TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol.12, No.5, May 2014, pp
. 3407 ~ 34
1
5
DOI: http://dx.doi.org/10.11591/telkomni
ka.v12i5.4953
3407
Re
cei
v
ed O
c
t
ober 2
1
, 201
3; Revi
se
d Decem
b
e
r
4, 2013; Accepte
d
De
cem
ber
23, 2013
Cost-Emission Scheduling under Uncertainty in a
Smart Grid with Wind Power and PHE
V
s
Zhang Xiaoh
u
a*
1,2
, Xie Jun
3
, Li Zhenkun
4
1
School of Infor
m
ation Sci
enc
e & Engin
eer
in
g,
Chan
gzh
ou
Univers
i
t
y
, C
h
a
ngzh
ou 2
131
6
4
, Chin
a
2
Jiangs
u Ke
y
L
abor
ator
y
of Po
w
e
r T
r
ansmissi
on a
nd Distri
butio
n Equ
i
pm
ent T
e
chnolo
g
y
3
Nanji
ng U
n
ive
r
sit
y
of Posts a
nd T
e
lecommu
nicati
ons, Nan
j
i
ng, 210
04
6
4
Shang
hai U
n
i
v
ersit
y
of El
ectric Po
w
e
r, Sha
ngh
ai, 20
009
0
Corresp
on
din
g
author, e-mai
l
: zhang
_8
103
0
1
@1
63.
com*, eej
xi
e@
gmai
l.com,
lzk021
@1
63.com
A
b
st
r
a
ct
T
he
rap
i
d de
velo
p
m
ent of
plu
g
-in el
ectric
vehic
l
es (P
HEVs) an
d w
i
nd
pow
er br
i
ngs n
e
w
challenges to
power system security and
econom
ic o
per
ation. Traditional dete
r
m
inist
i
c models fail
to
capture th
eir
e
x
tra character
i
stics. In this p
aper,
PHEVs, w
i
nd
pow
er an
d
ther
ma
l u
n
its
are stud
ie
d. T
he
sched
uli
ng
mo
del w
i
th PHEVs and w
i
nd po
w
e
r is more
co
mp
lex, w
h
ich mi
ni
mi
z
e
s the cost-emissio
n
w
h
i
l
e
consi
deri
n
g
th
e u
n
certa
i
nty
of w
i
nd
pow
er
an
d
lo
a
d
, th
e s
m
art c
har
g
i
ng/d
i
schar
gi
ng
of PHE
V
s, the
coord
i
nati
on of
w
i
nd pow
er a
nd PHEVs. T
he mu
lti-sce
nari
o
simul
a
tion is
prese
n
ted i
n
the rand
o
m
vari
a
b
l
e
discreti
z
at
ion.
Nu
mb
ers of re
prese
n
tative s
c
enar
ios is
c
h
osen, so t
hat the or
igi
nal
ob
j
e
ctive of th
e s
m
art
grid is
w
i
thin a
n
acce
ptab
le le
vel.
T
hen
the
mu
lti-ag
ent sys
tem (MAS) te
c
hno
logy
is pr
o
pose
d
to
divi
de
d
a
day is into 24 t
i
me interva
l
s, and eac
h time interval
is
man
age
d by a w
o
rk agent to prod
uce a soluti
on
set
for the
time
int
e
rval. T
h
e w
i
n
d
p
o
w
e
r, PHE
V
s an
d th
er
ma
l u
n
its
are c
o
o
r
din
a
ted
by t
h
e w
o
rk
age
nt.
24
w
o
rk age
nts ar
e
ma
nag
ed
a
coord
i
nat
i
o
n
a
gent th
at w
oul
d co
ordi
nate
t
he s
o
luti
ons
of
the w
o
rk
age
nts.
F
i
nally,
a s
m
ar
t grid
of 1
0
th
e
r
ma
l u
n
its, a
w
i
nd far
m
an
d P
H
EVs ar
e us
e
d
to
de
mo
nstra
t
e the
effective
of
the pro
pose
d
mo
de
l. T
he re
sults
show
tha
t
the smart gri
d
can us
e the
w
i
nd pow
er a
nd PHEVs
mo
st
effectively, ca
n
gre
a
tly c
u
t th
e o
per
ation
co
st and
car
bon
emissio
n
. By t
he tra
d
e
o
ff bet
w
een th
e w
e
i
g
ht
factor of cost and e
m
iss
i
on, th
e bal
anc
e of cost and e
m
issi
o
n
can reac
h.
Ke
y
w
ords
:
PHEVs, m
u
lti-scenario simulation,
MAS, cost-em
i
ss
ion
dispatching
Copy
right
©
2014 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
The po
we
r a
nd ene
rgy in
dustry in te
rm of eco
n
o
m
ic imp
o
rtan
ce a
nd envi
r
onmental
impact i
s
on
e of the mo
st importa
nt se
ctors in
th
e wo
rld, sin
c
e every asp
e
ct of indu
st
rial
prod
uctivity a
nd d
a
ily life
are
dep
end
e
n
t on
elect
r
ic
ity. It repre
s
e
n
ts a
majo
r
portion
of gl
o
bal
emission.
With in
cre
a
si
ng con
c
ern
over glo
bal
clim
ate
cha
n
ge, poli
c
y make
rs
are promoting
rene
wa
ble en
ergy, whi
c
h
i
s
con
s
id
ere
d
as a
m
ean
s
of meetin
g e
m
issi
on
re
du
ction ta
rg
ets.
So
environ
ment f
r
iendly
mode
rn di
spat
chin
g
is
esse
nt
ial. Ho
wever, po
wer
sy
stem rese
arche
r
s
h
a
ve
addresse
d o
n
ly tradition
al unit commit
m
ent (UC)
p
r
oble
m
s to
minimize cost in the exist
i
ng
article
s
. The
y
consi
der e
m
issi
on in UC pro
b
lem
s
rarely, thoug
h it is an important facto
r
as
mentione
d ab
ove.
A techni
cal
repo
rt fro
m
the National
Re
ne
wable
Energy La
borato
r
y (NREL) h
a
s
repo
rted
signi
ficant re
du
ctions in
CO
2
e
m
issi
on
s fro
m
PHEVs [1]. Con
s
ide
r
ing
co
st advanta
ges,
PHEVs h
a
ve
a sig
n
ifica
n
t
potential m
a
rket [2]. Be
ca
use
of its e
n
e
rgy
saving
p
o
tential, PHE
V
s
’
resea
r
ch an
d
appli
c
ation
has be
com
e
the fo
cus a
ttention of countrie
s
. T
h
e
co
rrespondi
ng
resea
r
chers
have mai
n
ly con
c
e
r
ne
d o
n
the inte
rconn
ection
of vehi
cle
ene
rgy
storag
e a
nd g
r
i
d
s.
