Indonesi
an
Journa
l
of El
ect
ri
cal Engineer
ing
an
d
Comp
ut
er
Scie
nce
Vo
l.
13
,
No.
1
,
Jan
uar
y
201
9
,
pp.
347
~
353
IS
S
N: 25
02
-
4752, DO
I: 10
.11
591/ijeecs
.v1
3
.i
1
.pp
347
-
353
347
Journ
al h
om
e
page
:
http:
//
ia
es
core.c
om/j
ourn
als/i
ndex.
ph
p/ij
eecs
Resoluti
on of
econ
omic di
spatch p
ro
bl
em of the m
oroc
ca
n
network
usin
g crow sea
rch algo
rith
m
Ra
c
hid H
abac
hi, A
c
hraf T
oui
l, Abdel
lah
Boulal,
Abdel
ka
bir
Chark
aoui
,
Ab
delw
ahed
Ec
hchatbi
La
bora
tor
y
of
Mec
han
ic
a
l Engi
n
ee
ring
,
Industr
ial
Mana
g
ement a
nd
Innova
t
i
on
The
Fa
cul
t
y
of
S
ci
en
ce
s
and
T
echnolog
y
,
Hass
an
1st Uni
ver
si
t
y
,
PO
Box
577,
Se
t
ta
t
,
Moroc
co
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
J
un
1, 20
18
Re
vised Jul
10,
2018
Accepte
d
J
ul
25, 2
018
The
ec
onom
ic
d
ispat
ch
probl
em
of
power
play
s
a
ver
y
i
m
porta
n
t
role
in
the
expl
oitati
on
of
el
e
ct
ro
-
en
erg
y
s
y
stems
to
j
udic
iousl
y
distr
ibut
e
power
gene
ra
te
d
b
y
a
l
l
pla
nts.
Thi
s
pape
r
proposes
the
use
of
Cr
ow
Sear
ch
Algorit
hm
(CSA
)
for
solving
the
e
conomic
dispatch
probl
em
of
two
el
e
ct
ri
ci
t
y
net
wo
rks:
a
t
esti
ng
s
ystem
6
unit
s
and
the
m
oroc
co
ne
twork.
T
h
e
cro
w
sea
rch
al
g
orit
hm
(CSA
)
is
a
recentl
y
develope
d
m
et
ah
eur
i
stic
sea
r
c
h
al
gorit
hm
inspir
ed
b
y
the
intelli
gent
beh
avi
or
of
cro
ws
.
The
resu
lt
s
obtained
b
y
CS
A
are
c
om
par
ed
with
var
ious
result
s
obta
ine
d
in
t
he
li
t
era
tu
re.
Sim
ula
ti
on
resu
lt
s
show
tha
t
u
sing
CS
A
ca
n
le
ad
to
f
indi
ng
stabl
e
and
ade
qua
te
power
gene
r
at
ed
that
ca
n
fu
lfi
l
l
th
e
nee
d
of
both
t
he
ci
v
il
and
industri
al a
r
ea
s.
Ke
yw
or
d
s
:
Crow sea
rch al
gorithm
(
CSA)
Eco
no
m
ic
d
isp
at
ch
pr
ob
le
m
Sm
art g
rid
Copyr
ight
©
201
9
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
:
Ra
chid Ha
bachi
,
Lab
or
at
ory
of
Me
chan
ic
al
E
nginee
rin
g,
Ind
us
tria
l M
an
agem
ent an
d I
nnovat
ion Ha
s
san 1st
Un
i
versi
ty
,
Faculty
of S
ci
e
nces a
nd Tec
hnol
og
y,
P
O
B
ox
577, Sett
at
, M
orocco
Em
a
il
: hab
achi
rach
i
d@gm
ai
l.
com
1.
INTROD
U
CTION
Sm
art
gr
ids
a
re
a
set
of
te
chnolo
gies,
co
ncep
ts
an
d
a
ppr
oac
hes,
al
lo
wing
the
inte
gr
at
io
n
t
he
gen
e
rati
on,
t
ra
ns
m
issi
on
,
distrib
ution
an
d
use
into
one
i
ntern
et
by
f
ull
use
of
a
dv
a
nce
d
sens
or
m
easurem
ent
te
chnolo
gy,
c
om
m
un
ic
at
ion
s
te
chnolo
gy,
in
f
or
m
at
ion
te
ch
nolo
gy,
c
om
pu
te
r
te
ch
nolo
gy,
con
t
ro
l
te
c
hnol
og
y,
new
e
nergy
te
chnolo
gies
[
1].
H
ow
e
ve
r,
Sm
art
G
rid
use
s
di
gital
te
chnolo
gy
to
co
ntr
ol
gri
d
a
nd
c
hoos
i
ng
the
best
m
od
e
of
powe
r
distri
b
ut
ion
to
reduce
energy
co
nsu
m
pt
ion
,
reduce
costs,
i
ncr
eas
e
reli
abili
ty
and
al
s
o
increase
tra
nspare
ncy
in
the
netw
ork.
T
he
refor
e
,
the
sys
tem
intelli
gen
t
will
hav
e
wil
l
hav
e
a
sig
nificant
i
m
pact
in
the
fiel
ds
of
fi
nan
c
e
and
ec
onom
i
cs
of
the
powe
r
industry
[
2
]
.
Althou
gh,
The
tradit
ion
al
net
work
i
s
a
on
e
-
way
net
work
in
wh
ic
h
the
el
ect
rical
ener
gy
pro
duced
in
power
plants
is
chan
nele
d
to
co
nsum
ers
without i
nf
or
m
at
ion
to
create
an
a
uto
m
at
ed
and d
ist
ri
bu
te
d netw
ork of ad
va
nced p
ower
s
upplies.
ED
is
al
so
ap
pl
ie
d
i
n
t
he
inte
gr
at
e
d
syst
em
fo
r
sche
duli
ng
powe
r
plants.
A
few
m
et
ho
ds
ha
ve
bee
n
publ
ished
to so
l
ve
t
he
E
D prob
le
m
an
d O
pti
m
al
Po
we
r
Flo
w (
OP
F
).
Re
searche
rs ha
ve publi
sh
e
d
a
few
m
et
hods
t
o
s
olv
e
ED
an
d
OP
F
pro
blem
s.
Direct
m
et
ho
d
is
accurate
a
nd
ve
ry
si
m
ple
bu
t
lim
it
ed
by
the
qu
a
dr
at
ic
ob
je
ct
ive
functi
on [3].
