SAU
Fen Bilin1leri EnstitüsüDergisi
7.Cilt,2.Sayt (Tenımuz 2003)
Design Optimization Of Mechanical Systems
Using Genetic Algorithnıs H.Saruhan, İ. Uygur
DESIGN OPTIMIZATION
OF
MECHANICAL SYSTEMS USING GENETIC
ALGORITHMS
Harnit
SARUHAN,
İlyas
UYGUR
Abstract-
This papeı- prcsents an algorithnı foı· the design of minimum weight of speed reducer., gear train, subject to a specificd set of constraints. The study is primari1y aimed to expose the potential of genetic algorithms, to discuss their application capabilities, and to show the concept of these algorithms as optiınization techniques and their scope of application by implementing them to thespeed reducer. Results obtained for the minimum \Veight of speed reducer are prescnted to provide insight into the capabilities of these tecbniques. Genetic algorithms are efficient search techniqucs which are inspircd fronı natural genetics selection process to cxplore a given search space.
Keywords-
Genetic al
gorith
ms, design, optimization ••Ozet-Bu
makalede sınır şartları verilen bir hız rcdüktörünün nıininıum ağırlığını hesaplayan bir algoritn1a tanıtılmaktadır. Bu çalışmanın asıl amacı genetik
algoritmaların potansiyellerini ve uy
gulanıa kabiliyetlerini, bir optinnun hız redüktörü tasarımında göstermektir. Bu tasanın için elde edilen sonuçlar bu tekniklerin uygunluğunu göstermektedir. Genetik algoritmalar tabii seleksiyon (seçim) teknikleri kullanarak tanınılanmtş sınırlar içinde tarama yapan ve genetik fikrine da)'ah uygun araştırma teknikleridir.A�ıalıtar k elim eler-
Genetik aJgoritnıalar,
tasarım, optinıizasyonI. INTRODUC1lON
Many nurner
i
calopti
miza
tion
a
l
g
oıi
thms
have been developed
and used ford
esi
g
n op
tinnzati
o
n of en
g
in
eeri
ng
pro
ble
ms
. Mostof
thcse
optiınization algoritluns so]ve engi
neer
ing
proble
msfor
fı
ndi
ng
optinıiındesign.
Solving engineeıing
pr
oblems can be con1plexand
a timeconsuming
process
when there are large nu
mber
sof
desi
g
n variables and constraints. Thus, there is a need forn
ıore
efficient and reliablea
l
g
or
ithms
that solve such problen1S. The developmentof faster
conıputers
has
a.ll
o\ve
d
developn1entof morc
1 l.Saruharı,
İ.
Uygur; Aba ntİzzet
BaysalÜniversitesi,
TeknikEgitim
l"akültcsiMakine E
ği
li ın i Bölüınü, 14550, DOzce,TURKEYrobust and effıcient
o
ptimi
zation methods. One of thesemethods
is the g
e
netic
algor
ithrns
.
The genetic
algorithnıs are
search
procedures
based onthe
id
eaof
natuı�aı
sel ec tionand
g
en
etics [ 1]. Genetic
alg
ori
tlu
nscan
be app
li
cd to conceptualand
preli
mi
n
aryengineering de
s
ig
n studies. Genetic a1gorithms havebeen increasingly
r
ecog
ni
zed and appl
iedin
manyapp1ications.
Interested reader
c
an refer studies by[2],
[3]. This
pa per shows h
o
w genetic algorithınss
ea
rc
htlu·
o
ugh a desig
n space to fırıd theminimun1
valueof
the
objective
fu
n
cti
onfor
engineering
desi
g
n problems.II.GENETIC ALGORITHMS
In this seetion of the paper, the
fundamental intuition
of
geneti
c alg
ori
thms
and how
they process areg
i
ven.