Ahmed Yo
usuf Sabe
r
co
nsid
ers the
UC on
CO
2
emission
s of
V2G (Vehi
cle
to
G
r
id
), and
analyzes the
influen
ce of CO
2
emi
ssi
o
n
s an
d PHE
V
s discha
rg
e
in different situations; ele
c
tri
c
vehicle
s
(EV)
can
re
pla
c
e conve
n
tional small uni
ts fo
r po
we
r g
ene
ration, the
r
eb
y redu
cin
g
th
e
operation
cost and
emi
ssi
o
n
of
pollutant
s. But it
i
s
assume
d th
at th
e cha
r
gin
g
d
e
m
and
of EV
h
a
s
been
provide
d
by the
re
newable
ene
rgy, and
ch
argin
g
loa
d
cha
r
a
c
teri
stics of EV is
not
con
s
id
ere
d
[
3
]. Several o
t
her
resea
r
ch
efforts
of PHEVs in
re
cent years [4
-8] examine t
he
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 5, May 2014: 3407 – 34
15
3408
impact of PHEVs on the p
o
we
r syste
m
but do not
take
wind e
n
e
rgy into account and d
o
not
prop
ose ope
rational metho
d
s.
Ho
wever, PHEVs can’t co
mpletely solv
e the em
issio
n
pro
b
lem alo
ne; they need
electri
c
power, which
is one of ma
in sou
r
ce of emission.
Th
e NRE
L exa
m
ines th
e lon
g
-term i
n
tera
ctio
n
betwe
en
win
d
ene
rgy an
d
PHEVs [9] b
y
assuming i
n
crea
sing
pe
netration
of PHEVs
com
p
a
r
ed
with the
current vehi
cle f
l
eet for future yea
r
s.
Th
e effective
control
of V2
G cha
r
gin
g
,
the
formation
of
rene
wa
ble e
nergy
and P
H
EVs effe
cti
v
e are
co
mpl
e
menta
r
y [10
]. Literature [11]
take
s the
Da
nish
po
wer system a
s
an
example, a
n
a
l
yzes th
e
cha
r
ging
control f
o
r the
promot
ion
of wind po
wer to ab
sorb
and red
u
ce
green
hou
se
gas emi
s
sio
n
s. Wa
ng et
al. [12] uses a
determi
nisti
c
method to a
ddre
s
s coord
i
nation
of wi
nd po
we
r an
d PHEV cha
r
ging. Li
sa [
13]
investigate
s
consequ
en
ce
s of integrating
PHEVs in
a wind
-the
rmal
power sy
ste
m
. Four different
PHEV integration st
rategi
es, with
different im
p
a
ct
s
have be
en in
vestigated. T
he stu
d
y sh
o
w
s
that PHEVs can impa
ct the CO
2
emissio
n
. Soare
s
[14] analyze
s
PHEVs a
s
a way to maximize
the integratio
n of variable rene
w
able e
n
e
rgy in po
we
r system
s.
Determini
s
tic UC d
eal
s
with the
unit
gen
eratio
n
sched
ule i
n
a po
we
r
system. The
purp
o
se of
such
a sch
e
d
u
le is to
min
i
mize o
p
e
r
ati
on cost
s an
d emi
ssio
n
s
while
sati
sfying
prevailing constrai
nts such
as
load
bal
ance, sy
stem
spi
nning reserve, et al
ov
er a set of ti
me
perio
ds. Com
pare
d
with d
e
termini
s
tic UC and disp
atch
m
e
thod
s,
sto
c
ha
stic UC studi
es h
a
ve
been mo
stly perfo
rmed in
aca
demia. Lit
e
ratu
re [
15] and [16] develop a st
ocha
stic UC mo
del
to
study the im
pact
s
of PHEVs on po
wer sy
stem op
eration a
nd
sched
uling. T
he un
certai
nty is
addresse
d in the prop
osed
model by
gen
erating diffe
re
nt sce
nari
o
s.
Traditio
nal
UC only
can
d
i
spat
ch g
ene
rator
but not
load. Lo
ad
disp
atch
ca
n
play an
important rol
e
in reducing t
he
operation
cost of power system
by
shaving the pea
k and filling the
valley of load
profile
s. Th
e
su
cce
s
s of
pra
c
ti
cal
ap
pl
ication
of PHEVs greatly
depe
nd
s on
the
maximum utilization of re
newable e
n
e
r
gy in the
smart gri
d
so
that emissi
o
n
and cost
are
redu
ce
d. In this pap
er, th
e PHEVs, wi
nd power an
d thermal uni
ts are stu
d
ie
d, the unce
r
tainty
sma
r
t g
r
id di
spatchi
ng m
o
d
e
l is form
ulat
ed a
s
a
stoch
a
stic cost
-emi
ssi
on
red
u
cti
on mo
del. In
the
sched
uling, t
he fo
re
castin
g loa
d
a
nd
wi
nd p
o
we
r
are
used, b
u
t th
e a
c
tual
win
d
po
wer an
d l
oad
usu
a
lly differs from th
e foreca
sted
one
s. So t
he un
ce
rtainties of lo
ad an
d wi
nd
power a
r
e ta
ken
into accou
n
t. The PHEVs
charg
e
/disch
arge co
ntro
l, the coo
r
din
a
tio
n
of PHEVs and wi
nd po
wer
are
con
s
id
ere
d
. First, the
multi-sce
nari
o
simul
a
tion
i
s
u
s
ed i
n
the
rand
om vari
a
b
le di
scretization.
Numb
ers of repre
s
e
n
tative scena
rio
s
is
cho
s
e
n
, so
th
at the origi
nal
obje
c
tive of the sm
art g
r
id
is
within a
n
a
cceptable l
e
vel. Then
a d
a
y i
s
divide
d into
24 time i
n
tervals, and
ea
ch time inte
rva
l
is
manag
ed by
a wo
rk a
gent
to produ
ce
a solutio
n
se
t for the time interval. The
work ag
ent i
s
pre
s
ente
d
to
coo
r
din
a
te th
e wi
nd
po
we
r, PHEVs
and
therm
a
l
unit
s
. Th
e a
d
ju
stment of
wei
g
ht
factors can re
ach the effe
ctive coordinati
on between
CO
2
emi
ssi
on
s and
co
sts.
2. Stochas
t
i
c
Cos
t-emis
s
ion Redu
cti
on Model
2.1. Multi-sc
enario Simulation
In multi-scenario, a large number of discre
te
probability distributions are formed t
o
simulate the
uncertainty of rando
m varia
b
les.
