You are on page 1of 16

QUALITY AND RELIABILITY ENGINEERING INTERNATIONAL

Qual. Reliab. Engng. Int. 2007; 23:3–18


Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/qre.822

Special Issue Fortification of Six Sigma:


Expanding the DMAIC Toolset
Loon Ching Tang1,∗,† , Thong Ngee Goh1 , Shao Wei Lam1 and Cai Wen Zhang2
1 Department of Industrial and Systems Engineering, National University of Singapore, 1 Engineering Drive 2,
Singapore 117576, Republic of Singapore
2 Hitachi Global Storage Technologies (Singapore) Pte Ltd, 4 Kaki Bukit Avenue 1, Singapore 417939,
Republic of Singapore

Six Sigma as a quality improvement framework cannot remain static if it is to sustain


its value for businesses beyond the first waves of applications. This paper explores
the possibilities of enhancing the usefulness and effectiveness of Six Sigma by
the integration of established Operations Research/Management Science (OR/MS)
techniques. In this paper, we elucidate the needs for OR/MS techniques to enhance
Six Sigma deployment in operational and transactional environments and propose
a new training roadmap for core Six Sigma professionals (Six Sigma Black Belts)
which incorporates these techniques. A matrix relating the components of the
proposed training curriculum to the actual deliverables during implementation for
a hybrid of operational and transactional environments is also presented. A practical
case study is also presented to demonstrate the usefulness of the OR/MS tools in a
typical transactional environment. Copyright c 2007 John Wiley & Sons, Ltd.
Received 10 November 2005; Revised 10 February 2006; Accepted 21 July 2006

KEY WORDS: Six Sigma; operations research; management science; Black Belt training; queueing

1. INTRODUCTION

S
ix Sigma is a systematic, highly disciplined, customer-centric and profit-driven organization-wide
strategic business improvement initiative that is based on a rigorous process-focused and measurement-
driven methodology. Six Sigma makes use of sound statistical methods and quality management
principles to improve processes and products via the Define–Measure–Analyze–Improve–Control (DMAIC)
quality improvement framework to meet customer needs on a project-by-project basis. With many high-profile
adoptions by companies such as General Electric (GE) in the 1990s, Six Sigma has spread like wild fire towards
the end of the 20th century1.
As an approach to organizational excellence, Six Sigma as it has been practiced can certainly be enhanced1,2 .
Despite its current success, Six Sigma has its inherent limitations and cannot be a universal solution for any
organization in any situation1 . Consequently, there have been many developments in enhancing the power
of Six Sigma for the improvement of business processes (see, for example, Recker and Bolstorff3 , Smith4 ,
Bossert5 , Murugappan and Keeni6, Revere and Black7 , Catherwood8, Dalgeish9 and Gurunatha and Siegel10 ).

∗ Correspondence to: Loon Ching Tang, Department of Industrial and Systems Engineering, National University of Singapore, 1 Engineering
Drive 2, Singapore 117576, Republic of Singapore.
† E-mail: isetlc@nus.edu.sg

Copyright 
c 2007 John Wiley & Sons, Ltd.
4 L. C. TANG ET AL.

Most of these efforts described enhanced frameworks and processes, but did not emphasize a suitable training
approach that can be related to the deliverables during implementation. In this paper, apart from proposing
the systematic inclusion of Operations Research/Management Science (OR/MS) techniques into the existing
Six Sigma DMAIC framework, emphasis has been placed on the design of a well-integrated Six Sigma Black
Belt (BB) training curriculum. The proper training of core Six Sigma professionals has been identified as one
of the fundamental success factor for the effective deployment of Six Sigma11,12 .
OR/MS techniques have been extensively applied to a wide range of areas including transportation,
manufacturing, telecommunication, the military, financial planning, public services and healthcare. They are
effective tools for improving the efficiency and productivity of organizations. OR/MS techniques, as implied by
the name, are concerned with the conduct and improvement of ‘operations’ as well as the practical management
of organizations. Another important characteristic of OR/MS is that, rather than simply improving the status
quo, its primary goal is to identify a best possible course of action13. This is also a fundamental goal of all
Six Sigma projects, commonly exemplified through the ‘Improve’ phase. Consequently, it is quite natural to
integrate OR/MS techniques into the framework of Six Sigma.
This paper explores the possibilities of further enhancing the usefulness and effectiveness of Six Sigma via
the integration of OR/MS techniques into Six Sigma deployment. To effectuate this integration and motivated
from the consideration that proper capability development of Six Sigma BBs is one of the key success factors in
the implementation of Six Sigma, a new roadmap for training BBs which contains an expanded curriculum
that includes new OR/MS techniques is proposed. A matrix that relates the training deliverables to actual
project deliverables is also presented to explicate the close relationship between training objectives and project
deliverables. A case study to improve the efficiency of processes in the transactional environment through more
effective manpower allocation is then presented to illustrate the usefulness of OR/MS techniques within the
proposed roadmap.

