Int
ern
at
i
onal
Journ
al
of El
e
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
10
,
No.
4
,
A
ugus
t
2020
,
pp.
4176
~
41
88
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v10
i
4
.
pp
4176
-
41
88
4176
Journ
al h
om
e
page
:
http:
//
ij
ece.i
aesc
or
e.c
om/i
nd
ex
.ph
p/IJ
ECE
Energy e
ffici
ent chaotic
wh
ale opti
mizati
on
techni
qu
e for da
ta
gathe
ring in wi
re
less sens
or netwo
rk
Sridhar
R
.
1
,
N
.
Gur
upr
asad
2
1
Depa
rtment of I
nform
at
ion
Sci
e
nce
and Engi
ne
e
ring,
Glob
al Academ
y
of
T
ec
hno
log
y
,
Ind
ia
2
Depa
rtment of
Com
pute
r
Scie
n
ce
and Engi
ne
ering,
Globa
l
Ac
ad
em
y
of
T
ec
hnol
og
y
,
Indi
a
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
J
un
28, 2
019
Re
vised Feb
5,
2020
Accepte
d
Fe
b 25,
2020
A
W
ire
le
ss
Sen
sor
Network
inc
lude
s
the
distr
ibut
ed
sensor
nodes
using
li
m
it
ed
ene
rg
y
,
t
o
m
onit
or
the
ph
y
sic
al
envi
ronm
ent
s
and
forward
to
the
sin
k
node.
Ene
rg
y
is
the
m
aj
or
reso
urc
e
in
W
SN
for
inc
r
ea
sing
th
e
net
work
li
fetim
e.
Sever
a
l
works
have
be
e
n
done
in
thi
s
fi
el
d
bu
t
th
e
ene
r
g
y
eff
i
cient
dat
a
g
at
her
ing
is
stil
l
not im
prove
d.
In
ord
er
to
a
m
end
the
data
ga
the
ring
wit
h
m
ini
m
al
ene
rg
y
consum
pti
on,
a
n
eff
ic
i
ent
t
ec
h
nique
call
ed
ch
aot
i
c
whale
m
et
ahe
urist
ic
en
erg
y
opt
imize
d
dat
a
gat
h
e
ring
(
CW
MEODG)
is
int
rodu
ce
d
.
T
he
m
at
hematica
l
m
odel
cal
le
d
Chaot
i
c
t
ent
m
ap
is
appl
i
ed
to
the
p
ara
m
eters
used
in
th
e
C
W
MEODG
te
ch
nique
for
findi
n
g
the
global
opti
m
um
soluti
on
and
fast
conve
rge
nc
e
rate.
Sim
ula
ti
on
of
the
proposed
CW
MEODG
te
chni
que
is
per
f
orm
ed
with
diffe
ren
t
par
ameters
such
as
ene
rg
y
consum
pti
on,
da
ta
pa
cket
del
iv
er
y
ratio,
dat
a
pa
cke
t
loss
rat
io
an
d
del
a
y
wi
th
def
er
enc
e
to
ded
ic
a
ted
quant
ity
of
se
nsor
nodes
and
num
ber
of
pac
ke
ts.
Th
e
con
seque
nce
s
discu
ss
ion
show
s
tha
t
the
CW
MEOD
G
te
chni
q
u
e
progre
ss
es
the
d
at
a
gat
h
eri
ng
an
d
net
work
li
fe
tim
e
with
m
ini
m
um
del
a
y
as
well
as
pa
cke
t
lo
ss
tha
n
th
e
st
at
e
-
of
-
the
-
ar
t
m
et
ho
ds.
Ke
yw
or
d
s
:
Chaotic
tent m
ap
Chaotic
wh
al
e
m
et
aheu
risti
c
op
ti
m
iz
ation
Data gat
her
i
ng
Re
sidu
al
e
nerg
y
WSN
Copyright
©
202
0
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed
.
Corres
pond
in
g
Aut
h
or
:
Sr
id
har R.,
Dep
a
rtm
ent o
f Info
rm
at
ion
Sc
ie
nce and En
gi
neer
i
ng,
Global
A
ca
de
m
y of
Tec
hnol
og
y,
Be
ng
al
uru,
I
ndia
.
Em
a
il
:
sri
m
ln77@
gm
ai
l.co
m
1.
INTROD
U
CTION
Now
a
day’
s
adv
a
nces
in
m
iniat
ur
iz
at
i
on
ye
t
reas
onably
ef
fici
en
t
wireless
co
m
m
un
ic
at
ion
equ
i
pm
ent
and
i
m
pr
ov
e
d
sm
a
ll
-
scal
e
ener
gy
su
ppli
es
that
hav
e
c
om
bin
ed
with
re
duced
m
anu
factu
rin
g
costs
to
m
ake
a
ne
w
te
chnolo
gical
visio
n
cal
le
d
w
irel
ess
sens
or
netw
ork.
I
n
W
SN
,
a
group
of
sens
ors
are
de
plo
ye
d
arb
it
ra
rily
for
m
on
it
or
ing
an
d
c
ollec
ti
ng
th
e
physi
cal
conditi
on
s
a
nd
a
rrang
i
ng
the
c
ollec
te
d
data
at
si
nk
no
de
for
f
urt
he
r
pro
cessi
ng.
T
he
WSN
is
em
ploy
ed
f
or
sev
eral
real
-
ti
m
e
app
li
cat
ion
s
s
uch
a
s
m
i
li
ta
ry,
m
o
nitor
i
ng
env
i
ronm
ent,
agr
ic
ultur
e,
hom
e
autom
at
ion
,
sm
art
trans
portat
io
n,
healt
h
ca
re
a
nd
s
o
on.
In
WSN
s
,
var
i
ous
sens
or
node
s c
onvey thei
r
c
ollec
te
d
inf
orm
ation
t
o
a
far
-
aw
ay
b
ase sta
ti
on
via sin
k n
od
e
.
Figure
1
il
lustr
at
es
the
WSN
arch
it
ect
ure
w
her
e
t
he
nodes
are
scat
te
re
d
i
n
the
netw
ork
.
In
a
ty
pical
WSN
topolo
gy
,
on
e
ca
n
disti
nguis
h
betwee
n
ordi
nar
y
wir
el
ess
sensor
node
s
an
d
base
sta
ti
on
s
nam
ed
sink
s
.
The
sin
k
is
us
ually
con
necte
d
to
a
power
s
upply
and
it
is
capab
le
of
pe
rfor
m
ing
m
or
e
com
plex
op
er
at
ion
s.
W
i
reles
s
se
nso
r
node
s
ha
ve
a
bili
ty
to
transfer
ra
w
sen
sed
data
to
the
sin
k.
Du
e
t
o
eco
nom
ic
reaso
ns
,
nodes
are
usual
ly
po
wer
e
d
by
sm
all
siz
e
batte
ries
that
in
m
os
t
app
li
cat
ion
sce
na
rios
a
re
e
ven
i
m
po
ssible
to
r
eplace
or
rec
harge.
Ever
y
node
i
n
WSN
s
har
e
s
the
i
nfo
rm
at
ion
to
a
no
t
he
r
node
t
hro
ugh
the
ra
dio
wav
e
s.
Data
gathe
rin
g
in
W
S
N
is
the
process
of
accum
ulatin
g
the
sta
ti
st
ic
s
from
the
sens
or
no
des
an
d
directs
towa
rd
s
a si
nk
per
i
od
ic
al
ly
and
gu
i
ded to
wards t
he base stat
ion
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ener
gy
ef
fi
ci
en
t chaoti
c wh
ale
opti
miza
ti
on t
echn
i
qu
e
for
data g
ath
eri
ng
…
(
Sr
idhar
R.
)
4177
Figure
1
.
WSN
a
rc
hitec
ture
The
e
n
er
gy
supp
li
ed
to
eac
h
sens
or
node
a
r
e
the
esse
ntial
chall
enges
in
WSNs
du
e
t
o
high
pa
cket
loss
a
nd
delay
.
The
refor
e
,
e
ff
i
ci
ent
pac
ket
de
li
ver
y
rati
o
is
a
lso
a
si
gn
i
fican
t
pro
blem
in
WSN.
A
no
t
her
issue
is
the
net
work
m
ay
con
sist
of
a
la
rg
e
num
ber
of
r
at
her
dif
f
eren
t
nodes
i
n
te
rm
s
of
sens
ors,
c
om
pu
ti
ng
powe
r,
and
m
e
m
or
y.
Re
cent
re
sea
rch
pro
ve
d
th
at
the
opti
m
i
zat
ion
te
c
hn
i
ques
pro
vid
e
d
as
ef
fici
ent
energy
conser
vation f
or d
at
a
gathe
ri
ng in WS
N.
A
pa
rtic
le
swar
m
op
tim
iz
at
io
n
-
base
d
sel
ect
ion
(P
S
OBS)
Me
thod
wa
s
de
velo
ped
i
n
[1
]
fo
r
am
assin
g
the
sta
ti
sti
cs
fr
om
the
senso
r
nodes
with
the
help
of
the
m
ob
il
e
sin
k.
T
hough
the
m
et
ho
d
re
du
ce
s
the
e
nergy
util
iz
at
ion
,
the
pack
et
loss
dur
ing
the
dat
a
ga
therin
g
was
not
m
ini
m
iz
ed.