It generally has two
steps to gen
erate scena
rio
s
.
The proba
bility distribution
of rando
m variabl
e i
s
obtai
ned by Monte
Carlo
simul
a
tion.
In order to minimize the information
loss
, the probability distri
bution of the random
variable i
s
dispersed by the
approxim
ate method.
Due to th
e st
och
a
sti
c
prop
erties of win
d
pow
er a
nd lo
ad, the wi
nd
power a
nd th
e load i
s
very difficult to predi
ct pre
c
isely. Und
e
r multi-
sc
en
ar
io
s
i
mu
la
tio
n
,
s
o
me
re
presentative discrete
scena
rio
s
are extracte
d for the o
p
timization i
n
a smart gri
d
wit
h
win
d
gen
erator an
d PHEV
s
unde
r un
ce
rtainty, as it is hard to
con
s
i
der a
ll
contin
uou
s state
s
. Ho
wever, the
total numbe
r of
scena
rio
s
gro
w
s exp
one
ntially
with state
variable.
For u
n
certain
t
y, discrete p
r
oba
bility
dist
ribution
set
s
for load
dem
and (
D
) a
nd
wind
r
e
sour
ce (
w
) are given as foll
ows:
11
2
2
{
(
,)
;
(
,
)
;
(
,)
;
(
,
)
}
ss
n
d
n
d
Dd
d
d
d
d
d
d
d
pp
p
p
(1)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Co
st-Em
i
ssio
n
Sched
uling
unde
r Un
ce
rt
ainty in a Sm
art Grid
with Wind
… (Zh
a
ng Xiaohu
a)
3409
(,
)
ss
dd
p
is load an
d the co
rrespon
ding proba
b
ili
ty of uncertai
n
load at sce
nario
s
;
nd
is the set of p
o
ssible
scena
rios d
e
rive
d from load.
12
1
nd
dd
d
(2)
11
22
{
(
,)
;
(
,)
;
(
,
)
;
(
,
)
}
s
s
nw
nw
w
w
in
d
w
in
d
w
in
d
w
ind
p
ppp
(3)
(,
)
ss
wi
nd
p
is wind and the corresponding probab
ility of uncertain wind at scenario
s
;
nw
is the set of possibl
e scen
ario
s de
rived
from win
d
po
wer.
12
1
nw
(4)
SC
is a set of po
ssi
ble sce
nari
o
s de
rived fro
m
wind po
we
r and loa
d
.
Dw
SC
(5)
1
d
sS
C
(6)
sd
w
(7)
D
,
w
are sets
of discrete
distrib
u
tion
of load, wi
nd po
we
r;
d
,
are the
corre
s
p
ondin
g
prob
ability of unce
r
tain
load, win
d
;
s
is the co
rrespondi
ng prob
ability of the
sma
r
t gri
d
system at
scen
ario
s
. Differe
n
c
e
between t
he
scena
rio
model
and
th
e ori
g
inal
mo
del
is a
di
screte
prob
ability di
stributio
n a
d
o
p
ted. Cu
rve
s
rep
r
e
s
entin
g
the ori
g
inal
probability de
n
s
ity
distrib
u
tion, rectan
gula
r
b
a
rs
represent
the
scena
ri
os; the
re
ct
a
ngula
r
b
a
r
h
e
ight rep
r
ese
n
ts
prob
ability of
co
rrespon
di
ng
scena
rio.
Becau
s
e
of wind
po
we
r and
l
oad un
certainty,
and
EV
cha
r
gin
g
/dischargi
ng in
sm
art gri
d
co
ntrol, so t
he trad
itional optimi
z
ation problem
is tran
sfo
r
me
d
into uncertai
nty smart gri
d
di
spat
ching. To
c
apture volatility,
we assu
me the
wind power
and
load a
r
e
subj
ect to the di
stribution
2
(,
)
N
with
their expe
cte
d
value (
) and their volatilit
y (
).
Five scen
ario
s a
r
e
con
s
id
e
r
ed fo
r the
wi
nd po
we
r
a
n
d
load
uncertai
n
ty, the scen
ario
distri
buti
on
of wind po
we
r and loa
d
are sho
w
n in Fi
gure
1
and 2
resp
ectively.
{
(
1
00%
,
0
.
5
)
;
(
99%
,0
.
1
5
)
;
(
10
1%
,0.
1
5
)
;
(
9
7.
5
%
,
0
.
1
)
;
(
1
02
.
5
%
,
0
.
1
)
}
ww
d
d
dd
pp
p
pp
(8)
{
(
100%
,
0
.
6
)
;
(
9
8.
5%
,
0
.
1
5
)
;(
1
02%
,
0
.
1
5
)
;
(
98%
,
0
.
0
5
)
;(
103%
,
0
.
0
5
)
}
Dd
d
d
dd
pp
p
pp
(9)
w
p
,
d
p
are the pre
d
i
ct value of wi
nd po
wer a
n
d
load.
Figure 1. The
Scena
rio Di
stribution of Wi
nd Powe
r
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 5, May 2014: 3407 – 34
15
3410
Figure 2. The
Scena
rio Di
stribution of Lo
ad
2.2 Cos
t-emi
ssion Re
duc
tion Model u
nder Un
cer
tainties
A quad
rati
c
function
is consi
dered fo
r the fuel
fun
c
tion
of the
r
mal unit
s
u
n
der t
he
determi
nisti
c
ca
se:
2
()
(
)
tt
t
ii
i
i
i
i
i
FC
P
a
b
p
c
p
(10)
Con
s
id
erin
g t
he u
n
certaint
y of load
and
win
d
po
we
r,
the fuel
co
st f
unctio
n
i
s
co
nverted
into the scen
ario mo
del:
2
[(
)
,
]
[
(
)
,
]
ts
t
s
t
ii
s
i
i
i
i
i
s
FC
P
a
b
p
c
p
(11)
st
i
p
is the po
wer of the
r
ma
l unit
i
at time
t
co
nsi
d
e
r
ing
scena
rio
s
,
s
is the
corre
s
p
ondin
g
pro
bability
;
i
a
,
i
b
,
i
c
are
cost coefficie
n
ts of unit;
i
.It is assum
ed tha
t
conve
n
tional thermal units
are co
al-fire
d
.