2. INTEGRATION OF OR/MS INTO SIX SIGMA DEPLOYMENT

Six Sigma is not only a performance target that applies to a single critical-to-quality (CTQ) characteristic14,
but has evolved into a strategic business-improvement approach that seeks to find and eliminate causes of
mistakes or defects in business processes by focusing on outputs that are of critical importance to customers15.
The ultimate goal of Six Sigma is business improvement, particularly in terms of the bottom-line results,
customer satisfaction and market share.
Customer satisfaction is a close reflection of the state of the business and depends on the ability of a business
to achieve three key objectives: (1) delivering a defect-free product or service; (2) delivering a product or
service on schedule; and (3) delivering a product or service at the lowest possible cost14 . Although standard
Six Sigma tools and methods are effective in handling the first problem, they are typically found lacking in
the ability to resolve the latter two issues. Most of the existing Six Sigma tools are statistical methods and
quality management tools, such as Design of Experiments (DOE), Response Surface Methodology (RSM),
Robust Design, Statistical Process Control (SPC), Quality Function Deployment (QFD), Failure Mode and
Effect Analysis (FMEA), Capacity Analysis, Hypothesis Testing, Analysis of Variance (ANOVA), Regression
Analysis, and so on16,17 . Common statistical methods for quality engineering and quality management tools are
usually insufficient for tackling many business-improvement related problems such as production and service
planning and scheduling, inventory control and management, supply chain management, operations scheduling
and workforce scheduling. OR/MS techniques are well positioned to effectively solve these problems, among
others. Many techniques in the possession of OR/MS practitioners could and should be integrated into Six Sigma
applications to complement the existing standard Six Sigma tools. Existing quality management tools such
as QFD can also be greatly enhanced with OR/MS tools18 . In general, common concepts and procedures of
operations management and project management would play an extremely useful role in Six Sigma projects.
Considering the fact that the Six Sigma programs adhere strictly to a systems perspective towards
quality improvements, it is quite natural to observe the trend of integrating Six Sigma with other business
improvement tools and methods such as Lean Manufacturing3–5 , the Capability Maturity Model (CMM)6 ,

Copyright 
c 2007 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2007; 23:3–18
DOI: 10.1002/qre
FORTIFICATION OF SIX SIGMA 5

Total Quality Management7, ISO9000 and ISO90018,9 , reliability engineering10 and human resource
functions11. However, there has been no proposed systematic incorporation of relevant and powerful OR/MS
techniques to drive product and process improvements through these existing quality tools. OR/MS techniques
also play important roles in operational management systems such as Material Requirements Planning (MRP I),
Manufacturing Resource Planning (MRP II) and Enterprise Resources Planning (ERP). All of these systems,
in turn, are very useful to improve the overall business performance and should be integrated with Six Sigma
deployment, probably in a parallel manner (see, for example, Adams et al.19).
Many of the OR/MS tools suggested in this article are not new. It is the ‘integration’ of these techniques
in a purposeful, systematic and sequential manner through the new fortified Six Sigma roadmap that will
achieve synergistic results far exceeding what is possible with isolated application of single tools or methods.
The ‘Define’ phase of Six Sigma involves scores of problems, such as project selection and planning, production
and service planning, training and education planning, resource allocation and investment decision making,
which conventional Six Sigma tools cannot handle but OR/MS techniques can. One of the most useful purposes
of applying OR/MS techniques is for ‘improvement’. OR/MS techniques such as mathematical programming,
queueing and simulation techniques can improve the analysis and operations of manufacturing and transactional
systems in Six Sigma projects. Consequently, various OR/MS techniques are well fitted into the ‘Improve’
phase in Six Sigma deployment. In the ‘Control’ phase of Six Sigma deployment, OR/MS techniques can
also be applied to optimize the design of control charts and control schemes, and to improve maintenance
management20–23 , etc.
In the long run, to really accomplish the simultaneous objectives of Six Sigma and powerful business
improvement tools such as Lean Manufacturing, practitioners will not only have to gain a solid understanding
of additional statistical tools, but knowledge of industrial engineering and operations research techniques, such
as systems simulation and factory modeling, mathematical optimization methods and queueing networks2.
Although some of the techniques may have been included in the training of Master Black Belts (MBBs), the
number of MBBs is far too small in a typical enterprise to have an impact. Future successes of Six Sigma could
only be brought about by dedicated teams of BBs mastering a set of synergistic tools arranged in a compact and
logical sequence for problem solving.

3. A NEW SIX SIGMA BB TRAINING ROADMAP


In developing a new training program for Six Sigma BBs, we first consider the training needs in different
environments to derive the relevant OR/MS tools for each of these environments. In Table I, we present
some important tools proposed for the design of an expanded curriculum for Six Sigma BBs training under
operational and transactional environments. This is a broad guideline as some degree of customization in the
design of Six Sigma curriculum is inevitable given the wide spectrum of different operational and transactional
environments. Note that only the basic OR/MS techniques have been incorporated in the new curriculum so
that these topics could be covered within a usual four-week (or approximately 160 hours) BB training program.
Furthermore, some fringe materials in the current DMAIC curriculum should be trimmed in order to incorporate
new and more relevant OR/MS techniques within the four-week training schedule according to the needs of
different environments. Table II summarizes all supplementary OR/MS techniques extracted from Table I.
A brief description of these OR/MS tools with emphasis on their applicability within the Six Sigma framework
is given later in this section.
Six Sigma is a process-focused quality improvement initiative. The nature of ‘processes’ in
manufacturing/operational and transactional environments are somewhat distinct and thus demand slightly
different tool sets during the implementation of Six Sigma as well as in a BB training program. From Table I,
it can be seen that the major difference between manufacturing/operational and transactional roadmap is
in ‘Analyze’ and ‘Improve’ phases, and some slight difference in other phases. OR/MS techniques such
as forecasting, queueing, simulation and modeling are essential tools in the ‘Analyze’ phase since system-
level analysis is usually needed in a transactional environment. In the ‘Improve’ phase, major tools used
in manufacturing/operational environment are DOE techniques; in contrast, queueing and mathematical

Copyright 
c 2007 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2007; 23:3–18
DOI: 10.1002/qre
6 L. C. TANG ET AL.

