Expert Systems with Applications 36 (2009) 7594–7603

Contents lists available at ScienceDirect

Expert Systems with Applications

journal homepage: www.elsevier.com/locate/eswa

Six-Sigma project selection using national quality award criteria and Delphi
fuzzy multiple criteria decision-making method
Taho Yang *, Chiung-Hsi Hsieh
Institute of Manufacturing Engineering, National Cheng Kung University, No. 1 University Road, Tainan 70101, Taiwan

a r t i c l e i n f o a b s t r a c t

Keywords: Six-Sigma is a tactical tool of significant value in achieving operational excellence. The project selection
Delphi method decision, under a resources constraint, is the early stage of implementation for a Six-Sigma intervention.
Fuzzy set The project selection decision is challenging due to its fuzzy group decision-making aspect inherent to
Multiple criteria decision-making the problem. The present study proposes to adopt national quality award criteria as the Six-Sigma project
National quality award
selection criteria, and proposes a hierarchical criteria evaluation process. The strategic criteria are eval-
Six-Sigma project
uated by the management team using a Delphi fuzzy multiple criteria decision-making method. Then, the
tactical sub-criteria which contain additional operational issues are evaluated by the Six-Sigma Cham-
pion. The proposed methodology is successfully applied in solving the project selection problem deriving
from a component manufacturer. The empirical outcomes are promising. Moreover, the results show that
the higher a project’s priority is, the greater the financial gains will be on completion of the project.
Accordingly, the proposed methodology can prioritize the financial gain – which is the key performance
indicator for a Six-Sigma project. Additionally, the quality status of the case company has been signifi-
cantly improved through implementation of the Six-Sigma project. The systematic evaluation process
also influences employees to adopt an analytical operations philosophy. Moreover, the commercial objec-
tives of the company are brought into focus by the proposed methodology.
Ó 2008 Elsevier Ltd. All rights reserved.

1. Introduction cited as the first winner of America’s Malcolm Baldrige national

Continuous improvement towards business performance excel-
lence is the competitive edge for commercial firms to survive in
highly competitive markets (Deming, 1986). Among the many
business improvement approaches available, the Six-Sigma ap-
proach has been recognized as one of the most effective methods.
Six-Sigma has been launched all over the world and many com-
panies testify to its pivotal role in their success (Hutchins, 2000).
Its application focuses resources on reducing variation in all pro-
cesses, including manufacturing processes, administrative pro-
cesses etc. The unambiguous measure on the improvement work
is referred to as Six-Sigma. Eminent examples of Six-Sigma compa-
nies include Motorola, General Electric, AlliedSignal (now Honey-
well), Lockheed Martin, Polaroid, Sony, Honda, America Express,
Ford, and Solectron. Indication of significant success at Motorola,
quickly become apparent. From 1987 to 1997 Motorola achieved
a fivefold growth in sales, with profits climbing nearly 20% per
year, cumulative saving at 14 billion US dollars and stock price
gains compounded to an annual rate of 21.3%. Motorola was also
* Corresponding author. Tel.: +886 6 2090780; fax: +886 6 2085334.
E-mail addresses: tyang@mail.ncku.edu.tw, z8708045@email.ncku.edu.tw
(T. Yang).
quality award in 1988 (Eckes, 2006).
Six-Sigma is a tactical tool of great value in achieving opera-
tional excellence. Operational excellence is required for the overall
attainment of business excellence – a notion that also requires cus-
tomer-related, financial, and marketplace performance excellence
(Edgeman, 2000). Six-Sigma explicitly links the tactical activities
with those strategic ones. Thus, an appropriately configured and
deployed Six-Sigma program may be highly consistent with the re-
sults-orientation underlying international quality awards, such as
the European Quality Award, America’s Malcolm Baldrige national
quality award (MBNQA), Canada’ Excellence Award, and the Aus-
tralian Quality Award (Klefsjö, Wiklund, & Edgeman, 2001).
One of the key emphases of a quality award is, for a company to
achieve sustainable financial success. In the instance of MBNQA,
the award winning firms reported a 44% higher stock-price return,
48% higher growth in operating income, and a 37% higher growth
in sales than the control group of firms (Davis & Stading, 2005).
Organizations are using various criteria to help them during
implementation efforts to evaluate themselves against criteria to
determine how well their improvement efforts are progressing.
Sets of criteria that the majority of organizations uses include
Deming prize categories, Juran’s ten points, Crosby’s fourteen
points, and the MBNQA criteria (Ritter, 1993).