A
quad
rati
c fun
c
tion i
s
con
s
id
ere
d
f
o
r the
emi
ssi
on
c
u
rve [17] as
follows
:
(
2
ci
[(
)
,
]
[
(
)
)
,
]
st
s
t
st
t
i
s
ci
ci
i
c
i
i
i
s
EP
p
p
u
(12)
ci
,
ci
,
ci
are
CO
2
e
m
issi
on
co
ef
f
i
cient
s of
u
n
it
i
.Therefore, th
e obje
c
tive fu
nction
for
co
st-emi
ssion
optimization
con
s
id
erin
g a
set of sce
na
rios
s
in a s
m
art grid is
:
(
21
11
2
mi
n
[
(
(
)
)
(
1
)
()
)
]
TN
ss
t
s
t
t
t
t
sc
i
i
i
i
i
i
i
i
i
sS
t
i
st
s
t
t
ec
i
c
i
i
c
i
i
i
TC
W
a
b
p
c
p
u
S
u
u
Wp
p
u
(13)
t
i
u
is de
cisi
on variable of unit
i
at time
t
, 1 fo
r up, 0 for do
wn;
i
S
is start-u
p
co
st of unit
i
.
N
is total numbers of the
r
mal unit
s
;
T
is numbe
rs
of perio
ds
unde
r study;
c
W
,
e
W
is the
weig
ht factor
of operatio
n cost (f
uel co
st
plus startup cost),
CO
2
emissi
on;
c
W
+
e
W
=
1
(14)
Con
s
trai
nts:
PHEVs a
r
e
consi
dered
as load
s o
r
sou
r
ce
s. Po
we
r
sup
p
lied f
r
om
distri
buted
g
eneration
s
m
u
st
satisfy the loa
d
deman
d:
PHEVs
disc
harging
2
1
1,
2
,
,
N
tt
t
t
ii
v
v
G
d
i
p
up
N
p
t
T
(15)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Co
st-Em
i
ssio
n
Sched
uling
unde
r Un
ce
rt
ainty in a Sm
art Grid
with Wind
… (Zh
a
ng Xiaohu
a)
3411
PHEVs
charging
2
1
1,
2
,
,
N
tt
t
t
ii
d
v
v
G
i
p
up
p
N
t
T
(16)
All registe
r
ed
PHEVs take part in sm
art
gr
id op
eration
s
duri
ng a scheduli
ng pe
ri
od,
ma
x
22
1
1,
2
,
,
T
t
vG
vG
t
NN
t
T
(17)
ma
x
2
vG
N
is the total re
gistered PHE
V
s;
2
t
vG
N
is num
ber
of vehicle
s
co
nne
cted to the grid at
hour
t
To main
tain system reliability, adequat
e spinni
n
g
reserve
s
are requi
re
d:
PHEVs
disc
harging
ma
x
2
1
1,
2
,
,
N
tm
a
x
t
t
t
ii
v
v
G
d
i
up
p
N
p
R
t
T
(18)
PHEVs charg
i
ng
ma
x
2
1
1,
2
,
,
N
tt
m
a
x
t
t
ii
d
v
v
G
i
up
p
p
N
R
t
T
(19)
ma
x
i
p
is the maximum output limit of unit;
i
,
ma
x
v
p
is the capa
city of PHEVs;
t
d
p
is
system
dema
nd at time
t
;
t
R
is sy
stem
spi
nning
re
se
rve req
u
iremen
t at time
t
;
ma
x
i
p
/
mi
n
i
p
is
maximum/ minimum ge
neration level of unit
i
;·Num
ber of charging/d
i
scharging P
H
EVs limit.
ma
x
22
1,
2
,
,
tt
vG
v
G
NN
t
T
(20)
All the PHEVs can
not ch
a
r
ge/di
scha
rge
at
the same time. For reliable ope
rati
on and
control, limite
d
num
be
r of
vehicl
es
will
ch
arg
e
/discharg
e
at
a ti
me.
ma
x
2
t
vG
N
is the
maximum
numbe
r of ch
argin
g
/disch
a
r
ging at ho
ur
t
.
Gene
ration li
mits, ramp
rate, minimu
m up and
down time con
s
trai
nts a
r
e al
so
c
o
ns
ide
r
ed
.
3. Proposed
Solution Ap
proach
The total
sch
edulin
g pe
rio
d
is
24h, a
n
d
it co
ntain
s
24 work a
g
e
n
ts in th
e scheduli
ng
perio
d. Each
work a
gent u
s
e
s
gen
etic a
l
gorithm to p
r
odu
ce a
solut
i
on set fo
r the time interv
al.
24
wo
rk ag
en
ts a
r
e
man
a
g
ed by
a
co
op
erative
agent
that
woul
d
coordi
nate
the
sol
u
tion
s of
the
work
age
nts.
The
relatio
n
s
hip
amo
ng
all the a
gent
s is sho
w
n i
n
Error
!
Re
ference source
not
fou
nd.
.
As
shown
in the
figure, except relating wi
th the
coo
r
dination
age
nt every work ag
ent
had info
rmati
on exchan
ge
d with the
previous
and
f
o
llowin
g
adj
a
c
ent
work a
g
ents. Ea
ch
work
agent i
s
resp
onsi
b
le for coordi
nating t
he stati
c
sch
edulin
g of wi
nd po
we
r, th
ermal
units
a
nd
PHEVs, their
relation
shi
p
is sh
own in Figure 4. Its
go
al is the mini
mum of fuel consumption a
nd
emission
s in
this peri
od, the co
nstraint
s are static f
o
r the co
rre
s
pondi
ng time
interval, without
con
s
id
erin
g the dynami
c
t
i
me co
uplin
g
con
s
trai
nts.
Then the
ge
netic al
gorith
m
is u
s
ed. T
he
target of
the
coo
perative a
gent i
s
the
m
i
nimum
of co
st an
d emi
s
si
ons for th
e
whole
sche
duli
n
g
cycle, the con
s
traint
s are the
dynamic
co
upling
con
s
traints on the e
n
tire sche
duli
ng peri
od.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 5, May 2014: 3407 – 34
15
3412
A1
…
…
Co
op
e
r
a
t
i
v
e
ag
e
n
t
A2
A3
A4
A2
3
A2
4
Figure 3. MAS Archite
c
ture of the Smart Grid
Opt
i
mal Di
sp
at
chin
g
Figure 4. The
Work Agent
Synergi
stic Effect
Diag
ram
4. Numerical
Example
An inde
pend
ent syste
m
o
perato
r
of
10
-unit
sy
stem
is con
s
ide
r
ed
for si
mulatio
n
wit
h
wind po
we
r a
nd 5000
0 PHEVs. Load d
e
mand a
nd u
n
it c
haracte
ri
stic of the 10
-unit syste
m
are
colle
cted fro
m
[18]. Assu
me the re
se
rve to be
10
% of the load dema
nd. It is ne
ce
ssary to
integrate
win
d
in the sust
ainabl
e sm
art
grid to
redu
ce co
st
and emission.
