Hindawi

Mathematical Problems in Engineering

Volume 2020, Article ID 5682587, 26 pages
https://doi.org/10.1155/2020/5682587

Research Article
A Control Chart for Exponentially Distributed Characteristics
Using Modified Multiple Dependent State Sampling

Ahmad O. Albazli , Muhammad Aslam , and Saeed A. Dobbah

Department of Statistics, Faculty of Science, King Abdulaziz University, Jeddah 21551, Saudi Arabia

Correspondence should be addressed to Muhammad Aslam; aslam_ravian@hotmail.com

Received 13 April 2020; Revised 12 November 2020; Accepted 5 December 2020; Published 28 December 2020

Academic Editor: Emilio Insfran Pelozo

Copyright © 2020 Ahmad O. Albazli et al. This is an open access article distributed under the Creative Commons Attribution
License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is
properly cited.
In this paper, a t-control chart based on modified multiple dependent state sampling is proposed for monitoring processes that
assume time between events following exponential distribution. The chart has double control limits and employs information
from a previous sample and the current sample. The control chart coefficient “constants” are estimated by considering different
values of the in-control average run lengths. The detection ability of the proposed control chart is found to be better than that of
control charts based on multiple dependent state sampling in terms of average run lengths and the standard deviation of run
lengths and better than generalized multiple dependent state sampling in terms of average run lengths. Case studies with real data
are included as illustrative examples for the implementation of the proposed chart.

1. Introduction is a common practice for monitoring quality and detecting

assignable causes of process variation that may cause out-of-
In this era of business globalization, organizations tend to control conditions [5].
adjust their strategies and tendencies to allow them to Shewhart’s control charts are efficient at detecting large
succeed and remain ahead of rivals and to expand into new shifts in a process [6], but they are not sufficiently sensitive to
markets. Thus, ensuring the effectiveness and efficiency of detecting small process shifts. Accordingly, the probability
the products or services they provide should boost customer of declaring a process out-of-control while it is actually in-
satisfaction and build loyalty. In addition, implementing control (type I error or false alarm) using Shewhart’s control
policies to monitor and control business processes and charts is high [7]. In addition, a Shewhart control chart based
procedures, as a function of management, should help or- on a single sampling scheme assumes a large average run
ganizations augment their profits and remain economically length (ARL) and involves a delay in indicating when a
sustainable [1]. Accordingly, monitoring and controlling process will be out-of-control. Therefore, it is possible to
processes are among the primary tasks that industries need enhance the performance of Shewhart’s control charts by
to embark on [2]. Furthermore, the importance of the using different and well-structured sampling schemes and
control function is not limited to the development and implementing supplementary run rules. For instance, Al-
improvement of quality; rather, it also provides important Marshadi et al. [8] proposed a control chart using a repetitive
information that informs decision-making processes and sampling scheme for a service performance index based on
aids in the application of sufficient corrective plans [3]. the proportion of customer complaints and proved that the
Statistical process control (SPC) techniques and tools have, proposed control chart is more efficient than single sam-
for decades, been applied to processes in a variety of fields to pling. Abujiya and Lee [9] evaluated the performance of
facilitate the creation of a decision-making road map [4]. In Shewhart’s X chart using ranked set sampling (RSS) in
particular, employing control charts, which were first de- comparison with traditional simple random sampling and
veloped and introduced in 1924 by Dr. Walter A. Shewhart, concluded that the control chart based on RSS is more
2 Mathematical Problems in Engineering

robust. The performance of a combined Shewhart expo- control if all specified subgroups are in a controlled state,
nential weighted moving average control chart based on while, in MMDS, the process is declared in-control state
double median RSS is analyzed by Abujiya et al. [10]. The even though one subgroup from the specified subgroups is in
results of the study show that the detection performance is an in-decision area. Therefore, the MMDS is more flexible
enhanced when a sampling method on combined schemes is than the MDS in making decisions about the state of the
used. Tran [11, 12] showed that applying run rules control process.
charts could notably improve the performance of Shewhart’s By exploring the literature and to the best of our
control charts. knowledge, there is no work on designing t-charts using
Usually, the assumption when designing any control MMDS for skewed process output. In this paper, MMDS is
chart for process monitoring is that real data values are at used to develop a new t-chart to monitor skewed variables,
least approximately normally distributed [13]. However, in assuming that the TBE follows an exponential distribution.
many real-life situations, this assumption is invalid and the In terms of ARL and the standard deviation of run length
process output is skewed. In this case, using a Shewhart (SDRL), a performance comparison between the control
control chart based on an assumption of normality can lead charts using MMDS and MDS is conducted. Additionally, a
to fallacious results. Therefore, when the assumption of performance comparison between the control charts using
normality is not met in the case of skewed data, additional MMDS and GMDS in terms of ARL is conducted. It is
steps may be necessary to adjust the data so that this ex- expected that the control chart using MMDS would be the
pectation is justified. For instance, Santiago and Smith [14] most efficient in terms of ARL and SDRL among the charts
introduced a control chart to monitor the time interval included in this study.
between events, known as the time-between-events (TBE) The rest of the paper is organized as follows: the
chart or t-control chart, to be used when data follows the design of the proposed control chart is presented in
exponential distribution. They used the variable transfor- Section 2. Section 3 discusses the efficiency of the pro-
mation provided by Nelson [15] to transform the data to be posed control chart when the data are skewed and follow
approximately normal. an exponential distribution. The performance of the
The multiple dependent state (MDS) sampling scheme is proposed chart is evaluated compared to other existing t-
proposed by Wortham and Baker [16]. In the inspection charts in Section 4. Real examples from industry are
procedure of the MDS plan, decision criterion is formed to presented in Section 5, and some concluding remarks are
accept the sample drawn from a lot, reject it, or accept or given in the last section.
reject the lot depending on the sample itself in combination
with the states of the preceding lots. According to Balamurali
and Jun [17], it is recommended that between two and five 2. The Proposed t-Chart Using the
preceding subgroups be used. The number of these sub- MMDS Design
groups (say, m) is set in advance by the industrial engineers/
quality control inspectors. The smaller values of m are set for Suppose that T is a random variable with an exponential
the strict inspection of the process. distribution that represents a quality characteristic of in-
Various studies were carried out on control charts terest and has the following probability density function:
under MDS. Based on this sampling plan, Aslam et al. 1
[18, 19] proposed a control chart for the exponential f(t) � e−t/θ , t > 0, (1)
θ
distribution and gamma distribution. Additionally, Yan
et al. [20] developed an MDS sampling scheme based on a where θ is the scale and shape parameter of the exponential
coefficient of variation of normally distributed quality distribution, since it is equal to the mean and standard
characteristics and showed the superiority of the proposed deviation. Let θ0 be the mean when the process is in-control.
plan over the single sampling plan. The form of MDS is According to Johnson and Kotz [23], if T has an ex-
then generalized by Aslam et al. [21, 22] to generalize ponential distribution with a scale parameter θ, the trans-
multiple dependent state sampling (GMDS) for the ap- formed random variable T∗ � T1/β follows the Weibull
plication of variable and attribute quality characteristics. It distribution with shape parameter β and scale parameter
is proven that introducing an additional parameter to MDS θ1/β . Nelson [15] provided transformation techniques to
to form GMDS increases the monitoring sensitivity convert exponential data to Weibull distribution with a
significantly. shape parameter based on this fact. Nelson [15] suggested
As pointed out earlier, MDS and GMDS are conditional that β � 3.6 makes an approximately normal curve; there-
sampling schemes that use information from the current and fore, symmetric control limits can be implemented when
previous subgroups to make decisions about the state of the proposing a control chart for nonconforming items. Thus, it
process. Similarly, the modified multiple dependent state may be more convenient to monitor the process using this
(MMDS) sampling scheme uses information from the transformed variable.
current and previous subgroups to make decisions about the In this section, based on the statistic T∗ , we present the
state of the process in a more flexible manner, as illustrated following t-chart using the MMDS for individual mea-
in the next section. MMDS is developed to enhance the surements (sample size � 1) considering two pairs of
detection ability of process shifts. The difference between control limits: a pair of outer control limits, UCL1 and
MMDS and MDS is that, in MDS, the process is declared in- LCL1 , and a pair of inner control limits, UCL2 and LCL2 .
Mathematical Problems in Engineering 3

(i) Step 1: When the process is in-control, we have the mean E(T∗ )
Measure the quality characteristic T of a randomly and the second moment E[(T∗ )2 ], for statistic T∗ as follows:
selected sample of the production process. Then, 1
calculate T∗ : μT∗ � E T∗ 􏼁 � θ∗0 Γ􏼠1 + 􏼡,
β
(3)
∗ 2 2 2
T∗ � T1/3.6 . (2) E􏽨 T 􏼁 􏽩 � 􏼂θ∗0 􏼃 Γ􏼠1 + 􏼡,
β

