Module – 4 (numerical solution)

NUMERICAL SOLUTIONS FOR ATTRIBUTES SAMPLING PLANS

Problem 15.1: Draw the Type-B OC curve for the single-sampling plan n = 50, c = 1.
Solution: Since the value of Lot Size (n), Producer’s Risk (α) and Consumer’s Risk (β) are
not given; hence, we will use the following formula to determine the values of Pa w.r.t. the
tabulated values of p in an iterative manner. Here, Pa is the probability of acceptance of a lot
and p is the lot fraction defective.
𝑐
𝑛!
𝑃𝑎 = 𝑃 𝑑 ≤ 𝑐 = 𝑝𝑑 1 − 𝑝 𝑛−𝑑
𝑑! 𝑛 − 𝑑 !
𝑑=0

1
50!
⇒ 𝑃𝑎 = 𝑃 𝑑 ≤ 1 = 𝑝𝑑 1 − 𝑝 50−𝑑
𝑑! 50 − 𝑑 !
𝑑=0

50! 50!
= 𝑝0 1 − 𝑝 50−0
+ 𝑝1 1 − 𝑝 50−1
0! 50 − 0 ! 1! 50 − 1 !

50! 50
50! 49
= ×1× 1−𝑝 + ×𝑝× 1−𝑝
1 × 50! 1 × 49!
50 49 49
= 1−𝑝 + 50 × 𝑝 × 1 − 𝑝 = 1−𝑝 × 1 − 𝑝 + 50𝑝

𝟒𝟗
𝑷𝒂 = 𝟏 − 𝒑 × 𝟏 + 𝟒𝟗𝒑

Table 4.1: Pa vs. p

Lot fraction Probability of Lot fraction Probability of
defective (p) Acceptance (Pa) defective (p) Acceptance (Pa)
0.000 1.00 0.030 0.56
0.001 1.00 0.040 0.40
0.002 1.00 0.050 0.28
S0.003 0.99 0.060 0.19
0.004 0.98 0.070 0.13
0.005 0.97 0.080 0.08
0.006 0.96 0.090 0.05
0.007 0.95 0.100 0.03
0.008 0.94 0.110 0.02
0.009 0.93 0.120 0.01
0.010 0.91 0.130 0.01
0.020 0.74 0.140 0.00

1
 Module – 4 (numerical solution)

Fig. 4.1: Type-B OC curve for the single sampling plan: n = 50, c = 1

Problem 15.7: A company uses the following acceptance-sampling procedure. A sample

equal to 10% of the lot is taken. If 2% or less of the items in the sample are defective, the lot
is accepted; otherwise, it is rejected. If submitted lots vary in size from 5000 to 10,000 units,
what can you say about the protection by this plan? If 0.05 is the desired LTPD, does this
scheme offer reasonable protection to the consumer?
Solution: Given,

Sample size = 10% of lot size => n = 0.01N

Acceptance number = 2% of sample size => c = 0.02n
LTPD = 0.05 or 5%
Therefore,

When N = 5000, n = 500, c = 10

When N = 10000, n = 1000, c = 20

To make the calculations handy, let’s decrease everything by 90%. [Never do this in exam]

Therefore,

When N = 500, n = 50, c=1

When N = 1000, n = 100, c=2

Now we need to plot the entire OC curve for both the sampling plans. This is because, the OC
curve gives the description of the protection provided by the plan to both the producer and the
consumer.

2
 Module – 4 (numerical solution)

The ideal OC curve looks like a Z. Probability of acceptance is 1.00 at the AQL or better lot
quality and 0.0 at the LTPD or worse lot quality. Usually the producer tries for a high
probability of acceptance at the AQL; and the consumer wants a low probability of
acceptance at the LTPD.

When N = 500, n = 50, c=1

1
50!
𝑃𝑎 = 𝑃 𝑑 ≤ 1 = 𝑝𝑑 1 − 𝑝 50−𝑑
𝑑! 50 − 𝑑 !
𝑑=0

50! 50!
⇒ 𝑃𝑎 = 𝑝0 1 − 𝑝 50−0
+ 𝑝1 1 − 𝑝 50−1
0! 50 − 0 ! 1! 50 − 1 !

50! 50
50! 49
⇒ 𝑃𝑎 = ×1× 1−𝑝 + ×𝑝× 1−𝑝
1 × 50! 1 × 49!
𝟒𝟗
⇒ 𝑷𝒂 = 𝟏 − 𝒑 𝟏 + 𝟒𝟗𝒑

Table 4.4: Pa vs. p

When N = 500, n = 50, c = 1
Lot fraction Probability of Lot fraction Probability of
defective (p) acceptance (Pa) defective (p) acceptance (Pa)
0.00 1.00 0.06 0.19
0.01 0.91 0.07 0.13
0.02 0.74 0.08 0.08
0.03 0.56 0.09 0.05
0.04 0.40 0.10 0.03
0.05 (LTPD) 0.28 0.15 0.00

Fig. 4.4: Type-B OC curve when N = 500, n = 50, c = 1

3
 Module – 4 (numerical solution)

From the above table and graph, at LTPD = 0.05, Pa = 0.28

When N = 1000, n = 100, c=2

2
100!
𝑃𝑎 = 𝑃 𝑑 ≤ 2 = 𝑝𝑑 1 − 𝑝 100−𝑑
𝑑! 100 − 𝑑 !
𝑑=0

100! 100!
⇒ 𝑃𝑎 = 𝑝0 1 − 𝑝 100−0
+ 𝑝1 1 − 𝑝 100−1
0! 100 − 0 ! 1! 100 − 0 !
100!
+ 𝑝2 1 − 𝑝 100−2
2! 100 − 0 !