T
h
e amou
nt of co
st
and e
m
issio
n
red
u
ctio
ns
mainly de
pen
ds
on m
a
xi
mum utilization
of re
ne
wabl
e ene
rgy th
ro
ugh
PHEVs. PHEVs are
c
harging/di
scharging i
n
telligently so
that both cost
and
emi
s
sion are
minimum. Lo
ad dem
and
and con
s
trai
nts are fulf
illed. Maximu
m battery capa
city=25
k
Wh,
minimum bat
tery capa
city=10
k
Wh, averag
e ba
ttery capacity
=
1
5
kWh, maximum numb
e
r of
cha
r
gin
g
/dischargi
ng P
H
E
V
s at e
a
ch h
our,
ma
x
2
t
vG
N
=
10% tot
a
l PHEVs
. Total number
of
PHEVs
in
the system,
ma
x
2
vG
N
=50
000. Cha
r
ging
-di
s
cha
r
ging fre
que
n
c
y=1
per
day
; sch
edulin
g
perio
d=24h,
depa
rture state
of
cha
r
ging/di
scharg
i
ng
=50%, efficien
cy
=85%. A PHEV need
s
8.22kWh/d
ay, an excess
of 8.22*50
000
=411M
Wh
po
wer will
be n
eede
d for th
e
sma
r
t gri
d
[1
9].
And the win
d
farm can p
r
ov
ide 500
MWh/day ene
rgy.
A typical day forecast
s of
wind a
r
e give
n in
[20]. This p
a
per
analyzes two
ca
ses,
one d
o
e
s
not
con
s
id
er th
e
uncertai
n
ty of load a
nd
wind
power, the other con
s
ide
r
s the unce
r
tain
ty of load and wind po
we
r for sm
art gri
d
.
1) Co
st-emission redu
ctio
n disp
atchi
n
g
without
the u
n
ce
rtainty of load an
d win
d
powe
r
Co
st-emi
ssio
n
re
du
ction weig
hts ca
n give
de
ci
sion
-ma
k
ers th
e i
n
tuitive analy
s
is of the
con
c
e
r
ne
d fa
ctors. Th
e eff
e
ct of th
e
wei
ght
chan
ge
s
on the
optimi
z
ation
sch
e
d
u
ling i
s
a
naly
z
ed
below.By this
way, it verifies
the effec
t
ivene
ss of the
co
st-emi
ssion
redu
ction mo
del.
CO
2
is
on
e
o
f
the
mai
n
di
scharge
in
th
e electri
c
po
wer p
r
od
ucti
on pro
c
e
s
s, it has a
signifi
cant im
pact o
n
the e
n
vironm
ent. The relation
ship of therm
a
l co
st-emi
ssio
n obje
c
tives
and
weig
hts with
o
u
t/with PHEVs ca
n be see
n
in Table 1, 2.
Table 1. The
Relatio
n
ship of Therm
a
l Cost-e
mi
ssio
n Obje
ctives an
d Weig
hts wit
hout PHEVs
we
ig
ht
s
obj
ec
tiv
e
(1
,0)
(0
.9
, 0.1)
(0
.8
, 0.2)
(0
.7
, 0.3)
(0
.6
, 0.4)
(0
.5
, 0.5)
(0
.4
, 0.6)
(0
.3
, 0.7)
(0
.2
, 0.8)
(0
.1
, 0.9)
F
/
$
5
628
77
.68 5
652
23
.52 5
652
77
.32 5
660
47
.70
5
671
42
.24
5
696
50
.16
5
713
98
.77
5
735
09
.10 5
749
78
.12 5
806
65
.09
E
c
/t
2
699
06
.39 2
587
51
.20 2
585
11
.43 2
561
40
.50
2
541
10
.02
2
511
07
.34
2
496
87
.99
2
486
11
.43 2
480
61
.91 2
472
06
.69
Table 2. The
Relatio
n
ship of Therm
a
l Cost-e
missio
n Obje
ctives an
d Weig
hts wit
h
PHEVs
we
ig
ht
s
obj
ec
tiv
e
(1
, 0)
(0
.9
, 0.1)
(0
.8
, 0.2)
(0
.7
, 0.3)
(0
.6
, 0.4)
(0
.5
, 0.5)
(0
.4
, 0.6)
(0
.3
, 0.7)
(0
.2
, 0.8)
(0
.1
, 0.9)
F
/
$
5
582
96
.90 5
588
20
.94 5
633
74
.18 5
669
55
.17
5
709
39
.72
5
768
05
.18
5
853
59
.22
5
954
55
.63 6
010
28
.84
6
274
16
.16
E
c
/t
2
733
26
.41 2
659
99
.73 2
392
29
.49 2
285
02
.65
2
217
47
.58
2
143
12
.78
2
069
59
.47
2
021
41
.48 2
003
50
.52
1
972
59
.18
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Co
st-Em
i
ssio
n
Sched
uling
unde
r Un
ce
rt
ainty in a Sm
art Grid
with Wind
… (Zh
a
ng Xiaohu
a)
3413
From
Tabl
e
1 an
d 2, the
weig
ht fact
ors of
co
st and emissio
n
are
(1,0), (0.9,0.1),
(0.8,0.2), (0.7,0.3), (0.6,0.4), (0
.5,0.5), (0.4,0.6), (0.3,
0
.7), (0.2
,0.8), (0.1,0.9) respec
tively. With
the wei
ght fa
ctor
of the o
peratio
n
co
st
c
decre
asi
ng,
co
st is i
n
cre
a
sin
g
, but th
e variatio
n is
small whi
c
h can be
a
c
ce
pted.
Increa
si
ng
the weig
h
t
factor
e
, CO
2
emission
ca
n be
red
u
ce
d
sub
s
tantially.
When
(
c
,
e
)
is
(0.8,0.2), the
operatio
n cost is 56
337
4.18$, CO
2
e
m
issi
on is
2392
29.49t (Table 2). On
the other han
d, when PHE
V
s are not co
nsid
ere
d
in the same
syste
m
,
the ope
ration
co
st is 56
52
27.32$,
CO
2
emission i
s
2
5851
1.43t in
the sam
e
sy
stem (Tabl
e 1
)
.
PHEVs save 1853.1
4$ an
d
redu
ce 19
28
1.94t emissio
n
. Compa
r
e
d
with Table 1,
CO
2
emi
ssi
on
s
sub
s
tantially redu
ce
in
T
a
ble
2 with others we
ig
hts.