(ii) Step 2: where θ∗0 � θ1/3.6

0 and Γ (.) is a gamma function. Based on the
Declare the process to be in-control if above equations, we have the standard deviation of the
LCL2 ≤ T∗ ≤ UCL2 . statistic T∗ as follows:
􏽶�������������������
Declare the process to be out-of-control if 􏽴
2
T∗ > UCL1 or T ∗ < LCL1 . ∗ 2 1
σ T∗ � θ0 Γ􏼠1 + 􏼡 − Γ􏼠1 + 􏼡 . (4)
Otherwise, go to Step 3. β β
(iii) Step 3:
If T has an exponential distribution with the scale pa-
Declare the process to be in-control if there are at rameter θ, then T∗ follows a normal distribution with
least (m − 1) out of the m proceeding subgroups θ∗0 � θ1/3.6 .
0
declared as in-control, and one m may be in an in- The outer control limits for the proposed chart are given
decision area (i.e., UCL2 < T∗ ≤ UCL1 , or by
LCL1 < T∗ ≤ LCL2

UCL1 � μT∗ + k1 σ T∗
􏽶�������������������
􏽴
2
1 2 1
� θ∗0 Γ􏼠1 + 􏼡 + k1 θ∗0 Γ􏼠1 + 􏼡 − Γ􏼠1 + 􏼡
β β β

􏽶�������������������
􏽴
2
∗⎢
⎢
⎡
⎢ 1 2 1 ⎤⎥⎥⎥
⎢
⎢
� θ0 ⎢ ⎥⎥⎥
⎣ Γ􏼠1 + β􏼡 + k1 Γ􏼠1 + β􏼡 − Γ􏼠1 + β􏼡 ⎦

� θ∗0 CU1 ,
(5)
LCL1 � μT∗ − k1 σ T∗
􏽶�������������������
􏽴
2
1 2 1
� θ∗0 Γ􏼠1 + 􏼡 − k1 θ∗0 Γ􏼠1 + 􏼡 − Γ􏼠1 + 􏼡
β β β

􏽶�������������������
􏽴
2
⎢
⎡
⎢ 1 2 1 ⎤⎥⎥⎥
� θ∗0 ⎢
⎢
⎢
⎢
⎣ Γ􏼠 1 + 􏼡 − k 1 Γ􏼠 1 + 􏼡 − Γ􏼠 1 + 􏼡 ⎥⎥⎥
⎦
β β β

� θ∗0 CL1 .
4 Mathematical Problems in Engineering

The inner control limits for the proposed chart are given LCL2 � μT∗ − k2 σ T∗
by
􏽶�������������������
􏽴
UCL2 � μT∗ + k2 σ T∗ , 2 (9)
1 2 1
􏽶������������������� � θ∗0 Γ􏼠1 + 􏼡 − k2 θ∗0 Γ􏼠1 + 􏼡 − Γ􏼠1 + 􏼡 ,
􏽴 β β β
2
∗ 1 ∗ 2 1
� θ0 Γ􏼠1 + 􏼡 + k2 θ0 Γ􏼠1 + 􏼡 − Γ􏼠1 + 􏼡 , 􏽶�������������������
β β β 􏽴
2
⎢
⎡
⎢ 1 2 1 ⎤⎥⎥⎥
(6) � θ∗0 ⎢⎢⎢⎢⎣ Γ􏼠1 + 􏼡 − k2 Γ􏼠1 + 􏼡 − Γ􏼠1 + 􏼡 ⎥⎥⎥,
⎦ (10)
β β β
􏽶�������������������
􏽴
2
⎢
⎡
⎢ 1 2 1 ⎤⎥⎥⎥
� θ∗0 ⎢
⎢
⎢
⎢
⎣ Γ 􏼠1 + 􏼡 + k 2 Γ􏼠 1 + 􏼡 − Γ􏼠 1 + 􏼡 ⎥⎥⎥,
⎦
� θ∗0 CL2 . (11)
β β β
In the above, k1 and k2 (k1 > k2 ) are control coefficients
(7) to be determined by considering the target in-control ARL,
say r0 .
� θ0∗ CU2 , (8) The probability of being declared in-control for the
proposed control chart, say P0in , when the process is actually
in an in-control state (θ � θ0 ), is given as follows:

P0in � P LCL2 ≤ T∗ ≤ UCL2 􏼁 + 􏼈P UCL2 ≤ T∗ ≤ UCL1 􏼁 + P LCL1 ≤ T∗ ≤ LCL2 􏼁􏼉

∗ m ∗ ∗
􏼈P LCL2 ≤ T ≤ UCL2 􏼁􏼉 + m􏼈P UCL2 ≤ T ≤ UCL1 􏼁 + P LCL1 ≤ T ≤ LCL2 􏼁􏼉 (12)
⎢
·⎡
⎣ ⎤⎥⎦,
∗ m− 1
P LCL2 ≤ T ≤ UCL2 􏼁

where

β β
P LCL2 ≤ T∗ ≤ UCL2 􏼁 � exp􏽮−CL2 􏽯 − exp􏽮−CU2 􏽯,
β β
P UCL2 ≤ T∗ ≤ UCL1 􏼁 � exp􏽮−CU2 􏽯 � exp􏽮−CU1 􏽯, (13)
β β
P LCL1 ≤ T ∗ ≤ LCL2 􏼁 � exp􏽮−CL1 􏽯 − exp􏽮−CL2 􏽯.

In this case, (8) can be written as follows:

β β β β β β
P0in � exp􏽮−CL2 􏽯 − exp􏽮−CU2 􏽯 + 􏽮exp􏽮−CU2 􏽯 − exp􏽮−CU1 􏽯 + exp􏽮−CL1 􏽯 − exp􏽮−CL2 􏽯􏽯
β β m β β β β β β m− 1
· 􏼔􏽮exp􏽮−CL2 􏽯 − exp􏽮−CU2 􏽯􏽯 + m􏽮exp􏽮−CU2 􏽯 − exp􏽮−CU1 􏽯 + exp􏽮−CL1 􏽯 − exp􏽮−CL2 􏽯􏽯􏼐exp􏽮−CL2 􏽯 − exp􏽮−CU2 􏽯􏼑 􏼕.
(14)
Mathematical Problems in Engineering 5

The in-control ARL and SDRL are defined as Now, suppose that the process mean is shifted “equal to
1 θ1 .” Then, the probability of declaring the process as out-of-
ARL0 � , control for the proposed control chart, say P1in , is given as
1 − P0in follows:
􏽶���������
􏽴 (15)
P0in
SDRL0 � .
0 2
􏽨1 − Pin 􏽩

P1in � P LCL2 ≤ T∗ ≤ UCL2 | θ � θ1 􏼁 + 􏼈P UCL2 ≤ T∗ ≤ UCL1 |θ � θ1 􏼁 + P LCL1 ≤ T∗ ≤ LCL2 |θ � θ1 􏼁􏼉

m
· 􏼂􏼈P LCL2 ≤ T∗ ≤ UCL2 |θ � θ1 􏼁􏼉 + m􏼈P UCL2 ≤ T∗ ≤ UCL1 |θ � θ1 􏼁 + P LCL1 ≤ T∗ (16)
∗ m− 1
≤ LCL2 |θ � θ1 􏼁􏼉P LCL2 ≤ T ≤ UCL2 |θ � θ1 􏼁 􏽩.

The ARL and SDRL for the shifted process are defined as exact statistics of the proposed process are not available.
1 Thus, this approach is performed in this research to compare
ARL1 � , the detection sensitivity of the aforementioned t-charts.
1 − P1in
􏽶���������
􏽴 (17)
P1in 4.1. Comparison of the Proposed Chart with the MDS Chart.
SDRL1 � 2 . A simulation approach is adopted in this study to compare
􏽨1 − P1in 􏽩
the proposed chart and the t-chart under MDS sampling
(available here). First, an exponential distribution of 1000
observations is generated. Then, the limit coefficients under
3. Performance Evaluation of the MMDS and MDS are computed. Choosing r0 � 200, 300, 370,
Proposed Chart and 500 and m � 2 to 6, the ARL and SDRL values of the t-chart
under both schemes are compared for different shift ratios of
To measure the performance of any control chart, ARL is θ0 /θ1 , as shown in Tables 5–8. From Figures 1–8, it is observed
used as a sole measure or is combined with other measures that the proposed chart detects process shifts more quickly; for
[24]. The ARL is the number of in-control observations, on example, when r0 � 370 and m � 6, it took the chart almost 221
average, before a change in process level or an out-of-control samples to detect the out-of-control signal after the occurrence
observation is indicated [25]. of the shift of “θ0 /θ1 � 0.9” in the process mean, while, under
Usually, along with the ARL, the SDRL is also computed. the MDS scheme, it required 230 samples to detect the shift.
When the process behaves consistently over time, greater Similarly, the SDRL of the proposed chart required 220.43
values of ARL and SDRL are expected, and the process is samples, compared to the 230.79 samples for the t-chart under
considered to be statistically in-control. By contrast, smaller the MDS scheme. In the same manner, when the shift in the
values of ARL and SDRL are desired when the process shifts process mean increases, the number of samples needed to
or is declared out-of-control. detect the shift in the proposed chart becomes smaller than in
Using the R programming language and software for sta- the MDS sampling scheme. To illustrate, the control charts
tistical computation and simulation [26], the values of k1 and k2 under the MMDS and MDS are constructed and shown in
and the values of the ARL and SDRL for the in-control and out- Figures 9 and 10. It is evident that observation 36 falls in the in-
of-control processes are determined and presented in Tables 1–4 decision area with the t-chart under the MDS, whereas it is
for r0 � 200, 300, 370, and 500, respectively. It is noted that the considered out-of-control under the MMDS.
values of the ARL and SDRL remain as targeted when the
process is in-control (the ratio θ0 /θ1 is equal to 1); however, the
values of the out-of-control ARL and SDRL decrease as the shift 4.2. Comparison of the Proposed Chart with the GMDS Chart.
in the process mean increases (the ratio θ0 /θ1 decreases). GMDS sampling is a generalized form of the MDS sampling
scheme that is more efficient at making decisions about the
4. Performance Comparison state of the process. Like MDS, GMDS relies on information
about current and previous observations. The performance
In this section, the efficiency of the proposed t-chart under evaluation of GMDS is more efficient than that of MDS.
MMDS sampling is evaluated compared to the t-chart under When compared with the GMDS results presented by Aslam
the MDS and GMDS sampling schemes. In many studies, a et al. [21], the performance of the proposed chart when using
simulation approach is extensively employed to demonstrate simulated data is investigated in this section when r0 � 200,
the performance and effectiveness of control charts [27–31]. 300, and 370, as shown in Tables 9–11. The decreasing
Using simulated data is convenient and is usually done when pattern of ARL demonstrates that the performance of the
 6