Genetic alg
ori
thms nıaintain a po
pulatio
n of encode
dsolutions,
an
d guide the populati
on
tow
a
rd
s th
e opti mumsolution [ 4]. Thus,
they sear
ch thespace
of
possible indivi
d
uals
and seek
to fınd the best fitnesss
tring
.Rather than
st
artin
g
from as
ing
le point so
lu
ti
on within the search space as in traditional optim
iza
tio
n
n1ethods, genetic al
g
orit
hıns arei
nitiali
z
e
d with a population ofsolutions.
Viewingthe genetic
al
g
orithms
as
o
ptimization tcchniques, t hey bolo
ng
to theclass of
zero-orde
r optinıizati
onmethods [ 5],
[ 6].
The description
of
as
i
mpl
e genetical
g
oritlımis
outl
i
nedin
Figure 1. An
initial
po
pulati
onis chosen
randamly
at the begiıu1in
g
. Then an iterative proce�·starts until the termination criteri a have b
e
en
satisfıed. After the evaluation of each individualfi
tness
in thep
o
pu latio
n, theg
eneticoperators, selection, crossover,
and
mutation, are applied
top
roduc
ea
new generation.Other
genetic operatorsa
re appli
e
d as needed. Thenewly
c
r
eat
ed
indıvidua]s
replace th
eexisting
generation,
and reevaluation is started forthe
fitness of
new individuals.
T
h
e
loop isrepeated
unttl anacceptab
l
esolution is found. Genetic
al
g
ori
tluns differfroın
t
ra
diti
o
na
lsearch
tec
h
niq
ue
s in thefollowing
ways [ 4]:
-Genetic algorithnıs
workwith
a co
di
ng
of
desi
g
nvariablcs and not the de
s
ig
n
variables thernselves.-Genetic algor
i
th
rnsuse
obj
ect
ive
function or fitness fuııction information. No derivatives are necessaıy asSAU Fen
Bilimleri
Enstitüsü Dergisi 7.Cllt, 2.Sayı(Temmuz
2003)-Genetic algorithms search from a population of points
not
a
single point.-Genetic algorithms ga
t
her
information fron1 current search points and direct them to the subsequent search. -Genetic algorithms can be used with discretc,integer,
continuous, or amix of
th
e
se
three d esign variables.r- --ı ı 1 ıÖ ı.Z 1 ... ıQ 'W ;ı.ı.ı ı en ı ı
(
STAR'T)
--·--- - ---- - - · - - - � ı ı ı Inpu.t: ı'
De� v.u-W.>lro co&JG ı ı I ui. ti.al :ro pcla tion ı ıObjecti,re :fiuıctio:n ı ' ı ı , __ _ --- - - �---:-- - --- -.- ... _ .... ız 'O ı ... ı ı-' o :� ı o 'O :ı:t: ıı:ı... • ı.ı.ı ·� 1 ... ·-. ' Z :o 1 ... •ı ' <( !;:ı
:�
;p.. ıııl -\ Genu.ıtl..."'ln.: 1 · - -- - - - -- - - -� . . -- --- ---.. --·- . ... ----.---. --- --- , ı 1 Perfonn sel.echorı 1 1Pa.rent I Perfonll cros�ovcr ı ı
Pai9nt ll ' •
Ir
' ı 1 1 ,-- Perfonu ıııı.ıtation ı ı U :rd.il teınpar.n:y ı\f
ıpo pı.ıla t:io h i; full • ı
..._ P�rfonn other gerı.etr. ı
• oııerator 1 1 1 -- ... ____ ---- .... ---.. --- -.. ---... - - - -- -... ----. -- --- ··---·---·; - - - ·-·--·-··-·--· • ı ı Updating txi5ling 1 1 generat ion 1 ı
�
ı ' ı ı EY'.ılı.ı.a.tıon ı ' 1 ı l. · -- ___ .. _ ____ _ _ ---\ �.--. ---. ---·- --.ıT �mtiııat io rı YES Write
1
)
cntuia - f ı.na1 rıısıı.l ts \ END
sah.siied NO
�neratiı::mc�:t\eratioıı+ 1
1
Figure 1 Flo\'V chart for a simple genetic algorithms.