It sho
w
s tha
t
the sche
dul
ing with
PHE
V
s
can
effectivel
y red
u
ce the
differe
nce b
e
twee
n p
e
a
k
and
valley
p
o
we
r
system,
save
co
sts
and
redu
ce
emission, in
crea
se the com
p
rehen
sive
be
nefit in the 10-u
n
it thermal syste
m
. By
cho
o
si
ng pro
per wei
ght
f
a
ctors
of co
st
an
d
e
m
ission o
n
the
b
a
si
s of th
e
deci
s
io
n-m
a
kers’
willingn
ess, satisfacto
ry scheduli
ng re
su
lts of
coo
r
din
a
ting co
st an
d emission
ca
n be rea
c
h
ed.
The rel
a
tion
ship of sma
r
t grid co
st-emission o
b
je
ctive and wei
ghts
with PHEVs a
nd
wind p
o
wer is sho
w
n in Ta
ble 3. PHEVs optimal
cha
r
ge/disch
a
rg
e power un
de
r the determini
stic
load an
d win
d
power with
weig
hts (0.9,
0.1) is
sho
w
n
in Figure 5
Table 3. The
Relatio
n
ship of Smart Grid
Cost
-emi
ssi
o
n
Obje
ctives
and Weight
s
with PHEVs
and Wind Power
we
ig
ht
s
obj
ec
tiv
e
(1
, 0)
(0
.9
, 0.1)
(0
.8
, 0.2)
(0
.7
, 0.3)
(0
.6
, 0.4)
(0
.5
, 0.5)
(0
.4
, 0.6)
(0
.3
, 0.7)
(0
.2
, 0.8)
(0
.1
, 0.9)
F
/
$
5
488
91
.79
5
495
46
.20
5
542
52
.04 5
573
02
.88
5
628
02
.51
5
701
60
.03
5
721
50
.96
5
808
46
.21 6
000
41
.87
6
168
46
.89
E
c
/t 2
727
20
.36
2
603
72
.43
2
328
46
.27 2
231
81
.17
2
131
11
.80
2
064
24
.36
2
046
99
.40
2
008
16
.72 1
951
08
.10
1
934
41
.20
Effect of both co
st and e
m
i
ssi
on in the
d
e
te
rmini
s
tic
model
with PHEVs a
nd wi
nd po
we
r
is
sho
w
n
in
Table
3.
Co
mpared
with
Table
2, in
th
e same
weig
hts of
cost
s
and
emi
ssio
n
,
the
operation
co
sts a
nd emi
s
sion a
r
e
rapi
dly decrea
s
in
g (Ta
b
le 3
)
; co
st is redu
ced by 912
2.1
4$,
and emi
ssi
on
is redu
ce
d by 6383.22t in the weight
of (0.8,0.2) (Ta
b
le 3). Com
p
ared
with Ta
ble
1, cost is re
d
u
ce
d rapi
dly, emissi
on increa
se
s sl
o
w
ly
((1,0), (0.9,0
.1)); the co
st is redu
ced b
y
1102
5.28$, t
he emi
s
sion
i
s
redu
ce
d by
256
65.15t
((
0.8,0.2)) . Bo
th the
cost
a
nd emi
s
sion
are
redu
ce
d in
T
able 3
than
those of T
abl
e 1 a
nd
2.
Prope
r u
s
in
g
of PHEVs a
nd
wind
po
wer,
PHEVs
ca
n
charg
e
from th
e g
r
id
with
wi
nd p
o
wer at
off-pea
k
hou
rs a
n
d
di
scharge to
the
grid
at
pea
k hou
rs,
whi
c
h are co
mpleme
ntary for each othe
r.
Figure 5. PHEVs Optimal
Cha
r
ge/
Disch
a
rge Po
we
r u
nder the
Dete
rmini
s
tic Lo
a
d
and Wi
nd
Power
As you can
see from Figu
re 5, PHEVs chargi
ng in the
low load p
e
ri
od, in the pea
k load
stage di
scha
rge, cha
r
ge/di
scharge po
wer thro
ugh ef
f
e
ctive co
ntrol
of PHEVs, can reali
z
e the
minimization
of cost
-emi
ssi
on in the sma
r
t grid.
2) Co
st-emission redu
ctio
n disp
atchin
g
co
n
s
ide
r
ing
the uncertain
ty of load and wind
power
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 5, May 2014: 3407 – 34
15
3414
The relatio
n
s
hip
of co
st-emi
ssi
on ob
jectives
an
d
wei
ghts wit
h
the
un
cert
ainty of
PHEVs/PHE
Vs and wi
n
d
powe
r
ca
n be see
n
in Table 4 a
nd Table 5.
PHEVs optimal
cha
r
ge/di
sch
a
rge
po
we
r u
nder the u
n
certainty of lo
a
d
and
PHEVs with the
wei
ghts
(0.9, 0.1
)
i
s
s
h
ow
n
in
F
i
gu
r
e
6
.
Table 4. The
Relatio
n
ship of Therm
a
l Cost-e
miss
io
n Obje
ctive and
Weight
s with
the Uncertai
nty
of PHEVs
we
ig
ht
s
obj
ec
tiv
e
(1
,0)
(0
.9
,0
.1) (0
.8
,0
.2) (0
.7
,0
.3) (0
.6
,0
.4) (0
.5
,0
.5) (0
.4
,0
.6) (0
.3
,0
.7) (0
.2
,0
.8) (0
.1
,0
.9)
F
/
$
5
782
23
.53 5
789
36
.68 5
823
29
.60 5
864
81
.65
5
922
18
.67
5
992
84
.96
6
052
92
.07
6
158
94
.74 6
364
82
.46 6
472
22
.85
E
c
/t
2
801
08
.44 2
685
64
.82 2
489
20
.36 2
366
54
.97
2
256
03
.96
2
175
72
.76
2
129
67
.85
2
062
50
.55 2
029
43
.65 1
987
71
.44
Table 5. The
Relatio
n
ship of Cost
-emi
ssion
Obje
ctive and Weight
s with the Un
ce
rtainty of
PHEVs and
Wind Po
we
r
we
ig
ht
s
obj
ec
tiv
e
(1
, 0)
(0
.9
, 0.1)
(0
.8
, 0.2)
(0
.7
, 0.3)
(0
.6
, 0.4)
(0
.5
, 0.5)
(0
.4
, 0.6)
(0
.3
, 0.7)
(0
.2
, 0.8)
(0
.1
, 0.9)
F
/
$
5
686
47
.30 5
693
99
.55 5
731
36
.83 5
773
66
.93
5
825
08
.40
5
898
49
.29
5
960
51
.48
5
984
41
.91 6
179
21
.86 6
358
98
.43
E
c
/t
2
749
33
.62 2
638
43
.80 2
421
64
.38 2
295
76
.25
2
199
53
.49
2
112
18
.50
2
062
18
.27
2
052
42
.61 1
982
09
.91 1
964
59
.40
Table 4 shows the re
sults
of
cost and e
m
issi
on when
only PHEV is con
s
ide
r
ed,
simila
rly
Table 5 sho
w
s the resul
t
s of cost
s and emi
ssi
o
n
whe
n
both
PHEV and wind p
o
we
r
are
con
s
id
ere
d
. Comp
ared wi
th the re
sults of Table 4,
t
he cost an
d
emission
are
cut with the
same
weig
hts (T
abl
e 5). Wi
nd po
wer
save
s 91
92.77$, an
d
redu
ce
s 675
5.98t with the
weig
ht (0.8,0.