Table 1: The ARL and SDRL values of the proposed chart at r0 � 200 with the m value varying from 2 to 10.
m 2 3 4 5 6 7 8 9 10
k1 2.65298 2.83972 3.01942 2.98755 3.11139 3.12916 2.76820 3.16861 2.71363
k2 1.80389 1.64421 1.68708 1.76101 1.79793 1.83897 1.97868 1.90199 2.09203
θ0 /θ1 ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL
1 199.99 199.49 200.00 199.50 200.01 199.51 200.00 199.50 200.00 199.50 200.01 199.51 200.01 199.51 200.02 199.52 200.02 199.51
0.90 133.16 132.66 130.57 130.07 132.51 132.01 129.58 129.08 130.73 130.23 129.59 129.09 124.71 124.21 127.92 127.42 125.92 125.42
0.80 81.26 80.76 77.72 77.22 79.24 78.74 75.79 75.29 76.56 76.06 75.08 74.58 70.76 70.26 72.90 72.40 72.10 71.60
0.70 46.69 46.19 43.26 42.76 43.76 43.26 41.09 40.59 41.24 40.74 40.01 39.51 37.75 37.25 38.23 37.73 38.80 38.29
0.60 25.84 25.34 23.10 22.59 22.96 22.45 21.31 20.80 21.14 20.64 20.36 19.85 19.59 19.08 19.25 18.75 20.30 19.79
0.50 14.01 13.50 12.11 11.60 11.79 11.28 10.93 10.41 10.72 10.21 10.32 9.80 10.24 9.73 9.77 9.26 10.71 10.20
0.40 7.56 7.05 6.40 5.88 6.13 5.61 5.74 5.22 5.62 5.09 5.45 4.92 5.60 5.08 5.25 4.73 5.92 5.40
0.30 4.14 3.60 3.51 2.96 3.36 2.81 3.22 2.68 3.17 2.63 3.14 2.59 3.32 2.77 3.12 2.57 3.54 3.00
0.20 2.36 1.79 2.07 1.49 2.02 1.43 2.00 1.42 2.00 1.42 2.02 1.43 2.15 1.57 2.06 1.47 2.27 1.70
0.10 1.45 0.81 1.37 0.71 1.37 0.71 1.38 0.73 1.39 0.74 1.41 0.76 1.46 0.82 1.43 0.79 1.51 0.87
Mathematical Problems in Engineering
 Mathematical Problems in Engineering

Table 2: The ARL and SDRL values of the proposed chart at r0 � 300 with the m value varying from 2 to 10.
m 2 3 4 5 6 7 8 9 10
k1 2.80952 2.71631 2.99340 2.89292 2.95491 3.08240832 2.95759 2.86055 2.94165
k2 1.64840 2.18966 1.76962 1.87672 1.89838 1.90957829 1.97186 2.04903 2.03151
θ0 /θ1 ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL
1 300.00 299.50 300.00 299.50 299.98 299.48 300.02 299.51 300.01 299.51 300.03 299.53 300.02 299.52 300.01 299.51 300.00 299.50
0.90 187.23 186.73 191.28 190.78 186.86 186.36 181.02 180.51 180.56 180.06 183.17 182.67 177.47 176.96 175.92 175.42 174.86 174.36
0.80 107.38 106.88 112.26 111.76 105.23 104.73 99.49 98.99 98.42 97.92 100.01 99.51 94.97 94.47 93.99 93.48 92.32 91.81
0.70 58.18 57.68 62.35 61.85 55.09 54.59 51.36 50.85 50.16 49.66 50.49 49.99 47.61 47.11 47.33 46.83 45.80 45.30
0.60 30.45 29.94 33.45 32.94 27.60 27.10 25.62 25.11 24.69 24.19 24.48 23.97 23.18 22.68 23.25 22.75 22.22 21.71
0.50 15.66 15.15 17.57 17.07 13.62 13.11 12.71 12.20 12.14 11.63 11.88 11.37 11.41 10.90 11.57 11.06 11.00 10.49
0.40 8.05 7.53 9.17 8.65 6.83 6.31 6.48 5.96 6.20 5.68 6.03 5.51 5.92 5.40 6.08 5.56 5.82 5.30
0.30 4.22 3.68 4.84 4.31 3.62 3.08 3.53 2.99 3.42 2.88 3.35 2.81 3.38 2.83 3.51 2.97 3.41 2.87
0.20 2.33 1.76 2.66 2.10 2.11 1.53 2.13 1.55 2.11 1.53 2.10 1.52 2.15 1.57 2.23 1.65 2.20 1.63
0.10 1.42 0.77 1.57 0.95 1.40 0.74 1.43 0.78 1.43 0.79 1.43 0.79 1.46 0.82 1.49 0.85 1.48 0.85
7
 8

Table 3: The ARL and SDRL values of the proposed chart at r0 � 370 with the m value varying from 2 to 10.
m 2 3 4 5 6 7 8 9 10
k1 2.78334 2.86733 3.01155 2.9984146 3.06503 3.10255 3.13290 2.87685 2.86028
k2 1.86932 1.80537 1.80629 1.8756383 1.90764 1.94119 1.97057 2.09426 2.13496
θ0 /θ1 ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL
1 370.10 369.52 370.06 369.56 369.99 369.49 369.99 369.49 370.00 369.50 370.00 369.50 370.02 369.52 370.04 369.54 370.00 369.50
0.90 225.34 224.84 221.27 220.77 224.24 223.74 220.17 219.67 220.93 220.43 220.41 219.91 212.52 212.02 211.30 210.79 211.13 210.63
0.80 126.83 126.33 121.59 121.09 122.95 122.45 118.60 118.10 118.42 117.91 117.28 116.78 110.77 110.27 110.31 109.81 110.22 109.71
0.70 67.87 67.37 63.16 62.66 62.79 62.29 59.66 59.15 58.87 58.37 57.70 57.20 54.24 53.73 54.40 53.90 54.38 53.87
0.60 35.22 34.72 31.73 31.23 30.74 30.23 28.88 28.37 28.10 27.60 27.27 26.77 25.82 25.32 26.16 25.66 26.19 25.68
0.50 17.97 17.46 15.72 15.21 14.83 14.32 13.88 13.37 13.36 12.85 12.90 12.39 12.42 11.91 12.72 12.21 12.78 12.27
0.40 9.13 8.62 7.85 7.33 7.28 6.76 6.86 6.34 6.59 6.07 6.39 5.87 6.30 5.78 6.52 6.00 6.59 6.07
0.30 4.71 4.18 4.07 3.53 3.77 3.24 3.63 3.09 3.53 2.99 3.48 2.94 3.52 2.98 3.67 3.13 3.74 3.20
0.20 2.54 1.97 2.27 1.70 2.16 1.58 2.14 1.57 2.13 1.55 2.14 1.56 2.20 1.62 2.29 1.72 2.34 1.77
0.10 1.49 0.86 1.43 0.78 1.41 0.76 1.43 0.78 1.43 0.79 1.45 0.80 1.47 0.84 1.51 0.88 1.53 0.90
Mathematical Problems in Engineering
 Mathematical Problems in Engineering

