Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
9
, No
.
5
,
Octo
ber
201
9
, pp.
3384
~3
390
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ijece
.
v9
i
5
.
pp3384
-
33
90
3384
Journ
al h
om
e
page
:
http:
//
ia
es
core
.c
om/
journa
ls
/i
ndex.
ph
p/IJECE
Eco
n
omi
c and em
ission di
spatc
h u
sing cuc
koo sea
rch al
gorith
m
R
ac
hid
H
abac
hi
1
,
Ach
ra
f
T
oui
l
2
,
Ab
d
el
lah
Boulal
3
,
Abde
lkabir
Chark
aoui
4
, A
bdelw
ahed
Ec
hc
hatbi
5
La
bora
tor
y
of
Mec
han
ic
a
l Engi
n
ee
ring
,
Industr
ial
Mana
g
ement a
nd
Innova
t
ion
,
The
Fa
cul
t
y
of
S
ci
en
ce
s
and
T
echnolog
y
,
Hass
an
1st Uni
ver
si
t
y
,
Morocc
o
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Feb
14
, 201
9
Re
vised
A
pr
4
,
201
9
Accepte
d
Apr
4
, 2
01
9
The
ec
onom
ic
d
ispat
ch
probl
em
of
power
play
s
a
ver
y
importan
t
role
in
the
expl
oitati
on
of
el
e
ct
ro
-
e
n
erg
y
s
y
stems
to
j
udic
iousl
y
distr
ibut
e
power
gene
ra
te
d
b
y
all
pla
nts
.
Thi
s
p
ape
r
proposes
t
he
use
of
Cuck
oo
Sear
ch
Algorit
hm
(CSA
)
for
solving
the
e
conomic
a
nd
Emiss
ion
dispat
ch.
The
eff
ective
n
ess
of
the
proposed
appr
oac
h
has
b
ee
n
t
este
d
on
3
gene
r
at
or
sy
stem.CSA
is
a
new
m
et
a
-
heur
i
stic
opti
m
izat
ion
m
et
hod
inspired
from
the
obli
gate
brood
par
asitism
of
so
me
cuc
koo
spec
ie
s
b
y
lay
i
ng
the
ir
eggs
in
the
nests
of
othe
r
ho
st
birds
of
othe
r
spec
ie
s
.
The
r
esult
s
show
s
tha
t
pe
rform
ance
of
the
propose
d
appr
oa
ch
r
ev
ea
l
the
eff
ic
i
en
tly
and
robust
ness
when
compare
d
r
esult
s
of
oth
er
op
ti
m
izati
on
al
gor
it
hm
s
rep
orte
d
in
liter
a
ture
.
Ke
yw
or
d
s
:
Cuck
oo
s
ea
rch
a
lgorit
hm
Eco
no
m
ic
d
isp
at
ch
p
r
ob
le
m
Em
issi
on
c
os
t
Fu
el
c
os
t
S
m
art g
rid
Copyright
©
201
9
Instit
ute of
Ad
v
ance
d
Engi
n
e
eri
ng
and
Sc
ie
n
ce
.
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
ng
i
neer
i
ng,
I
nd
us
tria
l M
ana
ge
m
ent an
d I
nnovat
ion,
Faculty
of S
ci
e
nces a
nd Tec
hnol
og
y,
Hassa
n 1st Uni
ver
sit
y
,
PO
B
ox 57
7,
S
et
ta
t, Moro
c
co
.
Em
a
il
:
hab
achi
dep
a
rtem
entgeg
m
@g
m
a
il
.co
m
1.
INTROD
U
CTION
Sm
art
gr
ids
are
a
set
of
te
chnolo
gies,
c
oncepts
a
nd
ap
proac
hes,
al
lo
wing
the
inte
gr
at
io
n
th
e
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
chno
lo
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
buti
on
to
reduce
energy
co
n
s
um
pt
ion
,
reduce
costs,
i
ncr
eas
e
reli
abili
ty
and
al
s
o
increase
tra
nsp
aren
cy
in
t
he
netw
ork.
T
he
r
efore,
the
syst
e
m
intel
li
gen
t
will
hav
e
a
sig
nificant
im
pact
in
the
fiel
ds
of
fina
nc
e
an
d
ec
onom
ic
s
of
the
po
wer
i
ndus
try
[
2].
Alth
ough,
The
tra
diti
on
al
netw
ork
is
a
on
e
-
way
netw
ork
i
n
w
hich
t
he
el
ect
rical
energy
pro
duced
in
po
wer
pla
nts
is
cha
nn
el
e
d
to
co
ns
um
ers
w
it
ho
ut
inf
or
m
at
ion
to c
reate an
au
t
om
at
ed
and d
ist
rib
uted netw
ork of ad
va
nced
powe
r
s
upplies
.
ED
P
is
al
so
ap
plied
in
the
integrate
d
syst
e
m
fo
r
sc
heduli
ng
po
w
er
pla
nts.
A
few
m
et
ho
ds
hav
e
bee
n
publishe
d
to
s
olv
e
t
he
E
D
pro
blem
and
O
pti
m
al
Po
we
r
Flow
(
OP
F
).
Re
searche
rs
ha
ve
publishe
d
a
fe
w
m
et
ho
ds
to
s
ol
ve
ED
a
nd
O
PF
pro
blem
s.
Direct
m
e
tho
d
is
accurate
and
ver
y
sim
pl
e
bu
t
lim
it
ed
by
th
e
qu
a
drat
ic
obj
e
ct
ive
functi
on
[3
]
.T
he
eco
no
m
ic
disp
at
ch
(
ED)
is
one
of
the
powe
r
m
anag
em
ent
too
ls
that
is
us
e
d
to
deter
m
ine
real
pow
er
outp
ut
of
t
her
m
al
gen
erat
ing
unit
s
to
m
eet
req
ui
red
lo
ad
dem
and
.
T
he
ED
resu
lt
s
in
m
ini
m
u
m
fu
el
gen
e
rati
on
c
os
t,
m
i
nim
u
m
transm
i
ssion
powe
r
l
oss
wh
il
e
sat
isf
yi
ng
al
l
un
it
s,
as
well
as syst
em
con
strai
nts [4
-
5].
