Unit 7 Single Sampling Plans: Structure
Unit 7 Single Sampling Plans: Structure
Unit 7 Single Sampling Plans: Structure
Structure
7.1 Introduction
Objectives
7.2 Single Sampling Plan
7.3 Operating Characteristics (OC) Curve
7.4 Producer’s Risk and Consumer’s Risk
7.5 Average Outgoing Quality (AOQ)
7.6 Average Sample Number (ASN) and Average Total Inspection (ATI)
7.7 Design of Single Sampling Plans
Stipulated Producer’s Risk
Stipulated Consumer’s Risk
Stipulated Producer’s Risk and Consumer’s Risk
Larson Binomial Nomograph
7.8 Summary
7.9 Solutions/Answers
7.1 INTRODUCTION
In Units 5 and 6, you have learnt the various features of a sampling inspection
plan such as AQL, LTPD, producer’s risk, consumer’s risk, OC curve, ASN,
ATI, etc. The same features characterise different types of sampling plans. In
Unit 5, you have also learnt that the main types of acceptance sampling plans
for attributes are:
i) Single sampling plan,
ii) Double sampling plan,
iii) Multiple sampling plan, and
iv) Sequential sampling plan
In this unit, we focus on the single sampling plans for attributes. In Sec. 7.2,
we explain the single sampling plan and its implementation. We describe
various features of the single sampling plan such as the operating characteristic
(OC) curve, producer’s risk, consumer’s risk, average sample number (ASN)
and average total inspection (ATI) in Secs. 7.3 to 7.6. Finally, we describe the
design of the single sampling plans in Sec. 7.7. In the next unit, we shall
discuss the double sampling plans for attributes.
Objectives
After studying this unit, you should be able to:
describe a single sampling plan;
compute the probability of accepting or rejecting a lot of given incoming
quality in a single sampling plan;
construct the operating characteristics (OC) curve of a single sampling
plan;
compute the consumer’s risk and producer’s risk in a single sampling plan;
1
Product Control compute the average sample number (ASN) and the average total
inspection (ATI) for a single sampling plan; and
design single sampling plans.
2
Single Sampling Plans
For rapid calculation, we can use Table I for obtaining this probability.
From Table I, for n = 12, x = c = 1 and p = 0.01, we have
5
Product Control 1
n n x
Cx p x 1 p 0.9938
x 0
6
If X represents the number of defective candles in the sample, the consumer Single Sampling Plans
accepts the lot if X ≤ c, i.e., X ≤ 1. Therefore, the probability of accepting the
lot is given by
1
Pa p P X c P X 1 P X x
x 0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20
Proportion Defective (p)
1 Pa p1 … (6)
We can compute Pa p1 from equation (3) by replacing the quality level p with
p1 as follows:
c c Np1
C x N Np1 C n x
Pa p1 P X x N
x 0 x 0 Cx
Therefore, from equation (6), the producer’s risk is given by
c Np1
C x N Np1 Cn x
Pp 1 Pa p1 1 N
… (7)
x 0 Cx
For rapid calculation of the producer’s risk for a single sampling plan, we can
also use approximations as we have discussed in Sec. 7.4. Therefore, if we use
the approximation of the hypergeometric distribution to the binomial
distribution with parameters n and p1, the producer’s risk is given by
c
nx
Pp 1 Pa p1 1 n C x p1x 1 p1 … (8)
x 0
8
We can compute the consumer’s risk for the single sampling plan from Single Sampling Plans
equation (4) by replacing the quality level p with p 2 as follows:
c c Np 2
Cx N Np2 Cn x
Pc Pa p 2 P X x N
… (9)
x0 x 0 Cx
If we approximate the hypergeometric distribution to the binomial distribution
with parameters n and p 2, the consumer’s risk is given by
c
nx
Pc Pa p 2 n Cx p 2x 1 p 2 … (10)
x 0
9
Product Control For practice you can also compute the producer’s risk and consumer’s risk in
the following exercise.
E3) Suppose in E2, the acceptance quality level (AQL) and lot tolerance
percent defective (LTPD) are 0.04 and 0.10, respectively. Calculate the
producer’s risk and consumer’s risk for this plan.
ii) If the lot is rejected, all units of the lot go for 100% inspection and all
defective units found in the lot are replaced by non-defective units. So there
is no defective unit at the outgoing stage. The probability that the lot will be
rejected is (1 – Pa). Therefore, the expected number of defective units per
lot at the outgoing stage is 0 1 Pa 0.