2
).
Becau
s
e
of the un
ce
rtaint
y, the sy
stem
co
st and e
m
i
ssi
on
s are in
cre
a
sed, but i
t
is clo
s
e
r
to the
actual situatio
n.
PHEVs
can
red
u
ce de
p
ende
nci
e
s
o
n
sm
all
exp
ensive
units in existin
g
system
s,
resulting in
redu
ced o
p
e
r
ation co
st an
d emissi
o
n
. It can al
so in
crease re
se
rve
and reli
abilit
y
o
f
existing po
we
r system
s.
Figure 6. PHEVs Optimal
Cha
r
ge/
Disch
a
rge Po
we
r u
nder the
Un
certainty of Load and
Wind
Power
5. Conclusio
n
Wind
po
wer
and PHEV g
r
id-conn
ecte
d
cap
a
city
exp
ansi
on ha
s b
e
com
e
an i
n
evitable
trend, will exe
r
t a far-rea
c
hi
ng influen
ce
on po
we
r
system. To bring
the oppo
rtuni
ty to power g
r
id
co
st-emi
ssion
redu
ction o
peratio
n, ran
dom
an
d loa
d
dema
nd a
nd output of
wind p
o
wer
has
increa
sed the
difficulty of sche
duling.
The
optimal
sched
uling
wi
th win
d
p
o
we
r,
PHEVs an
d conventio
n
a
l
thermal un
its
un
der
uncertainty is presented in
this pape
r to
illustrate
cost and emissio
n
redu
ction
s
. The un
ce
rtai
nty
of wind p
o
we
r and lo
ad, P
H
EVs
cha
r
ge
/discharge
co
ntrol, the coo
r
dinatio
n of PHEVs an
d wi
nd
power
i
s
co
n
s
ide
r
ed. The
multi-sce
nari
o
sim
u
lati
on
i
s
u
s
e
d
for a
c
comm
odatin
g
the volatility
o
f
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Co
st-Em
i
ssio
n
Sched
uling
unde
r Un
ce
rt
ainty in a Sm
art Grid
with Wind
… (Zh
a
ng Xiaohu
a)
3415
wind
po
we
r
a
nd lo
ad. T
h
e
n
the
MAS i
s
used
to g
e
n
e
rate
a
su
cce
ssful
sch
edul
e. It co
ntain
s
24
work ag
ents
in the sched
uling pe
riod.
Each work
a
gent uses g
e
netic alg
o
rith
m to produ
ce a
solutio
n
set f
o
r the time i
n
terval. Win
d
power, PHE
V
s and th
ermal units
are
coo
r
din
a
ted.
24
work
age
nts
are
man
age
d
by a
co
ope
rative age
nt that would
co
ordin
a
te the
solutio
n
s of t
h
e
work
age
nts.
Valid sce
n
a
r
i
o
s
are
de
rive
d from
p
r
ior
s
t
at
ist
i
cs,
heu
ri
st
ic
s
a
nd th
e
experie
nce. T
he
results
sho
w
that the alg
o
rithm i
s
an
efficient
ap
proach an
d the
solutio
n
is
reasona
ble. T
h
is
optimizatio
n
with un
ce
rtai
nties for
sch
edulin
g
nee
d
s
mo
re cost
and lon
ger
executio
n time;
however, it is more reliabl
e in real envi
r
o
n
ment.
Ackn
o
w
l
e
dg
ement
s
This work is
sup
porte
d by National Natural Scie
nce Found
ation o
f
China (G
ran
t
No: 512070
74
)
and the
project of Ji
ang
su Key la
boratory of
po
wer tra
n
smissi
on & di
stribu
tion equi
pme
n
t
tec
h
nology (Grant
No: 2011J
SSPD10).
Referen
ces
[
1
]
Lunz B,
Yan Z
X
,
Gerschl
e
r JB,
et
al.
I
n
f
l
uen
ce of
pl
u
g
-in
h
y
br
id e
l
ect
r
ic v
ehicl
e ch
argi
ng
st
rat
egies o
n
charg
i
ng a
nd b
a
t
t
e
r
y
d
egra
dat
ion cost
s.
Ener
gy pol
ic.
201
2;
46:
511-
51
9.
[2]
Parks K, Denholm P, Mark
el
T.
Cost
s and
emissio
n
s
ass
o
ciat
e
d
w
i
t
h
p
l
ug-i
n
hy
brid
el
ect
r
ic veh
i
cl
e
charg
i
ng i
n
t
he xcel en
ergy
Colora
do ser
v
ice t
e
rrit
o
ry.
Color
a
d
o
:
Nat
i
ona
l Ren
e
w
a
b
l
e Ener
g
y
Lab
orat
or
y.
20
07.
[
3
]
Saber AY,
Ve
na
ya
gamo
o
rt
h
y
GK.
I
n
t
e
llig
e
n
t
unit
commit
m
ent
w
i
t
h
veh
i
cle-t
o
-gr
i
d-A c
o
st
-emissio
n
opt
imiz
at
ion.
J
ourn
a
l of
Pow
e
r Sources.
201
0;
195(1):
8
98-
911.
[
4
]
Hadl
e SW,
T
s
vet
k
ova A.
P
o
t
ent
ial
imp
a
ct
s of
pl
ug-i
n
h
y
brid
el
ect
r
ic v
ehicl
es
on r
e
g
i
on
al
po
w
e
r
gen
erat
io
n.
ORNL/
T
M-2007/
1
50.
200
8.
[5
]
K M
e
y
e
r, KK Sch
n
e
id
e
r
,
R
Pra
tt. Im
p
a
c
ts a
sse
ssm
e
n
t
o
f
p
l
u
g
-
in
hy
brid veh
i
cle
s
on
ele
c
tric u
t
ilitie
s
a
nd
regi
ona
l U.