0957-4174/$ - see front matter Ó 2008 Elsevier Ltd. All rights reserved.
doi:10.1016/j.eswa.2008.09.045
 T. Yang, C.-H. Hsieh / Expert Systems with Applications 36 (2009) 7594–7603
A comparison between the national quality award and the Six-
Sigma program conducted by Przekop (2006) shows that the core
emphases of both are similar. For example, similarities are evident
in the procedure-focusing, customer-focusing, cooperation, data
driven management, and the strategic planning. Therefore, the
quality award criteria are also logically exacting for the project
selection criteria of a successful Six-Sigma program. The project
selection is a priori for the implementation of a Six-Sigma program.
In fact, the project selection for Six-Sigma program is often the
most important and difficult part (Pande, Robert, & Roland, 2002).
The Six-Sigma project selection problem falls within fuzzy mul-
tiple criteria decision-making (FMCDM). The present study pro-
poses to adopt the national quality award as the selection
framework and proposes a FMCDM method for the criteria evalu-
ation and project selection. The FMCDM methods have been devel-
oped due to the imprecision in assessing the relative importance of
criteria and the performance ratings of alternative techniques.
Imprecision may arise from a variety of reasons: unquantifiable
information, incomplete information, unobtainable information,
and partial ignorance. To resolve this difficulty, fuzzy set theory
has been adopted for the decision-making process (Bellman & Za-
deh, 1970).
For the FMCDM problem, the decision-makers use linguistic
variables to evaluate the importance of criteria and the ratings of
alternatives (or projects) with respect to various criteria. The pres-
Appendix A1
Taiwan national quality award criteria and sub-criteria
Criteria Sub-criteria
C1 Leadership C11 Business concepts/values
C12 Organization mission/vision
C13 Senior executive leadership
C14 Total quality culture
C15 Corporate citizenship
C2 Strategic management C21 Innovation values
C22 Business model and strategic
planning
C23 Strategy development and
development
C3 Research and innovation C31 Research and innovation strategy
and process
C32 Research and innovation input
C33 Research and innovation result
measurement
C4 Customer/market development C41 Product/service and market strategy
C42 Customer and business information
management
C43 Customer relationship management
C5 Human resource and knowledge C51 Human resource planning
management C52 Human resource development
C53 Human resource usage
C54 Employee relationship management
C55 Knowledge management
C6 Information strategy application and C61 Information strategy planning
management C62 Application of computer network
C63 Information application
C7 Process management C71 Product/process management
C72 Management for off-line activity
C73 Organizational relationship
management
C8 Business result C81 Customer satisfaction
C82 Market development performance
C83 Company financial performance
C84 Performance of human resource
development
C85 Information management
performance
C86 Process management performance
C87 Performance on innovation and core
competitive
C88 Social evaluation
7595
ent study will adopt a case from Taiwan for empirical examples.
Accordingly, Taiwan National Quality Award (TNQA) criteria will
be adopted as the Six-Sigma project selection criteria (as shown
in Appendix A1). The TNQA has a hierarchical structure. It has eight
strategic criteria. Each strategic criterion has its associated sub-cri-
teria. In total, there are 33 sub-criteria.
For the strategic criteria, it is logical to incorporate group opin-
ions from the different management positions. Accordingly, it is a
group decision-making problem that will be solved by the Delphi
method in this study. The fuzzy concept was embedded in Delphi
methods by calculating the average weighting of all the criteria
based on experts’ experience (Chang, Huang, & Lin, 2000). Then,
the sub-criteria, with respect to each candidate project, will be
evaluated by the Champion who is a qualified Six-Sigma expert
(Master Black-belt).
The remainder of this paper is organized as follows: Section 2
reviews the pertinent literature. Section 3 provides details of the
proposed Delphi FMCDM methods. The background information
for the case study and empirical illustrations are discussed in Sec-
tion 4. Conclusions and future research opportunities are ad-
dressed in the final section.
2. Literature review
The Six-Sigma method utilizes a well-disciplined approach. The
unique features of the Six-Sigma approach are as follows: (1) se-
quences and links improvement-tools into an overall approach
(known as DMAIC), (2) integration of the human and process ele-
ments for improvement using a belt-based organization (Cham-
pion, Black Belt, and Green Belt), (3) attention to bottom-line
results and the sustaining of gains over time (Su, Chiang, & Chiao,
2005).
Six-Sigma is already successfully applied in individual activities
and industries such as witnessed by the improvement in the auto-
mobile industry’s manufacturing flow (Kalamdani & Khalaf, 2006),
and in quality of integrated-circuit design (Su et al., 2005). Das
(2005) applied Six-Sigma to reduce procurement delay. Six-Sigma
is applied using a project management, under resource constraints.
The project selection-decision, to maximize the financial outcomes,
is often challenging for a company.
Breygogle (1999) suggested that companies can consider four
dimensions of the balanced score card, namely financial, customer,
internal business process and learning, and growth as the criteria
for project selection. Snee and Rodebaugh (2002) identified that
projects need to link with the strategic goal. Mark (2001) stated
that projects should focus on activities critical to quality (CTQ)
and financial performances. Brue (2002) considers that project
selection should acknowledge resources and time. George, Row-
lands, Price, and Maxey (2006) argues for recognition of the busi-
ness voice, customer voice, and process voice for project selection.
Przekop (2006) argued that Six-Sigma has the same content
with that of the American national quality award criteria. Seethar-
aman, Sreenivasan, and Boon (2006) found that a national quality
award winner also showed improved performance in both sales
and revenue. Thus, national quality award criteria should be a po-
tential framework for the Six-Sigma project selection criteria.
The project criteria evaluation is a FMCDM problem where fuz-
zy assessments and multiple expert opinions can be considered.
Human opinions are often in conflict because of group decision-
making in a fuzzy environment. Various approaches to different as-
pects of decision problems with vague data have been published,
and a significant amount of literature is available on FMCDM, such
as: Chang, Wang, and Wang (2006), Chou, Chang, and Shen (2008),
Coffin and Taylor (1996), Greco, Matarazzo, and Slowinski (2002),
Ölcer and Odabasi (2005), Wang and Lin (2003), Wang (2008),
7596 T. Yang, C.-H. Hsieh / Expert Systems with Applications 36 (2009) 7594–7603

Xu and Chen (2007), Yang and Chou (2005), and Yang and Hung This function maps the elements of a fuzzy set into a real value
(2007). belonging to the interval between 0 and 1.
Recent development has extended the FMCDM to a group deci- ~ in a universe of discourse X is characterized by a
A fuzzy set a
sion-making problem, as investigated by Chang, Tasuhiro, and Toz- membership function la~ ðxÞ which associates with each element x
awa (1995), Chang et al. (2000), Chang and Wang (2006), Chang, in X, a real number in the interval [0, 1]. The function value la~ ðxÞ
Wu, and Chen (2008), Cheng and Lin (2002), Liu and Chen is termed the grade of membership of x in a ~ (Yang, Chen, & Hung,
(2007), Yeh, Cheng, and Chi (2007), and Zeng, An, and Smith 2007; Zadeh, 1965).
(2007). Among the commonly used fuzzy numbers, triangular and trap-
Cheng and Lin (2002) utilized a fuzzy Delphi method to adjust ezoidal fuzzy numbers are likely to be the most common, due to
the fuzzy rating of every expert, and so achieve the consensus con- their simplicity in modeling and ease of interpretation (Petroni &
dition. The experts’ opinions are described by linguistic terms Rizzi, 2002).
which are expressed in trapezoidal fuzzy numbers. It then took A triangular fuzzy number a ~ can be defined by a triplet (a1, a2,
the operation of fuzzy numbers to calculate the mean of fuzzy rat- a3). Its mathematical form is shown as Eq. (1) (Kaufmann & Gupta,
ings and the mean of weight. The aggregated fuzzy numbers were 1985) and Fig. 2.
solved by multiplying the fuzzy decision matrix with the corre- 8
>
> 0; x 6 a1
sponding fuzzy attribute weightings. >
> xa1
< ; a1 < x 6 a2
Zeng et al. (2007) used standardized trapezoidal fuzzy number a2 a1
la~ ðxÞ ¼ a3 x
ð1Þ
(STFN) to capture and convert experts’ fuzzy information and sub- >
> ; a2 < x 6 a3
>
> a3 a2
jective judgment for the group FMCDM problem. The experts’ opin- :
0; x > a3
ions can be expressed by a precise numerical value, a range of
numerical values, a linguistic term or fuzzy number. In other ~ be defined by a triplet (b , b , b ). Then, the basic operations on
Let b 1 2 3
words, the members of the decision-making team have the flexibil- fuzzy triangular numbers are shown as Eqs. (2) and (3).
ity to use different evaluation measures depending on their indi- ~ ¼ ða  b ; a  b ; a  b Þ for multiplication
~b
a 1 1 2 2 3 3 ð2Þ
vidual knowledge and confidence. The STFN is employed to
convert these experts’ judgments into a universal format for the ~ ~
a þ b ¼ ða1 þ b1 ; a2 þ b2 ; a3 þ b3 Þ for addition ð3Þ
composition of group references. The fuzzy aggregation is used to
A trapezoidal fuzzy number a~ = (a1, a2, a3, a4), a1 6 a2 6 a3 6 a4, and
create group decisions, and then defuzzification is employed to
its membership function is shown as Eq. (4) and Fig. 3 (Jang, Sun, &
transform the STFN scales into numerical scales for the computa-
Mizutani, 1997).
tion of priority weights.
8
> 0; x 6 a1
>
>
3. Proposed methodology >
> xa1
; a 6 x 6 a2
>
< a2 a1 1

