TELKOM
NIKA
, Vol. 13, No. 4, Dece
mb
er 201
5, pp. 1187
~1
193
ISSN: 1693-6
930,
accredited
A
by DIKTI, De
cree No: 58/DIK
T
I/Kep/2013
DOI
:
10.12928/TELKOMNIKA.v13i4.3166
1187
Re
cei
v
ed
De
cem
ber 2
3
, 2014; Re
vi
sed
Jan
uar
y 29, 2
015; Accepte
d
March 12, 2
015
Comparing Performances of Evolutionary Algorithms
on the Emission Dispatch and Economic Dispatch
Problem
AN Afa
ndi
1
, Irham Fadlik
a
2
, Andoko
3
1,2
Electrical En
gin
eeri
ng, Un
iv
ersitas Ne
geri
Mala
ng, Jl
. Semarang 5, Building
G4, Malang, Ja
w
a
T
i
mur,
Indon
esi
a
3
F
a
cult
y
Memb
er of Engin
eer
i
ng, Univ
ersitas
Neger
i Mala
ng
, Jl. Semarang
5, Build
ing H
5
, Mala
ng, Ja
w
a
T
i
mur, Indones
ia
e-mail: an.afa
nd
i
@
u
m
.a
c.id
1
, an.afan
di@
i
e
e
e
.org
1
, irh
a
m.el
ektro.um@gm
a
il.com
2
, andoko.ft@um.ac.id
3
A
b
st
r
a
ct
Evoluti
onary methods do
mi
na
te
in
the co
mputatio
n for br
eaki
ng o
u
t the
real pro
b
l
e
ms
. F
o
r a
coup
le of ye
ar
s, it is more p
o
pul
ar
than cl
as
sical
meth
ods f
o
r solvi
ng
ma
n
y
cases. T
e
chn
i
cally, o
ne of r
eal
prob
le
ms
is th
e e
m
iss
i
o
n
d
i
s
patch
and
ec
o
n
o
m
ic
dis
patch
(EDED)
prob
l
e
m. T
h
e ED
ED
pro
b
le
m
is
us
ed to
opti
m
i
z
e
the
p
o
w
e
r syste
m
o
perati
on (PSO)
at a
certa
i
n
ti
me
u
nder
so
me co
nstraints.
This p
aper
pr
e
s
ent
s
perfor
m
a
n
ce c
o
mparis
on
bet
w
een Harv
est
Seaso
n
Arti
fici
al Be
e C
o
lo
ny
(HSABC) a
nd
Genetic Al
gor
ithms
on the E
D
ED problem
applied to t
he IEEE-62 bus system
. Sim
u
lation res
u
lt
s show that the tested methods
have d
i
fferenc
e character
i
stic
s and ab
iliti
e
s for
opti
m
i
z
i
n
g th
e PSO based o
n
the EDED pr
obl
e
m
.
Ke
y
w
ords
: ec
onomic dispatch, em
iss
i
on dispatch, genet
ic algorit
hm
, HSABC algorithm
,
power system
oper
ation
Copy
right
©
2015 Un
ive
r
sita
s Ah
mad
Dah
l
an
. All rig
h
t
s r
ese
rved
.
1. Introduc
tion
At present years, many im
porta
nt decision
s
are m
ade from de
scribin
g
the optimal
solutio
n
for measuri
ng the real probl
e
m
of
t
he real operation ba
sed on de
sirable solutio
n
s
to
meet operational co
nst
r
ai
nts while ex
isting in
the
certain con
d
ition. In particula
r
,
man
y
mathemati
c
al
and o
p
timization techniq
ues
hav
e b
e
en propo
se
d
to solve the
probl
em
s a
nd
several
p
r
evi
ous wo
rks ha
ve
been su
ccessfully
appl
i
ed to
carry o
u
t real
proble
m
s [1], [2], [3].
Other
studie
s
have al
so
repo
rted m
any tec
hni
q
ues fo
r obt
aining the o
p
timal soluti
on
categ
o
ri
zed i
n
to cla
ssi
cal
method
s and
evolutiona
ry method
s [4], [5], [6],
[7],
[8], [9].
In general,
evolutiona
ry method
s are
con
s
iste
d of many
algori
t
hms ba
se
d on natu
r
al e
n
tity behaviors
whi
c
h a
r
e ad
opted a
s
opti
m
ization m
e
thod
s for imp
r
oving pe
rform
ances of
cla
s
sical tech
niqu
es.
More
over
, e
v
olutionary
method
s do
minate in
computation
s
for determi
ning sol
u
tion
s of
optimizatio
n probl
em
s on
many application
s
with
various the
m
es.
The
s
e
methods a
r
e
comm
only de
veloped a
s
an intelligent computation fo
r sea
r
ching the optimal so
lution usin
g an
optimizatio
n tech
niqu
e in multiple co
nst
r
aints.