Thus, the expected number of defective units per lot after sampling inspection
is p N n Pa 0 p N n Pa .
Hence, the average proportion defective in the outgoing stage or average
outgoing quality (AOQ) is given by
Number of defective units in the lot after the inspection
AOQ
Total number of units in the lot
p N n Pa
or AOQ … (12)
N
If the sample size n is very small in proportion to the lot size N, i.e., n / N 0,
equation (12) for AOQ becomes
n
AOQ p 1 Pa pPa … (13)
N
10
Let us now discuss the construction of the AOQ curve for a single sampling Single Sampling Plans
plan.
As you know, the AOQ curve is constructed by taking the quality level
(proportion defective) on the X-axis and the AOQ on the Y-axis. So for
constructing the AOQ curve, we first consider different quality levels such as
p = 0.01, 0.02, 0.03… and then calculate the corresponding AOQ.
Let us consider some examples for calculating AOQ and constructing the AOQ
curve.
Example 5: Suppose in Example 3 the submitted lot quality is p = 0.02. The
rejected lots are screened and all defective candles are replaced by the
non-defective candles. Calculate the average outgoing quality (AOQ) for this
plan.
Solution: The submitted lot quality is p = 0.02 and we have to calculate the
AOQ for this single sampling plan.
It is given that
N 500, n 20, c 1and p 0.02
From equation (12), the AOQ for the single sampling plan is
p N n Pa
AOQ
N
where Pa is the probability of accepting the lot of quality p. Therefore, for
calculating AOQ, we have to calculate Pa. We have already calculated this
probability in Example 3. So we directly use the result:
Pa 0.9401
We now construct the AOQ curve by taking the quality level (proportion
defective) on the X-axis and the corresponding AOQ values on the Y-axis as
shown in Fig.7.3.
0.025
Average Outgoing Quality(AOQ)
0.02
0.015
0.01
0.005
0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
Proportion Defective (p)
Fig. 7.3: The AOQ curve for Example 5.
You may now like to calculate the AOQ and construct the AOQ curve. Try the
following exercises.
E4) Assuming that the lot size is large relative to the sample size, calculate
the approximate average outgoing quality (AOQ) for the single sampling
plan with n = 10 and c = 0 containing 20% defective units.
E5) A computer manufacturer purchases computer chips from a company in
lots of 200. Twelve computer chips are sampled at random and inspected
for defects. The computer manufacturer accepts the lot if the inspected
sample contains at most one defective chip. Otherwise, he/she rejects the
lot. If the rejected lots are screened and all defective computer chips are
replaced by non-defective chips. Construct the AOQ curve for this plan.
You have learnt about the OC curve, producer’s risk, consumer’s risk and
AOQ for a single sampling plan. We now discuss ASN and ATI of the plan.
ATI n Pa N 1 Pa
This can also be written as
ATI n 1 Pa N n … (15)
The curve drawn between ATI and lot quality (p) is known as the ATI curve.
Let us take up an example to illustrate this concept.
Example 7: Calculate the ASN for the plan given in Example 4. If the rejected
lots are screened and all defective tyres are replaced by non-defective tyres,
construct the ATI curve for this plan.
Solution: It is given that
N 400, n 15, c 0
In the first case, the decision of acceptance or rejection of the lot is taken only
on a single sample of size n = 15. Therefore, the ASN for this plan is simply
the sample size, i.e., ASN = n = 15.
For construction of the ATI curve, we first calculate the probabilities of
accepting the lot corresponding to different quality levels using Table I. Then
we calculate the ATI for each quality level using equation (15).
We have already calculated these probabilities in Example 6 and we use those
results. Substituting the values of N, n, p and Pa in equation (15) we can
calculate ATI.
The probabilities of accepting the lot and the ATIs corresponding to different
quality levels are given in the following table:
Probability of
Incoming
Accepting the ATI
Lot Quality
Lot
0 1 15.00
0.01 0.8601 68.86
0.02 0.7386 115.64
0.04 0.5421 191.29
0.06 0.3953 247.81
0.08 0.2863 289.77
0.10 0.2059 320.73
0.12 0.1470 343.41
0.14 0.1041 359.92
0.16 0.0731 371.86
0.18 0.0510 380.37
0.20 0.0352 386.45
We now construct the ATI curve by taking the quality level (proportion
defective) on the X-axis and the corresponding ATI values on the Y-axis as
shown in Fig.7.5.
