Int J Adv Manuf Technol (2006) 27: 1217–1222

DOI 10.1007/s00170-004-2314-1

ORIGINAL ARTICLE

Kun-Lin Hsieh · Lee-Ing Tong

Incorporating process capability index and quality loss function into analyzing
the process capability for qualitative data

Received: 13 January 2004 / Accepted: 22 June 2004 / Published online: 13 April 2005
© Springer-Verlag London Limited 2005

Abstract Process capability analysis (PCA) is frequently em- capability in most manufacturing industries. PCIs, such as Ca ,
ployed to evaluate a product or a process if it can meet the cus- C p and C pk are commonly used for most manufactures [3, 7–11],
tomer’s requirement. In general, process capability analysis can can frequently measure the process capability for the quantita-
be represented by using the process capability index (PCI). Until tive response. Herein, Ca evaluates the related scale of the process
now, the PCI was frequently used for processes with quantitative mean with the tolerance specification (i.e. the difference between
characteristics. However, for process quality with the qualitative the upper tolerance limit and the lower tolerance limit). C p eval-
characteristic, the data’s type and single specification caused lim- uates the related scale of the specification’s tolerance with pro-
itations of using the PCI. When the product can not meet the cess’s tolerance. While C pk simultaneously evaluates the center-
target, even if it lies in the specified range, it should lead to the ing degree and the dispersion degree. These PCIs will make some
corresponding quality loss. Taguchi developed a quadratic qual- adjustments if there are necessary particulars like the unilateral
ity loss function (QLF) to address such issues. In this study, we specification [10]. For the quantitative type, the theories on PCA
intend to construct a measurable index which incorporates the and PCIs are well developed [1, 8, 11]. The qualitative data type
PCI philosophy and QLF concept to analyze the process ca- may exist during the manufacturing environment, e.g. the inte-
pability with the consideration of the qualitative response data. grated circuit (IC) manufacturing industry uses the defect count
The manufacturers can not only employ the proposed index to on a wafer to analyze their product’s yield and control their pro-
self-assess the process capability, but they also can make com- cess, the process capability analysis for qualitative data will be an
parisons with the other competitors. important issue to study. However, most studies only focus on the
PCA application for the quantitative response data, and the qual-
Keywords Process capability analysis (PCA) · Process itative response data is seldom mentioned [6, 7]. Several difficul-
capability indexes (PCIs) · Qualitative data · Quality ties can be mentioned as: (1) the target of the qualitative data may
loss function (QLF). lead to unobvious centering evaluation, e.g. the target will be set as
zero defect, (2) the limitation of the unilateral specification, espe-
cially only the upper specification exist, e.g. the defect rate may be
1 Introduction less than 1% and (3) the quantitative data utilizes the process mean
(µ) and process deviation (σ) to compute the PCIs, however, the
Process capability analysis (PCA) [1, 4, 5, 11] is frequently em- qualitative data can not directly utilize them to compute the PCIs.
ployed by the manufacturers to evaluate if the capability of pro- Under the global market environment, to realize the pro-
cess can meet the customer’s requirement. Process capability in- cess capability comparison with other competitions can provide
dexes (PCIs) [1, 4] are a quantitative measurement of the process helpful information for enhancing organizational competence or
making strategic decisions. Especially, the PCA for the different
K.-L. Hsieh (u) manufacturers will be a significant factor to seek for the col-
Department of Information Management,
National Taitung University, laborators during the consideration of supply chain management
684 Chunghua Rd., Sec. 1, Taitung, Taiwan, R.O.C. (SCM). In this study, we intend to construct a process capability
E-mail: klhsieh2644@mail200.com.tw index, the PCI on qualitative response data, to evaluate the pro-
Tel.: +886-89-318855 ext. 2656
Fax: +886-89-321981
cess capability for the qualitative response data. The logical idea
is to combine the PCI philosophy and QLF concept. The rest of
L.-I. Tong this study is organized as follows. Sect. 2 clearly demonstrates
Department of Industrial Engineering and Management,
National Chaio Tung University, the construction procedure of the quantitative measurement we
HsinChu, Taiwan, R.O.C. proposed. Sect. 3 will employ the numerical examples to demon-
1218

strate the effectiveness of the proposed approach. Concluding where θ denotes the process parameter of the actual process and
remarks are finally made in Sect. 4. θc denotes the process parameter of the customer’s expectation.
In fact, the qualitative data can be described well by several
distributions like the binomial and Poisson distributions. In this
2 Construction of PCI for the qualitative response section, we will clearly describe the PCI value of both distri-
data butions, and the features hidden in the proposed PCI also are
explained in the next section.
2.1 Construction concept of PCI

