INDU 372

Quality Control and Reliability

Final Exam Crash 2023 - Part I

Given By: Asif Choudhury

Upgrade Tutorials
 Courses Given By: Asif Choudhury • 3495 de la Montagne, suite 609, Montreal, Quebec, Canada
Tel. No.: 514-594-2395 • e-mail: [email protected]

INDU 372
Quality Control Reliability

Chapter I
I 1
Chanter 3 Chapters

Introduction to statistical Methods Statistical Process

Quality Control for Quality Improvement Control Basics
Part D u
Chapter 2 Chapter 4 Chapter 6

Quality Improvement Statistical Methods Control Charts

DMAIC for Quality Improvement for Variables
part
Chapter 7

Control Charts
for Attributes
v v
Chapter
8 Chapter 15 Extra Noles
Process Measurement Acceptance Sampling Reliability Models
Capability Analysis Sampling Plans Analysis
Chapter
a
CUSUM EWMA Charts Course Review

Review for Final Exam

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Chapter 1 Introduction to Quality Control

Objective

I to learn the concept of quality and how quality is measured

2 To introduce the reeder to the concept of quality control

and the different modes of quality control

General Concept

1 The concept of quality is different depending on one's perspective

and one's desired result

2 Therefore as an introduction The dimensions and modes of

quality need to be defined

I Definition of Quality
duality cen be defined in many ways and is dependent upon ones
perspective

Quality Perspectives

Manufacturer End User

Quality and its control represents a key factor for success is the
manufacturing and service sectors

lil Dimensions of Quality

duality is a complex characteristic having numerous dimensions which
cen be summarized as follows
a performance

Will the product perform its intended use

Customers will
usually evaluate a product in order to determine
whether the product performs certain functions

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
bl Reliability
How often does the product fail
How many detective items is a production line producing

automobile industry is an exemple how often do

we need to repair the car

c Durability

How long does the product last lifespan

This represents the effective service life of the product
Does the product perform satisfactorily over time

applies to automobile and appliance industries

d servileability

How easy is it to repair the product

For example how easy is it to repair your laptop

el Aesthetits

An important dimension of quality is whet the product look like

is it pleasing to the eye
Style Color Shape Packaging er

f Features

Refers to the product's capabilities

customers associate high
quality with the number of features
and capabilities of the product

g Conformance to standards
Does the product meet with the exact requirements or specifications
placed on it
for example a laser printer cogiststeotpigny contains specifications
thereby alibiing
for proper menulecture and use

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Ii Definitions Of Quality

Due to the presence of different dimensions Of quality different

definitions of quality can result

Productsor services meets the needs of the client lusNmers and

make customers feel satisfied

Fitness Dr use entails quality of clevis

quality of conformance

Quality of Destin deets with intentional design differences with

respect Ni

The materialused I
specification of the components
Reliability
Appearance
Product performance

Quality of conformance deals with how cute the product confirms

to specifications required by the design These factors are a function of

Manufacturing Process
Training supervision of the workforce
Quality assurance procedures and standards UCL Lcl

Problems With This Definition

This definition hes become more associated with the conformance aspect
rather than the quality of design

Focuses less on the whiner since the quality level is established

by the manufacturer

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 II

Fundamental concept if variability of the important characteristics

decreases inequality increases

Figure it shows a comparison of an an mobile transmission made

in the USA Japan Wrt warranty cost and variability

with respect to

Warranty Cost Variability

Figure 1.1. Automobile Transmission Warranty Cost and Variability

Conclusions:

The transmissions made in the USA have a higher variability then those made in Japan → as a, result,
this increased variability has resulted in more warranty cost in the USA → lower quality.

this definition of quality is very precise by reducing variability

warranty costs decrease less cost for the company and consumer satisfaction
increases

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 III

Excessive variability in process performance often results in waste This

increases production costs and will have significant environmental impacts

Reduction ol waste is often used as an indicator of Quality Improvement

particularly true for service industries

Iii Quality Engineering Terminology

In order to establish quality we need to make sure that we have a

comprehension ofquality engineering terminology The terminology can be sub
divided into two celergones

al Quality characteristics

1 Physical length weight voltage viscosity height diameter

2 Sensory taste appearance color
3 Time orientation related to longevity reliability durability
serviceability
4 Attributes Data number of counts lie it of detectives usually a
discrete random variable

Hypergeometric
Discrete Distributions Binomial
Poisson

5 Variables Data continuous random variable voltage spectication power

specification

Exponential Distribution
Continuous Distributions Normal Distribution
Weibull Distribution

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 b Quality Engineering

Much better representation of quality since it

a set encompasses
of operational managerial and engineering activities Thet a company uses
to ensure that the quality characteristics of a product are at normal
and required levels and then variability is at a minimum

Variability cen be measured using statistical methods

statistical measures of attributes or variables

Quality measured relative to the giver or established

specifications

Cl other Quality Engineering Terms

Other terms Thet ar important towards quality control

i specifications desired measurement range for quality characteristic

2 Nominal or target value desired value for a quality characteristic

s upper specification limit usu largest allowable valve

4 Lower specification limit Lse smallest allowable valve
s won confirming Detective a product that does not meet
product specifications

2 Statistical Methods for Quality Control and Improvement

this course will deal with using statistical methods for Quality control
and quality improvement These methods will be introduced 17 this chapter
with all the detail equations described through subsequent chapters

statistical methods for

Quality

statistical Process control Design of Experiments acceptance sampling

see

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 it Statistical Process control SPU

Uses control Cherts with LCL UCL with samples taken 15 Order to
ensure thot samples are within LCL VCL
A typkel control chert is shown in figure 1.2 We will be learning
how to setup control charts for attributes variables

Upper control limit

Lower Control Limit

Figure 1.2. A Typical Control Chart

His Designed Experiment

A designed experiment an experiment

represents where key variables influencing
the quality 0L a product are identified
Controllable varied in order to see their effect on the product
factors are
It the number of control variables is high fecund design is used
where two or more variables are varied at the same time

Figure 1.3 shows two three variable factorial experimentation

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Indu 475

Figure 1.3. Two and Three Variable Factorial Design

Ciii Acceptance sampling

Acceptance sampling represents the third area of quality control and

improvement

Acceptance sampling cutely related to inspection and

testing of the product

The inspection of the product Can be performed immediately after

production Outgoing Inspection Figure 1.4 a

Incoming Inspection where lots are inspected as they are received by the
supplier figure 1.4b

Lot Dispositionrig decisions are shown in figure 1.6C where decisions

are made releny to quality and shipping

Various Probability Distributions Statistics Ulli aid Is

The decision making process

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
2 Describing Central Tendency Numerial Summaries of Dota

In addition to describing the shape of the data graphically as explained

previously we would like to describe the data set's central tendency
A measure of central tendency represents the center or middle of
the data This could be the typical value of a data set
Examples The class average
The average height of a popuchog
The average number of children in a typical family

The central tenancy can be determined for a population or

a sample but in both Ceres is a meen

i Population Meen 111

Me population men y mu is the average of an entire
population also now as the process man

If population measurements are given the population meen cen

be celeucted as the sum of all population measurements divided
the number of class divisions n

I
q measurements

where n of group dinsions

the population meen is known as a population parameter since

it is a valve parameter calculated based on population measurements

Population Parameters Population Meen

Typical y
Punuction Variance 04
Population Proportion p
Problem 1.5 Shows this celeuction

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
His Point Estimate

A point estimate is a one valve estimate of a population

parameter For example we have many point estimators of he poinction
mean Y
sample mean I y pollution mean
Media Xo
Mode

A point estimate is an canceled guess of a population parameter

and is done by obtaining a sample from a population

With this sample we measure

a sample statistic

a Sample Mean I
One example of a sample statistic is the sample mean I the
sample mean is used as a point estimate of the population
mean
male
sample mean I population mean m

supposewe take a number of measurements obtained from a

population The n is known as the sample size The sample
mean cen be written as

InXi
n
i

where x sample mean

Xi the ith measurement taken
n sample size

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
3 Measures of Variation

when estimating the central tendency of a data set it is important

to estimate is variation

The variation tells us how far the data is spread out from
its central tendency

There are 3 ways of measuring sample variation

I
2
Ihara
Variance
ones
sample crannies

we will now discuss each of these quantities

lil Range r

consider a population of asample of measurements The range is

the largest measurement minus the smallest measurement

r Max exit min Xi

While the range is a simple auction it is not accurate since

its accuracy decreases in the presence of outliers
Beetle measures of variability that use all of the data
tend to be better

population ranence
Sta demotion

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Ii Population Variance 104 Sta Deviation O

It population measurements are taken we can use all of the acts

to determine o o more accurate measures of variability

Problem 2.3.

Eduardo is a commerce student and he is taking 5 classes. The classes and their class sizes are shown
below. Determine the variance and standard deviation for the average number of people in Eduardo’s
classes, shown below:

General concepts

Since the entire class for each class was obtained these
measurements are population measurements

I
Step calculate the Population Meen y

60 t 41 15 30 t 34
Y t

s classes

18,0

36

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Step 2 Determine 02 population Variance

o 2 Ki y
n

60 361 t 141 361 t 15 36 130 361 t 34 36

s

10g

I 02 216.4

Step 3 Determine o Population Std Deviation

o I Ki y
n

T I 216.4

0 14.71 I value 36 I 14.71

y

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
a Point Estimates Nr O d o
when the population is too large we take a sample of size
n and determine the sample variance sa and the sample
standard deviation s

point Estimate
puttied's variance
sample Vanence sa Coy
sample Sta Denchon S Population Sta Denchon o
finals

lil sample variance s

s
it Xi I
h i
Xi
n 1
Enki

Ii Sample Sta Denchon s

the sample std deviation is simple the square root of he

sample variance

s Xi
Enki
I n I

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 2.4. Calculation of the Sample Mean and Variance

General concepts

I we are given a sample of data we must determine

the sample men I and sample standard deviation s

step I sample mean X

x
I Xi
n

2,13 t 2.96 t t 2.60

9

19.956

I 2.17 my sample mean

step 2 Calculate the sample variance sa

2
s Xi
ji
n I

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 2 Xi 2.132 2.962 t 2.6012 45.95

E Xi 19.5612 382.9

82 45.95 3892.6

9 I

82 0.43 mm

Sample Std Deviation S 06556 mm

Interpretation
1 Our mean X 2.17 mm with an 5 06556 MM

2 17 I 0.6556 MM

To determine it your sample Std deviation s is high

always check it relative N T
0.6556 x 100 I 30010 relative error
2 17

2 It we were asked to find the range of the sample

Range Max lxi multi

3.02 mm 1.15 MM

Range 1,87 mm

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
iii Probability Plots

Probability plots of sample data are used to determine whether

a particular probability distribution is a reasonable model for data

The normal distribution is the most important distribution in

statistics As such this section deets with using probability plots
in order to determine whether a data models a normal distribution

Why do we care if our acta is Normal

we cen assign probabilities to any outcome of a random
variable

2 we can represent the data in many different ways if

we know the date is normally distributed

In order to construct a probability plot for determining whether

a set of data obey a normal distribution let us look at
the following example

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 8.2: Probability Plot to Determine if Data Obeys a Normal Distribution

Ten observations on the effective service life in minutes of batteries used in a portable personal computer are as
follows: 176, 191, 214, 220, 205, 192, 201, 190, 183,185. Determine if the data obeys a normal distribution.

Step I Rank the Data From Lowest to Highest

Table A III

j x j 0 51
n
Z

I 176 005 1.64

2 183 0.15 1.04

3 185 0.25 0.67

4 190 0.35 0.38

5 191 0.45 0.12

6 192 0.55 0.12

7 201 0.65 0.38
8 205 0.75 0.67
9 214 0.85 1.04
10 220 0.95 1.64

Ranking the t rates corresponding

data ptigitiites to probabilities

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
742
Tape
Appendix A/Statistical Tables and Charts

z
⌠ 1 − 12 u2
Φ( z) = P( Z ≤ z) = 
 e du
⌡−∞ 2π

F (z)