100! 100!
⇒ 𝑃𝑎 = × 1 × 1 − 𝑝 100 + ×𝑝× 1−𝑝 99
1 × 100! 1 × 100!
100!
+ × 𝑝2 × 1 − 𝑝 98
2 × 100!
𝟗𝟖
⇒ 𝑷𝒂 = 𝟏 − 𝒑 𝟒𝟖𝟓𝟏𝒑𝟐 + 𝟗𝟖𝒑 + 𝟏

Table 4.5: Pa vs. p

When N = 1000, n = 100, c = 2
Lot fraction Probability of Lot fraction Probability of
defective (p) acceptance (Pa) defective (p) acceptance (Pa)
0.00 1.00 0.06 0.06
0.01 0.92 0.07 0.03
0.02 0.68 0.08 0.01
0.03 0.42 0.09 0.00
0.04 0.23 0.10 0.00
0.05 (LTPD) 0.12 0.15 0.00

4
 Module – 4 (numerical solution)

Fig. 4.5: Type-B OC curve when N = 1000, n = 100, c = 2

From the above table and graph, at LTPD = 0.05, Pa = 0.12
Since, the consumer wants a low probability of acceptance at LTPD, therefore, the sampling
plant N = 1000, n = 100, c = 2 is offering better protection to the consumer than that of the
other sampling plan: N = 500, n = 50, c = 1.

Problem 15.9: Find a single-sampling plan for which p1 = 0.01, α = 0.05, p2 = 0.10, and β =
0.10. Suppose that lots of N = 2000 are submitted. Draw the ATI curve for this plan. Draw
the AOQ curve and find the AOQL.
Solution: On the nomograph, one line is drawn connecting p1 & (1 – α) and another line
connecting p2 & β. The pint of intersection of these two lines gives the values n = 45 and
c = 1 as shown in the below figure.

5
 Module – 4 (numerical solution)

Fig. 4.6: Binomial nomograph for attributes sampling plan

Now for the single sampling plan, the formula the following formula:

For probability of acceptance of the lot is:

𝑐
𝑛!
𝑃𝑎 = 𝑑 ≤ 𝑐 = 𝑝𝑑 1 − 𝑝 𝑛−𝑑
𝑑! 𝑛 − 𝑑 !
𝑑=0

1
45!
⇒ 𝑃𝑎 = 𝑃 𝑑 ≤ 1 = 𝑝𝑑 1 − 𝑝 45−𝑑
𝑑! 45 − 𝑑 !
𝑑=0

45! 45!
⇒ 𝑃𝑎 = × 𝑝0 × 1 − 𝑝 45−0
+ × 𝑝1 × 1 − 𝑝 45−1
0! × 45 − 0 ! 1! × 45 − 1 !

𝟒𝟒
⇒ 𝑷𝒂 = 𝟏 − 𝒑 𝟏 + 𝟒𝟒𝒑

And

𝑃𝑎 𝑝 𝑁 − 𝑛 𝑃𝑎 𝑝 2000 − 45
𝐴𝑂𝑄 = = ⇒ 𝑨𝑶𝑸 = 𝟎. 𝟗𝟖𝑷𝒂 𝒑
𝑁 2000

Also,

6
 Module – 4 (numerical solution)

𝐴𝑇𝐼 = 𝑛 + 1 − 𝑃𝑎 𝑁 − 𝑛 = 45 + 1 − 𝑃𝑎 2000 − 45 ⇒ 𝑨𝑻𝑰 = 𝟒𝟓 + 𝟏𝟗𝟓𝟓 𝟏 − 𝑷𝒂

The below table shows the values of Pa, AOQ and ATI for different values of p:

Table 4.6: Pa vs. p; AOQ vs. Pa.p and ATI vs. Pa

Lot fraction Probability of
AOQ ATI
defective (p) acceptance (Pa)
0.00 1.00 0.0000 45.00
0.01 0.93 0.0091 190.92
0.02 0.77 0.0151 489.04
0.03 0.61 0.0179 812.62
0.04 0.46 0.0180 1104.66
0.05 0.33 0.0164 1345.16
0.06 0.24 0.0141 1532.40
0.07 0.17 0.0115 1672.62
0.08 0.12 0.0090 1774.60
0.09 0.08 0.0069 1847.09
0.10 0.05 0.0051 1897.62
0.15 0.01 0.0009 1988.35
0.20 0.00 0.0001 1998.96
0.25 0.00 0.0000 1999.93
0.30 0.00 0.0000 2000.00
0.35 0.00 0.0000 2000.00
0.40 0.00 0.0000 2000.00
0.45 0.00 0.0000 2000.00

From the above table, the value of AOQL = 0.0180 or 1.8%. The curves are drawn below:

Fig. 4.7: Type-B OC curve (Pa vs. p) for n = 45, c = 1

7
 Module – 4 (numerical solution)

Fig. 4.8: AOQ curve (AOQ vs. p) for n = 45, c = 1

Fig. 4.9: ATI curve (ATI vs. p) for n = 45, c = 1

Problem 15.11: Suppose that a supplier ships components in lots of size 5000. A single-
sampling plan with n = 50 and c = 2 is being used for receiving inspection. Rejected lots are
screened, and all defective items are reworked and returned to the lot.
(a) Draw the OC curve for this plan.
(b) Find the level of lot quality that will be rejected 90% of the time.
(c) Management has objected to the use of the above sampling procedure and wants to
use a plan with an acceptance number c = 0, arguing that this is more consistent with
their zero-defects program. What do you think of this?
(d) Design a single-sampling plan with c = 0 that will give a 0.90 probability of rejection
of lots having the quality level found in part (b). Note that the two plans are now

8
 Module – 4 (numerical solution)

matched at the LTPD point. Draw the OC curve for this plan and compare it to the
one for n = 50, c = 2 in part (a).
(e) Suppose that incoming lots are 0.5% nonconforming. What is the probability of
rejecting these lots under both plans? Calculate the ATI at this point for both plans.
Which plan do you prefer? Why?