The
hybri
d
ant
c
olony
op
ti
m
iz
a
ti
on
and
par
ti
cl
e
sw
arm
op
tim
iz
at
i
on
(A
C
OP
S
O)
Ce
ntered
E
nergy
Eff
ect
ive
C
lusterin
g
te
ch
ni
qu
e
wa
s
intr
oduce
d
in
[2
]
to
stre
ngthe
n
the
inte
r
-
cl
ust
er
data
a
m
ass
m
ent
and
en
ha
nce
the
netw
ork
li
feti
m
e.
The
AC
O
PSO
te
chn
iq
ue fai
le
d
to
im
pr
ov
e
th
e p
ac
ket del
ive
ry r
at
io
with m
i
ni
m
u
m
d
el
ay
.
A
m
axi
m
u
m
lif
et
i
m
e
data
ag
gr
e
gatio
n
tree
sche
du
li
ng
(M
LDATS)
al
gor
it
h
m
was
intr
oduce
d
[
3]
to
save
the
se
nso
r
node
s
ene
rg
y
for
at
ta
ining
t
he
im
pr
ov
e
d
ne
twork
li
fes
pa
n.
But
the
al
gorithm
fail
ed
to
ob
ta
in
eff
ic
ie
nt
data
aggre
gation
i
n
dynam
ic
W
S
N
.
A
Cl
us
te
r
-
Ri
ng
Me
th
o
d
was
intr
oduc
ed
[
4]
to
i
m
pro
ve
the
energy
prof
ic
ie
nt
gat
heri
ng
of
data
f
or
exte
ns
ive
WSN.
Wh
il
e
changin
g
the
topolo
gy
of
ne
twor
k
dynam
ic
al
l
y, the a
ppr
oach pe
rfor
m
ed
the
da
ta
ag
gre
gation
bu
t
fail
ed
t
o prov
i
de
acc
ur
at
e
resu
lt
s.
An
A
u
ct
ion
-
B
ased
Sc
hem
e
was
de
vel
op
e
d
[5
]
f
or
gathe
r
ing
the
data
f
r
om
W
S
N
with
m
ini
m
u
m
energy
co
nsu
m
pt
ion
.
T
he
s
chem
e
fail
ed
to
sel
ect
the
op
ti
m
al
senso
r
nodes
for
in
creasin
g
the
ne
twork
li
fetim
e.
A
Bayesi
an
com
pr
essive
sensi
ng
ga
therin
g
of
da
ta
m
et
ho
d
wa
s
intr
od
uced
[6
]
for
ens
uri
ng
th
e
data
gathe
rin
g
an
d
prolo
ng
the
ne
twork
li
fetim
e.
The
appr
oach
fail
ed
to
at
ta
i
n
higher
data
pack
et
delive
r
y
at
the
sink
no
de.
A
Distribu
te
d
Op
ti
m
a
l
Mov
e
m
ent
te
chn
iq
ue
was
de
velo
ped
[
7]
for
col
le
ct
ing
the
data
with
m
ini
m
al
loss
r
at
e.
But
the
energy
eff
ic
ie
nt
data
gathe
r
ing
wa
s
not
perform
ed
to
inc
rease
the
li
fetim
e
of n
et
work.
The
rendez
vous
-
bas
ed
data
c
ongregati
on
protoc
ols
w
ere
desig
ne
d
in
[
8]
fo
r
delay
in
hib
it
ed
dat
a
gathe
rin
g
in
WSN
.
B
ut
the
li
fetim
e
of
network
was
not
enri
ched
since
it
f
ai
le
d
to
sel
ect
the
energy
eff
i
ci
ent
nodes
.
An
ef
fici
ent
struc
ture
-
fr
ee
data
a
ggre
gation
an
d
deli
ver
y
(ESD
AD)
te
chn
i
que
was
intr
oduce
d
[9
]
f
or
i
m
pr
ovin
g
the
data
deli
ver
y
with
le
sse
r
e
ne
rg
y
c
onsu
m
ption.
T
he
reli
abi
li
ty
in
the
data
ag
gr
e
gatio
n
proces
s
was
no
t
i
ncr
ea
sed.
A
sp
a
rsity
feedbac
k
-
ba
sed
c
om
pr
essive
data
gat
heri
ng
proce
dure
was
hoste
d
[10]
f
or
balancin
g
the
vital
it
y
a
m
id
t
he
sen
sor
no
de
s.
The
proce
dure
did
not
m
inim
iz
e
the
total
energy
co
nsu
m
pt
ion
for fu
rthe
r
im
pr
ovin
g data
gather
i
ng p
e
rfor
m
ance.
The
m
ajo
r
iss
ue
s
identifie
d
from
the
existi
ng
li
te
ratur
e
a
re
higher
pac
ket
loss
rate,
l
onge
r
delay
,
la
ck
of
im
pr
ov
i
ng
t
he
netw
ork
li
fe
tim
e,
fail
ur
e
to
i
m
pr
ov
e
the
da
ta
pack
et
delivery
an
d
so
on.
In
orde
r
to
address
su
c
h
kind
of
issues,
a
n
ef
fici
ent
te
chn
i
qu
e
cal
le
d
chao
t
ic
wh
al
e
m
etah
eu
risti
c
ene
rg
y
opti
m
iz
ed
data
gathe
rin
g
(C
WMEO
D
G)
i
s
intr
oduce
d.
The
m
ai
n
ai
ds
of
the
sugg
e
ste
d
C
W
MEOD
G
te
c
h
nique
are
su
m
m
arized as foll
ow
s;
To
i
m
pr
ove
th
e
data
gather
in
g
with
m
ini
m
um
ener
gy
cons
um
ption
,
C
WMEODG
te
ch
ni
qu
e
is
dev
el
oped.
Fo
r
e
ach
se
nsor
node
i
n
th
e
popula
ti
on
,
the
rem
ai
nin
g
energy
is
cal
culat
ed.
In
fitness
cal
culat
io
n,
the
resi
du
al
str
eng
t
h
of
the
node
is
relat
ed
with
the
t
hr
es
ho
l
d
val
ue.
If
the
ene
rg
y
of
the
node
is
m
or
e
pr
e
pondera
nt t
han the t
hr
es
ho
ld v
al
ue,
t
he
n
t
he node is
d
esi
gn
at
e
d
as
a c
urren
t
best.
The
ene
r
gy
of
the
current
be
st
is
co
m
par
ed
with
the
othe
r
sens
or
node
s
to
fin
d
the
glo
bal
op
ti
m
um
thr
ough
the
fi
tness
m
easur
e.
The
global
optim
u
m
is
det
erm
ined
by
app
ly
in
g
the
c
hao
ti
c
te
nt
m
ap
m
at
he
m
at
ic
a
l m
o
del to th
e al
gorithm
.
To
im
pr
ove
t
he
delive
ry
of
da
ta
pac
ket,
m
ini
m
iz
e
the
loss
r
at
e
an
d
delay
,
t
he
s
ource
no
de
pro
pels
t
he
da
ta
pack
et
t
ow
a
r
ds t
he hig
her ene
rg
y se
nsor
no
de
to
the
sin
k n
ode.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
4
,
A
ugus
t
2020
:
4176
-
4188
4178
The
rest
of
t
he
arti
cl
e
is
arr
an
ge
d
into
five
dif
fe
ren
t
s
ect
ion
s.
S
ect
io
n
2
outl
ines
the
al
li
ed
w
orks.
In
sect
io
n
3,
the
pro
posed
C
W
MEO
DG
proce
dure
is
e
xp
la
ine
d
with
a
neat
dia
gr
a
m
.
In
sect
ion
4,
si
m
ulati
on
set
ti
ng
is
ex
hib
it
e
d
with
var
i
ou
s
par
am
et
ers.
Si
m
ula
ti
on
outc
om
es
are
deli
be
rated
i
n
sect
io
n
5.
The de
du
ct
io
n
of p
a
per is
pr
es
ented
i
n
sect
io
n
6.
2.
RE
LATE
D
W
ORK
A
n
a
r
t
i
f
i
c
i
a
l
b
e
e
c
o
l
o
n
y
(
A
B
C
)
o
p
t
i
m
i
z
e
d
p
a
r
t
i
c
l
e
s
w
a
r
m
o
p
t
i
m
i
z
a
t
i
o
n
p
r
o
c
e
d
u
r
e
(
A
B
C
-
P
S
O
)
w
a
s
i
n
t
r
o
d
u
c
e
d
i
n
[
1
1
]
f
o
r
i
m
p
r
o
v
i
n
g
t
h
e
l
i
f
e
s
p
a
n
o
f
n
e
t
w
o
r
k
b
y
c
o
n
s
u
m
i
n
g
t
h
e
m
i
n
i
m
u
m
e
n
e
r
g
y
.
T
h
o
u
g
h
t
h
e
a
l
g
o
r
i
t
h
m
i
m
p
r
o
v
e
s
t
h
e
r
o
b
u
s
t
n
e
s
s
a
n
d
t
h
e
r
e
l
i
a
b
i
l
i
t
y
,
e
n
e
r
g
y
u
t
i
l
i
z
a
t
i
o
n
r
a
t
e
s
w
e
r
e
n
o
t
m
i
n
i
m
i
z
e
d
a
t
t
h
e
r
e
q
u
i
r
e
d
l
e
v
e
l
.