For a
co
uple
of years, g
e
n
e
tic alg
o
rithm
(GA)
i
s
mo
re pop
ular th
a
n
other
algo
ri
thm in
the impleme
n
t
ation of evolutionary met
hod
s in ma
n
y
variants a
s
pre
s
ente
d
in
many previo
us
studie
s
. In p
r
i
n
cipl
es, thi
s
algorith
m
ha
s been
inspire
d
by a p
hen
o
m
enon
of a n
a
tural
evolution
and many p
r
e
v
ious works h
a
ve use
d
to carry out
optim
ization p
r
obl
e
m
s appli
ed to solve man
y
topics. Its proced
ure
s
for selectin
g the optimal
solutio
n
are perfo
rm
ed by several
steps cove
re
d
popul
ation, natural sele
ction, cro
s
sove
r
,
and mutation.
In detail,
prin
ciple
s
an
d procedu
re
s of
GA
are di
scu
s
sed
clea
rl
y in literatures [3], [
10], [1
1], [12]. Rec
e
ntly
, the lates
t
variant
of
evolutiona
ry method
s is harvest se
aso
n
artifi
cial bee colony (HS
ABC) algorith
m
promoted i
n
2013.
Thi
s
al
gorithm i
s
co
mposed u
s
in
g bee b
ehav
i
ours an
d the
harve
st se
ason situ
ation,
and
its function i
s
discussed cl
early
as the
computing abilit
y in [13].
Re
cently
, the
applicatio
n of evolutiona
ry
algorithm
s to the power syste
m
operatio
n
(PSO) is m
o
re popul
ar tha
n
cla
ssi
cal m
e
thod
s for b
r
inging o
u
t technical pro
b
le
ms. One of t
h
e
most imp
o
rta
n
t com
b
ined
probl
em
s is t
he emi
ssi
on disp
atch and
eco
nomi
c
di
spat
ch (EDE
D)
probl
em fo
r searchin
g an
optimal bal
an
ce in th
e a
cceptable
econ
omic
ope
ratio
n
with va
riou
s
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No
. 4, Decem
b
e
r
2015 : 118
7 – 1193
1188
techni
cal
co
n
s
traint
s a
nd
condition
al limi
t
s. In these st
udie
s
, this p
r
oblem i
s
u
s
e
d
for
com
pari
ng
GA
and HSA
B
C on both
perfo
rman
ce
s through
out the com
putin
g ability whil
e sea
r
ching t
he
optimal sol
u
tion und
er vari
ous
con
s
trai
n
t
s and limitations.
2. Genetic a
nd Harv
est Season
Arti
fic
i
al Bee Colo
n
y
Algorithms
As me
ntione
d befo
r
e, int
e
lligent
com
p
utations
are
con
s
i
s
ted of
ev
olutiona
ry method
s
comp
osed u
s
ing a p
opul
ation ba
se
d alg
o
rithm. Many
method
s hav
e bee
n intro
d
uce
d
to attem
p
t
the natural p
henom
eno
n for creating va
riou
s evol
utio
nary algo
rith
ms and it has been su
cce
s
sful
applie
d to th
e optimizatio
n pro
b
lem
s
[2], [3], [4]
,
[8],
[10], [14], [15]. Spec
ific
ally
, for the las
t
ten
years, the m
o
st pop
ular
of ev
olutiona
ry algorithm
s is GA
inspi
r
ed by a phe
nomen
on of the
natural
evolut
ion. GA’
s
p
r
o
c
ed
ure
s
fo
r
selectin
g the
optimal
soluti
on a
r
e
pe
rformed
by seve
ral
step
s, su
ch a
s
, popul
ation;
natural
sele
ction;
cro
s
so
ver; and mut
a
tion, as di
scussed
clea
rly
in
[3], [10], [15]. Many previo
us
wo
rks h
a
ve used GA fo
r solving o
p
ti
mization
prob
lems
of the P
S
O
and thi
s
alg
o
rithm
ha
s
been
ap
plied
to
solve v
a
riou
s topics of di
spat
ch
ing p
o
wers
for
determi
ning
the o
p
timal
so
lution a
nd fo
r sche
duling
p
o
we
r o
u
tputs
of gen
eratin
g
units. In
deta
il,
prin
ciple
s
an
d pro
c
ed
ure
s
of GA are clearly disc
u
ssed in literatures [3],
[10],
[
16], [15]. Thes
e
studie
s
are condu
cted to those
previo
u
s
wo
rks for G
A
’s pr
in
cipl
es and hierarchi
e
s.
In partic
ular,
HSABC algorithm
was
propo
s
e
d on
Marc
h
2013
after developing and
in
tr
o
d
u
c
i
ng
ea
r
l
y its
s
t
r
u
c
t
u
r
es
in
2
0
12 b
a
s
ed
on the harve
st season situ
ation co
nsi
s
ted
of
bloomin
g flowers i
n
its are
a
and bee’
s be
haviors whil
e sea
r
ching fo
r foods. In deta
il, flowers
are
expre
s
sed
u
s
ing
multiple
food
source
s
relate
d
to
the first food
so
urce
an
d
the
other fo
od
sou
r
ces. E
a
ch food
so
urce is l
o
cated
randomly
at a
ce
rtain
po
sition u
s
ing
a
h
a
rvest
ope
rat
o
r
[13], [17]. In princ
i
ple, the s
equenc
i
ng computation of
HSABC
algorithm is dis
t
ributed into
several
pro
c
e
s
ses to
sel
e
ct
the
optimal
solution
as foll
owin
g o
r
de
rs
as
ado
pted
in
these
studi
e
s
[5], [14]:
Gene
rating
p
opulatio
n: create initial p
opulat
io
n set
s
, evaluate i
n
itial popul
ation sets, an
d
define the po
pulation.