The
de
m
and
f
or
el
ect
rici
ty
i
s
increasi
ng
in
a
la
rg
e
fact
or
in
tod
ay
‟s
li
f
e,
w
hich
m
akes
it
hig
hly
cru
ci
al
to
r
un
gen
e
rato
rs
at
ver
y
m
ini
m
al
cost.
T
his
is
t
he
m
ai
n
factor
o
f
an
Eco
no
m
ic
di
sp
at
ch
pro
blem
.
W
it
h
the
une
xc
eptional
producti
on
of
ca
rbon
em
issi
on
s
in
therm
al
po
w
er
plant,
it
s
ne
eded
t
o
optim
i
ze
the
e
m
issi
on
to
get
her
with
the
optim
iz
at
ion
of
cost
wh
ic
h
act
s
as
t
wo
vital
par
ts
of
Ec
onom
ic
disp
at
ch
pro
blem
.
The
eco
no
m
ic
disp
at
c
h
so
lut
i
on
prov
i
des
th
e
best
m
ini
m
u
m
cost
of
fu
el
and
em
issi
on
.
This
ind
ire
ct
ly
m
akes
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
Eco
nomic
and emissi
on
dispa
tc
h
usi
ng c
uck
oo se
ar
ch
a
l
gorit
hm
(R
ach
i
dHa
bachi)
3385
lowe
r
cost
for
el
ect
rici
ty
and
m
akes
el
ect
rical
util
it
ie
s
m
or
e
com
petit
ive
i
n
the
m
ark
et
.
As
the
e
nergy
cannot
be
store
d,
it
re
qu
i
res
highly
eff
ic
ie
nt
est
im
a
t
ion
sce
nar
i
os
inclu
ding
tran
s
m
issi
on
an
d
di
stribu
ti
on
syst
e
m
s
to
m
ake th
e sam
e wor
k
e
ff
ect
ive
ly
.
Var
i
ou
s
te
ch
nolo
gies
hav
e
be
en
int
rod
uced
to
s
olv
e
the
optim
iz
at
ion
of
Eco
no
m
ic
Loa
d
Disp
at
c
h
pro
blem
s.
The
sel
ect
ion
of
the
optim
iz
at
i
on
al
gorithm
is
the
i
m
po
rta
nt
pa
rt
of
t
he
p
r
ob
le
m
involvin
g
econom
ic
dispa
tc
h.
T
he
E
DP
is
de
velo
ped
base
d
on
real
-
value
d
c
od
i
ficat
ion
.
In
m
od
ern
m
et
ho
dol
ogy
on
ly
the cost
fun
ct
i
on is eval
uated
and a
global
m
ini
m
u
m
so
luti
on
is c
om
pu
te
d,
i
nd
e
pe
nd
e
nt
ly
o
f
th
e c
os
t f
un
ct
io
n.
The
us
e
of
di
gi
ta
l
co
m
pu
te
rs
for
obta
inin
g
loadi
ng
sc
hedul
es
wer
e
in
vestigat
ed
an
d
us
e
d
tod
ay
.
T
he
E
D
is
a
sta
ti
c
pr
ob
le
m
is
to
say
we
m
us
t
def
i
ne
at
a
giv
e
n
powe
rs
gen
e
rated
by
each
powe
r
pla
nt
to
powe
r
a
load
as
econom
ic
al
l
y as p
os
sible
. T
o solve
this
pro
bl
e
m
the o
ptim
i
zat
ion
m
et
ho
ds are
us
e
d.
Conve
nti
on
al
op
ti
m
iz
ation
techn
i
qu
e
s [
6
,
7
]
. h
ave lon
g
bee
n
ap
plied to
s
ol
ve
the ED
pro
blem
su
ch
as
Qu
a
drat
ic
Pr
ogram
m
ing
[8
,
9
]
.
li
near
pro
gr
am
m
ing
[
10
]
sequ
e
ntial
appr
oach
with
a
m
at
rix
fr
am
ewor
k
(S
AM
F)
,
[
11
]
.
m
od
ifie
d
Lam
bd
a
-
it
erati
on
m
et
ho
d
[1
2
]
,
New
t
on
Ra
phs
on
a
nd
Lag
ra
ngia
n
m
ulti
plier
(LM)
al
gorithm
s
[1
3
]
,
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
so
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.
Ma
ny
determ
i
nisti
c
op
ti
m
iz
at
ion
ap
proac
he
s
wer
e
propos
ed
to
so
l
ve
th
e
ELD
pro
ble
m
,
including
lam
bd
a
it
erat
io
n
m
et
ho
d
[
1
1
]
,
gradie
n
t
m
et
h
od,
li
nea
r
pro
gram
m
ing
[
12
]
,
non
-
li
near
pro
g
ram
m
ing
,
dy
nam
ic
pro
gr
am
m
ing
[
13
]
a
nd
quad
ra
ti
c
pr
ogram
m
i
ng
[1
4
]
.
But
t
he
se
m
e
tho
ds
re
qu
i
re
en
or
m
ous
effor
ts
i
n
te
r
m
s
of
com
pu
ta
ti
on
.
Du
e
to
c
om
plexiti
es
of
com
pu
ti
ng,
there
f
ore
eff
i
ci
ent
al
gorithm
to
find
op
ti
m
al
so
l
ution
li
ke
gen
et
ic
al
gorithm
[1
5
,
1
6
]
,
pa
rtic
le
swar
m
op
ti
m
iz
ation
[
17
]
,
e
vo
l
ution
a
ry
program
m
ing
,
arti
f
ic
ia
l
bee
colon
y
op
ti
m
iz
ation
[
1
8
,
19
]
,
an
d
bi
og
e
ogra
phy
ba
sed
optim
iz
at
i
on
;
bacteria
l
f
or
a
ging
a
nd
al
so
t
heir
va
rian
t
s
cam
e
into
i
m
ple
m
ent.
Bi
o
-
insp
i
red
m
et
a
-
heu
risti
c
al
go
rithm
s
hav
e
rece
ntly
sh
own
the
ef
fici
ency
in
deali
ng
with
m
any n
onli
nea
r op
ti
m
iz
a
ti
on
s
constraine
d p
r
ob
le
m
s f
or f
i
ndin
g
the
opti
m
al
so
luti
on
.