z 0

III Cumulative Standard Normal Distribution

z −0.09 −0.08 −0.07 −0.06 −0.05 −0.04 −0.03 −0.03 −0.01 −0.00
−3.9 0.000033 0.000034 0.000036 0.000037 0.000039 0.000041 0.000042 0.000044 0.000046 0.000048
−3.8 0.000050 0.000052 0.000054 0.000057 0.000059 0.000062 0.000064 0.000067 0.000069 0.000072
−3.7 0.000075 0.000078 0.000082 0.000085 0.000088 0.000092 0.000096 0.000100 0.000104 0.000108
−3.6 0.000112 0.000117 0.000121 0.000126 0.000131 0.000136 0.000142 0.000147 0.000153 0.000159
−3.5 0.000165 0.000172 0.000179 0.000185 0.000193 0.000200 0.000208 0.000216 0.000224 0.000233
−3.4 0.000242 0.000251 0.000260 0.000270 0.000280 0.000291 0.000302 0.000313 0.000325 0.000337
−3.3 0.000350 0.000362 0.000376 0.000390 0.000404 0.000419 0.000434 0.000450 0.000467 0.000483
−3.2 0.000501 0.000519 0.000538 0.000557 0.000577 0.000598 0.000619 0.000641 0.000664 0.000687
−3.1 0.000711 0.000736 0.000762 0.000789 0.000816 0.000845 0.000874 0.000904 0.000935 0.000968
−3.0 0.001001 0.001035 0.001070 0.001107 0.001144 0.001183 0.001223 0.001264 0.001306 0.001350
−2.9 0.001395 0.001441 0.001489 0.001538 0.001589 0.001641 0.001695 0.001750 0.001807 0.001866
−2.8 0.001926 0.001988 0.002052 0.002118 0.002186 0.002256 0.002327 0.002401 0.002477 0.002555
−2.7 0.002635 0.002718 0.002803 0.002890 0.002980 0.003072 0.003167 0.003264 0.003364 0.003467
−2.6 0.003573 0.003681 0.003793 0.003907 0.004025 0.004145 0.004269 0.004396 0.004527 0.004661
−2.5 0.004799 0.004940 0.005085 0.005234 0.005386 0.005543 0.005703 0.005868 0.006037 0.006210
−2.4 0.006387 0.006569 0.006756 0.006947 0.007143 0.007344 0.007549 0.007760 0.007976 0.008198
−2.3 0.008424 0.008656 0.008894 0.009137 0.009387 0.009642 0.009903 0.010170 0.010444 0.010724
−2.2 0.011011 0.011304 0.011604 0.011911 0.012224 0.012545 0.012874 0.013209 0.013553 0.013903
−2.1 0.014262 0.014629 0.015003 0.015386 0.015778 0.016177 0.016586 0.017003 0.017429 0.017864
−2.0 0.018309 0.018763 0.019226 0.019699 0.020182 0.020675 0.021178 0.021692 0.022216 0.022750
−1.9 0.023295 0.023852 0.024419 0.024998 0.025588 0.026190 0.026803 0.027429 0.028067 0.028717
−1.8 0.029379 0.030054 0.030742 0.031443 0.032157 0.032884 0.033625 0.034379 0.035148 0.035930
−1.7 0.036727 0.037538 0.038364 0.039204 0.040059 0.040929 0.041815 0.042716 0.043633 0.044565
−1.6 0.045514 0.046479 0.047460 0.048457 0.049471 0.050503 0.051551 0.052616 0.053699 0.054799
−1.5 0.055917 0.057053 0.058208 0.059380 0.060571 0.061780 0.063008 0.064256 0.065522 0.066807
−1.4 0.068112 0.069437 0.070781 0.072145 0.073529 0.074934 0.076359 0.077804 0.079270 0.080757
−1.3 0.082264 0.083793 0.085343 0.086915 0.088508 0.090123 0.091759 0.093418 0.095098 0.096801
−1.2 0.098525 0.100273 0.102042 0.103835 0.105650 0.107488 0.109349 0.111233 0.113140 0.115070
−1.1 0.117023 0.119000 0.121001 0.123024 0.125072 0.127143 0.129238 0.131357 0.133500 0.135666
−1.0 0.137857 0.140071 0.142310 0.144572 0.146859 0.149170 0.151505 0.153864 0.156248 0.158655
−0.9 0.161087 0.163543 0.166023 0.168528 0.171056 0.173609 0.176185 0.178786 0.181411 0.184060
−0.8 0.186733 0.189430 0.192150 0.194894 0.197662 0.200454 0.203269 0.206108 0.208970 0.211855
−0.7 0.214764 0.217695 0.220650 0.223627 0.226627 0.229650 0.232695 0.235762 0.238852 0.241964
−0.6 0.245097 0.248252 0.251429 0.254627 0.257846 0.261086 0.264347 0.267629 0.270931 0.274253
−0.5 0.277595 0.280957 0.284339 0.287740 0.291160 0.294599 0.298056 0.301532 0.305026 0.308538
−0.4 0.312067 0.315614 0.319178 0.322758 0.326355 0.329969 0.333598 0.337243 0.340903 0.344578
−0.3 0.348268 0.351973 0.355691 0.359424 0.363169 0.366928 0.370700 0.374484 0.378281 0.382089
−0.2 0.385908 0.389739 0.393580 0.397432 0.401294 0.405165 0.409046 0.412936 0.416834 0.420740
−0.1 0.424655 0.428576 0.432505 0.436441 0.440382 0.444330 0.448283 0.452242 0.456205 0.460172
0.0 0.464144 0.468119 0.472097 0.476078 0.480061 0.484047 0.488033 0.492022 0.496011 0.500000
 Closer to 1, the better

Zi
r

Figure P8.2.2. Probability Plot for Example 2 Using the Standard Normal Variable Z

step 3 Make a conclusion with Regards to the Data Plot

the correlation coefficient is to 1
close Theatre the dots
can be assumed to be normally distributed

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
lil Probability Distributions

A probability distribution is a mathematical model that relates the value

or range of values of a random variable with the probability of occurrence
of ther valve in the population

al General Concepts

1 A continuous random variable is a random variable Thel has

infinite possible outlines A continuous distribution is one Thet models
The values of a continuous random variable and the probability of
the outcome

Example The current measured in a wire

The length of a semiconductor

2 A discrete random variable Xis a number assured to outcome

of a random experiment when the number of outcomes is finite
Example Flipping a coin twice let X of heads obtained

S AH HT TH TT

s continuous random variables are different then discrete random

variables since the number of outcomes cannot be quantified

Probability Distributions 15 Quality Control

Continuous Distributions Discrete Distributions

Normal Distribution Binomial Distribution

Exponential Distribution Hypergeometric
Weibull Distribution Poisson Distributor
Log normal Distribution

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
3 Continuous Random Variables

As explained in figure 2.4 a continuous random variable

is one containing infinite possible outcomes some examples

a the time between arrivals at a bus terminal

b the number of minutes required for a repair
C the distance between cracks in a road
at the diameter of a hole drilled in a part

What you will notice about continuous random variables is

that they deal with time and distance slice these cen here
infinite possible outcomes

A continuous probability distribution is represented

by
a continuous function fix not a table of x th values

A probability density function PDF is a graph of

X = random variable
f(x) = Probability associated with a range of values of x)
fix P(a < x< b)

f(x)
M

A b Ivo
Figure 2.4. Continuous Probability Density Function

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
lil properties of a continuous probability Distribution fix
there are some fundamental properties of tix

I 0 s flat I
any probability must have a valve
between 0 I inclusive

2 fix I sum integral of flat over the

f entire region of possible outcomes
B must be equal to 11

3 Pla ex e b fab fat ax area under fix between

a a b

Iiiil Special property of continuous fix

since finding the probability that a random variable x
lies between two values x a a x b involves an area
under the curve calculation

Pl X x O 1 for continuous
only

random valve assigned

variable x

Property i implies that for a continuous random variable you

will never be asked to find the probability that X equals
a valve unless they want to trick you
Figure 2.5 shows a graphical representation of finding PIX 50
It cen be seen that there is no area una x so Therefore
P x 50 01

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 X = random variable
fix P(X=50)
f(x) = Probability associated with a range of values of x)

f(x)

50 NO X

Figure 2.5. Continuous Probability Density Function - P(X=50) = 0

al Implications of Property 1

Since P x x D we have numerous consequences related

to finding probabilities involving intends a b

Pl a ex e b Pla e x I b Pl x a

Pl a e Xt b Pl a ex bl t pix b
o

Pl a e x e b Pl g x cb t Ply at Pix b

P a ex e b Pla ye b P a ex b Pla X b 2

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
S Meen Variance of a Continuous Probability Distribution

is mean 171
Given a random variable x and a corresponding distribution
fix discrete the men y of this distribution is
defined as

Mean
Y Etat
fix fix ax 6

expected value of X

Interpretation of y
The mean represents the valve one would expect to
obtain for y the random variable X

ii variance 10 s
The variance o of a probability distribution represents
the spread of the possible values for the random
variable X

02 E x y x2 fix ax
Vix
f y 7

iii standard Deviation o

The standard deviation is simply the square root of the

variance

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Notes Concerning Equations 6 7
I since each of the equations involve integrals of functions
multiplied together you will need to remember your
integration rules

a Integration by Parts
b substitution techniques
c Integration of Basic functions

2 The B in each integral refers to the entire region

over which feel is defined
3 Take note that these parameters at process parameters I population
parameters

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
6 Continuous Probability Distributions Models

Objectives

1 To discuss and describe the most famous probability

distribution model Normal Distribution

2 To look
at using a approximation to
normal
the Binomial Poisson Distributions where applicable

3 To discuss the Exponential Distribution and its

relationship to the Poisson Distribution

General concepts

1 As we did for discrete probability distributions we will

now look at two commonly used continuous distributions
Normal Exponential
These two distributions are used widely in engineering

2 Frequently when determining the probability related

to the number of outcomes becomes large the normal
distribution cen be used to approximate the results
intended for a binomial or poisson distribution

1 Normal Distribution Gaussian Distribution

When aexperiment is repeated
random the random variable
X that equals the average result over the experiments is
called the Normal Distribution
I Gaussian Distribution

the probability distribution fix is

fix e it I l N x c o
ppl

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
lit properties of the Normal Distribution

1 the normal distribution is a bell shaped curve charactericed

by two main quantities
Elx1 lpopulator Mean
y
VH1 02 variance

2 In short form the normal distributor Cen be written

as Nly O

3 The differences in shape ar based on differences

between
y o or both as shown in Figure 2.6

Figure 2.6. Characterization of Normal Distribution with Differing Mean and Variance

The Meer y tells us where the distribution is

centred The variance 02 tells us how wide the
bell curve will be large 02 wider and Halter

4 The normal is symmetric and moving away

distribution
in standard devichon away from the man encloses
a set amount of area line probability see figure 2.7

Every normal distribution regardless of its y 04

has this characteristic

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Figure 2.7. Area Under the Normal Distribution

Ii Standard Normal Random Variable Z

As a result of the properties mentioned
any random variable
x obeying normal distribution
a can be converted a
to
standard normal variable Z

Normal Random Variable X standard Normal Random variable Z

7 O 4 0 02 1

Z
Xyy
2 7 100 X 120
Represents the number a Ip
of standard deviations
the random variable X
is from its mean

E 120100 1.33

How rare is the

data point
PIX 21201 Pl E C 1.33

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Once the standard normal random variable is known we can
establish probabilities associated with 2 i x Using table III
Normal Distribution Tables shown in figure 5.3

Notes concerning figure 5.3 table A 111

1 The table give cumulative probabilities P Zet less than
or equal to probabilities
2 The tables are divided in two pages

Page 1 For 2 relics between 3.99 and O

2 For Z valves between 0 3.99

The complete two pages are shown in this document

3 The 1061207 of Z and the probabilities corresponding

N Z are Shown in figure 5.3
For example Pl 2 a 1.761 0.039204

4 Useful Property

P1 Z E Z I Pl Z E Z

For example

Pl Z E t 1.96 I Pl 2 E 1.96

This is shown on page 7 using Z 1.76 as an

example shown graphically

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 I

Z values - hundreth decimal place

Z values - up
to the tenth
decimal place

Probability - all of the numbers within the table are probability values

Figure 5.3. One-Half of the Normal Distribution Tables (Negative Z values, Probabilities from 0 to 0.5

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
4 Useful Property Graphically

Éi

PLZ 1.76

t
I

1.76

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
742
tbh
Appendix A/Statistical Tables and Charts

z
⌠ 1 − 12 u2
Φ( z) = P( Z ≤ z) = 
 e du
⌡−∞ 2π

F (z)

z 0

III Cumulative Standard Normal Distribution

z −0.09 −0.08 −0.07 −0.06 −0.05 −0.04 −0.03 −0.03 −0.01 −0.00
−3.9 0.000033 0.000034 0.000036 0.000037 0.000039 0.000041 0.000042 0.000044 0.000046 0.000048
−3.8 0.000050 0.000052 0.000054 0.000057 0.000059 0.000062 0.000064 0.000067 0.000069 0.000072
−3.7 0.000075 0.000078 0.000082 0.000085 0.000088 0.000092 0.000096 0.000100 0.000104 0.000108
−3.6 0.000112 0.000117 0.000121 0.000126 0.000131 0.000136 0.000142 0.000147 0.000153 0.000159
−3.5 0.000165 0.000172 0.000179 0.000185 0.000193 0.000200 0.000208 0.000216 0.000224 0.000233
−3.4 0.000242 0.000251 0.000260 0.000270 0.000280 0.000291 0.000302 0.000313 0.000325 0.000337
−3.3 0.000350 0.000362 0.000376 0.000390 0.000404 0.000419 0.000434 0.000450 0.000467 0.000483
−3.2 0.000501 0.000519 0.000538 0.000557 0.000577 0.000598 0.000619 0.000641 0.000664 0.000687
−3.1 0.000711 0.000736 0.000762 0.000789 0.000816 0.000845 0.000874 0.000904 0.000935 0.000968
−3.0 0.001001 0.001035 0.001070 0.001107 0.001144 0.001183 0.001223 0.001264 0.001306 0.001350
−2.9 0.001395 0.001441 0.001489 0.001538 0.001589 0.001641 0.001695 0.001750 0.001807 0.001866
−2.8 0.001926 0.001988 0.002052 0.002118 0.002186 0.002256 0.002327 0.002401 0.002477 0.002555
−2.7 0.002635 0.002718 0.002803 0.002890 0.002980 0.003072 0.003167 0.003264 0.003364 0.003467
−2.6 0.003573 0.003681 0.003793 0.003907 0.004025 0.004145 0.004269 0.004396 0.004527 0.004661
−2.5 0.004799 0.004940 0.005085 0.005234 0.005386 0.005543 0.005703 0.005868 0.006037 0.006210
−2.4 0.006387 0.006569 0.006756 0.006947 0.007143 0.007344 0.007549 0.007760 0.007976 0.008198
−2.3 0.008424 0.008656 0.008894 0.009137 0.009387 0.009642 0.009903 0.010170 0.010444 0.010724
−2.2 0.011011 0.011304 0.011604 0.011911 0.012224 0.012545 0.012874 0.013209 0.013553 0.013903
−2.1 0.014262 0.014629 0.015003 0.015386 0.015778 0.016177 0.016586 0.017003 0.017429 0.017864
−2.0 0.018309 0.018763 0.019226 0.019699 0.020182 0.020675 0.021178 0.021692 0.022216 0.022750
−1.9 0.023295 0.023852 0.024419 0.024998 0.025588 0.026190 0.026803 0.027429 0.028067 0.028717
−1.8 0.029379 0.030054 0.030742 0.031443 0.032157 0.032884 0.033625 0.034379 0.035148 0.035930
−1.7 0.036727 0.037538 0.038364 0.039204 0.040059 0.040929 0.041815 0.042716 0.043633 0.044565
−1.6 0.045514 0.046479 0.047460 0.048457 0.049471 0.050503 0.051551 0.052616 0.053699 0.054799
−1.5 0.055917 0.057053 0.058208 0.059380 0.060571 0.061780 0.063008 0.064256 0.065522 0.066807
−1.4 0.068112 0.069437 0.070781 0.072145 0.073529 0.074934 0.076359 0.077804 0.079270 0.080757
−1.3 0.082264 0.083793 0.085343 0.086915 0.088508 0.090123 0.091759 0.093418 0.095098 0.096801
−1.2 0.098525 0.100273 0.102042 0.103835 0.105650 0.107488 0.109349 0.111233 0.113140 0.115070
−1.1 0.117023 0.119000 0.121001 0.123024 0.125072 0.127143 0.129238 0.131357 0.133500 0.135666
−1.0 0.137857 0.140071 0.142310 0.144572 0.146859 0.149170 0.151505 0.153864 0.156248 0.158655
−0.9 0.161087 0.163543 0.166023 0.168528 0.171056 0.173609 0.176185 0.178786 0.181411 0.184060
−0.8 0.186733 0.189430 0.192150 0.194894 0.197662 0.200454 0.203269 0.206108 0.208970 0.211855
−0.7 0.214764 0.217695 0.220650 0.223627 0.226627 0.229650 0.232695 0.235762 0.238852 0.241964
−0.6 0.245097 0.248252 0.251429 0.254627 0.257846 0.261086 0.264347 0.267629 0.270931 0.274253
−0.5 0.277595 0.280957 0.284339 0.287740 0.291160 0.294599 0.298056 0.301532 0.305026 0.308538
−0.4 0.312067 0.315614 0.319178 0.322758 0.326355 0.329969 0.333598 0.337243 0.340903 0.344578
−0.3 0.348268 0.351973 0.355691 0.359424 0.363169 0.366928 0.370700 0.374484 0.378281 0.382089
−0.2 0.385908 0.389739 0.393580 0.397432 0.401294 0.405165 0.409046 0.412936 0.416834 0.420740
−0.1 0.424655 0.428576 0.432505 0.436441 0.440382 0.444330 0.448283 0.452242 0.456205 0.460172
0.0 0.464144 0.468119 0.472097 0.476078 0.480061 0.484047 0.488033 0.492022 0.496011 0.500000
 Appendix A/Statistical Tables and Charts 743
n
z 1 − 12 u2
Φ( z) = P( Z ≤ z) = ∫−∞ e du
2π

