I
nd
o
ne
s
ia
n J
o
urna
l o
f
E
lect
rica
l En
g
ineering
a
nd
Co
m
pu
t
er
Science
Vo
l.
24
,
No
.
3
,
Dec
em
b
er
2
0
2
1
,
p
p
.
1
5
8
9
~
1
5
9
5
I
SS
N:
2
5
0
2
-
4
7
5
2
,
DOI
: 1
0
.
1
1
5
9
1
/ijeecs.v
24
.i
3
.
pp
1
5
8
9
-
1
5
9
5
1589
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ee
cs.ia
esco
r
e.
co
m
New scaled
alg
o
rithm f
o
r
non
-
linea
r conjug
a
te
g
ra
di
ents in
unco
nstra
ined op
timiza
tion
G
ha
da
M
.
Al
-
N
ae
mi
,
Ahm
e
d H
.
Sh
ee
k
o
o
De
p
a
rtme
n
t
o
f
E
n
e
rg
y
M
a
th
e
m
a
ti
c
s,
F
a
c
u
lt
y
o
f
C
o
m
p
u
ter
S
c
ien
c
e
a
n
d
M
a
t
h
e
m
a
ti
c
s,
M
o
su
l
Un
i
v
e
rsity
,
M
o
su
l
,
Ira
q
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Mar
10
,
2
0
2
1
R
ev
is
ed
Oct
11
,
2
0
2
1
Acc
ep
ted
Oct
18
,
2
0
2
1
A
n
e
w
sc
a
led
c
o
n
ju
g
a
te
g
ra
d
ien
t
(S
CG
)
m
e
th
o
d
is
p
ro
p
o
se
d
t
h
ro
u
g
h
o
u
t
t
h
is
p
a
p
e
r,
th
e
S
CG
tec
h
n
iq
u
e
m
a
y
b
e
a
sp
e
c
ial
imp
o
rtan
t
g
e
n
e
ra
li
z
a
ti
o
n
c
o
n
ju
g
a
te
g
ra
d
ie
n
t
(CG
)
m
e
th
o
d
,
a
n
d
it
is
a
n
e
fficie
n
t
n
u
m
e
rica
l
m
e
th
o
d
f
o
r
so
lv
i
n
g
n
o
n
li
n
e
a
r
larg
e
sc
a
le
u
n
c
o
n
stra
in
e
d
o
p
ti
m
iza
ti
o
n
.
As
a
re
su
lt
,
we
p
ro
p
o
se
d
t
h
e
n
e
w
S
CG
m
e
th
o
d
with
a
stro
n
g
Wo
lfe
c
o
n
d
it
io
n
(
S
WC)
li
n
e
se
a
rc
h
is
p
ro
p
o
se
d
.
T
h
e
p
r
o
p
o
se
d
tec
h
n
i
q
u
e
'
s
d
e
sc
e
n
t
p
ro
p
e
rty
,
a
s
we
ll
a
s
it
s
g
lo
b
a
l
c
o
n
v
e
rg
e
n
c
e
p
ro
p
e
rty
,
a
re
sa
ti
sfie
d
with
o
u
t
th
e
u
se
o
f
a
n
y
l
in
e
se
a
rc
h
e
s
u
n
d
e
r
so
m
e
su
it
a
b
le
a
ss
u
m
p
ti
o
n
s.
T
h
e
p
ro
p
o
se
d
tec
h
n
iq
u
e
'
s
e
fficie
n
c
y
a
n
d
fe
a
sib
i
li
ty
a
re
b
a
c
k
e
d
u
p
b
y
n
u
m
e
rica
l
e
x
p
e
rime
n
ts
c
o
m
p
a
rin
g
th
e
m
to
tra
d
it
i
o
n
a
l
CG
tec
h
n
iq
u
e
s
.
K
ey
w
o
r
d
s
:
C
G
m
eth
o
d
L
ar
g
e
-
s
ca
le
n
o
n
lin
ea
r
SC
G
m
eth
o
d
Su
f
f
icien
t d
escen
t p
r
o
p
er
ty
Un
co
n
s
tr
ain
ed
o
p
tim
izatio
n
T
h
is i
s
a
n
o
p
e
n
a
c
c
e
ss
a
rticle
u
n
d
e
r th
e
CC B
Y
-
SA
li
c
e
n
se
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Gh
ad
a
M.
Al
-
Nae
m
i
Dep
ar
tm
en
t o
f
Ma
th
em
atics,
Facu
lty
o
f
C
o
m
p
u
ter
Scien
ce
s
a
n
d
Ma
th
em
atics
Un
iv
er
s
ity
o
f
Mo
s
u
l
Mo
s
u
l,
I
r
aq
E
m
ail:
d
r
g
h
a
d
aa
ln
ae
m
i@
u
o
m
o
s
u
l.e
d
u
.
iq
1.
I
NT
RO
D
UCT
I
O
N
C
G
m
eth
o
d
is
u
n
iv
er
s
al
m
e
th
o
d
f
o
r
s
o
lv
in
g
n
o
n
lin
ea
r
l
ar
g
e
-
s
ca
le
u
n
co
n
s
tr
ain
ed
o
p
t
im
izatio
n
p
r
o
b
lem
s
,
b
ec
a
u
s
e
it
h
as
s
im
p
le
iter
atio
n
s
,
lo
w
m
em
o
r
y
r
eq
u
ir
em
en
ts
an
d
v
er
y
f
ast
co
n
v
er
g
en
ce
p
r
o
p
er
ties
[
1
]
.
