You are on page 1of 48

Design of

Experiments
Dr. James Wakiru
Mechanical Engineering

DR J.W 16-1
DATA ANALYSIS

➢ A credible amount of data has to be collected to allow for a substantial


analysis
➢ Information collected must now be described, analyzed and interpreted
➢ You have to look beyond the raw data to ask important questions about
what the results mean and whether they are significant
➢ You have to check the data to ensure that it is “clean” and look
for inconsistenciess

DR J.W 16-2
SURVEY DATA ANALYSIS – Descriptive

DR J.W 16-3
DATA ANALYSIS

Wakiru et al. (2020). Towards an innovative Lubricant Condition Monitoring strategy for maintenance of ageing multi-unit systems.
Reliability Engineering & System Safety 204: 107200.

DR J.W 16-4
DATA ANALYSIS – Qualitative data

Steps for evaluating qualitative data

1. Read through the responses.


2. Create response categories.
3. Label each comment with one or several categories.
4. Look at what you have.
5. Think what are the responses about?
6. Identifying the patterns and trends.
Try and avoid turning qualitative data into quantitative data!

DR J.W 16-5
DATA ANALYSIS TO REPORTING

➢ As you analyse your data, you are arriving at your own assessment of what
your findings mean -key questions to ask:
• Patterns and themes are emerging – do you have enough evidence to justify
your conclusions?
• Have you ruled out other explanations and examined
exceptions to the patterns you are seeing?
• Do you need to run correlation analysis (SPSS) to test if the differences
observed are statistically significant?
• Do you need to have a colleague double-check your findings against your
data?

➢ You should then be ready to move to writing your report

DR J.W 16-6
Design of Experiments (DoE)
• Design of Experiments (DOE) is a multi-purpose technique. It
consists in a series of tests in which changes are made to input
variables of a process, so that we may observe and identify
corresponding changes in the output response.
• It can be applied to physical experiments, to simulation
experiments or to other decision problems where the effects of
several factors are examined.
• Experimental Design methods may be used to design processes or
products, to improve process performance or to obtain a robust
process, insensitive to sources of external variations.

DR J.W
Design of Experiments (DoE)
• Collection of statistical techniques providing systematic
way to sample design space
• Useful when tackling new problem in which you know
very little about design space
• Study effects of multiple input variables on 1 or more
output parameters

DR J.W
Design of Experiments (DoE)
DOE:DesignOf Experiments
•Experiment: a learning process for
▪ knowledge gathering,
▪ problem solving,
▪ conjecture testing
•DOE offers a systematical procedure to
▪ construct efficient experiments,
▪ obtain accurate conclusions,
▪ economically consume resources

DR J.W 16-9
Design of Experiments (DoE) (cont.)
• Often used before setting up formal optimization problem
– Identify key drivers among potential design variables
– Identify appropriate design variable ranges
– Identify achievable objective function values
• Often, DoE used in context of robust design
• Today, discussion is about design space exploration

DR J.W 16-10
Design of Experiments (DoE) (cont.)
• (controllable)factors: process input variables that an investigator can
manipulate to cause a change in theoutput
• qualitative (also called nominal)
• quantitative
• Levels: settings or values of a factor in the experiment
• Treatments: combinations of levels for all(controllable)factors in a trial
• Response: the output variable that provides useful information about
the system/process

Factors
& “Experiment” Observation
levels
(Often analysis
DR J.Wcode) 16-11
DOE Guidlines
• Statement of the problem
• Choice of the factors to be varied (independent variables)
and the levels to be investigated
• Selection of the response (dependent variable)
• Choice of the experimental design
• Performing the experiment/series of experiments
• Data analysis
• Conclusions and recommendations

DR J.W 16-12
Basic DOE principles
• Randomization: Each unit of the investigated population
has the same chance of being selected in a sample of
units on which the experiment is performed.
• Statistical replications: As measurements are subject to
measurement uncertainty, to better estimate the true
effects of treatments, experiments are replicated.