F (z)

0 z

III Cumulative Standard Normal Distribution (Continued)

z 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
0.0 0.500000 0.503989 0.507978 0.511967 0.515953 0.519939 0.532922 0.527903 0.531881 0.535856
0.1 0.539828 0.543795 0.547758 0.551717 0.555760 0.559618 0.563559 0.567495 0.571424 0.575345
0.2 0.579260 0.583166 0.587064 0.590954 0.594835 0.598706 0.602568 0.606420 0.610261 0.614092
0.3 0.617911 0.621719 0.625516 0.629300 0.633072 0.636831 0.640576 0.644309 0.648027 0.651732
0.4 0.655422 0.659097 0.662757 0.666402 0.670031 0.673645 0.677242 0.680822 0.684386 0.687933
0.5 0.691462 0.694974 0.698468 0.701944 0.705401 0.708840 0.712260 0.715661 0.719043 0.722405
0.6 0.725747 0.729069 0.732371 0.735653 0.738914 0.742154 0.745373 0.748571 0.751748 0.754903
0.7 0.758036 0.761148 0.764238 0.767305 0.770350 0.773373 0.776373 0.779350 0.782305 0.785236
0.8 0.788145 0.791030 0.793892 0.796731 0.799546 0.802338 0.805106 0.807850 0.810570 0.813267
0.9 0.815940 0.818589 0.821214 0.823815 0.826391 0.828944 0.831472 0.833977 0.836457 0.838913
1.0 0.841345 0.843752 0.846136 0.848495 0.850830 0.853141 0.855428 0.857690 0.859929 0.862143
1.1 0.864334 0.866500 0.868643 0.870762 0.872857 0.874928 0.876976 0.878999 0.881000 0.882977
1.2 0.884930 0.886860 0.888767 0.890651 0.892512 0.894350 0.896165 0.897958 0.899727 0.901475
1.3 0.903199 0.904902 0.906582 0.908241 0.909877 0.911492 0.913085 0.914657 0.916207 0.917736
1.4 0.919243 0.920730 0.922196 0.923641 0.925066 0.926471 0.927855 0.929219 0.930563 0.931888
1.5 0.933193 0.934478 0.935744 0.936992 0.938220 0.939429 0.940620 0.941792 0.942947 0.944083
1.6 0.945201 0.946301 0.947384 0.948449 0.949497 0.950529 0.951543 0.952540 0.953521 0.954486
1.7 0.955435 0.956367 0.957284 0.958185 0.959071 0.959941 0.960796 0.961636 0.962462 0.963273
1.8 0.964070 0.964852 0.965621 0.966375 0.967116 0.967843 0.968557 0.969258 0.969946 0.970621
1.9 0.971283 0.971933 0.972571 0.973197 0.973810 0.974412 0.975002 0.975581 0.976148 0.976705
2.0 0.977250 0.977784 0.978308 0.978822 0.979325 0.979818 0.980301 0.980774 0.981237 0.981691
2.1 0.982136 0.982571 0.982997 0.983414 0.983823 0.984222 0.984614 0.984997 0.985371 0.985738
2.2 0.986097 0.986447 0.986791 0.987126 0.987455 0.987776 0.988089 0.988396 0.988696 0.988989
2.3 0.989276 0.989556 0.989830 0.990097 0.990358 0.990613 0.990863 0.991106 0.991344 0.991576
2.4 0.991802 0.992024 0.992240 0.992451 0.992656 0.992857 0.993053 0.993244 0.993431 0.993613
2.5 0.993790 0.993963 0.994132 0.994297 0.994457 0.994614 0.994766 0.994915 0.995060 0.995201
2.6 0.995339 0.995473 0.995604 0.995731 0.995855 0.995975 0.996093 0.996207 0.996319 0.996427
2.7 0.996533 0.996636 0.996736 0.996833 0.996928 0.997020 0.997110 0.997197 0.997282 0.997365
2.8 0.997445 0.997523 0.997599 0.997673 0.997744 0.997814 0.997882 0.997948 0.998012 0.998074
2.9 0.998134 0.998193 0.998250 0.998305 0.998359 0.998411 0.998462 0.998511 0.998559 0.998605
3.0 0.998650 0.998694 0.998736 0.998777 0.998817 0.998856 0.998893 0.998930 0.998965 0.998999
3.1 0.999032 0.999065 0.999096 0.999126 0.999155 0.999184 0.999211 0.999238 0.999264 0.999289
3.2 0.999313 0.999336 0.999359 0.999381 0.999402 0.999423 0.999443 0.999462 0.999481 0.999499
3.3 0.999517 0.999533 0.999550 0.999566 0.999581 0.999596 0.999610 0.999624 0.999638 0.999650
3.4 0.999663 0.999675 0.999687 0.999698 0.999709 0.999720 0.999730 0.999740 0.999749 0.999758
3.5 0.999767 0.999776 0.999784 0.999792 0.999800 0.999807 0.999815 0.999821 0.999828 0.999835
3.6 0.999841 0.999847 0.999853 0.999858 0.999864 0.999869 0.999874 0.999879 0.999883 0.999888
3.7 0.999892 0.999896 0.999900 0.999904 0.999908 0.999912 0.999915 0.999918 0.999922 0.999925
3.8 0.999928 0.999931 0.999933 0.999936 0.999938 0.999941 0.999943 0.999946 0.999948 0.999950
3.9 0.999952 0.999954 0.999956 0.999958 0.999959 0.999961 0.999963 0.999964 0.999966 0.999967
Problem 2.4. Normal Distribution

E
I Findthis
N ly 04 Probability

Let X tensile strength of the wood

Step I Convert X to Z
Z
Ey
Z 35 2.5
240

Step 2 End P x 351

PIX 3351 1 P X 35
I PL Z L 2.5

I 0.006210

71 2,351 0.99379

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
iii central limit theorem

suppose we ar sampling from a population of unknown distribution

The sampling distribution of the sample men will still be normally
distributed with men of y ly and variance of o n
as long as the sample size is large enough In 30

Central Limit Theorem

(most important theorem in statistics)

If we take a sample of size n from a population with

mean y and variance o and if I is the sample mean the
limiting form of the distribution

z X MI
OI
chapter 4 z
Z X y 3 xp
o

Notes concerning Equation 3

I Eavation s gets increasingly more accurate as n so lie

as the sample size increases It is a good approximation when
no 30

2 Figure 2.8 and 2.9 Shows this concept

3 Equation s cen be used in order to establish probabilities

related to the sample mean I
Make sure you understand the difference between whet
we learned in chapter Y and what we just teemed

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Figure 2.8. Distribution of the Sample Mean and Population Mean

4 It a sample of she n is taken from a normally distributed

population The population of all sample means will be normally
distributed

Any number of samples of size n taken from a normal

distribution will result in sample means that are normally
distributed

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Figure 2.9. Central Tendency as the Number of Dice (n) Increases

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 2.5.

General concepts

I the population is normally distributed t

2 The question refers to the sample meed so we are dealing
with a Semple

Z I Y
Orn

step I Find t
z
To In

75.75 psi 75.5 psi

3S psi 6T

0.175

Step 2 Find P I 75.75

Pl I 75.75 PLZ 0.1751 1 P 7 0.175

I 0.571424

0.428576

probability that the sample mean exceeds 75.75

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 2.6.

my

General concepts

1 Although we ae not told that the data is normally distributed

it cen readily be assumed slice n 30 n na

2 Gim population parameters

y 8.2 Mit
G 1.5min
Z T Y
on
step I P Ic 10 min

Pl I P
I
10 Min Z s 10 min 8.2min
45
jug
P Z 8.4
I
C

P I N min PLZ 8.4 1.0

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Step 2 P Sc x 10

P 5 X 10 PL I CIO Pl I 5

Zz 10 8.2 8.4
1.5
VIS

Z 5 8.2 14.93

P 5 X 10 PL I CIO Pl I 5

PI Z 8.41 P Z C 14.931

1.0 O

1.0

Step 3 PLI 6

PL X G P
I Z 6 8.2
s
1.5 49

t
P
I Z 10.3
a

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
I Exponential Distributor
The exponential distribution is a continuous distribution
where the random variable x represents

let X time space between occurrences

Continuous random variable

this is the inverse of the Poisson Distributor where the

random variable x of occurrences in a given time discrete
As such they are related

the exponential distribution is used in engineering applications

a Serna times wait times in a queueing system

b Traffic flow 5197611261707
C Markov chains Stochastic Processes

the random variable x that denotes the time between

occurrences obeys the Exponential Distribution

d
fly de 4

lil Meon Vanance

The mean and variance of the Exponential Distribution

are as follows

E 49
I
time between
y x average
oesurrenie

02 Vix 4 2 4b

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Notes concerning Equation 4 d
1 the exponential distribution uses the Poisson Meen
The mean must be normalized to the unit of
study

2 You cen be asked Poisson Exponential questions

in the same problem keep the following in mind

Number of Occurrences Poisson

Time space between occurrences Exponential

3 the role of can be known as the Meen Time N Failure Imitt

It
lit we modelled faire time as an exponential distribution

continuing on the idle of point S X Invasion men could be

represented at the failure rose I of failures in a given amount of
time

s these two distributionare modelled as having a tech of memory

property which means the probability of ofluence is one Miera is
the same as in previous or subsequent intervals

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 3.1. Exponential Distribution

General concepts

1 The exponential mien is given 1 400

t fine
2 Theater the Poisson Meen if reveal I felines
go he
t
e OWLS felloes

0002
Pl x 1001
fi d 0025 e

000254 100
ax
feat ax
e e eat
0 fix I
I 0 2212 I

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Step 2 Part b

PIX soul l Pi SW
500 O0025t
I 0.002g e dy
go

500
e 0
0025 x
I
o I
I O00254001
I e

0,2865
Step 3 Part C

PIX 100 x 4001 i lack of memory properly

PIX civil

DIX c 1001 0.2212 see part a

Step 4 Part d

Let 4 be Ot failure ri 10 tested

binomial i n 10
P 431 I Ply c 1 p p x civil 0.2212

1 10.2212141 0,2212 10
Ig
0.9179

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Part e
let us first And the probability that One will fail in less
then 800 hours
O0025
PIX C 8001 1,800 0.002g e ax

0.0025k 800
e
O
0002518001
I e

0 8647

let W the number of failed assemblies out of 10 assembled

binomial

Plw 101 0.8647 I 0.864710

0.2337

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 3.2. Exponential Distribution and Time to Failure

An electrical component in a radar system has a usual life described by an exponential distribution with a
failure rate of 10^-4 number of failures/hour. Find the probability that the component will fail before it's
expected life.