T
h
er
ef
o
r
e,
in
th
is
wo
r
k
,
we
co
n
s
id
er
ed
th
is
g
en
er
al
u
n
co
n
s
tr
ain
ed
o
p
ti
m
izatio
n
p
r
o
b
lem
:
i
n
d
ex
in
g
an
d
a
b
s
tr
ac
tin
g
s
er
v
i
ce
s
d
ep
en
d
o
n
t
h
e
ac
c
u
r
ac
y
o
f
th
e
titl
e,
ex
tr
ac
tin
g
f
r
o
m
it
k
ey
wo
r
d
s
u
s
ef
u
l
i
n
cr
o
s
s
-
r
ef
er
en
cin
g
an
d
co
m
p
u
t
er
s
ea
r
ch
in
g
.
An
im
p
r
o
p
er
ly
titl
ed
p
ap
er
m
ay
n
ev
er
r
ea
ch
th
e
au
d
ien
ce
f
o
r
wh
ich
it wa
s
in
ten
d
ed
,
s
o
b
e
s
p
ec
if
ic.
{
(
)
:
∈
}
(
1
)
W
h
er
e
:
→
i
s
s
m
o
o
t
h
an
d
its
g
r
a
d
ien
t
v
ec
to
r
d
ef
in
ed
=
(
)
,
an
d
th
e
in
itial
p
o
in
t
0
∈
is
u
s
u
ally
s
o
lv
ed
iter
ativ
ely
ac
c
o
r
d
in
g
to
th
e
r
ec
u
r
s
iv
e
f
o
r
m
u
la
,
+
1
=
+
,
≥
0
(
2
)
w
h
er
e
is
cu
r
r
en
t
iter
atio
n
,
>
0
is
th
e
s
tep
-
s
ize
ca
lcu
lated
b
y
t
h
e
SW
C
,
(
+
)
≤
(
)
+
|
(
+
)
|
≤
−
(
3
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
24
,
No
.
3
,
Dec
em
b
er
2
0
2
1
:
1
5
8
9
-
1
5
9
5
1590
w
h
e
r
e
0
<
<
<
1
a
n
d
i
s
a
s
e
a
r
c
h
d
i
r
e
c
t
i
o
n
.
T
h
e
c
l
a
s
s
i
c
a
l
s
e
a
r
c
h
d
i
r
e
c
t
i
o
n
+
1
a
r
e
f
r
e
q
u
e
n
t
l
y
d
e
f
i
n
e
d
b
y
,
+
1
=
{
−
+
1
,
=
0
−
+
1
+
,
≥
1
(
4
)
g
en
er
ally
,
t
h
e
p
a
r
am
eter
is
s
elec
ted
s
o
th
at
if
f(
x)
is
a
s
tr
ictly
co
n
v
e
x
q
u
ad
r
atic
f
u
n
c
tio
n
an
d
if
is
ca
lcu
lated
b
y
th
e
e
x
ac
t
lin
e
s
ea
r
ch
,
th
en
(
2
)
an
d
(
4
)
ca
n
b
e
s
im
p
lifie
d
to
th
e
lin
ea
r
co
n
ju
g
ate
g
r
ad
ien
t
tech
n
iq
u
e
[
2
]
.
Sev
er
al
f
o
r
m
u
las,
s
u
ch
as
h
esten
es
a
n
d
s
tief
el
(
HS)
,
f
letch
er
an
d
r
ee
v
e
s
(
FR
)
,
co
n
ju
g
ate
d
escen
t
(
C
D)
,
Po
lak
-
R
ib
ier
e
(
PR
P)
,
L
iu
an
d
Sto
r
ey
(
L
S)
an
d
Dai
-
Yu
a
n
m
eth
o
d
(
DY)
,
h
av
e
b
ee
n
p
r
o
p
o
s
ed
[
3
]
-
[
9
]
.
As d
em
o
n
s
tr
ated
b
y
th
e
f
o
r
m
u
la
,
=
(
−
−
1
)
(
−
−
1
)
−
1
;
=
−
1
−
1
;
=
(
−
−
1
)
|
|
−
1
|
|
2
=
−
−
1
−
1
;
=
−
1
−
−
1
−
1
;
=
−
1
−
1
t
h
e
p
r
im
ar
y
d
is
tin
ctio
n
b
etwe
en
SC
G
an
d
C
G
is
th
e
ca
lcu
la
tio
n
o
f
th
e
s
ea
r
ch
d
ir
ec
tio
n
.
SC
G
'
s
ty
p
ical
s
ea
r
ch
d
ir
ec
tio
n
is
as f
o
llo
ws
,
+
1
=
{
−
+
1
,
=
0
−
+
1
+
,
≥
1
(
5
)
w
h
er
e
d
en
o
tes
a
s
p
ec
tr
al
p
ar
a
m
eter
.