Food so
urce
exploratio
n: prod
uce
the first
food
so
u
r
ce,
pro
d
u
c
e
the othe
r fo
od sou
r
ce
s,
evaluate the
multiple food sour
ces, apply the greed
y process, and calc
ulate t
he probability
value.
Food
sele
ctio
n: prod
uce a
new fo
od, p
r
odu
ce n
e
ighb
or food
s, eval
uate food
s, a
nd ap
ply the
gree
dy pro
c
e
ss.
Abando
ned
repla
c
eme
n
t: determi
ne an
aba
ndo
ned
food, repla
c
e
with
a
ne
w ra
ndomly
o
ne,
and mem
o
ri
ze the food.
3. Emission
Dispa
t
ch an
d Economic Dispa
t
ch
In gen
eral, th
e PSO i
s
st
ru
cture
d
u
s
in
g
variou
s p
a
rt
s
and
equip
m
e
n
ts for e
s
tabli
s
hin
g
an
integrate
d
sy
stem to deliver ele
c
tric e
n
e
rgy fr
om ge
neratin
g we
b
s
to usag
e areas un
de
r ce
rtain
operational
constraints wit
h
the
di
strib
u
t
ed o
w
n
po
we
r con
s
umptio
n. Techni
call
y, this ope
rati
on
sho
u
ld be p
r
ovided in re
a
s
on
able bu
dg
ets for a
ll pro
c
e
s
ses in 24
hours of the operating pe
riod
force
d
in
envi
r
onm
ental
re
quire
ment
s [1
4]. Re
cently, the e
n
viron
m
e
n
t prote
c
tion
i
s
con
s
id
ere
d
in
the PSO to
redu
ce
polluta
nt produ
ction
s
at
ther
mal
power plant
s
discha
rge
d
in
vario
u
s pa
rticl
e
s
and ga
sse
s
, like CO, CO
2
, S
O
x
and NO
x
[2],
[3], [12],
[14], [15],
[18]. To c
o
ver t
hese problems
,
the PSO con
s
ide
r
s fina
nci
a
l and environmental
a
s
pect
s
for se
arching the
balan
ce de
ci
sio
n
betwe
en e
m
issi
on di
sp
atch and
econo
mic di
sp
atch
with its
different o
r
ientati
on tro
ugh
out
the
EDED fo
rmul
ation a
s
singl
e obj
ective fu
nction
for
det
ermini
ng the
optimal
soluti
on a
nd pl
otting a
committed
schedul
e of gen
erating u
n
it outputs.
Basically, the
EDED is
ad
dre
s
sed to
m
i
nimize
the to
tal fuel
cost i
n
clu
ded th
e
pollutant
redu
ction
wit
h
con
s
id
erin
g seve
ral lim
itations
for t
he PSO. Moreove
r
, the EDED p
r
obl
e
m
is
comp
osed u
s
ing
pe
nalty and com
p
ro
mised
facto
r
s
fo
r combi
n
g
finan
cial
a
nd
e
n
viron
m
enta
l
asp
e
ct
s. The
penalty fact
or i
s
u
s
ed t
o
sh
ows
the
coeffici
ent rate of ea
ch
gene
rating.
Th
e
comp
romi
se
d
factor sh
ows the cont
rib
u
tion of
pollutants and
costs in the
comp
utation.
In
addition, eco
nomic di
sp
atch; emissio
n
dis
pat
ch; and the EDED pro
b
lem
are form
ed
in
expre
ssi
on (1
); (2); an
d (3
).
Econo
mic di
spatch:
F
∑
c
b
.P
a
.P
,
(
1
)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Com
pari
ng Perform
a
n
c
e
s
of Evolutiona
ry Al
gorithm
s
on the Em
issi
on Di
spat
ch
… (AN Afandi
)
1189
Emissi
on di
spatch:
E
∑
γ
β
.P
α
.P
,
(
2
)
EDED proble
m
:
w
.
F
1
w
.h
.
E
,
(
3
)
whe
r
e P
i
is a
output po
we
r of i
th
generat
ing unit (M
W), a
i
; b
i
; c
i
are fuel co
st co
efficients
of i
th
generating unit, F
tc
is a total fuel cost ($/h
), ng
is num
ber
of gene
rating
u
n
it,
α
i
;
β
i
;
i
are
emission
coef
ficients of i
th
g
enerating
unit
,
E
t
is a t
o
tal
emission
(kg/
h),
i
s
the E
D
ED
($/h
), w
is
a comp
romi
sed facto
r
, and
h is a penalt
y
factor.