General concepts

Gwen the Poisson mean I 10 4 failures 41011107 man

how

Find the probability ther the component will fail before its expected
life

Stell I Find Me Expected Life of the Rear System

The expected life of the radar is known as the mean time
N failure MTF

MTTF
I I
10000
4 failures 1pm
hour

Step 2 Find probability that the component will fail before its
expected life

Pl x a tt
mite e p x 100001
good e dy

Where 4 10 4 Poisson Meen

fly
10
Plx looool 1100010 4 e ax

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 10000
do Ya
Pl X 10000 e
0

Phx 10000 0.368

2 Discrete Probability Distributions

Discrete probability distributions are critical to setting up control charts

and monitoring overall quality control

Probability Distributions 15 Quality Control

Continuous Distributions Discrete Distributions

Normal Distribution Binomial Distribution

Exponential Distribution Hypergeometric
Poisson Distribution

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
2 Meen Variance of a Discrete Probability Distribution

is mean 171
Given a random variable x and a corresponding distribution
fix discrete the men y of this distribution is
defined as

Mean
y Ely Xiflxi
Xi o

expected value of X

Interpretation of y
The mean represents the valve one would expect to
obtain for y the random variable X

ii variance 10 s
The variance o of a probability distribution represents
the spread of the possible values for the random
variable X

02 Vix E x y x fix y

iii standard Deviation o

The standard deviation is simple the square root of the

variance

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
3 Specific Discrete Probability Distributions

Objectives

i to extend our discussion of discrete probability distributions

and to describe some common distributions
2 To be able to distinguish between each distribution and
its corresponding random variable

3 To be able to use the correct distribution for a

given problem

General concepts

1 In this lecture we will discuss J discrete probability

models that can be used to model real life processes

2 These models are very useful for determining probabilities

very easily and quickly

1 Discrete Probability Distributions Overview

In this course 6 different discrete probability distributions

will be discussed These are described in Figure 3.1

Discrete Probability Distributions

Uniform Geometric Hypergeometric

Distribution Distribution Distribution

Binomial Negative Binomial Poisson’s

Distribution Distribution Distribution

Figure 3.1. Discrete Probability Distributions

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
3 Binomial Distribution
The Binomial Distribution has the following properties

I the experiment consists of n trials

2 Each trial has only two possible outcomes
success or failure
Bernoulli
3 Each trial is independent the outcome of a Trials
trial is unaffected by any other trials
4 Each trial has a constant probability of
success P

Let X number of trials that result in a success out

of the n trials

the binomial distribution takes the following form

flat p i p 2
y

Where i n total of trials

x number of sucesses out of the n trials
p probability of success for each trial

it Meen of a Binomial Distribution

7 Ela Np za

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Interpretation of men

This represents the expected number of successes we

would get in n trials

Ii variance of a Binomial Distribution

02 Vix ppl 1 p 2b

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 3.2.

General Concepts

1 Number of trials n 25
2 we wish to determine the performance of the rule
stopping theprocess it one or more non conforming parts are
fund
decision

Ski I Define the Random Variable X

Let X of non conforming parts in a sample ol 25

Stell 2 Find the Required Probability

Find the probability that the process will stop

Ply 1 I PI XLI I Ply O

I 10.01 0.99125
128
0.222
This means that we will stop the production 22.2 of the time

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Some Possible Outcomes Not Possible Outcomes

Binomial DD NNN
DNNND
N DD NN
NNDD N
NN DND
DNNND

Negative Binomial DNNND DD NNN x 21

DNNND N DD NN X 31
NN DND NN DD N X 4

6 Hypergeometric Distribution

This is the only distribution where each trial is dependent without

replacement

we have the following properties

A set of N objects kit of successes

NK II of failures

r we choose n objects without

replacement

Let X of successes that result out of the n

trials

fix
I YY 5

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 we actually dealt with this distribution in Enon 171
However we did not call it the hypergeometric distribution

keep in mind this is the only distribution whee each trial

is dependent

lil Meon y of the Hypergeometric Distribution

Y Ehl Xi fixe

Elx AP Sa where p KIN

good for when
N is very large

Interpretation of men

This represents the expected number of success we

would get in n trials

ii variance to of the Hypergeometric Distribution

02 V x Xi fixi y

Vix Noll pl sb
in p

The complicated equation for the variance is due to the fact

that we have w items and we are only doing n trials
N n

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
lil Comparison Between Binomial Hypergeometric Distribution

When is the Binomial Distribution similar to the hypergeometric

distribution

It would have N be the case when the probability of

Sukets does not change very much on each bill

when K is small such that D say

N is large I
relatively constant

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
4 Poisson Distribution

The Poisson Distribution allows for the determination of the

probability related to the number of occurrences in a given amount
of time or space
It is used frequently to model traffic flow queueing systems
and factory production

i properties of the Poisson Distribution

Gives an interval of time hr see mint or space m mi ms

a the probability of one event occurring in a given

subinterval is the same as any other subliterval and
is proportional to the size time or space of the
subinterval

b let d
be the average number of occurrences in
the subliterval ie calls hr cracks km cars hr

Let X of occurrences within an interval of time or

space

the Poisson distribution takes the following form

fix PIX x e f 6
x

Notes concerning Eg 6

1 The t Poisson meant must be normalized to the interval

of stray
2 X 0,1 2,3 N

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Hit Meen ly of the Poisson Distribution

Ek Xi flxi

E x e X 6A

Interpretation of me Meon

The men of the Poisson Distribution represents the expected

of successes in a given interval of time or space

iii variance 021 of the Poisson Distributor

2
02 VIX Xi fix y

VH1 X Gb

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 3.3. More Complicated Problem

General concepts

This is a continuation 0L Problem 1 but is a much more practices

problem

Step 1 Develop the anchor

let X of detective components A of components to here

in Stock

PIKE Xl 0.95 xn
0.021 10.981

step 2 Trial Error

n e 104

PIX E y PIX O t PIX L t PIX n 21 t Pil 3 t PIX 4

4t 103 2 102
1004110.0210 10.981 10.02 0.981 t 0.02 0.981
f 1,04 w
24

3 101 t w
104 0.02 0.981 wy 0.02 0.981
3 4

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 0.122 t 0.0245 t 0.2729 0.1893 0.0976

0.706

n 105

Plxes PIX OI PIX l PIX 21 t PIX 3 PIX y PIX 5

10.0210 0.981105 10.02 0.981104 0.0212 0.981103

10,5 11,05 ios

0.0213 0.981102 10.021 0.98110 t 10.021 10.981100

ios 145 gs

0.1199 0.2569 0.2726 0.1910 0,0994 0.0410

0.981

he should keep 105 components in stock

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 3.6.

step I Choose Distribution

Assume each trial occurs without replacement Hypergeometric

Dish

fat
1 Y
i
X of non conforming units in a sample of 5

Stell 2 Find Plx i

k 3
No 30

fill
É fill 0.3695

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Step 3 And PIX 11

PIX 31 1 PL x ai

I Dex 01

I
8 I
Y
I 0.5665

10.43351

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 3.7.

Step I Define the Random Variable X

Let X it errs in a customer's bill

Step 2 Find Probability Oh One Emr

fix a ed fx
x
out
e f 01
I

0.0099

What probability that in the next 10 customer’s bills, that two of them will have no errors?

Step I Find Probability of one funner having No Fmr

o 00
Pix O flop e 10.01 e 0.99

probability of Juices p
for our binomial

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Step 2 And the Probability the 2 out 10 Bills will have
NO Emes

Let y it of bills that have no errors

PLY 21 0.99 1 0.9978

to
4.41 10 15

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Lecture 4 the use of statistics for Process control

Objects
I to describe the statistical methods used from samples in
order to conduct Process control
2 to look at sampling distributions originating from different
probability distributions

3 N describe a intrepret confidence intents and hypothesis tests

for different cases

General concepts

1 In the previous chapter we discussed different probability distributions

used in the modelling of a process It was assumed Thel the
process parameters i.e population men y pop variance Coat or p
were known
2 In practical applications 15 quality control the process parameters are
rally known Il we did know them there would be no need
for quality control
3 Therefore we need statistics in order to predict our process
parameters and make decisions regarding the process

1 Basic Statistics
Making conclusions based upon statistics is celled statistical inference
This forms the basis of quality control

are measured from random samples of size n which are

statistics
independently or identically distributed large populations with replacement
or are not independent without replacement

i sample Mean Sample Variance Range

when a sample of size n is obtained containing n observations

of the random ranable X hi Xz ks Xn we can obtain the
sample mean variance as follows

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
a sample mean x

I E Ii
n

b Sample variance s2 use this form

2
s
II n
x I
I Xi
n 1
Exit 2

The standard deviation of the sample s is simply the square

root of the sample variance
c Range r

The range of a sample is the difference between the maximum

r
and minimum valves in the sample

r Max exit min xi s

2 Sampling Distributions

The probability distribution that a statistic obeys is called a sampling

distribution The statistics described above I s r all obey
some probability distribution these distributions were discussed in lectures
2 3

i Sampling from a Normal Distribution

suppose we are studying a random variable X originating from a

population obeying a Normal Distribution with a population mean Y
and Variance 02

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
a Sampling Distributor Of The Sample Men I
Now suppose a sample of size n is taken from this population
The sample men I is equal to
s Provided 02
is known
X X t Xz t Xn a ply
n

The sample men follows a normal distribution with mean Y

and variance sa and with the standard normal random
variable Z g

Z it o is known
top
If 82 population variance is unknown the sampling distribution of
the men will obey the T distribution with V n t degrees of
freedom

T I y z if 02 is minion
sit
sample Std deviation

b sampling Distribution of the sample variance 84

Now suppose we have the same sample of size n for a random
variable X originating from a Normal Distribution The sample
variance can be calculated as shown below

2
S2
IT x I
I Xi
Exit
n n 1

the probability distribution then the sample variance Obeys is

known as the Chi squared X distribution

x
54
Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
3 Estimation of Process Parameters

Recall from Gwon 371 statistics are used to measure or predict

population parameters

In Quality Control we cell population parameters process parameters

ENGR 371 INDU 372

Process Parameters cen be estimated in 3 different ways

1 Point Estimates
a Interel Estimates confidence Intervals
a Hypothesis testing Ho Hi
Each of these methods will be discussed with specific reference to
quality control

i Point Estimates

Point estimates are one valve statistics used as a predictor of

a process parameter

I point estimate
0 population parameter process parmeter
For example

The sample men is a point estimate of 7

I
process
the sample variance 82 is a point estimate of G perimeter
the sample Sta deviation s is a point estimate of o

a point Estimates of 8 in Quality Control

In quality control the population std denation o can be estimated

as follows I provided X ha x xp is from a normal distribution

whee S sample Std deviation

8 point estimate Ot process Roemer o

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 If the range r is calculated r Ymca Xmas another point
estimate for o cen be used

8 R S
dz

where ca and da are given 15 Table 4,1 shown below the

Table has also been attached to the end of the document

Table 4.1. Factors for Control Charts and Point Estimates

Yammile
size
If n>25, us the formulas at the bottom of Table 4.1

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 4.1. Point Estimation
The service life of a battery used in a cardiac pacemaker is assumed to be normally distributed. A random sample
of 10 batteries is subjected to an accelerated life test by running them continuously at an elevated temperature
until failure. The following lifetimes (in hours) were obtained: 25.5, 26.1, 26.8, 23.2, 24.2, 28.4, 25.0, 27.8, 27.3 and
25.7.

a) Determine a point estimate for the standard deviation using the sample standard deviation and Table 4.1.
b) Determine a point estimate for the standard deviation using the range and Table 4.1.

Step I find the script Std Devicton Range

s Exit Exit 6783.8 67600 1.63 hrs

10
I n 1 10

R Xmas Xm 28.4 23.2 5.2 hrs

Step 2 Rid 8 for Fah Case

Table 4.1 n 10
Cy I 9ft

p 1.63 1.68
g 0.9127

F I 5.2 1.69
I 3.078

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Iii Confidence Intervals Interval Estimates

Before we develop confidence intervals for different cases we need

a general idea of what a confidence interval entails

i Confidence intervals establish interval over which we

some
are some percent confident usually 95 9801 or 99 1
that our population parameter 17 o or p will lie

2 We have 3 general types of confidence intervals

a two Sided Confidence Interval

establish
In this case we a percentage of confidence
that our population parameter will lie between a lower
rate e and upper value V

L t y U
02
E E U L E O EU
Le p E U

b One Sided Upper Confidence Interval

In this case we establish a percentage of confidence

that our population parameter will lie below a specified
upper value V

7EV
OLE U O E U
p e u

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 b one sided lower Confidence Interval

In this case we establish a percentage of confidence

that our population parameter will lie above a specified
lower value V
L s y
L E or
L E P

we here 4 cases of intervals

confidence For each case we can
generate two sided one sided upper and one sided lower confidence
intervals

These 4 cases are as follows

case
i confidence interval for y with 02 known

case 2 Confidence interval for with G unknown

y
case 3 Confidence interval for o or o

case 4 Confidence interval for the population proportion P

Note Each of these confidence internees are for the y and o

of a normal distribution only
How do we know from a problem that are are dealing with
a normal distribution

a we are told directly

b it our sample size is large enough n 30

c it we plot the data gives from the sample and

do a normal probability plot and fee then the
deta plots to straight line

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
2 Confidence Interval on the Mees y O known

knowing o means we have less uncertainty regarding our

confidence interval As such the standard normal Z cen be
used

it two sided CI
A 10011 2 confidence interval for the mean y with o
known

X E E X t 1
24,9 Y 24,9

Where a alpha level of the CI the probability

of error with our CI
if 951 CI 2 0.05
9801 CI 2 0.02

x sample Meen
n sample size
Zai standard normal at a probability of 42
o population standard deviation

IF

25010 2.5010

th's 742
a p.gg

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Iii One Sided CIs
sometimes a two sided CI is not required As such we
can form one sided CIS

a one sided Upper

A 10011 21 upper CI for y can be written as

E X t Za 0 2
Y I n

whee a alpha level of the CI the probability

of error with our CI
if 957 CI 2 0.05
9801 CI 2 0.02

x sample Meen
n a
sample size
Za standard normal at a probability of d
o population standard denaton

MI

si

72
L OV5

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
b one sided lower

A 10011 21 lower CI for y can be written as

X ta s s
g y

whee a alpha level of the CI the probability

of error with our CI
if 951 CI 2 0.05
9801 CI 2 0.02

x sample Meen
n sample size
Za standard normal at a probability of d
o population standard deviation

57

Za
2 0V5

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Hii Emir of a confidence Interval
the error of a confidence interval represents the quantity
we are adding or subtracting from x to form the CI