The
rem
ai
nin
g
organ
iz
at
io
n
of
this
pa
per
is
as
fo
ll
ows.
Se
ct
ion
2
pr
e
sent
s
the
pro
blem
form
ulati
on
of
t
he
E
D
P
.
H
and
li
ng
of
c
onstrai
nts
a
nd
im
plem
entat
ion
of
the
pro
pose
d
CSA
t
o
E
D
prob
le
m
are
ad
dressed
in
Sect
.
3.
Sec
ti
on
4
re
ports
r
esults
of
the
pr
opos
e
d
CSA
m
et
ho
d.
A
nu
m
ber
of
case
stud
ie
s
us
i
ng
st
and
a
r
d
te
st
syst
e
m
s
are
us
e
d
t
o
te
st
t
he
pro
po
s
ed
m
et
hod.
The
c
om
par
ison
s
of
r
esults
betwee
n
the
pro
posed
m
et
ho
d
and
e
xisti
ng
m
et
ho
ds
are
al
so
car
ried
out
in
this
sect
ion
.
The
discu
ssio
n
is
fo
ll
owed
in
Sect
.
5.
A
ft
er
al
l,
the concl
usi
on
is give
n.
2.
PROBLE
M
FOR
M
ULAT
I
ON
The
ge
ne
rati
ng
un
it
s
are
loa
de
d
eco
nom
ic
a
lly
su
ch
a
way
t
o
re
du
ce
t
he
operati
ng
co
st.
Con
si
der
i
ng
the v
al
ve po
i
nt
eff
ect
t
he
ec
on
om
ic
d
ispatc
h form
ulate
d
the
obj
ect
ive
fun
ct
ion
a
s
giv
e
n be
low
(
)
=
+
+
2
(1)
w
he
re
a
i
,b
i ,ci a
re t
he f
uel cost
co
e
ff
ic
ie
nts of gene
ra
tor
i
P
i
is t
he p
ow
e
r g
ener
at
e
d by un
it
I
i ,M
W
F
i
(
P
i
)
is t
he fuel c
os
t
fun
ct
io
n of u
ni
t i
The o
bj
ect
ive
fun
ct
io
n o
f
the
ED pr
oble
m
is
to m
ini
m
iz
e the total
prod
uction co
st,
wh
ic
h be
wr
it
te
n
as:
(
)
=
∑
(
)
=
1
,
2
,
…
.
,
=
1
(2)
The ne
w objec
ti
ve
f
un
ct
io
n b
y consi
der
i
ng
valve p
oin
t l
oa
ding alo
ng
with total
fu
el
c
ost
b
ecom
es,
(
)
=
∑
+
+
2
+
(
sin
(
(
−
)
)
)
=
1
(3)
w
he
re
,b
i
,c
i
d
i
a
nd e
i
are the
fuel c
ost
co
ef
fici
ents
of
gen
e
rato
r
i
is t
he power
g
e
ner
at
e
d by unit
i ,
M
W
is t
he
m
ini
m
u
m
g
ener
arti
on l
i
m
i
t of
unit
i
,M
W
(
)
is t
he
total
fu
el
co
st
$/hr
The
so
l
ution
of
ED
P
ca
n
be
highly
i
m
pr
oved
by
introd
uc
ing
hi
gh
e
r
ord
er
generato
r
co
st
fu
nctio
ns.
Cub
ic
cost
f
unct
ion
disp
la
ys
the
act
ual
respon
s
e
of
the
rm
al
gen
erat
or
sm
or
e
accu
ratel
y.The
cu
bic
fu
el
cost
functi
on
of
a
th
erm
al
g
ener
at
ing u
nit i
s r
e
pre
sented
as
fo
ll
ows
[
2
0
]:
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.
9
, N
o.
5
,
Oct
ober
20
19
:
3
3
8
4
-
3
3
9
0
3386
(
)
=
+
+
2
+
3
(4)
In
orde
r
to
m
ini
m
iz
e
the
poll
utants,
e
m
issi
on
is
co
ns
ie
re
d
al
on
g
with
ec
onom
ic
disp
at
ch
.
The
g
e
ner
at
or
can
be
m
od
el
le
d
as
ha
ving
a
qu
a
drat
ic
relat
i
on
betwee
n
th
e
a
m
ou
nt
of
poll
utants
releas
ed
an
d
the po
wer ge
ne
rated.T
he
m
at
he
m
at
ic
al
f
or
m
ulati
on
for ge
ne
rator is
giv
e
n b
y,
(
)
=
+
+
2
(5)
w
he
re
α
i
,
β
i
,
γ
i
are
the em
i
ssi
on co
e
ff
ic
ie
nts of gene
rato
r
i
P
i
is t
he power
g
e
ner
at
e
d by unit
i ,
M
W
Ei
(
Pi
)
is t
he
f
uel c
os
t
functi
on
of
unit
i
The
t
otal em
issi
on
f
or
t
he
e
ntire syst
em
o
f
N
gen
e
rato
rs
ca
n t
hen
be
cal
c
ula
te
d
as,
(
)
=
∑
(
)
=
∑
+
+
2
=
1
=
1
(6)
The ne
w
Em
is
sion f
un
ct
io
n b
ecom
es
,
E
T
(
P
T
)
=
∑
E
i
(
P
i
)
=
∑
α
i
+
β
i
P
i
+
γ
i
P
i
2
+
ε
i
exp
(
δ
i
Pi
)
N
i
=
1
N
i
=
1
(7)
w
he
re
α
i
,
β
i
,
γ
i
,
ε
i
,
δ
i
,are
t
he
em
is
sion coe
ff
ic
ie
nt
s o
f
g
e
ne
rator
i
P
i
is t
he p
ow
e
r g
ener
at
e
d by un
it
i ,
M
W
Ei
(
Pi
)
is t
he fuel c
os
t
fun
ct
io
n of u
ni
t i
E
T
(
P
T
)
is
t
he
total
em
issio
n , t
on
/
hr.
N
is t
he nu
m
ber
of
ge
ner
at
in
g un
it
s
su
bject
t
o
r
eal
powe
r
balanc
e
eq
uatio
n
.