Table I. An expanded list of Six Sigma tools

Manufacturing/operational environment Transactional environment


Define Project selection Project selection
Probabilistic risk thinking and strategic planning Probabilistic risk thinking and strategic planning
Decision analysis Decision analysis
Process mapping Process mapping
Project management tools Project management tools
Measure QFD and Kano analysis QFD and Kano analysis
Gap analysis
Sampling (data quantity and data quality) Sampling (data quantity and data quality)
Measurement system analysis Measurement system analysis
SPC Part I (concepts, implications of instability)
Run charts (or time series graphs)
Capability analysis Capability analysis
Monte Carlo simulation and statistical distributions
Analyze Basic graphical improvement tools Basic graphical improvement tools
FMEA FMEA
Hypothesis testing Hypothesis testing
Confidence intervals
ANOVA ANOVA
Correlation and regression analysis Correlation and regression analysis
Reliability models and measures
Cost analysis
Forecasting
Basic queueing systems
Simulation and modeling
Improve DOE (factorial, fractional factorial, DOE (factorial, fractional factorial
blocking, nested and RSM) and blocking)
Robust design
Optimization and control of queues
Mathematical programming techniques
Heuristics
Sensitivity analysis Sensitivity analysis
Control Mistake proofing Mistake proofing
Validation testing Validation testing
Control plans Control plans
SPC Part II (control charts) Basic control charts

Table II. A summary of OR/MS techniques


integrated into Six Sigma phases

OR/MS tools
Define Mathematical programming techniques for
resource allocation and project selection
Decision analysis
Project management tools
Analyze Forecasting
Basic queueing systems
Simulation and modeling
Improve Optimization and control of queues
Mathematical programming techniques
Heuristics

Copyright 
c 2007 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2007; 23:3–18
DOI: 10.1002/qre
FORTIFICATION OF SIX SIGMA 7

programming techniques are usually needed for transactional environments. Although, OR/MS techniques
appear to be much more applicable in transactional environment, it should be noted that Six Sigma BBs working
in manufacturing sectors are also expected to tackle transactional issues. This underscores the importance and
necessity of integrating OR/MS techniques with Six Sigma. The current evolution of Six Sigma is not simply
a transition from the original manufacturing sectors to service sectors, but a vehicle for making deep cultural
change, inculcating system thinking and problem solving that lead to quantifiable benefits.

3.1. Mathematical programming


Mathematical programming techniques include various mathematically rigorous optimization tools such as
Linear/Nonlinear/Integer/Mixed Integer/Dynamic Programming, Multiobjective Mathematical Programming
(MMP), Goal Programming (GP) and Stochastic Programming. Problems selected for Six Sigma projects are
not limited to engineering topics but also cover quality issues in transactional, commercial and financial areas,
with an explicit and strong customer focus24. Mathematical programming techniques, sometimes in conjunction
with sensitivity analysis, can be exploited to solve such problems. These techniques have been prevalently used
in production planning and operations management and may take a wide variety of forms depending on the
particular problem situation and various objectives involved. For example, given some limited capital budget,
the decision of how to select a subset of proposed Six Sigma projects to invest in can be readily modeled as
a single or multiobjective knapsack problem25. They can also be deployed in Six Sigma projects for project
selection and planning during the ‘Define’ phase of the Six Sigma deployment for selecting an optimal number
of projects or to achieve profit maximization or cost minimization goals in general. In particular, as the objective
of mathematical programming techniques is ‘optimization’, various techniques can naturally be weaved into the
‘Improve’ phase of Six Sigma deployment to solve various optimization problems. For example, a general
framework for dual response problem can be cast using MMP26,27 . Nonlinear optimization techniques can be
applied, for example, to optimize mechanical design tolerancing28 and product design capability29, as well as to
estimate various statistical parameters. In the ‘Control’ phase of Six Sigma, nonlinear optimization techniques
have been applied to optimize the design of control charts, including economic design, economic-statistical
design and robust design, design of sampling schemes and control plans, etc. Examples of these applications
can be found in Tagaras30, Crowder31, Rahim32, Chung33, McWilliams et al.34 and Rohleder and Silver35 , to
name a few. In addition, a brief introduction of meta-heuristics, which is a class of effective solution techniques
for solving various mathematical programming and combinatorial optimization problems, can also be included
into the training of Six Sigma BBs. Detailed treatment, however, can be deferred to a MBB program.

3.2. Queueing
Queueing theory is concerned with understanding the queueing phenomenon and how to operate queueing
systems in the most effective way. Providing too much service capacity to operate a system incurs excessive
costs; however, insufficient service capacity can lead to annoyingly long waiting times resulting in dissatisfied
customers and loss of business. Within the context of business improvement, queueing techniques have been
applied to solve problems pertaining to effectively planning and operating of service and production systems.
Some application areas include maintenance management, resource scheduling such as in wafer fabrication and
the healthcare, telecommunication, transportation and airline industries, and so on. Queueing techniques can
play a useful role in Six Sigma deployment particularly in analyzing and improving a system providing services
in the ‘Analyse’ and ‘Improve’ phases. A case study on the application of queueing techniques in these two
phases is presented in Section 4.