4. Teste
d
Sy
stem and Pr
ocedur
es
To compare
performances of
GA and
HSABC, desi
gned prog
rams of both
al
gorithm
s
are
appli
ed t
o
a
stan
da
rd
model f
r
om
Institution
of Electri
c
al and
Electroni
cs Enginee
rs
(IE
EE).
IEEE’s models are often
adopted as
sample
sy
stems for
simul
a
ting many problems.
These
stand
ard m
o
dels a
r
e al
so
useful an
d effective to
test related prob
lems of the PSO beca
u
se of
available
pro
v
ided techni
cal data. In
p
a
rticul
ar, m
a
ny pro
b
lem
s
of the PSO
are
app
roa
c
h
ed
usin
g the
s
e
standard mo
d
e
ls in
ord
e
r t
o
kn
ow
ch
aracteri
stic; p
e
rforman
c
e
s
; a
nd re
sult
s of
the
problems
.
In thes
e
s
t
udi
es, the IEEE-62 bus
s
y
s
t
em is
s
e
lec
t
ed as a tes
t
ed s
y
s
t
em for both
algorith
m
s.
Table 1. Fuel
Co
st and Emi
ssi
on Coeffici
ents of Gen
e
rators
Bus G
e
n
a, x10
-
3
($/MW
h
2
)
b
($/MW
h
)
c
α
(k
g/MW
h
2
)
β
(kg/MWh)
1 G1
7.00
6.80
95
0.0180
-1.8100
24.300
2 G2
5.50
4.00
30
0.0330
-2.5000
27.023
5 G3
5.50
4.00
45
0.0330
-2.5000
27.023
9 G4
2.50
0.85
10
0.0136
-1.3000
22.070
14 G5
6.00
4.60
20
0.0180
-1.8100
24.300
17 G6
5.50
4.00
90
0.0330
-2.5000
27.023
23 G7
6.50
4.70
42
0.0126
-1.3600
23.040
25 G8
7.50
5.00
46
0.0360
-3.0000
29.030
32 G9
8.50
6.00
55
0.0400
-3.2000
27.050
33 G10
2.00
0.50
58
0.0136
-1.3000
22.070
34 G11
4.50
1.60
65
0.0139
-1.2500
23.010
37 G12
2.50
0.85
78
0.0121
-1.2700
21.090
49 G13
5.00
1.80
75
0.0180
-1.8100
24.300
50 G14
4.50
1.60
85
0.0140
-1.2000
23.060
51 G15
6.50
4.70
80
0.0360
-3.0000
29.000
52 G16
4.50
1.40
90
0.0139
-1.2500
23.010
54 G17
2.50
0.85
10
0.0136
-1.3000
22.070
57 G18
4.50
1.60
25
0.0180
-1.8100
24.300
58 G19
8.00
5.50
90
0.0400
-3.000
27.010
In these studi
e
s, the IEEE-62
bus
system is adopted
as a
sample model of
the PSO
for
compari
ng ability of GA and HSABC. In
detail, this
system
is devel
oped usin
g 19 generators, 62
buses
and 8
9
line
s
. One
line diag
ram
of the IEEE
-62 b
u
s
syst
em is
sho
w
n
in Figure 2.
Its
gene
rating
coefficient
s a
r
e liste
d in
Ta
ble 1
for f
uel
co
nsumptio
n
s
a
n
d
polluta
nt produ
ction
s
of
19 units.
Main p
r
o
c
ed
ure
s
of
simu
lations fo
r
p
e
rformi
ng
G
A
and
HSA
B
C a
r
e d
e
scrib
ed
by
followin
g
expl
anation
s
a
s
ill
ustrate
d
in
Figure
1. The
first
step i
s
an
EDED fo
rmat
ion con
s
ide
r
e
d
emission
and
eco
nomi
c
a
s
pe
cts i
n
si
n
g
le obj
ecti
ve
function. T
h
e
se
con
d
ste
p
is an
algo
rithm
comp
ositio
n f
o
r b
o
th evol
utionary
met
hod
s
con
s
tr
u
c
ted
by its
p
a
ram
e
ters. T
he third
step
is
prog
ram
m
ing
developme
n
t
s for all co
mputing proc
esse
s asso
ci
ated with ref
e
ren
c
e
s
for the
tes
t
ed algorit
h
ms
[3], [5], [
10], [14], [15].
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No
. 4, Decem
b
e
r
2015 : 118
7 – 1193
1190
Figure 1. Sequen
cing p
r
o
c
edures fo
r the comp
ari
s
o
n
Figure 2. One-line diagram of IEEE-62 bus
s
y
s
t
em
In these
sim
u
lation
s, ope
rational con
s
traints a
r
e al
so use
d
to pu
t desired sol
u
tions in
the fea
s
ible
rang
es of th
e PSO. In
d
e
tail, ope
rati
onal
co
nst
r
ai
nts a
r
e
defin
ed in
5%
of
maximum an
d minimum li
mits of fluctuated vo
ltage
s; 90% of the maximum power tra
n
sf
er
capability on transmission lines; an
equality of
total powers bet
ween
generating units, power
losse
s
, and
load
s; maxi
mum an
d m
i
nimum p
o
wer limits of
gene
rating
u
n
its; 15%
of the
maximum total power lo
ss;
and the emission
stand
ard
.