T
he
total
act
ive
powe
r
ou
t
pu
t
of
ge
ner
at
i
ng
un
it
s
m
us
t
be
equ
al
t
o
total
p
owe
r
loa
d dem
and
p
l
us
powe
r
loss:
∑
=
=
1
+
(8)
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
+
(9)
Gen
e
rato
r
ca
pa
ci
ty
lim
it
s Th
e act
ive pow
e
r o
utput o
f ge
ner
a
ti
ng
un
it
s m
us
t be
within
the
al
lowed li
m
it
s:
.
≤
≤
.
(10)
3.
CUCK
OO
SE
ARCH
A
LG
O
RITH
M (CSA
)
Cuck
oo
searc
h
(CS)
is
ins
pir
ed
by
s
om
e
sp
eci
es
of
a
bi
rd
fam
il
y
ca
ll
ed
cucko
o
beca
use
of
thei
r
sp
eci
al
li
festy
l
e
andag
gr
e
ssiv
e
reprod
uctio
n
strat
egy.
T
his
al
gorithm
was
pro
posed
by
Yang
a
nd
De
b
[2
1
]
.
The
CS
is
an
optim
iz
at
ion
al
go
rithm
based
on
the
bro
od
pa
rasit
is
m
of
cuck
oo
sp
e
ci
es
by
la
yi
ng
their
eg
gs
in
the
com
m
un
al
nests
ofothe
r
ho
st
bir
ds
,
th
ough
they
m
a
y
re
m
ov
e
othe
rs’
eg
gs
to
i
nc
rease
the
ha
tc
hin
g
pro
bab
il
it
y
of
their
own
e
ggs.
So
m
e
ho
st
bird
s
do
not
beh
a
ve
fr
ie
nd
ly
aga
inst
intruders
a
nd
e
ng
a
ge
in
di
rect
confli
ct
with
them
.
If
a
ho
st
bir
d
disc
ov
e
rs
the
eg
gs
are
nott
heir
ow
n,
it
will
ei
ther
throw
these
f
or
ei
gn
eg
gs
away o
r
sim
ply a
band
on it
s nest
and
bu
il
d a
new n
e
st el
sew
her
e
[2
2
].
The
Cuc
koo
se
arch
al
go
rithm
con
ta
ins
a
popula
ti
on
of
nests
or
e
gg
s
.
Each
egg
in
a
nest
r
epr
ese
nts
a
so
luti
on
a
nd
a
cucko
o
e
gg
re
pr
ese
nts
a
ne
w
so
luti
on.
I
f
th
e
cuc
koo
e
gg
is
ve
ry
sim
il
ar
t
o
the
host’s
,
t
he
n
thi
s
cucko
o
eg
g
is
le
ss
li
kely
to
be
disco
ver
e
d;
thu
s
,
the
fit
nes
s
shou
l
d
be
relat
ed
to
the
difference
i
n
so
l
ution
s
.
The
bette
r
ne
w
so
luti
on
(c
uc
koo)
is
re
placed
with
a
so
l
utio
n
w
hich
is
no
t
so
go
od
i
n
the
nest.
I
n
the
sim
plest
form
,
each
ne
st
has
one
e
gg.
When
ge
ne
rati
ng
ne
w
s
ol
utions
f
or
x
(
t+1
)
,
say
cucko
o
i
,
a
Lé
vy
fli
gh
t
is
perform
ed:
+
1
=
+
⊕
(
)
(11)
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
Eco
nomic
and emissi
on
dispa
tc
h
usi
ng c
uck
oo se
ar
ch
a
l
gorit
hm
(R
ach
i
dHa
bachi)
3387
w
he
re
α
>
0
is
the
ste
p
siz
e
w
hich
sho
uld
be
relat
ed
t
o
the
scal
es
of
the
pro
blem
of
inte
rest.
In
m
os
t
cases,
wecan
us
e
α
=
O
(
1).
T
he
pro
du
ct
⊕
m
eans
entry
-
wise
m
ulti
plica
ti
on
s.
L
évy
flig
hts
ess
entia
ll
y
pr
ovid
e
a
rand
om
walk
w
hile
Levy
(
λ
)
=
u
=
t
−
λ
,
1
≤
λ
≤
3
(12)
rand
om
walk
proces
s
wh
ic
h
obey
s
a
powe
r
-
l
awstep
-
le
ngth
distrib
ution
with
a
hea
vy
ta
il
.
The
r
ules
for
CSA
are
des
cri
bed a
s foll
ow
s:
-
Each c
uc
koo
la
ys one e
gg at a
tim
e, an
d d
umps it i
n ara
ndom
ly
ch
os
en
n
e
st;
-
The best
nests
with
high
qu
al
i
ty
o
f
e
ggs
(so
l
utions)
will
carry
o
ve
r
t
o
the
next ge
ne
rati
ons
;
-
The
num
ber
of
avail
able
host
nests
is
fixe
d,
an
d
a
host
can d
isc
over
a
f
ore
ign
e
gg
w
it
h
a
pro
bab
il
it
y
paε[
0,
1].
I
n
this
case
,
the
host
bir
d
can
ei
ther
th
r
owthe
e
gg
a
way
or
a
band
on
th
e
nest
so
as
to
bu
il
d
ac
om
pletely
new n
e
st i
n
a
new locati
on.
The
la
te
r
as
sum
pt
ion
ca
n
be
ap
prox
im
at
ed
by
the
fr
act
io
n
pa
of
the
n
ne
sts
w
hich
a
re
rep
la
ce
d
by
new
ones
(w
it
h
ne
w
ra
ndom
so
luti
ons)
.