14
Single Sampling Plans
400
350
Average Total Inspection (ATI)
300
250
200
150
100
50
0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
Proportion Defective (p)
16
0.08 0.5719 0.3526 0.0445 Single Sampling Plans
0.10 0.4503 0.2086 0.0075
0.12 0.3460 0.1156 0.0010
0.14 0.2602 0.0607 0.0001
0.16 0.1920 0.0303 0.0000
0.18 0.1391 0.0145 0.0000
0.20 0.0991 0.0066 0.0000
1
Probability of Accepting the lot Pa(p)
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
Proportion Defective(p)
18
The values of R corresponding to various acceptance number (c) and α and β Single Sampling Plans
are also listed in Table III. We choose a value of R from Table III which is
exactly equal to the desired value of R corresponding to the desired α and β.
Generally, the tabulated value of R is not equal to the desired value of R. So in
such situations, we take the tabulated values of R between which the desired
value of R lies. Then we look up the corresponding values of acceptance
number (c) in Table III and find the value of sample size (n). The following
two approaches are used to find n:
1. Satisfy Producer’s Risk Stipulation Exactly and come Close to
Consumer’s Risk
According to this approach, we find the values of n as we have discussed in
Sec. 7.7.1. It means that we first find the value of np1 corresponding to c
and α with the help of Table III. Then we find n from equation (16). Thus,
np1
n
p1
The values of n are obtained for both values of c. In this way, we get two
sampling plans which satisfy producer’s risk exactly.
Out of these, we choose the sampling plan which is close to satisfying the
consumer’s risk. For that we find the value of p2 for each plan. We find the
values of np2 corresponding to the desired β and each c from Table III.
Then we find the value of p2 for each plan as follows:
np 2
p2 … (19)
n
We choose the sampling plan for which the calculated p2 is closer to the
desired p 2.
Another criterion for choosing the sampling plan is that we select the
sampling plan which has the smallest sample size in order to minimize
inspection costs. Alternatively, we can choose the sampling plan which has
the largest sample size in order to get maximum information.
2. Satisfy Consumer’s Risk Stipulation Exactly and come Close to
Producer’s Risk
According to this approach, we find the values of n as we have discussed in
Sec. 7.7.2. It means that we first find the value of np2 corresponding to c
and β with the help of Table III. Then we find n from equation (17). Thus,
np 2
n
p2
The values of n are obtained for both values of c. In this way, we get two
sampling plans which satisfy consumer’s risk exactly.
Out of these, we choose the sampling plan which is close to satisfying the
producer’s risk. For that we find the value of p 1 for each plan. We find the
values of np1 corresponding to the desired α and each c from Table III.
Then we find the value of p1 for each plan as follows:
np1
p1 … (20)
n
19
Product Control We choose the sampling plan for which the calculated p1 is closer to the
desired p1.
Another criterion for choosing the sampling plan is that we select the
sampling plan which has the smallest sample size in order to minimize
inspection costs. Alternatively, we can choose the sampling plan which has
the largest sample size in order to get maximum information.
We now describe this procedure with the help of an example.
Example 10: Suppose, in Example 8, the supplier and the quality control
inspector decide the acceptance quality level (AQL) to be 2% and the lot
tolerance percent defective (LTPD) to be 8%. Design a sampling plan which
ensures that lots of quality 2% will be rejected 5% of the time and lots of
quality 8% will be accepted 5% of the time.
Solution: It is given that
AQL p1 2% 0.02 and 5% 0.05
21
Product Control
Fig. 7.8
From Fig. 7.8 we have, n = 135 and c = 5.
You can now check your understanding of how to design a single sampling
plan by answering the following exercises.
E7) A ball bearing supplier and an automobile company have decided to
check the quality of ball bearings in lots of size 1000 with acceptance
quality level (AQL) as 1%. Design the single sampling plans using c = 2,
4 and 6 which ensure that lots of quality 1% will be rejected 1% of the
time.