DR J.W 16-13
Basic DOE principles
• Blocking: Arrangement of experimental units into
homogeneous groups to reduce variation between units
within such group and thus achieve greater estimating
precision.
• Orthogonality: An experimental design is orthogonal if
each factor can be evaluated independently at all the
other factors. It is achieved by matching each level of each
factor with an equal number of each level of the other
factors.

DR J.W 16-14
Design of Experiments (DoE) (cont.)
• Factors & variables
– Design variables = factors
– Values of design variables = levels
– Noise factors = variables over which we have no control:
e.g., manufacturing variation in blade thickness
– Control factors = variables we can control
e.g., nominal blade thickness
– Outputs = observations (= objective functions)

Factors
& “Experiment” Observation
levels
(Often analysis code)

DR J.W 16-15
Matrix Experiments
• Each row of matrix corresponds to 1 experiment
• Each column of matrix corresponds to 1 factor
• Each experiment corresponds to different combination of factor levels &
provides 1 observation

Expt No. Factor A Factor B Observation

1 A1 B1 h1
2 A1 B2 h2
3 A2 B1 h3
4 A2 B2 h4

Here, we have 2 factors, each of which can take 2 levels

DR J.W 16-16
OFAT
• The one-factor-at-a-time method, also known as one-
variable-at-a-time, OFAT, OF@T, OFaaT, OVAT, OV@T, OVaaT,
or monothetic analysis
• It is a method of designing experiments involving the
testing of factors, or causes, one at a time instead of
multiple factors simultaneously.
• The approach adopted is to change one cause, ceteris
paribus i.e. while keeping everything else (all other causes)
constant. Hypothesis testing is the most commonly used
tool for OFAT.

DR J.W 16-17
One Factor At a Time (OFAT)
• Change 1st factor, all others at base value
• If output is improved, keep new level for that factor
• Move on to next factor & repeat
Expt Factor
n factors No. A B C D
1+n(l – 1) 1 A1 B1 C1 D1
evaluations 2 A2 B1 C1 D1
l levels 3 A3 B1 C1 D1
4 A* B2 C1 D1
4 factors, 3 levels each: 5 A* B3 C1 D1
1 + n(l – 1) = 6 A* B* C2 D1
7 A* B* C3 D1
1 + 4(3 – 1) = 9 expts
8 A* B* C* D2

• Result depends on order of factors 9 A* B* C* D3

DR J.W 16-18
OFAT 100

95

90
1800
1600
1400
f Ao Tr

Ao (%)
1200

Tr (Hrs)
70 69.52 1193 85
1000
75 68.603 1166 80
80 72.8 1250 800
75
85 83.86 1403 600
Ao Tr
90 90 1526 70 400
92 89.8 1504 65 200
94 87.4 1474 60 0
96 91.7 1593 70 75 80 85 90 92 94 96 98 99 100
98 93.2 1572 f
99 94.7 1640
100 95.1 1656 94 2000
Ao Tr
1800
tbo Ao Tr 92
1600
7000 82.8 1763 90
1400
7500 85.6 1682 88 1200
8000 91.6 1735

Tr (Hrs)
Ao (%)

86 1000
8500 85.5 1517
84 800
9000 90 1526
600
9500 83.4 1375 82
400
10000 87 1376 80 200
10500 89.97 1360
78 0
11000 84.4 1187
7000 7500 8000 8500 9000 9500 1000010500110001150012000
11500 87.6 1197
12000 87.56 1176 tbo (Hrs)

Wakiru, et.al.,(2019). A simulation-based optimization approach evaluating


maintenance and spare parts demand interaction effects. International Journal of DR J.W 16-19
Production Economics, 208:329-342.
Factorial experiments
• A full factorial experiment is an experiment whose design
consists of two or more independent variables (factors),
each with discrete possible values or ‘levels’, and whose
experimental units take on all possible combinations of
these levels across all such factors.

DR J.W 16-20
Factorial experiments
Why use
1. Factorial designs can test limits;
to test whether an independent variable effects different
kinds of people, or people in different situations, the
same way.
2. Factorial designs can test theories;
can test generalizability of a causal variable and also test
theories.