B
ar
zilai
an
d
B
o
r
wien
[
1
0
]
p
r
o
p
o
s
ed
t
h
e
SC
G
m
eth
o
d
an
d
d
e
v
elo
p
e
d
th
eir
u
n
co
n
s
tr
ain
ed
o
p
tim
izati
o
n
.
I
n
s
tead
o
f
g
lo
b
al
c
o
n
v
e
r
g
e
n
ce
,
th
e
id
ea
is
to
u
s
e
o
n
ly
tea
s
in
g
tr
en
d
s
.
B
ir
g
in
an
d
Ma
r
tin
ez
[
1
1
]
p
r
o
p
o
s
ed
an
u
n
co
n
s
tr
ain
e
d
o
p
tim
izati
o
n
m
eth
o
d
,
b
u
t
it
lack
ed
a
s
u
f
f
icien
t
d
escen
t
co
n
d
itio
n
.
As
a
r
esu
lt,
An
d
r
ai
[
1
2
]
p
r
o
p
o
s
ed
an
ac
ce
ler
ate
d
C
G
tech
n
o
lo
g
y
th
at
u
s
es
th
e
New
to
n
m
eth
o
d
to
im
p
r
o
v
e
th
e
C
G
m
eth
o
d
'
s
p
er
f
o
r
m
an
ce
.
Fo
llo
win
g
o
n
f
r
o
m
t
h
is
th
o
u
g
h
t,
Far
v
a
n
eh
an
d
Key
v
an
[
1
3
]
p
r
o
p
o
s
ed
a
n
ew
SC
G
[
14]
-
[
2
0
]
c
o
n
tain
ad
d
itio
n
al
r
ef
e
r
en
ce
s
in
th
is
f
i
eld
.
2.
NE
W
A
L
G
O
RI
T
H
M
AND
T
H
E
D
E
SC
E
N
T
P
RO
P
E
R
T
Y
Ob
v
io
u
s
ly
,
f
o
r
SC
G,
th
e
m
eth
o
d
f
o
r
s
elec
tin
g
th
e
s
p
ec
tr
al
p
ar
am
eter
an
d
co
n
ju
g
ate
p
a
r
a
m
eter
is
cr
itical.
I
n
th
is
s
ec
tio
n
,
we
ex
p
lain
e
h
o
w
o
u
r
p
r
o
p
o
s
ed
SC
G
is
d
ep
en
d
en
t
o
n
th
e
p
ar
am
eter
p
r
o
p
o
s
ed
b
y
W
ei
et
a
l.
[
2
1
]
,
w
h
ich
is
d
ef
in
ed
as
(
6
)
.
=
|
|
|
|
2
−
|
|
|
|
|
|
+
1
|
|
+
1
|
|
+
1
|
|
2
(
6
)
T
h
e
n
ew
s
p
ec
tr
al
p
a
r
am
eter
is
p
r
p
o
s
ed
b
y
(
7
)
,
=
1
+
+
1
−
(
+
1
)
(
+
1
)
|
|
|
|
|
|
+
1
|
|
|
|
|
|
2
(
7
)
n
o
te
th
at,
if
a
n
ex
ac
t lin
e
s
ea
r
c
h
is
u
s
ed
th
en
=
1
,
s
o
(
5
)
r
ed
u
ce
d
to
(
4
)
.
A
lg
o
r
ith
m
SC
G
Step
1
:
Select
a
s
tar
tin
g
p
o
in
t
0
∈
,
g
iv
en
co
n
s
tan
d
0
<
<
<
1
,
s
to
p
p
in
g
cr
iter
i
a
=
1
0
−
6
>
0
;
Set
0
=
−
0
.
Step
2
: Co
m
p
u
te
|
|
|
|
,
if
|
|
|
|
≤
,
s
to
p
.
Oth
er
wis
e,
co
n
tin
u
es.
Step
3
: Calcu
late
,
,
b
y
(
6
)
a
n
d
(
7
)
r
esp
ec
tiv
ely
an
d
co
m
p
u
te
s
tep
len
g
th
b
y
(
3
)
.
Step
4
: U
p
d
ate
th
e
n
ew
p
o
in
t
b
y
(
2
)
.
C
o
m
p
u
te
+
1
=
(
+
1
)
;
if
|
|
+
1
|
|
≤
,
s
to
p
; O
th
er
wis
e,
co
n
tin
u
es.
Step
5
: Co
m
p
u
te
s
ea
r
ch
d
i
r
ec
tio
n
+
1
b
y
(
5
)
.
Step
6
: I
f
th
e
Po
well
r
estar
t c
r
i
ter
ia
|
+
1
|
≥
0
.
2
|
|
+
1
|
|
2
(
8
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
N
ew s
ca
led
a
lg
o
r
ith
m
fo
r
n
o
n
-
lin
ea
r
co
n
ju
g
a
te
g
r
a
d
ien
ts
in
u
n
co
n
s
tr
a
in
ed
…
(
G
h
a
d
a
M.
A
l
-
N
a
emi
)
1591
is
s
ati
s
f
ied
,
s
et
+
1
=
−
+
1
an
d
g
o
b
ac
k
t
o
Step
3
; o
th
er
wis
e
co
n
tin
u
es.
Step
7
:
P
u
t
=
+
1
an
d
g
o
to
s
tep
3
.