5. Results a
nd Discu
ssi
on
These wo
rks
are ad
dresse
d to
obtain the optimal sol
u
tion of
the PSO throug
h the EDED
solved using
HSABC and
GA. These si
mulations
are also used to compare its
ability based
on
the EDED. T
o
demonst
rat
e
GA and
HSABC, thes
e simulations
have considered 2,912 M
W
of
the power d
e
m
and; 0.5 of
the com
p
ro
mi
sed fa
ctor
; 0.
85 kg/M
Wh o
f
the emissi
o
n
stand
ard; the
domina
n
t pe
n
a
lty factor.
F
o
r
exe
c
utin
g desi
gne
d
p
r
o
g
ram
s
of
HS
ABC,
this alg
o
rithm ha
s
b
een
applie
d to sol
v
e the EDED using th
e co
lony size=
50
; food sou
r
ce
s= 2
5
; and 2
00 of foragi
n
g
Tech
nical
para
m
eters
Environme
n
tal
requi
rem
ents
EDED
probl
em
HSABC
GA
Re
sults
Nume
ri
cal
c
o
mpa
r
is
on
Grap
hical
comp
ari
s
o
n
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Com
pari
ng Perform
a
n
c
e
s
of Evolutiona
ry Al
gorithm
s
on the Em
issi
on Di
spat
ch
… (AN Afandi
)
1191
cycle
s
. Ea
ch
pro
c
e
s
s of G
A
has
co
nsi
d
ered
its
p
r
o
c
e
dure
s
and hie
r
archi
e
s as di
scusse
d
cl
ea
rly
in several ref
e
ren
c
e
s
[2], [10], [11].
In these simul
a
tions, GA ha
s been al
so im
plemente
d
using
its
qu
alified p
a
ram
e
ters co
vered pop
ulat
ion;
natu
r
al
selectio
n; cro
s
sover;
mutati
on; an
d oth
e
rs.
In detail, ma
in paramete
r
s of GA h
a
v
e use
d
po
p
u
lation
= 50;
natural sele
ction
=
roulett
e
;
cro
s
sove
r=
scattere
d; mutation= G
a
u
s
sian; and maxi
mum gen
erati
on= 2
00.
In parti
cula
r,
nume
r
ical results a
r
e
provi
ded
in
Ta
ble
2. Performan
c
e
s
of the
E
D
ED
are
also present
ed graphicall
y in Fi
gure 3 and Figure
4. These fi
gures illustrate
performances in
terms of
co
n
v
ergen
ce sp
e
eds and
time
con
s
um
pt
ion
s
.
Figu
re
3 shows conve
r
gen
ce spe
e
d
s
of
use
d
intellig
e
n
t com
putati
ons for t
w
o t
y
pes
of
evol
utionary algo
rithms
a
s
soci
ated with
ea
ch
sele
ction
fo
r obtainin
g
the optimal soluti
on
from
avail
able
can
d
idat
e sol
u
tion
s in
the po
pulatio
n.
This figu
re ill
ustrate
s
p
r
og
ressin
g optim
al solutio
n
s o
f
the EDED probl
em com
puted u
s
ing
GA
and HSABC for 200 cycles. By considering all
param
eters and operation
al constrai
nts for
solving
the
problem,
optim
al solution
s
o
f
the EDE
D
h
a
ve be
en
initialed
at differe
nt point
s
befo
r
e
leveling at th
e optimal
sol
u
tion a
s
sho
w
n i
n
Fig
u
re
3 with
the fa
stest is HSAB
C. In a
ddition
, the
executio
n of desi
gne
d pro
g
ram
s
for e
a
c
h alg
o
rithm
has
con
s
u
m
e
d
a ce
rtain time to com
p
le
te all
comp
utation
s
whil
e o
b
taini
ng the
o
p
timal solution
of
the E
D
ED with differe
nt chara
c
te
risti
c
s a
s
given in Fi
gu
re 4. In total,
both time
consumpt
ion
s
are li
sted i
n
Table
2 with
the shorte
st
is
HSABC. According to this table, the optimal so
l
u
tion of the EDED is
settled
at 11,585.13
$/h
with different
budg
eting fee
s
for polluta
nts and fuel
s.
Figure 3. Con
v
ergen
ce
spe
eds of EDE
D
’
s
co
mputatio
ns
Figure 4. Time con
s
um
ptio
ns of EDED’
s
computatio
n
s
11,
450
11,
500
11,
550
11,
600
11,
650
11,
700
11,
750
11,
800
1
2
1
4
1
6
1
8
1
101
121
141
161
181
EDED ($
/h
)
For
a
ging cy
cle
GA
HSABC
0.
