights weights weights
Cost 0.5560 Brand 0.529 Reputation 0.219 0.1289
Guarantee 0.387 0.2279
Service 0.393 0.2314
Price 0.471 Unit cost 0.600 0.1571
Accessory C 0.400 0.1048
Quality 0.444 Style 0.643 Style design 0.348 0.0994
Weight 0.652 0.1861

16
Journal of China University of Science and Technology Vol.__-2011.11

Specification 0.357 Frame M 0.206 0.0327


Shifters 0.255 0.0404
Shock ABS 0.142 0.0225
Tire 0.170 0.0269
Brake 0.227 0.0360
system

Table4
Application of the AHP model to folding bike selection of a data switching network
Alternatives
Factors Dahon Shimano KHS F20- Merida DPG
E21 D lx9 R 18 VW-B FX04
Reputation Superior Superior Superior Superior Good
Brand Guarantee 5 Years 5 Years 1 Year 1 Year 1 Year
Service Better Better Better Better Better
Unit cost NT$15000 NT$54800 NT$26500 NT$14999 NT$12000

Price Accessory
No NT$1200 No No No
cost
Style
50 80 80 80 75
Style design
Weight 13.3KG 10.3KG 11.5KG 11.87KG 12.6KG
Frame
Aluminum Aluminum Aluminum Aluminum Aluminum
material
Shifters 21speed 9speed 18speed 24speed 7speed
Shock No No No No
Specification Adjuster
ABS Adjuster Adjuster Adjuster Adjuster
Tire 26×1.5 18×1.5 20×1.5 20×1.5 20×1.5
Brake PROMAX F&R F&R F&R F&R
system V Aluminum stronger V Aluminum stronger C

Table 5
Application of the AHP model to product selection of a data switching network
SCI GW Dahon E21 Shimano D lx9 KHS F20-R 18 Merida VW-B DPG FX04

17
Score GW Score GW Score GW Score GW Score GW

Cost
BR
RPT 0.064 0.113 0.007 0.113 0.007 0.113 0.007 0.113 0.007 0.076 0.005
GRT 0.114 0.067 0.008 0.067 0.008 0.040 0.005 0.040 0.005 0.040 0.005
SVC 0.116 0.750 0.087 0.750 0.087 0.750 0.087 0.750 0.087 0.750 0.087
Price
UC 0.157 0.850 0.134 0.548 0.086 0.735 0.115 0.850 0.134 0.870 0.137
AC 0.105 0.010 0.001 0.050 0.005 0.010 0.001 0.010 0.001 0.050 0.005
QLT
Style
SD 0.099 0.101 0.010 0.151 0.015 0.151 0.015 0.151 0.015 0.143 0.014
WT 0.186 0.061 0.011 0.151 0.028 0.151 0.028 0.151 0.028 0.131 0.024
SPF
FM 0.033 0.050 0.002 0.050 0.002 0.050 0.002 0.050 0.002 0.050 0.002
SFT 0.040 0.108 0.004 0.027 0.001 0.087 0.004 0.128 0.005 0.014 0.001
SABS 0.023 0.059 0.001 0.118 0.003 0.059 0.001 0.059 0.001 0.059 0.001
Tire 0.027 0.050 0.001 0.081 0.002 0.081 0.002 0.071 0.002 0.071 0.002
BS 0.036 0.042 0.002 0.059 0.002 0.084 0.003 0.059 0.002 0.071 0.003
TV* 0.268 0.246 0.270 0.289 0.285
RS* 0.197 0.181 0.199 0.213 0.210
*TV denotes total value, *RS represents renormalized scores.

18