See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/245529111

Six Sigma based multicriteria approach to improve decision settings

Article  in  International Journal of Quality Engineering and Technology · January 2009

DOI: 10.1504/IJQET.2009.030503

CITATIONS READS

9 67

6 authors, including:

Lotfi Azzabi Donia Ayadi

University of Angers National center os sports medicine and sciences , Tunis
27 PUBLICATIONS   26 CITATIONS    7 PUBLICATIONS   22 CITATIONS   

SEE PROFILE SEE PROFILE

Abdessamad Kobi Christian Robledo

University of Angers University of Angers
128 PUBLICATIONS   1,248 CITATIONS    32 PUBLICATIONS   196 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Correction of capability indices taking into account measurement errors View project

Supply chain resilience View project

All content following this page was uploaded by Lotfi Azzabi on 30 June 2020.

The user has requested enhancement of the downloaded file.

Int. J. Quality Engineering and Technology, Vol. 1, No. 1, 2009 99

Six Sigma based multicriteria approach to improve

decision settings

Lotfi Azzabi*, Dorra Ayadi and

Younes Boujelbenne
Faculté des Sciences Économiques et Gestions de Sfax,
Université de Sfax – Tunisia,
Route de l’Aérodrome Km 4,
BP 1088, Sfax 3018, Tunisia
E-mail: lotfi_azzabi@yahoo.fr
E-mail: dorra_ayadi@yahoo.fr
E-mail: younes-boujelbene@fsegs.rnu.t
*Corresponding author

Abdessamad Kobi and Christian Robledo

LASQUO/ISTIA,
62, Avenue Notre Dame du Lac,
49000 Angers, France
E-mail: kobi@istia.univ-angers.fr
E-mail: Christian.Robledo@univ-angers.fr

Habib Chabchoub
Institut Supérieur de Gestion Industrielle de Sfax,
Route Mharza Km 1.5 BP 954, 3018 Sfax, Tunisie
E-mail: habib.chabchoub@fsegs.rnu.tn

Abstract: The present competitive market is focusing industrial efforts on

producing high-quality products with the lowest possible cost. The total
performance of the process and the quality of its production depend on the one
hand, of the characteristics of the intermediate products, and on the other hand,
of the operation parameters of the manufacturing. To help accomplish this
objective, various quality improvement philosophies have been put forward in
recent years and of these, Six Sigma has emerged as perhaps the most viable
and efficient technique for process quality improvement. The objective of this
paper is to propose a method that puts in obviousness the enforcement
performances improvement Six Sigma to assure high-level quality products and
to make firm a level of improvement of the long-term performance. The
application of the Six Sigma methods enforced with multicriteria approach to
permit classification the betters’ choices of a Tunisian industry.
Keywords: Six Sigma; quality engineering; methods ELECTRE; multicriteria
approach; methods PROMETHEE; control chart.
Reference to this paper should be made as follows: Azzabi, L., Ayadi, D.,
Boujelbenne, Y., Kobi, A., Robledo, C. and, Chabchoub, H. (2009) ‘Six Sigma
based multicriteria approach to improve decision settings’, Int. J. Quality
Engineering and Technology, Vol. 1, No. 1, pp.99–123.

Copyright © 2009 Inderscience Enterprises Ltd.

100 L. Azzabi et al.

Biographical notes: Lotfi Azzabi has a Doctoral with the University of

Sfax-Tunisia and the University of Angers, France. Her global research include
Six Sigma, statistic control; operational research especially multicriteria
approach, performance product management. She obtained her Master’s in
Quantitative Methods and Product Management from Faculty Economic
Science and Management Sfax-Tunisia. She has published in journal Reliability
Engineering and System Safety.

Dorra Ayadi has a Doctorate with the University of Sfax-Tunisia and the
University of Angers, France. Her global research include quality, reliability
and dependability; operational research especially multicriteria approach, risk
analysis. She obtained her Master’s in Quantitative Methods and Product
Management from Faculty Economic Science and Management Sfax-Tunisia.
She has published in journal Reliability Engineering and System Safety.

Abdessamad Kobi is a Professor from the University Angers (ISTIA), his axles
of searches are the Federal Bureau of Investigation of systems working
Bayésiennes methods of trial reduction, accelerated trials, laws of mechanisms
acceleration of damaging. The valuation, the optimisation and the statistical
mastery process (statistic cases control and optimisation) multivariés and
non-Gaussians, and analyses him of the gross performance of the contract
(quality and organisations, study of their performances). His majority publisher
in International Journal of Advanced Manufacturing and Technologies, Quality
Engineering, International Journal of Materials & Product Technology,
Journal Européen des Systèmes Automatisés and Journal of Coatings
Technology.

Younes Boujelbenne is a Conference Master and Director of the ISAS

Sfax-Tunisia, his axles of searches are the valuation of the performance, the
order to pay of the extraction, performance management, analysis of given, the
management of logistics and operations, the management of the quality, the
mathematical programming and statistics. His has published in La Revue du
Financier, Revue Gestion Management, Revue de la Littérature.

Christian Robledo is Conference Master and Director to the ISTIA, University

Angers, his axles of searches are the Federal Bureau of Investigation of systems
working (working stock and software) (dependability of products and software)
Bayésiennes methods of trial reduction, accelerated trials, laws of mechanisms
acceleration of damaging. The valuation, the optimisation and the statistical
mastery processes (statistic cases control and optimisation) multivariés and
non-Gaussiens, and analyses him of the gross performance of the contract
(quality and organisations, study of their performances). He has published in
Quality Engineering.