W
it
h
these
th
ree
ru
l
es,
the
basic
ste
ps
of
the
CS
ca
n
be
s
umm
arized
as
the pseu
do
-
c
odesh
ownbel
low,
1)
Def
i
ne
the
ob
je
ct
ive fun
ct
io
n
Td
f
(x),
x=
(x
1
,x
2
,x
3
,…..,
x
d
)
T
2)
Set
n
,
pa
, a
nd
Ma
xG
e
ner
at
io
n param
et
ers
3)
Gen
e
r
at
e init
ia
l pop
ulati
on of
n
a
vaila
ble
nes
ts
4)
Mov
e
a c
ucko
o (
i
) ran
dom
l
y by Lé
vy f
li
ghts
5)
Evaluate t
he fit
ness
fi
6)
Ra
ndom
ly
ch
oo
se a
n
est
(
j
)
a
m
on
g
n
avail
a
ble n
e
sts
7)
If
fi
>
f
j
the
r
e
pl
ace
j
by th
e
ne
w
s
olu
ti
on
8)
Ab
a
ndon a f
ra
ct
ion
pa
of wor
se n
est
s and cre
at
e the
sa
m
e f
racti
on
of
new
nests at new
lo
cat
ion
v
ia
Lévy
fligh
ts
9)
Keep t
he best s
olu
ti
ons
(or ne
sts wit
h q
ualit
y
so
l
ution
s
)
10)
So
rt
the s
olu
ti
ons a
nd f
i
nd the
b
est
c
urre
nt sol
ution
11)
If
st
oppi
ng crite
ria is n
ot sati
s
fied, i
ncr
e
ase
ge
ner
at
io
n n
umber
and
go to
s
te
p
4
12)
Po
st
-
proces
s r
e
su
lt
s and
fin
d
t
he best s
olu
ti
on am
on
g al
l.
The pse
udo
c
ode
of CSA
for
EDP
is
sho
wn
in Figu
re
1
.
Alg
o
rith
m
1
C
uc
k
o
o
Search
Algo
rith
m
via
levy
f
lig
ht
a
lg
o
rith
m
1
:
Beg
in
2
:
Ob
jectiv
e f
u
n
ctio
n
f
(x),
x
=(x
1
,x
2
,x
3
,….
.,
x
d
)
T
3
:
Gen
erate
in
itial
po
p
u
latio
n
o
f
n h
o
st
n
ests
xi( i=1,2
,3,
……,n)
4
:
While
(
t<M
a
x G
e
n
era
tio
n
)
o
r
(
sto
p
criter
io
n
)
do
5
:
beg
in
6
:
Get a cu
ck
o
o
r
an
d
o
m
l
y
b
y
lev
y
f
lig
h
t
7
:
Evalu
ate its qu
alit
y
/
f
itn
ess
Fi
8
:
Ch
o
o
se a nest a
m
o
n
g
n (say
,j
)
rand
o
m
l
y
9
:
If
(F
i
>Fj
)
Repla
ce
j by th
e new
so
lu
tio
n
;
10
:
A f
raction
(
p
n
)
o
f
worse nes
ts are
ab
an
d
o
n
ed
and
new
o
n
es are
bu
ilt;
11
:
Keep
the b
est so
lu
tio
n
s (
o
r n
ests
with
q
u
a
lity so
lu
tio
n
s );
12
:
Ran
k
the so
lu
tio
n
s
and
f
in
d
the cu
rr
en
t bes
t
13
:
end w
hile
14
:
Po
st p
rocess
r
esu
lts an
d
vis
u
alizatio
n
15
:
End
Figure
1. Pse
udoc
ode
of Cuc
koo
s
earc
h
al
gorithm
f
or E
D
P
4.
RESU
LT
S
A
ND AN
ALYSIS
In
t
his
sect
ion,
we
pr
e
sent
t
he
res
ults
obt
ai
ned
base
d
on
Cuc
koo
Se
arch
Algorit
hm
(CSA
)
f
or
so
lvi
ng
the
ec
onom
ic
and
Em
issi
o
n
disp
at
ch
an
d
com
pare
this
resu
lt
s
with
the
CM
(
Conve
ntion
al
Me
thod)
[2
3
]
an
d
Pa
rtic
le
Sw
arm
Optim
iz
at
ion
[
2
4
]
.
A
t
hr
ee
unit
s
powe
r
unit
syst
e
m
to
exp
l
ore
our
i
dea
on
us
i
ng
CSA
t
o
fi
nd
th
e
opti
m
a
l
set
of
power
gen
e
ra
ti
on
of
the
syst
e
m
.
CSA
will
be
us
e
d
in
this
pa
per
to
so
l
ve
the
econom
ic
an
d Em
issi
on
disp
a
tc
h
.T
he pr
ogra
m
s ar
e d
evelo
pe
d
in
MAT
LA
B 7.9 e
nviro
nm
ent.
The
a
dopted
s
yst
e
m
is
exp
ec
te
d
to
pro
du
ce
dem
and
po
wer
of
40
0
M
W.
The
Ta
ble
1
s
hows
the
c
os
t
coeffic
ie
nt
of
the
three
ge
ne
r
at
or
s,
under
st
ud
y,
wh
i
le
the
m
at
rix
is
the
loss
coeffic
ie
nt
m
a
trix
of
the
three
un
it
s
power
sy
stem
..
Gen
e
rato
r
em
issi
on
coe
ff
ic
ie
nts
f
or
IE
EE
-
30
-
bus
sys
tem
is
pr
ov
i
de
d
in
Ta
ble
2
gi
ven
belo
w.
From
t
he
re
su
lt
s
of
T
able
3,
we
not
ic
e
that
CSA
gi
ve
us
the
sam
e
pro
du
ct
io
n
c
os
t,
a
nd
C
M
gi
ves
a
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.
9
, N
o.
5
,
Oct
ober
20
19
:
3
3
8
4
-
3
3
9
0
3388
sli
gh
tl
y
lowe
r
cost
of
$
0.5
/
h,
CS
A
giv
es
us
a
good
pro
duct
ion
cost
a
nd
good
acc
ur
acy
,.T
hroug
h
the
r
esults
fou
nd,
we
can
say
that
ther
e
is
a
sm
all
diff
e
re
nce
between
the
op
ti
m
al
gen
erate
d
powe
rs,
t
he
cost
of
pro
du
ct
io
n
as
well
as the
t
ransm
issi
on
losse
s
b
et
wee
n
t
he diffe
ren
t al
gorith
m
s.Th
e o
ptim
i
zat
ion
of
pro
du
ct
ion
cost
is
bette
r
wh
e
n
a
pp
ly
in
g
CSA
c
om
par
ed
to
CM
.
T
his
le
ads
us
to
c
om
par
e
the
resul
ts
of
CS
A
wit
h
th
os
e
of
PS
O.