00
0.
01
0.
02
0.
03
0.
04
0.
05
0.
06
0.
07
1
2
1
4
1
6
1
8
1
101
121
141
161
181
Tim
e
(m
in
)
For
a
ging cy
cle
GA
HSABC
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 16
93-6
930
TELKOM
NIKA
Vol. 13, No
. 4, Decem
b
e
r
2015 : 118
7 – 1193
1192
Table 2. Co
m
puting Perfo
r
mances
Subjects
G
A
HSABC
Subjects
G
A
HSABC
Fuels ($/h)
8,547.47
8,192.53
Range ($
/h)
176.66
72.41
Pollutants ($/h)
3,037.66
3,392.60
Optimal c
y
cle
104
11
EDED ($/h
)
11,585.13
11,585.13
Optimal time (min)
3.04
0.21
Start point ($/
h
)
11,761.85
11,657.60
Total time (min)
6.40
5.32
Table 3. Power and Poll
utant Perform
a
nce
s
Units
Powers (MW)
Pollutant (kg/h)
G
A
HSABC
G
A
HSABC
G1
140.24
105.75
124.48
34.18
G2
168.18
305.14
540
2,336.86
G3
258.16
380.22
1,580.89
3,847.27
G4
91.89
91.89
17.45
17.45
G5
140.48
146.48
125.25
145.39
G6
161.14
252.25
481.06
1,496.22
G7
145.27
108.62
91.37
23.97
G8
190.16
195.10
760.34
814.08
G9
357.32
119.92
3,990.73
218.55
G10
91.89
91.89
17.45
17.45
G11
110.22
147.94
54.11
142.30
G12
130.1
105.35
60.67
21.58
G13
281.01
220.83
937.08
502.41
G14
115.13
113.14
70.46
66.50
G15
340.31
399.78
3,177.35
4,583.36
G16
121.26
80.13
75.82
12.10
G17
91.89
91.89
17.45
17.45
G18
237.13
153.7
607.23
171.33
G19
130.47
94.84
316.52
102.30
Total 3,302.25
3,204.88
13,045.71
14,570.74
Table 4. Co
st
Performa
nce
s
Units
T
o
tal costs (
$
/hr
)
F
uel costs (
$
/h)
Emission co
sts (
$
/h)
G
A
HSABC G
A
HSABC
G
A
HSABC
G1
1,244.27
908.30
1,186.30
892.38
57.97
15.92
G2
1,109.79
2,850.97
858.31
1,762.69
251.48
1,088.28
G3
2,180.38
4,152.70
1,444.16
2,361.03
736.22
1,791.67
G4
117.35
117.35
109.22
109.22
8.13
8.13
G5
842.95
890.27
784.62
822.56
58.33
67.71
G6
1,101.41
2,145.77
877.38
1,448.98
224.03
696.79
G7
904.46
640.37
861.91
629.18
42.55
11.16
G8
1,622.30
1,685.12
1,268.01
1,306.01
354.29
379.11
G9
5,142.66
998.55
3,284.18
896.77
1,858.48
101.78
G10
128.96
128.96
120.83
120.83
8.13
8.13
G11
321.35
466.45
296.03
400.18
25.32
66.27
G12
259.16
205.34
230.91
195.29
28.25
10.05
G13
1,412.01
950.31
975.61
716.34
436.4
233.97
G14
361.65
354.60
328.84
323.63
32.81
30.97
G15
3,911.95
5,132.31
2,432.26
2,997.84
1,479.69
2,134.47
G16
361.24
236.71
325.93
231.08
35.31
5.63
G17
117.35
117.35
109.22
109.22
8.13
8.13
G18
940.22
457.02
657.43
377.23
282.79
79.79
G19
1,091.19
731.25
943.79
683.61
147.4
47.64
Total
23,170.65
23,170.65
17,094.94
16,385.05
6,075.71
6,785.60
Based o
n
the
EDED probl
em, final results for sch
e
duling g
ene
rating units of
the PSO
are lis
t
ed in
Table
3, which
have been
optimiz
ed us
ing GA
and
HSABC. This table
s
h
ows
real
con
d
ition
s
of
gene
rating
u
n
its to
com
p
o
s
e th
e
commi
tted po
wer o
u
tput con
s
ide
r
ing
2,912
M
W
.
