Académique Documents
Professionnel Documents
Culture Documents
A “Lot” of
goods
z Purposes
¾ Determine the quality level of an incoming shipment
or at the end of production
¾ Judge whether quality level is within the level that has
been predetermined
Acceptance
Acceptance Sampling:
Sampling: the the inspection
inspection and
and classification
classification ofof aa
sample
sample of
of nits
nits selected
selected at
at random
random from
from aa larger
larger batch
batch or
or lot
lot
and
and ultimate
ultimate decision
decision about
about disposition
disposition of
of the
the lot
lot ––
Lot
Lot Disposition
Disposition or
or Lot
Lot Sentencing
Sentencing
Whenparts
When partsare
arereceived
received
Two common points of inspection
Afterproduction
After production
Copyright Tapan P Bagchi 5
Sampling Plans
z Acceptance sampling: Form of inspection applied
to lots or batches of items before or after a
process, to judge conformance with
predetermined standards
z Sampling plans: Plans that specify lot size,
sample size, number of samples, and
acceptance/rejection criteria
¾ Single-sampling
¾ Double-sampling
¾ Multiple-sampling and Sequential sampling
Count
N Accept or
n Number
(Lot) Conforming Reject Lot
1. Single
1. Single sampling
sampling
• For low cost or low impact material on the subsequent process
• From trusted supplier
3. Double/multiple
3. Double/multiple sampling
sampling
• Inspect every item in the lot, then remove the defective units by
discarding, reworking, or returning to suppliers
• For critical process. Defective input may result high failure cost
• From new supplier
Copyright Tapan P Bagchi 16
Types of Sampling Plan
SamplingPlan
Sampling Planfor
forAttributes
Attributes SamplingPlan
Sampling Planfor
forVariables
Variables
z The most common and easiest plan to use but not most
efficient in terms of average number of samples needed
z One sample drawn from the lot and 100% inspected
z Single sampling plan
N = lot size
n = sample size (randomized)
c = acceptance number
d = number of defective items in sample
z Rule: If d ≤ c, accept lot; else reject the lot
Yes
d≤c? Accept lot
No
Reject lot
Sample
exactly n parts
N = 10,000
n = 89 and inspect
every part
Copyright Tapan P Bagchi 23
Producer’s & Consumer’s Risks
due to mistaken sentencing of lot
z Producer’s risk
¾ Risk associated with a lot of acceptable quality rejected
z Alpha α
= Prob (committing Type I error)
= P(rejecting lot at AQL quality level)
= producers risk
z Beta β
= Prob (committing Type II error)
= Prob (accepting a lot at RQL quality level)
= consumers risk
The Operating Characteristic (OC) curve for a
sampling plan quantifies these risks
Yes
d≤c? Accept lot
No
Reject lot
75%
50%
25%
RQL
AQL
Go
β = 0.10 od Indifferent Bad
.03 .06 .09
Copyright Tapan P Bagchi 30
Lot Quality (Fraction Defective)
OC Curve Terms
z Acceptable Quality Level (AQL)
¾ Percentage of defective items a customer is willing
to accept from you (a property of mfg. process)
z Lot Tolerance Percent Defective (LTPD)
¾ Upper limit on the percentage of defects a customer
is willing to accept ( a property of the consumer)
z Average Outgoing Quality (AOQ)
¾ Average of rejected lots and accepted lots
z Average Outgoing Quality Limit (AOQL)
¾ Maximum AOQ for a range of fractions defective
OC Curves come in
Probability of Accepting Lot
various shapes
75% depending on the
sample size and risk of
α and β errors
50%
This curve is more
discriminating
0.80
0.40
0.20
β = 0.10 {
0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.18 0.20
Proportion defective p
AQL Copyright Tapan P Bagchi
LTPD 35
OC Curve helps visualize
producer’s and consumer’s points
z Type I and Type II decision errors correspond logically to
the two decision points on the oc-curve.
z Type I error -- Wrongful Rejection
A type I error is associated with the producer's point -- to
reject when the true value of the quality characteristic is
AQL. The risk of rejecting an AQL lot is the producer's
risk (α = alpha risk)
z Type II error -- Wrongful acceptance
A Type II error is to accept when the true value of the
quality characteristic is RQL -- at the consumer's point.
The risk of accepting a lot, if it is an RQL lot, is the
consumer's risk (ß = beta risk).
Copyright Tapan P Bagchi 36
Decision Criteria
1.00
Probability of accepting lot
Ideal
Not very
discriminating
“Good” “Bad”
1
α = .10
Probability of accepting lot
0.9
0.8
0.7
0.6
0.5
0.4 LTPD
0.3
0.2 “Good” Indifferent “Bad”
0.1 β = .10
0
0 .05 .10 .15 .20 .25
AQL Lot quality (fraction defective)
Copyright Tapan P Bagchi 38
Example: QC Curve for n = 10, c = 1
1 .9139
0.9
Probability of acceptance
0.8 .7361
0.7
0.6 .5443
0.5
0.4 .3758
0.3 .2440
0.2 .1493
0.1 .0860
0
0 .10 .20 .30 .40 .50
Fraction defective in lot
Copyright Tapan P Bagchi 39
Types of OC Curves
z Type A
¾ Gives the probability of acceptance for an individual lot
coming from finite production
z Type B
¾ Give the probability of acceptance for lots coming from a
continuous process or infinite size lot
0.0
AQL pÆ
z Theoretically , it can be
achieved by 100%
inspection of the lot
z Due to sampling, ideal
OC curve cannot be
perfectly achieved.
z If sample size (n) is large,
the OC curve shape will
approach the ideal OC
curve
z Increasing
acceptance number
movement of OC (c) does not
significantly change
the shape of OC
z But Producer’s risk
(β) increases
z Poisson parameter: λ = np
z P(r) = (np)r e-np/r! = Prob(exactly r defectives in n)
z This formula may be used to formulate equations involving
AQL,RQL, α and β to given (n, c).
We can use Poisson tables to approximately solve these
equations. Poisson can approximate binomial probabilities if
n is large and p small.
Q. If we sample 50 items from a large lot, what is the probability
that 2 are defective if the defect rate (p) = .02? What is the
probability that no more than 3 defects are found out of the
50?
Rectified
Passed
⎝ N ⎠
0.015
AOQL
Average
Outgoing 0.010
Quality
0.005
0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09 0.10
AQL LTPD
(Incoming) Percent Defective
where
z It is an economic decision
z Single sampling plans
¾ high sampling costs
z Double/Multiple sampling plans
¾ low sampling costs
¾ More complex to implement
Yes
d≤c? Accept lot
No
Reject lot
z Binomial distribution:
P(x defectives in n) = [n!/(x!(n-x))!]px(1- p)n-x
Recall n!/(x!(n-x))! = ways to choose x in n
Q. If 4 samples (items) are chosen from a population
with a defect rate = .1, what is the probability that
a) exactly 1 out of 4 is defective?
b) at most 1 out of 4 is defective?
Two
simultaneous
conditions that
the (n, c) plan
should meet
Designing a
single sampling
plan by Larson’s
Binomial
Nomogram
http://www.bioss.sari.ac.uk/smart/unix/mseqacc/slides/frames.htm
Acceptance Sampling Overview Text and Audio
http://iew3.technion.ac.il/sqconline/milstd105.html
Online calculator for acceptance sampling plans
http://www.stats.uwo.ca/courses/ss316b/2002/accept_02red.pdf
Acceptance sampling mathematical background
http://assist.daps.dla.mil/
The new US military sampling procedure