Solution: (a) OC curve for the single sampling plan: n = 50 and c = 2

𝑐
𝑛!
𝑃𝑎 = 𝑑 ≤ 𝑐 = 𝑝𝑑 1 − 𝑝 𝑛−𝑑
𝑑! 𝑛 − 𝑑 !
𝑑=0

2
50!
⇒ 𝑃𝑎 = 𝑃 𝑑 ≤ 2 = 𝑝𝑑 1 − 𝑝 50−𝑑
𝑑! 50 − 𝑑 !
𝑑=0

50! 50!
⇒ 𝑃𝑎 = × 𝑝0 × 1 − 𝑝 50−0
+ × 𝑝1 × 1 − 𝑝 50−1
0! × 50 − 0 ! 1! × 50 − 1 !
50!
+ × 𝑝2 × 1 − 𝑝 50−2
2! × 50 − 2 !

𝟒𝟖
⇒ 𝑷𝒂 = 𝟏 − 𝒑 𝟏𝟏𝟕𝟔𝒑𝟐 + 𝟒𝟖𝒑 + 𝟏

The below table shows the values of Pa for different values of p:

Table 4.8: Pa vs. p

Lot fraction Probability of Lot fraction Probability of
defective (p) acceptance (Pa) defective (p) acceptance (Pa)
0.000 1.000 0.090 0.161
0.010 0.986 0.100 0.112
0.020 0.922 0.101 0.108
0.030 0.811 0.102 0.104
0.040 0.677 0.103 0.100
0.050 0.541 0.150 0.014
0.060 0.416 0.200 0.001
0.070 0.311 0.250 0.000
0.080 0.226 0.300 0.000

9
 Module – 4 (numerical solution)

Fig. 4.13: Type-B OC curve for single sampling plan: n = 50, c = 2

(b) Level of lot quality that will be rejected 90% of the time.

Rejection probability 90% means acceptance probability equal to 10%. Therefore, on the OC
curve, draw a horizontal line from Pa = 0.10. From the point where this line cuts the OC
curve, draw a vertical line. It is seen that this vertical line cuts the X-axis at around 0.103.
Therefore, the required lot quality of p = 0.103 or 10.3% for which the lot will be rejected
90% of the time. This is shown in the above figure.

(c) Efficiency of the new sampling plan given by the management: n = 50, c = 0.
𝑐
𝑛!
𝑃𝑎 = 𝑑 ≤ 𝑐 = 𝑝𝑑 1 − 𝑝 𝑛−𝑑
𝑑! 𝑛 − 𝑑 !
𝑑=0

0
50!
⇒ 𝑃𝑎 = 𝑃 𝑑 ≤ 0 = 𝑝𝑑 1 − 𝑝 50−𝑑
𝑑! 50 − 𝑑 !
𝑑=0

50! 50!
⇒ 𝑃𝑎 = × 𝑝0 × 1 − 𝑝 50−0
= ×1× 1−𝑝 50
0! × 50 − 0 ! 1 × 50!

𝟓𝟎
⇒ 𝑷𝒂 = 𝟏 − 𝒑

The below table shows the values of Pa for different values of p.

10
 Module – 4 (numerical solution)

Table 4.9: Pa vs. p

Lot fraction Probability of Lot fraction Probability of
defective (p) acceptance (Pa) defective (p) acceptance (Pa)
0.00 1.00 0.06 0.00
0.01 0.00 0.07 0.00
0.02 0.00 0.08 0.00
0.03 0.00 0.09 0.00
0.04 0.00 0.10 0.00
0.05 0.00 0.20 0.00

Fig. 4.14: Type-B OC curve for single sampling plan: n = 50, c = 0

As can be seen from the above OC curve, the plan gives full protection to the consumer as
there is no chance of any effective item get accepted. Thus, the sampling plan can be said as
consumer centric. As such, the plan does not favour the producer at all. Also it is very hard
for a producer to make all the items perfect, especially in huge mass production. Thus,

 The sampling plan is consumer centric.

 The plan is suitable if the items in the lot are very costly and their quality is very
much critical to check any high cost failure in a later stage.
 If the items in the lot are not so much costly and accepting a few defective items is not
going to pose any huge loss to the consumer, then this plan is going to impose
unnecessary pressure on the supplier by rejecting the lots.

(d) Designing a single sampling plan with c = 0 for p = 0.103 with probability of
rejection 0.90 or Pa = 0.10.

11
 Module – 4 (numerical solution)

It is given that the existing sampling plan n = 50 and c = 2 and the plan to be designed have
the same LTPD value on the OC curve. Therefore, LTPD = 0.103.