la~ ðxÞ ¼ 1; a2 6 x 6 a3 ð4Þ

The implementation of the proposed methodology has four >
>
>
> xa4
; a3 6 x 6 a4
steps as shown in Fig. 1. >
> a a
: 3 4
The first step determines the project selection criteria. Then, the 0; x P a4
strategic criteria weightings are determined by fuzzy group deci- ~ be defined by a triplet (b , b , b , b ). Then, the basic operations
Let b 1 2 3 4
sion-making method – Delphi FMCDM method. The STFN approach on fuzzy trapezoidal numbers are shown as Eqs. (5) and (6).
from Zeng et al. (2007) is adapted for the fuzzy group decision-
making. Next, the sub-criteria weights with respect to each alter- a ~ ¼ ða1  b1 ; a2  b2 ; a3  b3 ; a4  b4 Þ for multiplication
~b ð5Þ
native (project) are evaluated by fuzzy linguistic variables. The fuz- ~ ¼ ða1 þ b1 ; a2 þ b2 ; a3 þ b3 ; a4 þ b4 Þ for addition
~þb
a ð6Þ
zy Delphi method from Cheng and Lin (2002) is adopted for this
evaluation. Finally, the project selection decision is analyzed by
fuzzy defuzzification. What follows in the following section is the 3.2. The project selection criteria decision
introduction of basic fuzzy theory, followed by detailed discussion
of the step-by-step process. TNQA are adopted as the Six-Sigma project selection criteria for
the present study. The top level contains the strategic criteria
3.1. The basic fuzzy set theory which are evaluated by Delphi FMCD. The second level contains
tactical criteria which are evaluated by FMCDM. Based on the
A fuzzy set contains elements that have varying degrees of TNQA criteria, there are three decision modules: (i) the group
membership in the set. Elements of a fuzzy set are mapped to a
universe of membership value using a function-theoretic form.

Step 1: Determine the Step 3: Use fuzzy

project selection criteria linguistic variables to
evaluate the sub-criteria
weights with respect to
each project candidate
Step 2: Use fuzzy group
decision-making method
to determine the strategic Step 4: The project
criteria weights ranking by fuzzy
defuzzification
a1 a2 a3

Fig. 1. The proposed Six-Sigma project selection methodology. Fig. 2. Triangular fuzzy membership function.
 T. Yang, C.-H. Hsieh / Expert Systems with Applications 36 (2009) 7594–7603
a a a a
1 2 3 4
Fig. 3. Trapezoidal fuzzy membership function.
decision-making for the top level criteria, (ii) the sub-criteria eval-
uation by Champion, and (iii) the aggregated project ranking for
the project selection decision. The details of the proposed method-
ology are discussed next.
3.3. The group decision-making for the strategic criteria weights
The Delphi FMCDM is a methodology in which subjective data
of experts are transformed into quasi-objective data using the sta-
tistical analysis and fuzzy operations.
~ i be the fuzzy weight for the ith criterion, i = 1, 2, ... , n.
Let W
There are p persons in the decision-making group who are usually
from the company executives. As different experts have different
impacts on the final decision, a contribution factor (CF) is intro-
duced into the group decision-making process to distinguish an ex-
pert’s competence. CF will be allocated to experts on the basis of
their experience, knowledge and expertise. The kth expert Ek is
assigned a contribution factor ck, where ck 2 [0, 1], and
c1 + c2 +    +cp = 1.
Members in the decision group are required to provide their
judgments on the basis of their knowledge and expertise for each
strategic criterion. There are four options for an expert to evaluate
the criteria, they are: (i) a precise numerical value, (ii) a range of
numerical values, (iii) a linguistic term and (iv) a fuzzy number. The
present study proposes to use STFN to normalize the above alter-
native evaluation methods (Zeng et al., 2007). The conversion pro-
cedure is discussed as follows.
3.3.1. Convert preference into STFN for each expert’s evaluation
Let U be the universe of discourse, U = [0, u] where u is an inte-
ger number. A STFN can be defined as A* = (al, ar, as, au), where
0 6 al 6 ar 6 as 6 au 6 u, and its membership function is shown as
Eq. (7) and Fig. 4.
8 taking over a project. Champions are advocates who ‘fight’ for the
>
> ðx  al Þ=ðar  al Þ for al 6 x 6 ar
>
< 1 for ar 6 x 6 as
uA ðxÞ ¼
>
> ðau  xÞ=ðau  as Þ for as 6 x 6 au
>
: jects that align with the company strategy and can be understood
0 for otherwise
For the four input types, the conversion is as follows. When al = ar =
as = au, a STFN is a numerical value; when al = ar and as = au, a STFN
is a range of numerical values; when ar = as, a SFTN becomes a tri-
angular fuzzy number. Finally, the expert can directly input a trap-
ezoidal fuzzy number.The choice among the different evaluation
methods is dependent on the level of confidence among the differ-
ent experts and on the individual’s preference. As a rule of thumb, if
adequate information is obtained and the criterion is quantitative
measurable, an expert is likely to provide a precise numerical value
or a possible range of numerical values. However, experts some-
times find that it is hard to give numerical values due to uncertain-
7597
A*
l r l
(x-a )/(a -a ) u u s
(a -x)/(a -a )
a
l
a
r
a
s
a
u x
Fig. 4. The SFTN membership function.
ties involved or the criterion is quantitative immeasurable, then a
linguistic term or a fuzzy number can be used in the proposed
model.
3.3.2. Aggregate experts’ evaluations
Let ~Sik be the STFN for criterion i evaluated by expert Ek. Each
STFN represents a preference provided by an expert in the decision
group on the basis of available information and subjective judg-
ments. Let Si ; i = 1, 2, . . . , n, be the fuzzy aggregated score of the cri-
terion i. Then, Si is solved by applying the fuzzy weighted
trapezoidal averaging operator, which is defined by Eq. (8).
Si ¼ ~Si1  c1 þ ~Si2  c2 þ    þ ~Sik þ    þ ~Sip  cp ð8Þ
Si
The resulting will be the weight for criterion i. By convention, the
criteria weights are usually ranging from 0 to 1. Thus, Si is con-
verted to W~ i by Eq. (9).
~ i ¼ S =u; i ¼ 1; 2; :::; n
W ð9Þ
i
3.4. The sub-criteria evaluation by Champion
The sub-criteria weights are determined by a Six-Sigma project
Champion who is responsible for the success of a project. The
Champion is the decision-maker for the project resources alloca-
tion, conflict management, communication, etc.
When the strategic criteria have been determined by the com-
pany’s management, the sub-criteria are then determined by the
Champion. For a Six-Sigma project, the Master Black Belt (MBB)
is an expert in all areas of Six Sigma. Within the company, the
MBB serves as an official teacher for Champions and Black Belts.
The MBB should be skilled at facilitating problem solving, without
cause and remove barriers to success. They pave the way for Black
Belts to do their jobs. Black Belts are dedicated to assigned projects
ð7Þ
and work full time to drive project success. Champions select pro-
and embraced by project teams. Champions select Black Belt can-
didates and projects and ultimately establish clear and measurable
goals for projects (Przekop, 2006).
In the present study, the Champion uses a trapezoidal fuzzy
number to measure the weights of the 33 sub-criteria which are
clustered into eight criteria groups, as shown in Appendix A2. As-
sume that there are m projects for the project selection process. Let
Pj, j=1, 2, . . . , m, be the jth project for the priority ranking and
evaluation.
Let X ij ; i ¼ 1; 2; :::; n and j = 1, 2, . . . , m, be the fuzzy preference
measure for the jth project with respect to the ith strategic
criterion. Note that there are different numbers of sub-criteria
 7598
Appendix A2
The sub-criteria evaluation by the Champion