DR J.W 16-21
Factorial experiments
Advantages
• Ability to study interactions between independent variables
• Significantly more efficient than OFAT
• Provide more information at a similar or lower cost.
• It can also help find optimal conditions quicker than OFAT
experiments can.

Disadvantages
• More chance of having confounds than single IV designs &
same problems controlling for them.
• Interpretations are no better than correlational studies if
factors are not manipulated.
DR J.W 16-22
Full-Factorial Experiment
• Specify levels for each factor
• Evaluate outputs at every combination of values
– Complete, but expensive!

n factors Factor
Expt No.
A B
ln observations
1 A1 B1
l levels 2 A1 B2
3 A1 B3
2 factors, 3 levels each: 4 A2 B1
ln = 32 = 9 expts 5 A2 B2
6 A2 B3
4 factors, 3 levels each: 7 A3 B1
8 A3 B2
ln = 34 = 81 expts
9 A3 B3

DR J.W 16-23
Fractional Factorial Experiments
• Due to combinatorial explosion, you cannot usually perform
full factorial experiment
• Instead, consider just some possible combinations
• Questions to be answered
– How many experiments do I need?
– Which combination of levels should I choose?
• Need to balance experiment cost with design space coverage

DR J.W 16-24
DoE Techniques Overview
TECHNIQUE COMMENT EXPENSE
(l=# levels, n=# factors)
Full factorial design Evaluates all possible ln grows exponentially with
designs number of factors

Orthogonal arrays Don’t always seem to Moderate—depends on


work - interactions? which array

1 at a time (OFAT) Order of factors? 1 + n(l – 1)—cheap

Latin hypercubes Not reproducible, poor l—cheap


coverage if divisions are
large
Parameter study Captures no interactions 1 + n(l – 1)—cheap

DR J.W 16-25
Parameter Study
• Specify levels for each factor
• Change 1 factor at a time, all others at base level
• Consider each factor at every level Expt Factor
No. A B C D
1 A1 B1 C1 D1
n factors 2 A2 B1 C1 D1
1 + n(l – 1) 3 A3 B1 C1 D1
evaluations 4 A1 B2 C1 D1
l levels 5 A1 B3 C1 D1
6 A1 B1 C2 D1
4 factors, 3 levels each: 7 A1 B1 C3 D1
8 A1 B1 C1 D2
1 + n(l – 1) =
9 A1 B1 C1 D3
1 + 4(3 – 1) = 9 expts
Baseline : A1, B1, C1, D1

DR J.W 16-26
Parameter Study (cont.)
• Select the best result for each factor

Expt Factor Observatio


No. A B C D n 1. Compare h1, h2, h3
1 A1 B1 C1 D1 h1  A*
2 A2 B1 C1 D1 h2 2. Compare h1, h4, h5
3 A3 B1 C1 D1 h3  B*
4 A1 B2 C1 D1 h4 3. Compare h1, h6, h7
5 A1 B3 C1 D1 h5  C*
6 A1 B1 C2 D1 h6 4. Compare h1, h8, h9
7 A1 B1 C3 D1 h7  D*
h8
8 A1 B1 C1 D2
“Best design” is
9 A1 B1 C1 D3 h9
A*, B*, C*, D*
• Does not capture interaction between variables

DR J.W 16-27
Orthogonal Arrays
• Specify levels for each factor
• Use arrays to choose subset of full-factorial experiment
• Subset selected to maintain orthogonality between factors

n factors
subset of ln evaluations
l levels

• Does not capture all interactions, but is efficient


• Experiment is balanced

DR J.W 16-28
Orthogonal Arrays (cont.)
Expt Factor
Factor No. A B C D
Expt
No. A B C 1 A1 B1 C1 D1

1 A1 B1 C1 2 A1 B2 C2 D2
3 A1 B3 C3 D3
2 A1 B2 C2
4 A2 B1 C2 D3
3 A2 B1 C2 5 A2 B2 C3 D1
4 A2 B2 C1 6 A2 B3 C1 D2
7 A3 B1 C3 D2

L4(23) 8
9
A3
A3
B2
B3
C1
C2
D3
D1

4 expts 3 factors L9(34)

2 levels
9 expts 4 factors

DR J.W
3 levels 16-29
Effects
• Once experiments have been performed, results can be
used to calculate effects
• Effect of factor is change in response as level of factor is
changed:
– Main effects: averaged individual measures of effects of
factors
– Interaction effects: effect of factor depends on level of
another factor
• Often, effect is determined for change from minus level (–)
to plus level (+) (2-level experiments)

DR J.W 16-30
Effects (cont.)