Figurc 2 Seltematic Diagram of Speed Reducer.
111.1.
Design Variables�
ıı
I
X
ı !<.--'ı
ı
/
Pinion
ı
ı
/'
ı
1 /1
Design Optiınization Of l\1cchanical Systems
l1sing Genetic Algoritbms
H.Saruhan,
İ.Uygur
DI. PROBLEM
STATEMENTFigu re
2
shows the configuration of a compound geartrain which was takeı1 from Ra o [7].
lt is
desired to o btajn the lowest weight of the geart1'ain
subject toa
set of constraints. The statement of the design optimization of the problem is forn1ulated as:
Objective function
\1ininlize
F
(X)
Subject to
j
=1,
· · · ·NIC
(
numbe
ı- of inequality constraints)X
�ower
<X
. <X
�tp per
ı - l - 1 where X;={X1,X2,
. • . . . .,X,}
i= 1,
. ..
. . . nFa�;ective
=P(X)
=0.7854Xı X i (3.3333XJ
+14.9334X
3-43 .0934)
-1.508X1
(X;+
Xi)+7.477(Xı +Xi)+
0.7854(X4X
;
+X
5Xf )
(1)
The design
variables
used for fınding theminimum
N
ı ı�
�ı
ı
ı
ı
ı,., Gear 2
ı
/'
t _.r...,ı
� -----r--1---,.._ ____________ ---.. -t--- -ı •• ı J ı1
ı
ı
ı
ı
ı
....
.......1
1
'-....ı
<-'" ı ı:-- ...Shaft 2
ı
I
X
!
�
Shaft 1 """
ı ı ı 1"-
,
ı1
Bearing
�ı
ı
ı
�
ı ı)
t-" - ı 'w
ei
gh
t
of the
gea
r train include:SAL Fen Bilinıleri Enstitüsü Dergisi /.Cilt,
2.Sayı (Temmuz 2003)
.�
.. 1 is the face width. 2.6 < X1� 3.6
..
f., is
modu
le of teeth.0.7
<X
2 < 0.8X
3is
nu
nıb
er of teeth on pinion.17
<X
3 <28
X
4is
length of shaft1
between bearings. 7.3<X4 <8.3.. r
5 is length ofshaft 2
betwcen bearings.7.3<X5<8.3
X
6 is dai
meter of shaft1.
2.9<X5
�3.9
.A...
1 is daimeter of shaft 2.5
.0
<X
5< 5.5
III.2.
Constraints(2)
(3)
(4)
(5)
(6)
(7)
(8)
Constraints are conditions that must be met in the
optiınum design and inc
lu
de restrictions on the designvariabfes value and optimum design of the function. These constraints define the boundaries of the fea
s
ible
and infcasible
design space domain. The constraintsconsidered for the optin1un1 design of gear train
include
the
following:-
2 7
V -1v-2.x
}<
1
gl -
. /\ ı .. '1 2 3-g,
J
2745.X'
4 +(16.9 )106
X2.r\'3. l
2go=
74SX5
+(157.5)106
)[ıX
3g7
=x
2X1< 40
t;T =(ı
"Y
+ı
9 )x
-ı<
ı
o� ·�· ./\ 6.
4-(9)
( 1 O)
( J ı)
( 12)
0.5o.ıx�
<ı
( 13)
0.5o.ıx?
<ı
( 14)
(15)
(16)
( 1 7)
C
icnctic
algorithms are unconstrained optin1İzationprocedure. 'Therefore, th
e
constrained optimization79
Design
Optiınization Of Mectıanical SystemsUsing Genetic Algorithrns
H.Saruhan, İ.Uygur
prob
l
em has bee
n transforn1ed into an unconstrainedoptimization problem by
p
enalizing the objectivefunction
v alue with the quadratic penalty function. In
case of any violation of a constraint boundaıy, the
fitness of corresponding solutj on is pe
n
alized, and thu
s kept within feasible regions of the design space byincreas
i
ng the value of the obj
ective function when constraint violations are encoun
tered. The penaltycoefficients, ri, for the j -th constraint have to be
judiciously selected. The fitness function provides a
measure of the performance
of
an individu
al, which isuse d t
o
bais the se leetion process in fa vor of the mostfıt me mb ers of
the cuıTeni
popu
lation.