T
he
t
ot
al
cost
gi
ven
by
PS
O
is
20
813
$/h
an
d
a
rou
nd
2081
2.5
$/h
f
or
CSA
,
w
hich
re
pr
es
ents
a
diff
e
re
nce
of
0.5
$/h
.
Th
e
tra
ns
m
issi
on
loss
es
evaluate
d
by
the
three
a
pp
ro
ac
hes
rem
ai
n
ed
ve
ry
cl
os
e
f
or
t
he
three al
gorithm
s r
es
pecti
vely
(PSO
) 7.
5681M
W,
(CM)
7.5
687 M
W,
(
C
SA)
7.568
1
M
W.
In
Fi
gure
2,
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
[
2
5
]
.
T
he
com
par
iso
n
is
rep
re
se
nte
d
by
the
gr
a
ph
in
F
ig
ure
3
an
d
F
ig
ur
e
4.
It
ca
n
be
see
n
that
CSA
pro
vid
e
d
the
m
ini
m
u
m
fu
el
cost
and
Em
issi
on
cost
i
n
this
case
c
om
par
ed
t
o
oth
e
r
r
eported
m
et
ho
ds
in
the
li
te
rat
ur
e
.
T
his
s
how
s
that
the
C
SA
is
m
or
e
eff
ect
ive
in
fin
ding the
b
es
t l
oad
f
or
t
he
t
hr
ee
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
14
nodes
with
3
pr
oduct
ion
unit
s.
T
he
total
dem
and
of
the
netw
ork
is
equ
al
to
40
0
M
W
an
d
los
s
coeffic
ie
nts
are
as
fo
ll
ows:
=
10
−
5
[
7
.
1
3
.
0
2
.
5
3
.
0
6
.
9
3
.
2
2
.
5
3
.
2
8
.
0
]
Table
1
. T
he pa
ram
et
ers
of
th
e
co
st
functi
on a
nd g
e
ne
rators l
i
m
i
ts of
t
he
t
hree
-
un
it
syst
em
Un
its
a (
$
/M
W
2
)
b
(
$
/MW)
c (
$
)
P
m
in
(
M
W
)
P
m
ax
(
M
W
)
1
0
.03
5
4
6
3
8
.30
5
5
3
1
2
4
3
.5
3
1
1
35
210
2
0
.02
1
1
1
3
6
.32
7
82
1
6
5
8
.5
6
9
6
130
325
3
0
.01
7
9
9
3
8
.27
0
4
1
1
3
5
6
.6
5
9
2
125
315
Table
2.
Sam
pl
e Em
issi
on
Co
eff
ic
ie
nts
of th
e
three
-
un
it
sy
stem
Un
its
1
4
.09
1
-
5
.55
4
6
.49
2
.00
E
-
03
2
.85
7
2
2
.54
3
-
6
.04
7
5
.63
8
5
.00
E
-
04
3
.33
3
3
4
.25
8
-
5
.09
4
4
.58
6
1
.00
E
-
06
1
Table
3
. Res
ults o
f
the
eco
no
m
ic
d
ispatc
hing
of
three
-
un
it
syst
e
m
CM
PSO
CSA
P1
(
M
W
)
8
2
.08
7
0
8
2
.07
8
6
8
2
.07
8
3
P2
(
M
W
)
1
7
5
.0042
1
7
5
.0050
1
7
5
.00
.4
8
P3
(
M
W
)
1
5
0
.4938
1
5
0
.5002
1
5
0
.4961
Pl (
M
W
)
7
.56
8
7
7
.56
8
1
7
.56
8
1
Fu
el cos
t (
$
/h
)
2
0
8
1
3
2
0
8
1
3
208
1
2
.5
Total E
m
iss
io
n
(ton
/h
r
)
0
.21
3
9
2
1
0
.21
3
8
4
1
0
.21
3
8
4
1
Total e
m
iss
io
n
co
st ($/h
)
3
3
0
.018
3
2
9
.018
3
2
8
.7687
Figure
2. Cost
conve
rg
e
nce c
har
act
erist
ic
of 3
-
ge
ner
at
or syst
e
m
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
Eco
nomic
and emissi
on
dispa
tc
h
usi
ng c
uck
oo se
ar
ch
a
l
gorit
hm
(R
ach
i
dHa
bachi)
3389
Figure
3
.
Com
par
is
on F
uel c
os
t (
$/h)
gr
a
ph
betwee
n
CSA a
nd PSO
Figure
4
.
Com
par
is
on gra
ph
Total
Em
issi
on
co
st
($
/
h)
Bet
ween
CSA a
nd PSO
5.
CONCL
US
I
O
N
In
this
pa
per,
we
propose
d
a
Cuckoo
search
al
gorith
m
to
so
lve
the
econom
ic
and
Em
iss
ion
disp
at
c
h
.
T
he
pr
act
ic
al
it
y
of
the
pro
po
se
d
m
et
aheu
risti
cs
CSA
was
te
ste
d
f
or
th
ree
powe
r
ge
n
erat
ors.
T
he
gaine
d
res
ults
wer
e
com
par
e
d
to
e
xisti
ng
re
su
lt
s
ba
sed
on
PSO
an
d
CM
m
et
ho
ds.
It
wa
s
sho
wn
that
C
SA
a
r
e
su
pe
rio
r
in
ob
t
ai
nin
g
a
c
om
bi
nation
of
powe
r
load
s
that
fu
l
fill
the
pr
oble
m
con
strai
nts
a
nd
m
ini
m
iz
e
the
total
fu
el
c
os
t
an
d
e
m
issi
on
cost
.
CSA
f
ound
to
be
ef
fici
ent
in
fin
ding
the
op
t
i
m
al
po
wer
ge
ner
at
io
n
loa
ds.
CSA
was
capa
ble
of
handling
th
e
non
-
li
near
it
y
of
ED
pro
blem
.
T
he
ev
olv
e
d
po
wer
us
in
g
CS
A
m
ini
m
iz
ed
bo
th
the
cost
of
ge
ner
at
ed
po
wer
,
t
he
total
powe
r
los
s
in
the
tra
ns
m
issi
on
a
nd
m
ax
i
m
iz
es
the
reliab
il
it
y
of
the
pow
e
r
pro
vid
e
d
to
t
he
custom
ers.