Habib Chabchoub is a Professor from the University of Sfax-Tunisia, his axles

of searches are: the valuation of the performance, the order to pay of the
extraction, multicriteria decision, the management of logistics and operations,
the management of the quality, the mathematical programming and statistics.
His majority publisher in: European Journal of Operational Research, Journal
of Multicriteria Decision Aid.
 Six Sigma based multicriteria approach to improve decision settings 101

1 Introduction

The manufacturing industry is characterised by many restrictions and complex operations

required in depositing metallic layers to nanometre level thickness (Meziane et al., 2000).
Read/write head producers must manufacture two separate devices incorporated in a
single product. Therefore, process engineers must not only control processing conditions
of the head build, but they must also control any possible interactions between the read
and write aspects of the head, for example, the shielding of stray magnetic flux that may
affect the reading procedure from the media (Proudlove et al., 1998). Several business
improvement techniques have been developed over the past few decades that enable
producers to identify and eradicate any discrepancies within manufacturing
environments. One such technique is the Six Sigma breakthrough manufacturing
methodology; as a methodology, it uses existing problem solving tools to eradicate
system defects (Breyfogle, 1999). The discipline of Six Sigma views every business
activity as a process, that once optimised and controlled, reduces cost. Six Sigma uses a
group of improvement specialists, typically referred to as champions, master black belts,
black belts, and green belts (Linderman et al., 2003). Those specialists receive intensive
differentiated training that is tailored for their ranks and is designed to improve their
knowledge and skills in statistical methods, project management, process design,
problem-solving techniques, leadership skill, and other managerial skills (Linderman et
al., 2003; Barney, 2002; Gowen and Tallon, 2005, Snee and Hoerl, 2003) with assigning
the improvement specialists to take different levels of roles and responsibilities in leading
the continuous improvement efforts. Then, the proposed paper permits to contribute of
the high level of quality produced by Six Sigma as method of analysis.

2 Six Sigma methods

Six Sigma has been characterised as the latest management fad to repackage old quality
management principles, practices, and tools/techniques (Clifford, 2001). At first glance,
Six Sigma looks strikingly similar to prior quality management approaches. However,
leading organisations with a track record in quality have adopted Six Sigma and claimed
that it has transformed their organisation. For example, 3M’s Dental Division won the
Baldrige Award (Adler et al., 1999) and then later adopted Six Sigma to improve
performance even further (McClenahen, 2004). The financial performance of 3M since
Six Sigma adoption has been very impressive (Fiedler, 2004). Other organisations with a
quality track record, such as Ford, Honeywell, and American Express, have adopted Six
Sigma as a way to further enhance business performance (Hahn et al., 2000). This creates
a dilemma: on the one hand, sceptics argue that Six Sigma lacks discriminate validity
over prior approaches to quality management; on the other hand, quality-mature
organisations adopt Six Sigma to enhance performance.
102 L. Azzabi et al.

2.1 Definition of Six sigma

Wacker (2004) developed rules for constructing conceptual definitions. With those rules
in mind, we propose the following rigorous base definition that captures the theoretical
aspects of Six Sigma from the case study data and literature: Six Sigma is an organised,
parallel-meso structure to reduce variation in organisational processes by using
improvement specialists, a structured method, and performance metrics with the aim of
achieving strategic objectives. We do not suggest one definition for everyone. Companies
may choose variations of this base definition when implementing Six Sigma in order to
customise it to their situation. Later, we suggest some possible variations. Contingency
theory implies that the base definition will not fit every company, but nonetheless, it is a
starting point for research and implementation.
Six Sigma, which is ‘an organised and systematic method for strategic process
improvement and new product and service development that relies on statistical methods
and the scientific method to make dramatic reductions in customer defined defect rates’
(Sahoo et al., 2008).
The simplest definition for Six Sigma is to eliminate waste and to mistake proof the
processes that create value for customer. The elimination of waste led to yield
improvement and production quality; higher yield increased customer satisfaction
(Naumann and Hoisington, 2001).
Then, the Six Sigma of today speaks the language of management: bottom-line
results. It institutionalises a rigorous, disciplined, fact-based way to deliver more money
to the bottom line through process improvement and process design projects – selected by
the top leadership and led by high potentials trained as black belts or master black belts in
Six Sigma – that aim to create near-perfect processes, products, and services all aligned
to delivering what the customer wants. In successful implementations, the majority of Six
Sigma projects are selected for measurable bottom line or customer impact that is
completed within two to six months. The projects deliver through the application of a
well-defined set of statistical tools and process improvement techniques by well-trained
people in an organisation that has made it clear that Six Sigma is a career accelerator.
In our practice, we see companies viewing Six Sigma in two ways: as a set of
powerful tools for improving processes and products and as an approach for improving
both the process- and people-related aspects of business performance. Six Sigma is used
as a hands-on approach to developing leadership and change management skills. The
companies that achieve the greatest benefits from Six Sigma leverage the linkages
between people, processes, customer, and culture. In its 2000 annual report, GE describes
the changes brought by Six Sigma this way: “Six Sigma has turned the Company’s focus
from inside to outside, changed the way we think and train our future leaders and moved
us toward becoming a truly customer-focused organization.”

2.2 Six Sigma practices

Based on a review of both research studies and practitioner literature on Six Sigma, we
identify three practices that are critically associated with Six Sigma implementation.
These practices are Six Sigma role structure, Six Sigma structured improvement
procedure, and Six Sigma focus on metrics.
 Six Sigma based multicriteria approach to improve decision settings 103

2.2.1 Six Sigma role structure

Six Sigma uses a group of improvement specialists, typically referred to as champions,
master black belts, black belts, and green belts (Henderson and Evans, 2000; Linderman
et al., 2003). Those specialists receive intensive differentiated training that is tailored for
their ranks and is designed to improve their knowledge and skills in statistical methods,
project management, process design, problem-solving techniques, leadership skill, and
other managerial skills (Barney, 2002; Gowen and Tallon, 2005; Linderman et al., 2003;
Snee and Hoerl, 2003). With assigning the improvement specialists to take different
levels of roles and responsibilities in leading the continuous improvement efforts, the
organisation builds a Six Sigma role structure for quality improvement. In the Six Sigma
role structure, there is a hierarchical coordination mechanism of work for quality
improvement across multiple organisational levels (Sinha and Van de Ven, 2005). This
mechanism helps to coordinate and control work across organisational levels to ensure
that the tactical tasks match with the overall business strategy (Sinha and Van de Ven,
2005).