Fitness0�jcctive
=F-
(F(x)
+P)
NCONP
=L
ri
(
nıaxf O,
g
i]
)
2 .i=l(18)
(19)
where
F
is an arbitrarylarge
enou
gh that is greaterthan
F(X)
+P
to cxclude negative fitness functionvalues and
P
is the penalty function.ITI.3 C
onstruction of Designvariab I
es and GeneticAlgoritlını Parameters
In optin
ıi
zation probleın, a design of variables, x(i) ,represents a solution that n
1i
nirni
zes or maximizes an obj
ective function. The first stepfor
applying the genetic algoritluns toa
ssigned design problem isencoding of the
de
sign variables.Genetic a
l
gorithms require the design variables of theopti
m.i
zation problem to be coded. Binary coding, as afınite l ength strings, is generally used although other coding schemes hav
e
bcen used. These
sırings arerepresented as chron1osonıes. Each design variable has
a specified range
so
thatx(i)ıower :5;x(i)<x(i)upper·
'The continuous design variab
l
es can be representeuand discrctized to a precision of e . Genetic algoıithms
have abihty to deal with integer and discrete desig�
variables. The nun1ber of the digits in the binary strin
g
s,l
, is cstiınated
from the following relationship[8].
(
')
(
')
x z -x ı1
upper
lower
2 >+1
&(20)
where
x(i)1
and x(i) are the lower andower
upper
upper
bou
nd for design
vari ables respectively. Thedesign variab
I
es are code d in to the binary eligit{O,
1
} .
The
physical value of the design variables. x(i), can beSAU Fen
Bilimleri Enstitüsü Dergısi
7.Cilt, 2.Sayı (Ten1muz 2003)
x(i)
up per
-x(i) 1
OltVer .x(i)
= x(i)lo1ver
+
1
d(ı)
2
-1
(21)
where d(i)
represents t
he decin1al value ofs
trin
gfor
design var
i
ab
leswhich is
obtaine
d
using
base-2
forn1.Tablc 1 Design variables mapping.
Design
Variables Lo"�er Limit1
The face
width 2.6rv1odule of
teeth
0.7Number of
teetlı
on pinia
n 17L
eng
th of
sh
aft1 between
b
eari
ngs
7.3Length
of shaft2
between bearings
7.3Daimeter of shaft
1
2.9Dainıeter or shaft
2
5.0To start the algoritlun, an injtial
p
op
ul
ati
on set isra
n
do
ml
y assigned. This
set of initialized population
isa p
o
tenti
al
solution to the problen1. Forexaınple,
theb
i
nary
string
representationfor
the desi
g
n vari
ab
les,Table 2 The binary string repre.sentation of the vat·iablcs.
Dc.sign Optimization Of :vıechanical Systems
l Ts ing Genetic Algorithms •
H.Saruhan, I.Uygur
Design variabtes
arerepresented in different level of
p
r
ecision. Ta bl e 1 gives
descriptions ofthese
mappi
ng
.Upper Limit
Precision
String Leııgth3.6 0 .01
7
0.8O.
12
28ı
4
8.3 0.017
8.3 0.017
3.9 0.0ı
7
5.5
0.01 6xi,
in Tab
le 2 gives an example
of a chroınosomethat
re
p
resent
s design vari
ab
les
accordıngly. This designstring
is co
mposed of 40 on
es and zeros.Design Variables
-x(l)
x(2)
x(3)
x(4)
x(5)
x(6)
x(7)
-, 1 000 1010 ' 1 1 0 0 0000 0 00011o
0001100 1000000 100100 ' -- .Concatenated
Vaıiables Head-to-Tail
-0 -0 -01-01-0 -0 -0 -0 -0-0-0-0 -0 -0 -011-0-0-0-011-0-01-0-0-0-0 -0-01 -0 -0 1-0-0
In T
a
bl
e
2, thestring
of 40-bit str
i
ng length represents
one of 240
alt
eınati
ve individual soluti
ons
existingin
the
designspace.