E8) A consumer receives lots of 5000 candles from a new supplier. To check
the quality of lots, the consumer and supplier want to use the single
sampling plan which satisfies a consumer’s risk of 5% for lots of quality
5% . Determine sampling plans for the specified consumer’s risk and
LTPD for acceptance number c = 3 and 6.
We end this unit by giving a summary of what we have covered in it.
7.8 SUMMARY
1. The main acceptance sampling plans for attributes are:
i) Single sampling plan,
ii) Double sampling plan,
Product Control iii) Multiple sampling plan, and
iv) Sequential Sampling Plan.
2. A sampling plan in which a decision about the acceptance or rejection of a
lot is based on a single sample that has been inspected is known as a single
sampling plan. There are two parameters of a single sampling plan:
n – size of the sample, and
c – acceptance number for the sample.
3. In a single sampling plan, if number of defective units (d) in the sample is
less than or equal to the stated acceptance number (c), i.e., if d ≤ c, we
accept the lot and if d > c, we reject the lot under acceptance sampling
plan.
4. In a single sampling plan, if d ≤ c, we accept the lot and replace all
defective units found in the sample by non-defective units and if d > c, we
accept the lot by inspecting the entire lot and replacing all defective units in
the lot by non-defective units under rectifying sampling plan.
5. The probability of accepting a lot of quality p for a single sampling plan is
given by
c
n x
Pa p P X c n C x p x 1 p
x 0
6. The produce’s risk and consumer’s risk for a single sampling plan are given
by
c
n x
Pp 1 Pa p1 1 n C x p1x 1 p x and
x 0
c
nx
Pc Pa p 2 n C x p 2x 1 p 2
x 0
7.9 SOLUTIONS/ANSWERS
E1) To check the quality of the lots, the buyer randomly draws 50 silicon
chips from each lot. After that he/she inspects each and every chip
drawn from the lot for certain defects and classifies each chip of the
sample as defective or non-defective. At the end of the inspection,
he/she counts the number of defective chips (d) found in the sample and
then compares the number of defective chips (d) with the acceptance
number (c). If d ≤ c = 2, he/she accepts the lot and if d > c = 2, he/she
rejects the lot on the basis of the inspected sample. It means that if the
24
buyer finds 0 or 1 or 2 defective chips in the sample, he/she accepts the Single Sampling Plans
lot. Otherwise, he/she rejects the lot.
E2) It is given that
N 2000, n 25, c 2
For constructing the OC curve, we have to calculate the probabilities of
accepting the lot corresponding to different quality levels.
If X represents the number of defective syringes in the sample, the
quality inspector accepts the lot if X ≤ c = 2. Therefore, the probability
of accepting the lot is given by
2
Pa p P X c P X 2 P X x
x 0
Since N 10n, we use the binomial distribution. We can use Table I for
calculating the probabilities of accepting the lot corresponding to
different quality levels such as p = 0.01, 0.02, 0.03… .These
probabilities are shown in the table given below:
Incoming Lot Probability of Accepting
Quality the Lot
0 1
0.01 0.9980
0.02 0.9868
0.04 0.9235
0.06 0.8129
0.08 0.6768
0.10 0.5371
0.12 0.4088
0.14 0.3000
0.16 0.2130
0.18 0.1467
0.20 0.0982
25
Product Control 1
We can use Table I for calculating the probabilities of accepting the lot
corresponding to different quality levels such as p = 0.01, 0.02, 0.03…
Then we calculate AOQ for each quality level by using equation (12).
The probabilities of accepting the lot and AOQs corresponding to
different quality levels are given in the following table:
Incoming Probabilty of AOQ
Lot Quality Accepting the Lot
0 1.0000 0
0.01 0.9938 0.0093
0.02 0.9769 0.0184
27
Product Control 0.04 0.9191 0.0346
0.06 0.8405 0.0474
0.08 0.7513 0.0565
0.10 0.6590 0.0619
0.12 0.5686 0.0641
0.14 0.4834 0.0636
0.16 0.4055 0.0610
0.18 0.3359 0.0568
0.20 0.2749 0.0517
0.05
0.04
0.03
0.02
0.01
0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
Proportion Defective(p)
We now construct the ATI curve by taking the quality level (proportion
defective) on the X-axis and the corresponding ATI values on the
Y-axis as shown in Fig.7.11.
140
Average Total Inspection (ATI)
120
100
80
60
40
20
0
0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 0.18 0.2
Proportion Defective (p)
30