We know,
𝑐
𝑛!
𝑃𝑎 = 𝑃 𝑑 ≤ 𝑐 = 𝑝𝑑 1 − 𝑝 𝑛−𝑑
𝑑! 𝑛 − 𝑑 !
𝑑=0

0
𝑛! 𝑑 𝑛−𝑑
⇒ 0.10 = 0.103 1 − 0.103
𝑑! 𝑛 − 𝑑 !
𝑑=0

𝑛! 0 𝑛−0
𝑛! 𝑛
⇒ 0.10 = 0.103 1 − 0.103 = × 1 × 0.897
0! 𝑛 − 0 ! 1 × 𝑛!
𝑛
⇒ 0.897 = 0.10

Taking log on both sides, we get,

𝑛
log 0.10 −1.000
log 0.897 = log 0.10 ⇒ 𝑛 log 0.897 = log 0.10 ⇒ 𝑛 = =
log 0.897 −0.047

⇒ 𝒏 ≈ 𝟐𝟏

Again,
𝑐
𝑛!
𝑃𝑎 = 𝑃 𝑑 ≤ 𝑐 = 𝑝𝑑 1 − 𝑝 𝑛−𝑑
𝑑! 𝑛 − 𝑑 !
𝑑=0

0
21!
⇒ 𝑃𝑎 = 𝑃 𝑑 ≤ 0 = 𝑝𝑑 1 − 𝑝 21−𝑑
𝑑! 21 − 𝑑 !
𝑑=0

21! 21!
⇒ 𝑃𝑎 = 𝑝0 1 − 𝑝 21−0
= ×1× 1−𝑝 21
0! 21 − 0 ! 1 × 21!

𝟐𝟏
⇒ 𝑷𝒂 = 𝟏 − 𝒑

12
 Module – 4 (numerical solution)

Table 4.10: Pa vs. p

Probability of Probability of
Lot fraction acceptance (Pa) Lot fraction acceptance (Pa)
defective (p) defective (p)
n = 50, c = 2 n = 21, c = 0 n = 50, c = 2 n = 21, c = 0
0.000 1.00 1.00 0.090 0.16 0.14
0.010 0.99 0.81 0.100 0.11 0.11
0.020 0.92 0.65 0.101 0.11 0.11
0.030 0.81 0.53 0.102 0.10 0.10
0.040 0.68 0.42 0.103 0.10 0.10
0.050 0.54 0.34 0.150 0.01 0.03
0.060 0.42 0.27 0.200 0.00 0.01
0.070 0.31 0.22 0.250 0.00 0.00
0.080 0.23 0.17 0.300 0.00 0.00

Fig. 4.15: Type-B OC curves for the sampling plans: n = 50, c = 2 & n = 21, c = 0

The probabilities of acceptance offered by both the sampling plans are displayed numerically
in the above table and the same is displayed graphically in the above figure.

(e) Probability of rejection at p = 0.05 given by both the sampling plans.

When n = 50, c = 2, p = 0.05

From the OC curve, at Pa = 0.95, p = 0.016, i.e. AQL = 0.016.

The probability of acceptance,

48
𝑃𝑎 = 1 − 𝑝 1176𝑝2 + 48𝑝 + 1

13
 Module – 4 (numerical solution)

48 2
⇒ 𝑃𝑎 = 1 − 0.05 × 1176 × 0.05 + 48 × 0.05 + 1
48
⇒ 𝑃𝑎 = 0.95 × 1176 × 0.0025 + 2.40 + 1

⇒ 𝑃𝑎 = 0.08526 × 2.94 + 2.40 + 1 = 0.08526 × 6.34

⇒ 𝑷𝒂 = 𝟎. 𝟓𝟒

Therefore the required probability of rejection is (1 – 0.54) = 0.46.

When n = 21, c = 0, p = 0.05

From the OC curve, at Pa = 0.95, p = 0.002, i.e. AQL = 0.002.

The probability of acceptance,

21 21 21
𝑃𝑎 = 1 − 𝑝 = 1 − 0.05 = 0.95

⇒ 𝑷𝒂 = 𝟎. 𝟑𝟒

Therefore the required probability of rejection is (1 – 0.34) = 0.66.

The probability of acceptance (Pa) should be higher at the AQL, or the rejection probability
should be lower at the AQL. The sampling plan n = 50, c = 2 is offering higher Pa value at
the AQL than that of the other sampling plan n = 21, c = 0. Therefore, the former sampling
plan n = 50, c = 2 is more preferable.

Problem 15.12: Draw the primary and supplementary OC curves for a double-sampling plan
with n1 = 50, c1 = 2, n2 = 100, c2 = 6. If the incoming lots have fraction nonconforming
p = 0.05, what is the probability of acceptance on the first sample? What is the probability of
final acceptance? Calculate the probability of rejection on the first sample.

Solution:

Let PaI and PaII be the probabilities of acceptance on the first and second sample
respectively. Therefore,
𝑐1
𝑛1 !
𝑃𝑎𝐼 = 𝑃 𝑑1 ≤ 𝑐1 = 𝑝𝑑 1 1 − 𝑝 𝑛 1 −𝑑 1
𝑑1 ! 𝑛1 − 𝑑1 !
𝑑 1 =0

2
50!
⇒ 𝑃𝑎𝐼 = 𝑃 𝑑1 ≤ 2 = 𝑝𝑑 1 1 − 𝑝 50−𝑑 1
𝑑1 ! 50 − 𝑑1 !
𝑑 1 =0

14
 Module – 4 (numerical solution)

50! 50!
⇒ 𝑃𝑎𝐼 = 𝑝0 1 − 𝑝 50−0
+ 𝑝1 1 − 𝑝 50−1
0! 50 − 0 ! 1! 50 − 1 !
50!
+ 𝑝2 1 − 𝑝 50−2
2! 50 − 2 !