Table 4: The ARL and SDRL values of the proposed chart at r0 � 500 with the m value varying from 2 to 10.
m 2 3 4 5 6 7 8 9 10
k1 2.92776 2.88286 2.84840 2.86525 2.93012 2.90614 3.02547 2.87998 2.88962
k2 1.69567 1.90976 2.09168 2.09376 2.03635 2.10125 2.04767 2.20013 2.20653
θ0 /θ1 ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL
1 500.04 499.54 500.00 499.50 499.99 499.49 499.99 499.49 500.06 499.59 500.00 499.50 500.00 499.50 499.99 499.49 500.01 499.51
0.90 292.08 291.58 285.90 285.40 286.47 285.97 283.00 282.50 278.35 277.85 277.08 276.58 276.95 276.45 276.92 276.42 274.92 274.42
0.80 157.62 157.12 151.64 151.14 152.41 151.91 148.41 147.91 142.59 142.09 141.58 141.07 139.19 138.68 141.54 141.04 139.34 138.84
0.70 80.74 80.24 76.62 76.12 77.27 76.77 74.03 73.53 69.19 68.69 68.61 68.10 65.68 65.18 68.63 68.13 66.98 66.48
0.60 40.05 39.54 37.59 37.08 38.00 37.50 35.82 35.32 32.60 32.10 32.32 31.82 30.11 29.61 32.37 31.86 31.38 30.88
0.50 19.53 19.03 18.19 17.68 18.42 17.91 17.15 16.64 15.33 14.82 15.23 14.72 13.94 13.43 15.31 14.80 14.82 14.31
0.40 9.52 9.00 8.85 8.34 8.98 8.47 8.34 7.82 7.45 6.93 7.44 6.93 6.81 6.29 7.56 7.04 7.37 6.85
0.30 4.73 4.20 4.45 3.92 4.54 4.01 4.28 3.74 3.90 3.36 3.94 3.41 3.68 3.14 4.08 3.54 4.03 3.49
0.20 2.48 1.92 2.41 1.84 2.49 1.92 2.41 1.84 2.29 1.71 2.34 1.77 2.25 1.68 2.44 1.88 2.44 1.88
0.10 1.45 0.81 1.47 0.83 1.52 0.89 1.51 0.88 1.49 0.85 1.51 0.88 1.49 0.85 1.56 0.93 1.56 0.93
9
 10

Table 5: The ARL and SDRL comparison of the proposed chart with the MDS chart at r0 � 200 with the m value varying from 2 to 6.
MMDS MDS MMDS MDS MMDS MDS MMDS MDS MMDS MDS
m 2 3 4 5 6
k1 2.65298 3.05325 2.83972 2.93820 3.01942 3.17867 2.98755 2.81403 3.11139 2.64592
k2 1.80389 1.95328 1.64421 2.04096 1.68708 2.06290 1.76101 2.16003 1.79793 2.36430
θ0 /θ1 ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL
1 199.99 199.49 199.99 199.49 200.00 199.50 200.00 199.50 200.01 199.51 200.00 199.50 200.00 199.50 200.01 199.51 200.00 199.50 200.00 199.50
0.90 133.16 132.66 139.12 138.62 130.57 130.07 134.54 134.04 132.51 132.01 137.09 136.58 129.58 129.08 130.97 130.47 130.73 130.23 134.26 133.76
0.80 81.26 80.76 87.56 87.06 77.72 77.22 82.12 81.61 79.24 78.74 84.39 83.89 75.79 75.29 78.16 77.65 76.56 76.06 82.58 82.08
0.70 46.69 46.19 50.92 50.42 43.26 42.76 46.68 46.17 43.76 43.26 47.79 47.29 41.09 40.59 43.79 43.29 41.24 40.74 47.84 47.34
0.60 25.84 25.34 28.06 27.56 23.10 22.59 25.41 24.90 22.96 22.45 25.68 25.17 21.31 20.80 23.73 23.22 21.14 20.64 26.75 26.25
0.50 14.01 13.50 15.03 14.52 12.11 11.60 13.59 13.08 11.79 11.28 13.51 13.00 10.93 10.41 12.76 12.25 10.72 10.21 14.73 14.22
0.40 7.56 7.05 8.00 7.49 6.40 5.88 7.32 6.80 6.13 5.61 7.18 6.66 5.74 5.22 6.99 6.47 5.62 5.09 8.15 7.63
0.30 4.14 3.60 4.34 3.80 3.51 2.96 4.06 3.52 3.36 2.81 3.97 3.44 3.22 2.68 3.99 3.45 3.17 2.63 4.61 4.08
0.20 2.36 1.79 2.45 1.89 2.07 1.49 2.38 1.81 2.02 1.43 2.35 1.78 2.00 1.42 2.41 1.85 2.00 1.42 2.71 2.15
0.10 1.45 0.81 1.50 0.86 1.37 0.71 1.50 0.87 1.37 0.71 1.50 0.87 1.38 0.73 1.54 0.91 1.39 0.74 1.64 1.02
Mathematical Problems in Engineering
 Mathematical Problems in Engineering

Table 6: The ARL and SDRL comparison of the proposed chart with the MDS chart at r0 � 300 with the m value varying from 2 to 6.
MMDS MDS MMDS MDS MMDS MDS MMDS MDS MMDS MDS
m 2 3 4 5 6
k1 2.80952 3.03098 2.71631 2.71297 2.99340 3.10876 2.89292 3.10105 2.95491 3.11713
k2 1.64840 2.04333 2.18966 2.45933 1.76962 2.14471 1.87672 2.17995 1.89838 2.20569
θ0 /θ1 ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL
1 300.00 299.50 300.01 299.51 300.00 299.50 300.01 299.51 299.98 299.48 300.02 299.52 300.02 299.51 300.04 299.54 300.01 299.51 300.00 299.50
0.90 187.23 186.73 197.31 196.81 191.28 190.78 193.15 192.65 186.86 186.36 194.22 193.72 181.02 180.51 191.94 191.44 180.56 180.06 190.73 190.23
0.80 107.38 106.88 117.23 116.73 112.26 111.76 114.71 114.21 105.23 104.73 112.75 112.25 99.49 98.99 110.12 109.62 98.42 97.92 108.59 108.09
0.70 58.18 57.68 64.64 64.14 62.35 61.85 64.66 64.15 55.09 54.59 60.52 60.01 51.36 50.85 58.51 58.01 50.16 49.66 57.27 56.77
0.60 30.45 29.94 34.01 33.50 33.45 32.94 35.33 34.82 27.60 27.10 31.03 30.53 25.62 25.11 29.81 29.31 24.69 24.19 29.02 28.52
0.50 15.66 15.15 17.49 16.99 17.57 17.07 18.98 18.47 13.62 13.11 15.68 15.17 12.71 12.20 15.05 14.54 12.14 11.63 14.64 14.13
0.40 8.05 7.53 8.99 8.48 9.17 8.65 10.13 9.62 6.83 6.31 8.04 7.52 6.48 5.96 7.76 7.25 6.20 5.68 7.59 7.07
0.30 4.22 3.68 4.72 4.19 4.84 4.31 5.44 4.91 3.62 3.08 4.30 3.77 3.53 2.99 4.22 3.68 3.42 2.88 4.17 3.64
0.20 2.33 1.76 2.59 2.03 2.66 2.10 2.98 2.43 2.11 1.53 2.47 1.91 2.13 1.55 2.47 1.90 2.11 1.53 2.47 1.91
0.10 1.42 0.77 1.53 0.91 1.57 0.95 1.70 1.09 1.40 0.74 1.54 0.91 1.43 0.78 1.55 0.92 1.43 0.79 1.56 0.93
11
 12

Table 7: The ARL and SDRL comparison of the proposed chart with the MDS chart at r0 � 370 with the m value varying from 2 to 6.
MMDS MDS MMDS MDS MMDS MDS MMDS MDS MMDS MDS
m 2 3 4 5 6
k1 2.78334 2.91731 2.86733 2.98925 3.01155 3.06655 2.99841 3.19171 3.06503 3.18454
k2 1.86932 2.13330 1.80537 2.16569 1.80629 2.19134 1.87564 2.20979 1.90764 2.23733
θ0 /θ1 ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL
1 370.10 369.52 370.05 369.55 370.06 369.56 370.04 369.54 369.99 369.49 370.05 369.55 369.99 369.49 370.01 369.51 370.00 369.50 370.01 369.51
0.90 225.34 224.84 230.10 229.60 221.27 220.77 230.11 229.61 224.24 223.74 230.93 23.43 220.17 219.67 233.81 233.31 220.93 220.43 231.29 230.79
0.80 126.83 126.33 130.81 130.31 121.59 121.09 129.83 129.32 122.95 122.45 129.73 129.23 118.60 118.10 131.60 131.10 118.42 117.91 128.81 128.31
0.70 67.87 67.37 70.07 69.57 63.16 62.66 68.58 68.08 62.79 62.29 67.79 67.29 59.66 59.15 68.24 67.74 58.87 58.37 66.22 65.72
0.60 35.22 34.72 36.26 35.76 31.73 31.23 34.91 34.40 30.74 30.23 34.04 33.53 28.88 28.37 33.82 33.32 28.10 27.60 32.66 32.16
0.50 17.97 17.46 18.50 17.99 15.72 15.21 17.54 17.03 14.83 14.32 16.90 16.39 13.88 13.37 16.59 16.08 13.36 12.85 16.02 15.51
0.40 9.13 8.62 9.46 8.94 7.85 7.33 8.89 8.38 7.28 6.76 8.53 8.01 6.86 6.34 8.32 7.80 6.59 6.07 8.09 7.57
0.30 4.71 4.18 4.93 4.41 4.07 3.53 4.66 4.13 3.77 3.24 4.49 3.96 3.63 3.09 4.40 3.87 3.53 2.99 4.34 3.81
0.20 2.54 1.97 2.69 2.13 2.27 1.70 2.59 2.03 2.16 1.58 2.54 1.98 2.14 1.57 2.52 1.96 2.13 1.55 2.53 1.97
0.10 1.49 0.86 1.57 0.95 1.43 0.78 1.56 0.93 1.41 0.76 1.56 0.93 1.43 0.78 1.56 0.94 1.43 0.79 1.57 0.95
Mathematical Problems in Engineering
 Mathematical Problems in Engineering