The product can not meet the target, even if it lies in the speci- 2.2 Binomial distribution
fied range, and it should lead to quality loss. Taguchi developed
a quadratic quality loss function to address such a case [2]. The First, let y ∼ Ber(1, p), Ber(1, p) denotes the Bernoulli’s trial,
quadratic quality loss function is defined as follows:
θc = pc , θ = p, µ = p, σ 2 = p(1 − p),
L(y) = k(y − T) 2
(1)
hence,
and the expected quality loss can be described as:
QL(θc ) K [ p2c + ( pc − p2c )] pc
PCI = = = (5)
QL = E[L(y)] = E[k (y − T) ] 2
QL(θ) k[ p + ( p − p )]
2 2 p
= kE[(y − µ + µ − T)2 ] = k((µ − T)2 + σ 2 ) (2) where p denotes the non-conforming rate (the parameter of the
binomial distribution), pc denotes the acceptable quality level of
where y denotes the response data, T denotes the target value
the customer for the non-conforming rate and n denotes inspec-
or the nominal value, k denotes the constant of the quality loss
tion count.
when the process is within the allowable tolerance, µ denotes the
The features include:
process mean and σ denotes the process deviation.
For the qualitative response data set, the target situation 1. When pc < p, it means the capability of the process can not
should be zero defect (T = 0) or not non-conforming. Hence, ap- meet the customer’s requirement; that is, it is a “bad” pro-
plying it into the quality loss function, Eq. 2 can be modified cess, PCI < 1.
as: 2. When pc = p, it means that the capability of the process ex-
actly meets the customer’s requirement, PCI = 1.
QL = k(µ2 + σ 2 ) (3) 3. When pc > p, it means that the capability of the process ab-
solutely satisfies the customer’s requirement, PCI > 1.
Then, the quality loss function can be represented as QL =
4. Figure 1 graphically depicts the relationship between process
QL(θ) if the process parameter θ is involved. For the qualita-
parameter p and the PCI value with different pc .
tive data, according to the concept of the quality loss function,
the quantitative measurement of the process capability can be Then, let y ∼ Ber(1, p), and if the lot count is n, the related
constructed. That is, we can take the ratio of the customer’s parameters can be denoted as:
allowable quality loss and the actual quality loss. Hence, the gen-

n
eralized PCI of the attribute data can be defined as: D= yi ∼ B(n, p),
i.i.d.
QL(θc ) i=1
PCI = (4)
QL(θ) θc = pc , θ = p, µ = np, σ 2 = np(1 − p)

Fig. 1. The diagram for Ber-

noulli distribution with dif-
ferent pc , p and PCI
 1219

where B(n, p) will denote the binominal distribution. Hence, the the lot count n; the pc = p will be the saddle point in such
PCI formula for binominal distribution can be represented as cases. Furthermore, from Fig. 2a and b, we can clearly realize
follows: the change of the PCI curve when the customer’s allowable
parameter changes.
QL(θc ) K [(n pc )2 + n( pc − p2c )] [(n − 1) p2c + pc ]
PCI = = =
QL(θ) k[(np)2 + n( p − p2 )] [(n − 1) p2 + p]
2.2.1 Poisson distribution
(6)
Let y ∼ P(λ), herein, P denotes the Poisson distribution, and
Features:
the related distribution’s parameters can be represented as: θc =
1. When pc < p, it means the capability of the process can not λc , θ = λ, µ = λ, σ 2 = λ,
meet the customer’s requirement; the PCI will decrease with
respect to the inspection count n. QL(θc ) K [λ2c + λc ] [λ2c + λc ]
PCI = = = 2 (7)
2. When pc = p, it means that the capability of the process ex- QL(θ) k[λ2 + λ] [λ + λ]
actly meets the customer’s requirement.
3. When pc > p, it means that the capability of the process ab- where λ denotes the defect rate of the Poisson distribution for the
solutely satisfies the customer’s requirement; the PCI will actual process and λc denotes the acceptable quality level of the
increase with respect to the lot count n. customer for the defect rate of the Poisson distribution.
4. From Fig. 2a, when the process capability can meet the cus- The features include:
tomer’s requirement, the PCI will increase with respect to the 1. When λc < λ, it means the process capability can not meet
lot count n; while the process capability can not meet the cus- the customer’s requirement; that is, it is a “bad” process, that
tomer’s requirement, the PCI will decrease with respect to is, PCI < 1.