S.
po
w
e
r gri
d
s part
1:
T
e
chnical a
nal
ysis.
Pac
i
f
i
c
Nort
hw
est
Lab
orat
ory.
200
7.
[
6
]
Zhong
WZ,
Hao Y,
Y
i
ng
L
L
.
P
aramet
ric
mat
c
hin
g
of
dr
ivet
rai
n
f
o
r pa
ralle
l h
y
bri
d
e
l
ect
r
ic
ve
hicl
e.
TELKOMNI
KA I
ndon
esi
an Jou
r
nal of
Elect
r
ic
al Eng
i
ne
eri
n
g
.
2013;
1
1
(10):
578
9-57
96.
[7
]
Isla
m
FR
, Po
t
a
H
R
.
PH
EVs Pa
rk a
s
Virtua
l U
P
FC
.
TEL
K
OMNI
KA I
n
d
ones
ian
Jo
urn
a
l of
E
l
ect
r
ic
al
Engi
neer
in
g
.
2012;
10(
8):
228
5-22
94.
[
8
]
Siosh
ansi
R,
F
agi
ani
R,
Mar
a
no V.
C
o
st
a
n
d
emissi
ons
imp
a
ct
s of
pl
ug-
in
h
y
bri
d
ve
hicl
es
on t
h
e Oh
io
po
w
e
r s
y
st
em.
Energy p
o
licy.
201
0;
38:
670
3
-
671
2.
[9]
NREL, plug-
i
n
hy
br
id electric vehicl
es
and
w
i
nd ener
gy
. Av
ailabl
e: http://www
.
n
rel.gov/
ana
l
y
sis/
w
i
nds/
pdf
s/
w
i
n
d
_
phe
v_post
e
r.
p
d
f
[
10]
Willet
t
K,
Jasn
a T
.
Vehicle-t
o
-grid
po
w
e
r im
plem
ent
at
io
n:
From st
abiliz
in
g t
he gr
id t
o
s
u
pport
i
ng l
a
rg
e-
scale re
ne
w
a
b
l
e ener
g
y
.
Jo
ur
nal of
Pow
e
r Sources
.
20
05;
144(
1):
280-
29
4.
[1
1
]
H
e
n
r
ik L
,
Willett K. In
te
g
r
a
t
ion
o
f
re
ne
w
a
b
l
e e
n
e
r
gy
in
to the
tra
n
s
p
o
r
t
a
n
d
ele
c
tricit
y
secto
r
s th
ro
ugh
V2G.
Energy Policy
.
20
08;
36(
9):
3578-
35
87.
[
12]
Wang J,
Liu C,
T
on D,
et
al.
Impact
of
plug-
i
n
h
y
bri
d
elect
r
i
c
vehicl
es on p
o
w
e
r s
y
st
ems w
i
t
h
dem
and
respo
n
se
a
nd w
i
nd po
w
e
r.
Energy p
o
licy.
2
011;
39(
7):
4
0
1
6
-40
21.
[13]
Lisa G, Sten K
,
Filip J. Integr
ation of
plug-in hy
br
id
el
ectr
ic vehicles
i
n
a r
egi
onal
w
i
nd-thermal
po
w
e
r
sy
s
t
e
m
.
Ener
g
y
Policy.
201
0;
38(1
0
):
548
2-5
492.
[
14]
Soares MC,
Br
uno B,
Ale
x
an
dre S,
et
al.
Plug-i
n
h
y
bri
d
el
ect
r
ic vehic
l
es
as a
w
a
y t
o
maximize t
h
e
int
egr
at
ion
of
varia
b
le r
e
n
e
w
a
ble
en
erg
y
i
n
p
o
w
e
r
s
y
st
ems:
T
he case of
w
i
n
d
gen
erat
i
on i
n
nort
h
e
a
st
ern B
r
azil.
Ener
gy
.
2012;
37(
1):
469
-481.
[
15]
Liu
C,
Wan
g
JH,
Bot
t
e
rud
A,
et
,
al.
Asse
ssment
of
imp
a
ct
s of
PHEV
char
gin
g
p
a
t
t
e
rns
on
w
i
n
d
-
t
hermal sc
he
d
u
lin
g b
y
st
och
a
s
t
i
c unit
c
o
mmi
t
m
ent
.
IEEE
Transacti
ons on Sm
art
Gri
d
.
20
12;
3(2):
675-
683.
[
16]
Khod
a
y
ar
ME,
Wu L,
Sha
h
i
d
ehp
our
M.
Ho
url
y
c
oord
i
n
a
t
i
on
of
e
l
ect
r
ic v
ehicl
e
op
erat
io
n a
n
d
vol
a
t
i
l
e
w
i
nd
po
w
e
r ge
nerat
i
on in SC
UC.
IEEE Transactions on Sm
art Grid
.
20
1
2
;
3(3):
127
1-1
279.
[
17]
Venkat
es
h
P,
Gnana
dass,
P
adh
y NP.
Co
mparis
o
n
an
d
ap
plic
at
io
n of
evo
l
ut
io
nar
y
progr
ammin
g
t
e
chni
qu
es t
o
combi
ned ec
o
nomic emiss
i
o
n
disp
at
ch
w
i
t
h
line f
l
o
w
c
o
nst
r
aint
s.
IEEE Transactions
Power Systems
.
2003;
18(
2):
688-
697.
[
18]
Ongsaku
W,
Pet
c
haraks
N.
Un
it
commit
m
ent
b
y
e
nha
nced
a
dapt
iv
e
la
gran
gi
an r
e
la
xat
i
on.
IEEE
Transactions on power system
s
.
20
04;
19(
1
)
:
620-62
8.
[
19]
Saber AY,
Ve
na
ya
gamo
o
rt
h
y
GK.
I
n
t
e
llig
e
n
t
unit
commit
m
ent
w
i
t
h
veh
i
cle-t
o
-gr
i
d-A c
o
st
-emissio
n
pt
imizat
i
on.
Jo
urna
l of
Pow
e
r Sources
.
20
10;
195(1):
89
8-9
11.
[
20]
Ahmed YS,
V
ena
ya
g
a
moort
h
y
GK.
R
e
sou
r
ce sche
dul
in
g
und
er unc
ert
a
int
y
i
n
a sm
art
grid
w
i
t
h
rene
w
a
b
l
es a
n
d
plu
g
-in ve
hic
l
es.
IEEE syste
m
Jour
nal
.
201
2;
6(1):
103-1
0
9
.
Evaluation Warning : The document was created with Spire.PDF for Python.