X E s e x t E two sided
y
y e x t E one sided upper

x E s
y one sided lower

therefore the Emirs are calculated as follows

a Emir E for a Two Sided CI

E 4 X X
Zha
y
E E t
24,9 Y 24,9

a Emir E for a One Sided CI

The error sided CI is the same for upper a

lower confidence forgery
E Za s e x t ta
no y G

iii length L of a two sided CI

Recall the general form of the two sided CI

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 X E E
Y E Xt E

L U

The length Lee of a two sided confidence interval is simply

the difference between the U l values

Length Lee U L
Itt T E

Lee
24,9 I
2E 2

3 Confidence Interval on population Meen y O Unknown

In must reel life cases particularly in quality control the
population process variance o is unknown
This requires us to use the sample standard deviation
s in all CI calculations

We identify two subceses depending on sample size n

case za when the sample size is large n 401

Case 2b when the sample size is small ne 401

case 29 when the sample size is large In 40

when a large sample size can be obtained n 40 the

Z parameter can still be used along with the sample standard
deviation

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
as follows

a Two Sided CI
A 10011 2 confidence interval for the mean y with o
unknown and a yo

X e X t
th 1
I
Zay sp s
y

Whee a alpha level of the CI the probability

of error with our CI
if 951 CI 2 0.05
9801 CI 2 0.02

x sample Meen
n sample size
7 normal a probability of
at a
Iggy standard deviation

If

2.5010 2.5

th's 742
a p.gg

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
b one sided Upper

A 10011 21 upper CI for y with o unknown and

n 401 can be written as

E X t ta 2
Y sp

Where a alpha level of the CI the probability

of error with our CI
if 957 CI 2 0.05
9801 CI 2 0.02

x sample Meen
n sample size
Za standard normal at a probability of a
s sample standard denaton

HM

51

ta
a o.rs

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
C one sided lower

A10011 21 lower CI for y with o unknown and 7340

can be written as

X th 3
is
s
y

Whee a alpha level of the CI the probability

of error with our CI
if 957 CI 2 0.05
9801 CI 2 0.02

x sample Meen
n sample size
Za standard normal at a probability of a
s population standard deviation

57

Za
2 0V5

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Iii Case 2b when the sample size is smell In 40
when 02 is unknown and the sample size n is less then
40 using the z valve is no longer suitable This is due
to the double uncertainty of not Inowing o and having a
small Sample size
As a result we need to use a new distribution

Student t T Distribution

The T table is shown in figure 4.2 on the following page

Some characteristics must be discussed

I the t distribution is shaped like a bell core and

is symmetrical
2 The distribution has a lower peek and wider variance
than the stanaera normal distribution
it will yield more conservative
results for the same probability

ta ta
3 The t valve is based on L N nu

a based on the confidence level required 110011 a

degrees of freedom n t

4 As the sample size n t v increases the t relies

approach the Z valves at the same probability

Same as the Z values

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Table 4.2. T-Distribution Table (v=n-1)

DOF

V NI

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
as follows

a Two Sided CI
A 10011 2 confidence interval for the mean y with o
unknown and na yo

X than i e X t
thin i 1
i I
t
y

Where a alpha level of the CI the probability

of error with our CI
if 957 CI 2 0.05
9801 CI 2 0.02

x sample Meen
n sample size
value a probability of
at a
thing standitinenution
standard deviation

IF

25010 2.5010

the in l thin l
a p.gg

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
b one sided Upper

A 10011 21 upper CI for y with o unknown and

n 401 cen be written as

X taint
f
s t 2
y

Where a alpha level of the CI the probability

of error with our CI
if 951 CI 2 0.05
9801 CI 2 0.02

x sample Meen
n sample size
stantiereastandaramatiation
s

MI

si

tain t
a o.us

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
C one sided lower

A10011 21 lower CI for y with o unknown and he 40

can be written as

x s s
taints y

whee a alpha level of the CI the probability

of error with our CI
if 957 CI 2 0.05
9801 CI 2 0.02

x sample Meen
n a
sample site
at a probability of d

Mi

Za
2 0V5

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 4.2. Confidence Interval

Construct a two-sided 90% confidence interval for the mean life in the accelerated
o test.

General concepts

I we went to establish a two sided 90 t confidante

the men battery life ly
interval for

2 901 lonhaence interval 10011 a Y CI

90 10011 4
us

3 Sind 82 is unknown we must Vil the t dismbono

lend n yo

step 1 Equation for the 901 CI

f to.us s I to.ve 9
In
s e t
In y

Step 2 Determine the Semple mean II a sample Std Dench07102

I 25.5 t 26.1 25.7

Igi
t
10

I 26.0 ht

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
714 Appendix

■ APPENDIX IV
Percentage Points of the t Distribution

0 tα,ν

α
v 0.40 0.25 0.10 0.05 0.025 0.01 0.005 0.0025 0.001 0.0005
1 0.325 1.000 3.078 6.314 12.706 31.821 63.657 127.32 318.31 636.62
2 0.289 0.816 1.886 2.920 4.303 6.965 9.925 14.089 23.326 31.598
3 0.277 0.765 1.638 2.353 3.182 4.541 5.841 7.453 10.213 12.924
4 0.271 0.741 1.533 2.132 2.776 3.747 4.604 5.598 7.173 8.610
5 0.267 0.727 1.476 2.015 2.571 3.365 4.032 4.773 5.893 6.869
6 0.265 0.727 1.440 1.943 2.447 3.143 3.707 4.317 5.208 5.959
7 0.263 0.711 1.415 1.895 2.365 2.998 3.49 4.019 4.785 5.408
8 0.262 0.706 1.397 1.860 2.306 2.896 3.355 3.833 4.501 5.041
9 0.261 0.703 1.383 1.833 2.262 2.821 3.250 3.690 4.297 4.781
10 0.260 0.700 1.372 1.812 2.228 2.764 3.169 3.581 4.144 4.587
11 0.260 0.697 1.363 1.796 2.20 2.718 3.106 3.497 4.025 4.437
12 0.259 0.695 1.356 1.782 2.179 2.681 3.055 3.428 3.930 4.318
13 0.259 0.694 1.350 1.771 2.160 2.650 3.012 3.372 3.852 4.221
14 0.258 0.692 1.345 1.761 2.145 2.624 2.977 3.326 3.787 4.140
15 0.258 0.691 1.341 1.753 2.131 2.602 2.947 3.286 3.733 4.073
16 0.258 0.690 1.337 1.746 2.120 2.583 2.921 3.252 3.686 4.015
17 0.257 0.689 1.333 1.740 2.110 2.567 2.898 3.222 3.646 3.965
18 0.257 0.688 1.330 1.734 2.101 2.552 2.878 3.197 3.610 3.992
19 0.257 0.688 1.328 1.729 2.093 2.539 2.861 3.174 3.579 3.883
20 0.257 0.687 1.325 1.725 2.086 2.528 2.845 3.153 3.552 3.850
21 0.257 0.686 1.323 1.721 2.080 2.518 2.831 3.135 3.527 3.819
22 0.256 0.686 1.321 1.717 2.074 2.508 2.819 3.119 3.505 3.792
23 0.256 0.685 1.319 1.714 2.069 2.500 2.807 3.104 3.485 3.767
24 0.256 0.685 1.318 1.711 2.064 2.492 2.797 3.091 3.467 3.745
25 0.256 0.684 1.316 1.708 2.060 2.485 2.787 3.078 3.450 3.725
26 0.256 0.684 1.315 1.706 2.056 2.479 2.779 3.067 3.435 3.707
27 0.256 0.684 1.314 1.703 2.052 2.473 2.771 3.057 3.421 3.690
28 0.256 0.683 1.313 1.701 2.048 2.467 2.763 3.047 3.408 3.674
29 0.256 0.683 1.311 1.699 2.045 2.462 2.756 3.038 3.396 3.659
30 0.256 0.683 1.310 1.697 2.042 2.457 2.750 3.030 3.385 3.646
40 0.255 0.681 1.303 1.684 2.021 2.423 2.704 2.971 3.307 3.551
60 0.254 0.679 1.296 1.671 2.000 2.390 2.660 2.915 3.232 3.460
120 0.254 0.677 1.289 1.658 1.980 2.358 2.617 2.860 3.160 3.373
q 0.253 0.674 1.282 1.645 1.960 2.326 2.576 2.807 3.090 3.291

v = degrees of freedom
Source: Adapted with permission from Biometrika Tables for Statisticians, Vol. 1, 3rd ed., by E. S. Pearson and H. O. Hartley,
Cambridge University Press, Cambridge, 1966.
 d

I n I no

Step 3 Find takin

I
0.05
I 01 to 05,9 1.83

M I 10 g g

step 4 Find 9010 II

26 1.833 1.63 t 26 3
yo Y 1.8331

25.06 s I 26.94
y

Interreleton we are 90t continent that the men better life

lies between 25.06 a 2694 hrs

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Lectures Hypothesis Testing Chenery
Objective

I To use a third technique called hypothesis testing in order

to mane conclusions related to population permits
2 To learn about me major wmponents end outcomes of a
hypothesis test including errors
3 N consult a hypothesis test to the population mean and
y
population variance o l and a population proportion Ip

General concepts

I In the past two chapters we here attempted to use statistics

to make conclusions about population parameters

Lectures point estimates

Chapters Y Y
Lecture 4 Interval Estimates L ey e u
Chipley I 102 s u

2 In this lecture that some population

we will hypothesize
parameter is some value we will then use some analysis
to see if this hypothesis is correct

I Hypothesis Testing

Many Engineering problems require which of two competing claims

about some population parameter are true
For example suppose that you are designing ball bearings and
you speculate that the men ball bearing diameter of all the
ball bearings is so mm

Question Is the men ball bearing diameter equal to so mm

or is it another valve

we cen send two types of hypothesis tests to answer this question

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
i types of Hypothesis tests

we have in general two types of hypothesis tests a Wu sicced

and b one sided

a two Sided Hypothesis Test

Null Hypothesis Ho y so
Alternative Hypothesis H t 50
y
not equal meers
it can be or
two sided

b One Sided Hypothesis Test

Null Hypothesis Ho y so
Alternchie Hypothesis H 50 OR y 50
y

Could be or
could be

Where Ho Null Hypothesis what is speculated to be true

H Alternative Hypothesis what we would choose if

we reject IN

ii testing the Hypothesis

so how we cen test our hypothesis

1 We would take a sample of ball bearings with

sample size h and calculate the sample meen I

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 2 We would establish a
region when we would fail
to reject Ho Suppose we would fail to reject it
48.5 X E 51.5 we would setup the regions shown
in Figure 5.1

Reject Ho Fail to Reject It Reject it

z 50 7 so 7150
48.5 51.5

critical
values

Figure 5.1. Two-Sided Hypothesis Test

iii Possible outcomes of a Hypothesis test

there ar 4 possible outcomes of a hypothesis test and two

of them are errors Table 5.1 summarizes the outcomes

Table 5.1. Summary of Possible Outcomes in a Hypothesis Test

s error

terror

i Error Type 1 Error

Rejecting the null hypothesis Ho when the null hypothesis

is true

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Pl Type I Emr d 1

ii B Error Type II Error

Not rejecting Ito when Ito is false incorrectly rejecting
the null hypothesis

Pl Type II Emir Pl not Rejecting Ito when Ho is false 2

Iv Calculating 4 A Emr
suppose we go back to our ball bearing hypothesis test where

Ho y 50

It y 50 Two Sided

suppose we take a sample of size 1 10 and our population

standard deviation was 0 2.5

a calculation of 2 Error
the a emir would be the probability of incorrectly rejecting
the null hypothesis Ito Graphically we can represent this probability
by the shaded region shown in Figure 5.2

a emir probability of rejecting Ito is error

PIT I 48.5 PI I 351.5

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Figure 5.2. Calculation of Type I (α) Error

i
p Zz s XT y t p Z E X y
o G
En En
P Zz 3 51.5 50 t P Z E 48.5 50
2.5 2.5
I w s I 10

d Error P Z 3 1.90 t P l 2 c l901

0.0287 t 0.0287

0 0574

How can we Decrease the 4 Error

we can widen the acceptance region By widening the acceptance

region we reduce our probability of rejecting Ito in error
This is not always good Stice p error will increase

2 Increase the sample size n l the best choice if possible

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
b calculation of B Error
Remember that
p error is

Pl Type II Emir Pl not Rejecting Ito when Ho is false

since the true mean is not so in order to calculate p

error the true men must be known

Suppose The true men is 52 how would the B emir

be celeucted

we would need to calculate the probability of not rejecting

Ito when Ho is false and the true men is 52

B Emr Pl 48.5 X 51.5 y 52

This is shown graphically in figure 5

48.5 50 51.5 52

True mean has shifted to 52

Figure 5.3. Graphical Representation of β Error

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 AN

48.5 SO 51.5 52
True mean of 52

AN

48.5 SO 51 51.5
True mean of 51

As the the man gets civier to the hypothesiced meen the probability
of 13 error increases

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 P X 51.5
p 48.5
y 52

Pl X 51.5 Pl X 48.5 Y 82 I

P Z L 51.5 52 P Z L 48.5 52
25 2.5
no I 10

P P Z 4.43
I z a 0.63 a

0.2643 O

B 0.2643

Notes concerning
A error
1 the p error increases
rapidly as the true men approaches the
hypothesized mean

2 B emir cen be reduced by vicresing sample size

3 Table 5.2 Shows the effect of changing critical values
and sample size on a a p errors

Table 5.2. Effect of Sample Size and Acceptance Region on α and β Error

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
VI Power of the test
The power a of statistical hypothesis test is the probability
of correctly rejecting the null hypothesis

Probability of correctly rejecting a false null hypothesis

Probability 07 not making a p error

Power I B

Gil p valve of a Hypothesis Test

The P valve represents the smallest level of significance that

would lead to the rejection of the null hypothesis

This will mace more sense when we actually conduct

a hypothesis test

2 Hypothesis Testing for the Meon Y

we have two cases for testing the Mun ri a hypothesis test
within each case a one sided or two sided hypothesis test
can be used

case 2A Hypothesis Test for with 02 known

y
case 2b Hypothesis test for y with 02 unknown

lil General steps and Guidelines for Hypothesis testing

1 Parameter of interest y 02 pl
2 Establish your null hypothesis Ho
3 Establish your alternative hypothesis Hi

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 decide on one sided or two sided