3.3. Simulation and modeling


Simulation is an exceptionally versatile technique and can be used (with varying degrees of difficulty)
to investigate virtually any kind of stochastic system13 . For instance, simulation can help to improve the
design and development of products, and manufacturing and service processes for a wide variety of systems

Copyright 
c 2007 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2007; 23:3–18
DOI: 10.1002/qre
8 L. C. TANG ET AL.

(e.g. queueing, inventory, manufacturing, distribution, etc.). Simulation has been successfully deployed in
Design for Six Sigma to replace costly preliminary prototype testing and tolerancing. Also, simulation provides
an attractive alternative to more formal statistical analysis in, for example, assessing how large a sample is
required to achieve a specified level of precision in a market survey or in a product life test36 . Bayle et al.37
reported the approach of integrating simulation modeling, DOE and engineering and physical expertise to
successfully design and improve a braking subsystem that would have not been accomplished by any individual
tool or method alone. For system operations analysis, simulation is an indispensable companion to queueing
techniques as it is much less restrictive in modeling assumption38. Queueing and simulation techniques also
play important roles in inventory control39 and supply chain management in organizations.

3.4. Forecasting
Every company needs to at least do some forecasting in order to strategize and plan; the future success of
any business depends heavily on the ability of its management to forecast well13 . However, the availability
of ‘good’ data is crucial for the use of forecasting methods; otherwise, it would turn into ‘garbage in,
garbage out’. The accuracy of forecasts and the efficiency of subsequent production and service planning
are related to the stability and consistency of the processes which are, in turn, influenced by successful
applications of standard Six Sigma tools. Six Sigma tools and methods identify and eliminate process defects
and diminish process variation. Six Sigma also requires that data be collected in an accurate and scientific
manner. The combination of defect elimination, variation diminishing and more accurate and scientific data
collection allows forecasting to be conducted more easily and effectively, which will, in turn, help improve
the effectiveness of production and service planning, operations scheduling and management. In contrast, if
the processes are erratic, then forecasting and subsequent production and service planning and operations
scheduling will be much less effective or useful. Important applications of forecasting techniques within the
context of operations management include demand forecasting, yield forecasting and inventory forecasting that
is essentially the conjunction of the first two. In addition, forecasting results are important inputs to other OR/MS
techniques such as mathematical programming, queueing, simulation and modeling, etc.

3.5. A roadmap that integrates OR/MS techniques


In the development of a new Six Sigma roadmap which is fortified with OR/MS tools, emphasis is placed on
the effective capability development of BBs who are the core Six Sigma practitioners. The type of training BBs
should receive is a function of the environment in which they work; and training curricula should be designed
accordingly. It is also important in the presentation of the tools to provide roadmaps and step-by-step procedures
for each tool and overall methods16. Characteristics of Six Sigma that make it effective are the integration of
the tools within a systematic framework such as the DMAIC process. While most curricula proposed in the
literature manifest the integration (see, for example, Hoerl16 and Hahn et al.40), the linking and sequencing
of the tools and the relationship between the components of the curricula with actual project deliverables are
less apparent. In this paper, leveraging on previous programs and our consulting experiences, a sequence of
deliverables and the associated tools needed in a typical BB project is conceived; bearing in mind the tasks that
need to be accomplished in DMAIC phases. A training roadmap which relates the training deliverables to actual
project deliverables during implementation is proposed in a matrix form shown in Table III.
Table III presents a matrix that summarizes the DMAIC framework under both manufacturing/operational
and transactional environments. The vertical dimension of the matrix lists the deliverables in each DMAIC
phase and the horizontal dimension lists the tools/techniques that could be used to serve the purposes in the
vertical dimension. The flow of deliverables is self-explanatory as they represent tasks/milestones in a typical
DMAIC process. The tool set across the horizontal dimension has been fortified with OR/MS techniques to
meet the higher expectation of Six Sigma program in delivering value to an enterprise. It should be noted that
while it is conceivable that a specific OR/MS technique could be applied in multiple phases we have made
each basic OR/MS technique appear with intentional precision within the DMAIC process corresponding to its
major area of application for the purpose of conciseness. Nonetheless, the placement of various techniques is
by no means rigid, due to broad scope of coverage in Six Sigma projects. The matrix can be used as a roadmap

Copyright 
c 2007 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2007; 23:3–18
DOI: 10.1002/qre
Table III. A roadmap integrated OR/MS tools that BBs can follow

Copyright 
FORTIFICATION OF SIX SIGMA

c 2007 John Wiley & Sons, Ltd.


Use in manufacturing/operational environment; use in transactional environment; x use in both environments.

Qual. Reliab. Engng. Int. 2007; 23:3–18


DOI: 10.1002/qre
9
10 L. C. TANG ET AL.

for BBs to implement their projects and as a training curriculum for a new breed of Six Sigma BBs. While more
elaborate techniques can also be included, it is felt that the current tool set is the most essential and can be
covered in a typical four-week training program for BBs.