• Consider this experiment:


– We are studying effect of 3 factors on price of aircraft
– Factors are number of seats, range & aircraft manufacturer
– Each factor can take 2 levels:
Factor 1: Seats 100 < S1 < 150 150 < S2 < 200

Factor 2: Range (nm) 2,000 < R1 < 2,800 2,800 < R2 < 3,500

Factor 3: Manufacturer M1 = Boeing M2 = Airbus

DR J.W 16-31
Main Effects
• Main effect of factor is effect of that factor on output averaged
across levels of other factors
Expt Seats Range Mfr Price
No. (S) (R) (M) (observation)
1 S1 R1 M1 P1
L8(23) 2 S1 R1 M2 P2
(full factorial 3 S1 R2 M1 P3
design)
4 S1 R2 M2 P4
5 S2 R1 M1 P5
6 S2 R1 M2 P6
7 S2 R2 M1 P7
8 S2 R2 M2 P8
DR J.W 16-32
Main Effects (cont.)
• Question: What is main effect of manufacturer?
• That is, from experiments, can we predict how price is affected by whether
Boeing or Airbus makes aircraft?

Expt Seats Range Mfr Price


No. (S) (R) (M) (observation)
1 S1 R1 M1 P1 Expts 1 & 2 differ only
2 S1 R1 M2 P2 in manufacturer
3 S1 R2 M1 P3
4 S1 R2 M2 P4
5 S2 R1 M1 P5
6 S2 R1 M2 P6
7 S2 R2 M1 P7
8 S2 R2 M2 P8

(P2 -P1)+ (P4 -P3 )+ (P6 -P5 )+ (P8 -P7 ) Main effect of
4 manufacturer
DR J.W 16-33
Main Effects: Another Interpretation
Overall mean P1 + P2 + P3 + P4 + P5 + P6 + P7 + P8
response: m=
8
Avg over all expts when P1 + P3 + P5 + P7
M = M1: mM 1 =
4

Effect of mfr
level M1
= mM 1 − m Effect of mfr = mM 2 − m
level M2

Effect of factor level can be defined for multiple levels

Main effect of factor is defined as


Main effect of mfr
= mM 2 − mM 1 difference between 2 levels

NOTE: Main effect should be interpreted individually only if variable does not
appear to interact with other variables
DR J.W 16-34
Main Effect Example
Expt Seats Range Mfr Price
Aircraft
No. (S) (R) (M) ($M)
1 717 S1 R1 M1 24.0
2 A318-100 S1 R1 M2 29.3
3 737-700 S1 R2 M1 33.0
4 A319-100 S1 R2 M2 35.0
5 737-900 S2 R1 M1 43.7
6 A321-200 S2 R1 M2 48.0
7 737-800 S2 R2 M1 39.1
8 A320-200 S2 R2 M2 38.0

Sources:
Seats/Range data: Boeing Quick Looks 100 < S1 < 150 150 < S2 < 200
Price data: Aircraft Value News 2,000 < R1 < 2,800 2,800 < R2 < 3,500
Airline Monitor, May 2001. Reprinted with M1 = Boeing M2 = Airbus
permission of Prof. Olivier de Weck

DR J.W 16-35
Interaction Effects
• Can also measure interaction effects between factors
• Answers question: Does effect of factor depend on level of another factor?
– For example, does effect of manufacturer depend on whether we
consider shorter range or longer range aircraft?

Avg mfr effect Avg mfr effect with


Mfr  range with range 2 – range 1
interaction =
2

DR J.W 16-36
Analysis of variance -ANOVA
▪ Investigator controls one or more factors of interest
◦ Each factor contains two or more levels
◦ Levels can be numerical or categorical
◦ Different levels produce different groups
◦ Think of each group as a sample from a different
◦ population

▪ Observe effects on the dependent variable


◦ Are the groups the same?