The
ec
onom
ic
disp
at
c
h
(E
D)
is
on
e
of
t
he
powe
r
m
anag
e
m
ent
too
ls
that
is
us
ed
t
o
det
erm
ine
real
powe
r
out
pu
t
of
t
her
m
al
generati
ng
unit
s
to
m
eet
req
uire
d
loa
d
dem
and.
T
he
E
D
r
esu
lt
s
in
m
ini
m
u
m
fu
el
gen
e
rati
on
co
st,
m
ini
m
u
m
transm
issi
on
power
loss
wh
il
e
sat
isfyi
ng
al
l
un
it
s,
as
well
as
syst
e
m
const
raints
[4
-
5].
The
rise
of
e
ne
rg
y
dem
and
an
d
ins
uffici
ent
of
ene
r
gy
res
our
ces
are
require
d
f
or
qual
it
y
and
secu
re
d
disp
at
c
h
[
6].
A
well
-
c
oor
di
nated
a
nd
opt
i
m
iz
ed
po
wer
syst
e
m
op
era
ti
on
help
i
n
s
at
isfyi
ng
Ec
on
om
i
c
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
13
, N
o.
1
,
Ja
nu
a
ry 20
19
:
347
–
353
348
Disp
at
c
h
(E
D)
a
m
on
g
us
e
rs
of
powe
r
netw
orks
.
He
nce,
st
udie
s
nee
d
to
be
cond
ucted
in
order
t
o
analy
z
e
and
dev
el
op
ne
w
to
ols
so
t
hat
the
op
ti
m
iz
ation
issues
in
E
D
c
ould
be
ov
e
rc
ome
.
Ba
sic
al
ly
,
the
pr
i
ncipal
ob
je
ct
ive
of
l
oad
dis
patch
is
to
m
inim
iz
e
the
total
fu
el
c
os
t
wh
i
le
sat
isfyi
ng
t
he
re
quirem
ents
of
s
om
e
i
m
portant
op
e
rati
onal
pa
ram
et
ers.
In
tod
ay
’s
en
vir
onm
ent,
eff
ic
ie
nt
load
dis
pat
ch
re
qu
ire
s
no
t
on
ly
to
sche
du
le
the
powe
r
ge
ner
at
ion
at
the
le
ast
cost
bu
t
al
so
to
consi
der
oth
er
pe
r
form
ance
factor
s
to
be
op
ti
m
iz
ed
in
power
flo
w
over
the
networ
ks
.
T
he
obli
gatio
n
of
so
ci
al
at
te
ntions
has
i
nf
l
uen
ce
d
the
reducti
on
of
energy
conser
vation a
nd poll
utio
n
e
m
issi
on
pr
oduc
ed by p
ower
p
l
ants [7]
.
Faci
ng
t
he
el
ect
ric
bu
li
m
ia
exp
e
rience
d
by
the
world
a
nd
as
a
n
urge
nt
an
d
ef
fici
ent
so
luti
on
is
so
ug
ht,
it
is
e
ssentia
l
to
op
ti
m
iz
e
the
cost
of
produci
ng
e
le
ct
rici
ty
.
As
su
ch
,
ti
ny
cuts
costs
co
nceal
huge
po
te
ntial
savin
gs
,
this
is
pa
rt
of
this
pa
per
,
a
nd
we
lo
ok
at
t
he
ov
e
rall
opti
m
iz
at
ion
pur
poses
kn
own
ec
onom
ic
load dist
rib
utio
n (O
P
F) o
r
ec
onom
ic
d
ispatc
h (E
D) [
8].
The
E
D
is a sta
ti
c p
roblem
is t
o
say
we m
us
t def
i
ne
at
a
giv
e
n powers
g
e
ne
r
at
ed
by eac
h p
ow
e
r plant
to po
wer a loa
d
as
eco
nom
icall
y as po
s
sible
. T
o
s
olv
e t
his
pro
b
le
m
the o
pt
i
m
iz
at
ion
m
eth
ods
are
us
e
d.
Conve
ntion
al
op
ti
m
iz
ation
t
echn
i
qu
e
s
[
9
-
10]
.
ha
ve
l
ong
bee
n
a
pp
li
ed
to
s
olv
e
t
he
ED
pro
blem
s
uch
as
Qu
a
drat
ic
Pro
gr
am
m
ing
[
11
-
12]
.
li
nea
r
program
m
ing
[13]
seq
ue
ntial
appr
oach
with
a
m
at
rix
fr
a
m
ewor
k
(S
AM
F)
,[14
]
.
m
od
i
fied
La
m
bd
a
-
it
erati
on
m
e
tho
d
[
15
]
,
New
t
on
Ra
phs
on
an
d
La
gran
gian
m
ulti
plier
(LM
)
al
gorithm
s
[1
6]
,
In
the
real
-
de
sign
cases,
th
e
nu
m
ber
of
de
ci
sion
va
riabl
es
(i.e.
powe
r
un
it
s)
of
the
E
D
area
are very l
ar
ge.
The object
ive
crit
erion to
be m
ini
m
iz
ed
co
ul
d
al
s
o ha
ve
to
o
m
any local
m
ini
m
u
m
w
hich
m
igh
t
no
t
le
ad
t
o
the
m
ini
m
u
m
cos
t
and
the
best
gen
e
rati
on
sch
edu
le
of
powe
r
syst
e
m
un
it
s.
Ther
e
fore,
e
ffi
ci
ent
search
alg
or
it
hm
s ar
e n
eede
d.
Natu
re
-
in
sp
ire
d
m
e
ta
heu
rist
ic
search
al
gorithm
s
gain
a
po
pula
rity
du
e
to
thei
r
prom
isi
ng
perform
ance
on
so
l
ving
m
any
real
-
w
or
l
d
optim
iz
at
ion
prob
le
m
s
wh
ic
h
are
com
plex,
nonlinea
r
an
d
m
ul
ti
-
m
od
el
. I
n t
he p
ast
two deca
de
s,
the
li
te
ratur
e
of m
et
aheu
risti
c search ha
s e
xp
a
nded
ex
te
nsi
vely
.
So
m
e
of
the
well
-
kn
own
m
et
aheurist
ic
app
r
oac
hes
are
G
eneti
c
Algorith
m
s
[1
7],
Ge
net
ic
Pr
ogram
m
in
g
[
18
-
20
]
,
Pa
rtic
le
Sw
arm
Op
ti
m
izati
on
[21
-
22]
,
Si
m
ulate
d
A
nneal
ing
[
23]
,
Ar
ti
fici
al
Be
e
Colo
ny
(
ABC)
[
24
]
,
Cuck
oo Sea
rc
h
[
25
-
26]
, etc.
The
rem
ai
ning
org
anizat
io
n
of
this
pa
per
is
as
fo
ll
ows
.