4. CASE STUDY: MANPOWER RESOURCE PLANNING


Here, we give an illustration of how the Six Sigma framework is implemented to reduce the waiting times in
a hospital retail pharmacy through more efficient manpower allocation strategies. This case study is discussed
with particular emphasis on the application of queuing techniques in the ‘Analyse’ and ‘Improve’ phases.
Prior to the introduction of Six Sigma, continuous efforts have been made to expedite the work flow by the
pharmacy staff in order to reduce the waiting times without a systemic examination of relevant work processes.
These efforts turned out to be insufficient in achieving a waiting time target set by the management. In the
‘Define’ phase, Six Sigma consultants together with the hospital management selected this project given its
urgency and proximity of the process to the customers. Furthermore, as there were multiple satellite pharmacies
with similar processes as the central pharmacy, the project will be able to reap benefits beyond this particular
department. The success of this project would enable the Six Sigma team to garner more extensive buy-in and
support from the management and other hospital staff.
Four major tasks, namely, ‘Typing’, ‘Packing’, ‘Checking’ and ‘Checking and Dispensing’ in the current
process were identified. The entire process can be represented by an open queueing network shown in Figure 1.
With reasonable estimates of data on the arrival, service and rework rates, the mean waiting times and value-
added times for each process can be computed. The arrival rate profile measured based on the number of arrival
every 10 minutes is shown in Figure 2. After accounting for outliers in the arrival rate profile, the profile was
discretized taking into account several practical constraints related to manpower deployment. Estimated service
and rework rates are shown in Table IV. With these data collected, estimates of mean waiting times and queue
lengths can be derived using basic queueing results. Steady state queueing analysis is adequate here as the
arrival and service rates are fast enough to ensure the system reaches its steady state in a short time. Inter-arrival
and service times were assumed exponentially distributed. These assumptions were validated with exponential
probability plots. The entire process can be represented as a system of interconnected M/M/1 and M/M/s
service stations. Mean total waiting times for the entire drug dispensing process in the pharmacy can then be
easily computed using basic queueing results41 .

4.1. Sensitivity analysis


At the ‘Improve’ phase, many different types of sensitivity analysis can be conducted to assess the quality of
different process configurations using theoretical models derived from basic queuing theory. Quality can be
defined in terms of the ability to meet a desired target and its robustness under varying operational conditions.
The impact of different manpower allocations and process configurations together with robustness under varying
operational conditions on the overall waiting times are investigated. Results for the lower arrival rate extracted
from the estimated arrival rate profile are presented.
Figure 3 shows the impact of varying the number of packers and dispensing pharmacists on the mean total
waiting times. It was observed that waiting times would be increased significantly if the number of packers was
reduced to eight. With number of packers above eight, the waiting times were observed to be relatively stable
over the different numbers of dispensing pharmacists experimented. It was further observed that by having one
additional dispensing pharmacist, the targeted 15 minutes of mean total waiting times could potentially be met.
Further sensitivity analysis was conducted to assess the implication of having different process configurations.
A new ‘Screening’ process was implemented at the point when the pharmacy receives prescriptions from
patients (prior to the existing ‘Typing’ process). It was expected that this would help to alleviate problems
associated with errors in prescriptions and medicine shortages. A pilot run was implemented and service and
arrival rates estimates were obtained from these runs and mean waiting times predicted for each process. Table V
shows the improvements in the average sojourn times of each sub-process and the mean total waiting time for

Copyright 
c 2007 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2007; 23:3–18
DOI: 10.1002/qre
FORTIFICATION OF SIX SIGMA 11

Source

Checking
Typing
Packing Dispensing
Legend:
: Queues : External Flows

: Servers : Internal Flows


Sink

Figure 1. Queueing network representation of drug dispensing process

Outliers

Figure 2. Profile of estimated arrival rates

Copyright 
c 2007 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2007; 23:3–18
DOI: 10.1002/qre
12 L. C. TANG ET AL.

Table IV. Service and rework rates


Estimated service Estimated proportion
Process rates (jobs min−1 ) Rework routing of rework
Typing 1.92 Packing → Typing 0.025
Packing 0.20 Checking → Typing 0.025
Checking 1.70 Checking → Packing 0.025
Checking and Dispensing 0.19 Checking and Dispensing → Typing 0.001
Checking and Dispensing → Packing 0.001

View B
View A

89.0 90.0

79.0 80.0

69.0 70.0

59.0
Waiting Times

60.0
Waiting Times

49.0 50.0

39.0 40.0

29.0
30.0

19.0
20.0

9.0
10.0
8 9 10 11
8 9 10 11
Number of Packers
Number of Dispensing Pharmacists
8 Dispensing Pharmacists 9 Dispensing Pharmacists

10 Dispensing Pharmacists 11 Dispensing Pharmacists 8 Packers 9 Packers 10 Packers 11 Packers

View A View B
Figure 3. Sensitivity of total waiting times to variations in number of packers and dispensing pharmacists

processes before and after the addition of the ‘Screening’ process. There was an observable increase in the
productive time of pharmacists and the mean queue length in front of the dispensing process was shortened
from 13 to 3. As observed in the relatively flatter contours in Figure 4, the proposed new configuration is also
more robust against changes in manpower deployment over the packing and dispensing sub-processes.
Robustness of different manpower allocations and process configurations against varying arrival rates can also
be assessed through sensitivity analysis. Figure 5 shows the sensitivity of mean total waiting times subjected
to small perturbations in arrival rates over different number of dispensing pharmacists. It was observed that
the original configuration with only eight dispensing pharmacists would result in large increases in waiting
times when there were only small increases in the number of prescriptions arriving per hour. This provided the
management with insights on the frequently experienced phenomenon of doubling in waiting times on some
‘bad’ days.