⇒ 𝑷𝑰𝒂 = 𝟏 − 𝒑 𝟒𝟖
× 𝟏𝟏𝟕𝟔𝒑𝟐 + 𝟒𝟖𝒑 + 𝟏

Calculations for PaII

(1) When d1 = 3 and d2 = 0 to 3. Therefore,

𝑃𝑎𝐼𝐼 1 = 𝑃 𝑑1 = 3 × 𝑃 𝑑2 ≤ 3

3
50! 100!
⇒ 𝑃𝑎𝐼𝐼 1 = 𝑝3 1 − 𝑝 47
× 𝑝𝑑 2 1 − 𝑝 100−𝑑 2
3! × 47! 𝑑2 ! 100 − 𝑑2 !
𝑑 2 =0

50!
⇒ 𝑃𝑎𝐼𝐼 1 = 𝑝3 1 − 𝑝 47
6 × 47!
100! 100! 1
× 𝑝0 1 − 𝑝 100 + 𝑝 1−𝑝 99
1 × 100! 1 × 99!
100! 2 100! 3
+ 𝑝 1 − 𝑝 98 + 𝑝 1 − 𝑝 97
2 × 98! 1 × 97!

⇒ 𝑃𝑎𝐼𝐼 1 = 19600𝑝3 1 − 𝑝 47

100 99
× 1−𝑝 + 100𝑝 1 − 𝑝 + 4950𝑝2 1 − 𝑝 98
+ 161700𝑝3 1 − 𝑝 97

⇒ 𝑷𝑰𝑰
𝒂 𝟏 = 𝟏𝟗𝟔𝟎𝟎𝒑𝟑 𝟏 − 𝒑 𝟒𝟕
× 𝟏−𝒑 𝟗𝟕
𝟏𝟓𝟔𝟖𝟒𝟗𝒑𝟑 + 𝟒𝟕𝟓𝟑𝒑𝟐 + 𝟗𝟕𝒑 + 𝟏

(2) When d1 = 4 and d2 = 0 to 2. Therefore,

𝑃𝑎𝐼𝐼 2 = 𝑃 𝑑1 = 4 × 𝑃 𝑑2 ≤ 2

2
50! 100!
⇒ 𝑃𝑎𝐼𝐼 2 = 𝑝4 1 − 𝑝 46
× 𝑝𝑑 2 1 − 𝑝 100−𝑑 2
4! × 46! 𝑑2 ! 100 − 𝑑2 !
𝑑 2 =0

⇒ 𝑃𝑎𝐼𝐼 2 = 230300𝑝4 1 − 𝑝 46

100! 100! 1
× 𝑝0 1 − 𝑝 100 + 𝑝 1−𝑝 99
1 × 100! 1 × 99!
100! 2
+ 𝑝 1 − 𝑝 98
2 × 98!

⇒ 𝑷𝑰𝑰
𝒂 𝟐 = 𝟐𝟑𝟎𝟑𝟎𝟎𝒑𝟒 𝟏 − 𝒑 𝟒𝟔
× 𝟏−𝒑 𝟗𝟖
𝟒𝟖𝟓𝟏𝒑𝟐 + 𝟗𝟖𝒑 + 𝟏

15
 Module – 4 (numerical solution)

(3) When d1 = 5 and d2 = 0 to 1. Therefore,

𝑃𝑎𝐼𝐼 3 = 𝑃 𝑑1 = 5 × 𝑃 𝑑2 ≤ 1

1
50! 100!
⇒ 𝑃𝑎𝐼𝐼 3 = 𝑝5 1 − 𝑝 45
× 𝑝𝑑 2 1 − 𝑝 100−𝑑 2
5! × 45! 𝑑2 ! 100 − 𝑑2 !
𝑑 2 =0

100! 100! 1
⇒ 𝑃𝑎𝐼𝐼 3 = 2118760𝑝5 1 − 𝑝 45
× 𝑝0 1 − 𝑝 100
+ 𝑝 1−𝑝 99
1 × 100! 1 × 99!

⇒ 𝑷𝑰𝑰
𝒂 𝟑 = 𝟐𝟏𝟏𝟖𝟕𝟔𝟎𝒑𝟓 𝟏 − 𝒑 𝟒𝟓
× 𝟏−𝒑 𝟗𝟗
𝟗𝟗𝒑 + 𝟏

(4) When d1 = 6 and d2 = 0. Therefore,

𝑃𝑎𝐼𝐼 4 = 𝑃 𝑑1 = 6 × 𝑃 𝑑2 = 0

50! 100!
⇒ 𝑃𝑎𝐼𝐼 4 = 𝑝6 1 − 𝑝 44
× 𝑝0 1 − 𝑝 100
6! × 44! 0! × 100!

⇒ 𝑷𝑰𝑰
𝒂 𝟒 = 𝟏𝟓𝟖𝟗𝟎𝟕𝟎𝟎𝒑𝟔 𝟏 − 𝒑 𝟒𝟒
× 𝟏−𝒑 𝟏𝟎𝟎

Hence,

𝑷𝑰𝑰 𝑰𝑰
𝒂 = 𝑷𝒂 𝟏 + 𝑷𝑰𝑰
𝒂 𝟐 + 𝑷𝑰𝑰
𝒂 𝟑 + 𝑷𝑰𝑰
𝒂 𝟒

And, final probability of lot acceptance at p = 0.05 is,

𝑷𝒂 = 𝑷𝑰𝒂 + 𝑷𝑰𝑰
𝒂

The following table gives the required probabilities:

Table 4.11: PaI & PaII vs. p

𝒑 𝑷𝑰𝒂 𝑷𝑰𝑰
𝒂 𝑷𝒂 𝑷𝑰𝑰
𝒂 𝟏 𝑷𝑰𝑰
𝒂 𝟐 𝑷𝑰𝑰
𝒂 𝟑 𝑷𝑰𝑰
𝒂 𝟒

0.000 1.000 0.000 1.000 0.000 0.000 0.000 0.000

0.010 0.986 0.013 1.000 0.012 0.001 0.000 0.000
0.020 0.922 0.063 0.985 0.052 0.010 0.001 0.000
0.030 0.811 0.104 0.915 0.082 0.019 0.003 0.000
0.040 0.677 0.103 0.780 0.079 0.021 0.003 0.000
0.050 0.541 0.075 0.616 0.057 0.016 0.002 0.000
0.060 0.416 0.045 0.461 0.033 0.010 0.002 0.000