2.2.2 Six Sigma structured improvement procedure

Six Sigma applies a structured approach to managing improvement activities,
which is represented by define-measure-analyse-improve-control (DMAIC) used in
process improvement or define-measure-analyse-design-verify (DMADV) used in
product/service design improvement (Linderman et al., 2003). Both of these procedures
are grounded in the classic plan-do-check-act (PDCA) cycle, but Six Sigma specifies the
QM tools and techniques to use within each step, which is unique to Six Sigma
(Linderman et al., 2003). The Six Sigma structured improvement procedures provide
teams a methodological framework to guide them in the conduct of improvement projects
[Pande et al. (2002) or Pyzdek (2003) for detailed explanations of the steps in these
procedures]. Extensive use of the Six Sigma structured procedures and the associated
tools and techniques in quality improvement projects is shown to facilitate the teams in
learning and knowledge acquisition (Choo et al., 2007a, 2007b).

2.2.3 Six Sigma focus on metrics

Six Sigma emphasises using a variety of quantitative metrics in continuous improvement,
such as process Sigma measurements, critical-to-quality metrics, defect measures, and
improvement measures as well as traditional quality measures like process capability
(Breyfogle et al., 2001; Dasgupta, 2003; Linderman et al., 2003; Pyzdek, 2003). Six
Sigma metrics are used to set improvement goals (Linderman et al., 2003; Pande et al.,
2002). Using objective data should reduce corporate use of political agendas to drive
solutions (Brewer, 2004). As suggested by Linderman et al. (2003), using explicit,
challenging goals in Six Sigma projects can increase the magnitude of improvements;
reduce performance variability of the projects, and increase employees’ improvement
efforts and commitment to quality. Moreover, Six Sigma integrates business-level
performance, process measures, and project metrics into a systematic review process so
that managers can manage the organisation quantitatively and translate the business
strategy into tactical tasks (Barney, 2002).
104 L. Azzabi et al.

3 Six Sigma process: DMAIC

Six Sigma is a process that is often briefly described by the acronym DMAIC. First, the
production testing process, or process issue, needs to be defined. Second, since stability
testing itself is a measuring process, its capability needs to be measured. Third, the
capability of the process needs to be analysed in order to determine if it is delivering what
is required (accurate stability predictions or estimates), and if not, improve. Finally,
control the stability testing process by insuring that the improvements that have been
implemented are maintained through time. Both of these procedures are grounded in the
classic PDCA cycle, but Six Sigma specifies the quality management tools and
techniques to use within each step, which is unique to Six Sigma (Linderman et al.,
2006). The Six Sigma gait methodological as follows:

Figure 1 Six Sigma process (see online version for colours)

3.1 Define
To collect indispensable dates, to serve as low working for all the project, for our gait
method collected data concern the product system; after that to decompose working axes
(Dave 2002).
The steps define is realised by the multicriteria approach: ELECTRE.
ELECTRE method provides a different approach. This method concentrates the
analysis on the dominance relations among the alternatives. That is, this method is based
on the study of outranking relations, exploiting notions of concordance (EAA, 2003;
Rogers and Bruen, 1996, 1998]. These outranking relations are built in such a way that it
 Six Sigma based multicriteria approach to improve decision settings 105

is possible to compare alternatives. The information required by ELECTRE consists of

information among the criteria and information within each criterion (EAA, 2003).
Then, ELECTRE IV consists of classification procedures, which result in a ranking of
all possible candidate materials in relation to each other. The method is designed to rank
alternatives without using the relative criteria importance coefficients and it is equipped
with embedded outranking relations framework (Shanian and Savadogo, 2006). This
procedure offers the most satisfactory overall resolution to conflicts between the possible
candidate materials, which exist at the level of individual selection criteria. The method
presents an incomparability relation, which is useful when the material designer is not
able to compare two candidate materials. On contrary, the method makes the procedures
sensitive when a set of candidate materials very close to each other perform in an almost
identical manner with all others (Rogers et al., 2000). To make the pairwise comparison
of candidate materials, the following notations are defined:
mp(Mk, Mi) is the number of performance indices for which candidate
material Mi is strictly preferred to candidate material Mk
mq(Mk, Mi) is the number of performance indices for which candidate
material Mi is preferred to candidate material Mk
min(Mk, Mi) is the number of performance indices for which candidate
materials Mi and Mk are considered indifferent, although
candidate material Mi has a better performance than
candidate material Mk
mo(Mi, Mk) = mo(Mk, Mi) is the number of performance indices for which candidate
material Mi and candidate material Mk perform identically.
If m is the total number of criteria, it is clear that:

m = mp ( Mk , Mi ) + mq ( Mk , Mi ) + min ( Mk , Mi ) + mo ( Mk , Mi )
+ min ( Mk , Mi ) + mp ( Mk , Mi ) + mq ( Mk , Mi )

The four levels of domination are represented in the outranking relation as follows:

• Quasi-domination Sq: We are talking about quasi-domination when there is a beach

indifference [Mi, Mk]. There is indifference between actions a and b when away
from their assessment falls within the range [Mi, Mk].
Mi outranks Mk with quasi-domination if

⎧ up ( Mk , Mi ) + uq ( Mk , Mi ) = 0
⎪
MiSqMk ⎨uin ( Mk , Mi ) ≤ 1 + uin ( Mk , Mi )
⎪ +up ( Mk , Mi ) + uq ( Mk , Mi )
⎩

• Canonical domination Sc: We are talking about canonical-domination is a real test by

Mi and Mk that has introduced a range preferably low.
Mi outranks Mk with canonical-domination if
106 L. Azzabi et al.