Sect
ion
2
presen
ts
t
he
m
at
hem
atica
l
form
ulati
on
of
the
ED.
H
an
dling
of
co
ns
tr
ai
nt
s
and
im
ple
mentat
ion
of
the
pr
op
os
e
d
CSA
to
ED
pr
ob
l
e
m
are
addresse
d
in
S
ect
.
3.
Sect
io
n
4
re
ports
res
ults
of
the
pro
po
s
ed
CSA
m
et
hod.
A
num
ber
of
case
stu
dies
us
in
g
sta
nd
a
rd
te
st
syst
e
m
s
are
us
e
d
to
te
st
the
pr
opos
e
d
m
e
tho
d.
The
com
par
ison
s
of
res
ults
betwee
n
the
propo
s
e
d
m
et
ho
d
a
nd
e
xi
sti
ng
m
et
ho
ds
are
al
so
car
ried
out
in
this
se
ct
ion
.
T
he
disc
us
sio
n
is
f
ollo
wed
i
n
Sect
.
5.
Af
te
r
al
l, the c
on
cl
usi
on
is
g
i
ven.
2.
MA
T
HEM
AT
ICA
L
FO
R
M
ULATIO
N O
F ED
The
ov
e
rall
res
earch
m
et
ho
do
log
y
in
vo
l
ve
d
in
the
Ec
onom
i
c
Disp
at
c
h
(E
D)
was
cl
assifi
ed
into
f
ou
r
sta
ges.
T
he
fir
st
ta
sk
was
to
achieve
the
obj
ect
iv
e
of
the
stud
y
w
hich
was
to
est
abli
sh
a
ne
w
te
ch
nique
par
ti
cula
rly
to
so
lve
E
D
optim
iz
ation
pro
bl
e
m
.
The
re
se
arch
a
ppr
oach
was
to
desi
gn
a
ne
w
opti
m
iz
at
ion
te
chn
iq
ue
ta
ki
ng
s
om
e
insp
irat
ion
from
th
e
Me
ta
-
EP
m
utati
on
strat
e
gy
.
The
de
velo
pm
ent
al
so
include
d
identify
in
g
su
it
able
obj
ect
ive
functi
ons
w
hich
wer
e
si
gn
ific
ant
to
ED
pro
bl
e
m
a
lon
g
with
so
m
e
con
strai
nts
a
s
discusse
d
in
th
e
fo
ll
owin
g
se
ct
ion
[
27
]
.
I
n
order
to
ac
hie
ve
the
resea
rc
h
obj
ect
i
ve,
th
e
dev
el
op
m
ent
of
the
new
sin
gle
obje
ct
ive
te
ch
nique
was
t
o
be
a
ccom
plished
.
The
perf
or
m
ance
of
the
de
ve
lop
e
d
te
c
hn
i
que
wa
s
evaluate
d
a
nd
com
par
ed
with
oth
e
r
te
c
hn
i
ques
nam
el
y
the
AI
S
an
d
Me
ta
-
EP
al
on
g
wit
h
Ba
se
te
chn
i
qu
e
.
T
he
dev
el
op
e
d
te
ch
niques
we
re
te
ste
d
on
the
sta
nd
a
r
d
IEEE
26
and
57
bus
syst
e
m
in
or
de
r
to
m
ini
m
iz
e
the
total
fu
el
c
os
t, em
issi
on d
is
per
se
d and sy
ste
m
losses.
The
ob
j
ect
ive
f
un
ct
io
n
of
the
ED
pr
ob
le
m
i
s
to
m
ini
m
iz
e
t
he
total
pro
du
c
ti
on
cost,
w
hich
be
wr
it
te
n
as:
=
∑
(
)
=
1
,
2
,
…
.
,
=
1
(1)
Ma
them
a
ti
call
y, the
fu
el
c
os
t
of a the
rm
al
g
ener
at
io
n u
nit i
s r
e
pr
ese
nted
as
quadrat
ic
func
ti
on
[4
]
:
(
)
=
+
+
2
(2)
The
s
olu
ti
on
of
ED
c
a
n
be
hi
gh
ly
i
m
pr
ove
d
by
intr
oduc
ing
highe
r
orde
r
ge
ner
at
or
c
ost
functi
ons
.
Cub
ic
c
os
t
fun
ct
ion
disp
la
ys t
he
act
ual
r
es
po
ns
e
of the
rm
al
gen
e
rato
rsm
or
e accu
ratel
y.
The
c
ubic
fuel
cost fu
nction o
f
a t
her
m
al
g
en
erati
ng unit
is re
pr
ese
nted
as
f
ollows [
28]
:
(
)
=
+
+
2
+
3
(3)
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Resolutio
n of e
cono
mic d
is
patc
h pr
ob
le
m
of the mor
occ
an
ne
tw
or
k
…
(
Rac
hid
H
abac
hi
)
349
S
ubj
ect
to
Re
a
l
po
we
r
bala
nc
e
equ
at
io
n
The
total
active
po
we
r
outp
ut
of
gen
e
rati
ng
unit
s
m
us
t
be
equ
al
t
o
total
powe
r
l
oad d
em
and p
l
us
powe
r
loss:
∑
=
=
1
+
(4)
wh
e
re t
he pow
er lo
s
s PL
is ca
lc
ulate
d by the
belo
w form
ulatio
n [
4]:
=
∑
∑
=
1
=
1
+
∑
=
1
+
(5)
Gen
e
rato
r
cap
aci
ty
lim
it
s
Th
e
act
ive
po
we
r
ou
tp
ut
of
generati
ng
un
it
s
m
us
t
be
within
t
he
al
lowe
d
lim
it
s:
.
≤
≤
.
(6)
3.
CROW
SEA
R
CH ALG
ORIT
HM
In
this
sect
io
n,
it
is
exp
la
ined
the
res
ul
ts
of
resea
rch
and
at
the
sam
e
tim
e
is
giv
en
th
e
com
pr
ehe
ns
ive
discussi
on.
Re
su
lt
s
can
be
presented
i
n
fi
gures,
gr
a
phs,
ta
bl
es
and
ot
her
s
t
hat
m
ake
the
r
eade
r
unde
rstan
d
eas
il
y
[2
]
, [5]
. T
he
d
isc
us
sio
n
ca
n be m
ade in
s
ever
al
s
ub
-
c
ha
pters.
The
c
row
sear
ch
al
gorithm
(CSA
)
is
a
ne
w
popula
ti
on
-
ba
sed
stoc
hastic
search
al
gorith
m
recently
pr
opose
d
by
[29].