16
 Module – 4 (numerical solution)

0.070 0.311 0.022 0.333 0.017 0.005 0.001 0.000

0.080 0.226 0.010 0.236 0.007 0.002 0.000 0.000
0.090 0.161 0.004 0.165 0.003 0.001 0.000 0.000
0.100 0.112 0.002 0.113 0.001 0.000 0.000 0.000
0.150 0.014 0.000 0.014 0.000 0.000 0.000 0.000
0.200 0.001 0.000 0.001 0.000 0.000 0.000 0.000
0.250 0.000 0.000 0.000 0.000 0.000 0.000 0.000
0.300 0.000 0.000 0.000 0.000 0.000 0.000 0.000

Fig. 4.16: Type-B OC curve, Primary & Supplementary OC curve

Problem 15.13:
(a) Derive an item-by-item sequential-sampling plan for which:
p1 = 0.01, α = 0.05, p2 = 0.10, and β = 0.10.
(b) Draw the OC curve for this plan.

Solution: The equations for the acceptance and rejection line for specified values of p1, 1 – α,
p2 and β are:

𝑋𝐴 = −ℎ1 + 𝑠𝑛 acceptance line

𝑋𝐵 = ℎ2 + 𝑠𝑛 rejection line

17
 Module – 4 (numerical solution)

Here,

𝑝2 1 − 𝑝1 0.10 × 1 − 0.01 0.099

𝑘 = log = log = log = log 11 ⇒ 𝒌 = 𝟏. 𝟎𝟒𝟏
𝑝1 1 − 𝑝2 0.01 × 1 − 0.10 0.009

1−𝛼 1 − 0.05
ℎ1 = log 𝑘 = log 1.041 = log 9.5 1.041 ⇒ 𝒉𝟏 = 𝟎. 𝟗𝟑𝟗
𝛽 0.10

1−𝛽 1 − 0.10
ℎ2 = log 𝑘 = log 1.041 = log 180 1.041 ⇒ 𝒉𝟐 = 𝟐. 𝟏𝟔𝟔
𝛼 0.05

𝑠 = log 1 − 𝑝1 1 − 𝑝2 𝑘 = log 1 − 0.01 1 − 0.10 1.041

⇒ 𝑠 = log 0.99 0.90 1.041 = log 1.1 1.041 ⇒ 𝒔 = 𝟎. 𝟎𝟒𝟎

Therefore,

𝑋𝐴 = −ℎ1 + 𝑠𝑛 acceptance line

⇒ 𝑿𝑨 = −𝟎. 𝟗𝟑𝟗 + 𝟎. 𝟎𝟒𝟎𝒏

And

𝑋𝐵 = ℎ2 + 𝑠𝑛 rejection line

⇒ 𝑿𝑩 = 𝟐. 𝟏𝟔𝟔 + 𝟎. 𝟎𝟒𝟎𝒏

The below table shows the values of acceptance number and rejection number for different
values of n based on the equation of XA and XB. In the table, “a” means “Acceptance not
possible”; and “b” means “Rejection not possible”.

Table 4.12: Values of XA and XB

Number of items Acceptance Rejection
𝑿𝑨 𝑿𝑩
inspected (n) Number (c) Number (r)
1 -0.9 a 2.2 b
2 -0.9 a 2.2 2
3 -0.8 a 2.3 2
4 -0.8 a 2.3 2
5 -0.7 a 2.4 2
6 -0.7 a 2.4 2
7 -0.7 a 2.4 2
8 -0.6 a 2.5 3
9 -0.6 a 2.5 3
10 -0.5 a 2.6 3

18
 Module – 4 (numerical solution)

11 -0.5 a 2.6 3
12 -0.5 a 2.6 3
13 -0.4 a 2.7 3
14 -0.4 a 2.7 3
15 -0.3 a 2.8 3
16 -0.3 a 2.8 3
17 -0.3 a 2.8 3
18 -0.2 a 2.9 3
19 -0.2 a 2.9 3
20 -0.1 a 3.0 3
21 -0.1 a 3.0 3
22 -0.1 a 3.0 3
23 0.0 0 3.1 3
24 0.0 0 3.1 3
25 0.1 0 3.2 3
26 0.1 0 3.2 3
27 0.1 0 3.2 3
28 0.2 0 3.3 3
29 0.2 0 3.3 3
30 0.3 0 3.4 3
31 0.3 0 3.4 3
32 0.3 0 3.4 3
33 0.4 0 3.5 4
34 0.4 0 3.5 4
35 0.5 0 3.6 4
36 0.5 0 3.6 4
37 0.5 0 3.6 4
38 0.6 0 3.7 4
39 0.6 0 3.7 4
40 0.7 0 3.8 4
41 0.7 0 3.8 4
42 0.7 0 3.8 4
43 0.8 0 3.9 4
44 0.8 0 3.9 4
45 0.9 0 4.0 4
46 0.9 0 4.0 4
47 0.9 0 4.0 4
48 1.0 1 4.1 4
49 1.0 1 4.1 4
50 1.1 1 4.2 4

19
 Module – 4 (numerical solution)

Items have been drawn one at a time till the acceptance number becomes equal to 1.
Therefore, from the above table, the required sampling plan is:

n = 48, c = 1, r = 4

Fig. 4.17: Graphical performance of the item-by-item sequential sampling

(a) The OC curve

𝑃𝑎 = ℎ2 ℎ1 + ℎ2 = 2.166 0.939 + 2.166 = 2.166 3.105 ⇒ 𝑷𝒂 = 𝟎. 𝟕𝟎

Therefore, the required three points to form the OC curve are:

 (p1 , 1 – α) = (0.01 , 1 – 0.05) = (0.01 , 0.95)

 (p2 , β) = (0.10 , 0.10)
 p = s = 0.040

Problem 15.16: A supplier ships a component in lots of size N = 3000. The AQL has been
established for this product at 1%. Find the normal, tightened, and reduced single-sampling
plans for this situation from MIL STD 105E, assuming that general inspection level II is
appropriate.
Solution: Given,

N = 3000 ; AQL = 1% = 0.010

Initially, we will start with Normal Inspection with Inspection Level-II.