P1 P2 P3 P4 P5 P6 P7 P8
C1 C11 VL VL VH VH VL L H H
C12 M M H VH VL L H H
C13 VL VL VL VL VL VL VL VL
C14 H L H M VL H L L
C15 VL VL VL VL VL VL VL VL
Mean (1.60,2.00,3.00,4.00) (0.80,1.20,2.20,3.20) (3.40,4.00,5.00,6.00) (3.40,4.00,5.00,6.00) (0.00,0.00,1.00,2.00) (1.40,2.00,3.00,4.00) (2.20,2.80,3.80,4.80) (2.20,2.80,3.80,4.80)
C2 C21 M M M M H M VH VH

T. Yang, C.-H. Hsieh / Expert Systems with Applications 36 (2009) 7594–7603

C22 H H M M M M H H
C23 L VL M M M M H H
Mean (3.00,4.00,5.00,6.00) (2.67,3.33,4.33,5.33) (3.00,4.00,5.00,6.00) (3.00,4.00,5.00,6.00) (3.67,4.67,5.67,6.67) (3.00,4.00,5.00,6.00) (5.67,6.67,7.67,8.67) (5.67,6.67,7.67,8.67)
C3 C31 M M L L VH H VH VH
C32 H M L L H VH H VH
C33 L L VL VL VH H VH VH
Mean (3.00,4.00,5.00,6.00) (2.33,3.33,4.33,5.33) (0.67,1.33,2.33,3.33) (0.67,1.33,2.33,3.33) (6.33,7.33,8.33,9.33) (5.67,6.67,7.67,8.67) (6.33,7.33,8.33,9.33) (7.00,8.00,9.00,10.0)
C4 C41 VH VH H VH VL M H VH
C42 L L M VH L M M M
C43 H M VH VH VL H H VH
Mean (4.33,5.33,6.33,7.33) (3.67,4.67,5.67,6.67) (5.00,6.00,7.00,8.00) (7.00,8.00,9.00,10.0) (0.33,0.67,1.67,2.67) (3.67,4.67,5.67,6.67) (4.33,5.33,6.33,7.33) (5.67,6.67,7.67,8.67)
C5 C51 VL H L VL M M M M
C52 VL VL L VL M H M M
C53 VL H L L M H M M
C54 VL VL VL VL L H M M
C55 VL VL VL VL VH H VH M
Mean (0.00,0.00,1.00,2.00) (2.00,2.40,3.40,4.40) (0.60,1.20,2.20,3.20) (0.20,0.40,1.40,2.40) (3.40,4.40,5.40,6.40) (4.60,5.60,6.60,7.60) (3.80,4.80,5.80,6.80) (3.00,4.00,5.00,6.00)
C6 C61 VL VL VL VL VH VL VH VL
C62 VL VL VL VL H VL VL VL
C63 VL VL VL VL VH VL VH VL
Mean (0.00,0.00,1.00,2.00) (2.33,3.33,4.33,5.33) (0.00,0.00,1.00,2.00) (0.00,0.00,1.00,2.00) (6.33,7.33,8.33,9.33) (0.00,0.00,1.00,2.00) (4.67,5.33,6.33,7.33) (0.00,0.00,1.00,2.00)
C7 C71 VH H L L H H M H
C72 M M L VL VL M M M
C73 M M L VL VL L L L
Mean (5.00,6.00,7.00,8.00) (4.33,5.33,6.33,7.33) (1.00,2.00,3.00,4.00) (0.33,0.67,1.67,2.67) (1.67,2.00,3.00,4.00) (3.67,4.67,5.67,6.67) (3.00,4.00,5.00,6.00) (3.67,4.67,5.67,6.67)
C8 C81 VH M VH VH VH H M VH
C82 VH VH VH VH M M H VH
C83 VH VH H M L M H VH
C84 M M L VL L H M M
C85 M M L VL VH VL M M
C86 H VH M M L H M H
C87 H M H L H L H H
C88 VL VL VL VL VL VL VL VL
Mean (4.63,5.50,6.50,7.50) (4.13,5.00,6.00,7.00) (3.63,4.50,5.50,6.50) (2.63,3.25,4.25,5.25) (3.13,4.00,5.00,6.00) (2.13,2.75,3.75,4.75) (3.38,4.25,5.25,6.25) (4.63,5.50,6.50,7.50)
 T. Yang, C.-H. Hsieh / Expert Systems with Applications 36 (2009) 7594–7603 7599