A
n
t
C
o
l
o
n
y
O
p
t
i
m
i
z
a
t
i
o
n
A
l
g
o
r
i
t
h
m
w
a
s
d
e
s
i
g
n
e
d
i
n
[
1
2
]
f
o
r
d
i
p
p
i
n
g
t
h
e
n
o
d
e
e
n
e
r
g
y
a
n
d
e
n
h
a
n
c
i
n
g
t
h
e
n
e
t
w
o
r
k
l
i
f
e
s
p
a
n
.
T
h
e
o
p
t
i
m
i
z
a
t
i
o
n
a
l
g
o
r
i
t
h
m
d
i
d
n
o
t
e
f
f
e
c
t
i
v
e
l
y
p
r
o
v
i
d
e
t
h
e
g
l
o
b
a
l
o
p
t
i
m
u
m
s
o
l
u
t
i
o
n
.
In
[
13
]
,
Vi
vac
it
y
Su
pe
rv
isi
on
a
nd
Cr
os
s
-
la
ye
r
O
ptim
iz
a
ti
on
Me
th
od
was
i
ntrod
uce
d
inte
nded
at
reducin
g
the
ti
m
e
-
aver
a
ge
rat
e
of
e
nergy
de
pleti
on
.
The
al
gorithm
fail
ed
to
perf
or
m
the
delay
reducti
on
i
n
the
data
gat
he
rin
g
pract
ic
e.
The
pa
rtic
le
swar
m
op
ti
m
izati
on
-
base
d
cl
us
te
rin
g
al
go
rithm
was
dev
e
lope
d
in
[
14
]
for
pe
rfor
m
ing
da
ta
colle
ct
ion
wi
th
le
sser
c
ons
um
ption
of
e
ne
rg
y
a
nd
tra
nsm
issi
on
delay
us
in
g
the m
ob
il
e sin
k.
The
alg
or
it
h
m
f
ai
le
d
to m
ini
m
iz
e the
packet
loss
in
d
at
a
colle
ct
ion
.
An
ada
ptive
a
ncho
r
sel
ect
io
n
al
go
rithm
was
intr
oduce
d
in
[
15]
us
i
ng
m
ob
il
e
colle
cto
r
,
to
am
ass
the
da
ta
f
ro
m
the
sel
ect
ed
se
ns
ors
an
d
m
ain
ta
in
t
he
c
ons
um
ption
of
en
erg
y.
T
he
al
go
rithm
fail
ed
to
sel
e
ct
op
ti
m
al
ener
gy
eff
ic
ie
nt
sen
s
or
nodes
.
A
P
areto
-
O
pti
m
al
Cl
us
te
rin
g
Me
thod
was
intr
oduced
i
n
[
16
]
f
or
data
aggre
gation
a
nd
i
m
pr
ovin
g
th
e
networ
k
li
fetim
e
by
achieving
bette
r
energ
y
saving
s
.
But
the
perform
ance
of
delay
whil
e p
e
rfor
m
ing
the
dat
a g
at
he
rin
g was
no
t
m
ini
mize
d.
In
[17],
delay
-
eff
ic
ie
nt
traf
fic
adap
ti
ve
(DE
TA)
m
et
ho
d
w
as
introd
uced
aim
ed
at
collecti
ng
the
data
from
senso
r
node
s
us
in
g
le
s
ser
co
nsum
pti
on
of
e
nergy.
Th
ough
the
m
et
hod
co
ns
ide
r
ably
reduces
the
data
c
o
l
l
e
c
t
i
o
n
d
e
l
a
y
,
t
h
e
d
a
t
a
p
a
c
k
e
t
d
e
l
i
v
e
r
y
p
r
o
p
o
r
t
i
o
n
w
a
s
n
o
t
e
n
r
i
c
h
e
d
.
A
u
n
i
c
a
s
t
t
r
e
e
-
b
a
s
e
d
d
a
t
a
g
a
t
h
e
r
i
n
g
p
r
o
t
o
c
o
l
(
U
T
D
G
)
w
a
s
d
e
s
i
g
n
e
d
i
n
[
1
8
]
t
o
i
m
pr
o
v
e
t
h
e
m
e
s
s
a
g
e
d
i
s
t
r
i
b
u
t
i
o
n
r
a
t
e
s
w
i
t
h
m
i
n
i
m
um
c
om
m
u
n
i
c
a
t
i
o
n
o
v
e
r
h
e
a
d
a
s
w
e
l
l
a
s
d
e
l
a
y
.
B
u
t
t
h
e
d
a
t
a
l
o
s
s
r
a
t
e
w
a
s
n
o
t
m
i
n
i
m
i
z
e
d
d
u
r
i
n
g
t
h
e
d
a
t
a
c
o
l
l
e
c
t
i
o
n
.
An
ant
c
olony op
ti
m
iz
ation
(
ACO
)
procedu
re
wa
s
de
sig
ne
d
in
[
19
]
u
sin
g
a
m
ob
il
e
sink
for
g
at
he
rin
g
the
data
and
in
creasin
g
the
ne
twork
li
fes
pan.
The
al
gorithm
fail
ed
to
trans
m
it
the
data
pack
et
with
m
inim
u
m
delay
.
An
on
-
dem
and
m
ob
il
e
si
nk
trave
rsal
(ODMST
)
proce
dure
was
desi
gn
e
d
[2
0]
to
ag
gregate
the
inf
or
m
at
ion
from
CHs
throu
gh
the
m
obil
e
sink
.
The
c
onve
rg
e
nce
of
proce
dures
in
la
rg
e
scal
e
of
WSN
was
no
t
at
ta
in
ed
in
an
e
ff
e
c
ti
ve
m
ann
er.
A
novel
ba
ct
eria
f
or
a
ging
optim
iz
at
ion
(BFO
)
al
gor
it
hm
was
introd
uced
in
[21]
f
or
ene
r
gy
eff
ic
ie
nt
c
om
m
un
ic
at
ion
i
n
WSN.
The
al
gorithm
m
a
i
ntains
good
st
abili
ty
betwee
n
the
com
pu
ta
ti
on
al
and
com
m
un
ic
at
ion
de
m
a
nd
s
of
a
se
nsor
node.
Howev
e
r,
the
c
os
t
of
com
pu
ta
ti
on
al
com
plexity
was
no
t
s
uffici
ent
.
A
no
vel
ap
proach
of
a
nt
col
on
y
op
ti
m
iz
a
tio
n
(A
C
O)
al
gorithm
was
im
ple
m
ented
in
[
22]
f
or
inf
or
m
at
ion
tra
ns
m
issi
on
in
t
he
WSN.
I
n
t
hi
s
m
e
tho
d
op
ti
m
al
ro
utin
g
is
base
d
on
the
bo
t
h
th
e
node
m
ob
il
ity
and
the
ene
rg
y
of
the
no
de.
T
h
e
i
n
t
r
o
d
u
c
e
d
m
e
t
h
o
d
o
b
t
a
i
n
i
m
p
r
o
v
e
m
e
n
t
i
n
t
h
e
e
n
e
r
g
y
c
o
n
s
u
m
p
t
i
o
n
o
f
t
h
e
n
o
d
e
s
p
e
r
t
r
a
n
s
m
i
s
s
i
o
n
.
H
o
w
e
v
e
r
,
t
h
e
p
e
r
f
o
r
m
a
n
c
e
o
f
d
e
l
a
y
d
u
r
i
n
g
i
n
f
o
r
m
a
t
i
o
n
t
r
a
n
s
m
i
s
s
i
o
n
w
a
s
n
o
t
c
o
n
s
i
d
e
r
e
d
.
A
hi
gh
-
le
vel
a
dap
ti
ve
po
wer
m
anag
em
ent
ci
rcu
it
(P
MC
)
was
pr
ese
nted
in
[
23
]
to
ac
hi
eve
ene
rgy
eff
ic
ie
nt
data
transm
issi
on
.
Howe
ver,
the
m
ini
m
iz
at
ion
of
en
d
-
to
-
en
d
delay
was
not
su
ff
ic
ie
nt.
A
novel
routin
g
protoc
ol
f
or
WSN’s
was
desig
ne
d
in
[
24]
.
I
n
t
he
intr
oduce
d
protoc
ol,
tra
ns
m
i
tt
ing
da
ta
proc
ess
i
s
done
in
a
hie
rar
c
hal
way
with
im
pr
ov
e
d
Q
oS
i
n
the
netw
ork.
H
ow
eve
r,
t
he
pe
rfor
m
ance
of
energy
m
ini
m
iz
at
ion
was
not
en
ough.
A
n
e
nergy
-
e
ff
ic
ie
nt
ha
ndov
er
m
echan
is
m
was
intr
oduce
d
in
[25]
for
ach
ie
ving
higher
pac
ket
delivery
rati
o
durin
g
data
tra
ns
m
issi
on
in
WSN.
B
ut
the
pack
et
l
os
s
rat
e
was
not
c
ons
idere
d.
The
issue
s
i
de
ntifie
d
f
ro
m
t
he
ab
ove
-
sai
d
rev
ie
w
s
are
ov
e
rc
om
e
by
pr
ese
ntin
g
a
ne
w
proce
dure
cal
le
d
C
W
MEO
DG
t
echn
i
qu
e
.