The
CSA
is
a
new
l
y
dev
el
ope
d
o
pti
m
iz
at
ion
te
c
hn
i
qu
e
to
s
olve
com
plex
engi
neer
in
g
opti
m
iz
at
ion
pro
blem
s
[3
0
-
31
]
.
It
is
ins
pi
red
by
the
i
nt
el
li
gen
t
be
havi
or
of
cr
ows.T
he
pr
in
ci
ples
of
CS
A
a
re
li
ste
d
a
s
fo
ll
ows
[
29
]
:
a.
Crows li
ve
in
t
he fo
rm
o
f
the
floc
k.
b.
Crows m
e
m
or
iz
e the
posit
ion
of their
hid
i
ng
places.
c.
Crows
fo
ll
ow
each
oth
e
r
to
c
omm
it
thievery.
Crows
protect
their cac
hes fr
om
b
ei
ng
pilfe
red thr
ough
pr
ob
a
bili
ty
.
Fo
ll
owin
g
the
above
ass
um
ption
s
,
the
c
or
e
m
echan
ism
of
the
CSA
c
onsist
s
of
th
ree
ba
sic
ph
ase
s,
nam
ely
init
ia
lizat
ion
,
ge
nerat
e
a
new
posit
ion
,
a
nd
up
da
ti
ng
t
he
m
e
m
or
y
of
cr
ows.
At
first,
the
init
ia
l
popula
ti
on
of
c
rows
re
pr
es
ent
ed
by
n
dim
ension
is
ra
ndoml
y
gen
erated
.
At
it
erati
on
t,
the
posit
ion
of
crow
is
sp
eci
fied
by
x
i
,
t
=
[
x
1
i
,
t
,
x
2
i
,
t
,
…
.
.
,
x
n
i
,
t
]
and
it
is
assum
ed
t
hat
t
his
cr
ow
has
m
e
m
or
i
zed
it
s
best
e
xperie
nce
th
us
far
in
it
s
m
e
m
or
y
m
i
,
t
=
[
m
m1
i
,
t
,
m
2
i
,
t
,
…
.
.
,
m
n
i
,
t
]
To
generat
e
a
ne
w
posit
ion,
c
r
ow
i
s
el
ect
ra
ndom
ly
a
crow
j,
for
e
xam
ple,
from
the
popula
ti
on
a
nd
at
te
m
p
ts
to
f
ollow
it
to
fi
nd
the posit
ion
of
it
s
hid
i
ng p
la
ce
(m
j
).
In
this
case, acc
ordin
g
to
a
par
am
et
e
r nam
ed
awar
e
ness pro
ba
bili
t
y (A
P
), t
w
o
sta
te
s m
a
y happe
n:
a.
Stat
e
1:
Crow
j
does
not
know
that
cr
ow
i
is
fo
ll
owin
g
it
.
As
a
re
su
lt
,
th
e
crow
i
will
de
te
rm
ine
th
e
hid
in
g place
of cr
ow j.
b.
Stat
e
2
:
Crow
j
knows
that
crow
j
is
f
ollow
i
ng
it
.
As
a
resu
lt
,
to
pr
otect
it
s
cache
fr
om
bein
g
pilfere
d,
t
he
c
r
ow j will
fo
ol c
row
i
by
go
i
ng
to anothe
r p
os
it
ion
wh
it
in t
he se
arch s
pace.
Accor
ding to
S
ta
te
s 1
a
nd 2, t
he posi
ti
on
of the c
rows
is
up
dated
a
s foll
ow
s:
x
i
,
it
e
r
+
1
=
{
x
i
,
it
e
r
+
1
+
r
i
×
fl
i
,
iter
×
(
m
i
,
ite
r
−
x
i
,
it
e
r
)
,
,
r
j
≥
A
P
j
,
i
te
r
A
ra
ndo
m
p
osi
t
ono
f
searc
h
s
pa
c
eot
he
rwi
se
(7)
Wh
e
re
rj
is
a
un
i
form
l
y
distr
ibu
te
d
f
uzzy
num
ber
f
ro
m
[0;
1]
an
d
A
P
j
,
iter
de
note
s
the
a
wa
ren
e
ss
pro
ba
bili
ty
of cr
ow j at
it
erati
on it
er. Fi
na
ll
y, the cr
ow
s
update thei
r
m
e
m
or
y as
fo
ll
ows:
m
i
,
iter
+
1
=
{
x
i
,
iter
+
1
,
if
f
(
x
i
,
iter
)
is
be
tter
th
a
n
f
(
m
i
,
iter
)
m
i
,
ite
r
,
othe
rwi
se
(8)
Wh
e
re f(
-
)
de
note
s the object
ive fun
ct
io
n value.
It is seen
th
at
if th
e fitness f
unct
ion
value
o
f
the
new
po
s
it
io
n
of
a
cr
ow
is
be
tt
e
r
than
the
fitness
f
un
ct
io
n
va
lue
of
t
he
m
e
m
or
iz
ed
posit
ion,
the
cr
ow
updates
it
s
m
e
mo
ry
by
the
ne
w
posit
ion.
T
he
a
bove
process
is
re
pe
at
ed
unti
l
a
gi
ve
n
te
rm
inati
on
crit
erion
(ite
rm
ax
)
is
m
et
.
Finall
y,
the
best
so
luti
on
of
the
m
e
m
o
ries
is
returned
as
the
op
ti
m
a
l
so
luti
on
fou
nd
by
the
CSA.
T
he
m
a
in
ste
ps
of
th
e
CSA a
re
ou
tl
in
ed
in
A
l
gorith
m
:
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
13
, N
o.
1
,
Ja
nu
a
ry 20
19
:
347
–
353
350
Algori
thm
Cro
w Sea
rch
Algori
thm
1
:
Randoml
y
initial
iz
e
the posit
ion of a
floc
k
of
(N
P
)
cro
ws
in
the se
arc
h
spac
e.
2
:
Eva
lu
at
e
th
e
pos
it
ion
o
f
th
e
Cro
ws
3
:
Init
iali
ze
the m
emory
of ea
ch
Cr
ow
4
:
While
(
iter
≤
iter
max
)
do
5
:
for
i
=
1:
to
N
P
do
6
:
Randoml
y
choos
e
one
of the
cro
ws
to
foll
ow
(for
exa
m
ple, j)
7
:
Defi
ne an awareness
proba
bilit
y
8
:
if
(r
j
≥
AP
j,i
ter
)
then
9
:
x
i,i
ter+
1
=x
i,i
ter
+r
i
∗
fl
i,i
ter
∗
(m
j,i
ter
−x
i,i
ter
)
10
:
el
se
11
:
x
i,i
ter+
1
=a ra
ndom
positi
on
of
se
ar
ch
spac
e.