Fig. 2. a The PCI curve dia-

gram for different inspection count
when pc = 0.01 b The PCI curve
diagram for different inspection
count when pc = 0.1
1220

2. When λc = λ, it means that the process capability exactly 3. PCIcompetitor: manufacturer > 1, it means the manufacturer’s
meets the customer’s requirement; that is, PCI = 1. process capability is better than the competitor’s process ca-
3. When λc > λ, it means that the process capability absolutely pability.
satisfies the customer’s requirement; that is, PCI > 1.
4. Figure 3 depicts graphically the relationship between the pro-
cess parameter λand PCI with different λc . 3 Numerical analysis and the conclusions
If the inspection count be m and it is denoted as the case of
the average defect rate, y ∼ P(λ), P denotes the Poisson distri- 3.1 Illustrative example 1
bution, and the related parameter can be denoted as: θc = λc , θ =
λ, µ = mλ , σ 2 = mλ2 , and the PCI formula is: A lead frame manufacturer in Taiwan expects to realize if their
process capability can meet the customer’s requirement (the
λ2
QL(θc ) K [ mc + λmc ] [λ2c + λc ] packaging fabrication). Several hundreds of lead frame types are
PCI = = = 2 (8) produced in lead frame manufacturing. The packaging fabrica-
QL(θ) k[ λm + mλ ]
2
[λ + λ]
tions expect the defect count of the lead frame in their in-line
where µ denotes the average defect rate and µc denotes the ac- quality control (IQC) must be less than ten strips per 500 in-
ceptable quality level of the customer for the average defect rate. spection strips, and then the yield of the packaging product can
We can find out that it has the same structure as the PCI of the be enhanced. Hence, the lead frame manufacturer plans to study
Poisson distribution. Hence, the features are also the same as their process capability and make a suitable compromise with the
Poisson’s. customers. The following data listed in Table 1 are collected for

2.3 Construction of PCI for comparison

with different competitors Table 1. The non-conforming count
We will make some integration according to the qualitative PCI Day Non-conforming Day Non-conforming
we previously proposed for the case of comparison between count per 500 strips count per 500 strips
different competitors. The concept is to substitute the manufac-
turer’s quality loss by the competitors’ quality loss. Hence, the 1 7 16 10
constructed comparison PCI is given as follows: 2 5 17 7
3 13 18 9
QL(θcompetitor ) 4 11 19 14
PCI competitor: manufacturer = (9) 5 12 20 12
QL(θmanufacturer) 6 9 21 11
7 10 22 8
Features: 8 14 23 9
1. PCIcompetitor: manufacturer < 1, it means the competitor’s pro- 9 10 24 12
cess capability is better than the manufacturer’s process ca- 10 6 25 8
11 13 26 10
pability. 12 9 27 9
2. PCIcompetitor: manufacturer = 1, it means the competitor’s pro- 13 12 28 7
cess capability is equal to the manufacturer’s process capabil- 14 8 29 8
15 12 30 10
ity.

Fig. 3. The PCI curve dia-

gram for the process parame-
ter λc , λ and PCI
 1221

30 lots as the same QFP lead frame, and the constant k of the 3.2 Numerical example 2: Comparison
manufacturer is 180 (that is the necessary cost for rework). The with different competitors
quality (result) of the product can be divided into two categories:
conforming and non-conforming. Each inspection is independent For the same numerical example, besides, the manufacturer also
of the others. About 30 records were collected in Table 1. Obvi- collected the related competitor’s information. They also ex-
ously, the data obey the binomial distribution. pect to realize the difference between their process capability
First, we must make sure that the process is in-control. The and competitors’ process capability. The collected information is
data type is owing to the qualitative type. Hence, the np-chart is given as follows.
applied to process control (see Fig. 4). The total non-conforming For competitor A: the parameter θa (the non-conforming rate
count can be computed as 295. Hence, we can use the following pa ) is 0.015, the inspection count is 600 strips and the constant k
formula to estimate the non-conforming rate p. is 200.
For competitor B: the parameter θb (the non-conforming rate
total number of non-conforming units 295
p̄ = = = 0.0197 pb ) is 0.025, the inspection count is 400 strips and the constant k
total number inspected 30 × 500 is 250.
Then, the np-chart is constructed as follows. For manufacturer: the parameter θb (the non-conforming rate
 pb ) is 0.0197, the inspection count is 500 strips and the constant
UCL = n p̄ + 3 n p̄(1 − p̄) = 19.17 k is 180.
CL = n p̄ = 9.85 Compare with competitor A:

LCL = n p̄ − 3 n p̄(1 − p̄) = 0.53.
QL(θ A )
PCI A:M =
It is in-control for screening out the np-chart. Then, we will QL(θ M )
employ the proposed PCI formula to study the process capabil- k A [(n A p A )2 + n A ( p A − p2A )]
=
ity of the lead frame manufacturer. As the customer’s require- k M [(n M p M )2 + n M ( p M − p2M )]
ment, the parameter θc ( pc ) is 0.02, and the manufacture’s es-
200[(0.015 × 600) 2 + 600 × (0.015 − 0.015 2 )]
timated parameter θ ( p) is 0.0197. When the inspection count =
is 500, the average count of the non-conforming unit is 9.85 180[(0.0197 × 500) 2 + 500 × (0.0197 − 0.0197 2 )]
(0.0197 × 500). Then, the proposed PCI value can be computed 17973
= = 0.936 < 1
as follows: 19202.22

QL(θc ) K [(n pc )2 + n( pc − p2c )] Conclusion: PCI A:M < 1, it means the competitor’s process
PCI = =
QL(θ) k[(np)2 + n( p − p2 )] capability is better than the manufacturers’ process capability.
[(n − 1) p2c + pc ] 499 × 0.022 + 0.02 That is, competitor A’s quality loss is less than the manufactur-
= = = 1.0393. er’s quality loss.
[(n − 1) p + p]
2 499 × 0.0197 2 + 0.0197
Compare with competitor B:
We can find out that the PCI value exceeds 1, and it means
the current process capability can meet the customer’s require- QL(θ B )
ment. However, the ratio is not significantly larger than 1, so the PCI B:M =
QL(θ M )
lead frame manufacturer still need to pay more attention to their
process. k B [(n B p B )2 + n B ( p B − p2B )]
=
k M [(n M p M )2 + n M ( p M − p2M )]
250[(0.025 × 400) 2 + 400 × (0.025 − 0.025 2 )]
=
18[(0.0197 × 500) 2 + 500 × (0.0197 − 0.0197 2 )]
27437.5
= = 1.429 > 1
19202.22

Conclusion: PCI B:M > 1, it means the competitor’s process

capability is worse than the manufacturer’s process capability.
That is, competitor B’s quality loss is larger than the manufactur-
ers’ quality loss.
According to the comparison, we can make the conclusion:
“Competitor B’s process capability is worse than the manufac-
turer, while competitor A’s process capability is significantly bet-
ter than the manufacturer. Furthermore, the sequence of process
capability from the best to the worst is Competitor A → Man-
Fig. 4. The constructed NP-chart ufacturer → Competitor B”. Restated, the manufacturer should
1222

work hard to enhance their process capability by performing the

necessary quality improvement. References
1. Montgomery DC (2001) Introduction to statistical quality control, 4th
edn. Wiley, New York, NY
2. Phadke MS (1989) Quality engineering using robust design. Prentice-
4 Concluding and remarks Hall International, New Jersey, NJ
3. Chen KS, Pearn WL (1997) An application of non-normal process ca-
pability indices. Qual Reliab Eng Int 13:355–360
In this study, we construct a quantitative measurement PCI for 4. Kane VE (1986) Process capability indices. J Qual Technol 18:41–52
the qualitative response. The quantitative measurement is based 5. Bissell AF (1990) How reliable is your capability index? Appl Stat
on the Taguchi’s quality loss function philosophy and PCI con- 39:331–340
cept. It is a ratio deriving from the customer’s quality loss with 6. Kaminsky FC, Davis RD (1989) Statistical measures of process capa-
bility and their relationship to non-conforming product. In: Proc Third
respect to the actual process’s quality loss. By employing the Biennial International Manufacturing Research Forum
proposed PCI, the manufacturers can employ it to assess if 7. Kaminsky FC, Dovich RA, Burke RJ ( 1998)Process capability indices:
the process capability can meet the customer’s requirement. Be- now and in the future. Qual Eng 10:445–453
sides, the constructed PCI can also be employed to make the 8. Kureková K (2001) Measurement process capability – trends and ap-
proaches. Meas Sci Rev 1(1):43–46
comparison between the manufacturer and the competitors. The 9. Tong LI, Chenn KS, Chen HT (2001) Statistical testing for assessing
PCI formulas for different quality data obeying the binomial the performance of lifetime index of elcetronic component with expo-
distribution or Poisson distribution are proposed in this study. nential distribution. Int J Qual Reliab Manage 19(6/7):812–824
10. Chen JP, Ding CG (2001) A new process capability index for non-
The other advantage is that the practitioners do not need com-
normal distribution. Int J Qual Reliab Manage 18(7):762–770
plicated computation to obtain the attribute PCIs by using the 11. Kotz S, Lovelace CL (1998) Process capability indices in theory and
proposed PCIs. practice. Arnold, London