The
proce
dure
of
the
C
W
ME
ODG
te
chn
i
qu
e
is
exp
la
ine
d
wit
h
the
neat
dia
gram
in
the n
e
xt secti
on.
3.
RESEA
R
CH MET
HO
DOL
OGY
In
WSN,
se
ve
ral
se
ns
or
node
s
ha
ve
t
he
ca
pab
il
it
y
to
se
nse
an
d
f
orwards
t
he
i
nfor
m
at
ion
t
o
si
nk
node.
The
sin
k
node
tu
rn
s
a
s
a
data
colle
c
tor
to
assem
bl
e
the
facts
fr
om
all
senso
r
node
s
for
fu
t
ure
us
e
.
Thro
ughout
th
e
data
gat
her
in
g
process
,
t
he
energy
reso
urc
e
is
esse
ntial
to
upsurg
e
the
li
fesp
an
of
net
work.
The
sin
k
node
and
sen
sor
no
des
bala
nce
th
e
energy
in
the
data
gather
i
ng
process
.
Ba
sed
on
this
obje
ct
ive,
chao
ti
c
wh
al
e
m
et
aheu
risti
c
energy
op
ti
m
i
zed
data
gat
h
e
rin
g
(C
WME
ODG)
is
intr
oduce
d.
T
he
f
ol
lowi
ng
syst
e
m
m
od
el
is u
se
d for
orga
nizing t
he
pro
pose
d
C
W
ME
O
DG tec
hniq
ue.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ener
gy
ef
fi
ci
en
t chaoti
c wh
ale
opti
miza
ti
on t
echn
i
qu
e
for
data g
ath
eri
ng
…
(
Sr
idhar
R.
)
4179
3.1.
Org
an
iz
at
io
n
mod
el
Con
si
der
the
wireless
sen
sor
netw
ork
(
W
SN
)
a
rr
a
nged
in
the
gr
a
ph
‘
=
(
,
)
’
w
her
e
‘
’
represe
nts
a
ve
rtic
es
i.e.
num
ber
of
se
nsor
node
s
de
no
te
d
a
s
‘
=
1
,
2
,
3
…
,
’
and
a
set
‘
’
de
no
te
s
an
ed
ges
i.e.
c
onnecti
on
bet
ween
t
he
node
s
in
a
sensing
zon
e
.
The
node
s
are
orga
nize
d
in
the
s
quare
area
for
m
on
it
or
ing
a
nd
colle
ct
in
g
the
data
pac
ke
ts
1
,
2
,
3
…
and
t
ran
s
m
itti
ng
to
the
sink
node
(
)
thr
ough
the
n
ei
ghbori
ng
no
de
s
1
,
2
,
3
,
…
.
.
Ba
se
d
on
t
he
ab
ove
-
sai
d
sys
tem
m
od
el
,
the
pro
pose
d
C
W
MEO
DG t
echn
i
qu
e
is
des
ign
e
d.
3.2.
Chaotic
whale
met
ah
euri
s
tic energ
y op
tim
iz
ed da
t
a gath
eri
ng
in
WS
N
The
C
W
ME
O
DG
te
ch
nique
is
e
m
plo
ye
d
to
pe
r
form
the
data
colle
ct
ion
in
W
S
N
wit
h
m
ini
m
u
m
energy
co
nsu
m
pt
ion
.
T
he
conve
ntion
al
wh
al
e
op
ti
m
izati
on
al
gorith
m
is
sti
l
l
no
t
eff
ic
ie
nt
to
perform
the
bette
r
so
lu
ti
on
.
I
n
orde
r
to
at
ta
in
the
go
od
co
nver
gence
rate
and
ob
ta
in
the
glo
bal
ly
op
tim
al
so
luti
on,
c
hao
ti
c
wh
al
e
m
et
aheu
risti
c
op
ti
m
iz
ation
is
em
plo
ye
d
by
t
un
i
ng
the
ce
rta
in
pa
ram
et
ers
us
e
d
in
t
he
al
gorithm
.
The
cha
otic
w
hale
op
ti
m
iz
ati
on
al
go
rithm
i
s
the
m
et
aheu
r
ist
ic
wh
ic
h
pro
vid
es
bette
r
sol
ution
s
with
m
ini
m
al
com
pu
ta
ti
on
al
effor
t.
T
he
C
WMEO
D
G
m
et
hod
ex
pe
nd
s
the
energy
as
m
ajo
r
r
eso
urc
e
to
upsurge
ne
twork
li
fesp
an
in data
g
at
he
rin
g proc
ess. T
he
a
rch
it
ect
ur
e
diag
ram
of C
W
ME
O
D
G
te
ch
nique is
sh
ow
n
in
F
i
gur
e 2
.
In
it
ia
ll
y,
the
de
dicat
ed
qu
a
ntit
y
of
se
nsor
no
des
is
posit
ione
d
in
net
work
for
c
ollec
ti
ng
data
pac
kets
from
the en
vir
on
m
ental
co
nd
it
ion
s. Afte
r
ac
cum
ulati
ng
the
d
at
a
pack
et
s,
the se
nsor n
od
e
s en
e
rg
y i
s c
om
pu
te
d
to
fi
nd
opti
m
a
l
node
us
i
ng
c
hao
ti
c
wh
al
e
m
et
aheu
risti
c
op
ti
m
iz
ation
t
echn
i
qu
e
.
Af
te
r
fi
nd
i
ng
the
energy
eff
ic
ie
nt
nodes
,
the
c
ollec
te
d
data
pa
ckets
a
r
e
sent
t
ow
a
r
ds
sin
k
to
en
rich
the
netw
ork
li
fesp
a
n.
T
he
de
ta
il
ed
process
of C
W
MEOD
G
te
c
hniqu
e is
d
esc
rib
ed
in
the
f
ollo
wing secti
on.
Figure
2
.
A
rch
i
te
ct
ur
e
diag
ra
m
o
f
pro
posed
C
W
MEO
DG t
echn
i
qu
e
The
po
pu
la
ti
ons
of
t
he
sens
or
nodes
1
,
2
,
3
…
,
are
init
ia
li
zed
in
t
he
WSN
f
or
c
ollec
ti
ng
the
data
packet
s
1
,
2
,
3
…
from
the
e
nviro
nm
ent.
At
first,
al
l
t
he
node
s
in
ne
tw
ork
hav
e
e
qu
i
vale
nt
e
n
e
r
g
y
l
e
v
e
l
.
S
e
n
s
o
r
n
o
d
e
e
n
e
r
g
y
i
s
p
r
o
c
e
s
s
e
d
b
y
t
h
e
p
r
o
d
u
c
t
o
f
p
o
w
e
r
a
n
d
t
i
m
e
.
T
h
e
e
n
e
r
g
y
i
s
m
a
t
h
e
m
a
t
i
c
a
l
ly
com
pu
te
d
as
fo
ll
ow
s,
=
∗
(1)
Fr
om
(1),
repr
esents
the
e
nergy
of
the
node
s,
de
note
s
a
powe
r
in
watt
s,
denotes
a
ti
m
e
in
seco
nds
(
Se
c).
The
e
ne
rg
y
of
ever
y
si
ng
le
no
de
is
quantifi
e
d
as
unit
of
joul
e
(
J)
.
Af
te
r
th
e
data
c
ollec
ti
on
,
t
he
e
nergy
of
eac
h
sens
or
node
ge
ts de
gr
a
ded an
d
the
r
em
ai
nin
g (i.e.
resid
ual)
en
e
rg
y
of the
nodes
are
c
ompu
te
d
as
foll
ows,
=
−
(2)
In
(
2)
,
repres
ents
the
resid
ua
l
energy,
de
no
te
s
a
total
e
nergy,
si
gn
i
fies
the
c
onsu
m
ed
e
nergy
of
sens
or
nod
e
s.
The
nodes
resi
du
al
e
nergy
is use
d
f
or
ide
ntifyi
ng
se
nsor
no
des
wh
ic
h
util
i
ze
the
m
or
e
en
erg
y
i
n
the
data
colle
ct
ion
.
T
her
e
f
or
e
,
an
eff
ic
ie
nt
optim
iz
at
ion
te
c
hn
i
qu
e
is
us
e
d
to
sel
ect
ener
gy
eff
ic
ie
nt
node
s
for
da
ta
gathe
rin
g
in
W
S
N.
T
he
C
W
MEO
DG
t
echn
i
qu
e
c
ompu
te
s
the
fit
ne
ss
of
eac
h
sen
so
r
node
f
or
s
el
ect
ing
the opti
m
al
to
perform
d
at
a ga
therin
g
i
n WS
N.
The
f
it
ness
is com
pu
te
d
as
foll
ow
s
,
=
≥
(3)
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
4
,
A
ugus
t
2020
:
4176
-
4188
4180
Fr
om
(
3)
,
denotes
a
resi
du
a
l
ene
rg
y,
re
presents
the
th
r
esh
old
f
or
the
resi
du
al
ene
r
gy.
Ba
se
d
on
the
fitness
m
e
asur
e
,
the
curr
ent
best
op
ti
m
al
ener
gy
eff
ic
ie
nt
sensor
no
de
is
sel
ect
ed.
In
orde
r
to
identify
the
gl
ob
al
best
optim
al
sensor
no
des,
th
ree
ph
a
ses
a
re
car
r
ie
d
out
s
uc
h
as
encircli
n
g
pre
y,
ex
plo
it
at
ion
ph
a
se,
and ex
plorat
io
n ph
a
se.