4 Take a sample of size n and calculate your test

statistic to To Xo Po
s state your rejection criteria either vic a level or
establishing critical values for X
6 Draw Conclusions Reject Ho
Do Not Reject Ho

ii case za for y
Hypothesis test o known

Figure 5.3 shows a summary for the calculation of the test

statistic and the acceptance do not reject and rejection criteria

a Based on a comparison of ta one sided or th

two sided versus the value of our test statistic Zo
OR

b comparison of the p valve to our d level

Slice o is known the standard normal z distribution can

be used

Zo I
Xp Mo
I n

whee X sample mean

yo speculated Meen in Ho

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Figure 5.3. Hypothesis Test on the Mean with Population Variance Known

al B timer for case he

the A emir for case ca cen be determined using the

equations below

it two sided

P O ta 0 Zay 2
Jin sign

Normal Probability Z-Value Z-Value

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Gil one sided lower

13 0 ta
sign
3

fit one sided Upper

0
13
It ta
sign 4

Where f yo y
Speculated mean in Ho Actual Mean

b Sample size n required

when continuing our own hypothesis test we might ask

ourselves what sample she is required a a gives a
error Ep

i two sided Test

N 2 da t Zp
2
20 s f n
y yo
f

Ii One sided test

N 2a 2 pl f b f
t
2 y Mo
j

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
ii case 2b for y
Hypothesis test o Unknown

figure 5.4 shows a summary for the calculation of the test

statistic and the acceptance do not reject and rejection criteria

a Based on a comparison of taint one sided or taint

two sided versus the value of our test statistic to
OR

b comparison of the p valve to our d level

Slice o is unknown the t test statistic to must be used

To X Yo I
s
I n

whee X sample mean

t
yo speculated Meen in Ho

Figure 5.4. Hypothesis Test on the Mean, Variance Unknown

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Appendix A/Statistical Tables and Charts
a
745

0 t a, n

V Percentage Points t α,v of the t Distribution

α
v .40 .25 .10 .05 .025 .01 .005 .0025 .001 .0005
1 .325 1.000 3.078 6.314 12.706 31.821 63.657 127.32 318.31 636.62
2 .289 .816 1.886 2.920 4.303 6.965 9.925 14.089 23.326 31.598
3 .277 .765 1.638 2.353 3.182 4.541 5.841 7.453 10.213 12.924
4 .271 .741 1.533 2.132 2.776 3.747 4.604 5.598 7.173 8.610
Degrees of
Freedom 5 .267 .727 1.476 2.015 2.571 3.365 4.032 4.773 5.893 6.869
= n-1 6 .265 .718 1.440 1.943 2.447 3.143 3.707 4.317 5.208 5.959
7 .263 .711 1.415 1.895 2.365 2.998 3.499 4.029 4.785 5.408
8 .262 .706 1.397 1.860 2.306 2.896 3.355 3.833 4.501 5.041
9 .261 .703 1.383 1.833 2.262 2.821 3.250 3.690 4.297 4.781
10 .260 .700 1.372 1.812 2.228 2.764 3.169 3.581 4.144 4.587
11 .260 .697 1.363 1.796 2.201 2.718 3.106 3.497 4.025 4.437
12 .259 .695 1.356 1.782 2.179 2.681 3.055 3.428 3.930 4.318
13 .259 .694 1.350 1.771 2.160 2.650 3.012 3.372 3.852 4.221
14 .258 .692 1.345 1.761 2.145 2.624 2.977 3.326 3.787 4.140
15 .258 .691 1.341 1.753 2.131 2.602 2.947 3.286 3.733 4.073
16 .258 .690 1.337 1.746 2.120 2.583 2.921 3.252 3.686 4.015
17 .257 .689 1.333 1.740 2.110 2.567 2.898 3.222 3.646 3.965
18 .257 .688 1.330 1.734 2.101 2.552 2.878 3.197 3.610 3.922
19 .257 .688 1.328 1.729 2.093 2.539 2.861 3.174 3.579 3.883
20 .257 .687 1.325 1.725 2.086 2.528 2.845 3.153 3.552 3.850
21 .257 .686 1.323 1.721 2.080 2.518 2.831 3.135 3.527 3.819
22 .256 .686 1.321 1.717 2.074 2.508 2.819 3.119 3.505 3.792
23 .256 .685 1.319 1.714 2.069 2.500 2.807 3.104 3.485 3.767
24 .256 .685 1.318 1.711 2.064 2.492 2.797 3.091 3.467 3.745
25 .256 .684 1.316 1.708 2.060 2.485 2.787 3.078 3.450 3.725
26 .256 .684 1.315 1.706 2.056 2.479 2.779 3.067 3.435 3.707
27 .256 .684 1.314 1.703 2.052 2.473 2.771 3.057 3.421 3.690
28 .256 .683 1.313 1.701 2.048 2.467 2.763 3.047 3.408 3.674
29 .256 .683 1.311 1.699 2.045 2.462 2.756 3.038 3.396 3.659
30 .256 .683 1.310 1.697 2.042 2.457 2.750 3.030 3.385 3.646
40 .255 .681 1.303 1.684 2.021 2.423 2.704 2.971 3.307 3.551
60 .254 .679 1.296 1.671 2.000 2.390 2.660 2.915 3.232 3.460
120 .254 = n-1
.677 1.289 1.658 1.980 2.358 2.617 2.860 3.160 3.373
∞ .253 .674 1.282 1.645 1.960 2.326 2.576 2.807 3.090 3.291

v = degrees of freedom.
Alobaidi, M. H. Chapter 9: Tests of Hypotheses for a Single Sample

Suppose that the analyst wishes to design the test so that if the true mean burning rate differs from
50 centimeters per second by as much as 1 centimeter per second, the test will detect this (i.e.,
reject : = 50) with a high probability, say, 0.90. Now, we note that = 2, = 51 − 50 = 1, =
0.05, and = 0.10. Since = = 1.96 and = = 1.28, the sample size required to
detect this departure from : = 50 is found by Equation 9-22 as

The approximation is good here, since

, which is small relative to .

BEE

Practical Interpretation: To achieve a much higher power of 0.90 you will need a considerably large
sample size, = 42 instead of = 25.

Using Operating Characteristic Curves

When performing sample size or type II error calculations, it is sometimes more convenient to use
the operating characteristic (OC) curves in Charts and . These curves plot as calculated
from Equation 9-20 against a parameter for various sample sizes . Curves are provided for both
= 0.05 and = 0.01. The parameter is defined as

How many standard deviations apart are the true

and hypothesized means.

so one set of operating characteristic curves can be used for all problems regardless of the values of
and . From examining the operating characteristic curves or from Equation 9-20, we note that

1. The further the true value of the mean is from , the smaller the probability of type II error
for a given and . That is, we see that for a specified sample size and , large
differences in the mean are easier to detect than small ones.
2. For a given and , the probability of type II error decreases as increases. That is, to
detect a specified difference in the mean, we may make the test more powerful by
increasing the sample size.

20
 83022020
0 3- Winter
McGill University CIVE 302 - Probabilistic Systems
0 3

Beta Error

( ) O.C. curves for different values of for the two-sided Normal test for a level of significance =
0.05.

Beta error

( ) O.C. curves for different values of for the two-sided Normal test for a level of significance =
0.01.

Example 9-4

Consider the propellant problem in Example 9-2. Suppose that the analyst is concerned about the
probability of type II error if the true mean burning rate is = 51 centimeters per second. We may

21
Problem 5.1

General concepts

I our data is normally distributed and O is known Measure

to can be used

step I construct the Hypothesis test

Ho y 40
It y 40 we want to see if battery life
exceeds 40 hours

step 2 Choose Calculate the test statistic

since we are doing a hypothesis test on the meen
y with
O known use to

Zo I 40 5 40
If 1 25
10

1.26

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
step 3 Determine th
Zo.us 1.64 1.64

Acceptance rejection region

Begin

0 20.05 164
20 1.26

Step 4 Make a conclusion

slice Zo Za
126 C 164

we do not reject Ho

there is no enaence showing that the men

battery life is greater then 40 hrs

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Step s p value part b

to no

Pl Zo 1.261 1 Pl Z C 1.26
I 0.896165

0.104

p value 0.104

p valve s a race we do not reject

step 6 B Emr part C

0
p
it ta
sign

whee f Mo y
Actual Mean
Speculated mean in Ho

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 p e
O 1.64 142 110
4

O 3.42

look up 2 3.42

p 0.000313

Step 7 Sample Size Refuted n

2 0.05
3 0.10

The meer hey

Z
N 2a Ep g
g2

Zo.us t Zo.io 1.2512

44 40 2

1.64 t 1.2812 1.2512

16

0.83

n I

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Step 8 An Equivalent CI
we would setup a 95 lower confidence interval as shr
below

I t
y
Zag
40.5 1.64 1251 E Y

39 85 E
y

we would conclude that with 95 confidence the men is

not greater than 40 this confirms the conclusion of the
hypothesis test

we would need about 90 confidence a 0.101 to conclude

that y 40

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 5.2.

General concepts

1 We are not Nld that our date is normally distributed We should

technically do a probability plot to determine whether the
data is normally distributed
For simplicity we will assume met me data is normally
distributed

2 Slice 0 is unknown we must use the t distributor

Step 1 Construct the Hypothesis Test

Ho 22.5
y
It y F 22.5 two sided

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Step 2 Find Sample Meen I Sample Std Denno S

x
ti
23.01 t 22.22 t 22.04 22.62 t 22.59
S

22.50

s xi It 123.01 27.1012 124,11 22.512 t 127.59 22.112

n I

0.3783

step 3 Determine To

To X Mo
s
i n

2.2.5 22.5 0
0.3783

4
Step Determine
then 1
L 0.05 OV25
L
n 1 4 to025,4 2.776

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Step 5 Make Conclusion

2.776 to.us To O 2 776 tov25.4

To c
takin p value I

i Do not reject Ho
Evidence shows that meen temperature 22.51

Step 6 Power of the test

Yo 22.5
y 22.75 d
7270
0.25 0.66
0.518

From ve 0.8
come on next page B

Power I B 0.2

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 B

d 0.66

d mo ul
o

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Step 7 Sample Siu Rewired
Power 0.9 13 0.1
i d 066
Y 22.75
Yo 22.5

FNM 0C Curve

n should be between 20 IO

n 30

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 5.5.

Stell I conduct a Hypothesis test two tidal

since our sample size is smell It 301 and we do not

linow O we must use the t distribution

Ho 12
It
y I 12 two sided
y
1
Q 20 0.05

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Step 2 Find Sample Meen I Sample Std Denno S

10.2 t 10.8 t 10.9 t 11.5 t 17.8

S 82 I 2.44

1.56

step 3 Determine To

To X Mo
s
i n

11.6 12 0.96
1.50

4
Step Determine
then 1
2 0.05 OV25
E
n 1 13 to025,1 2.160

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Step 5 Make Conclusion

2.160 to025,13 To 0.96 2160 tov25is

To tale n t

i Do not reject Ho
Evidence shows that the explitenon is realistic

stop 6 Contact C Twi Jina lonfidence In level

A 10011 2 confidence interval for the mean y with o

unknown and he yo

X tain x that
i
e 1
I
i s
y

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Whee a alpha level of the CI the probability
of error with our CI
if 957 CI 2 0.05
9801 CI 2 0.02

x sample Meen
n sample size
Zan standard normal at a probability of 42
s sample standard deviation

IF

2.5010 2.5

th's 742
a p.gg

11.6 2.160 1.561 E E 11.6 2.160 1.561

y

10.7 I E 12.5
Y

Sing 12 falls within the 95 t II Thi expliteto

of raising 1200 is Mesunibll

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Appendix A/Statistical Tables and Charts 745

0 t a, n

V Percentage Points t α,v of the t Distribution

α
v .40 .25 .10 .05 .025 .01 .005 .0025 .001 .0005
1 .325 1.000 3.078 6.314 12.706 31.821 63.657 127.32 318.31 636.62
2 .289 .816 1.886 2.920 4.303 6.965 9.925 14.089 23.326 31.598
3 .277 .765 1.638 2.353 3.182 4.541 5.841 7.453 10.213 12.924
4 .271 .741 1.533 2.132 2.776 3.747 4.604 5.598 7.173 8.610
5 .267 .727 1.476 2.015 2.571 3.365 4.032 4.773 5.893 6.869
6 .265 .718 1.440 1.943 2.447 3.143 3.707 4.317 5.208 5.959
7 .263 .711 1.415 1.895 2.365 2.998 3.499 4.029 4.785 5.408
8 .262 .706 1.397 1.860 2.306 2.896 3.355 3.833 4.501 5.041
9 .261 .703 1.383 1.833 2.262 2.821 3.250 3.690 4.297 4.781
10 .260 .700 1.372 1.812 2.228 2.764 3.169 3.581 4.144 4.587
11 .260 .697 1.363 1.796 2.201 2.718 3.106 3.497 4.025 4.437
12 .259 .695 1.356 1.782 2.179 2.681 3.055 3.428 3.930 4.318
13 .259 .694 1.350 1.771 2.160 2.650 3.012 3.372 3.852 4.221
14 .258 .692 1.345 1.761 2.145 2.624 2.977 3.326 3.787 4.140
15 .258 .691 1.341 1.753 2.131 2.602 2.947 3.286 3.733 4.073
16 .258 .690 1.337 1.746 2.120 2.583 2.921 3.252 3.686 4.015
17 .257 .689 1.333 1.740 2.110 2.567 2.898 3.222 3.646 3.965
18 .257 .688 1.330 1.734 2.101 2.552 2.878 3.197 3.610 3.922
19 .257 .688 1.328 1.729 2.093 2.539 2.861 3.174 3.579 3.883
20 .257 .687 1.325 1.725 2.086 2.528 2.845 3.153 3.552 3.850
21 .257 .686 1.323 1.721 2.080 2.518 2.831 3.135 3.527 3.819
22 .256 .686 1.321 1.717 2.074 2.508 2.819 3.119 3.505 3.792
23 .256 .685 1.319 1.714 2.069 2.500 2.807 3.104 3.485 3.767
24 .256 .685 1.318 1.711 2.064 2.492 2.797 3.091 3.467 3.745
25 .256 .684 1.316 1.708 2.060 2.485 2.787 3.078 3.450 3.725
26 .256 .684 1.315 1.706 2.056 2.479 2.779 3.067 3.435 3.707
27 .256 .684 1.314 1.703 2.052 2.473 2.771 3.057 3.421 3.690
28 .256 .683 1.313 1.701 2.048 2.467 2.763 3.047 3.408 3.674
29 .256 .683 1.311 1.699 2.045 2.462 2.756 3.038 3.396 3.659
30 .256 .683 1.310 1.697 2.042 2.457 2.750 3.030 3.385 3.646
40 .255 .681 1.303 1.684 2.021 2.423 2.704 2.971 3.307 3.551
60 .254 .679 1.296 1.671 2.000 2.390 2.660 2.915 3.232 3.460
120 .254 .677 1.289 1.658 1.980 2.358 2.617 2.860 3.160 3.373
∞ .253 .674 1.282 1.645 1.960 2.326 2.576 2.807 3.090 3.291

v = degrees of freedom.
Problem 5.5.