Table 8: The ARL and SDRL comparison of the proposed chart with the MDS chart at r0 � 500 with the m value varying from 2 to 6.
MMDS MDS MMDS MDS MMDS MDS MMDS MDS MMDS MDS
m 2 3 4 5 6
k1 2.92776 3.09800 2.88286 3.05617 2.84840 2.93402 2.86525 3.15044 2.93012 3.15310
k2 1.69567 2.13929 1.90976 2.21263 2.09168 2.31140 2.09376 2.26863 2.03635 2.29419
θ0 /θ1 ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL
1 500.04 499.54 500.00 499.39 500.00 499.50 500.00 499.50 499.99 499.49 500.01 499.51 499.99 499.49 500.01 499.50 500.06 499.59 500.04 499.54
0.90 292.08 291.58 309.62 309.12 285.90 285.40 300.82 300.32 286.47 285.97 290.69 290.19 283.00 282.50 300.60 300.10 278.35 277.85 297.97 297.47
0.80 157.62 157.12 172.79 172.29 151.64 151.14 163.89 163.39 152.41 151.91 155.61 155.11 148.41 147.91 161.70 161.20 142.59 142.09 158.83 158.33
0.70 80.74 80.24 89.70 89.20 76.62 76.12 83.60 83.10 77.27 76.77 79.12 78.62 74.03 73.53 80.75 80.25 69.19 68.69 78.69 78.19
0.60 40.05 39.54 44.59 44.09 37.59 37.08 41.11 40.61 38.00 37.50 39.15 38.64 35.82 35.32 38.80 38.29 32.60 32.10 37.60 37.10
0.50 19.53 19.03 21.75 21.24 18.19 17.68 19.97 19.46 18.42 17.91 19.24 18.73 17.15 16.64 18.52 18.01 15.33 14.82 17.93 17.42
0.40 9.52 9.00 10.62 10.11 8.85 8.34 9.80 9.28 8.98 8.47 9.58 9.07 8.34 7.82 9.05 8.54 7.45 6.93 8.81 8.30
0.30 4.73 4.20 5.31 4.78 4.45 3.92 4.97 4.45 4.54 4.01 4.96 4.43 4.28 3.74 4.68 4.15 3.90 3.36 4.61 4.08
0.20 2.48 1.92 2.78 2.23 2.41 1.84 2.69 2.13 2.49 1.92 2.73 2.17 2.41 1.84 2.62 2.06 2.29 1.71 2.62 2.06
0.10 1.45 0.81 1.58 0.96 1.47 0.83 1.58 0.96 1.52 0.89 1.62 1.00 1.51 0.88 1.59 0.97 1.49 0.85 1.60 0.98
13
14 Mathematical Problems in Engineering

200

150

ARLs
100

50

0
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
The shift ratio
MMDS
MDS
Figure 1: The ARL values for the MMDS and MDS at m � 4 and r0 � 200.

200

150
SDRLs

100

50

0
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
The shift ratio

MMDS
MDS
Figure 2: The SDRL values for the MMDS and MDS at m � 4 and r0 � 200.

300

250

200
ARLs

150

100

50

0
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
The shift ratio
MMDS
MDS
Figure 3: The ARL values for the MMDS and MDS at m � 5 and r0 � 300.
Mathematical Problems in Engineering 15

300

250

200

SDRLs
150

100

50

0
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
The shift ratio
MMDS
MDS
Figure 4: The SDRL values for the MMDS and MDS at m � 5 and r0 � 300.

400

350

300

250
ARLs

200

150

100

50

0
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
The shift ratio

MMDS
MDS
Figure 5: The ARL values for the MMDS and MDS at m � 6 and r0 � 370.

400

350

300

250
SDRLs

200

150

100

50

0
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
The shift ratio
MMDS
MDS
Figure 6: The SDRL values for the MMDS and MDS at m � 6 and r0 � 370.
16 Mathematical Problems in Engineering

500
450
400
350
300

ARLs
250
200
150
100
50
0
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
The shift ratio
MMDS
MDS
Figure 7: The ARL values for the MMDS and MDS at m � 6 and r0 � 500.

500
450
400
350
300
ARLs

250
200
150
100
50
0
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
The shift ratio

MMDS
MDS
Figure 8: The SDRL values for the MMDS and MDS at m � 6 and r0 � 500.

2.0 2.0
UCL 1 = 1.80037
UCL 1 = 1.73236

UCL 2 = 1.52556
1.5 UCL 2 = 1.42063
1.5
T∗

1.0
T∗ 1.0

0.5
0.5 LCL 2 = 0.37912
LCL 2 = 0.28865
LCL 1 = 0.06739
LCL 1 = 0.01384
0.0
0.0
0 10 20 30 40 50 60
0 10 20 30 40 50 60 Observations
Observations
Figure 10: The MMDS control chart for the simulated data at
Figure 9: The MDS control chart for the simulated data at m � 5. m � 5.
Mathematical Problems in Engineering 17

Table 9: The ARL comparison of the proposed chart with GMDS at r0 � 200.
MMDS GMDS MMDS GMDS MMDS GMDS MMDS GMDS
m 4 5 6 7
K — 2 — 3 — 4 — 4
k1 3.01942 2.7187 2.98755 2.7197 3.11139 2.8679 3.12916 2.7106
k2 1.68708 1.4458 1.76101 1.5486 1.79793 1.5502 1.83897 1.4557
θ0 /θ1 ARL ARL ARL ARL
1 200.01 199.99 200.00 199.99 200.00 200.00 200.01 200.00
0.90 132.51 133.58 129.58 131.89 130.73 130.12 129.59 130.62
0.80 79.24 81.86 75.79 79.60 76.56 77.04 75.08 77.93
0.70 43.76 47.12 41.09 45.01 41.24 42.45 40.01 43.45
0.60 22.96 25.99 21.31 24.34 21.14 22.29 20.36 23.09
0.50 11.79 13.96 10.93 12.83 10.72 11.45 10.32 11.92
0.40 6.13 7.41 5.74 6.72 5.62 5.93 5.45 6.13
0.30 3.36 3.97 3.22 3.60 3.17 3.22 3.14 3.27

Table 10: The ARL comparison of the proposed chart with GMDS at r0 � 300.
MMDS GMDS MMDS GMDS MMDS GMDS MMDS GMDS
m 4 5 6 7
K — 2 — 3 — 4 — 4
k1 2.74652 2.7682 2.89292 2.8651 2.95491 2.7608 3.08241 2.9145
k2 2.03462 1.5442 1.87672 1.5572 1.89838 1.7252 1.90958 1.4368
θ0 /θ1 ARL ARL ARL ARL
1 300.01 300.00 300.02 300.00 300.01 300.00 300.03 299.99
0.90 186.19 190.63 181.02 188.21 180.56 187.83 183.17 185.92
0.80 106.03 111.50 99.49 107.99 98.42 107.76 100.01 105.00
0.70 57.00 61.59 51.36 58.05 50.16 58.07 50.49 55.27
0.60 29.58 32.74 25.62 29.85 24.69 29.94 24.48 27.71
0.50 15.11 16.98 12.71 14.96 12.14 15.00 11.88 13.52
0.40 7.76 8.69 6.48 7.48 6.20 7.46 6.03 6.61
0.30 4.14 4.47 3.53 3.84 3.42 3.83 3.35 3.39

Table 11: The ARL comparison of the proposed chart with GMDS at r0 � 370.
MMDS GMDS MMDS GMDS MMDS GMDS MMDS GMDS
m 4 5 6 7
K — 2 — 3 — 4 — 4
k1 3.01155 2.8074 2.99841 2.894 3.06503 2.8527 3.10255 2.907
k2 1.80629 1.5739 1.87564 1.5902 1.90764 1.6918 1.94119 1.478
θ0 /θ1 ARL ARL ARL ARL
1 369.99 370.00 369.99 370.00 370.00 370.00 370.00 369.99
0.90 224.24 228.37 220.17 226.03 220.93 223.11 220.41 222.88
0.80 122.95 130.15 118.60 126.51 118.42 123.35 117.28 122.81
0.70 62.79 70.29 59.66 66.50 58.87 64.09 57.70 63.35
0.60 30.74 36.62 28.88 33.48 28.10 31.90 27.27 31.19
0.50 14.83 18.62 13.88 16.44 13.36 15.49 12.90 14.94
0.40 7.28 9.35 6.86 8.04 6.59 7.53 6.39 7.15
0.30 3.77 4.71 3.63 4.04 3.53 3.81 3.48 3.57
18 Mathematical Problems in Engineering