For ıunni
nggenetic
algorithms, anin
it
ialpopulation
is need to bea
ssig
ne
d
randomly at
thebegim1ing.
Population sizeinfluences
the nun1berof search p
oi
nts ineach
generation.A
guideline
for
an
a
p
p
rop
r
iate population sizeis suggested by Goldberg
[ 1 O]. The guideJine
for optimalpopulation size
dep en dson
the
i
nd
ivi
d
u
al chromosome len
gt
h
, \Vhich isvalid up
to 60 cxpressed as follows:
80
population size=
1.65 *2°·21*1
(23)
For
a string lcngth of 40 bits, an optinıal populationsize
of 558
nıay
be used [ 1 0]. Considering computation1üne� a randomly selected
set,
1
O strings,
of potentialso] u tion is us cd in this
study sine
c th
er
e was seen to noth:ıvc a significan1
iuıprovement
in
resul
ts
. SeeTable
3.
The genetic algoritlun
then
pr
oceeds by gener
at
i
ngnew
solutioııs with bit operations u
t
ili
zing
genetic algorithmSA U Fen Bilimleri Enstitüsü
Dergisi
7 Ci lt, 2.Sayı (Temmuz 2003)Ta Ille 3 Aset of starting popula tion.
Individual
n
umb
erDesign Optimization Of Mechanical Systen1s
Using Genetic Algoa·ithıns
H.Sarulıan,
İ.
UygurRandomized
bina
ı-y stri
ng
ı
10010100001000100010100110010000001010102
ooo1oıoıoooooooooııoıooıoooıooooooıooooo3
11000100100100010110000110001010001111014
oooıııoooooooo1oooııoooıııo11ooıooı1ııoo5
. 010001 00000100001110100010010100101001016
1001011001100100011000011100010010100101 7 oooıoıooıooo1ooooıoooıooo1ııoo1ooooıoooo 8 0001001001000000011000111011101000100100 9 1010010100001000011000011001000010100101lO
0001010000000010001000011001001000000000111.4. Genetic
AlgorithnıOpcrators
In
a sin1p
le
g
ene
tic alg
ori
t
hnı, there a
re
three basicoperators f
o
rcreating the
next generation. Eachof
these
ope
rators is
expl
ai
ned and
demonstrated
in the following:th
e selection operator shownin
this work is a touman1ent sele
ctio
n.
Toumament selectio n approach
vvorks as follows: a pair
of individuals
from
mating pool is rando
nuy
picked and t
he best-fıthvo
indiv
i
duals from thispair will
be
chosen as aparent.
Each
pair of thep
aren
tcreates t\vo Child as deseribed
in
the meth
od of unif
o
nncrossover
showni
nTable 4.
A
uni
f
omıcrossover
opcr
ato
r isused
inthis
study. Auniform
crossovcr
op
e
rat
or probability
ofO. S
isrecommended in ı
na
ny works such as [ 11] and
[
l
2].
Crossover
is
very
importantin
the
success of g
en
eti
calgorithıns. This
opera
ter
is theprimary
source ofthe
newcandidate
so
lutio
nsand provides
thesearch
nıechan1sm that
effı
cie
ntly
g
ui
de
sthe
evo
lut
ion through th
esolution
space towar
ds
the
optimum.In
unifoım
cross
ove
r,
every bitof
e
achparent stTing
has acbance
ofbe ing exchanged
wit
h
the
corres
pond
ing
bit
of the
other parent s
tri
ng
.Table 4 Oniform crossover.