▪ Experimental design: the plan used to collect the data

DR J.W 16-37
Hypotheses OFAT ANOVA

DR J.W 16-38
One Factor ANOVA

DR J.W 16-39
Variation - ANOVA
Total Variation = the aggregate variation
of the individual data values across the
various factor levels (SST)
Among-Group Variation = variation
among the factor sample means (SSA)
(SSR)
Within-Group Variation = variation that
exists among the data values within a
particular factor level (SSW) (SSE)

DR J.W 16-40
One-Way ANOVA Table

The F statistic is equal to the regression mean square (MSR/MSA) divided by


the error mean square (MSE/MSW).

DR J.W 16-41
One-Way ANOVA Table
Is the size of store influencing the annual sales?
Independent variable/Response: Annual sales
Dependent variable: Size of store
H0: Size of store is not significantly related to annual sales

Using a level of significance of 0.05, from Table, the critical value of the F distribution,
with 1 and 12 degrees of freedom. Because F = 113.2335 > 4.75 or because the p-value
= 0.0000 < 0.05, you reject H0 and conclude that the size of the store is significantly
related to annual sales.

DR J.W 16-42
Summary
▪ Role of designing experiments to explore design space
▪ With simple models or prototypes:
o Conduct simple experiments to understand main effects of model
o Understand interaction effects between factors

▪ Statistical techniques provide systematic way to sample design


space

DR J.W 16-43
Ave
Time to Spares Major repair Technician Engine Repair
Name Overhaul availability diagnosis time Capacity MA4IF availability Time
Full Factorial Experiment Base 9000 90 2 2 0.75 71.96 1545.766
----- 7000 80 0.5 1 0.6 58.65 1522.161
----+ 7000 80 0.5 1 0.9 59.64 1535.176
---+- 7000 80 0.5 4 0.6 58.65 1522.161
---++ 7000 80 0.5 4 0.9 59.64 1535.176
--+-- 7000 80 3.5 1 0.6 58.55 1504.398
Parameter Value Range --+-+ 7000 80 3.5 1 0.9 57.27 1503.237
--++- 7000 80 3.5 4 0.6 58.55 1504.398
Tbo 9000 7000 12000 --+++ 7000 80 3.5 4 0.9 57.27 1503.237
-+--- 7000 95 0.5 1 0.6 79.86 2101.909
Spares availability 90 80 95 -+--+ 7000 95 0.5 1 0.9 79.72 2106.964
-+-+- 7000 95 0.5 4 0.6 79.86 2101.909
Diagnosis time 2 0.5 3.5 -+-++ 7000 95 0.5 4 0.9 79.72 2106.964
-++-- 7000 95 3.5 1 0.6 78.31 2087.536
Technician capacity 2 1 4 -++-+ 7000 95 3.5 1 0.9 78.93 2106.35
-+++- 7000 95 3.5 4 0.6 78.31 2087.536
MA4IF 0.75 0.6 0.9 -++++ 7000 95 3.5 4 0.9 78.93 2106.35
+---- 12000 80 0.5 1 0.6 60.14 1068.573
+---+ 12000 80 0.5 1 0.9 60.13 1053.305
+--+- 12000 80 0.5 4 0.6 60.14 1068.573
+--++ 12000 80 0.5 4 0.9 60.13 1053.305
+-+-- 12000 80 3.5 1 0.6 60.14 1041.03
+-+-+ 12000 80 3.5 1 0.9 59.34 1062.462
+-++- 12000 80 3.5 4 0.6 60.14 1041.03
+-+++ 12000 80 3.5 4 0.9 59.34 1062.462
++--- 12000 95 0.5 1 0.6 80.48 1451.558
++--+ 12000 95 0.5 1 0.9 80.76 1470.463
++-+- 12000 95 0.5 4 0.6 80.48 1451.558
++-++ 12000 95 0.5 4 0.9 80.76 1470.463
+++-- 12000 95 3.5 1 0.6 79.35 1466.267
+++-+ 12000 95 3.5 1 0.9 80.11 1474.032
++++- 12000 95 3.5 4 0.6 79.35 1466.267
+++++ 12000 95 3.5 4 0.9 80.11 1474.032