⎧ up ( Mk , Mi ) = 0
⎪
⎪⎪ up ( Mk , Mi ) ≤ uq ( Mk , Mi )
MiScMk ⎨ uq ( Mk , Mi ) + uin ( Mk , Mi )
⎪ ≤ 1 + uin ( Mk , Mi )
⎪
⎪⎩+up ( Mk , Mi ) + uq ( Mk , Mi )

• Pseudo-domination Sp: In pseudo-domination, is introduced in addition to the beach

indifference of the quasi-domination, a preference for the beach weak.
Mi outranks Mk with pseudo-domination if

⎧ up ( Mk , Mi ) = 0
MiSpMk ⎨
⎩uq ( Mk , Mi ) ≤ up ( Mk , Mi ) + uq ( Mk , Mi )
• Veto-domination Sv: We are talking about true test when the action that receives the
highest score will be preferred over another. There is indifference if both scores are
equal.
Mi outranks Mk with veto-domination if
⎧ up ( Mk , Mi ) = 0 or
⎪
⎪⎪ uq ( Mk , Mi ) = 1
MiSvMk ⎨not MkPvjMi, ∀j and
⎪
⎪ uq ( Mk , Mi ) ≥ m
⎪⎩ 2
By representing the constant discrimination threshold s(λ), the material designer is able
to distinguish if one outranking relation is more credible than another, such that
• within the first step of classification, the strongest domination relations between
those established are taken into consideration;
• within the second step of classification procedure, it is the two strongest domination
relations that intercede in the procedure of ranking the remaining candidate
materials, etc.
The final ranks of possible candidate materials are thus derived from an exploiting
procedure. This procedure contains ascending and descending distillations and from these
come either partial or complete final pre-orders (Rogers et al., 2000). Whether the final
result is a partial pre-order (not containing a relative ranking of all of the possible
candidate materials), rather than a complete pre-order, depends on the level of
consistency between the rankings from the two orders. Briefly, the exploitation procedure
in ELECTRE IV starts by deriving from the fuzzy relation two complete pre-orders
(Figueira et al., 2005). A final partial pre-order Z is then built by the intersection of the
two complete pre-orders, Z1 and Z2. These pre-orders are obtained according to two
variants of the same principle, both acting in an antagonistic way on the floating actions.
The partial pre-order Z1 is defined as a partition on the set into q ordered classes,
B1 , K , Bh , K , Bq , where B1 is the head-class in Z1. Each class Bh is composed of ex-
aequo elements according to Z1. The complete pre-order Z2 is determined in a similar
 Six Sigma based multicriteria approach to improve decision settings 107

manner, using ordered classes, B1′, K , Bh′ , K , Bu′ , with Bu′ being the head-class. Each one
of these classes is obtained as a final distilled of a distillation procedure. The detailed
designed to compute Z1 starts (first distillation) by defining an initial set D0 = A (first
distillation). This leads to the first final distilled B1. After finding Bh , in the distillation
h + 1, the procedure sets D0 = A \ ( B1 ∪K Bh ) . According to Z1, the actions in class Bh
are, preferred to those of class Bh + 1, descending (top-down). The procedure leading to
Z2 is quite identical, except that now the actions in Bh + 1 are preferred to those in class
Bh ; ascending (bottom-up) distillation. A complete pre-order is finally suggested taking
into account the partial pre-orders (i.e., by intersecting Z1 and Z2) and some additional
consideration (Figueira et al., 2005). The detailed operations of the ELECTRE IV
methods and its exploitation procedure can be found in Collette and Siarry (2003),
Figueira et al. (2005) and Shanian and Savadogo (2006).

3.2 Measure
Next, the focus is on measuring the process. Key characteristics are categorised,
measurement systems are verified and data are collected. In this phase, the measure
contains the identification of appearance problems up to each some performance
situation.
This step is realised by the exponentially weighted moving average (EWMA) control
chart. Process control charts are chronological graphs of process data that are used to help
understand, control, and improve processes – such as infection control or adverse event
processes – and that, although based in statistical theory, are easy for practitioners to use
and interpret.
The EWMA is a statistic for monitoring the process that averages the data in a way
that gives less and less weight to data as they are further removed in time.
The EWMA is used extensively in time series modelling and in forecasting
(Murdoch, 1979; Braverman, 1981; Montgomery, 2001). Since the EWMA can be
viewed as a weighted average of all past and current observations, it is very insensitive to
the normality assumption.
In the EWMA control chart, the mean of the Xt = wxt + (1 − w ) Z t −1 with 1 ≤ w ≥ 1
Therefore, the EWMA control chart can be constructed by plotting the Zt versus the
sample number t (or time).
Upper control limit (UCL), control limit (CL) and lower control limit (LCL) for the
EWMA control chart are given below in the equation form respectively.

= w ⎡⎣1 − (1 − w )2t ⎤⎦
LC zt = x ± 3σ
n ( 2 − w)

= w
LCzt = x ± 3σ
n ( 2 − w)

If the sample means fall within the CLs, then we can say that the process is in control
with respect to the mean process and the process will be out of control if the sample
means exceeds the CLs (Singh and Singh, 2006).
108 L. Azzabi et al.