As listed in th
is table, ba
se
d on the co
m
b
inati
on of ge
neratin
g units for the PSO con
s
id
ere
d
the
minimum tota
l operating co
st, it is kno
w
n
that so
me g
e
neratin
g unit
s
are op
erated
in fixed powe
r
outputs. To
make th
e un
it commitme
n
t, generatin
g units p
r
od
uce diffe
rent
individual p
o
we
r
outputs. Mo
reover, some
gene
rating
units a
r
e o
p
erated i
n
th
e sam
e
cap
a
citie
s
for b
o
th
algorith
m
s. By discha
r
gi
ng
pollutant emi
ssi
on
s wh
ile
prod
uci
ng po
wer o
u
tputs t
o
meet the total
power dem
a
nd, gene
ratin
g
units have
used comp
ens
ation fee
s
as liste
d in Table 4. These
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
1693-6
930
Com
pari
ng Perform
a
n
c
e
s
of Evolutiona
ry Al
gorithm
s
on the Em
issi
on Di
spat
ch
… (AN Afandi
)
1193
payments ha
ve been
co
n
s
ide
r
ed i
n
th
e PSO. Acco
rding to
the
s
e table
s
, it is indi
cate
d that
gene
rating
u
n
its u
s
e
vario
u
s
payment
s
for p
r
od
uc
i
n
g
po
wer outp
u
ts. Rega
rdin
g
in combin
atio
ns
of powe
r
outp
u
ts, gene
ratin
g
units have
been o
per
ate
d
usin
g different budg
ets.
Focu
se
d on t
he
total ope
ratin
g
cost, it h
a
s bee
n o
p
timized
e
c
on
omi
c
ally in
23,1
7
0.65 $/h
dete
r
mine
d by
bo
th
comp
utation
s
, althoug
h it
has spent
different to
tal
costs of fuel
con
s
um
ption
s
a
nd
polluta
nt
comp
en
sat
i
o
n
s.
6. Conclusio
n
This
paper compares
GA and HSABC while
solving
the EDED
usi
ng IEEE-62 bus as a
sampl
e
sy
st
e
m
.
Obt
a
ine
d
res
u
lt
s sh
ow
t
hat
bot
h
al
g
o
rithm
s
h
a
ve
different
ch
a
r
acte
ri
stics a
n
d
perfo
rman
ce
s for the EDED pro
b
lem. Its co
nver
g
e
n
c
e spee
ds a
r
e
smooth an
d quick to sele
ct
optimal sol
u
tions. Focused on the soluti
on quality
and the comput
ational efficiency, HSABC has
sea
r
ched fo
r
the optimal solution in the
fastest
spee
d and the
sh
ortest time. Numeri
cally, b
o
th
algorith
m
s h
a
v
e produ
ce
d simila
r re
sults.
Referen
ces
[1]
Mian
Xin
g
, Li
n
g
ji,
Baiti
ng
Xu. Multi
p
le
Ob
je
ctive
Optimiz
a
tions
for En
e
r
g
y
Ma
nag
em
ent S
y
stem
und
er Uncerta
i
nties.
Te
l
k
om
nika
. 2013; 1
1
(1
2): 7044-
70
51.
[2]
Yunzh
i
Ch
eng,
W
e
ipin
g
Xia
o
, W
e
i-Jen L
ee a
nd Min
g
Yan
g
(200
9).
A New
Appro
a
ch for
Missions
an
d
Security Co
nstrain
ed Eco
n
o
m
ic Dispatc
h
. NAPS, IEEE Conferenc
e Pub
l
i
c
ation. Starkvil
l
e
. 2009: 1-
5.
[3]
R Gopal
akrish
nan, A Krishn
an. A Novel
Co
mbi
n
e
d
Econom
ic and E
m
ission D
i
sp
atch Probl
e
m
Solvi
ng
T
e
chni
que usi
ng Non
-
domin
ated
Ra
nked G
enetic
Algorit
hm.
Eur
ope
an J
our
nal
of Scie
ntifi
c
Research.
20
1
1
; 64(1): 14
1-1
51.
[4]
K Sathish Kum
a
r, V
T
a
milselv
an, N Murali,
R Rajar
a
m, N Shanm
ug
a Su
ndar
am and T
Ja
yab
a
rathi.
Econom
ic Loa
d Dispatch
w
i
th Emission
Cons
tra
i
nts u
s
ing Vari
ous
PSO Algorith
m
.
WSEAS
T
r
ansactio
n
on
Pow
e
r System
. 2008; 9(3): 59
8-60
7.
[5]
AN Afandi. Opt
i
mal So
lutio
n
o
f
the EPED Problem
C
onsi
deri
ng Spac
e Area
s of HSABC on
the Po
w
e
r
S
y
stem Oper
ation.
Internati
o
n
a
l Jour
nal of E
ngi
neer
in
g and
T
e
chnol
ogy
. 2
015; 7(5): 1
824
-183
0.
[6]
AA El-Kei
b, H
Ma, and
JL H
a
rt. Environm
e
n
tall
y
Constra
i
ned E
D
us
ing
the La
gra
ngi
an
Rel
a
xatio
n
Method.
IEEE Trans. Power System
s
.
19
94
; 9(4): 1723-1
7
29.
[7]
Ahmed F
a
r
ag,
Samir Al-B
ai
yat, T
C
Che
n
g
. Econ
omic
Loa
d Dis
patch
Multio
bjectiv
e
Optimizatio
n
Proced
ures us
ing Lin
ear
Pr
o
g
rammin
g
T
e
chni
ques.