The
pr
oced
ure isto ob
tainany
co
ınb
inati
on oft
wo
parent string s
(
chronı.osomes)from
the matingpoo
l
at
random and generate
ne
w Child string
s
from theseparcnt strings
by perfoıming bit-by-bit
crossover chosen accordingto a randomly
g
ene
rated
crossovermask [1 3]. Where
there is
a1 in
the
crossover mas k, the Ch
i
ld bit is cop
icd froın the
first parent string, and \vhere there is aO
in the mask, the Child bit
is cop
ied
from the
se
co
nd parent string. The second Child str
ing
uses theopposite
nıle
tothe p
revi
ousone as shown
inTab le 4 .
For
ea
ch pair ofparent stringsa n
e
w
crossover nıask is ran
do
rnJyge
ncrat
ed.
Preve
ntin
g the geneticalgoıithm f
rom
thepremature
convergence to
anon
optimal
solution, whichmay lose
diversity by re
peat
ed
ap
pli
catio
n of sele
ctio
n and crossover
oper
ator
s, amutation operator is used.
M
uta
tion operator is
basicallya p
ro
ce
ss ofr
andamlya
lte
ı·ing a paı1 of anindividua] to
p
r
od
uce anew
individualby S\Vitching
the bit
p
ositio
n from aO
to
a1 or
vice versa as seen in
Table
5.
Crossover mask
.. . . '··
-
c.t
o
:
o:r
:
o
_
�(lO
o o o
o ooo.oo
ıı,o�o
o
·
·
o
�
ııto
·
oJ.o
·
o
·
q·o
·
o
·
o:ı
·
o
'·
ö
l
i
·
o·cf.:
�
;-
,-'
��
·, !_ : • \ 1. ·i ı. •. . 'f ' 1' ' ... - . • t 1 .. ·� 1 · ••• ,, :" '· c ,Parent
I
Parent
II
Chil d
I Child
II
T bl 5 M t t" a e. ı u a ıon opera t or.
Be fare
i\fter . . 1101010001000100001000001001000100110000 oıoıooooooıcıooooı ıoıooı1ooooo1oooıooıoı 010101000010100000101000100�001000100001 1101000001000100011000011000000100110100 1001010000100000011000001001000100100100 100101000010001001100000100100010010010081
SAU Fen Bi1imleri Enstitüsü Dergisi
7 .Cilt,
2.Say1 (Temnıuz 2003)
'Ine
mutatio n
rate
suggested by
Back(14)
is:
1/
<"
P
.
<
1
/
/
population size
... mutatron/
c
h
ronıoso1n
e
length
(22)
A
specialized
mechanisnı,elitis1n,
is added
to
the
g
e
neti
c algorithın.Elitism forces
the
genetic
algorithmto retain the best
individualin
a gi veng
en
cra
tion and
proceed unch
ange
dinto the following
generatian [15].The parameters
ofgenetic
algorithmfor t his
study
have
chosen as in
Tabl
e6.
Table 6 Genetic search algorithm parameters.
Genetic
algorithm
pa rameters
Chromoson1e length
40
P
op
ula
tio
n
s
i
z
e
lONumber of generatian
200
Crossover probabib
ty
0.5Mutation pr obability
0.01
There
aremany different
waysto
deterınine
when
to
stop
ıunningthe
gen
e
t
ic algorithm.One
ınethod is
to
De_,ign Optinıization Of Mechanical Systems
lJsing Genetic AJgorithm!
H.Saruhan, i.Uygur
stop after a presct number of gencration
which is used
in this
study
or atime lin-ıit. Anather is to stop
afterthe
ge
n
etic algorithın has converged.
Convergenceis
the
progression towards
uniformity. A string is saidto
haveconverged vvhen
9 5
%
of the population
share the samevalue
[16].Thus, n1ost or all strings in
thepopulation
are identical or
simi
1
a
r whenpopulation is co
nverg
ed.
TV.
RESULTSFigure
3
sho,vsthe plots of
the
no
rm
ali
zed
minimum,average, and best
fitness
function values
in eachg
ene
rat
ianas
optiınizationproceeds.