W
e
wil
l
d
is
cu
s
s
th
e
s
u
f
f
icien
t
d
escen
t
p
r
o
p
er
ty
o
f
th
e
Alg
o
r
ith
m
SC
G
ab
o
v
e
with
o
u
t
d
ep
en
d
in
g
to
an
y
lin
e
s
ea
r
ch
.
2
.
1
.
T
heo
rm
I
t
can
b
e
co
n
clu
d
ed
t
h
at
th
e
S
C
G
m
eth
o
d
with
th
e
lin
e
s
ea
r
ch
d
ir
ec
tio
n
(
5
)
,
,
d
ef
in
ed
in
(
6
)
an
d
(
7
)
r
esp
ec
tiv
ely
,
an
d
t
h
en
,
+
1
+
1
≤
−
|
|
+
1
|
|
2
,
≥
0
(
9
)
h
o
ld
s
f
o
r
∀
≥
0
.
Pro
o
f
:
T
o
s
tim
u
late
th
is
co
n
f
ir
m
atio
n
,
we
u
s
e
in
d
u
ctio
n
,
if
=
0
,
th
en
0
=
−
|
|
0
|
|
2
,
as
a
r
esu
lt
;
co
n
d
itio
n
(
9
)
is
estab
lis
h
ed
.
No
w,
co
n
d
itio
n
(
9
)
is
also
tr
u
e
in
o
r
d
er
to
n
o
tify
th
at
e
v
er
y
≥
0
is
tr
u
e.
Mu
ltip
ly
b
o
th
s
id
es o
f
(
5
)
b
y
+
1
to
o
b
tain
,
+
1
+
1
=
−
(
1
+
+
1
−
(
+
1
)
(
+
1
)
|
|
|
|
|
|
+
1
|
|
|
|
|
|
2
)
|
|
+
1
|
|
2
+
|
|
+
1
|
|
2
−
|
|
+
1
|
|
|
|
|
|
+
1
|
|
|
|
2
+
1
=
−
|
|
+
1
|
|
2
−
|
|
+
1
|
|
2
−
|
|
+
1
|
|
|
|
|
|
+
1
|
|
|
|
2
+
1
+
|
|
+
1
|
|
2
−
|
|
+
1
|
|
|
|
|
|
+
1
|
|
|
|
2
+
1
=
−
|
|
+
1
|
|
2
(
1
0
)
t
h
er
ef
o
r
e
,
th
e
Alg
o
r
ith
m
SC
G
ca
n
s
atis
f
y
th
e
s
u
f
f
icien
t
d
esce
n
t c
o
n
d
itio
n
s
with
o
u
t u
s
in
g
a
n
y
lin
e
s
ea
r
ch
es.
3.
T
H
E
G
L
O
B
AL
CO
N
VE
RG
E
NC
E
ANA
L
YSI
S
T
h
e
g
en
er
al
s
itu
atio
n
o
f
th
e
o
b
jectiv
e
f
u
n
ctio
n
r
e
q
u
ir
e
d
f
o
r
t
h
e
o
v
er
all
g
lo
b
al
co
n
v
e
r
g
en
ce
o
f
g
en
er
al
C
G
in
p
s
y
ch
o
lo
g
ical
an
al
y
s
is
is
as f
o
llo
ws.
3
.
1
.
Ass
um
ptio
n
−
T
h
e
f
u
n
ctio
n
(
)
is
co
n
s
tr
ain
ed
f
r
o
m
b
elo
w
to
th
e
lev
el
s
et
=
{
:
∈
/
(
)
≤
(
0
)
}
,
wh
er
e
th
e
p
o
in
t o
f
d
ep
a
r
tu
r
e
is
0
.
i.e
.
,
th
e
r
e
is
a
co
n
s
tan
t α
>0
,
wh
ich
m
ea
n
s
‖
‖
≤
∀
∈
.
−
I
n
ce
r
tain
n
eig
h
b
o
r
h
o
o
d
Ν
o
f
th
e
lev
el
s
et
Φ
,
th
e
f
u
n
ctio
n
f
(
x
)
is
co
n
tin
u
o
u
s
ly
d
i
f
f
er
en
tiab
le
an
d
its
g
r
ad
ien
t
(
)
is
L
ip
s
ch
itz
co
n
tin
u
o
u
s
,
i.e
.
∃
a
co
n
s
tan
t,
>
0
s
.
t.
+
1
=
{
−
+
1
,
=
0
−
+
1
+
,
≥
1
(
1
1
)
Ass
u
m
p
tio
n
(
I
)
clea
r
l
y
im
p
lies
th
e
ex
is
ten
ce
o
f
a
co
n
s
tan
t
>
0
,
s
.
t.
,
0
<
‖
+
1
‖
≤
,
∀
∈
[
2
2
]
(
1
2
)
t
h
e
f
o
llo
win
g
L
em
m
a,
k
n
o
wn
as
th
e
Z
o
u
n
te
n
d
ijk
c
o
n
d
itio
n
Z
o
u
n
ten
d
ijk
[
2
3
]
,
p
r
o
p
o
s
ed
i
t
an
d
is
f
r
eq
u
en
tl
y
u
s
ed
to
d
em
o
n
s
tr
ate
g
lo
b
al
c
o
n
v
er
g
e
n
ce
o
f
C
G
tech
n
iq
u
es.