General concepts

I we are doing a hypothesis test for the men yl with

O known Z will be used

Step i construct the Hypothesis test

Ho Y 1.5 two sided

s
It y 1g

a dOl

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Step 2 Find Zo 242
242

4 0.01 0.005
I

Zo.us 2.57 table A II

Zo

Zo 1.4975 1.5 1.25

Ig 0.01

Stell I Make conclusion

we must compare our test statistic to to our 24

Zo.ws 2.57 20 1.25

20.995 2.57

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Do not reject Ito there is evidence that shows
Thet y 1.5

step 4 P reive

Zo 1.25 Zo 1.25

since we have a two sided test

P value 2 P I 20 C 1.25
2 0.105650

P valve 0.2113

step s power of the test

Power I
p

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 f Yo 7
1.5 1.495

O Out

0 O
B
I 2.57 0.008,125 I 2.57 0.005125
0.01

O 10.071 0 1 8.071

0.527903 O

B 0.527903

Power I
B 0.472097

Stell 6 Sample size Beavin'd n

n Zant G f y y

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Power I B
0.9 I B B 0.1 to 1 1 g

Zai Zoour 2.57

I is gg

i n 12.57 t 1.28 0 01 59.3

O00512

n 60

step 7 Constructing a two sided confidence Interval

since 2 0.01 we can produce a 99 confidence interval

I Zo.ws e e I t Zoout
En y f

1.4975 011 E
Y E 1.4975
2.571,10 2.12.10.011

1.4924 t e 1.5026
y

since our Ito y 1.5 falls within our confidence interval

199 1 we would not reject

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Appendix A/Statistical Tables and Charts 743

z 1 − 12 u2
Φ( z) = P( Z ≤ z) = ∫−∞ e du
2π

F (z)

0 z

III Cumulative Standard Normal Distribution (Continued)

z 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.09
0.0 0.500000 0.503989 0.507978 0.511967 0.515953 0.519939 0.532922 0.527903 0.531881 0.535856
0.1 0.539828 0.543795 0.547758 0.551717 0.555760 0.559618 0.563559 0.567495 0.571424 0.575345
0.2 0.579260 0.583166 0.587064 0.590954 0.594835 0.598706 0.602568 0.606420 0.610261 0.614092
0.3 0.617911 0.621719 0.625516 0.629300 0.633072 0.636831 0.640576 0.644309 0.648027 0.651732
0.4 0.655422 0.659097 0.662757 0.666402 0.670031 0.673645 0.677242 0.680822 0.684386 0.687933
0.5 0.691462 0.694974 0.698468 0.701944 0.705401 0.708840 0.712260 0.715661 0.719043 0.722405
0.6 0.725747 0.729069 0.732371 0.735653 0.738914 0.742154 0.745373 0.748571 0.751748 0.754903
0.7 0.758036 0.761148 0.764238 0.767305 0.770350 0.773373 0.776373 0.779350 0.782305 0.785236
0.8 0.788145 0.791030 0.793892 0.796731 0.799546 0.802338 0.805106 0.807850 0.810570 0.813267
0.9 0.815940 0.818589 0.821214 0.823815 0.826391 0.828944 0.831472 0.833977 0.836457 0.838913
1.0 0.841345 0.843752 0.846136 0.848495 0.850830 0.853141 0.855428 0.857690 0.859929 0.862143
1.1 0.864334 0.866500 0.868643 0.870762 0.872857 0.874928 0.876976 0.878999 0.881000 0.882977
1.2 0.884930 0.886860 0.888767 0.890651 0.892512 0.894350 0.896165 0.897958 0.899727 0.901475
1.3 0.903199 0.904902 0.906582 0.908241 0.909877 0.911492 0.913085 0.914657 0.916207 0.917736
1.4 0.919243 0.920730 0.922196 0.923641 0.925066 0.926471 0.927855 0.929219 0.930563 0.931888
1.5 0.933193 0.934478 0.935744 0.936992 0.938220 0.939429 0.940620 0.941792 0.942947 0.944083
1.6 0.945201 0.946301 0.947384 0.948449 0.949497 0.950529 0.951543 0.952540 0.953521 0.954486
1.7 0.955435 0.956367 0.957284 0.958185 0.959071 0.959941 0.960796 0.961636 0.962462 0.963273
1.8 0.964070 0.964852 0.965621 0.966375 0.967116 0.967843 0.968557 0.969258 0.969946 0.970621
1.9 0.971283 0.971933 0.972571 0.973197 0.973810 0.974412 0.975002 0.975581 0.976148 0.976705
2.0 0.977250 0.977784 0.978308 0.978822 0.979325 0.979818 0.980301 0.980774 0.981237 0.981691
2.1 0.982136 0.982571 0.982997 0.983414 0.983823 0.984222 0.984614 0.984997 0.985371 0.985738
2.2 0.986097 0.986447 0.986791 0.987126 0.987455 0.987776 0.988089 0.988396 0.988696 0.988989
2.3 0.989276 0.989556 0.989830 0.990097 0.990358 0.990613 0.990863 0.991106 0.991344 0.991576
2.4 0.991802 0.992024 0.992240 0.992451 0.992656 0.992857 0.993053 0.993244 0.993431 0.993613
2.5 0.993790 0.993963 0.994132 0.994297 0.994457 0.994614 0.994766 0.994915 0.995060 0.995201
2.6 0.995339 0.995473 0.995604 0.995731 0.995855 0.995975 0.996093 0.996207 0.996319 0.996427
2.7 0.996533 0.996636 0.996736 0.996833 0.996928 0.997020 0.997110 0.997197 0.997282 0.997365
2.8 0.997445 0.997523 0.997599 0.997673 0.997744 0.997814 0.997882 0.997948 0.998012 0.998074
2.9 0.998134 0.998193 0.998250 0.998305 0.998359 0.998411 0.998462 0.998511 0.998559 0.998605
3.0 0.998650 0.998694 0.998736 0.998777 0.998817 0.998856 0.998893 0.998930 0.998965 0.998999
3.1 0.999032 0.999065 0.999096 0.999126 0.999155 0.999184 0.999211 0.999238 0.999264 0.999289
3.2 0.999313 0.999336 0.999359 0.999381 0.999402 0.999423 0.999443 0.999462 0.999481 0.999499
3.3 0.999517 0.999533 0.999550 0.999566 0.999581 0.999596 0.999610 0.999624 0.999638 0.999650
3.4 0.999663 0.999675 0.999687 0.999698 0.999709 0.999720 0.999730 0.999740 0.999749 0.999758
3.5 0.999767 0.999776 0.999784 0.999792 0.999800 0.999807 0.999815 0.999821 0.999828 0.999835
3.6 0.999841 0.999847 0.999853 0.999858 0.999864 0.999869 0.999874 0.999879 0.999883 0.999888
3.7 0.999892 0.999896 0.999900 0.999904 0.999908 0.999912 0.999915 0.999918 0.999922 0.999925
3.8 0.999928 0.999931 0.999933 0.999936 0.999938 0.999941 0.999943 0.999946 0.999948 0.999950
3.9 0.999952 0.999954 0.999956 0.999958 0.999959 0.999961 0.999963 0.999964 0.999966 0.999967
 Appendix A/Statistical Tables and Charts 751

Chart VII Operating Characteristic Curves (Continued)

1.0

0.8

Probability of accepting H0
0.6

n
=
1
0.4

2
3
4
5 6
0.2

7 8
10
20 30
40 50

15
100
75

0
0 1 2 3 4 5
d

(a) O.C. curves for different values of n for the two-sided normal test for a level of
significance α = 0.05.

1.00

0.80
Probability of accepting H0

0.60

0.40
n
=
1

0.20
2
3
4
15

5
8 9
10

6
100

7
50

30

20
40
75

0
0 1 2 3 4
d

(b) O.C. curves for different values of n for the two-sided normal test for a level of
significance α = 0.01.
Source: Charts VIa, e, f , k, m, and q are reproduced with permission from “Operating Charac-
teristics for the Common Statistical Tests of Significance,” by C. L. Ferris, F. E. Grubbs, and
C. L. Weaver, Annals of Mathematical Statistics, June 1946.
Charts VIb, c, d, g, h, i, j , l, n, o, p, and r are reproduced with permission from Engineering
Statistics, 2nd Edition, by A. H. Bowker and G. J. Lieberman, Prentice-Hall, 1972.
752 Appendix A/Statistical Tables and Charts

Chart VII Operating Characteristic Curves (Continued)

1.00

Probability of accepting H0 0.80

0.60

0.40

n
=
1

0.20

2
3
4
5
7
30 40

6
15 20
8
50 75
100

9
10
0
–1.00 –0.50 0.0 0.50 1.00 1.50 2.00 2.50 3.00
d

(c) O.C. curves for different values of n for the one-sided normal test for a level of
significance α = 0.05.

1.00

0.80
Probability of accepting H0

0.60
n
=

0.40
1
2

0.20
3
4
5
6
15

7
9 10
30 40
50 75

8
100

20

0
–1.00 –0.50 0.0 0.50 1.00 1.50 2.00 2.50 3.00
d

(d) O.C. curves for different values of n for the one-sided normal test for a level of
significance α = 0.01.
 Appendix A/Statistical Tables and Charts 755

Chart VII Operating Characteristic Curves (Continued)

1.00

2
0.90

=
n

3
0.80

4
0.70
Probability of accepting H0

n=
2

5
0.60
3

6
7
8 4
0.50 9 5
10

6
0.40 15
20
0.30 30
40
0.20 50
75
0.10 100

0
0 0.40 0.80 1.00 1.20 1.60 2.00 2.40 2.80 3.20 3.60 4.00
l

(i) O.C. curves for different values of n for the two-sided chi-square test for a level of significance α = 0.05.

1.00
3

0.90
n=
4

0.80 n=
2
5

0.70
Probability of accepting H0

3
6

0.60 4
7
5

8
6

0.50 9
10
0.40 15
20
0.30 30
40
0.20 50
75
0.10 100

0
0 0.40 0.80 1.00 1.20 1.60 2.00 2.40 2.80 3.20 3.60 4.00
l

(j) O.C. curves for different values of n for the two-sided chi-square test for a level of significance α = 0.01.
Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Listen 6 i Methods a Philsophies of Statistical incest load in 51
Objectives

1 To teem the fundamental concerts for statistical process control

I variations whether a process is in control

2 To introduce you to the general model of a control chert

prelude to chapter 6 where we will look at all the different
control charts and their construction

General concepts

1 The general idles regarding process control must be understood

in order to make conclusions about a process i.e variability
in control out of control

2 The probability and statistics which were discussed in previous

lectures will be instrumental to the construction of control charts

This chapter represents an introduction to control cherts

I Statistics Process Control ISPs Basile

statistical process control uses statistical and probability calculations
and coupled with the use of control Charts in order to examine

al Procell Variability
1 Chance variability unavoidable causes Of variation
2 Assignchic louses of Variability
variability
machines methff Ypectors

b Process control

I process is it tonal making sure the only variability is

du to thin it

2 Process is out of control process is operating in the

presence of assignable levies of
vancho

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 11 Riduition or Elimination Of Variability
The goal of Spc is to identify the causes of variability in a
process fix them so that the variability is minimized or
totally eliminated

Figure 6.1 Shows graphically a process with chance and

assignable causes of variation

Increased
6 Variance

4 Shift in the
5 mean
3

Figure 6.1. Chance and Assignable Causes of Variation

Notes loncerning Figure b I

I LSI USL refer to lower specification limit and upper specification

limit respectively These represent the allowable variation Thet
the process on here
2 This graph is shoving assignable variations and their effect is
whether the process is in control or out of control

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
2 Basic Irrisiiles of a Control Chart
A control chart always has control limits set such that if
the process is in control nearly all points will lie between these
two limits
lower control limit LCL limits of a control
Upper Control Limit Lucy chart
Figure 6.2 Shows a typical control chert with a sample size ol n 18
All points lie between the LCL vice process is in control

In control process

• plot appears random

• all points are

between the control
limits

Figure 6.2. Typical Control Chart

mean of the sample
The centre line valve is the average of the sample and can
be expressed as
CL LCL t UCL
2

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
it out of Control Olivera
the process can be deemed to be out of control it
but not neleisenly

a At least one point plots outside of the control limits

b the points behave in a systematic or nonrandom manner indication
of a process Thet could be out of control

For practicality of a control chart will be given

a brief introduction
as an example This will be done just to show the general idle of a
control chart and the conclusions one cen make