Table 12: Urinary tract infection (UTI) data from [15]. fewer samples than GMDS to detect the shift. Likewise, when
Sample number T
m � 5 and r0 � 370, the proposed chart is found to be more
efficient.
1 0.57014
2 0.07431
3 0.15278 5. Applications
4 0.14583
5 0.13889 In this section, two sets of real data are employed to evaluate
6 0.14931 the performance of the proposed t-chart under MMDS
7 0.03333 sampling compared to the existing t-charts.
8 0.08681
9 0.33681
10 0.03819 5.1. Urinary Tract Infections. The time difference between
11 0.24653
admission and discharge of male patients with urinary tract
12 0.29514
13 0.11944 infections (UTIs) in a large hospital is discussed in Nelson’s
14 0.05208 research [15]. The observations (T), shown in Table 12, are
15 0.12500 fitted to an exponential distribution with a mean time of 0.21
16 0.25000 days. Here, the dataset is considered to compare the per-
17 0.40069 formance of the proposed chart under the MMDS scheme, in
18 0.02500 terms of the ARL and SDRL, with that of the existing t-chart
19 0.12014 when r0 � 370, as shown in Tables 13–14 and plotted in
20 0.11458 Figures 11 and 12. For instance, when m � 4, the proposed
21 0.00347 chart takes 219 samples to detect the out-of-control signal
22 0.12014
after the shift of “θ0 /θ1 � 9” occurs in the process mean,
23 0.04861
24 0.02778
while it takes almost 226 samples to detect the shift with the
25 0.32639 t-chart under the MDS scheme. Similarly, the SDRL of the
26 0.64931 proposed chart takes 218.73 samples compared to the 255.91
27 0.14931 samples for the t-chart under the MDS scheme. The control
28 0.01389 charts for the data are plotted in Figures 13 and 14. They
29 0.03819 show better detection ability for the proposed chart. Ob-
30 0.46806 servation 21, for example, is considered to be in-control with
31 0.22222 the t-chart under MDS, when in fact it is located in the in-
32 0.29514 decision area with the t-chart under the MMDS sampling
33 0.53472 scheme.
34 0.15139
35 0.52569
36 0.07986
37 0.27083 5.2. Customer Complaint Resolution Time. A Saudi com-
38 0.04514 pany, one of the Middle East’s largest utility companies by
39 0.13542 market value and among the top 15 in its field worldwide,
40 0.08681 provides communication channels for its customers to voice
41 0.40347 their complaints. In order to guarantee that the service
42 0.12639 provided by the company meets the highest international
43 0.18403 standards of quality and efficiency, the regulatory authority
44 0.70833 of Saudi Arabia requires it to successfully meet the standard
45 0.15625 for time to resolve complaints (TRC). Failure to do so leads
46 0.24653
to a payment to the customer once 10 working days have
47 0.04514
48 0.01736
passed. Further compensation is paid for each additional 10
49 1.08889 working days, whereby the required services are not com-
50 0.05208 pleted. The time needed to resolve the complaints, or the
51 0.02778 time from submission to resolution, is found to follow an
52 0.03472 exponential distribution based on a goodness-of-fit test.
53 0.23611 In this section, the proposed control chart is applied to
54 0.35972 monitor the time, in days, taken by the company to resolve
140 registered complaints in November 2019, as shown in
Table 15. The control chart under MMDS is constructed and
proposed chart exceeds that of the GMDS. For example, compared to the existing chart assuming a target ARL � 370
when m � 5 and r0 � 200, it takes 129 samples to detect the with the m value varying from 2 to 9. From this, the control
out-of-control shift of “θ0 /θ1 � 0.9” with the proposed chart coefficients k1 and k2 are obtained. Tables 16 and 17
t-chart, while it takes 131 samples with GMDS. Similarly, represent a comparison of the ARL and SDRL values of the t-
when m � 5 and r0 � 300, the proposed chart needs seven charts under both schemes for different shift ratios (θ0 /θ1 ).
 Mathematical Problems in Engineering

Table 13: The ARL and SDRL comparison of the proposed chart with MDS for UTI data at r0 � 370 with the m value varying from 2 to 5.
MMDS MDS MMDS MDS MMDS MDS MMDS MDS
m 2 3 4 5
k1 2.85412 3.09359 2.77324 3.18082 2.83068 2.95801 3.04052 3.03891
k2 1.67334 2.07679 2.06525 2.13187 1.94589 2.21982 1.86324 2.22984
θ0 /θ1 ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL
1 370.01 369.51 370.01 369.51 369.99 369.49 369.99 369.49 369.99 369.49 369.99 369.49 370.02 369.52 370.02 369.52
0.90 224.50 224.00 239.62 239.12 225.51 225.01 239.59 239.09 219.23 218.73 226.41 225.91 222.49 221.99 227.47 226.97
0.80 125.48 124.98 139.58 139.07 126.88 126.38 138.50 138.00 119.50 119.00 126.06 125.56 120.51 120.01 126.18 125.68
0.70 66.42 65.92 75.23 74.73 67.75 67.25 73.62 73.12 61.66 61.15 65.97 65.47 60.66 60.15 65.36 64.86
0.60 34.01 33.50 38.63 38.13 35.02 34.51 37.17 36.67 30.80 30.30 33.39 32.89 29.28 28.77 32.67 32.17
0.50 17.12 16.61 19.38 18.88 17.77 17.26 18.37 17.86 15.22 14.71 16.77 16.26 14.00 13.49 16.24 15.73
0.40 8.61 8.10 9.72 9.20 9.00 8.48 9.13 8.62 7.62 7.10 8.55 8.03 6.88 6.36 8.25 7.74
0.30 4.42 3.89 4.98 4.45 4.65 4.12 4.70 4.17 4.00 3.46 4.54 4.01 3.62 3.08 4.42 3.88
0.20 2.39 1.82 2.67 2.11 2.54 1.98 2.57 2.01 2.28 1.71 2.57 2.01 2.14 1.56 2.54 1.98
0.10 1.43 0.79 1.55 0.93 1.52 0.89 1.54 0.92 1.46 0.82 1.57 0.95 1.42 0.77 1.57 0.95
19
 20

Table 14: The ARL and SDRL comparison of the proposed chart with MDS for UTI data at r0 � 370 with the m value varying from 6 to 10.
MMDS MDS MMDS MDS MMDS MDS MMDS MDS MMDS MDS
m 6 7 8 9 10
k1 2.79727 3.10230 2.78925 2.86690 2.80485 2.99383 3.12831 2.99195 2.83832 2.95881
k2 2.13203 2.24591 2.19692 2.34644 2.17780 2.30674 2.00009 2.32315 2.16276 2.34671
θ0 /θ1 ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL
1 370.00 369.50 370.01 369.51 369.99 369.49 370.00 369.50 369.99 369.49 370.01 369.51 370.02 369.52 370.02 369.52 370.00 369.50 369.99 369.49
0.90 219.78 219.28 228.10 227.60 220.44 219.94 221.41 220.91 217.53 217.02 221.57 221.07 217.47 216.97 220.49 219.99 212.67 212.17 219.02 218.52
0.80 119.99 119.49 126.12 125.62 120.68 120.18 121.52 121.02 117.34 116.84 120.29 119.79 113.73 113.23 119.17 118.67 111.92 111.42 118.00 117.50
0.70 61.94 61.44 64.82 64.32 62.42 61.92 63.19 62.69 59.79 59.29 61.42 60.92 54.99 54.49 60.64 60.14 55.66 55.16 60.05 59.55
0.60 30.93 30.42 32.11 31.61 31.22 30.72 32.01 31.51 29.52 29.01 30.53 30.03 25.67 25.17 30.10 29.60 26.97 26.46 29.89 29.39
0.50 15.29 14.78 15.86 15.35 15.46 14.95 16.21 15.70 14.52 14.01 15.27 14.76 12.14 11.63 15.08 14.57 13.19 12.68 15.05 14.54
0.40 7.70 7.18 8.06 7.55 7.82 7.30 8.42 7.90 7.37 6.86 7.92 7.40 6.12 5.59 7.87 7.36 6.80 6.28 7.91 7.40
0.30 4.11 3.57 4.35 3.82 4.19 3.66 4.60 4.07 4.03 3.49 4.38 3.85 3.44 2.89 4.39 3.86 3.83 3.30 4.44 3.91
0.20 2.40 1.83 2.54 1.98 2.46 1.90 2.68 2.12 2.42 1.85 2.60 2.04 2.17 1.60 2.62 2.06 2.38 1.81 2.65 2.10
0.10 1.53 0.90 1.58 0.96 1.56 0.93 1.63 1.01 1.55 0.92 1.61 0.99 1.47 0.83 1.62 1.00 1.54 0.91 1.63 1.01
Mathematical Problems in Engineering
Mathematical Problems in Engineering 21

400

350

300

250

ARLs
200

150

100

50

0
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
The shift ratio
MMDS
MDS
Figure 11: The ARL values for the MMDS and MDS for UTI data at m � 2 and r0 � 370.