As
can
be seenfrom
Figure
3, thenormalized
fitnessfunction of
individuals
in a populationiınproves over
generations.The o
v
era
llresults sho\v that
the
best design rapidly
co nvergc
over
the first several generations
andrefıne
the
design over remaining
generations.
Thus, the
selectcd paramelers set has converged to
a stablesolutions
�rithsimilar values. The
results andtheir
comparison
witlı nurnerical nıethod used by
Rao [7]are
shown in Tab l
e
7.
As
can be seen from
theseresults,
the
genetic algorithn1Sproduced much
betterresults
than that the n
tınle
ri c
al
meth
od.
Table 7 The problem design vaı-iable� and
objective
function results.Speed Reducer Optinı.ization
Method
Design variables Nurnerical
Genetic
o
tiıniza tion
Algorithm
The face width
3.5
-2.6
:vfodule of teeth
0.7
0.7
r--
-Numbe r
ofteeth
on pinion
17 .O
ı
7.0
Length of
shaft1
between bearings 7. 3 7. 3Length of
shaft 2 betweenbearings
7.3 7. 3Diameter
ofshaft
1
3.35
3.40
Dian1eter
ofshaft
2
ı
'5.29
5.28
Design Objective
1\'linimum weight
of
ge artrain
2985.22
2654.19
.
1 •
,
c o ·--+-' u c :::.> u.. en C/) <D c -+-' ·-LL1.0
0.9
0.8
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
o
• •.
.
.
.
.
.
.
.
.
... . : . . . q
:
.. . . <t;:>. w
.
. . . �.
.. . <?. . . .: . �K��.
. �
. <?'?
.. : . . . <Ç) . : Ç)_
• • •�
• .<Doo
· · · · · · · · ·O.
.
·O
·O O
· · · · · • • • • • • • • • . . . .. . . .. . .o
· . .. . . .. . . ,.
.. .
.
. .. .
. .. .. .
. ..
.. . . • • • • • • • • • • • ·O
•((D . (()
aoo
• • •o
·(()
· • • • • • • . . · · · " · · · ·· · · ·· · · ·· · · · · · ' · · · ' · · · · · ·· · · • • • • 1 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • .. .
..
. .
.
..
.
. . .
..
...
.
. .
...
.. .
. '. .
.. .
.
. .
. . . .
.
. . . .
.
. .
.. .
.
.
.. .
• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • .. • .. • • • • .. • • • • • 1 • • • • • • • • • • • • • .. • • • • • 1 • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • · · · • • tt • • • • · · · · · · . . . • • • • • • • • • • • • • • • • • • • • • • • • • • •.
'
.
.
.
.
. 00 .
.
.
• • • • t • • • • • • • • • • • • • • • • • • • • • • 1 • • • • • • • * • • • • • • • • • • • • • • • • • • • • • • • • • • • • •.
.
.
. ".
..
.
. •[)
arn
o.
• • • . , ...
.
.
. .
.
.
.. .
..
.
• • • ..
,.
. .
...
.. .
.
. . .
..
20
40
60
• • • • • •([) G)D
• • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •.
.
.
. .
. .
'.
. . . .
. ...
.
.
• • • • • • •BO
100
1
20
Generatian
• •tD
:o
•co
• • • • • •o
Minirnum Fitness
o
Average Fitness
• • ••
Best Fitness
140
160
180
200
• •Figure 3 Co n' crgence process of genetic algoritlı ms for normalized minimum .. average, and best fitness function.
V. CONCLUSIONS
Systen1s :orithn1s
i.
UygurA genetic algori thın techııique was us ed to generate the
nıinin1uın
weiglıt of the gear train. The design variableswere sclected as those that influence the optimuın design. The results sho\v that genetic algoritlun provides goo d solutions when compared to a nurnerical
optinıization method. In this regard, the ef
fı
cacy
of genetic al
gorith
m optinrization techniques isdeınonstrated by employ
i
ng an engineering designproblen1. lt can be conc]uded that the genetic
algorithıns
can be successfully used for conceptual and pre1
in1inary des ign optimization of
the engineeringproblenıs.