Fo
ll
owed
b
y,
g
l
ob
al
optim
al
sen
sor
nod
e
s ar
e
sele
ct
ed
f
or d
at
a
gat
her
in
g
i
n WS
N.
3.2.1. E
ncir
cl
ing
prey
Encircli
ng
pr
e
y
is
the
fi
rst
le
vel
of
t
he
cha
otic
w
hal
e
m
e
ta
heu
risti
c
optim
iz
at
ion
te
chn
i
que.
In
t
h
is
phase
,
the
w
hale
(i.
e.
sen
sor
node
)
disc
overs
t
he
locat
ion
of
the
prey
an
d
encircle
s
them
since
the
locat
ion
is
no
t
known
previo
us
ly
in
the
search
sp
ace
.
Du
e
to
this,
th
e
pr
op
os
e
d
op
t
i
m
iz
ation
al
gorithm
consi
ders
that
the
current
bes
t
senso
r
node
is
cl
os
e
to
an
op
ti
m
al
.
The
c
urren
t
best
opt
i
m
al
senso
r
node
is
sel
ect
ed
base
d
on
fitness
com
pu
ta
ti
on.
Af
te
r
find
i
ng
the
c
urre
nt
best
sens
or
node,
the
posit
ion
of
t
he
w
hale
is
updated
to
wards
t
he
sea
rc
h
a
gen
t
f
or
com
par
in
g
with
oth
e
r
so
l
ution
(i.e
.
wh
al
e
)
to
fi
nd
the
e
nergy
op
ti
m
iz
ed
sens
or
node
. T
he posi
ti
on
up
date pr
ocess
is
expresse
d
as
fo
ll
ow
s,
(
+
1
)
=
(
)
−
.
(4)
=
|
.
(
)
−
(
)
|
(5)
In
(
4),
(
+
1
)
denote
s
an
update
d
l
ocati
on
of
the
sen
sor
node
s,
(
)
re
pr
e
sents
the
posit
io
n
vect
or
of
the
prey
.
is
a
coeffic
ie
nt
vect
or,
represents the
distanc
e
be
tween
t
he
posit
ion
v
ect
or o
f
t
he
pr
ey
(
)
an
d
the
posit
ion
ve
ct
or
of
t
he
w
hale
(
)
.
In
(
5)
,
represe
nts
the
coeffic
ie
nt
ve
ct
or
.
As
m
entione
d
a
bove,
an
d
pa
ram
eter
s
us
e
d
f
or
s
hr
i
nk
i
ng
the
e
ncircli
ng
m
ec
han
ism
.
O
n
c
on
t
rar
y
to
a
c
onve
ntion
al
w
hale
op
ti
m
iz
ation
a
lgorit
hm
,
the
pro
po
se
d
w
ha
le
op
ti
m
iz
ation
def
i
nes
th
e
two
pa
ram
eter
s
an
d
wi
th
the ch
a
otic m
a
p value.
T
hese
par
am
et
ers
are
def
i
ned as
fo
ll
ow
s
,
=
(
2
∗
−
1
)
(6)
=
2
∗
(7)
Fr
om
(
6)
,
de
no
te
s
com
pone
nt
wh
ic
h
is
dec
rem
ented
f
ro
m
2
t
o
0
t
hroug
hout
the
seq
ue
nc
e
of
rep
et
it
ion
s
.
denotes
a
n
assess
m
ent
obta
ine
d
from
the
c
hao
t
ic
m
ap.
T
he
pr
opos
e
d
opti
m
i
zat
ion
te
ch
niqu
e
us
e
s
t
he
te
nt
c
hao
ti
c
m
ap
f
or
identify
in
g
t
he
global
optim
um
so
luti
on.
T
he
te
nt
is
a
s
ha
pe
of
the
grap
h
w
hic
h
is sh
own
i
n
Fi
gure
3
.
Fi
gure
3 i
ll
us
trat
es a cha
otic t
ent m
ap.
Th
e
m
ap
funct
ion
is
expre
sse
d
as
foll
ows,
+
1
=
{
∗
<
0
.
5
∗
(
1
−
)
≥
0
.
5
(8)
Fr
om
(8
)
,
+
1
is
the
real
value
d
f
un
ct
io
n
of
the
chao
ti
c
te
nt
m
ap
in
the
range
from
0
to
1
in
the
un
it
inter
va
l.
Hen
ce
the
c
ha
otic
te
nt
m
ap
is
cal
le
d
as
a
di
screte
-
ti
m
e
dy
nam
ic
al
sys
tem
.
is
pa
ram
eter
ra
nge
s
from
0
t
o
1
and ch
oosin
g
t
he value
of
para
m
et
er
=
2
.
Figure
3
.
Cha
ot
ic
tent m
ap
v
n
V
n
+
1
0
1
0
.
5
1
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ener
gy
ef
fi
ci
en
t chaoti
c wh
ale
opti
miza
ti
on t
echn
i
qu
e
for
data g
ath
eri
ng
…
(
Sr
idhar
R.
)
4181
3.2.2. E
xp
l
oitation ph
as
e
The
ex
plo
it
at
io
n
phase is t
he
second level o
f t
he
op
ti
m
iz
a
ti
on
techn
i
qu
e
. Th
is segm
ent is
al
so
k
no
wn
as
a
Bu
bb
le
-
ne
t
at
ta
cking
te
ch
nique. I
n
t
he
e
xp
l
oitat
ion
p
ha
se,
the d
ist
ance
betwee
n
the whale
an
d
locat
i
on
of
the
prey
is
co
m
pu
te
d.
I
n
thi
s
phase,
a
hel
ix
eq
uatio
n
is
form
ulate
d
w
it
h
the
locat
io
n
of
t
he
prey
an
d
the
wh
al
e
’s
lo
cat
ion
to
fo
ll
ow
the
helix
-
s
hap
e
d
m
otion
of
hum
pb
ack
wh
al
es.
T
he
updatin
g
res
ults
are
expresse
d usin
g
the
ch
a
otic t
e
nt m
ap
as foll
ows,
(
+
1
)
=
′
∗
(
2
)
+
′
(
)
(9)
′
=
|
(
)
′
−
(
)
|
(10)
In
(
10),
(
+
1
)
re
presents
t
he
update
po
sit
io
n
of
the
sen
sor
nodes
,
′
denotes
an
up
dated
di
sta
nc
e
betwee
n
the
‘
’th
w
hale
posit
ion
(
)
to
the
pre
y
po
sit
ion
(
)
′
,
denotes
a
co
nst
ant
for
outl
inin
g
the
lo
gar
it
hm
i
c
curve
st
ru
ct
ur
e
.
E
xponenti
al
functi
on
‘
’
is
the
base
of
natu
ral
lo
gar
it
hm
s.
rep
re
sent
s
the v
al
ue of
th
e cha
otic t
ent
m
ap
(
0,1)
.
The
hum
pb
ack
wh
al
es
encl
ose
the
pr
ey
in
a
dw
in
dlin
g
ci
r
cl
e
and
al
so
a
coile
d
struct
ured
path
way
con
c
urre
ntly
.
In
orde
r
to
perf
or
m
these
con
c
urren
t
pe
rfo
rm
ances,
c
on
si
de
r
that
there
is
a
po
ssi
bili
ty
of
0.5
to
el
ect
eit
her
dw
ind
li
ng circl
e
or
helix
-
s
ha
ped
m
od
el
. Th
e
pre
ci
se m
od
el
is expresse
d
as
fol
lows
,
(
+
1
)
=
{
′
(
)
−
.
;
<
0
.
5
′
∗
(
2
)
+
′
(
)
;
≥
0
.
5
(11)
In (11),
(
+
1
)
repres
ents the
up
dated po
sit
io
n of t
he
sen
s
or
nodes
,
‘
’
is a
pro
bab
il
it
y rang
es
fro
m
[0, 1].
3.2.3. E
xp
l
oration
phase
The
final
ph
ase
of
the
pro
po
se
d
cha
otic
wh
al
e
m
et
aheu
risti
c
op
ti
m
iz
at
ion
te
chn
iq
ue
is
the
ex
plorat
ion
ph
a
se
i.e.
sea
r
ches
f
or
prey
.
The
global
opt
i
m
u
m
so
luti
on
is
ob
ta
ine
d
by
updatin
g
the
w
hale'
s
po
sit
io
n
with
a
ra
ndom
ly
chosen
w
hale
rath
er
t
han
the
c
urre
nt
best
wh
al
e.
T
his
m
echan
ism
us
es
t
he
|
|
>
1
wh
ic
h
hi
gh
li
ghts
the
ex
plorat
ion
a
nd
pe
rm
i
ts
the
pr
opose
d
c
hao
ti
c
wh
al
e
m
et
aheu
risti
c
optim
iz
at
io
n
al
gorithm
that perform
s a g
lo
bal searc
h. T
he
upd
at
in
g be
ha
vior is e
xpress
ed
as
foll
ows:
=
|
.
(
)
−
(
)
|
(12)
(
+
1
)
=
(
)
−
.
(13)
Fr
om
(13),
(
)
den
otes
a
n
a
rb
it
r
ary
locat
ion
ve
ct
or
of
a
ha
phazar
d
wh
al
e.