associated with the different strategic criteria. For each strategic 4.2. The group decision-making for the strategic criteria weights
criterion, X ij is the aggregated mean of its associated sub-criteria.
For instance, a ~ are defined as Eq. (6). Then, their aggregated
~ and b The strategic criteria are determined by a committee formed by
mean is shown in following equation. the top management team. The team has five people who have dif-
  ferent CF to the decision process (as shown in Table 1). There are
meanða ~ ¼ a1 þ b1 ; a2 þ b2 ; a3 þ b3 ; a4 þ b4
~ þ bÞ ð10Þ eight candidate projects, as shown in Table 2.
2 2 2 2
There are four evaluation options at this stage. Let u equals to
By convention, the criteria weights are usually ranging from 0 to 1. 10. Thus, the evaluation can be a precise number ranging from 0
~ ij by Eq. (11).
Thus, X ij is converted to X to 10, a range of numerical values defined by any two integer num-
~ ij ¼ X  =u; bers between 0 and 10, a fuzzy linguistic variable, or a triangular
X ij i ¼ 1; 2; :::; n; j ¼ 1; 2; :::; m ð11Þ
fuzzy number. The fuzzy linguistic variable has triangular mem-
bership functions as defined in Table 3. The triangular membership
3.5. The aggregated project ranking function is as shown in Fig. 5.
For the present study, E1 used crisp value for the weight evalu-
The fuzzy aggregated project weights are the multiplication of ation; E2 used fuzzy linguistic variables; E3 used range values; E4
the criteria weights and the project weightings with respect to used hybrid triangular fuzzy number and crisp value; E5 used tri-
~ j ; j ¼ 1; 2; . . . ; m, be the fuzzy aggre-
each strategic criterion. Let A angular fuzzy number.
gated weight of the jth project. Then, A ~ j can be calculated by Eq. All weight measures are converted into STFNs as defined in Eq.
(12) and the matrix form is shown as Eq. (13). (7). The resulting STFNs are the results of group decision-making.
~j ¼ W
~iX
~ ij ; i ¼ 1; 2; :::; n For each criterion, its associated STFNs from the five experts are
A ð12Þ
then aggregated by Eq. (8). It is then normalized into the range be-
2 3 2 3 tween 0 and 1 by Eq. (9). The resulting criteria weights from the
~1
A ~ 11    X
X ~ n1
6 . 7 6 group decision-making process are summarized, as shown in
6 . 7 ¼ ½W~ 1  W ~ n   6 .. .. .. 77
4 . 5 4 . . . 5 Table 4.
~m
A ~ 1m    X
X ~ nm
2 3 4.3. The sub-criteria evaluation by Champion
W~1X ~ 11 þ W~2 X~ 12 þ    þ W ~nX ~ 1m
6 .. 7
¼64 .
7
5 ð13Þ The project Champion evaluates the project rating using lin-
~ ~ ~ ~ ~ ~ guistic variables. The linguistic variables of the project evaluation
W 1  X n1 þ W 2  X n2 þ    þ W n  X nm
score system are defined as very low (VL), low (L), middle (M), high
Assume that the resulting fuzzy aggregated weight (H) and very high (VH). Their linguistic variables are shown as
A ~ j ; j ¼ 1; 2; :::; m, is transformed into a crisp
~ j ¼ ðaj ; aj ; aj ; aj Þ. Then, A Table 5. The score system is constructed by trapezoidal member-
1 2 3 4
value RðA ~ j Þ by the centroid-index defuzzification method proposed ship function as shown in Fig. 6.
by Yager (1980) as shown in following equation
! Table 1

~jÞ ¼ aj1 þ aj2 þ aj3 þ aj4 The decision-making team and their associated contribution factors
RðA ; j ¼ 1; 2; :::; m ð14Þ
4 Experts Position CF
E1 CEO c1 = 0.30
The higher a defuzzification value is, the higher a project priority
E2 Vice President c2 = 0.25
will be. Then, the project ranking is corresponding to the final E3 Manufacturing Manager (Champion) c3 = 0.15
Six-Sigma project selection priority, and accordingly, it determines E4 Quality Manager c4 = 0.15
the implementation sequence and resource allocation of the overall E5 Master Black Belt c5 = 0.15
improvement activities.

4. Empirical findings for the Six-Sigma project selection

methodology Table 2
The candidate projects

The proposed methodology is applied to a LCD manufacturer in No. Project names

solving the Six-Sigma project selection problem. Details are as P1 Improve 5-wire product yield rate
follows: P2 Increase productivity of 3.5 in. product
P3 Improve customer’s satisfaction for a anonymous customer
P4 Increase 5-wire market share
4.1. The case for the empirical illustrations
P5 Reduce maintain time of knowledge management system
P6 Improve engineer’s problem solving skills
A backlight module is an important component of a Thin-film- P7 Increase the number of product patents
transistor liquid-crystal-displays (TFT-LCD) panel. The applications P8 Reduce the new product release time
of TFT-LCD panel include: personal digital assistants (PDA), cellular
phones, digital cameras, computers, notebook computers, flat pa-
nel TVs, and various computer game units.
The case company is a leading backlight module manufacturer lo- Table 3
cated in Tainan, Taiwan and has more than 3000 employees. The The triangular fuzzy numbers for strategic criteria evaluation
company’s annual sales were more than 930 million US dollars in Linguistic variable Membership function
year 2007. Due to its past rapid growth, there is a compelling need Very low (VL) (0.0, 0.0, 2.5)
for productivity improvement. In addition, it is the goal of the com- Low (L) (0.0, 2.5, 5.0)
pany management to pursue TNQA in the near future. The company Middle (M) (2.5, 5.0, 7.5)
has deployed company-wide Six-Sigma projects since 2004. The High (H) (5.0, 7.5, 10.0)
Very high (VH) (7.5, 10.0, 10.0)
proposed methodology is used for the Six-Sigma project selection.
7600 T. Yang, C.-H. Hsieh / Expert Systems with Applications 36 (2009) 7594–7603

The evaluation is applied to each sub-criterion. The details are Table 5

collected in Appendix. For each strategic criterion, the mean value The trapezoidal fuzzy numbers for sub-criteria evaluation

of its associated sub-criteria fuzzy numbers is solved by Eq. (10). Linguistic variable Membership function
The results are summarized as Table 6 which is then normalized Very low (VL) (0,0,1,2)
into the values between 0 and 1 by Eq. (11). The resulting normal- Low (L) (1,2,3,4)
ized project weights are shown as Table 7. Middle (M) (3,4,5,6)
High (H) (5,6,7,8)
Very high (VH) (7,8,9,10)
4.4. The aggregated project ranking

