Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
Experiment)
Made By:
ISHA JAIN
NIDHI GAHLOT
Division of MPAE
Strategy of Experimentation
One factor at a time approach: keep
all other factors constant and change
any one (say A). This gives main
effect A ONLY.
Factorial: gives main effects as well
as interaction.
Basic Principles of
Experimentation
Basic principles
What do they mean? Why do we do them?
replication
Repetition of basic
experiment. NOT same
as repeated
measurements
Improves validity of
DOE
Reflects variability b/w
runs
randomisation
Allocation of
experimental material
and order of runs in
random
Assists in averaging
out the effects of
extraneous factors
blocking
A design technique to
improve precision with
which comparisons
among factors on
interest are made
Reduces effect of
nuisance factors
factors
levels
x y = (no. of levels) (no. of factors)
Main effect
Interaction
Regression Model
Refers to the equation establishing
quantitative relationship b/w factors
of interest (A & B) and response (y)
Interaction Plots
One factor interaction plot
Two factor interaction plot
Let us study two factor interaction
plot
Degrees
of
freedom
Mean
square
F - Value
P - Value
SSA
(a-1)
MSA
FA
PA
SSB
(b-1)
MSB
FB
PB
AB
SSAB
(a-1)(b-1)
MSAB
FC
PC
Error
SSE
ab(n-1)
MSE
Total
SST
abn-1
23 = 8
Consider 23= 8
half of 8= 4
for fractional factorial, we will perform 4
runs
ONLY.
AB
AC
BC
ABC
abc
ab
ac
bc
I = ABC
Known as principle
fraction
lA A + BC
lB B + AC
lc C + AB
I= -ABC
Known as alternate
fraction
lA A - BC
lB B - AC
lc C - AB
Sequential experimentation
Projection Property
A one half fractional
factorial design for 23
factorial can be perceived
as 22 full factorial design.