3.
Saruhan, H., Rouch,K.E.,
and Roso, C.A., Design Optinuzation of Tilting-Pad Journal Bearing Using a Genetic Algorithm Approach, The9th
ofInternational Symposium on Transport
Phcnomena and Dynaınics of Ro
t
atin
gMachinery, ISROMAC-9, Honolulu, Hawaii,
2002.
4.
Goldberg, D.E.,
Genet
ic Algorithms
in Search,Optiınization, and Machine Leaming, Addison
Wesley, Readiııg,
1989.
5.
Dracopoulos, D.C., Evolutionary LeamingAlgorithms for �eural Adaptive Contro'
Sp ring er-Verlag, London, 1997.
6. Louis, S.J., Zhoo,
F.,
and Zeng,X.,
FlawDctection and Configuration with
G
eneticAlgorithmsl Evolutionary Algorithms in
Engineering Applications, Springer-
V
erlag,1997.
Vl. REFERENCES
1.
Goldberg,D.E.,
TheD
esi
gn ofInnova
tion:Lessons from Genetic Algorithms. Lessons for the
Re al World, Univer
s
ity ofl
ll
in
ois
atU
rbanaChampaign, IlliGAL Report:
98004,
Urbana, ILı 998.
2.
Sanıhan,H.,
Rouch,K.E.,
and Roso, C.A., DesignOptirnization of Fixed Pad Journal Bearing for
Rotor Systen1 U s ing a Genetic Algorithın
Approach, The
1
st International Symp
osium onStability Control
of
Rotati
ng Machinery,ISCORMA-1,
I.,ake Tahoe,N
eva da,2001.
83
7.
Ra
o, S. S., Engineeıing Optiınization 11ıeory andPractice, New Age international (P) Limited,
Pub., New Delhi,
1999.
8.
Lin, C.Y.
and Hajela,P.,
Genetic Algorithmsin
Optinıization Problcms with Discrete and
Intege
rDesign Variables, Engineering Optiınization,
19,
309-327� 1992.
9.
Wu, S.J.
and Chow, P.l'., Genetic AJgorithms forNonlinear
Mixcd
Discrete-Integer Optinıization Problems via Meta-Genetic ParameterSAU Fen Bilimleri Enstitüsü
Dergisı
7 .Cilt, 2.Sayı(Temmuz 2003)
Optimization,
EngineeıingOptimization, 24,
1 3 7-1 59, 7-1 99 5 .1 O.
Goldberg,
D
.E.,
Optinıal Initial Pop
ul a
tio
nSize
for
Bi
nar
y
CodedGenetic Algorithms, The
Clearinghouse for Genetic Algorithms,
Universityof Alabama, TCGA
Rept. 8500 1 ,Tuscaloosa,
1 985 .
1
l . Syswerda,
G.,Un
i
form Crossover
inG
ene
ticAlgorithms, Proceedings of
the 3
rdInternational
Conference
on
Genetic Algorithnıs,
MorganKaufman, 2-9, 1 989.
1 2.
Spears,
W.M.,
andDe
Jong, K.A.,On the Virtues
ofParameterized
UnıformCrossover, Proceedings
of the
4
th InternationalConference on Genetic
A
lg
oıithms, MorganKaufinan,
230-236, 1 99 1 .84
Design Optimization Of Mechanical Systems Using Genetic Algorithms H.Saruhan, i.Uygur
1 3 .
B
easl
y
, D., Bull,D.R. , and
Martin,
R.R., An
Overview of Genetic A lgorithms: Part2,
Research
T
opic
s,University
Computing, 1 5(
4
)
,
1 70- 1 8 1 ,1 993.
1 4.
Back, T., Optimal Mutation Rates
inGenetic
Se
a
r
ch
,
Proceedings
ofthe
5thInternational
Conference on Genetic Algorithms, Morgan
Kaufmann,
Los Angeles, 2-8, 1 993 .1 5 .