3
.
2
.
L
emm
a
Su
p
p
o
s
e
Ass
u
m
p
tio
n
(
I
)
h
o
l
d
s
.
Su
p
p
o
s
e
a
g
en
er
al
iter
ativ
e
m
eth
o
d
(
2
)
an
d
th
e
d
ir
ec
tio
n
(
4
)
is
d
escen
t
d
ir
ec
tio
n
.
So
,
we
h
av
e
g
o
t
,
∑
(
)
2
|
|
|
|
2
∞
=
0
<
∞
(
1
3
)
a
cc
o
r
d
in
g
to
Ass
u
m
p
tio
n
s
(
3
.
1
)
,
T
h
eo
r
em
(
2
.
1
)
a
n
d
L
e
m
m
a
(
3
.
1
)
,
t
h
e
f
o
llo
win
g
r
esu
lts
ca
n
b
e
p
r
o
v
ed
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
24
,
No
.
3
,
Dec
em
b
er
2
0
2
1
:
1
5
8
9
-
1
5
9
5
1592
3
.
3
.
T
heo
re
m
Su
p
p
o
s
e
th
at
Ass
u
m
p
tio
n
(
I
)
h
o
ld
s
.
An
y
C
G
m
eth
o
d
o
f
t
h
e
f
o
r
m
(
2
)
an
d
(
5
)
with
is
a
d
escen
d
in
g
s
ea
r
ch
d
ir
ec
tio
n
a
n
d
s
atis
f
ies
SW
C
.
T
h
en
,
→
∞
|
|
|
|
=
0
(
1
4
)
o
r
,
∑
|
|
|
|
4
|
|
|
|
2
≥
1
<
+
∞
(
1
5
)
Pro
o
f
:
Ass
u
m
e,
f
o
r
th
e
s
ak
e
o
f
ar
g
u
m
en
t
th
at
t
h
e
co
n
clu
s
io
n
is
n
o
t
tr
u
e.
T
h
en
th
e
r
e
ex
is
t
s
a
p
o
s
itiv
e
co
n
s
tan
t
̅
>
0
s
.
t.
‖
+
1
‖
≥
̅
,
∀
.
W
e
ca
n
d
e
d
u
ce
f
r
o
m
(
5
)
th
at
+
1
+
+
1
=
.
W
h
en
we
s
q
u
a
r
e
b
o
th
s
id
es o
f
th
is
eq
u
atio
n
,
we
g
et
,
(
+
1
+
+
1
)
(
+
1
+
+
1
)
=
(
)
2
|
|
|
|
2
|
|
+
1
|
|
2
=
−
(
)
2
|
|
+
1
|
|
2
−
2
+
1
+
1
+
(
)
2
|
|
|
|
2
d
iv
id
in
g
b
o
th
s
id
es o
f
th
e
ab
o
v
e
eq
u
atio
n
b
y
|
|
+
1
|
|
4
,
an
d
u
s
e
(
1
0
)
we
g
et
,
|
|
+
1
|
|
2
(
+
1
+
1
)
2
=
|
|
+
1
|
|
2
|
|
+
1
|
|
4
=
−
(
)
2
|
|
+
1
|
|
2
−
2
,
|
|
+
1
|
|
2
+
(
)
2
|
|
|
|
2
|
|
+
1
|
|
4
=
−
(
(
)
2
+
2
)
|
|
+
1
|
|
2
+
(
)
2
|
|
|
|
2
|
|
+
1
|
|
4
=
−
(
(
)
2
+
2
+
1
−
1
)
|
|
+
1
|
|
2
+
(
)
2
|
|
|
|
2
|
|
+
1
|
|
4
=
1
|
|
+
1
|
|
2
+
(
)
2
|
|
|
|
2
|
|
+
1
|
|
4
−
(
1
|
|
+
1
|
|
2
+
(
+
1
)
2
|
|
+
1
|
|
2
)
≤
1
|
|
+
1
|
|
2
+
(
)
2
|
|
|
|
2
|
|
+
1
|
|
4
i
n
[
1
6
]
th
ey
p
r
o
v
ed
0
≤
≤
2
|
|
+
1
|
|
2
|
|
|
|
2
∀
≥
0
|
|
+
1
|
|
2
|
|
+
1
|
|
4
≤
1
|
|
+
1
|
|
2
+
(
2
|
|
+
1
|
|
2
|
|
|
|
2
)
2
|
|
|
|
2
|
|
+
1
|
|
4
=
1
|
|
+
1
|
|
2
+
4
|
|
|
|
2
|
|
|
|
4
|
|
+
1
|
|
2
|
|
+
1
|
|
4
≤
4
|
|
|
|
2
|
|
|
|
4
+
1
|
|
+
1
|
|
2
I
n
ter
m
s
o
f
|
|
1
|
|
2
(
1
1
)
2
=
1
|
|
1
|
|
2
,
to
g
et
h
er
with
th
e
ab
o
v
e
r
elatio
n
s
an
d
|
|
|
|
2
≥
,
we
h
av
e
,
|
|
+
1
|
|
2
|
|
+
1
|
|
4
≤
4
|
|
|
|
2
|
|
|
|
4
+
1
|
|
+
1
|
|
2
+
1
|
|
|
|
2
≤
⋯
≤
∑
1
|
|
|
|
2
=
1
≤
ῶ
2
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
N
ew s
ca
led
a
lg
o
r
ith
m
fo
r
n
o
n
-
lin
ea
r
co
n
ju
g
a
te
g
r
a
d
ien
ts
in
u
n
co
n
s
tr
a
in
ed
…
(
G
h
a
d
a
M.