 From Table 15.4, appropriate sample size code letter is K.

 From Table 15.5, consider the row for sample size code K and the column of AQL
0.01, a downward arrow is passing through the intersection cell. Move downward till

20
 Module – 4 (numerical solution)

the arrowhead and we see that the sample size code letter changes to Q and the
corresponding sampling plan for normal inspection becomes: n = 1250, c = 0, r = 1.
 From Table 15.6, the single-sampling plan for tightened inspection is n = 2000, c =
0, r = 1.
 From Table 15.7, the single-sampling plan for reduced inspection is n = 500, c = 0, r
= 1.

Problem 15.19: MIL STD 105E is being used to inspect incoming lots of size N = 5000.
Single sampling, general inspection level II, and an AQL of 0.65% are being used.
(a) Find the normal, tightened, and reduced inspection plans.
(b) Draw the OC curves of the normal, tightened, and reduced inspection plans on the
same graph.

Solution: Given, N = 5000 ; AQL = 6.5% = 0.065

Initially, we will start with Normal Inspection with Inspection Level-II.

 From Table 15.4, appropriate sample size code letter is L.

 From Table 15.5, consider the row for sample size code L and the column of AQL
0.10, intersection cell gives the sampling plan for normal inspection becomes: n =
200, c = 0, r = 1.
 From Table 15.6, the single-sampling plan for tightened inspection is n = 315, c = 0,
r = 1.
 From Table 15.7, the single-sampling plan for reduced inspection is n = 80, c = 0, r
= 1.

Problem 15.20: A product is shipped in lots of size N = 2000. Find a Dodge–Romig single-
sampling plan for which the LTPD = 1%, assuming that the process average is 0.25%
defective. Draw the OC curve and the ATI curve for this plan. What is the AOQL for this
sampling plan?
Solution: Given: N = 2000 ; LTPD = 1% = 0.010 ; p = 0.25%

From Table 15.9 which is the Dodge–Romig Single-Sampling Table for Lot Tolerance
Percent Defective (LTPD) = 1.0%, we have the following single-sampling plan,

n = 490, c = 2

AOQL = 0.21% assuming 100% inspection for the rejected lots and the defective items are
replaced with good ones.

21
 Module – 4 (numerical solution)

The probability of acceptance for this sampling plan is:

𝑐
𝑛!
𝑃𝑎 = 𝑃 𝑑 ≤ 𝑐 = 𝑝𝑑 1 − 𝑝 𝑛−𝑑
𝑑! 𝑛 − 𝑑 !
𝑑=0

2
490!
⇒ 𝑃𝑎 = 𝑃 𝑑 ≤ 2 = 𝑝𝑑 1 − 𝑝 490−𝑑
𝑑! 490 − 𝑑 !
𝑑=0

490! 490!
⇒ 𝑃𝑎 = 𝑝0 1 − 𝑝 490−0
+ 𝑝1 1 − 𝑝 490−1
0! 490 − 0 ! 1! 490 − 1 !
490!
+ 𝑝2 1 − 𝑝 490−2
2! 490 − 2 !

490! 490
490! 489
490!
⇒ 𝑃𝑎 = ×1× 1−𝑝 + 𝑝 1−𝑝 + 𝑝2 1 − 𝑝 488
1 × 490! 1 × 489! 2 × 488!
𝟒𝟖𝟖
⇒ 𝑷𝒂 = 𝟏 − 𝒑 × 𝟏𝟏𝟗𝟑𝟏𝟔𝒑𝟐 + 𝟒𝟖𝟖𝒑 + 𝟏

And

𝐴𝑇𝐼 = 𝑛 + 1 − 𝑃𝑎 𝑁 − 𝑛

⇒ 𝐴𝑇𝐼 = 490 + 1 − 𝑃𝑎 2000 − 490 ⇒ 𝑨𝑻𝑰 = 𝟒𝟗𝟎 + 𝟏𝟓𝟏𝟎 𝟏 − 𝑷𝒂

Table 4.14: Pa vs. p and ATI vs. Pa

Lot fraction Probability of Lot fraction Probability of
ATI ATI
defective (p) acceptance (Pa) defective (p) acceptance (Pa)
0.000 1.000 490.000 0.040 0.000 1999.999
0.010 0.132 1800.632 0.045 0.000 2000.000
0.020 0.003 1995.383 0.050 0.000 2000.000
0.030 0.000 1999.935 0.055 0.000 2000.000
0.035 0.000 1999.993 0.060 0.000 2000.000

22
 Module – 4 (numerical solution)

Fig. 4.19: Type-B OC curve for single sampling plan: n = 490, c = 2

Fig. 4.20: ATI curve for single sampling plan: n = 490, c = 2

Problem 15.22: A supplier ships a product in lots of size N = 8000. We wish to have an
AOQL of 3%, and we are going to use single sampling. We do not know the supplier’s
process fallout but suspect that it is at most 1% defective.
(a) Find the appropriate Dodge–Romig plan.
(b) Find the ATI for this plan, assuming that incoming lots are 1% defective.
(c) Suppose that our estimate of the supplier’s process average is incorrect and that it is
really 0.25% defective. What sampling plan should we have used? What reduction in
ATI would have been realized if we had used the correct plan?