Copyright 
c 2007 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2007; 23:3–18
DOI: 10.1002/qre
FORTIFICATION OF SIX SIGMA 13

Table V. Comparisons of sojourn times for each process and mean total
waiting times

Type of job Without screening With screening


Screening 0 2.6 (1 pharmacist)
Typing 0.6 (2 typists) 0.6 (2 typists)
Packing 5.9 (10 packers) 5.6 (10 packers)
Checking 6.2 (1 pharmacist) 5.0 (1 pharmacist)
Dispensing and checking 9.0 (8 pharmacists) 2.0 (7 pharmacists)
(only service times of this process)
Mean total waiting time 21.7 15.8

View B

View A

26.0 26.0
7 Dispensing Pharmacists With Screening 8 Packers With Screening 9 Packers With Screening
24.0 8 Dispensing Pharmacists With Screening 24.0 10 Packers With Screening 11 Packers With Screening
9 Dispensing Pharmacists With Screening
22.0 10 Dispensing Pharmacists With Screening 22.0
11 Dispensing Pharmacists With Screening

20.0 20.0
Waiting Times

Waiting Times

18.0 18.0

16.0 16.0

14.0 14.0

12.0 12.0

10.0 10.0
8 9 10 11 7 8 9 10 11
Number of Packers Number of Dispensing Pharmacists

View A View B
Figure 4. Sensitivity of total waiting times to variations in number of packers and dispensing pharmacists in new dispensing
process with screening

4.2. Relaxing exponential assumptions on service times distribution


Further sensitivity analysis can be performed to assess the effect of exponential assumptions on the distribution
of service times. This analysis demonstrated that the assumption of exponentially distributed service times
would result in more conservative system design choices. In this analysis, each service station i is assumed to
experience a Poisson arrival process with mean arrival rate λi . The same queueing network shown in Figure 1
with single and multiple servers service stations is assumed. Single server service stations, i, are assumed

Copyright 
c 2007 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2007; 23:3–18
DOI: 10.1002/qre
14 L. C. TANG ET AL.

View B

View A
44.0 44.0
85 prescriptions/hour 86 prescriptions/hour
8 Dispensing Pharmacists 9 Dispensing Pharmacists
39.0 87 prescriptions/hour 88 prescriptions/hour 39.0
89 prescriptions/hour 90 prescriptions/hour 10 Dispensing Pharmacists 11 Dispensing Pharmacists
91 prescriptions/hour
34.0 34.0

Waiting Times
Waiting Times

29.0 29.0

24.0 24.0

19.0 19.0

14.0 14.0

9.0 9.0
8 9 10 11 85 86 87 88 89 90 91
No. of Dispensing Pharmacists Mean Arrival Rates

View A View B
Figure 5. Sensitivity of total waiting times to variations in number dispensing pharmacists and patients arrival rates

to experience service processes that follows any general distributions with mean service rate µi (commonly
known as the M/G/1 queueing system). For service stations with multiple servers, there are si servers in
parallel with service processes of each server assumed to follow any general distributions that are independently
and identically distributed with mean service rate µi (commonly known as the M/G/s queueing system).
M/G/1
The mean queue length of each single server station, L̄i , can be easily computed with the well-known
M/G/1
Pollaczek–Khintchine result . Given L̄i
41 , the mean waiting times at each single server service station,
M/G/1
W̄i , can be computed using Little’s theorem42 . For computing the waiting times of queueing stations with
multiple servers (M/G/s), we apply an approximation by Cosmetatos43 . For this, we first compute the mean
M/M/s
waiting times of an M/M/s queueing system (W̄i ):

M/M/s M/M/s
W̄i = L̄i /λi

M/M/s
where the mean queue length of the M/M/s system, L̄i , is computed from

M/M/s (si ρi )2 ρi Pi,0


L̄i =
si !(1 − ρi )2

Copyright 
c 2007 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2007; 23:3–18
DOI: 10.1002/qre
FORTIFICATION OF SIX SIGMA 15

where ρi is the utilization of service station i given by λi /si µi and Pi,0 is the probability that station i would be
empty given by
si −1

(si ρi )n (si ρi )si −1
Pi,0 = +
n=0
n! si !(1 − ρi )
M/D/s
Next, we compute the mean waiting times of an M/D/s queueing system (W̄i ) as follows:

M/D/s 1 1 M/M/s
W̄i = · · W̄i
2 Ki
where
 √ 
4 + 5si − 2 −1
Ki = 1 + (1 − ρi )(si − 1)
16ρi si
The approximate mean waiting times for an M/G/s system can finally be computed from
M/G/s M/M/s M/D/s
W̄i ≈ CV 2i W̄i + (1 − CV 2i )W̄i , 0 ≤ CV 2i ≤ 1
where the squared coefficient of variation of service times (CVi ) is given by
Var(Ti )
CV 2i =
T̄i2

where Var(Ti ) is the variance of randomly distributed service times of server i and T̄i is the mean service times
by the server i (or reciprocal of service rate, µi ).
Figures 6(a) and (b) shows the difference in mean waiting times computed with and without exponential
service times assumptions. It was observed that the mean total waiting times was higher when service times
were assumed to follow some exponential distributions compared with when no such assumptions were made.
In many service processes, the exponential assumption of the distribution of service times usually results in
more conservative queueing system designs if the coefficient of variation of service times is less than unity.
Decisions based on mean waiting times and queue lengths predicted from such models would thus usually err
on the safe side.