Problem 6.1. General Control Chart Example

To assist in the understanding of a Control Chart, suppose we look at an important fabrication process in the
semi-conductor industry, called photolithography. This process consists of applying a light-sensitive
photoresist material using UV light. An important field characteristic is the flow width of the material when it
is applied. Suppose that the width can be controlled at a Mean of 1.5 microns and Standard Deviation of
0.15 microns. Every hour, a sample of 5 wafers are taken and the average width is calculated.

a) Establish the Control Limits

General concepts

1 Although a many different types of control charts for

there
demonstration purposes we will use the three signs want limits

SHD I Establish the control Limits

we will on the beat 30 limit representation

V11 My t
Joy
C1 MI 307

Recent met 2x Y 15

of 0.0671
441

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 I UCL 15 t 310.0671 1,70313

LCL 1.5 310.0671 1.2987

It the actual sample dele were given the average photoresist of tech
sample of 5 could be plotted to determine if the process is in control
Figure 6.3 shows a hypothetical plot

it sett

Each set has a sample size of 5 and the sample mean of each set is plotted

Figure 6.3. Three-Sigma Control Chart for UV Photoresists Width

For the hypothetical deter plot it cen be seen that the process is
in control

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 3 Stewart Control Charts General model of a control Chart

Stewart developed a general model for control charts and many different
control charts here been developed based on this model

lil General Model ol the control Chet

VCL Yu t Low
1
Centre Line Yw
LCL Yw LOW

Where L distance factor of the control limit from the centre line
Yw men of the sample statistic
ow standard deviation of the statistic

Figure 6.4 shows how a control Chet works Notice the difference
between the control chat limits and a standard normal distribution
with 2 1.5 and 8 0.15 The control chart shown is from problem
bl

Central
limit
theorem

Figure 6.4. Comparison of Normal Distribution with a Three-Sigma Control Chart

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Iiit types of truss ranebility
The type of variability exhibited by a process is important in a
diagnosis of whether a process is in or out of control Variability
can be classified as follows and are shown 15 Figure 6.6

Stationary Behavior process vanes around a fixed men in a

stable or predictable manner

implies an 15 control process

Stationary Uncorrelated observations here the apperance

of being drawn at random
Figure 6.6 a

Stationary An correlated successive Observations are

dependent on the previous
Observation Figure bb b

Non stationary Behavior is unstable and

process shows observations
Thel drift and wander
not hovering above and below the men
consistently

Stationary, Uncorrelated Stationary, Autocorrelated Non-Stationary

Figure 6.6. Variability in a Process

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Iv Types of Control Charts

There are two general types of control Charts

Types of Control
Charts

Variable Control Charts Attributes Control Charts

control charts on random control charts on random

variables that follow a variables that follow a
CONTINUOUS probability DISCRETE probability
distribution distribution

Y or o of a of detectives in a lot
normal distribution Ot products
process parameters
yes average of defectives is
RJ a lot

For both control chart types The larger the sample she and
freavency the better easier to detect small shifts is a process
Current practice smaller samples In is low
more frequent sampling m is high

V Average Ron Length ARL

a ARLO

The Average Run Length Arlo represents the number of samples that would
be sampled before an out of control signal falsealarm would be
generated

ARLO mem probably

P distribution I of trials until the first success

where p probability of any point to be outside the

control am

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 6.2. Calculating ARLo:

If the probability that a point on a process plots beyond the control limits is 0.0027, calculate the ARLo
and explain the result.

Since p 0.0027 Arlo 370

p gfoz

This implies than an out of control signal false alarm will occur
every 370 samples This also means that if you here more then one out
of control signal in the first 370 Samples one of thosepoints is probably
not a false alarm it is a true out of control signal
b ARL
The average run length Arlo uses the same concept but with respect
to detecting a shift is the men

ARL i
p

where p probability next sample following the

ther the
shift is able to detect the shifted mean
power 1 p B type II error I not rejecting Ho
detecting the when Ho is false
Shift inthe meen

ARLO and ARL with sample size since the LCL and VCL
vary
depend on sample size
probability of correctly detecting a shift in
the men
Figure bit shows p l B ther shifts is the mean will be
detected using different sample sizes and the 30 control chart
centred at 1.5

Figure 6.7 Shows the probability of detecting a shift in the

Meen when the Original men is 1.5
Each point on each graph is calculated by determining the B
error and using Y Yo S
true original
mean men

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 I A

honestly denting a shift in Mi mean

Figure 6.7. Probability of Detecting a Shift in the Mean versus Sample Size (Original Mean = 1.5)

Problem 6.3. ARL1 Calculation for a Shift in the Mean

Suppose that the true mean has shifted from 1.5 to 1.725. Determine the ARL1 for:

a) A sample size, n=5 (each set consists of 5 items)

b) A sample size, n=10 (each set consists of 10 items)

Use the graph shown in Figure 6.7, which is reproduced on the next page.

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 Est
l
P

sled I Determine P for each sample size for the new meen
For a new mean 1.725

n n p 0.1 11 81
n 5 P I V32 11 11

Step 2 Cellmate ARL for tech sample size

ARL n lo 10

ARL 11 5 3.125 2 4

Conclusion For n 10 take 10 samples of the 10 N detect a

shift in the men

For n s take 3.125 samples of size 5 to detect

a shift in the mean
24

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
IVI Type I Type II Emir in a control chert

similar to statistics in general we cen mere errors with respect

to whether our process is in control or not

a Type I Error : The probability of concluding the process is out of control, when in reality it is not.

• A point plots out of the control limits when in fact the process is still in control.

• High probability of Type I error or false alarms lead to excessive investigation costs, unnecessary process
adjustment and decreased credibility for SPC methods.

p Type II : The probability of concluding that the process is in control, when in reality, it is not.

• Results when a point plots within the control limits even though the process mean has shifted out of
control.

• This could result in an increase is non-conforming parts and extra costs.

a emir Producer's Rish

A era consumers Rish

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 6.5. Out of Control Process

stensteel Control limits control charts

YE
these limits are specified by statistics

LSI lower and upper specification limits

us
based on limits specified by the
chest and design engineer

Just because a process is it statistical control does not mean it will

meet specifications

Capability is a measure of process ability to meet specification

L LSL USC UCL

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Problem 9: Theory - Part I

Question 1:

In general, higher production quality:

a) leads to higher productivity since overall higher quality products will reduce production cost
b) leads to lower productivity since it takes more time to produce higher quality
c) has nothing to do with productivity
d) required more frequent operator interventions
e) none of the above

Question 2:

Quality improvement, from a statistical quality control point of view, can be considered as:

a) promoting a company's strong commitment to high quality products.

b) Increasing employee awareness on a quality program
c) reducing product variability
d) adding more attractive product feature
e) none of the above

Question 3:

In testing a product, it may be rendered that:

a) The product is defective as long as it has 1 or more defects Not required for
b) The product is not defective even if it has 1 or more defects midterm
c) The product is not defective even if has 2 or more defects
d) The product is conforming even if it has no defects.
e) None of the above

Question 4:

When multiple samples are taken from a process with a mean of X is calculated for each sample of su cient size:

a) X should follow a normal distribution if the process standard deviation is known.

b) X should follow a normal distribution even if the underlying distribution is not normal.
c) X should follow a normal distribution if the underlying distribution is normal.
d) X should follow a normal distribution only if the underlying distribution and process standard deviation is known
e) None of the above.

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Question 5:

In the use of a probability distribution in modelling the arrival process of an emergency room (ER) in a general hospital
serving a large population, assuming that the arrivals are independent of each other, generally:

a) The number of patients arriving to the ER should follow a Normal Distribution

b) The number of patients arriving to the ER should follow a Hypergeometric Distribution

c) The time between consecutive arrivals should follow a Poisson Distribution

d) The time between consecutive arrivals should follow an Exponential Distribution

e) None of the above

Question 6:

In a Six Sigma process, if sigma is the process standard deviation:

a) The design tolerances are to be widened to cover 6 times the standard deviation on each side of the process mean.

b) The process variability is to be reduced so that the designed tolerances will cover 6 times the standard deviation on
each side of the process mean.

c) The design tolerances are to be widened to cover a total range of 6 times the standard deviation centered at the
process mean.

d) The process variability is to be reduced so that the designed tolerances will cover a total range of 6 times the
standard deviation centered at the process mean.

e) None of the above.

Question 7:

In making an inference about the process mean:

a) Type I error is related to the probability that the process is at the targeted mean while the test result shows other wise

b) Type II error is related to the probability that the process is at the targeted mean while the test results show otherwise.

c) Type II error is related to the probability that the process is not at the targeted mean while the test results show
otherwise.

d) Type I and Type II errors are not allowed

e) None of the above

1a, 2c, 4b, 5c, 6d, 7b

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Midterm Sample Questions and Answers:

Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
 SAMPLE FORMULA SHEET FOR FINAL EXAM
1. Distributions
Bernoulli
p x 1
;   p ;   p(1  p )
2
p( x )  
1  p  q x0
Binomial
n! n
p x (1  p ) n  x =   p x (1  p ) n  x ;   np ,   np(1  p )
2
p( X  x ) 
x!( n  x )!  x
Hypergeometric

 D  N  D 
  
 x  n  x  nD nD  D  N  n 
, x = 0,1,2,… min (n,D);   ,  1  
2
p( x )  
N N N  N  N  1 
 
n 
Poisson
e  x
p( x )  , x  0,1,2,.... ;    ,  2  
x!
2. Confidence Interval
 
A) x  Z 2    x  Z 2
n n
S S
B) x  t 2 , n 1    x  t 2 ,n 1
n n
pˆ (1  pˆ ) pˆ (1  pˆ )
C) pˆ  Z 2  p  pˆ  Z 2
n n
3. Hypothesis Testing
A.1) Two-Sided: H 0 :   0 H 1 :    0 ; use: Z 0  x  0 and | Z 0 | Z 2
 n
A.2.1) One-sided: H 1 :    0 ; use: Z 0  x  0 and Z 0  Z
 n
A.2.2) One-sided: H 1 :   0 ; use: Z 0  x  0 and Z 0   Z
 n
B.1) Two-Sided: H 0 :   0 H 1 :    0 ; use: t0  x   0 and | t 0 | t 2,n 1
S n
B.2.1) One-sided: H 1 :    0 ; use: t0  x   0 and t 0  t ,n1
S n
x  0
B.2.2) One-sided: H 1 :   0 ; use: t 0  and t 0   t ,n1
S n
4. OC Functions
d |  | /  , 1   0   ,  0 is the target mean value
5. Variables Control Charts
x and R Control Charts
UCL= x  A2 R , Center Line = x , LCL= x  A2 R , UCL= D4 R , Center Line = R , LCL= D3 R
S Control Chart
UCL= B6 , Center Line = c4 , LCL= B5 ; UCL= B4 S , Center Line = S , LCL= B3 S
Average Run Length: ARL0  1 /  , ARL1  1 /(1   )

4
6. Attributes Control Charts

p Chart (with given standard p )

UCL= p  3 p (1  p ) n , Center Line = p , LCL= p  3 p (1  p ) n

np Chart (with given standard p )

UCL= np  3 np (1  p ) , Center Line = np , LCL= np  3 np (1  p )

p Chart, variable control limits (with given standard p )

UCL= p  3 p (1  p ) ni , Center Line = p , LCL= p  3 p (1  p ) ni
Use p if standard is not given

Standardized p Chart
UCL=3, Center Line =0, LCL= -3 , zi 
pˆ i  p . Use p if standard is not given
p (1  p ) ni
c Chart

UCL= c  3 c , Center Line = c , LCL= c  3 c . Use c if standard is not given

u Chart
ui  xi n , u   xi n  m , UCL= u  3 u n , Center Line = u , LCL= u  3 u n

u Chart, variable control limits

ui  x i n i , u   x i n i
, UCL= u  3 u ni , Center Line = u , LCL= u  3 u ni

Standardized u Chart
UCL=3, Center Line =0, LCL= -3 , zi 
ui  u
u ni
7. Capability Analysis

USL  LSL USL  LSL USL     LSL

Cp  , C pˆ  , C pu  , C pl  , C pk  min(C pu , C pl )
6 6ˆ 3 3
Cp  T 1 Cˆ p x T
C pm  with   , T  (USL  LSL) ; or Cˆ pm  with V 
1  2  2 1V 2 S
Sample Size and Critical Value Determination for Testing H 0 : C p  C p 0
    0.10     0.05
Sample Size C p ( High) / C p ( Low) C / C p ( Low) C p ( High ) / C p ( Low) C / C p ( Low)
10 1.88 1.27 2.26 1.37
20 1.53 1.20 1.73 1.26
30 1.41 1.16 1.55 1.21
40 1.34 1.14 1.46 1.18
50 1.30 1.13 1.40 1.16
60 1.27 1.11 1.36 1.15
70 1.25 1.10 1.33 1.14
80 1.23 1.10 1.30 1.13
90 1.21 1.10 1.28 1.12
100 1.20 1.09 1.26 1.11

5
8. Cusum Control Chart

Ci  max[ 0, xi  (  0  K )  Ci1 ] , Ci  max[0, (  0  K )  xi  Ci1 ] , C0  C0  0

UCL= 5 , Center Line =0, LCL=  5 , K  0.5 | 1   0 |

9. EWMA Chart
 
UCL   0  L [1  (1   ) 2i ] , Center Line   0 LCL   0  L [1  (1   ) 2i ]
(2   ) (2   )
 , Center Line   0 LCL    L  ; with 0.05    0.25 and L  3
UCL   0  L
(2   ) (2   )
0

10. Acceptance Sampling

AOQ  Pa p
ATI  n  (1  Pa )( N  n )

11. System Reliability

Ri  1  Fi , R( t )  e  t ,   1
MTTF

6
Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury
Upgrade Tutorials © 3495 de la Montagne, Suite 609, Montreal, PQ, H3G 2A5 By Asif Choudhury