3.3 Analyse
Once data concern performance situation are collected, it is analysed. The intent is to
convert the raw data into information that provides insights into the process. These
insights include identifying the fundamental and most important criteria to helps industry
taken the best choice of the performance situation.
The steps analyse is realised by the multicriteria approach the preference ranking
organisation method for enrichment evaluation (PROMETHEE) method is a multicriteria
decision-making method developed by Brans et al. (1986). It is a quite simple ranking
method in conception and application compared with other methods for multicriteria
analysis. It is well adapted to problems where a finite number of alternatives are to be
ranked considering several, sometimes conflicting criteria. The following steps are
required for the implementation of the method:
1 Alternatives are compared in pairs for each criterion. The preference is expressed by
a number in the interval [0, 1] (0 for no preference or indifference to, 1 for strict
preference). The function relating the difference in performance to preference is
called the generalised criterion and it is determined by the decision maker.
2 A multicriteria preference index is formed for each pair of alternatives as a weighted
average of the corresponding preferences computed in Step 1/for each criterion. The
index P(a, b) (in the interval [0, 1]) expresses the preference of alternative a over b
considering all criteria. The weighting factors express the relative importance of each
criterion and are chosen by the decision maker.
3 Alternatives can be ranked according to:
• The sum of indices π(a i) indicating the preference of alternative an over all the
others. It is termed ‘leaving flow’ U+ (a) and shows how ‘good’ is alternative a.
• The sum of indices P(i, a) indicating the preference of all other alternatives
compared to a. It is termed ‘entering flow’ U– (a) and shows how ‘inferior’ is
alternative a (Albadvi et al., 2007).
The decision multicriteria problem resulting can be expressed in the form of a decision
matrix (m × n) whose elements indicate the evaluation or the value of the alternative ai
according to the criterion fj. Thus, the PROMETHEE algorithm can be summarised as
follows (Brans et al., 1984):
1 To indicate for each criterion fj, a generalised preference function Pj(d) with

d = fj ( ai ) − fj ( ak ) .

2 To define a weighting vector, which is a measurements of the relative importance of

each criterion, W = {w1, ..., wj}.

3 To define for all the alternatives ai, ak ∈ A, the preference relation π:

⎧ A × A → [ 0 1]
⎪
π =⎨
∑ WjPj ( fj ( ai ) − f j ( ak ) )
n
⎪⎩ j =1
 Six Sigma based multicriteria approach to improve decision settings 109

The preference index π(ai, ak) is an intensity measurement of the total preference of
the decision maker for an alternative ai compared to an alternative ak and that by
taking into account all the criteria simultaneously. It is, then, a weighted average of
the preference functions Pj(d).

4 To calculate outgoing flow which is a measure of alternative force ai ∈ A like

1
∑
n
∅ + ( ai ) = π ( ai , ak )
n i =1/ i ≠ k

5 To calculate entering flow which is a measure of the outclassed character of an

alternative ai ∈ A, as

1
∑
n
( ai ) = i =1/ i ≠ k
π ( ak , ai )
n

6 Preference relation evaluation.

3.4 Improvement
The fourth step is to improve the process. Solutions to the problem are developed; and
changes are made to the process. Results of process changes are seen in the
measurements. In this step, the company can judge whether the changes are beneficial, or
if another set of changes is necessary (Luan, 2001).Then, the generation possible
solutions for detected data and this on the tool brainstorming.

3.5 Control
The objective of this step is the control and the supervision new improved process in
order to make firm that high quality levels are maintained on the long time (Azzabi et al.,
2007). This step is realised by the EWMA control chart.

4 Application of DMAIC process

The application of the Six Sigma method in a manufactured industrial Tunisian, of which
the objective is to help him to improve the product sheet metal process with the output
the sheet metal of steel, copper. The problem in sheet metal product process is the
variation of the outgoing datum, so that the tolerance of product sheet metal so be it
maximum (Figure 2) or minimum to the variation (Figure 3).
The sheet metal has produced by SSF (Figure 4) machine. It is responsible to product
the components thimbles. Then, the process of product is generally preferred for the
manufacture of high-value added products. This process facilitates the manufacture of
hollow products from sheet metal of steel, copper (Figure 5).
110 L. Azzabi et al.

Figure 2 Maximum tolerance variations

Figure 3 Minimum tolerance variations

Figure 4 SSF machine to product

 Six Sigma based multicriteria approach to improve decision settings 111

Figure 5 Output of sheet metal

The enforcement of the method Six Sigma is as follows:

4.1 Define
This step permits to identify six problems proceeding to the machine product SSF, these
problems has identified by the figure as following (Figure 6):

Figure 6 Identification problem products

112 L. Azzabi et al.

After that, to identify the problems, we classified this problem by ELECTRE IV, in order
to d, it is important to identify the list of criteria as shown in Figure 7:

Figure 7 Identification of criteria

The credibility matrix of ELECTRE IV is:

Figure 8 Credibility matrix

Then, the final classification of problems is in graphically following Figure 9:

 Six Sigma based multicriteria approach to improve decision settings 113

Figure 9 Final graph of classification problems (see online version for colours)

The important problem classified of prior is P4: height out of tolerance.

4.2 Measure
The measure step consists to identify with the graphics the problems dysfunctions and
non-conformities proceeding from the sheet metal. The control of the sheet metal as their
weight and thickness is determined through of a standard dimension of 18 parts, on the
reception of matters, the result of the reception control are as follows (Table 1):
Table 1 Measure of the thickness of sheet metal

W E Min Max
19.32 2.58 2.36 2.44
38.64 2.54 2.36 2.44
57.96 2.5 2.36 2.44
77.28 2.5 2.36 2.44
96.6 2.5 2.36 2.44
115.92 2.5 2.36 2.44
135.24 2.51 2.36 2.44
154.56 2.27 2.36 2.44
173.88 2.26 2.36 2.44
193.2 2.3 2.36 2.44
212.52 2.31 2.36 2.44
231.84 2.29 2.36 2.44
251.16 2.3 2.36 2.44
270.48 2.3 2.36 2.44
289.8 2.27 2.36 2.44
309.12 2.31 2.36 2.44
328.44 2.54 2.36 2.44
347.76 2.55 2.36 2.44
114 L. Azzabi et al.