Given the strategic criteria weights W ~ i ; i = 1, 2, . . . , 8, and the

normalized mean weight X ~ ij , i = 1, 2, . . . , 8, and j = 1, 2, . . . , 8, for pro-
ject Pj with respect to criterion Ci, the aggregated project ranking is The resulting project selection priority ranking according to the pro-
solved by Eq. (12). The resulting aggregated project weights A ~j, posed methodology is: P4 > P3 > P2 > P8 > P7 > P5 > P1 > P6.By the
j = 1, 2, . . . , 8, are shown as follows: adoption of the proposed methodology, the top management can
2 3 2 steer the productivity improvement through the inputs to the stra-
~1 3
A ð0:86; 1:03; 1:54; 2:31Þ tegic criteria decision. The Champion stands at the position to make
6~ 7 6
6 A2 7 6 ð1:19; 1:86; 2:45; 3:39Þ 7
7 the priority decisions for the sub-criteria. The overall selection pro-
6 7 6 7
6A 7 cess is stemmed from TNQA criteria; thus, the implementation of
6 ~3 7 6
6
ð1:22; 1:89; 2:49; 3:46Þ 7
7
6~ 7 6 the selected projects will also promote the management target to
6 A4 7 6 ð1:54; 2:25; 2:86; 3:85Þ 7
7
6 7¼6 ð15Þ pursue TNQA.
6 ~ 7 6 ð0:71; 1:14; 1:68; 2:49Þ 77
6 A5 7 6 7 Several managerial issues are also evident from the Six-Sigma
6 7 6
6 A6 7 6 ð0:22; 0:36; 0:85; 1:55Þ 7
~ 7 implementation process. First, we found that the higher the prior-
6 7 6
6 ~ 7 4 ð0:88; 1:44; 1:99; 2:87Þ 75 ity is, the greater the financial gains. Fig. 7 showed the financial
4 A7 5
~8 ð1:13; 1:72; 2:30; 3:22Þ gains from the implementation of the eight projects. The trend
A
supports the positive argument regarding the relationship between
Finally, the project ranking is solved by Eq. (14) as follows. the project ranking and the financial gains.
The Six-Sigma projects are considered to be successful due to
~ 1 Þ ¼ 1:39
RðA the resulting financial benefits. The project selection process is
~ 2 Þ ¼ 2:22
RðA one of the key success factors since it prioritized the limited re-
~ 3 Þ ¼ 2:27
RðA sources according to both the strategic and the tactical criteria.
~ 4 Þ ¼ 2:63 Second, the TNQA provide a self-assessment form for the appli-
RðA
ð16Þ cant to evaluate its own status with 1000 points as the full grade.
~ 5 Þ ¼ 1:51
RðA The score of 750 is considered to be the minimum requirements for
~ 6 Þ ¼ 0:74
RðA the TNQA award. The case company has improved the self-assess-
~ 7 Þ ¼ 1:79 ment score from 250 to 650 after two years’ efforts. In other words,
RðA
the quality status of the case company has significantly improved
~ 8 Þ ¼ 2:09
RðA through the implementation of the Six-Sigma project.

Fig. 5. The triangular fuzzy membership function for strategic criteria evaluation.

Table 4
The resulting of strategic criteria weight

E1 c1 = 0.30 E2 c2 = 0.25 E3 c3 = 0.15 E4 c4 = 0.15 E5 c5 = 0.15 Normalized weight

Score SFTN Score SFTN Score SFTN Score SFTN Score SFTN
C1 2 (2,2,2,2) Low (0.0,2.5,2.5,5.0) (2,4) (2,2,4,4) 4 (4,4,4,4) (2,4,6) (2,4,4,6) (0.18,0.27,0.30,0.40)
C2 7 (7,7,7,7) Low (0.0,2.5,2.5,5.0) (7,9) (7,7,9,9) (6,7,8) (6,7,7,8) (6,8,10) (6,8,8,10) (0.50,0.60,0.63,0.74)
C3 6 (6,6,6,6) Middle (2.5,5.0,5.0,7.5) (3,5) (3,3,5,5) 6 (6,6,6,6) (4,5,7) (4,5,5,7) (0.44,0.52,0.55,0.64)
C4 8 (8,8,8,8) High (5.0,7.5,7.5,10.0) (4,8) (4,4,8,8) 10 (10,10,10,10) (3,6,8) (3,6,6,8) (0.62,0.73,0.79,0.88)
C5 3 (3,3,3,3) Low (0.0,2.5,2.5,5.0) (2,5) (2,2,5,5) 5 (5,5,5,5) (1,3,5) (1,3,3,5) (0.21,0.30,0.35,0.44)
C6 9 (9,9,9,9) Middle (2.5,5.0,5.0,7.5) (3,7) (3,3,7,7) (6,7,8) (6,7,7,8) (5,7,9) (5,7,7,9) (0.54,0.65,0.71,0.82)
C7 2 (2,2,2,2) Middle (2.5,5.0,5.0,7.5) (0,2) (0,0,2,2) 3 (3,3,3,3) (1,2,3) (1,2,2,3) (0.18,0.26,0.29,0.37)
C8 8 (8,8,8,8) High (5.0,7.5,7.5,10.0) (6,9) (6,6,9,9) 6 (6,6,6,6) (4,7,10) (4,7,7,10) (0.61,0.71,0.76,0.87)
 T. Yang, C.-H. Hsieh / Expert Systems with Applications 36 (2009) 7594–7603 7601

Fig. 6. The trapezoidal fuzzy membership function for sub-criteria evaluation.

Table 6
The project weights evaluated by Champion

P1 P2 P3 P4
C1 (1.60,2.00,3.00,4.00) (0.80,1.20,2.20,3.20) (3.40,4.00,5.00,6.00) (3.40,4.00,5.00,6.00)
C2 (3.00,4.00,5.00,6.00) (2.67,3.33,4.33,5.33) (3.00,4.00,5.00,6.00) (3.00,4.00,5.00,6.00)
C3 (3.00,4.00,5.00,6.00) (2.33,3.33,4.33,5.33) (0.67,1.33,2.33,3.33) (0.67,1.33,2.33,3.33)
C4 (4.33,5.33,6.33,7.33) (3.67,4.67,5.67,6.67) (5.00,6.00,7.00,8.00) (7.00,8.00,9.00,10.0)
C5 (0.00,0.00,1.00,2.00) (2.00,2.40,3.40,4.40) (0.60,1.20,2.20,3.20) (0.20,0.40,1.40,2.40)
C6 (0.00,0.00,1.00,2.00) (2.33,3.33,4.33,5.33) (0.00,0.00,1.00,2.00) (0.00,0.00,1.00,2.00)
C7 (5.00,6.00,7.00,8.00) (4.33,5.33,6.33,7.33) (1.00,2.00,3.00,4.00) (0.33,0.67,1.67,2.67)
C8 (4.63,5.50,6.50,7.50) (4.13,5.00,6.00,7.00) (3.63,4.50,5.50,6.50) (2.63,3.25,4.25,5.25)
P5 P6 P7 P8
C1 (0.00,0.00,1.00,2.00) (1.40,2.00,3.00,4.00) (2.20,2.80,3.80,4.80) (2.20,2.80,3.80,4.80)
C2 (3.67,4.67,5.67,6.67) (3.00,4.00,5.00,6.00) (5.67,6.67,7.67,8.67) (5.67,6.67,7.67,8.67)
C3 (6.33,7.33,8.33,9.33) (5.67,6.67,7.67,8.67) (6.33,7.33,8.33,9.33) (7.00,8.00,9.00,10.0)
C4 (0.33,0.67,1.67,2.67) (3.67,4.67,5.67,6.67) (4.33,5.33,6.33,7.33) (5.67,6.67,7.67,8.67)
C5 (3.40,4.40,5.40,6.40) (4.60,5.60,6.60,7.60) (3.80,4.80,5.80,6.80) (3.00,4.00,5.00,6.00)
C6 (6.33,7.33,8.33,9.33) (0.00,0.00,1.00,2.00) (4.67,5.33,6.33,7.33) (0.00,0.00,1.00,2.00)
C7 (1.67,2.00,3.00,4.00) (3.67,4.67,5.67,6.67) (3.00,4.00,5.00,6.00) (3.67,4.67,5.67,6.67)
C8 (3.13,4.00,5.00,6.00) (2.13,2.75,3.75,4.75) (3.38,4.25,5.25,6.25) (4.63,5.50,6.50,7.50)