Finall
y,
the
ch
aotic
w
hale
m
et
aheu
risti
c
op
ti
m
iz
ation
al
gorithm
is
st
oppe
d
by
sat
isfyi
ng
the
te
r
m
inati
on
crit
erio
n.
By
th
is
way,
energy
ef
fici
en
t
sensor
nodes
are
sel
ect
ed
from
the
popu
la
t
ion
.
A
fter
t
hat,
the
le
sse
r
ene
rg
y
se
nsor
node
s
ar
e
e
m
plo
ye
d
to
di
rect
the
colle
ct
ed
data
packet
s
towards
ne
arest
highe
r
e
nergy
sen
sor
node
s
f
or
tr
ans
f
err
in
g
the
pack
et
s
to
the
sink
no
de
.
The
Eucli
dea
n
distance
m
easur
e
is
us
e
d
for
find
in
g
nei
ghborin
g
highe
r
energy
nodes
in
t
he
se
arch space
. T
he
Eu
cl
idea
n dis
ta
nce is c
om
pu
te
d
as
fo
ll
ows
,
=
√
∑
(
(
)
−
ℎ
(
)
)
2
(14)
Fr
om
(
14),
re
pr
ese
nts
t
he
E
uclidean
distance,
(
)
de
no
te
s
le
sser
e
nergy
sens
or
no
des
and
ℎ
(
)
represe
nts
the
higher
ene
r
gy
sens
or
no
des.
Lesser
e
ne
rg
y
sens
or
no
des
f
ind
t
he
neig
hb
or
i
ng
hi
gh
e
r
e
nergy
sens
or
nodes
t
o
tran
sm
i
t
the
pack
et
s
to
wa
rds
sink
.
Af
te
r
th
at
,
the
source
node
se
nds
accum
ulate
d
packet
s
to
the s
in
k
t
hroug
h
the
adjace
nt
higher
stre
ng
t
h sens
or no
des.
Figure
4
il
lustr
at
es
a
data
gat
her
i
ng
thr
ough
the
op
ti
m
al
e
nergy
ef
fici
ent
sens
or
nodes.
In
F
ig
ur
e
4,
the
sou
rce
node
(
S)
pro
pels
the
data
pac
ke
t
to
nea
rest
hi
gh
e
ne
rg
y
se
ns
or
nodes
w
hi
ch
is
re
presen
te
d
i
n
gr
ee
n
colo
r.
T
he
higher
e
nergy
sens
or
no
de
s
direct
the
data
pack
et
s
towards
sin
k
no
de
represe
nted
in
red
colo
r.
Th
e
sin
k
node
c
ollec
ts
data
pack
et
s
from
the
highe
r
energy
no
des
that
ensu
es
in
ref
inin
g
the
li
fetim
e
of n
et
work.
The
al
gorithm
i
c ex
planati
on
of C
W
ME
O
DG
te
chn
iq
ue
is
pr
esented
as
fo
ll
ow
s
,
Inp
ut: N
um
ber
of se
ns
or
node
s (
i.e.
whale
)
1
,
2
,
3
…
,
, N
um
ber
of d
at
a p
ac
kets
1
,
2
,
3
…
Ou
t
pu
t:
Im
pr
oved
en
e
r
gy ef
fici
ent d
at
a
gath
erin
g
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
4
,
A
ugus
t
2020
:
4176
-
4188
4182
Figure
4
.
Data
gathe
rin
g
in
WSN
Begin
1. Initialize the populations of sensor nodes
1
,
2
,
3
…
,
2. for each
3. Compute residual energy
4. Calculate the
5. If (
≥
)
then
6. Select the current best
solution
7. end if
8. While
(
<
)
9. if (
<
0
.
5
)
10. if (
|
|
<1) then
11. Update the location of the present search representative using (4)
12. else if (
|
|
≥
1) then
13. Select a random position of whale
(
)
14. update the location of the present search representative using (13)
15. end if
16. else if (
≥
0
.
5
)
17. update
the location of the present best solution (9)
18. end if
19. Go to step 4
20. t=t+1
21: end while
22. Yield the optimal energy efficient nodes
23: Less energy nodes finds neighboring high energy nodes using
24.
(
)
send
to
ℎ
(
)
25. Source node sends a
to sink through
ℎ
(
)
End
Algorithm 1 Chaotic Whale Metaheuristic Energy Optimized Data Gathering
Algorithm
1
descr
i
bes
t
he
chao
ti
c
wh
al
e
m
e
ta
heu
risti
c
ene
rg
y
op
ti
m
iz
ed
data
ga
therin
g
with
m
ini
m
u
m
dela
y
and
lo
ss
.
T
he
popula
ti
ons
of
the
se
nsor
nodes
are
ini
ti
al
iz
ed
arb
it
rar
il
y.
Fo
r
eac
h
node,
resid
ual
ene
rgy
is
assessed.
Rel
ia
nt
on
t
he
energy,
the
fitn
ess
is
com
pu
te
d
f
or
each
sen
s
or
nod
e
s.
T
he
r
esi
du
al
energy
is
c
ompare
d
by
m
eans
of
the
th
reshold
value
f
or
i
den
t
ify
in
g
t
he
current
best
se
arch
age
nt
(i
.e.
sens
or
nodes
).
Af
te
r
f
ind
in
g
the
c
urr
ent
best
age
nt,
the
pro
posed
C
W
MEO
DG
t
echn
i
qu
e
pe
rfo
rm
s
three
diff
e
ren
t
processes
by
tun
i
ng
the
certa
in
par
am
et
ers
us
in
g
cha
otic
t
ent
m
ap
values
.
Af
te
r
that,
th
e
fitness
is
com
pu
te
d
to
fin
d
the
gl
ob
al
e
nergy
eff
ic
ie
nt
sen
sor
nodes.
P
re
dicat
ed
on
the
te
nt
m
ap,
the
fast
conve
rg
e
nce
of
the
al
gorithm
i
s
at
ta
ined
an
d
sel
ect
s
the
ene
rg
y
ef
fici
ent
se
ns
or
nodes
am
ong
the
de
dica
te
d
quantit
y
of
nodes
dep
l
oyed
i
n
th
e
netw
ork
.
Af
t
er
fi
nd
i
n
g
the
energy
ef
fici
ent
no
des,
the
l
esser
e
nergy
s
ens
or
node
s
tr
ansfe
r
the
com
po
se
d
data
to
the
ne
arb
y
op
ti
m
al
e
nergy
ef
fici
ent
node
t
hroug
h
the
Eu
cl
idean
distance
m
easur
e
.
The
s
ource
no
de
the
n
directs
the
colle
ct
ed
pack
et
s
t
o
the
sink
ov
e
r
the
energy
eff
ic
ie
nt
sensor
no
des
with
m
ini
m
u
m
del
ay
.
This
pro
cess
ups
urges
the
pac
ket
delivery
p
r
oport
ion
a
nd
netw
ork
li
f
et
i
m
e.
The
ab
ove
-
e
xpla
ined
al
gorithm
ic
pr
oces
se
s
are
execu
te
d
in
the
si
m
ula
ti
on
to
sho
w
the
perform
ance
of
the
sugg
est
e
d
C
W
MEO
DG
t
echn
i
qu
e
t
han
the
e
xisti
ng
optim
iz
at
ion
te
c
hn
i
qu
e
.
The
si
m
ula
ti
on
res
ults
are
discusse
d i
n ne
xt secti
on.
4.
SIMULATI
O
N
SET
TI
NGS
An
ef
fici
ent
C
W
MEO
DG
te
chn
i
qu
e
an
d
existi
ng
m
eth
ods
PS
OBS
[1
]
a
nd
AC
O
PSO
[
2]
are
i
m
ple
m
ented
in
NS2.
34
net
work
sim
ulator.
Total
ly
50
0
sens
or
no
des
a
re
distrib
uted
i
n
a
s
qu
a
re
re
gi
on
of
2
(
15
00
m
*
15
00
m
)
fo
r
data
gat
her
i
ng
in
WSN
.
T
he
m
ob
il
it
y
m
od
el
in
t
he
sim
ula
ti
on
is
us
e
d
as
Ra
ndom
Wayp
oin
t.
T
he
nu
m
ber
of
data
pac
kets
us
e
d
f
or
th
e
si
m
ulati
on
pur
po
s
es
are
var
i
ed
f
ro
m
25
t
o
25
0.
The
se
nsor
no
des
s
pee
d
is
se
t
as
0
-
20m
/se
c
an
d
the
sim
ul
at
ion
ti
m
e
is
30
0
sec.
The
dy
nam
ic
so
ur
ce
r
ou
ti
ng
(D
SR
)
prot
ocol
is
app
li
ed
for
carrying
ou
t
e
nergy
eff
ect
ive
data
gather
i
ng
in
WSN.
The
var
i
ou
s
sim
ulati
on
factors a
nd the
ir v
al
ues
a
re lis
te
d
in
Table
1.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ener
gy
ef
fi
ci
en
t chaoti
c wh
ale
opti
miza
ti
on t
echn
i
qu
e
for
data g
ath
eri
ng
…
(
Sr
idhar
R.