The schematisation of the EWMA control chart indispensable for the values held in the
sample are in Figure 10:

Figure 10 EWMA control charts of the thickness of sheet metal

EWMA:Control charts of thikness of sheet metal

2,50

2,45 UCL=2,4453

_
_
EWMA

X=2,4072
2,40

LCL=2,3692

2,35

2,30
1 3 5 7 9 11 13 15 17
Sample

Figure 11 Variation of the thickness of sheet metal

 Six Sigma based multicriteria approach to improve decision settings 115

From the graphics of control, the metal sheet thickness to the reception, one establishes
that variations of the thickness for encoded arrivals of 135 as far as 320 yields to levels of
minimum demanded tolerance by norms of the contract. Then, the receiving metal
sheet process is no conformable, and it exist dysfunctions to levels of some parts of
sheet metal. The process is outside the control, it exist point presentation of the above
and below the CLs. This is well identified by the bell of Six Sigma following in
Figure 11.

4.3 Analyse
This step is identified by PROMETHEE methods to classify the six level of variation
tolerance of the thickness of sheet metal. Choice generalised criteria: generalised criteria
and following setter have been chosen. In Table 2, the different criteria setting by types
and parameters are shown:

Table 2 Criteria setter

Criterion Types Parameters

C1 (E) II q = 10
C2 (E) III p = 30
C3 (E) V q = 50
p = 500
C4 (E) IV q = 10
p = 60
C5 (E) I
C6 (E) VI σ=5

After identified the six level of types criterion in the Table 2, the following step, is
valuate variation tolerance of problem P4 in Table 3:

Table 3 Value thickness of sheet metal

Criteria Min/max E1 E2 E3 E4 E5 E6
C1 (E) Min 80 65 83 40 52 94
C2 (E) Max 90 58 60 80 72 96
C3 (E) Min 600 200 400 1000 600 700
C4 (E) Min 54 97 72 75 20 36
C5 (E) Min 8 1 4 7 3 5
C6 (E) Max 5 1 7 10 8 6

On foot of both preceding boards, we come to build the flow board of the following
manner. Let us consider the situations pair (E1, E2) for the level of thickness of sheet
metal to be evaluated in Table 4:
116 L. Azzabi et al.

Table 4 Construction of the flow board

 Six Sigma based multicriteria approach to improve decision settings 117

with:
q indifference threshold, to present is the largest value of ‘d’ below which the
adjudicator considers that there is indifference
p threshold of strict preference is the smallest value of ‘d’ above, which the adjudicator
considers that there is strict preference.
Synthesis:
Mean of Ci (E1, E2): π (E1, E2) = (0 + 1 + 0 + 0,5 + 0 + 0,274) / 6 = 0.296

Mean of Ci (E2, E1): π (E2, E1) = (1 + 0 + 0,778 + 0 + 1 + 0) / 6 = 0.462

The board summarises the reasoning above for all pairs (Ei, Ej), and calculating the
entering flow Φ–; the leaving flow Φ+, and the net flow Φ as resumed in Table 5:

Table 5 Calculating the flow

E1 E2 E3 E3 E5 E6
E1 0.296 0.250 0.268 0.100 0.185
E2 0.462 0.389 0.333 0.296 0.500
E3 0.236 0.180 0.333 0.056 0.429
E4 0.399 0.505 0.305 0.223 0.212
E5 0.444 0.515 0.487 0.380 0.448
E6 0.286 0.399 0.250 0.432 0.133
+
Φ 0.220 0.396 0.247 0.329 0.455 0.300
–
Φ 0.366 0.379 0.336 0.349 0.162 0.355
Φ –0.146 +0.017 –0.089 –0.020 +0.293 –0.055

with:

Leaving flow : Φ + ( E1 ) = ∑π ( E , E )
1 i (1)

Entering flow : Φ – ( E ) = ∑π ( E , E )
1 i 1 (2)

Net flow : Φ ( E1 ) = Φ + ( E1 ) – Φ − ( E1 ) (3)

The classification of the PROMETHEE I methods

On base relations of upgrade of the PROMETHEE methods, we arrive to the partial
classification following in Figure 12:
118 L. Azzabi et al.

Figure 12 Classification of the thickness of sheet metal

E E

E E
E
E

The different identification of the six thickness of the sheet metal is described in Table 6:

Table 6 Shortenings identification

E1 Sheet metal thickness of the first arrival

E2 Sheet metal thickness of the second arrival
E3 Sheet metal thickness of the third arrival
E4 Sheet metal thickness of the fourth arrival
E5 Sheet metal thickness of the fifth arrival
E6 Sheet metal thickness of the sixth arrival

It is clear that E5 outclasses is the level identified the height variation of the tolerance of
sheet metal thickness, then in the following step (improve), we search the solution of the
primary sheet metal thickness of the arrival n°5.

4.4 Improve
The cause’s improvement of variation of the tolerance of sheet metal thickness is effected
by brainstorming methods to find solutions summarise in the Figure 13:

Figure 13 List of solution by brainstorming

 Six Sigma based multicriteria approach to improve decision settings 119

4.5 Control
The control step consists to identify with the graphics the improvement of problems
dysfunctions and non-conformities proceeding from the sheet metal realised after step
improve. The objective of this control is to identify a level of realisation and application
the solution expressed The control of the sheet metal as their weight and thickness is
determined through of a standard dimension of 18 parts, on the reception of matters, the
result of the reception control are following in Table 7:

Table 7 Measure of the thickness of sheet metal

W E Min Max

20.67 2.52 2.49 2.51

41.34 2.52 2.49 2.51

62.01 2.52 2.49 2.51

82.68 2.49 2.49 2.51

103.35 2.49 2.49 2.51

124.02 2.5 2.49 2.51

144.69 2.5 2.49 2.51

165.36 2.5 2.49 2.51

186.03 2.49 2.49 2.51

206.7 2.5 2.49 2.51

227.37 2.52 2.49 2.51

248.04 2.53 2.49 2.51

268.71 2.53 2.49 2.51

289.38 2.53 2.49 2.51

310.05 2.49 2.49 2.51

330.72 2.51 2.49 2.51

351.39 2.49 2.49 2.51

372.06 2.51 2.49 2.51

The schematisation of the EWMA control chart indispensable for the values held in the
sample is shown Figure 14.
The graphics of control chart is conform of the minimum and maximum level
demanded by norms of the contract. Then, the receiving metal sheet process is
conformable, this is well identified by the bell of Six Sigma schematised in Figure 15
120 L. Azzabi et al.