Wakiru, et.al.,(2019). A simulation-based optimization approach evaluating


maintenance and spare parts demand interaction effects. International Journal of DR J.W 16-44
Production Economics, 208:329-342.
Major
repair
Main Effects Scenario
Time to
Reps Overhaul
Spares diagnosis
availability time
Technician
Capacity MA4IF
Engine Total
availability Repair time
Ave Repair
Time
----- 235 -1 -1 -1 -1 -1 58.65 1,522.16 126.85
----+ 235 -1 -1 -1 -1 1 59.64 1,535.18 127.93
---+- 235 -1 -1 -1 1 -1 58.65 1,522.16 126.85
---++ 235 -1 -1 -1 1 1 59.64 1,535.18 127.93
--+-- 235 -1 -1 1 -1 -1 58.55 1,504.40 125.37
--+-+ 235 -1 -1 1 -1 1 57.27 1,503.24 125.27
--++- 235 -1 -1 1 1 -1 58.55 1,504.40 125.37
--+++ 235 -1 -1 1 1 1 57.27 1,503.24 125.27
-+--- 235 -1 1 -1 -1 -1 79.86 2,101.91 175.16
-+--+ 235 -1 1 -1 -1 1 79.72 2,106.96 175.58
-+-+- 235 -1 1 -1 1 -1 79.86 2,101.91 175.16
-+-++ 235 -1 1 -1 1 1 79.72 2,106.96 175.58
-++-- 235 -1 1 1 -1 -1 78.31 2,087.54 173.96
-++-+ 235 -1 1 1 -1 1 78.93 2,106.35 175.53
-+++- 235 -1 1 1 1 -1 78.31 2,087.54 173.96
-++++ 235 -1 1 1 1 1 78.93 2,106.35 175.53
+---- 235 1 -1 -1 -1 -1 60.14 1,068.57 89.05
+---+ 235 1 -1 -1 -1 1 60.13 1,053.31 87.78
+--+- 235 1 -1 -1 1 -1 60.14 1,068.57 89.05
+--++ 235 1 -1 -1 1 1 60.13 1,053.31 87.78
+-+-- 235 1 -1 1 -1 -1 60.14 1,041.03 86.75
+-+-+ 235 1 -1 1 -1 1 59.34 1,062.46 88.54
+-++- 235 1 -1 1 1 -1 60.14 1,041.03 86.75
+-+++ 235 1 -1 1 1 1 59.34 1,062.46 88.54
++--- 235 1 1 -1 -1 -1 80.48 1,451.56 120.96
++--+ 235 1 1 -1 -1 1 80.76 1,470.46 122.54
++-+- 235 1 1 -1 1 -1 80.48 1,451.56 120.96
++-++ 235 1 1 -1 1 1 80.76 1,470.46 122.54
+++-- 235 1 1 1 -1 -1 79.35 1,466.27 122.19
+++-+ 235 1 1 1 -1 1 80.11 1,474.03 122.84
++++- 235 1 1 1 1 -1 79.35 1,466.27 122.19
+++++ 235 1 1 1 1 1 80.11 1,474.03 122.84

Main Effect (Engine


Availability) 1.19 20.46 -0.92 0.00 0.05
Main Effect (Total
Repair Time) -547.51 496.84 -8.10 -0.00 8.57