Table 7
The normalized project weights evaluated by Champion

P1 P2 P3 P4
C1 (0.16,0.20,0.30,0.40) (0.08,0.12,0.22,0.32) (0.34,0.40,0.50,0.60) (0.34,0.40,0.50,0.60)
C2 (0.30,0.40,0.50,0.60) (0.27,0.33,0.43,0.53) (0.30,0.40,0.50,0.60) (0.30,0.40,0.50,0.60)
C3 (0.30,0.40,0.50,0.60) (0.23,0.33,0.43,0.53) (0.07,0.13,0.23,0.33) (0.07,0.13,0.23,0.33)
C4 (0.43,0.53,0.63,0.73) (0.37,0.47,0.57,0.67) (0.50,0.60,0.70,0.80) (0.70,0.80,0.90,1.00)
C5 (0.00,0.00,0.10,0.20) (0.20,0.24,0.34,0.44) (0.06,0.12,0.22,0.32) (0.02,0.04,0.14,0.24)
C6 (0.00,0.00,0.10,0.20) (0.00,0.00,0.10,0.20) (0.00,0.00,0.10,0.20) (0.00,0.00,0.10,0.20)
C7 (0.50,0.60,0.70,0.80) (0.43,0.53,0.63,0.73) (0.10,0.20,0.30,0.40) (0.03,0.07,0.17,0.27)
C8 (0.46,0.55,0.65,0.75) (0.41,0.50,0.60,0.70) (0.36,0.45,0.55,0.65) (0.26,0.33,0.43,0.53)
P5 P6 P7 P8
C1 (0.00,0.00,0.10,0.20) (0.14,0.20,0.30,0.40) (0.22,0.28,0.38,0.48) (0.22,0.28,0.39,0.48)
C2 (0.37,0.47,0.57,0.67) (0.30,0.40,0.50,0.60) (0.57,0.67,0.77,0.87) (0.57,0.67,0.77,0.87)
C3 (0.63,0.73,0.83,0.93) (0.57,0.67,0.77,0.87) (0.63,0.73,0.83,0.93) (0.70,0.80,0.90,1.00)
C4 (0.03,0.07,0.17,0.27) (0.37,0.47,0.57,0.67) (0.43,0.53,0.63,0.73) (0.57,0.67,0.77,0.87)
C5 (0.34,0.44,0.54,0.64) (0.46,0.56,0.66,0.76) (0.38,0.48,0.58,0.68) (0.30,0.40,0.50,0.60)
C6 (0.63,0.73,0.83,0.93) (0.00,0.00,0.10,0.20) (0.47,0.53,0.63,0.73) (0.00,0.00,0.10,0.20)
C7 (0.17,0.20,0.30,0.40) (0.37,0.47,0.57,0.67) (0.30,0.40,0.50,0.60) (0.37,0.47,0.57,0.67)
C8 (0.31,0.40,0.50,0.60) (0.21,0.28,0.38,0.48) (0.34,0.43,0.53,0.63) (0.46,0.55,0.65,0.75)