T
he
pro
gr
am
s
we
re
de
vel
op
e
d
usi
ng
M
ATL
A
B
and
te
ste
d
a
netw
ork
of
14
nodes
.
The
res
ults
ha
ve
s
how
n
t
hat
our
CS
A
to
gi
ve
us
a
bette
r
perform
ance
w
it
h
opti
m
a
l
resu
lt
s
in
al
l
ca
se
s
an
d
resp
ect
in
g
t
he c
onstrai
nts im
po
se
d.
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.
H
.
Rui
and
P
.
Xu,
“
Stud
y
on
Sm
art
Grid
S
y
stem
Based
on
Sy
stem
D
y
n
amics,
”
TEL
KOMNIKA
Tele
communic
a
t
ion
Computing
El
e
ct
ronics
and
Control
,v
ol
/
issue:
12
(
12
)
,
pp
.
79
79
-
7986,
De
c
20
14.
[2]
Shahinz
ad
eh
H
.
and
Hasana
lizade
hK
.
A.
,“
Im
ple
m
ent
at
ion
of
Sm
art
Mete
ring
S
y
st
ems
:
Chal
le
ng
es
a
nd
Soluti
ons
,”
TEL
KOMNIKA
Telecomm
unic
ati
on
Computing
E
le
c
t
ronics
and
Cont
rol
,
vol
/i
ss
ue:
12
(7)
,
2014
.
[3]
Herm
aga
santos
Z
.
,
e
t
al.
,“
Im
ple
m
ent
at
ion
of
Elec
tr
ic
i
t
y
Com
petiti
on
Fram
ework
with
Ec
onom
ic
Dispatc
h
Dire
c
t
Method
,”
TEL
K
OMNIKA
Tele
c
omm
unic
ati
on
Compu
ti
ng
El
ectronics
and
Co
ntrol
,
vol/is
sue:
10(4)
,
pp.
667
-
674
,
2012
.
[4]
W
oll
enbe
rg
B
.
a
nd
W
ood
A
.,“
Pow
er
gen
era
t
ion,
oper
ation
and
co
ntrol
,”
New Yor
k
,
W
ile
y
,
1996
.
[5]
Dieu
V
.
N
.
,
e
t
al.
,“
Ps
eudo
-
gr
adi
en
t
base
d
p
art
i
cl
e
sw
arm
opti
m
iz
ation
m
et
hod
for
nonc
onvex
ec
onom
i
c
dispat
ch
,”
i
n
Po
wer,
con
trol
and
opti
mization
,
Sp
ringe
r, New Yor
k,
pp
.
1
-
27
,
201
3
.
[6]
K.
Aoki
and
T.
Satoh,
“
New
a
lgori
thms
for
class
ic
e
conomic
loa
d
d
ispat
ch
,
”
IEE
E
Tr
ansacti
ons
on
Power
Apparatus
and
S
yste
ms
,
vol
/i
ss
ue
:
PAS
-
103
(
6
)
,
p
p.
1423
-
1431
,
Ju
n
1984
.
[7]
J.
K.
Delson
and
S.
M.
Shah
ide
h
pour,
“
Li
ne
ar
pr
ogra
m
m
ing
applications
to
powe
r
s
y
stem
ec
onom
ic
s,
p
la
nning
an
d
oper
ations,”
IEEE
Tr
ansacti
ons
on
Powe
r S
ystem
s
,
vol
/i
ss
ue:
7
(
3
)
,
pp
.
1155
-
116
3,
Aug 1992
.
[8]
S.
Subram
ani
an
and
S.
Gan
esa
n,
“
A
sim
ple
a
pproa
ch
for
emiss
ion
constra
ined
ec
onom
ic
d
ispat
ch
prob
le
m
s,
”
Inte
rnational
Jo
urnal
of
Comput
er
Applications
,
vo
l/
issue:
8
(
11
)
,
pp.
3
9
-
45
,
Oc
t
2
010
.
[9]
D.
D.
Obiom
a
a
nd
A.
M.
Izu
ch
ukwu,
“
Com
par
at
iv
e
anal
y
s
is
of
te
chn
ique
s
for
ec
onom
ic
dispa
t
ch
of
gene
r
ated
power
with
m
odifi
ed
la
m
bda
-
i
t
era
t
ion
m
et
hod,
”
in
Proceedi
ng
s
of
the
2013
I
EE
E
In
te
rnation
al
Confe
ren
ce
o
n
Eme
rging Sustainable
Te
chnol
og
ie
s for Power
IC
T in
a
D
ev
e
lopi
n
g
Society (
NIGERCON)
,
pp.
231
-
237
,
Nov 2013
.
[10]
S.
K.
Mishra
an
d
S.
K.
Mishra,
“
A
com
par
at
iv
e
stud
y
of
soluti
on
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
Comput
er
S
ci
en
ce
,
vol
.
48
,
pp.
96
-
100
,
201
5.
[11]
S
.
K
.
Dewangan,
et
al.
,
“
A
Tr
adi
ti
on
al
Appro
ac
h
to
Solve
E
c
onom
ic
Loa
d
D
ispat
ch
Probl
em
C
onside
ring
th
e
Gene
rat
or
Const
rai
nts
,
”
IOSR
Jou
rnal
of
El
ectric
a
l
and
El
ectronic
s
Engi
nee
ring
(
IOSR
-
JE
E
E)
,
vol
/i
ss
ue:
10
(
2
)
,
pp
.
27
-
32
,
2015
.
[12]
B.
Stott
,
et
a
l.
,
“
Pow
er
S
y
stem
Secur
i
t
y
Contro
l
Cal
cula
t
ions
Us
ing
Li
ne
ar
Progr
amm
ing
,
”
Part
I
,
IEEE
Tr
ans.
on
Powe
r A
pparatu
s and
Syste
ms
,
v
ol
/i
ss
ue:
PAS
-
97
(
5
)
,
pp
.
1713
-
17
20,
Sep
1978.
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.
9
, N
o.
5
,
Oct
ober
20
19
:
3
3
8
4
-
3
3
9
0
3390
[13]
D.
L.