12
:
end if
13
:
end for
14
:
Chec
k
the fe
asib
il
ity
of
n
ew
posit
ions
15
:
Eva
lu
at
e
th
e
ne
w posit
ion
of
th
e
Crows
16
:
Update
the m
emor
y
of crows
17
:
end
w
hile
4.
RESU
LT
S
AND DI
SCUS
S
ION
In
t
his
sect
io
n,
we
prese
nt
t
he
re
su
lt
s
ob
t
ai
ned
base
d
on
CS
A
f
or
s
olv
in
g
the
ED
pro
blem
and
com
par
e
this
r
esults
wit
h
the
CM
(
C
onve
ntion
al
Me
th
od)
[15]
an
d
Partic
le
Sw
a
rm
Op
ti
m
iz
at
ion
[32].
A
si
x
un
it
s
po
wer
un
it
syst
e
m
to
exp
lo
re
our
i
dea
on
us
in
g
CS
A
to
fin
d
the
op
ti
m
al
set
of
pow
e
r
ge
ner
at
io
n
of
the
syst
e
m
.
CSA
will
be
us
e
d
i
n
this
pa
per
to
so
l
ve
the
pro
blem
of
eco
no
m
ic
disp
at
ch
of
a
te
st
netw
ork
of
26
nodes
and t
he
m
or
occo
netw
ork. The
pr
ogr
a
m
s ar
e d
e
vel
oped
in
M
ATL
AB 7.9 e
nv
ir
onm
ent.
The
ad
opte
d
s
yst
e
m
is
exp
ec
te
d
to
pro
duce
dem
a
nd
powe
r
of
700
M
W.
The
tu
ning
pa
r
a
m
et
ers
fo
r
CSA
a
re
giv
e
n
in
Ta
ble
1,
the
Table
2
s
hows
the
c
os
t
coe
ff
i
ci
ent
of
the
six
ge
ner
at
or
s
,
un
der
stu
dy,
wh
il
e
the
m
at
rix
is
the
loss
coeffic
ie
nt
m
at
rix
of
the
six
unit
s
power
s
yst
e
m
.
Fr
om
th
e
resu
lt
s
of
Ta
ble
3,
w
e
noti
ce
that
CSA
gi
ve
us
the
sam
e
pr
oduc
ti
on
cost,
a
nd
CM
giv
es
a
sli
gh
tl
y
lowe
r
co
st
of
$
0.0
8
/
h,
by
cons,
delta
giv
en
by
CM
is
la
rger
tha
n
that
gi
ven
by
CS
A,
m
ean
it
is
le
ss
accu
rate
al
though
we
al
s
o
se
e
that
it
is
the
fastest
.
CSA
giv
es
u
s
a
good p
ro
duct
ion co
st an
d g
ood ac
cu
racy.
Table
1.
p
ar
am
et
ers
of CS
A [
34
]
Alg
o
rith
m
/p
ara
m
et
ers
AP
fl
CSA
0
.2
2
In
Fi
gure
1,
w
e
sh
ow
the
c
on
verge
nce
of
th
e
m
et
aheu
risti
c
search
process
based
on
CSA
in
bo
t
h
th
e
best
an
d
aver
a
ge
cases.
T
o
see
the
diff
e
ren
c
e
between
our
new
a
ppr
oach
and
a
no
t
her
known
m
et
ho
d,
we
will
com
par
e
the
pr
oductio
n
co
st
f
ound
by
CS
A
to
that
f
ound
by
PSO
[
33]
.
T
he
com
par
iso
n
is
rep
re
sente
d
by
the
gr
a
ph
in
fi
gur
e
2.
It
ca
n
be
seen
t
hat
CS
A
pro
vid
e
d
t
he
m
ini
m
u
m
fu
el
cost
i
n
this
ca
se
com
par
e
d
t
o
oth
e
r
repor
te
d
m
et
ho
ds
in
th
e
li
te
ratur
e
.
This
s
how
s
that
the
CSA
is
m
or
e
eff
ect
ive
in
fi
nd
i
ng
t
he
be
st
load
for
th
e
three
ge
ner
at
or syst
e
m
.
In
this
case
,
w
e
will
te
st
the
op
e
rati
on
of
CSA.
F
or
this,
we
will
us
e
a
s
i
m
ple
netwo
r
k
of
26
nodes
with
6
pr
oduct
ion
unit
s.
T
he
total
dem
and
of
the
netw
ork
is
equ
al
to
70
0
M
W
an
d
los
s
coe
f
fici
ents
are
as
fo
ll
ows:
=
10
−
5
[
1
.
4
1
.
7
1
.
5
1
.
9
2
.
6
2
.
2
1
.
7
6
.
0
1
.
3
1
.
6
1
.
5
2
.
0
1
.
5
1
.
3
6
.
5
1
.
7
2
.
4
1
.
9
1
.
9
1
.
6
1
.
7
7
.
2
3
.
0
2
.
5
2
.
6
1
.
5
2
.
4
3
.
0
6
.
9
3
.
2
2
.
2
2
.
0
1
.
9
2
.
5
3
.
2
8
.
5
]
The
sim
ulati
on r
es
ults are
pre
sented
in Ta
ble
3.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Resolutio
n of e
cono
mic d
is
patc
h pr
ob
le
m
of the mor
occ
an
ne
tw
or
k
…
(
Rac
hid
H
abac
hi
)
351
Table
2.
T
he
param
et
ers
of
t
he
co
st
functi
on
and g
e
ne
rators
lim
it
s o
f
t
he
si
x
-
unit
syst
e
m
Un
its
a (
$
/M
W
2
)
b
(
$
/MW) c
(
$
)
P
m
in
(
M
W
)
P
m
ax
(
MW)
1
0
.15
2
4
0
3
8
.53
756.7
9
8
8
6
1
0
12
5
2
0
.10
5
8
7
4
6
.15
9
1
6
451.3
2
5
1
3
10
150
3
0
.02
8
0
3
4
0
.39
6
5
5
1049
.9977
35
225
4
0
.03
5
4
6
3
8
.30
4
4
3
1242
.5311
35
210
5
0
.02
1
1
1
3
6
.32
7
8
2
1658
.5696
13
0
3
2
5
6
0
.01
7
9
9
3
8
.27
0
4
1
1356
.6592
12
5
3
1
5
Table
3
. Res
ults o
f
the
eco
no
m
ic
d
ispatc
hing
of six
-
unit
syst
e
m
CM
PSO
CS
A
P1
(
M
W
)
2
9
.45
5
2
2
8
.28
2
9
.45
P2
(
M
W
)
10
10
10
P3
(
M
W
)
1
1
8
.814
1
1
9
.02
1
1
8
.702
P4
(
M
W
)
1
1
8
.420
1
1
8
.79
1
1
8
.339
P5
(M
W
)
2
3
0
.559
2
3
0
.78
2
3
0
.441
P6
(
M
W
)
2
1
2
.511
2
1
2
.56
2
1
2
.394
Pl (
M
W
)
1
9
.33
6
1
1
9
.43
1
9
1
9
.31
2
8
Delta (M
W
)
0
.00
0
7
0
2
0
.00
6
6
8
8
0
.00
0
6
7
7
Fu
el
co
st ($/h
)
3
6
9
2
6
3
6
9
1
2
.16
3
6
9
0
6
t (
s)
6
4
.61
4
3
.72
4
3
.45
Figure
1.