Finally, the proposed Six-Sigma project selection decision is a posed methodology. The resulting ranking score then is compared
rolling decision. When there is a new project added to the candi- against the other existing candidate projects to update the ranking
date list, it will be evaluated alone by the Champion using the pro- list.
7602 T. Yang, C.-H. Hsieh / Expert Systems with Applications 36 (2009) 7594–7603
170
162.50
138.44
150
Financial gains
130
(× $1,000)
110
90
70
50
P4 P3 P2
Fig. 7. The financial gains for the projects.
5. Conclusions
The proposed methodology was successfully applied in solving
the project selection problem derived from a backlight module
manufacturer. Under normal resource constraints in most empiri-
cal instances, selection of a project is always an important decision.
There is a need to rank the implementation priority among the
candidate projects. The present study, innovatively adopted na-
tional quality award criteria as the Six-Sigma project selection cri-
teria. In addition, the hierarchical decision process for the strategic
criteria and the tactical sub-criteria by the top management team
and by the Champion, respectively, can integrate both the strategic
and tactical objectives. Furthermore, results indicate that the Del-
phi FMADM is both flexible and robust for the group decision-mak-
ing process.
Empirical examples show promise for the proposed methodol-
ogy. The present study showed that the higher a project’s priority
is, the greater the financial gain from completion of the project.
Accordingly, the proposed methodology can prioritize the financial
gain – which is the key performance indicator for a Six-Sigma pro-
ject. In addition to the tactical benefits of the financial gains, the
selection process will also meet strategic benefits due to its delib-
erate criteria mapping with TNQA. Furthermore, the systematic
evaluation process will influence people with quality management
philosophy and reinforce the commercial objectives of the
company.
Acknowledgment
This work was supported, in part, by the National Science Coun-
cil of Taiwan, Republic of China, under Grant NSC-95-2221-E-006-
349-MY3.
References
Bellman, R. E., & Zadeh, L. A. (1970). Decision-making in a fuzzy environment.
Management Science, 17, B141–B164.
Breygogle, F. W. III, (1999). Implementing Six Sigma smarter solutions using statistic
methods. New York: Wiley.
Brue, G. (2002). Six Sigma for managers. New York: McGraw-Hill.
Chang, P. T., Huang, L. C., & Lin, H. J. (2000). The fuzzy Delphi method via fuzzy
statistics and membership function fitting and an application to the human
resources. Fuzzy Sets and Systems, 112, 511–520.
Chang, I. S., Tasuhiro, Y., & Tozawa, T. (1995). An efficient approach for large scale
project planning based on fuzzy Delphi method. Fuzzy Sets and Systems, 76,
277–288.
Chang, P. C., & Wang, Y. W. (2006). Fuzzy Delphi and back-propagation model for
sales forecasting in PCB industry. Expert Systems with Applications, 30, 715–726.
Chang, S. L., Wang, R. C., & Wang, S. Y. (2006). Applying fuzzy linguistic quantifier to
select supply chain partners at different phases of product life cycle.
International Journal of Production Economics, 100, 348–359.
Chang, C. W., Wu, C. R., & Chen, H. C. (2008). Using expert technology to select
unstable slicing machine to control wafer slicing quality via fuzzy AHP. Expert
Systems with Applications, 34, 2210–2220.
140.63
128.75
118.75
107.81
100.94
65.94
P8 P7 P5 P1 P6
Project
Cheng, C. H., & Lin, Y. (2002). Evaluating the best main battle tank using fuzzy
decision theory with linguistic criteria evaluation. European Journal of
Operational Research, 142, 174–186.
Chou, S. Y., Chang, Y. H., & Shen, C. Y. (2008). A fuzzy simple additive weighting
system under group decision-making for facility location selection with
objective/subjective attributes. European Journal of Operational Research, 189,
132–145.
Coffin, M. A., & Taylor, B. W. III, (1996). Multiple criteria R&D project selection and
scheduling using fuzzy logic. Computers and Operations Research, 23, 207–220.
Das, P. (2005). Reduction in delay in procurement of materials using six sigma
philosophy. Total Quality Management and Business Excellence, 16, 645–656.
Davis, R. A., & Stading, G. L. (2005). Linking firm performance to the Malcolm
Baldrige national quality award implementation effort using multiattribute
utility theory. Managerial Finance, 31, 19–34.
Deming, W. E. (1986). Out of the crisis (2nd ed.). Cambridge, MA: MIT Press.
Eckes, G. (2006). Six Sigma execution. New York: McGraw-Hill.
Edgeman, R. L. (2000). Quoted in ‘‘New voices of quality: 21 for the 21st century”.
In: M. Maguire, (Ed.), Quality Progress (Vol. 33, pp. 31–39).
George, M. L., Rowlands, D., Price, M., & Maxey, J. (2006). The lean Six Sigma pocket
tool book. New York: McGraw-Hill.
Greco, S., Matarazzo, B., & Slowinski, R. (2002). Rough set methodology for sorting
problems in presence of multiple attributes and criteria. European Journal of
Operational Research, 138, 247–259.
Hutchins, D. (2000). The power of Six Sigma in practice. Measuring Business
Excellence, 4, 26–33.
Jang, J.-S. R., Sun, C.-T., & Mizutani, E. (1997). Neuro-fuzzy and soft computing. Upper
Saddle River, NJ: Prentice Hall.
Kalamdani, R., & Khalaf, F. (2006). Application of design for six sigma to
manufacturing process design at Ford PTO. International Journal of Product
Development, 3, 369–387.
Kaufmann, A., & Gupta, M. M. (1985). Introduction to fuzzy arithmetic: Theory and
applications. New York: Van Nostrand Reinhold.
Klefsjö, B., Wiklund, H., & Edgeman, R. L. (2001). Six sigma seen as a methodology
for total quality management. Measuring Business Excellence, 5, 31–35.
Liu, Y. C., & Chen, C. H. (2007). A new approach for application of rock mass
classification on rock slope stability assessment. Engineering Geology, 89,
129–143.
Mark, D. G. (2001). Six Sigma program success factors. Six Sigma Forum Magazine, 1,
36–45.
_ & Odabasi, A. Y. (2005). A new fuzzy multiple attribute group decision
Ölcer, A. I.,
making methodology and its application to propulsion/manoeuvring system
selection problem. European Journal of Operation Research, 166, 93–114.
Pande, P. S., Robert, P. N., & Roland, R. C. (2002). The Six Sigma way team fieldbook.
New York: McGraw-Hill.
Petroni, A., & Rizzi, A. (2002). Fuzzy logic based methodology to rank shop floor
dispatching rules. International Journal Production Economics, 76, 99–108.
Przekop, P. (2006). Six Sigma for business excellence. New York: McGraw-Hill.
Ritter, D. (1993). A tool for improvement using the Baldrige criteria. National
Productivity Review, 12, 167–182.
Seetharaman, A., Sreenivasan, S., & Boon, L. P. (2006). Critical success factors of total
quality management. Quality and Quantity, 40, 675–695.
Snee, R. D., & Rodebaugh, W. F. Jr., (2002). The project selection process. Quality
Progress, 78–80.
Su, C. T., Chiang, T. L., & Chiao, K. (2005). Optimizing the IC delamination quality via
six-sigma approach. IEEE Transactions on Electronics Packaging Manufacturing,
28, 241–248.
Wang, Y. J. (2008). Applying FMCDM to evaluate financial performance of domestic
airlines in Taiwan. Expert Systems with Applications, 34, 1837–1845.
Wang, J., & Lin, Y. (2003). A fuzzy multicriteria group decision making approach to
select configuration items for software development. Fuzzy Sets and Systems,
134, 343–363.
Xu, Z. S., & Chen, J. (2007). An interactive method for fuzzy multiple attribute group
decision making. Information Sciences, 177, 248–263.
Yager, R. R. (1980). On a general class of fuzzy connectives. Fuzzy Sets and Systems, 4,
235–242.
 T. Yang, C.-H. Hsieh / Expert Systems with Applications 36 (2009) 7594–7603
Yang, T., Chen, M.-C., & Hung, C.-C. (2007). Multiple attribute decision-making
methods for the dynamic operator allocation problem. Mathematics and
Computers in Simulation, 73, 285–299.
Yang, T., & Chou, P. (2005). Solving a multiresponse simulation-optimization
problem with discrete variables using a multiple-attribute decision-making
method. Mathematics and Computers in Simulation, 68, 9–21.
Yang, T., & Hung, C.-C. (2007). Multiple-attribute decision making methods for plant
layout design problem. Robotics and Computer-Integrated Manufacturing, 23,
126–137.
7603
Yeh, D. Y., Cheng, C. H., & Chi, M. L. (2007). A modified two-tuple FLC model for
evaluating the performance of SCM: By the Six Sigma DMAIC process. Applied
Soft Computing, 7, 1027–1034.
Zadeh, L. A. (1965). Fuzzy sets. Information and Control, 8, 338–353.
Zeng, J., An, M., & Smith, N. J. (2007). Application of fuzzy based decision making
methodology to construction project risk assessment. International Journal of
Project Management, 25, 589–600.