Figure 14 EWMA control charts of the thickness of sheet metal

E W M A :C o ntr o l c ha r ts o f thi k ne s s o f s he e t me ta l
2 ,5 2 0

UC L= 2 ,5 1 7 6 8

2 ,5 1 5

2 ,5 1 0
EWMA

_
_
X = 2 ,5 0 7 7 8

2 ,5 0 5

2 ,5 0 0
LC L= 2 ,4 9 7 8 7

1 3 5 7 9 11 13 15 17
S a mp le

Figure 15 Variation of the thickness of sheet metal

The bell of Six Sigma designates although the process of checking of the variation
tolerance sheet metal thickness is centralised a top relative, then the process is
conformable.
 Six Sigma based multicriteria approach to improve decision settings 121

5 Conclusions

Six Sigma is the most fervent managerial methodology not only in manufacturing area
but also in the services industry. Many investigations have indicated that Six Sigma can
increase an organisation’s competitive capability and enhance the quality of products or
services by conducting the projects. Notably, project generation and priority performs the
most critical parts while carrying out Six Sigma initiations. A good decision for project
generation not only provides profits but also increases customer satisfaction. However,
there still lacks a well-structured approach to assist a company in creating the feasible
projects.
However, with a Six-Sigma approach, stability testing is viewed as a process where
the precision must be adequate to distinguish between good and bad product with a high
degree of confidence, the recourse to the method Six Sigma for the improvement of the
product system represents a progress importing for the concern, with such gait permits to
pursue simultaneously several qualitative quantitative objectives of discount product
system, in order to make firm a long-term improvement.
To boot the tools integration multicriteria in our gait methodological, in a few phases
of the method, Six Sigma has supplied a significant high-speed problems resolution and
situation’s choice often contradictory. Lastly, the gait methodological proposed for the
improvement of the product system is not cut up in time of specification of the
intervening of each phase of the method Six Sigma in under-systems, to boot there is
absence from enforcement of the two scale practical gait to distinguish retails in times
situation’s choice, as everything the process of resolution and of improvement product
system with the multicriteria tools enforcement.

References
Adler, P.S., Goldoftas,. B. and Levine, D.I. (1999) ‘ Flexibility versus efficiency? A case study of
model changeovers in the Toyota production system’‚ Organization Science, Vol. 10, No. 1,
pp.43–68.
Albadvi, A., Chaharsooghi, S.K. and Esfahanipour, A. (2007) ‘Discrete optimization decision
making in stock trading: an application of PROMETHEE’‚ European Journal of Operational
Research, Vol. 177, pp.673–683.
Azzabi, L., Ayadi, D., Kobi, A., Robledo, C., Boujelbene, Y. and Chabchoub, H. (2007)
‘Improvement of human safety in the complex system: integration of Six Sigma methods’‚ 5th
International Conference on Quality and Reliability, ICQR (2007).
Barney, M. (2002a) ‘Macro, meso, micro: Six Sigma’‚ The Industrial – Organizational
Psychologist, Vol. 9, No. 4, pp.104–107.
Brans, J.P., Mareschal, B. and Vincke, P. (1984) ‘PROMETHEE: a new family of outranking
methods in multicriteria analysis’, Operational Research, Elsevier Science Publishers B.V.,
North Holland.
Brans, J.P., Vincke, P. and Mareschal, B. (1986) ‘How to select and how to rank projects: the
PROMETHEE method’, European Journal of Operational Research, Vol. 24, pp.228–238.
Braverman, J.D. (1981) Fundamentals of Statistical Quality Control, Reston Publishing Co. Inc.,
Virginia.
Brewer, P.C. (2004) ‘Six Sigma helps a company create a culture of accountability’, Journal of
Organizational Excellence, Vol. 23, No. 3, pp.45–59.
122 L. Azzabi et al.