5. CONCLUSIONS

Regardless of which industrial sector a BB is being employed, they need to adopt a system-level view of the
operations of an enterprise. The current BB training contents are no longer adequate for increasingly demanding
customers of the 21st century (i.e. versus 1980s and early 1990s when Six Sigma was first formulated).
A new breed of BBs will need to integrate OR/MS techniques into their Six Sigma tool set to remain
relevant. A new roadmap is formulated and presented in Table III to meet the emerging needs.
Not all of the OR/MS tools will be used in a project, but they serve as a reminder/checklist. In this way, a BB
can remain focused on the project while being alerted on other tools that may be useful in providing the solution.
It could be argued that a Six Sigma BB armed with OR/MS techniques would operate like a ‘Super Black Belt’,
with breath and depth well beyond what is found in the routine toolkits of BB coming from a regular Six Sigma
training conveyor belt.
In addition to OR/MS techniques, there is also an emerging trend of integrating artificial intelligence and
information systems technologies, such as data mining10, fuzzy logic and neural networks, into Six Sigma
programs; in particular, Design for Six Sigma for software. As the scope of Six Sigma application expands with
time, more cross-functional tools will be integrated with Six Sigma to achieve even wider and deeper business
performance improvement. The current integration of OR/MS tools is only part of the itinerary in the journey
towards Six Sigma excellence.

Copyright 
c 2007 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2007; 23:3–18
DOI: 10.1002/qre
16 L. C. TANG ET AL.

89.0

79.0

69.0

59.0
Waiting Times
49.0

39.0

29.0

19.0

9.0
8 9 10 11
Number of Packers
8 Dispensing Pharmacists (M/M/S) 8 Dispensing Pharmacists (M/G/S)
9 Dispensing Pharmacists (M/M/S) 9 Dispensing Pharmacists (M/G/S)
10 Dispensing Pharmacists (M/M/S) 10 Dispensing Pharmacists (M/G/S)
11 Dispensing Pharmacists (M/M/S) 11 Dispensing Pharmacists (M/G/S)

(a)
90.0

80.0

70.0
Waiting Times

60.0

50.0

40.0

30.0

20.0

10.0
8 9 10 11
Number of Dispensing Pharmacists
8 Packers (M/M/S) 8 Packers (M/G/S)
9 Packers (M/M/S) 9 Packers (M/G/S)
10 Packers (M/M/S) 10 Packers (M/G/S)
11 Packers (M/M/S) 11 Packers (M/G/S)

(b)

Figure 6. Comparisons of mean total waiting times computed with and without assumptions of exponential service times

REFERENCES
1. Goh TN. A strategic assessment of Six Sigma. Quality and Reliability Engineering International 2002; 18:403–410.
2. Montgomery DC. Editorial: Beyond Six Sigma. Quality and Reliability Engineering International 2001; 17(4):iii–iv.
3. Recker R, Bolstorff P. Integration of SCOR with Lean & Six Sigma. Supply-Chain Council, Advanced Integrated
Technologies Group, 2003.
4. Smith B. Lean and Six Sigma—a one–two punch. Quality Progress 2003; 36(4):37–41.
5. Bossert J. Lean and Six Sigma—synergy made in heaven. Quality Progress 2003; 36(7):31–32.
6. Murugappan M, Keeni G. Blending CMM and Six Sigma to meet business goals. IEEE Software 2003; 20(2):42–48.

Copyright 
c 2007 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2007; 23:3–18
DOI: 10.1002/qre
FORTIFICATION OF SIX SIGMA 17