Wakiru, et.al.,(2019). A simulation-based optimization approach evaluating


maintenance and spare parts demand interaction effects. International Journal of DR J.W 16-45
Production Economics, 208:329-342.
Time to OL & Time to Time to Time to Spares &
Spares OL & Mjr OL & Tech Ol & Mjr Diag Spares & Spares
availaibility Dia time cap MA4IF time Tech Cap &MA4IF
Interaction Effects 1
1
1
1
1
1
1
-1
1
1
1
1
1
-1
1 1 -1 1 1 -1 1
1 1 -1 -1 1 -1 -1
1 -1 1 1 -1 1 1
1 -1 1 -1 -1 1 -1
1 -1 -1 1 -1 -1 1
1 -1 -1 -1 -1 -1 -1
-1 1 1 1 -1 -1 -1
-1 1 1 -1 -1 -1 1
-1 1 -1 1 -1 1 -1
-1 1 -1 -1 -1 1 1
-1 -1 1 1 1 -1 -1
-1 -1 1 -1 1 -1 1
-1 -1 -1 1 1 1 -1
-1 -1 -1 -1 1 1 1
-1 -1 -1 -1 1 1 1
-1 -1 -1 1 1 1 -1
-1 -1 1 -1 1 -1 1
-1 -1 1 1 1 -1 -1
-1 1 -1 -1 -1 1 1
-1 1 -1 1 -1 1 -1
-1 1 1 -1 -1 -1 1
-1 1 1 1 -1 -1 -1
1 -1 -1 -1 -1 -1 -1
1 -1 -1 1 -1 -1 1
1 -1 1 -1 -1 1 -1
1 -1 1 1 -1 1 1
1 1 -1 -1 1 -1 -1
1 1 -1 1 1 -1 1
1 1 1 -1 1 1 -1
1 1 1 1 1 1 1

Interactions (Engine -
Availability) 0.22 0.28 0.00 0.00 -0.11 0.00 0.33
Interactions (Total -
Repair Time) 87.60 8.07 0.00 -0.36 8.92 0.00 4.07
Wakiru, et.al.,(2019). A simulation-based optimization approach evaluating
maintenance and spare parts demand interaction effects. International Journal of DR J.W 16-46
Production Economics, 208:329-342.
Variables Estimate Nparm DF Sum of Squares F Ratio Prob > F

ANOVA tbo
f
n2
612.325
-548.138
-716.62
1
1
1
1
1
1
1499767.6 384.4216 0.00000
1201818.9 308.0512 0.00000
2054176.9 526.5282 0.00000
f*n2 -379.9 1 1 577303.6 147.9749 0.00000
• Main and interaction effects of tbo*n3 -78.58 1 1 49404 31.1416 0.00000
the decision variables, where n3 -209.828 1 1 176110.32 46.3816 0.00010
n2*n3 42.38 1 1 3591.71 36.6776 0.00380
D.f is the degrees of freedom. tbo*n2 -94.96 1 1 36069.6 9.2454 0.01600
• Main and Interaction effects tbo*f -92.073 1 1 33909.4 8.6917 0.01850
that are significant, means that f*n3 -112.273 1 1 25210.2 9.8302 0.03500
the variables are statistically tbo*f*n2 -76.0376 1 1 23126.8 5.9279 0.04090
tbo*Td2 54.12 1 1 11715.9 3.0856 0.11710
significant (P-Value ≤ 5%), in
n3*Td2 35.02 1 1 4907.7 1.2925 0.28850
relation to Tr. Td2 -11.54 1 1 531.99 0.1401 0.71790
• The significant interaction tbo*n3*Td2 -10.45 1 1 436.81 0.115 0.74320
effects indicate the validity of n2*MA2IF -60.735 1 1 14755 0.0741 0.79240
determining a balanced trade- MA2IF 50.76 1 1 10308.3 0.0518 0.82570
Td3 -37.35 1 1 5580.09 0.0478 0.83250
off solution while integrating Td2*Td3 -29.24 1 1 3418.74 0.0293 0.86840
the decision variables. n3*Td2*Td3 25.35 1 1 2570.49 0.022 0.88570
f*MA2IF -29.248 1 1 3421.7 0.0172 0.89890
f*n2*MA2IF 24.9926 1 1 2498.5 0.0125 0.91360
n3*Td3 15.71 1 1 987.22 0.0085 0.92900
tbo*MA2IF -6.15 1 1 151.3 0.0007 0.98020
tbo*n2*MA2IF 3.76 1 1 56.6 0.0002 0.98790

Wakiru, et.al.,(2019). A simulation-based optimization approach evaluating


maintenance and spare parts demand interaction effects. International Journal of DR J.W 16-47
Production Economics, 208:329-342.
OFAT and DOE Comparison

DR J.W 16-48

You might also like