Breyfogle, F.W. (1999) Implementing Six Sigma: Smarter Solutions Using Statistical Methods,
Wiley, New York.
Breyfogle, F.W., Cupello, J.M. and Meadows, B. (2001) Managing Six Sigma: A Practical Guide
to Understanding, Assessing, and Implementing the Strategy That Yields Bottom-Line Success‚
John Wiley & Sons, Danvers, MA.
Choo, A.S., Linderman, K.W. and Schroeder, R.G. (2007a) ‘Method and psychological effects on
learning behaviors and knowledge creation in quality improvement projects’‚ Management
Science, Vol. 53, No. 3, pp.437–450.
Choo, A.S., Linderman, K.W. and Schroeder, R.G. (2007b) ‘Method and context perspectives on
learning and knowledge creation in quality management’‚ Journal of Operations
Management, Vol. 25, No. 4, pp.918–931.
Clifford, L. (2001) ‘Why you can safely ignore Six Sigma’‚ Fortune, January, Vol. 140.
Collette, Y. and Siarry, P. (2003) Multiobjective Optimization, Springer, New York.
Dasgupta, T. (2003) ‘Using the Six-Sigma metric to measure and improve the performance of a
supply chain’‚ Total Quality Management & Business Excellence, Vol. 14, No. 3, pp.355–367.
Dave, N. (2002) ‘How to compare Six Sigma, lean and the theory of constraints, a frame work for
choosing what’s best for your organization‚ process improvement’‚ Quality Progress.
EEA (2003) ‘Assessment of information related to waste and material flows. A catalogue of
methods and tools’‚ Technical Report 96, European Environment Agency, Project Manager:
Dimitrios Tsotsos, Prepared by: Despo Fatta, Stephan Moll, European Topic Centre on Waste
and Material Flows, Copenhagen, Denmark, 2003.
Fiedler, T. ( 2004) ‘Mopping up profits: with 3M sitting on solid earnings, CEO James McNerney
handled his fourth annual meeting like a contented company veteran’‚ Star Tribune, Metro ed.,
May 12, Minneapolis, MN.
Figueira, J., Greco, S. and Ehrgott, M. (2005) ‘Multiple criteria decision analysis’, State of the Art,
Springer.
Gowen, C.R. III and Tallon, W.J. (2005) ‘Effect of technological intensity on the relationships
among Six Sigma design, electronic-business, and competitive advantage: a dynamic
capabilities model study’‚ Journal of High Technology Management Research, Vol. 16, No. 1,
pp.59–87.
Hahn, G.J., Doganaksoy, N. and Hoerl, R. (2000) ‘The evolution of Six Sigma’‚ Quality
Engineering, Vol. 12, No. 3, pp.317–326.
Henderson, K.M. and Evans, J.R. (2000) ‘Successful implementation of Six Sigma: benchmarking
General Electric company’‚ Benchmarking: An International Journal, Vol. 7, No. 4,
pp.260–281.
Linderman, K., Schroeder, R.G. and Choo, A.S. (2006) ‘Six Sigma: the role of goals in
improvement teams’‚ Journal of Operations Management, Vol. 24, No. 6, pp.779–790.
Linderman, K., Schroeder, R.G., Zahee, S. and Choo, A. (2003) ‘Six Sigma: a goal-theoretic
perspective’‚ Journal of Operations Management, Vol. 21, pp.193–203.
Luan, J. (2001) Multivariate Control Charts for Complex Processes in Multivariate Total Quality
Control, Springer.
McClenahen, J.S. (2004) ‘New world leader’‚ Industry Week, Vol. 253, No. 1, pp.36–39.
Meziane, F., Vadera, S., Kobbacy, K. and Proudlove, N. (2000) ‘Intelligent systems in
manufacturing: current developments and future prospects’‚ Integrated Manuf. Syst., Vol. 11,
No. 4, pp.218–38.
Montgomery, D.C. (2001) Introduction to Statistical Quality Control, 4th ed., John Wiley & Sons,
Inc., Canada.
Murdoch, J. (1979) Control Charts, The Macmillan Press Ltd., London.
Naumann, E. and Hoisington, S.H. (2001) ‘Customer centered Six Sigma linking customers process
improvement and financial results. ‘ASQ Quality Press. optimisation of decentralised energy
systems’’, Omega, Vol. 36, pp.766–776.
 Six Sigma based multicriteria approach to improve decision settings 123

Pande, P.S., Neuman, R.P. and Cavanagh, R.R. (2002) The Six Sigma Way, Team Fieldbook: An
Implementation Guide for Process Improvement Teams, McGraw-Hill, New York, NY.
Proudlove, N.C., Vadera, S. and Kobbacy, K.A.H. (1998) ‘Intelligent management systems in
operations: a review’‚ J. Oper. Res. Soc., Vol. 49, No. 7, pp.682–99.
Pyzdek, T. (2003) The Six Sigma Handbook: A Complete Guide for Green Belts, Black Belts, and
Managers at all Levels, McGraw-Hill, New York, NY.
Rogers, M. and Bruen, M. (1996) ‘Using ELECTRE to rank options within a highway
environmental appraisal – two case studies’‚ Civil Engineering Systems, Vol. 13, No. 3,
pp.203–21.
Rogers, M. and Bruen, M. (1998) ‘Using ELECTRE III to elect the best route option for the Dublin
port access motorway’‚ Proceedings Of The 16th European Conference on Operational
Research, EURO XVI, Environmental Planning Stream, Brussels, July.
Rogers, M., Bruen, M. and Maystre, L. (2000) Electre and Decision Support, Kluwer Academic
Publishers, London.
Sahoo, A.K., Tiwari, M.K. and Mileham, A.R. (2008) ‘Six Sigma based approach to optimize
radial forging operation variables’‚ Journal of Materials Processing Technology, pp.125–136.
Shanian, A. and Savadogo, O. (2006) ‘A non-compensatory compromised solution for material
selection of bipolar plates for polymer electrolyte membrane fuel cell (PEMFC) using
ELECTRE IV’‚ Electrochimica Acta, Vol. 51, pp.5307–5315.
Singh, N. and Singh, S. (2006) ‘EWMA control chart in detecting and diagnosing a persistent shift
in a process mean’‚ Proceeding of the 2nd IMT-GT Regional Conference on Mathematics,
Statistics and Applications, University Sains Malaysia, Penang.
Sinha, K.K. and Van de Ven, A.H. (2005) ‘Designing work within and between organizations’‚
Organization Science, Vol. 16, No. 4, pp.389–408.
Snee, R.D. and Hoerl, R.W. (2003) ‘Leading Six Sigma: a step-by-step guide based on experience
with GE and other Six Sigma companies’‚ Financial Times, Prentice Hall, Upper Saddle
River, NJ.
Wacker, J.G. (2004) ‘A theory of formal conceptual definitions: developing theory-building
measurement instruments’‚ Journal of Operations Management, Vol. 22, No. 6, pp.629–650.

View publication stats