The
c
onve
rg
e
nce
of
CSA in t
he
cas
e of ec
onom
ic
disp
at
c
h of
six
-
unit
syst
em
Figure
2
.
Com
par
is
on gra
ph
betwee
n
CS
A and PS
O
5.
CONCL
US
I
O
N
In
this p
a
per, w
e p
r
opose
d
a
crow sear
c
h
a
lgorit
hm
to
so
lve the p
roblem
of
ec
onom
ic
d
i
sp
at
ch of
the
m
or
occo
el
ect
r
ic
it
y
netwo
r
k.
The
pr
act
ic
al
it
y
of
the
pro
posed
m
et
aheu
ris
ti
cs
CSA
was
te
ste
d
for
six
powe
r
gen
e
rato
rs
te
st
case.
T
he
gai
ne
d
res
ults
we
re
com
par
ed
t
o
e
xisti
ng
res
ults
base
d
on
P
SO
and
CM
m
et
hods
.
It
was
s
how
n
th
at
CSA
are
s
up
e
rio
r
in
obta
ining
a
com
bin
at
io
n
of
power
l
oad
s
t
hat
fu
lfil
l
the
prob
le
m
const
raints
an
d
m
ini
m
iz
e
the
total
fu
el
cost.
CSA
f
ou
nd
to
be
eff
ic
ie
nt
in
find
i
ng
the
optim
al
power
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
13
, N
o.
1
,
Ja
nu
a
ry 20
19
:
347
–
353
352
gen
e
rati
on
loa
ds
.
CS
A
was
c
apab
le
of
ha
nd
li
ng
the
non
-
li
near
it
y
of
ED
pro
blem
.
The
evo
l
ved
powe
r
us
in
g
CSA
m
ini
m
iz
e
d
both
the
cost
of
ge
ner
at
e
d
powe
r,
the
total
powe
r
loss
in
the
transm
issi
on
an
d
m
axi
m
izes
the
reli
abili
ty
of
the
power
pro
vi
de
d
to
the
custo
m
ers.
The
pro
gram
s
wer
e
devel
op
e
d
us
in
g
M
ATLA
B
a
nd
te
ste
d
a
netw
ork
of
26
no
des.
The
r
esults
ha
ve
s
how
n
t
hat
our
CSA
to
gi
ve
us
a
bette
r
perf
or
m
ance
with
op
ti
m
al
resu
lt
s i
n
al
l ca
ses and
res
pecti
ng
t
he
c
onstra
ints im
po
sed.
ACKN
OWLE
DGE
MENTS
The
a
uthors
a
r
e
ver
y
m
uch
t
hank
fu
l
to
t
he
un
a
nim
ou
s
re
viewe
rs
of
the
pap
e
r
a
nd
edi
tors
of
t
he
j
ou
rn
al
for
t
heir
c
on
st
ru
ct
ive
and h
el
pful c
om
m
ents that im
pr
ov
e
d
the
quali
ty
o
f
t
he pa
per.
REFERE
NCE
S
[1]
C.
He
-
Rui,
P.
Xu,
"S
tud
y
on
Sm
a
rt
Grid
Sy
st
em
Based
on
Sy
st
e
m
Dy
namics",
T
EL
KO
MN
IKA
I
ndonesia
n
Journa
l
of
E
l
ec
tr
ical Eng
ine
er
ing
Vol
.
12
,
No.
12
,
pp
.
797
9
-
7986,
De
ce
m
ber
2014.
[2]
Shahinz
ad
eh
H,
Hasana
li
z
ade
h
-
Khos
roshahi
A.
"Im
ple
m
ent
at
io
n
of
S
m
art
Mete
ring
S
y
st
ems
:
Chal
le
ng
es
and
Soluti
ons"
.
T
ELK
OM
NI
KA
Indone
sian
Journa
l of E
l
ec
tr
ical Eng
ine
er
ing. 2014;
1
2(7).
[3]
Herm
aga
santos
Z,
Yus
ra
S,
Ali
M.
Im
ple
m
ent
a
tion
of
E
lectr
i
ci
t
y
Com
pet
it
ion
Fr
amework
with
E
conomic
Dispat
c
h
Dire
ct Met
hod
.
Te
lkomnika
.
20
12;
10(4):
667
-
6
74.
[4]
W
oll
enbe
rg
B
,
W
ood
A (1996) P
ower
gene
r
at
io
n,
oper
ation
a
nd
cont
rol
.
W
ile
y
,
New York
[5]
D
ie
u
VN
,
Schegne
r
P,
Ongs
akul
W
(2013)
P
seudo
-
gra
die
n
t
base
d
par
ticl
e
sw
arm
opti
m
iz
at
ion
m
et
hod
for
nonconve
x
ec
on
om
ic
dispatch.
I
n:
Pow
er,
cont
ro
l
and
opti
m
i
za
t
io
n.
Springer
,
N
e
w York,
pp
1
–
27
.
[6]
Muhtaz
aru
dd
in,
M.N.B.
,
B
ani,
N.A.,
Aris,
S
.
A.
M.,
Kaid
i,
H
.
M.
,
Fata
h
,
A.Y
.
A.,
Jam
ia
,
J.J.
,
Mu
hamm
ad
-
Sukki,
F.
and
Abu
-
Baka
r
,
S.H.
“
Distribution
power
loss
m
ini
m
iz
at
ion
vi
a
distri
bu
te
d
ge
ner
ation,
ca
p
ac
i
tor
and
net
wor
k
rec
onfigur
at
ion
”
.
Indone
si
an
Jou
rna
l
of
E
le
c
trica
l
Engi
n
ee
ring
an
d
Com
pute
r
Sci
e
nce
.
2017;
5(3):
488
-
495
[7]
Arriffi
n,
A.M
.
,
Othm
an,
M.M.,
Kam
aru
za
m
an,
A.A.M.
,
Mus
ir
in,
I
.