400 1.2 UCL 1 = 1.094647

350 UCL 2 = 0.935135

1.0
300
0.8
250
SDRLs

0.6
T∗

200

150 0.4
LCL 2 = 0.233514
100 0.2
LCL 1 = 0.74002
50
0.0
0 0 10 20 30 40 50
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
Observations
The shift ratio
Figure 14: The MMDS control chart for UTI data at m � 4.
MMDS
MDS
Figure 12: The SDRL values for the MMDS and MDS for UTI
data at m � 2 and r0 � 370.
Table 15: Data for time to resolve complaints (TRC).
Sample number T∗
1 1.040
1.2 UCL 1 = 1.117602 2 1.027
UCL 2 = 0.984521 3 1.334
1.0 4 1.449
5 0.519
0.8 6 0.905
7 1.847
0.6 8 1.461
T∗

9 1.554
0.4 10 1.595
11 1.207
LCL 2 = 0.02778
0.2 12 0.278
LCL 1 = 0.051046 13 0.603
14 1.551
0.0
15 1.129
0 10 20 30 40 50 16 0.928
Observations 17 1.466
18 0.937
Figure 13: The MDS control chart for UTI data at m � 4.
22 Mathematical Problems in Engineering

Table 15: Continued. Table 15: Continued.

Sample number T∗ Sample number T∗
19 0.820 80 0.928
20 0.616 81 1.754
21 1.246 82 0.558
22 0.983 83 1.240
23 1.219 84 1.026
24 0.897 85 1.206
25 0.863 86 0.976
26 1.078 87 0.513
27 1.188 88 1.454
28 1.237 89 0.166
29 1.010 90 0.855
30 1.007 91 1.032
31 0.897 92 0.620
32 0.391 93 0.819
33 1.621 94 1.210
34 1.026 95 0.440
35 1.040 96 0.270
36 1.277 97 1.218
37 0.608 98 1.135
38 0.987 99 1.070
39 1.319 100 1.458
40 0.915 101 0.485
41 1.780 102 0.857
42 1.020 103 0.596
43 0.948 104 0.885
44 1.205 105 0.239
45 0.626 106 1.326
46 1.511 107 0.797
47 1.202 108 1.152
48 0.671 109 0.995
49 0.452 110 0.972
50 1.083 111 0.240
51 0.911 112 0.282
52 1.165 113 1.002
53 1.239 114 0.947
54 0.718 115 1.248
55 1.215 116 0.785
56 0.664 117 1.330
57 1.267 118 0.833
58 0.960 119 0.531
59 0.712 120 0.937
60 0.955 121 0.921
61 0.877 122 0.523
62 1.075 123 0.885
63 0.653 124 1.048
64 1.313 125 0.282
65 1.051 126 0.841
66 0.607 127 0.614
67 0.667 128 0.520
68 1.284 129 0.974
69 0.876 130 0.396
70 0.751 131 1.560
71 1.505 132 1.470
72 1.026 133 0.780
73 0.996 134 1.264
74 0.507 135 1.233
75 1.301 136 0.996
76 1.419 137 0.928
77 1.070 138 0.648
78 1.228 139 0.251
79 1.052 140 1.386
Mathematical Problems in Engineering 23

Table 16: The ARL and SDRL comparison of the proposed chart with MDS for TRC data at r0 � 370 with m value varying from 2 to 5.
MMDS MDS MMDS MDS MMDS MDS MMDS MDS
m 2 3 4 5
k1 2.88690 3.10355 2.78183 2.89276 3.17406 3.18913 3.00510 3.11057
k2 1.63317 2.07520 2.01076 2.21062 1.77478 2.17615 1.87341 2.21810
θ0 /θ1 ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL
1 370.00 369.50 370.00 369.50 370.00 369.50 370.01 369.51 370.01 369.51 370.00 369.50 369.99 369.49 369.99 369.49
0.90 226.02 225.52 240.19 239.69 224.25 223.75 226.60 226.10 234.95 234.45 236.45 235.94 220.52 220.02 230.45 229.95
0.80 126.81 126.31 140.14 139.64 125.42 124.92 127.11 126.61 132.87 132.37 134.69 134.19 118.88 118.38 128.66 128.16
0.70 67.16 66.65 75.59 75.09 66.56 66.06 67.36 66.85 68.68 68.18 70.60 70.10 59.80 59.30 66.63 66.13
0.60 34.31 33.81 38.81 38.31 34.19 33.69 34.58 34.08 33.48 32.97 35.26 34.76 28.93 28.43 33.15 32.65
0.50 17.21 16.70 19.46 18.95 17.27 16.76 17.56 17.06 15.90 15.40 17.33 16.82 13.89 13.38 16.37 15.86
0.40 8.62 8.10 9.74 9.23 8.72 8.20 8.99 8.48 7.63 7.11 8.64 8.13 6.86 6.34 8.26 7.75
0.30 4.41 3.87 4.98 4.46 4.51 3.98 4.74 4.21 3.86 3.32 4.51 3.98 3.63 3.09 4.40 3.87
0.20 2.38 1.81 2.67 2.11 2.47 1.91 2.64 2.08 2.16 1.59 2.53 1.97 2.14 1.56 2.53 1.97
0.10 1.42 0.78 1.55 0.93 1.50 0.87 1.58 0.95 1.40 0.75 1.55 0.93 1.42 0.78 1.57 0.94

Table 17: The ARL and SDRL comparison of the proposed chart with the MDS chart for TRC data at r0 � 370 with m value varying from
6 to 9.
MMDS MDS MMDS MDS MMDS MDS MMDS MDS
m 6 7 8 9
k1 2.79685 3.15202 2.90606 3.10523 3.05044 3.08986 2.95615 3.18391
k2 2.13339 2.24025 2.01025 2.26763 1.98476 2.28843 2.04333 2.29514
θ0 /θ1 ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL ARL SDRL
1 370.00 369.50 370.00 369.50 370.01 369.51 370.01 369.51 370.01 369.51 370.00 369.50 370.01 369.51 370.01 369.51
0.90 219.84 219.33 230.05 229.55 213.25 212.75 226.52 226.02 215.68 215.18 224.52 224.02 210.90 210.40 226.34 225.84
0.80 120.06 119.55 127.75 127.25 112.22 111.72 124.35 123.85 112.93 112.43 122.35 121.85 109.17 108.67 123.37 122.87
0.70 61.99 61.49 65.66 65.16 55.67 55.17 63.52 63.02 55.06 54.56 62.21 61.70 53.15 52.65 62.35 61.85
0.60 30.96 30.46 32.43 31.93 26.84 26.33 31.35 30.85 25.98 25.48 30.66 30.16 25.24 24.73 30.49 29.99
0.50 15.30 14.80 15.95 15.44 12.99 12.48 15.49 14.98 12.37 11.86 15.20 14.69 12.17 11.66 15.03 14.52
0.40 7.71 7.19 8.08 7.56 6.56 6.04 7.93 7.41 6.23 5.70 7.83 7.32 6.23 5.71 7.75 7.23
0.30 4.11 3.58 4.34 3.81 3.61 3.07 4.33 3.79 3.46 2.92 4.32 3.79 3.53 2.98 4.30 3.77
0.20 2.40 1.83 2.53 1.97 2.21 1.64 2.55 1.99 2.17 1.59 2.57 2.01 2.22 1.65 2.58 2.01
0.10 1.53 0.90 1.58 0.95 1.47 0.84 1.59 0.97 1.46 0.82 1.60 0.98 1.49 0.85 1.60 0.98

400

350

300

250
ARLs

200

150

100

50

0
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
The shift ratio
MMDS
MDS
Figure 15: The ARL values for MMDS and MDS for TRC data at m � 2 and r0 � 370.
24 Mathematical Problems in Engineering

400

350

300

250

SDRLs
200

150

100

50

0
1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0
The shift ratio
MMDS
MDS
Figure 16: The SDRL values for MMDS and MDS for TRC data at m � 2 and r0 � 370.

UCL 1 = 1.99252
2.0
UCL 2 = 1.70835
1.5

1.0
T∗

0.5 LCL 2 = 0.31195

LCL 1 = 0.02778
0.0
0 20 40 60 80 100 120 140
Observations
Figure 17: The MDS control chart for TRC data at m � 6.

2.0 UCL 1 = 188183

UCL 2 = 1.67505
1.5
T∗

1.0

0.5
LCL 2 = 0.34525
LCL 1 = 0.13848
0.0
0 20 40 60 80 100 120 140
Observations
Figure 18: The MMDS control chart for TRC data at m � 6.