Tr
ave
rs
an
d
R.
Ka
y
e
,
“
D
y
namic
dispat
ch
b
y
co
nstru
ct
iv
e
d
y
nami
c
progra
m
m
ing
,
”
IEE
E
Tr
ans.
on
Powe
r
Syste
ms
,
vol
/i
ss
ue:
13
(
1
)
,
pp
.
72
-
78,
Feb1998
.
[14]
Gaing,
“
Parti
c
le
sw
arm
opti
m
iz
ati
on
to
solving
th
e
ec
onom
ic
d
ispat
ch
conside
ring
the
gen
era
tor
co
nstrai
nts
,
”
IEEE
Tr
ans.
Powe
r Sy
st.
,
vo
l. 18, pp. 1
187
-
1195,
Aug
2
003.
[15]
E.
Li
n
and
G.
L.
Vivia
n
i,
“
Hier
arc
h
ical
E
c
onom
ic
Dispatch
for
pie
ce
wis
e
quadr
at
i
c
cost
func
ti
ons
,
”
IE
EE
Tr
ansacti
ons on p
ower
apparatus and
systems
,
vo
l
/i
ss
ue:
PAS
-
103
(
6
)
,
pp
.
1170
-
11
75
,
1984
.
[16]
G
.
Kum
ar
and
R
.
Singh,
“
Ec
on
om
ic
Dispatc
h
of
Pow
er
S
y
st
e
m
Optimiza
ti
on
with
Pow
er
Ge
ner
ation
Schedu
le
Us
ing
Evol
uti
o
nar
y
Te
chn
ique
,
”
Inte
rnationa
l
Journal
of
Adv
ance
d
R
ese
arc
h
in
El
ec
tric
al
El
ectronics
and
Instrum
ent
ati
on
Engi
ne
ering
,
vol
/i
ss
ue:
3
(
7
)
,
Jul
2014.
[17]
K.
S.
Kum
ar,
et
al.
,
“
Ec
onom
ic
loa
d
dispatch
with
emiss
ion
c
onstrai
nts
using
var
ious
PS
O
al
g
orit
hm
s
,
”
WSE
A
S
Tr
ansacti
ons on Power
Syst
ems
,
vol
/i
ss
ue:
3
(
9
)
,
p
p.
598
-
607
,
200
8.
[18]
H.
Vennila
and
T.
R
.
D
.
P
rak
as
h
,
“
A
soluti
on
f
or
envi
ronm
enta
l
constraine
d
Econom
isc
Dispatch
Prob
le
m
s
using
Hone
y
Bee Algo
rit
hm
,
”
Int
ernational
journal
of Com
pute
r A
ppl
i
cat
ion
,
vol
/i
ss
ue
:
47
(
22
)
,
pp.
13
-
17,
2012
.
[19]
H.
Vennil
a
and
T
.
R
.
D
.
Praka
sh,
“
Parti
cl
e
Sw
am
Optimis
at
ion
Te
chni
qu
e
for
solving
Ec
onom
ic
Dispatch
Problems
,
”
El
s
e
vi
er
Journal
o
f
P
roce
dia Engi
ne
e
ring
,
v
o
l
.
48
,
pp
.
2009
-
2021
,
201
2
.
[20]
Jiang
A
.
,
et
al.
,“
Ec
onom
ic
disp
at
ch
wi
th
non
-
m
onotoni
ca
l
l
y
i
ncr
ea
sing
inc
r
e
m
ent
al
cost
units
a
nd
tra
nsm
ission
s
y
stem
losses
,”
Discuss
ion.
IEEE
Tr
ans P
ower
S
yst
,
vol
.
10
,
pp
.
8
91
-
897
,
1995
.
[21]
X.
S.
Yang
an
d
S.
Deb,
“
Cuc
koo
sea
rch
vi
a
Le
v
y
fl
ight
s
,
”
i
n
Proce
ed
ings
of
world
congre
ss
on
natur
e
&
biol
ogic
a
ll
y
insp
ired
comput
ing,U
SA;
IEEE Publ
ic
ati
ons
,
pp
.
210
-
214,
De
c
2009
.
[22]
X.
S
.
Yang
an
d
S.
Deb,
“
Engi
nee
r
ing
opti
m
i
za
t
ion
b
y
cuc
k
oo
sea
rch
,
”
Int
.
J.
Math
emati
c
al
Mode
lling
a
nd
Numerical
Opt
i
misation
,
vo
l
/i
ss
ue:
1
(
4
)
,
pp
.
330
-
343,
2010
.
[23]
D.
D.
Obiom
a
a
nd
A.
M.
Izu
ch
ukwu,
“
Com
par
at
iv
e
anal
y
s
is
of
te
chn
ique
s
for
ec
onom
ic
dispa
t
ch
of
ge
ner
ated
power
with
m
odifi
ed
la
m
bda
-
i
t
era
t
ion
m
et
hod,
”
in
Proceedi
ng
s
of
the
2013
I
EE
E
In
te
rnation
al
Confe
ren
ce
o
n
Eme
rging Sustainable
Te
chnol
og
ie
s for Power
IC
T in
a
D
ev
e
lopi
n
g
Society (
NIGERCON)
,
pp.
231
-
237
,
Nov 2013
.
[24]
A.
L.
Devi
and
O.
V.
Krishna,
“
Com
b
ine
d
ec
on
om
ic
and
emiss
ion
dispat
ch
usin
g
evol
uti
on
ar
y
a
lgori
thms
:
A
ca
se
stud
y
,
”
ARPN
Jo
urnal
of Engi
ne
e
ring a
nd
App
li
ed
Scienc
es
,
vol
/i
ss
ue:
3
(
6
)
,
pp
.
28
-
35,
2008
.
[25]
Z.
L
.
G
ai
ng
,
“
Particle
Sw
arm
Optimiza
ti
o
n
to
Solving
the
Ec
onom
ic
Dispatc
h
Consid
e
rin
g
the
Gene
r
at
o
r
Constrai
nts
,”
IEEE
transact
ions o
n
power
systems
,
vol
/i
ss
ue:
18(3)
,
pp.
1187
-
1195
,
2003
.
Evaluation Warning : The document was created with Spire.PDF for Python.