)
4183
T
h
e
m
o
d
e
l
i
s
p
e
r
f
o
r
m
e
d
w
i
t
h
r
e
v
e
r
e
n
c
e
t
o
q
u
a
n
t
i
t
y
o
f
s
e
n
s
o
r
n
o
d
e
s
a
n
d
d
a
t
a
p
a
c
k
e
t
s
.
T
h
e
v
a
r
i
o
u
s
p
e
r
f
o
r
m
a
n
c
e
m
e
t
r
i
c
s
s
u
c
h
a
s
e
n
e
r
g
y
u
t
i
l
i
z
a
t
i
o
n
,
p
a
c
k
e
t
d
e
l
i
v
e
r
y
p
r
o
p
o
r
t
i
o
n
,
l
o
s
s
r
a
t
e
o
f
d
a
t
a
p
a
c
k
e
t
a
n
d
d
e
l
a
y
a
r
e
c
o
m
p
u
t
e
d
u
s
i
n
g
t
h
e
a
b
o
v
e
s
a
i
d
s
i
m
u
l
a
t
i
o
n
s
e
t
t
i
n
g
s
.
T
h
e
r
e
s
u
l
t
s
o
f
d
i
f
f
e
r
e
n
t
p
a
r
a
m
e
t
e
r
s
a
r
e
d
i
s
c
u
s
s
e
d
i
n
n
e
x
t
s
e
c
t
i
o
n
.
Table
1
.
Sim
ul
at
ion
facto
rs
Si
m
u
latio
n
Factors
Valu
es
Si
m
u
lato
r
NS2
.34
Netwo
rk Ran
g
e
1
5
0
0
m
* 1
5
0
0
m
Qu
an
tity
of
sen
so
r
n
o
d
es
50,
100,
150,
2
0
0
,
250,
300,
3
5
0
,
4
0
0
,
450,
500
Mob
ility
m
o
d
el
Ran
d
o
m
W
a
y
po
in
t
m
o
d
el
Qu
an
tity
of
Data
p
acket
s
25,
50,
75,
100,
1
2
5
,
150,
175,
200,
2
2
5
,
250
Sen
so
r
n
o
d
es sp
eed
0
-
2
0
m
/s
Si
m
u
latio
n
T
i
m
e
3
0
0
sec
Nu
m
b
e
r
o
f
T
r
acks
10
Proto
co
l
DSR
5.
RESU
LT
S
AND DI
SCUS
S
ION
S
The
sim
ulati
on
ou
tc
om
es
of
pro
j
ect
ed
C
WM
EODG
te
ch
nique
an
d
e
xisti
ng
opti
m
iz
at
ion
te
chn
iq
ue
s
nam
ely
PSO
BS
[1
]
an
d
AC
OP
S
O
[
2]
are
discusse
d
with
var
io
us
par
am
et
ers
su
c
h
as
c
on
s
um
ption
of
energy,
data
pac
ket
de
li
ver
y
pro
portion,
data
pac
ke
t
loss
rate
an
d
delay
.
Per
for
m
ance
of
C
W
MEOD
G
t
ech
nique
is
evaluate
d
with
the
existi
ng
m
et
ho
ds
us
in
g
ta
ble
values
a
nd
gr
a
phic
al
resu
lt
s.
F
or
eac
h
sect
ion,
the
sam
ple
m
at
he
m
at
ic
a
l
cal
culat
ion
is
pro
vid
e
d
for
sh
owin
g
the
pe
rfor
m
ance
resu
lt
s
of
the
propose
d
C
W
M
EO
D
G
te
chn
iq
ue a
nd
existi
ng m
et
ho
ds
.
5.1.
Si
mulati
on
re
sults
of ener
gy u
tili
z
at
ion
Con
s
um
ption
of
e
ne
rg
y
is
ca
lc
ulate
d
as
a
n
a
m
ou
nt
of
vita
li
ty
sp
ent
by
s
ens
or
nodes
f
or
se
ns
in
g
a
nd
gathe
rin
g
the
data
from
the
env
i
ronm
ental
conditi
ons.
Th
e
m
at
he
m
atica
l
fo
r
m
ula
fo
r
c
al
culat
ing
the
energy
consum
ption
is
expres
sed
as
f
ollows,
=
∗
(
)
(15)
Fr
om
(15
),
de
note
s
e
nergy
co
ns
um
ption
of
s
ens
or
node
.
T
he
co
nsum
ption
of
e
nergy
is
r
est
r
ai
ned
in
te
r
m
s
of jo
ule (
J
).
Sa
m
ple
m
at
he
m
a
ti
cal
calc
ulati
on
for
e
nergy c
onsu
m
ption
:
Pr
op
os
e
d
C
WMEODG
te
ch
ni
qu
e:
Total
qua
ntit
y
of
sens
or
nodes
are
ta
ke
n
as
50,
ene
rg
y
con
s
um
ption
f
or
sing
le
se
nsor
nod
e
is
0.54
J
oule
, th
e
n
the
tota
l ener
gy c
on
s
um
pt
ion
is c
ompu
te
d
as
,
=
50
∗
0
.
54
=
27
Existi
ng
PS
O
BS:
Total
nu
m
ber
of
se
nsor
nodes
a
re
ta
ke
n
as
50,
e
nerg
y
con
s
um
ption
for
sin
gle
sen
so
r
node
is
0.65J
oule
, th
e
n
t
he
t
ot
al
en
er
gy cons
um
ption
is c
om
pu
te
d
as,
=
50
∗
0
.
65
=
32
.
5
≈
33
Existi
ng
ACO
PSO
:
T
otal
num
ber
of
sens
or
nodes
are
ta
ke
n
as
50,
e
nergy
consum
ption
f
or
sin
gle
se
ns
or
node
is
0.8Jo
ule, the
n
the
tota
l ener
gy c
on
s
um
pt
ion
is c
ompu
te
d
as
,
=
50
∗
0
.
8
=
40
The
sim
ulati
on
outc
om
es
of
the
en
er
gy
util
iz
at
ion
of
the
se
ns
or
nodes
via
thr
ee
diff
e
re
nt
op
ti
m
iz
ation
te
chn
i
qu
e
s
nam
el
y
CW
ME
O
DG,
PS
OBS
[
1]
an
d
AC
OPSO
[2
]
are
des
cribe
d
in
T
a
bl
e
2.
For
the
si
m
ulati
on
pur
po
ses
,
the
de
finite
qua
ntit
y
of
sen
sor
node
s
is
ta
ke
n
f
ro
m
50
to
500.
T
he
var
i
ous
nu
m
ber
of
sens
or
node
s
are
co
ns
i
der
e
d
an
d
it
re
ported
f
or
eac
h
in
sta
nce.
At
fir
s
t
50
num
ber
of
sens
or
no
de
s
are
consi
der
e
d,
27
J
ene
rg
y
c
ons
um
ption
is
at
ta
ined
i
n
the
pro
posed
C
WMEODG,
exis
ti
ng
P
SO
BS
[
1]
an
d
ACOPSO
[2
]
at
ta
ins
33
J
a
nd
40
J
resp
ect
i
ve
ly
.
In
this
cal
culat
ion
,
t
he
no
des
ene
r
gy
util
iz
at
ion
is
com
pu
te
d
after
se
ns
in
g
a
nd
gat
her
i
ng
t
he
data.
T
he
res
ults
il
lustrate
that
the p
r
opose
d
C
W
ME
O
DG
te
chn
i
qu
e
co
nsum
e
s
le
sser
e
nergy than
the c
onve
nt
ion
al
opti
m
iz
a
ti
on
tech
ni
qu
e
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
10
, No
.
4
,
A
ugus
t
2020
:
4176
-
4188
4184
Total
ly
te
n
vari
ou
s
runs
are
pe
rfor
m
ed
thr
ough
dif
fer
e
nt
qu
antit
y
of
sens
or
node
s.
T
he
outc
om
es
of
the
C
W
ME
O
DG
schem
e
is
com
par
ed
by
m
eans
of
tw
o
pr
ese
nt
te
ch
ni
qu
e
s
PS
OBS
[
1]
an
d
AC
OPSO
[2
]
.
T
he
c
om
par
iso
n
res
ults
sho
w
that
the
C
W
M
EODG
te
c
hn
i
que
m
ini
m
iz
es
c
on
s
um
ption
of
energy
by
12
%
and
19% whe
n rela
te
d
to e
xisti
ng
op
ti
m
iz
ation
te
chn
i
qu
e
s.
Table
2.
T
ab
ul
arizat
ion
f
or
e
ne
rg
y
util
iz
at
ion
Qu
an
tity
of
Sens
o
r
n
o
d
es
Energy
co
n
su
m
p
ti
o
n
(
Jo
u
le)
CWM
EOD
G
PSOBS
ACOPSO
50
27
33
40
100
35
39
45
150
39
47
51
200
42
50
54
250
46
53
58
300
48
56
60
350
55
59
63
400
57
61
64
450
61
65
69
500
63
68
73
5.2.
Simul
at
i
on
re
sults
of d
ata
p
acke
t
deli
ver
y ra
tio
Data
pac
ket
de
li
ver
y
rati
o
is
m
easur
ed
as
th
e
pro
portio
n
of
the
qu
a
ntit
y
of
pac
ket
c
orrec
tl
y
received
at
the sin
k
no
de
to
the total
q
uan
ti
ty
o
f
pa
ck
et
s p
ropell
ed fr
om
the so
urce
node.