,
Yah
y
a
,
A.,
&
La
t
ip,
M
.
F.A.
“
Stocha
st
i
c
Approac
h
of
V
olt
ag
e
Optimiz
a
ti
on
to
Maximiz
e
Pow
er
Saving
in
a
Bui
ldi
ng”
.
Indone
sian
Journal
of
Elec
tri
c
a
l
Engi
ne
eri
ng
and
Com
pute
r
Sci
en
ce
.
2017;
8(1):
2
68
-
272.
[8]
Moradi
-
Dalva
nd
M.,
B.
Mohammadi
-
Iva
tl
oo
,
A.
Naja
fi,
A.
Rabiee.
Cont
inuous
quic
k
group
sea
rch
opti
m
iz
er
fo
r
solving
non
-
con
vex
e
conomic
d
i
spatc
h
prob
le
m
s.
Elec
tr
ic Pow
er S
y
stems
Resea
r
c
h.
2012
,
Vol
93,
pp.
93
–
105
.
[9]
L.
Kir
chma
y
er
,
Ec
onom
ic
Op
erati
on
of
Pow
er
S
ystems
.
New Delhi:
W
ile
y
Ea
ster
n
Li
m
ited,
first
e
d.
,
1979
.
[10]
A.
J.
W
ood
and
B.
F.
W
oll
enber
g,
Pow
er
Gen
erati
on
Oper
at
ion
a
nd
Control.
New
York:
John
W
ile
y
&
Sons
,
Inc
.
,
sec
ond
ed
.
,
2006
.
[11]
G.
F.
Rei
d
and
L.
Hasdorff,
“
Ec
o
nom
ic
dispa
tc
h
using
quadr
at
i
c
progra
m
m
ing,
”
IE
EE
Tr
an
sac
ti
ons
on
Pow
er
Appara
tus a
nd
S
y
stems
,
vo
l. PAS
-
92,
no.
6,
pp.
2015
–
2023,
Nov 1973.
[12]
K.
Aoki
and
T
.
Satoh,
“
New alg
orit
hm
s for
c
la
ss
ic
ec
onom
ic l
oa
d
dispatch,
”
IE
E
E
Tr
ansa
c
ti
ons o
n
Pow
er
Apparat
u
s
and
S
y
st
ems
,
vol
.
PA
S
-
103,
no.
6
,
pp
.
1423
–
1431
,
June
1984
[13]
J.
K.
Delson
and
S.
M.
Shahide
hpour,
“
Li
nea
r
pr
ogra
m
m
ing
appl
ic
a
ti
ons
to
power
sy
st
em
ec
ono
m
ic
s,
pla
nning
a
nd
oper
ations,” IE
E
E
Tr
ansa
c
ti
ons o
n
Pow
er
S
y
st
ems
,
vol. 7, no. 3, p
p.
1155
–
1163
,
A
ug
1992
[14]
S.
Subram
ani
an
and
S.
Gane
sa
n,
“
A
sim
ple
appr
oac
h
for
em
ission
constra
in
ed
ec
onom
ic
d
i
spatc
h
probl
ems
,
”
Inte
rna
ti
ona
l
Journal
of
Com
puter
Applic
ations,
v
ol.
8,
no
.
11,
pp.
39
–
45,
Octobe
r
2010,
publi
shed
B
y
Foundati
on
of
Com
pute
r
Sci
enc
e
.
[15]
D.
D.
Obiom
a
and
A.
M.
Izu
c
hukwu,
“
Com
par
at
iv
e
an
aly
sis
of
technique
s
fo
r
ec
onom
ic
disp
at
ch
of
gen
erated
power
with
m
o
difi
ed
la
m
bda
-
i
t
era
t
ion
m
et
hod
,
”
in
Proce
edi
ng
s
of
the
2013
I
EE
E
Inte
rn
at
io
n
al
Conf
ere
n
ce
o
n
Emerging
Sus
ta
i
nabl
e
Technol
og
ie
s for
Pow
er
IC
T
in
a
Deve
lo
pin
g
Societ
y
(NIG
E
RCON
),
Nov 20
13,
pp
.
231
–
237
[16]
S.
K.
Mishra
an
d
S.
K.
Mishra,
“
A
compara
ti
v
e
stud
y
of
solut
ion
of
ec
onom
ic
loa
d
dispa
tc
h
p
roble
m
in
powe
r
s
y
stems
in the
e
nvironmenta
l
pe
rspec
ti
v
e,”
Proc
edi
a
Com
pute
r
S
ci
en
ce,
vol
.
48
,
pp.
96
–
100,
201
5.
[17]
Z.
L.
Gaing
,
“
Parti
cle
sw
arm
opt
imiza
ti
on
to
sol
ving
the
ec
onom
ic
dispat
ch
consi
der
ing
the
gen
er
at
or
constraints,”
IEE
E
Tr
ansa
ctio
ns on
Pow
er
S
y
s
te
m
s,
vol
.
18
,
pp
.
1187
–
1195
,
20
03.
[18]
H.
Faris,
A
.
Sheta,
and
E
.
¨
Ozne
rgiz,
“
Modell
ing
hot
ro
lling
m
anuf
acturi
ng
proc
ess
usi
ng
soft
comput
ing
te
chn
ique
s,”
Int
e
rna
ti
on
al
Journ
al of
Com
pute
r
Int
egr
ated
Manuf
acturi
ng,
vol
.
26
,
n
o.
8
,
pp
.
762
–
77
1,
2013
[19]
A.
F.
Sheta
,
H.
Faris,
and
E.
¨
O
zne
rgi
z,
“
Im
proving
produc
ti
on
qual
ity
of
a
hot
-
r
oll
ing
i
ndustri
al
proc
ess
via
genetic
progra
m
m
ing
mode
l,”
In
t. J.
Co
m
put.
Appl.
Tec
hnol.
,
vol
.
49
,
no
.
3/4
,
pp
.
239
–
25
0,
Jun.
2014.
[20]
H.
Faris
and
A.
F.
Sheta
,
“
A
compari
son
bet
wee
n
par
ametr
i
c
and
non
-
par
ame
tri
c
soft
comput
ing
appr
oa
che
s
t
o
m
odel
the
t
empe
rat
ur
e
of
a
m
et
a
l
cut
ti
ng
tool,”
In
te
rna
ti
ona
l
Journal
of
Com
pute
r
I
nte
gra
te
d
Manuf
ac
tur
ing,
vol
.
29
,
no.
1
,
pp
.
64
–
75
,
2016.
[21]
M.
Brai
k
,
A.
S
het
a
,
and
A
.