Solution: Given: N = 8000 ; AOQL = 3% ; p ≤ 1%

23
 Module – 4 (numerical solution)

(a) Appropriate Dodge-Roming Plan:

In Table 15.8, select the row for 7,001 – 10,000 (since p ≤ 1%) and column for p = 0.07 – 0
.60%, the required sampling plan is:

n = 46, c = 2 and LTPD = 11.6%

(b) Calculation of ATI at p = 1% or 0.01

𝑐
𝑛!
𝑃𝑎 = 𝑃 𝑑 ≤ 𝑐 = 𝑝𝑑 1 − 𝑝 𝑛−𝑑
𝑑! 𝑛 − 𝑑 !
𝑑=0

2
46!
⇒ 𝑃𝑎 = 𝑃 𝑑 ≤ 2 = 𝑝𝑑 1 − 𝑝 46−𝑑
𝑑! 46 − 𝑑 !
𝑑=0

46! 46!
⇒ 𝑃𝑎 = 𝑝0 1 − 𝑝 46−0
+ 𝑝1 1 − 𝑝 46−1
0! 46 − 0 ! 1! 46 − 1 !
46!
+ 𝑝2 1 − 𝑝 46−2
2! 46 − 2 !

46! 46
46! 45
46!
⇒ 𝑃𝑎 = 1−𝑝 + 𝑝 1−𝑝 + 𝑝2 1 − 𝑝 44
1 × 46! 1 × 45! 2 × 44!
𝟒𝟒
⇒ 𝑷𝒂 = 𝟏 − 𝒑 × 𝟗𝟗𝟎𝒑𝟐 + 𝟒𝟒𝒑 + 𝟏

And

𝐴𝑇𝐼 = 𝑛 + 1 − 𝑃𝑎 𝑁 − 𝑛

⇒ 𝐴𝑇𝐼 = 46 + 1 − 𝑃𝑎 8000 − 46 ⇒ 𝑨𝑻𝑰 = 𝟒𝟔 + 𝟕𝟗𝟓𝟒 𝟏 − 𝑷𝒂

Table 4.16: Pa vs. p and ATI vs. Pa

Lot fraction Probability of Lot fraction Probability of
ATI ATI
defective (p) acceptance (Pa) defective (p) acceptance (Pa)
0.00 1.00 46.00 0.07 0.37 5084.35
0.01 0.99 133.66 0.08 0.28 5797.48
0.02 0.94 557.73 0.09 0.20 6372.02
0.03 0.84 1313.76 0.10 0.15 6819.82
0.04 0.78 1775.03 0.15 0.02 7813.65
0.04 0.72 2266.63 0.20 0.00 7978.61
0.05 0.66 2771.50 0.25 0.00 7998.13
0.05 0.59 3275.13 0.30 0.00 7999.87
0.06 0.53 3765.86 0.30 0.00 7999.87
0.06 0.47 4234.83 0.30 0.00 7999.87

24
 Module – 4 (numerical solution)

Fig. 4.23: Type-B OC curve for single sampling plan: n = 46, c = 2

Fig. 4.24: ATI curve for single sampling plan: n = 46, c = 2

25
 Module – 4 (numerical solution)

NUMERICAL SOLUTIONS FOR VARIABLES SAMPLING PLANS

Problem 16.1: The density of a plastic part used in a cellular telephone is required to be at
least 0.70 g/cm3. The parts are supplied in large lots, and a variables sampling plan is to be
used to sentence the lots. It is desired to have p1 = 0.02, p2 = 0.10, α = 0.10, and β = 0.05. The
variability of the manufacturing process is unknown but will be estimated by the sample
standard deviation.
(a) Find an appropriate variables sampling plan, using Procedure 1.
(b) Suppose that a sample of the appropriate size was taken, and x̅ = 0.73, s = 1.05 × 10−2.
Should the lot be accepted or rejected?
(c) Sketch the OC curve for this sampling plan. Find the probability of accepting lots that
are 5% defective.

Solution:

(a) Appropriate variables sampling plan, using Procedure 1

Only lower value on the density is given as 0.70 g/cm3. Therefore, LSL = 0.70 g/cm3

On the nomograph for variables sampling plan, join p1 = 0.02, (1 – α) = 0.90; and p2 = 0.10, α
= 0.10, β = 0.05. The point of intersection gives the required values of n and k for the
sampling plan as:

n = 32 ; k = 1.6 (for σ unknown case)

(b) Lot Disposition

Given: x̅ = 0.73 ; s = 1.05 × 10−2.

Therefore,

𝑥 − 𝐿𝑆𝐿 0.73 − 0.70 0.03

𝑍𝐿𝑆𝐿 = = = ⇒ 𝒁𝑳𝑺𝑳 = 𝟐. 𝟖𝟔
𝜎 1.05 × 10−2 1.05 × 10−2

Since 𝒁𝑳𝑺𝑳 = 𝟐. 𝟖𝟔 > 𝑘 = 1.6, therefore accept the lot.

(c) Calculation for Pa at p = 0.05

 On the nomograph, from any point on the left axis draw a line passing through the
point of intersection and extend the line till it intercepts the right axis. The intercept

26
 Module – 4 (numerical solution)

on the right axis gives the probability of acceptance (Pa) for that particular value of p
on the left axis.
 In this way, for different values of p, find out the corresponding the values of Pa and
then draw the OC curve (Pa vs. p).

Fig. 4.25: Nomograph for designing variables sampling plan

--- 000 ---

27