7. Revere L, Black K. Integrating Six Sigma with Total Quality Management: A case example for measuring medication
errors. Journal of Healthcare Management 2003; 48(6):377–391.
8. Catherwood P. What’s different about Six Sigma? Manufacturing Engineer 2002; 81(8):186–189.
9. Dalgeish S. My ideal quality system. Quality 2003; 42(7):1.
10. Gurunatha T, Siegel RP. Applying quality tools to reliability: A 12 step Six Sigma process to accelerate reliability
growth in product design. Proceedings of the Annual Reliability and Maintainability Symposium. IEEE Press:
Piscataway, NJ, 2003; 562–567.
11. Wyper B, Harrison A. Deployment of Six Sigma methodology in human resource function: A case study. Total Quality
Management and Business Excellence 2000; 11(4–5):S720–S727.
12. Hahn GJ. Six Sigma: 20 key lessons learned. Quality and Reliability International 2005; 21:225–233.
13. Hillier FS, Lieberman GJ. Introduction to Operations Research (7th edn). McGraw-Hill: New York, 2001.
14. Harry MJ, Schroeder R. Six Sigma: The Breakthrough Management Strategy Revolutionizing the World’s Top
Corporations. Doubleday: New York, 2000.
15. Snee RD. Guest Editorial: Impact of Six Sigma on quality engineering. Quality Engineering 2000; 12(3):ix–xiv.
16. Hoerl RW. Six Sigma Black Belts: What do they need to know? (with discussion). Journal of Quality Technology 2001;
33(4):391–417.
17. Elshennawy AK. Quality in the new age and the body of knowledge for quality engineers. Total Quality Management
2004; 15(5–6):603–614.
18. Tang LC, Paoli PM. A spreadsheet-based multiple criteria optimization framework for Quality Function Deployment.
International Journal of Quality and Reliability Management 2004; 21(3):329–347.
19. Adams CW, Gupta P, Wilson CE. Six Sigma Deployment. Butterworth-Heinemann: Amsterdam, 2003.
20. Pintelon LM, Gelders LF. Maintenance management decision making. European Journal of Operational Research
1992; 58:301–317.
21. Jardine AKS (ed.). Operational Research in Maintenance. Manchester University Press: New York, 1970.
22. Schouten FA, Tapiero CS. OR models for maintenance management and quality control. European Journal of
Operational Research 1995; 82(5):211–213.
23. Scarf PA. On the application of mathematical models in maintenance. European Journal of Operational Research 1997;
99(3):493–506.
24. Goh TN. The role of statistical design of experiments in Six Sigma: Perspectives of a practitioner. Quality Engineering
2002; 14(4):659–671.
25. Zhang CW, Ong HL. Solving the biobjective zero–one knapsack problem by an efficient LP-based heuristic. European
Journal of Operational Research 2004; 159(3):545–557.
26. Tang LC, Xu K. A unified approach for dual response surface optimization. Journal of Quality Technology 2002;
34(4):437–447.
27. Lam SW, Tang LC. A graphical approach to the dual response robust design problem. Proceedings of the Annual
Reliability and Maintainability Symposium. IEEE Press: Piscataway, NJ, 2005; 200–206.
28. Harry M, Stewart R. Six Sigma Mechanical Design Tolerancing. Motorola University Press: Schaumburg, IL, 1988.
29. Harry MJ, Lawson JR. Six Sigma Producibility Analysis and Process Characterization. Addison-Wesley: Reading,
MA, 1992.
30. Tagaras G. Power approximations in the economic design of control chart. Naval Research Logistics 1989;
36:639–654.
31. Crowder SV. An SPC model for short production runs: Minimizing expected cost. Technometrics 1992; 34(1):64–73.
32. Rahim MA. Economic design of X̄ control charts assuming Weibull in-control times. Journal of Quality Technology
1993; 25(4):296–305.
33. Chung KJ. Algorithm for computing the economically optimal X̄ control chart for a process with multiple assignable
causes. European Journal of Operational Research 1994; 72(2):350–363.
34. McWilliams TP, Saniga EM, Davis DJ. Economic-statistical design of X̄ and R or X̄ and S charts. Journal of Quality
Technology 2001; 33(2):234–241.
35. Rohleder TR, Silver EA. Statistical process control subject to a labor resource constraint. International Journal of
Production Research 2002; 40(14):3337–3356.
36. Hahn GJ, Doganaksoy N, Hoerl R. The evolution of Six Sigma. Quality Engineering 2000; 12(3):317–326.
37. Bayle P, Farrington M, Sharp B, Hild C, Sanders D. Illustration of Six Sigma assistance on a design project. Quality
Engineering 2001; 13(3):341–348.
38. Taha HA. Operations Research: An Introduction (7th edn). Pearson Education International: Cambridge, 2003.
39. Prabhu NU. Queues and Inventories: A Study of their Basic Stochastic Processes. Wiley: New York, 1965.

Copyright 
c 2007 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2007; 23:3–18
DOI: 10.1002/qre
18 L. C. TANG ET AL.

40. Hahn GJ, Hill WJ, Hoerl RW, Zinkgraf SA. The impact of Six Sigma improvement—a glimpse into the future of
statistics. The American Statistician 1999; 53:1–8.
41. Ross SM. Introduction to Probability Models. Academic Press: New York, 2003.
42. Little J. A proof of a queueing formula. Operations Research 1961; 9(3):383–387.
43. Cosmetatos GP. Some approximate equilibrium results for multi-server queues (M/G/r). Operational Research
Quarterly 1976; 27(3):615–620.

Authors’ biographies
Loon Ching Tang, an Associate Professor in NUS, obtained his PhD from Cornell University in the field of
Operations Research with minors in Statistics. He has published numerous papers in more than 20 international
journals. In addition to being active in the forefront of research, he has been active in providing consulting
and training in the areas of Six Sigma, Design for Six Sigma, Statistical Methods, Operations Research
and Reliability Engineering. He is the Area Editor (Quality Engineering) of the International Journal of
Performability Engineering.
Cai Wen Zhang is a PhD candidate with the Department of Industrial and Systems Engineering. He is currently
working for Hitachi Global Storage Technologies (Singapore) Pte Ltd.
Thong Ngee Goh holds a PhD from the University of Wisconsin–Madison. He is an Academician of the
International Academy for Quality and Fellow of the American Society for Quality, as well as director of the
Quality and Innovation Research Centre, Industrial and Systems Engineering Department, National University
of Singapore. He is also recipient of the latest IEEE Engineering Management Society Educator of the Year
Award.
Shao Wei Lam is a Research Fellow with the Department of Industrial and Systems Engineering, National
University of Singapore. He holds a BEng in Mechanical Engineering and MEng in Industrial and Systems
Engineering. His research interests are in the field of quality-by-design, operations research in healthcare and
supply chain management. He has provided training and consultancy services in Design for Six Sigma.

Copyright 
c 2007 John Wiley & Sons, Ltd. Qual. Reliab. Engng. Int. 2007; 23:3–18
DOI: 10.1002/qre

You might also like