T
he
data
p
acket deli
ve
r
y rati
o
is pr
eci
sel
y c
om
pu
te
d
as foll
ow
s
,
=
.
.
∗
100
(
16)
Fr
om
(
16),
den
otes
data
pa
c
ket
delivery
r
at
io,
de
note
s
a
data
pac
ket
co
rr
ect
ly
r
ece
ived,
represe
nts a d
a
ta
p
acket
pro
pe
ll
ed
from
so
ur
c
e nod
e
. T
he
pa
cket d
el
ive
ry ra
ti
o
is m
easur
ed
in
pe
rcen
ta
ge
(
%)
.
Sam
ple
m
at
he
m
at
ic
al
cal
cula
ti
on
for dat
a
pa
cket d
el
i
ver
y
r
at
io:
Pr
op
os
e
d
C
WMEODG
te
ch
nique:
De
fin
it
e
qu
a
ntit
y
of
pac
kets
re
c
ei
ved
at
si
nk
node
is
21
an
d
the qua
ntit
y of
pack
et
s
sen
t i
s
25. T
he
n
the
dat
a p
ack
et
d
el
i
ver
y
rati
o
is c
om
pu
te
d
as,
=
21
25
∗
100
=
84%
Existi
ng
PS
OB
S:
Def
i
nite
qu
a
ntit
y
of
pac
ket
s
receive
d
at
sink
node
is
20
and
t
he
qua
ntit
y
of
pa
ckets
se
nt
is 2
5.
T
he
n
the
d
at
a
packet
d
e
li
ver
y rati
o i
s c
om
pu
te
d
as
,
=
20
25
∗
100
=
80%
Existi
ng
ACO
PSO
:
De
finite
qu
a
ntit
y
of
pa
ckets
receive
d
at
sink
node
is
19
an
d
the
qu
antit
y
of
pack
e
ts
sent is
25. T
he
n
the
d
at
a
pac
ke
t deli
ver
y
rati
o
is c
om
pu
te
d as,
=
19
25
∗
100
=
76
%.
Table
3
de
scri
bes
the
va
rio
us
si
m
ulati
on
resu
lt
s
of
data
pack
et
de
li
very
rati
o
with
de
fer
e
nce
to
a
de
finite
qua
nt
it
y
of
sen
sor
node
s
in
WSN.
The
m
od
el
ou
t
com
es
cl
early
sh
ow
t
hat
the
C
W
MEO
DG
m
et
ho
d
increases
the
data
pack
et
de
li
ver
y
rati
o
w
hen
an
al
yz
ed
t
o
oth
e
r
te
ch
ni
qu
e
s
P
SO
BS
[
1]
an
d
AC
OPSO
[
2].
In
T
able
3,
let
us
c
onsider
the
num
ber
of d
at
a p
ac
kets
are
25, 2
1
pa
ckets c
orrectl
y ac
cept
ed
at
sin
k n
od
e
u
si
ng
C
W
MEO
DG
t
echn
i
qu
e
a
nd
their
per
ce
nta
ge
is
84%.
T
he
de
finite
qu
a
ntit
y
of
data
pa
ckets
acce
pta
ble
at
the
sink
node
us
in
g
PS
OBS
[1
]
an
d
AC
OPSO
[
2]
are
20
and
19
res
pe
ct
ively
.
The
equ
i
valent
pe
rc
entage
value o
f
P
SO
B
S [1] an
d ACO
PSO [
2] are
80
% an
d 7
6% res
pecti
vely
.
Af
te
r
pe
rfor
m
ing
the
te
n
occ
urren
ce
s
wit
h
a
dif
fer
e
nt
c
ount
of
pack
et
s
,
the
pro
pose
d
data
pac
ket
delivery
res
ults
are
com
par
ed
us
in
g
the
exi
sti
ng
res
ults.
Then
the
a
ver
a
ge
value
is
ta
ken
for
the
com
par
iso
n
resu
lt
s.
T
he
a
ve
rag
e
resu
lt
s
pro
ve
that
the
pro
posed
C
WM
EODG
te
ch
nique
co
ns
ide
ra
bly
a
m
end
s
the
data
pack
et
d
el
ive
r
y
rati
o
by
6%
c
om
par
ed
t
o
the
PS
OBS [
1]. S
i
m
il
arly
,
the
propose
d
C
WM
EODG
te
c
hniq
ue
al
s
o
increases
the
da
ta
p
ack
et
d
el
i
ver
y
rati
o by 12% c
om
par
ed wit
h
the
o
t
her
existi
ng tech
niq
ue
A
C
OP
S
O [2
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Ener
gy
ef
fi
ci
en
t chaoti
c wh
ale
opti
miza
ti
on t
echn
i
qu
e
for
data g
ath
eri
ng
…
(
Sr
idhar
R.
)
4185
Table
3.
T
ab
ul
arizat
ion
f
or
da
ta
p
acket
d
el
iv
ery rati
o
Qu
an
tity
o
f
Data
P
ackets
Data pack
et deliv
e
ry
ratio (
%)
CWM
EOD
G
PSOBS
ACOPSO
25
84
80
76
50
90
86
82
75
89
84
77
100
88
83
79
125
85
81
77
150
83
79
75
175
89
81
77
200
90
86
81
225
88
84
82
250
89
82
78
5.3.
Simul
at
i
on
re
sults
of d
ata
p
acke
t
los
s r
at
e
Data
pac
ket
lo
ss
rate
is
asses
sed
as
t
he
pro
portio
n
of
def
i
nite
qu
a
ntit
y
of
pack
et
s
l
os
t
at
sink
node
t
o
the
total
quant
it
y
of
pac
kets
pro
pelle
d
f
ro
m
the
source
no
de.
T
he
data
pa
cket
loss
rate
is
com
pu
te
d
us
in
g
the m
at
he
m
atic
al
eq
uati
on,
=
.
.
∗
100
(17)
Fr
om
(
17),
den
otes
data
pac
ket
los
s
rate,
denotes
a
data
pac
ket
l
os
s,
r
epr
ese
nts
data
pack
e
t
sent
from
so
urce
node
.
The
data
pack
et
l
os
s
rate
is
as
sessed
in
per
c
entage
(
%).
S
a
m
ple
m
at
he
m
at
ic
a
l
cal
cul
at
ion
for dat
a p
ac
ket l
oss rate
:
Pr
op
os
e
d
C
WMEODG
te
ch
ni
qu
e:
De
finite
qu
a
ntit
y
of
data
pack
et
s
l
os
t
at
the
sink
node
is
4
an
d
qu
a
nt
it
y
of d
at
a
pac
kets
tran
sm
it
te
d
is 25. T
he pac
ket loss rat
e is cal
culat
ed
as
foll
ows,
=
4
25
∗
100
=
16%
Existi
ng
PS
OBS:
Def
init
e
qu
antit
y
of
data
pack
et
s
lo
st
at
the
sink
node
i
s
5
an
d
qua
ntit
y
of
data
packet
s
transm
itted is
25. T
he pac
ket loss rat
e is cal
culat
ed
as
foll
ows,
=
5
25
∗
100
=
20%
Existi
ng
AC
O
PSO
:
Def
i
nite
qu
a
ntit
y
of
dat
a
pac
kets
lost
at
the
sin
k
node
is
6
a
nd
qu
antit
y
of
pac
ke
ts
transm
itted is
25. T
he pac
ket loss rat
e is cal
culat
ed
as
foll
ows,
=
6
25
∗
100
=
24%
.
T
a
b
l
e
4
d
e
s
c
r
i
b
e
s
a
d
a
t
a
p
a
c
k
e
t
l
o
s
s
r
a
t
e
v
e
r
s
u
s
a
q
u
a
n
t
i
t
y
o
f
d
a
t
a
p
a
c
k
e
t
s
.
T
h
e
l
o
s
s
r
a
t
e
i
s
m
e
a
s
u
r
e
d
a
t
t
h
e
s
i
n
k
n
o
d
e
w
h
i
c
h
r
e
s
u
l
t
s
i
n
i
m
p
r
o
v
i
n
g
t
h
e
d
a
t
a
g
a
t
h
e
r
i
n
g
e
f
f
i
c
i
e
n
c
y
w
h
e
n
c
o
m
p
a
r
e
d
t
o
c
o
n
v
e
n
t
i
o
n
a
l
o
p
t
i
m
i
z
a
t
i
o
n
m
e
t
h
o
d
s
.
T
h
e
s
i
n
k
n
o
d
e
a
c
t
s
l
i
k
e
a
d
a
t
a
c
o
l
l
e
c
t
o
r
a
n
d
g
a
t
h
e
r
s
t
h
e
d
a
t
a
f
r
o
m
t
h
e
h
i
g
h
e
r
e
n
e
r
g
y
s
e
n
s
o
r
n
o
d
e
s
.
Table
4.
T
ab
ul
arizat
ion
f
or
da
ta
p
acket
loss
ra
te
Qu
an
tity
o
f
Data P
ackets
Data pack
et los
s r
a
te (
%)
CWM
EOD
G
PSOBS
ACOPSO
25
16
20
24
50
10
14
18
75
11
16
23
100
12
17
21
125
15
19
23
150
17
21
25
175
11
19
23
200
10
14
19
225
1
2
16
18
250
11
18
22
Evaluation Warning : The document was created with Spire.PDF for Python.