A
l
-
N
a
emi
)
1593
t
h
at
is
,
|
|
+
1
|
|
4
|
|
+
1
|
|
≥
ῶ
2
.
Hen
ce
∑
|
|
+
1
|
|
2
|
|
+
1
|
|
4
≥
1
≥
+
∞
,
th
is
is
c
o
n
t
r
ad
icts
lem
m
a
(
3
.
1
)
.
T
h
er
ef
o
r
e,
th
e
p
r
o
o
f
is
co
m
p
lete.
4.
T
H
E
NU
M
E
RICA
L
R
E
SU
L
T
S
I
n
th
is
s
ec
tio
n
,
we
will
p
r
esen
t
th
e
o
u
tco
m
es
o
f
v
ar
io
u
s
test
f
u
n
ctio
n
s
.
T
o
ev
alu
ate
th
e
n
e
w
m
eth
o
d
,
s
o
m
e
test
f
u
n
ctio
n
s
wer
e
ch
o
s
en
.
T
h
ese
f
u
n
ctio
n
s
ar
e
tak
e
n
in
to
ac
co
u
n
t
b
y
C
UT
E
test
f
u
n
ctio
n
[
2
4
]
,
[
2
5
]
.
Usi
n
g
SW
C
lin
e
s
ea
r
ch
,
th
e
n
ew
SC
G
m
eth
o
d
,
t
h
e
class
ic
[
2
1
]
(
W
YL
)
m
eth
o
d
,
th
e
FR
m
eth
o
d
,
an
d
th
e
L
S
m
eth
o
d
ar
e
co
m
p
a
r
ed
in
ter
m
s
o
f
th
e
n
u
m
b
er
o
f
iter
atio
n
s
(
NI
)
an
d
th
e
n
u
m
b
e
r
o
f
f
u
n
ctio
n
ev
alu
atio
n
s
(
NF)
.
All
s
y
m
b
o
ls
ar
e
wr
itten
in
F
OR
T
R
A
N
7
7
d
o
u
b
le
p
r
ec
is
io
n
an
d
c
o
llected
as
Vis
u
al
FOR
T
R
AN
(
F6
.
6
)
.
T
h
e
n
ew
SC
G
m
eth
o
d
is
im
p
lem
e
n
ted
u
s
in
g
th
e
SW
C
lin
e
s
ea
r
c
h
(
3
)
,
a
n
d
with
=
0
.
001
,
=
0
.
9
,
we
test
ed
1
5
well
-
k
n
o
wn
test
f
u
n
ctio
n
s
,
th
e
d
im
en
s
io
n
s
o
f
wh
ich
ar
e
g
iv
en
b
elo
w
(
1
0
0
0
,
5
0
0
0
,
1
0
0
0
0
,
5
0
0
0
0
,
a
n
d
1
0
0
0
0
0
)
.
T
h
is
alg
o
r
ith
m
’
s
s
to
p
p
in
g
cr
iter
io
n
is
|
|
+
1
|
|
≤
10
−
6
an
d
we
en
ter
6
0
0
if
th
e
(
NI
)
eq
u
al
to
o
r
m
o
r
e
th
an
6
0
0
.
T
h
e
r
esu
lts
o
b
tain
ed
b
y
th
e
n
ewly
p
r
o
p
o
s
ed
m
eth
o
d
o
u
tp
er
f
o
r
m
th
o
s
e
o
b
tain
ed
b
y
th
e
o
th
er
m
eth
o
d
s
m
en
tio
n
ed
i
n
t
h
e
T
ab
le
1
.
T
ab
le
1
.
