Probability and Statistics
Basic Probability concepts
Most inspection and quality control theory deals with statistics to make inference about a population based on
information contained in samples. The mechanism we use to make these inferences is probabi
Analyze Phase
1
Analyze Phase Elements
Identify Sources
of Variation
Use Analysis Tools
Select Analysis
Tools
Analyze
Basic Tools and Methods
Measuring and modeling the relationship between
Variables
Hypothesis Testing
Failure mode and effects analysis
IE 7610 Fundamentals of 6 Sigma
Objectives
Develop conceptual understanding of important aspects of six sigma
Develop basic skills in statistical data analysis
Develop basic skills in problem definition and analysis by using six sigma tools
Covers the
Introduction: Six Sigma
Six Sigma is a proven business strategy (structured according to the DMAIC phases) to
measure, analyze and improve the performance in terms of operational excellence.
The methodology, thanks to a wide range of qualitative and quant
Six Sigma: Define Phase
Define Phase Outcomes
Define the problem with a HighLevel Problem Statement
Specifically identify the process or product customers
impacted by the problem
Define CTQs (Critical to Quality) characteristics from the
customers poi
Team Management
Types of Teams
Improvement Teams
A group belonging to any department chooses to solve a quality/productivity problem. It will continue until a
reasonable solution is found and implemented. The problem may be management selected but the sol
Analysis of Variance and Comparative Experiment
An example:
An independent automobile evaluation company wants to compare the qualities of 4
different cars, Ford Taurus, Chevy Lumina, Honda Accord and Toyota Camry. Evaluation
cars are driven and evaluated
PM NETWORK JJRNUARI 2009 WWWPMLORG
Robert Forest faced some tough bar
gaining during the initial contract
negotiations for the Pearl River
Tower in Guangzhou. China while
an associate partner at Skidmore,
Owings 8C Merrill. Scheduled
For completion with
Basic Statistics
Objective:
Represents the main statistical properties of a set of data (sample
or population)
Lean6
Characteristics:
LEAN SIX SIGMA MINIBOOK
Location parameters
Mean, Mode, Median, Quartiles, Percentiles
Dispersion parameters
Range,
Legup Evei yeti on't follows
their advice, Its the best way to
keep team members on track.
by Elisa Ludwig * photo by Debbie Zimelman it comes right down to it, project
teams function much like a benevolent
dictatorship. Its up to project man
agers to ma
T THE
PEIP
IT'S A FINE LINE
BETWEEN CONTROL
FREAK AND
DISENOAGED
PUPPETEER. BUT
DELEGATINO CAN
HELP PROJECT
LEADERS SURVIVE IN
TODAY'S 24/7 GLOBAL
PROJECT WORLD. ustly or not, delegating is
stuck with a bad reputa
tion. To the Cynic, it can
come across as
a
n
I

_ _ 2w Ta Van, Vietnam
. train! a
: l, 'V II I
I A
r ,
, c,
a. _._
:11! Mum
_ . " o " 'Hf :Nl u ' I
I ' i I iun, .I1
I I I m LEI K
_ . a ' 0 1" I I _ I
16
h EmuN .
*We wanted to demonstrate that broadband can be done at a reasona
IHZIIIIHBER 2U"? \ \ITlmlllih 49 50 \I \!'.I'\HJIcfw_I\
Crunch Time
is no] a partinularly plowi
an! IXpi."ltlll't?. Long cfw_lays.
nights and \t'tlktlltl Spt'll! toilin;_r
Io [Ill't'l tight deadlines can erodv team
\k motivation and ultimately
Youve assembled
a fine crew of
professionals
ready, willing
and able to do
whatever it
takes to get the
project done.
Now how
do "you
keep II 1.0m.
that may?
team
by Simon Kent (11' illustration by Otto Steininger
38 \1\l~:l'\"l)cfw_h \UHJHIIJI 2m)? \
d
d
C
J
a
.1
C
F
a
V).
b 'Ihinking outside the border
is becoming a prerequisite among today's business
('ognoscenti. The emergence of an integrated
worldwide economy is spawning a slew of projects
in farung localesopening up a wealth of oppor
tunities
Bye] nfCuugrwa Puple Project managers
are in their own
race to handle a
mindnumbing mix
of details under
doordie deadlines.
II B4 \ \li'lWtJilK
The Big Event
This anicie is based on material in the white paper. "Event Management St the PMBOK
IGu
General Full Factorial vs 2 level full factorial
The number of experiments needed in a general full factorial design
will increase VERY fast with the number of factors, (variables), if we
have 4 factors, say A,B,C,D, A has 4 levels, B has 5 levels, C has
Estimation of Model Parameters
yijk i j ( ) ij ijk
y.
i yi. y.
j y. j. y.
( ) ij yij. yi. y. j. y.
Fitted Value
yijk i j ( ) ij yij.
The assumption of No Interaction in a two factor model
yijk i j ijk
Twoway ANOVA: Battery Life versus Material, Tempe
Design of Experiments
For
Several Factors
A Typical Process
Key Performance Metrics
(Response)
Input Materials
(Factors)
Output Materials
Process
Process
Parameters
(Factors)
(Product or Intermediate)
One Factor at a time
X1 (parameter1)
X2 (parameter 2)
Response: Rocket propellant burning rate
A,B,C,D,E: Formulations of propellant, the primary factor
For Latin square designs, the 2 nuisance factors are divided into a tabular
grid with the property that each row and each column receive each
treatment exac
_
BY
JOAN KNUTSON.
CONTRIBUTING EDITOR
FacilitationThe Core Competency
Needed in Conflict Resolution
Longer hours, greater stress and shorter tempers all increase conflict. Expect increased
clashes during these troubled times and be prepared to intervene
Confidence Intervals
Confidence Interval for mi
yi. ~ N ( m i ,
y i. m i
MSE / n
2
n
)
~ t N a
Confidence Interval for
yi. t
2
, N a
mi
:
MS E
n
Individual 95% CIs For Mean Based on
Pooled StDev
Level N Mean StDev ++++15 5 9.800 3.347 (*)
20 5 15.
The Blocking Principle
Blocking is a technique for dealing with nuisance
factors
A nuisance factor is a factor that probably has some
effect on the response, but its of no interest to the
experimenterhowever, the variability it transmits to
the response
Why Experimental Design?
Key Performance Metrics
(Response)
Input Materials
(Factors)
Output Materials
Process
(Product or Intermediate)
Process
Parameters
(Factors)
To find out the cause effect relationship in a process:
y=f(x1,x2,xn)+e
That is, to find