IEEE
T
r
ansactio
n
s
on Pow
e
r Syst
ems
. 199
5;
10(2): 73
1-7
3
8
.
[8]
Z
L
Gain
g. Part
icle s
w
a
rm
opti
m
izatio
n to s
o
l
v
ing t
he E
D
co
nsid
erin
g th
e g
ener
ator co
nstraints.
IE
EE
T
r
ans. Pow
e
r Systems.
20
03
; 18(3): 118
7-1
195.
[9]
K Cha
ndram,
N Subr
ahma
n
y
am, M S
y
dul
u. Secant
Met
h
o
d
for Econ
omi
c
Disp
a
tch
w
i
t
h
Gener
ato
r
Constra
i
nts a
n
d
T
r
ansmissio
n
Loss
e
s.
Jour
nal
of Electric
a
l
Eng
i
n
eeri
ng
a
nd T
e
ch
no
logy
. 200
8; 3(1):
52-5
9
.
[10]
H Ch
ahk
an
di
Neja
d, R
Ja
ha
ni, M M
oham
m
ad A
bad
i. G
APSO-Based
Econom
ic L
o
a
d
Dis
patch
o
f
Po
w
e
r S
y
stem.
Australia
n Jou
r
nal of Basic a
nd App
lie
d Sci
ences
. 20
11; 5
(
7): 606-6
11.
[11]
Xu
eso
ng Y
an,
W
e
i Li, W
e
i
Che
n
, W
enji
n
g
Luo, C
an Z
h
ang, Qin
g
h
ua
W
u
. W
e
ighted
K-Near
es
t
Neig
hb
or C
l
as
sificatio
n
Al
gor
ithm Bas
ed
o
n
Gen
e
tic Al
g
o
rithm.
Te
l
k
om
ni
ka
. 2
0
1
3
;
11(1
0
): 61
73-
617
8.
[12]
Mukesh Garg,
Surender Kumar. A Survey
o
n
Enviro
n
m
ental Eco
n
o
m
ic Loa
d Dis
patch usi
n
g
Lagr
ang
e Multi
p
lier M
e
tho
d
.
Internati
o
n
a
l Jo
urna
l of Electro
n
ics & Co
mmu
nicati
on T
e
ch
n
o
lo
gy
. 201
2;
3(1): 43-4
6
.
[13]
Arif N Afandi,
Hajim
e Mi
ya
uchi. Improv
e
d
Arti
ficial B
e
e Col
o
n
y
A
l
g
o
rithm Co
nsi
d
erin
g Harv
est
Seaso
n
for C
o
mputi
ng Eco
nomic D
i
spatc
h
on Po
w
e
r
S
y
stem.
T
r
an
saction o
n
El
ectrical a
n
d
Electron
ic Engi
neer
ing
. 2
014;
9(3): 251-
25
7.
[14]
AN
Afandi.
O
p
timal Sch
e
d
u
ling P
o
w
e
r Generati
ons us
in
g HSABC Al
g
o
rith
m Co
nsid
ered a N
e
w
Pena
lty F
a
ctor
Appr
oach
.
IEEE Co
nfere
n
c
e
o
n
Po
w
e
r
E
ngi
neer
in
g a
n
d
R
ene
w
a
b
l
e
Energ
y
. B
a
li.
201
4; 13-1
8
.
[15]
Ioann
is G Dam
ausis, An
astasi
os G Bakirtzis,
Petros S Dok
opo
ulos. N
e
t
w
ork Co
nstrain
e
d Econ
omic
Dispatc
h
usin
g
Real C
ode
d Genetic Al
gorith
m
.
IEEE
Trans. Power Systems
. 2003; 18(
1): 198-
205.
[16]
Sun Ho
ngb
in,
T
i
an Chu
ngu
ang. Optimizi
n
g
Mu
lti-a
gen
MicroGrid Res
ource Sch
e
d
u
l
i
ng b
y
C
o
-
Evoluti
onar
y w
i
th
Preference.
Te
lkom
n
i
ka
. 20
12; 11(1
2
): 722
2-72
29.
[17]
AN Afan
di, Ha
jime Mi
ya
uc
hi.
Solvi
ng
Com
b
in
ed Ec
on
omi
c
and
Emissi
o
n
Dis
patch
Us
ing
Harvest
Seaso
n
Artifici
al Be
e Co
lo
n
y
Algor
ithm Co
nsid
erin
g F
o
o
d
Source P
l
ac
e
m
ents an
d Mo
difie
d
Rates.
Internatio
na
l Journ
a
l on E
l
ect
r
ical En
gin
eeri
ng an
d Infor
m
a
t
ics
, 2014; 6(2)
; 266-27
9.
[18]
Yong F
u
, Mo
hamma
d Sha
h
id
ehp
our, Z
u
yi
Li. AC
Co
nting
enc
y D
i
s
patch Bas
ed
on Sec
u
ri
t
y
Constrained Unit Commitment.
IEEE
Transactions on Power System
s.
200
6; 21(2): 89
7-9
08.
Evaluation Warning : The document was created with Spire.PDF for Python.