T
h
e
co
m
p
ar
is
o
n
b
etw
ee
n
th
e
p
r
o
p
o
s
ed
m
eth
o
d
an
d
th
e
o
th
er
class
ical
m
eth
o
d
s
No
Te
st
F
u
n
c
t
i
o
n
D
i
me
n
si
o
n
S
C
G
me
t
h
o
d
W
Y
L
me
t
h
o
d
F
R
me
t
h
o
d
ni
nf
ni
nf
ni
nf
1
R
O
S
EN
1
0
0
0
26
67
30
78
30
78
1
0
0
0
0
26
67
30
78
30
78
1
0
0
0
0
0
27
70
32
83
31
81
2
W
O
LFE
1
0
0
0
1
1
0
2
2
5
1
1
6
2
3
0
1
1
1
2
2
9
1
0
0
0
0
1
2
0
2
4
4
1
2
3
2
4
8
1
3
5
2
8
0
1
0
0
0
0
0
1
2
1
2
4
7
1
3
0
2
5
5
1
3
4
2
7
6
EX
-
B
LO
C
K
D
I
A
G
O
N
A
L
B
D
1
1
0
0
0
23
49
21
45
22
46
1
0
0
0
0
24
51
23
49
26
48
1
0
0
0
0
0
26
53
25
53
27
52
4
S
H
A
LLO
W
1
0
0
0
10
25
10
25
10
25
1
0
0
0
0
10
25
10
25
10
25
1
0
0
0
0
0
11
27
11
27
11
27
5
W
O
O
D
1
0
0
0
28
64
29
66
29
66
1
0
0
0
0
28
64
29
66
29
66
1
0
0
0
0
0
30
68
29
66
30
68
6
B
EA
L
1
0
0
0
11
27
11
27
11
27
1
0
0
0
0
11
27
11
27
6
0
0
4
9
0
1
0
0
0
0
0
11
27
6
0
0
5
2
3
6
0
0
4
9
0
7
P
O
W
ELL
1
0
0
0
38
1
2
2
56
1
6
2
36
1
1
0
1
0
0
0
0
38
1
2
2
56
1
6
2
36
1
1
0
1
0
0
0
0
0
41
1
3
8
56
1
6
2
39
1
3
1
8
C
U
B
I
C
1
0
0
0
16
45
16
45
16
45
1
0
0
0
0
16
45
16
45
16
45
1
0
0
0
0
0
16
45
16
45
16
45
9
H
I
M
M
ELB
A
U
1
0
0
0
24
2
5
1
26
2
6
6
26
2
7
6
1
0
0
0
0
8
3
9
1
10
4
0
5
10
4
0
1
1
0
0
0
0
0
12
4
8
5
6
5
2
3
12
4
9
0
10
D
Q
D
R
TI
C
1
0
0
0
5
11
6
13
5
11
1
0
0
0
0
5
11
6
13
5
11
1
0
0
0
0
0
5
11
13
13
5
11
11
D
I
X
M
A
A
N
B
1
0
0
0
5
13
6
15
6
15
1
0
0
0
0
6
16
7
16
7
16
1
0
0
0
0
0
6
16
7
16
7
16
12
S
TR
A
I
T
1
0
0
0
6
14
8
21
7
18
1
0
0
0
0
6
14
8
21
6
15
1
0
0
0
0
0
6
14
8
21
6
15
13
B
EA
L
U
6
3
1
0
0
0
10
27
13
32
12
29
1
0
0
0
0
10
27
13
32
12
29
1
0
0
0
0
0
11
29
15
39
13
32
14
H
I
L
I
C
A
L
1
0
0
0
30
78
29
75
29
73
1
0
0
0
0
33
82
30
78
31
82
1
0
0
0
0
0
36
92
36
85
36
88
15
D
EN
S
C
H
N
B
1
0
0
0
9
21
9
21
9
21
9
21
9
21
9
21
9
21
9
21
9
21
To
t
a
l
1
0
6
9
3
5
8
9
1
7
6
0
4
3
3
9
2
2
9
7
4
6
2
9
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
5
0
2
-
4
7
5
2
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci,
Vo
l.
24
,
No
.
3
,
Dec
em
b
er
2
0
2
1
:
1
5
8
9
-
1
5
9
5
1594
T
ab
le
2
c
o
m
p
ar
es
th
e
p
er
f
o
r
m
an
ce
p
er
ce
n
tag
es
o
f
th
e
FR
,
W
YL
,
an
d
p
r
o
p
o
s
ed
SC
G
tech
n
o
lo
g
ies.
W
h
en
co
m
p
ar
ed
to
th
e
FR
-
m
eth
o
d
,
th
e
W
YL
tech
n
iq
u
e
s
av
es
(
NI
2
3
.
3
8
%),
(
NF
6
.
2
6
%)
an
d
th
e
SC
G
tech
n
iq
u
e
s
av
es
(
NI
5
3
.
4
6
%).
(
NF
2
2
.
4
7
%
)
.
Un
d
e
r
th
e
s
tr
o
n
g
W
o
lf
e
lin
e
s
ea
r
ch
,
th
e
p
r
o
p
o
s
ed
m
eth
o
d
o
u
tp
er
f
o
r
m
ed
th
e
ex
is
tin
g
m
et
h
o
d
s
in
ter
m
s
o
f
n
u
m
b
er
o
f
ite
r
atio
n
s
an
d
n
u
m
b
e
r
o
f
f
u
n
ctio
n
ev
alu
atio
n
s
.
T
ab
le
2
.
T
h
e
p
er
ce
n
tag
e
p
e
r
f
o
r
m
an
ce
o
f
th
e
p
r
o
p
o
s
ed
m
eth
o
d
s
M
e
a
su
r
e
s
F
R
me
t
h
o
d
W
Y
L
me
t
h
o
d
S
C
G
me
t
h
o
d
NI
1
0
0
%
7
6
.
6
2
%
4
6
.
5
4
%
NF
1
0
0
%
9
3
.
7
4
%
7
7
.
5
3
%
5.
CO
NCLU
SI
O
N
I
n
th
is
p
ap
er
,
a
n
ew
s
ca
led
c
o
n
ju
g
ate
g
r
a
d
ien
t
alg
o
r
ith
m
f
o
r
u
n
co
n
s
tr
ain
ed
o
p
tim
izatio
n
p
r
o
b
lem
s
is
p
r
o
p
o
s
ed
.