Ay
esh
,
“
Parti
c
le
sw
arm
opti
m
iz
at
ion
enh
ancem
ent
appr
oa
ch
for
improving
imag
e
qual
ity
,
”
Int
.
J.
I
nnov.
C
om
put.
Appl.
,
vol. 1, no. 2, pp. 138
–
145,
Jan.
2007
.
[22]
B.
Solai
m
an
and
A.
Sheta,
“
Evolving
a
h
y
brid
K
Mea
ns
cl
uster
in
g
al
gori
thm
for
wire
le
ss
sensor
net
work
using
PS
O
and
GA
,
”
Int
ern
at
ion
al
Journa
l
o
f
Com
pute
r
Sci
e
nce
Iss
ues,
vol
.
12,
no
.
1
,
pp
.
23
–
32,
2
015
.
[23]
J.
Sasikal
a
and
R.
M,
“
Optimal
_
base
d
e
cono
m
ic
emiss
ion
dispat
ch
using
si
m
ula
te
d
ann
eali
ng,
”
Int
ern
a
ti
on
al
Journal
of
Com
p
ute
r
Applicati
on
s,
vol.
1,
no.
10
,
pp.
55
–
63,
Febr
uar
y
2010
,
publis
hed
B
y
Founda
ti
on
of
Com
pute
r
Scie
nc
e.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Resolutio
n of e
cono
mic d
is
patc
h pr
ob
le
m
of the mor
occ
an
ne
tw
or
k
…
(
Rac
hid
H
abac
hi
)
353
[24]
H.
Hardi
ans
y
ah,
“
Artifi
cial
be
e
col
on
y
al
gori
th
m
for
ec
onom
ic
loa
d
d
ispat
ch
pr
oble
m
,
”
IAES
I
nte
rna
ti
ona
l
Jour
nal
of
Artificial
In
telli
gen
ce,
vo
l. 2,
no.
2
,
pp
.
90
–
98
,
2013
[25]
W
.
M.
Al
y
and
A.
Shet
a,
“
Para
m
et
er
esti
m
ati
on
of
nonl
ine
ar
s
y
stems
using
l`e
v
y
fli
gh
t
c
uckoo
sea
r
ch,”
in
Proce
edi
ngs
of
the
Th
irty
-
th
ird
SG
AI
Inte
rna
tional
Confer
ence
on
Innova
ti
v
e
Te
chn
ique
s
and
Applic
a
ti
ons
o
f
Artifi
c
ia
l
Int
el
l
ig
enc
e
,
C
ambridge
,
Eng
l
and
,
UK
,
2
013,
pp
.
443
–
44
9
[26]
“
Com
par
at
ive
a
naly
s
is
of
opti
m
al
lo
ad
dispa
tc
h
through
evol
u
ti
o
nar
y
al
gor
it
hm
s,
”
Ain
Sham
s
Engi
nee
r
ing
Journa
l,
vol.
6
,
no
.
1
,
pp
.
107
–
120,
2015
[27]
Ajeni
koko,
G.A.
,
&
Olabode,
O.E
.
“
Optimal
Pow
er
Flow
with
Rea
c
ti
ve
Pow
er
C
om
pensa
ti
on
for
Cost
and
Los
s
Minim
iz
at
ion
o
n
Niger
ia
n
Pow
er
Grid
Sy
st
em”.
Indone
sian
J
ourna
l
of
El
e
ct
r
ic
a
l
Engi
ne
eri
ng
anda
Inform
at
i
cs
(IJEEI
). 2017;
5(
3):
236
-
247.
[28]
Jiang
A,
Ertem
S,
Subir
S,
Koth
ari
D
(1995)
E
co
nom
ic
dispatch
with
non
-
m
onotoni
call
y
inc
re
asi
ng
inc
r
ementa
l
c
ost
unit
s a
nd
tr
ansmi
ss
ion
s
y
stem
losses.
Discuss
ion.
IEE
E
Tr
ans
Pow
er
S
y
s
t
10:891
–
8
97
[29]
A.
As
kar
za
d
eh,
“
A
novel
m
et
ah
eur
isti
c
m
et
hod
for
solving
constra
in
ed
eng
ineer
ing
opti
m
ization
proble
m
s:
cro
w
sea
rch
al
gor
it
hm
,
”
Com
puters
&
Struct
ure
s,
vol
.
169,
pp
.
1
–
12
,
2
016
[30]
A.
As
kar
za
deh
,
“
El
ec
tr
ic
a
l
po
wer
gene
ra
ti
on
b
y
an
opt
imis
e
d
au
-
tonomous
pv/wind/
ti
d
al
/b
a
tt
er
y
s
y
s
te
m
,
”
I
ET
Rene
wabl
e
Pow
er
Gen
er
-
ation,
2
016.
[31]
D.
Oliva,
S.
Hi
nojosa,
E
.
Cu
ev
as,
G.
Pa
ja
res
,
O.
Avalos,
and
J.
Ga
lvez,
´
“
Cross
ent
rop
y
base
d
thre
sholding
f
or
m
agne
ti
c
r
esona
nce
bra
in
images
using c
row sea
rch
al
gor
it
hm
,
”
E
xper
t
S
y
st
ems
w
it
h
Applicati
ons
,
vol.
79,
pp
.
164
–
180,
2017
[32]
A.
L.
Devi
and
O.
V.
Krishna
,
“
Com
bine
d
ec
on
om
ic
and
emiss
ion
dispat
ch
usin
g
evol
ut
iona
r
y
a
lgori
thms
:
A
c
ase
stud
y
,
”
ARP
N Journal
of
Eng
ineeri
ng
and
Appl
ied Sci
en
ce
s,
vol
.
3,
no
.
6
,
pp
.
28
–
35,
2008
[33]
GA
ING
,
Z.
L.
Parti
cle
Sw
arm
Optimiza
ti
on
to
Solving
th
e
Ec
onom
ic
Di
spatc
h
Consider
ing
the
Gene
r
a
tor
Constrai
nts.
IEEE
tr
ansa
ctions o
n
po
wer
s
y
s
te
m
s,
2003
,
vol
.
18
(
3),
pp
.
1187
-
119
5
[34]
R
.
Haba
chi
et
a
l
.
Ea
g
le
Strateg
y
Based
Crow
Sear
ch
Algorit
hm
for
Solving
Unit
Comm
it
m
ent
P
roble
m
in
Sm
art
Grid
S
y
stem
.
In
donesia
n
Journa
l
of
El
e
ct
r
ic
a
l
Engi
ne
eri
ng
and
Com
pute
r
Sci
e
nce
Vol
.
12
,
N
o.
1,
,
p
p.
17~2
9
Octobe
r
2018
Evaluation Warning : The document was created with Spire.PDF for Python.