This example shows that the proposed chart detects while the observation under MDS is obviously in the in-
process shifts more quickly than the t-chart under the MDS decision area.
scheme, as shown in Figures 15 and 16. For example, when
m � 2, the proposed chart takes almost 226 samples to detect 6. Conclusions
the out-of-control signal after the shift of “θ0 /θ1 � 0.9” oc-
curs in the process mean, while the t-chart under the MDS In this paper, a new t-control chart based on an MMDS
scheme takes almost 240 samples to detect the shift. Simi- sampling scheme is introduced to monitor processes under
larly, the SDRL of the proposed chart takes 225.52 samples exponential distribution. The performance evaluation of the
compared to the 234.69 samples for the t-chart under the proposed chart is investigated using the run length char-
MDS scheme. From Figures 17 and 18, it is apparent that the acteristics. A simulation approach using the R programming
proposed chart is more sensitive to small shifts. For example, language and software is used to find the control chart
observation 7 lies near the upper outer control limit with the coefficients k1 and k2 for the different values of m and at
t-chart under the MMDS; therefore, it needs more attention, specified values of r0 . Furthermore, the ARL and SDRL
Mathematical Problems in Engineering 25

values are also computed for various mean shifts to assess the checks using statistical process control,” PLoS One, vol. 11,
shift detection ability. The superiority of the proposed chart no. 2, Article ID e0147936, 2016.
is confirmed by comparing the values of the ARL and SDRL [5] J. C. Benneyan, R. C. Lloyd, and P. E. Plsek, “Statistical process
using the existing and proposed charts. The simulation study control as a tool for research and healthcare improvement,”
also shows that the proposed control chart uniformly out- Quality and Safety in Health Care, vol. 12, no. 6, pp. 458–464,
2003.
performs the other t-charts considered in this study. Em-
[6] S. Sukparungsee, Y. Areepong, and R. Taboran, “Exponen-
pirical examples based on the monitoring of process tially weighted moving average-Moving average charts for
performance in a large hospital and a utility company are monitoring the process mean,” PLoS One, vol. 15, no. 2,
also included for the practical implementation of the pro- Article ID e0228208, 2020.
posed chart. The control chart presented in this study has [7] S. Ali, A. Pievatolo, and R. Göb, “An overview of control
been proposed for the utility company and is now in the charts for high-quality processes,” Quality and Reliability
process of implementation. The proposed chart can be ex- Engineering International, vol. 32, no. 7, pp. 2171–2189, 2016.
tended for the testing of other distributions as well. [8] A. H. Al-Marshadi, M. Aslam, A. H. Alharbey, N. Khan, and
L. Ahmad, “Monitoring customer complaints using the re-
petitive sampling,” Communications in Statistics—Theory and
Data Availability Methods, pp. 1–15. In press, 2020.
[9] M. Abujiya and M. H. Lee, “The three statistical control charts
The data are included in the article.
using ranked set sampling,” in Proceedings of the 5th Inter-
national Conference on Modeling, Simulation and Applied
Conflicts of Interest Optimization (ICMSAO), pp. 1–6, Hammamet, Tunisia, April
2013.
The authors declare no conflicts of interest. [10] M. Abujiya, M. H. Lee, and M. Riaz, “An improved combined
shewhart-EWMA chart based on double median ranked set
Authors’ Contributions sampling,” Journal of Computer Science & Computational
Mathematics, vol. 4, no. 2, pp. 23–28, 2014.
A.O.A., M.A., and S.A.D. contributed to the con- [11] K. P. Tran, “Designing of run rules t control charts for
ceptualisation, data curation, formal analysis, investigation, monitoring changes in the process mean,” Chemometrics and
methodology, project administration, and project validation; Intelligent Laboratory Systems, vol. 174, pp. 85–93, 2018.
A.O.A. and M.A. were in charge of gathering resources, [12] K. P. Tran, “Run rules median control charts for monitoring
process mean in manufacturing,” Quality and Reliability
while A.O.A. contributed to the software integration, vis-
Engineering, vol. 33, no. 8, pp. 2437–2450, 2017.
ualisation, and the writing of the first draft. M.A. and S.A.D. [13] F. Y. Yen, K. M. B. Chong, and L. M. Ha, “Synthetic-type
maintained supervisory roles, and A.O.A., M.A., and S.A.D. control charts for time-between-events monitoring,” PLoS
contributed to the writing, reviewing, and editing of the One, vol. 8, no. 6, Article ID e65440, 2013.
document. [14] E. Santiago and J. Smith, “Control charts based on the ex-
ponential distribution: adapting runs rules for the t chart,”
Acknowledgments Quality Engineering, vol. 25, no. 2, 2013.
[15] L. S. Nelson, “A control chart for parts-per-million non-
This work was funded by the Deanship of Scientific Research conforming items,” Journal of Quality Technology, vol. 26,
(DSR), King Abdulaziz University, Jeddah. The authors, no. 3, pp. 239-240, 1994.
therefore, gratefully acknowledge the DSR technical and [16] A. W. Wortham and R. C. Baker, “Multiple deferred state
financial support. sampling inspection,” International Journal of Production
Research, vol. 14, no. 6, pp. 719–731, 1976.
[17] S. Balamurali and C.-H. Jun, “Multiple dependent state
Supplementary Materials sampling plans for lot acceptance based on measurement
data,” European Journal of Operational Research, vol. 180,
R codes are given in the Supplementary File. (Supplementary no. 3, pp. 1221–1230, 2007.
Materials) [18] M. Aslam, M. Azam, N. Khan, and C.-H. Jun, “A control chart
for an exponential distribution using multiple dependent state
References sampling,” Quality and Quantity, vol. 49, p. 462, 2015.
[19] M. Aslam, O.-H. Arif, and C.-H. Jun, “A control chart for
[1] D. C. Montgomery and C. M. Borror, “Systems for modern gamma distribution using multiple dependent state sam-
quality and business improvement,” Quality Technology & pling,” Industrial Engineering and Management Systems,
Quantitative Management, vol. 14, no. 4, pp. 343–352, 2017. vol. 16, no. 1, pp. 109–117, 2017.
[2] J. Yu and X. Yan, “Whole process monitoring based on [20] A. Yan, S. Liu, and X. Dong, “Designing a multiple dependent
unstable neuron output information in hidden layers of deep state sampling plan based on the coefficient of variation,”
belief network,” IEEE Transactions on Cybernetics, vol. 50, SpringerPlus, vol. 5, p. 1447, 2016.
no. 9, pp. 3998–4007, 2020. [21] M. Aslam, M. A. Raza, L. Ahmad, and C. H. Jun, “Design of a
[3] J. Yu and X. Yan, “Active features extracted by deep belief t-chart using generalized multiple dependent state sampling,”
network for process monitoring,” ISA Transactions, vol. 84, Quality and Reliability Engineering International, vol. 35,
2019. no. 6, pp. 1789–1802, 2019.
[4] E. Mezzenga, V. D’Errico, A. Sarnelli et al., “Preliminary [22] M. Aslam, S. Balamurali, and C.-H. Jun, “Determination and
retrospective analysis of daily tomotherapy output constancy economic design of a generalized multiple dependent state
26 Mathematical Problems in Engineering

sampling plan,” Communication in Statistics-Simulation and

Computation, pp. 1–8. In press, 2019.
[23] N. Johnson and S. Kotz, Continuous Univariate Distributions,
Houghton Mifflin, Boston, MA, USA, 1970.
[24] W. L. Teoh, M. B. C. Khoo, and S. Y. Teh, “Optimal designs of
the median run length based double sampling X̄ chart for
minimizing the average sample size,” PLoS One, vol. 8, no. 7,
Article ID e68580, 2013.
[25] J. Anhøj and A. V. Olesen, “Run charts revisited: a simulation
study of run chart rules for detection of non-random variation
in health care processes,” PLoS One, vol. 9, no. 11, Article ID
e113825, 2014.
[26] R Core Team, A Language and Environment for Statistical
Computing, R Foundation for Statistical Computing, Vienna,
Austria, 2016, https://www.R-project.org/.
[27] D. Santos, “Beyond six sigma—a control chart for tracking
defects per billion opportunities (DPBO),” International
Journal of Industrial Engineering: Theory Applications and
Practice, vol. 16, no. 3, pp. 227–233, 2009.
[28] L. Shu, W. Huang, and W. Jiang, “A novel gradient approach
for optimal design and sensitivity analysis of EWMA control
charts,” Naval Research Logistics (NRL), vol. 61, no. 3,
pp. 223–237, 2014.
[29] M. Flores, S. Naya, R. Fernández-Casal, S. Zaragoza, P. Raña,
and J. Tarrı́o-Saavedra, “Constructing a control chart using
functional data,” Mathematics, vol. 8, no. 1, p. 58, 2020.
[30] I. A. Raji, M. H. Lee, M. Riaz, M. Abujiya, and N. Abbas,
“Outliers detection models in Shewhart control charts; an
application in photolithography: a semiconductor
manufacturing industry,” Mathematics, vol. 8, no. 5, p. 857,
2020.
[31] J.-H. Chen and S.-L. Lu, “An enhanced auxiliary information-
based EWMA-t chart for monitoring the process mean,”
Applied Sciences, vol. 10, no. 7, p. 2252, 2020.