T
h
is
m
eth
o
d
,
in
d
e
p
en
d
en
t
o
f
th
e
lin
e
s
ea
r
ch
,
s
atis
f
ies
th
e
s
u
f
f
icien
t
d
escen
t
co
n
d
itio
n
.
T
h
e
p
r
o
p
o
s
ed
m
eth
o
d
h
as th
e
ad
v
a
n
tag
e
o
f
b
ein
g
ap
p
licab
le
to
la
r
g
e
-
s
ca
le
p
r
o
b
lem
s
.
T
h
e
s
tr
o
n
g
W
o
lf
e
lin
e
s
ea
r
ch
is
u
s
ed
to
p
er
f
o
r
m
n
u
m
e
r
ical
co
m
p
u
tatio
n
s
o
n
s
o
m
e
s
tan
d
ar
d
b
en
ch
m
a
r
k
p
r
o
b
le
m
s
.
Pre
lim
in
ar
y
f
i
n
d
in
g
s
in
d
icate
th
at
th
e
p
r
o
p
o
s
ed
m
e
th
o
d
is
b
o
th
ef
f
icien
t
a
n
d
p
r
o
m
is
in
g
.
As
a
r
esu
lt,
it
ca
n
b
e
u
s
ed
as
a
d
if
f
er
en
t
ap
p
r
o
ac
h
f
o
r
lar
g
e
-
s
ca
le
u
n
c
o
n
s
tr
ain
ed
o
p
tim
izatio
n
p
r
o
b
l
em
s
.
Fu
r
th
er
m
o
r
e,
f
u
tu
r
e
r
esear
ch
ca
n
f
o
c
u
s
o
n
d
em
o
n
s
tr
atin
g
t
h
e
co
n
v
er
g
e
n
c
e
o
f
th
is
m
eth
o
d
u
n
d
er
d
if
f
er
e
n
t lin
e
s
ea
r
ch
m
eth
o
d
s
.
ACK
NO
WL
E
DG
M
E
N
T
S
T
h
e
au
th
o
r
s
ar
e
g
r
atef
u
l
to
th
e
Un
iv
er
s
ity
o
f
Mo
s
u
l'
s
C
o
lleg
e
o
f
C
o
m
p
u
ter
Scien
ce
s
an
d
Ma
th
em
atics f
o
r
th
eir
en
c
o
u
r
a
g
em
en
t a
n
d
s
u
p
p
o
r
t.
REFE
RENC
E
S
[1
]
J
.
Jia
n
,
L
.
Ya
n
g
,
X
.
Jia
n
g
,
P
.
L
iu
a
n
d
M
.
Li
u
,
“
A
sp
e
c
tral
c
o
n
j
u
g
a
te
g
ra
d
ien
t
m
e
th
o
d
wi
th
d
e
sc
e
n
t
p
ro
p
e
rty
,
”
J
o
u
rn
a
l
o
f
M
a
th
e
m
a
ti
c
s
,
v
o
l.
8
,
n
o
.
2
,
p
.
2
8
0
,
2
0
2
0
,
d
o
i:
1
0
.
3
3
9
0
/m
a
th
8
0
2
0
2
8
0
.
[2
]
N.
S
.
M
o
h
a
m
e
d
,
M
.
M
a
m
a
t,
M
.
Riv
a
ie
a
n
d
S
.
M
.
S
h
a
h
a
ru
d
d
i
n
,
“
G
lo
b
a
l
Co
n
v
e
r
g
e
n
c
e
o
f
a
Ne
w
Co
e
fficie
n
t
Co
n
j
u
g
a
te
G
ra
d
ien
t
M
e
th
o
d
,
”
I
n
d
o
n
e
si
a
n
J
o
u
rn
a
l
o
f
El
e
c
trica
l
E
n
g
i
n
e
e
rin
g
a
n
d
Co
m
p
u
ter
S
c
ien
c
e
(IJ
EE
CS
)
,
v
o
l
.
1
1
,
n
o
.
3
,
p
p
.
1
1
8
8
-
1
1
9
3
,
2
0
1
8
,
d
o
i:
1
0
.
1
1
5
9
1
/i
jee
c
s.v
1
1
.
i
3
.
p
p
1
1
8
8
-
1
1
9
3
.
[3
]
M
.
R.
He
ste
n
e
s
a
n
d
E.
S
ti
e
fe
l,
“
M
e
th
o
d
s
o
f
c
o
n
ju
g
a
te
g
ra
d
ien
ts
f
o
r
so
l
v
i
n
g
li
n
e
a
r
s
y
ste
m
s
,
”
J
o
u
rn
a
l
o
f
Res
e
a
rc
h
o
f
th
e
Na
ti
o
n
a
l
B
u
re
a
u
o
f
S
t
a
n
d
a
r
d
s
,
v
o
l
.
4
9
,
p
p
.
4
0
9
-
4
3
6
,
1
9
5
2
.
[On
li
n
e
].
Av
a
il
a
b
le:
h
tt
p
s:/
/n
v
lp
u
b
s.
n
ist.
g
o
v
/n
istp
u
b
s/j
re
s/0
4
9
/
jres
v
4
9
n
6
p
4
0
9
_
A1
b
.
p
d
f
[4
]
R.
F
letc
h
e
r
a
n
d
C.
M
.
Re
e
v
e
s.
“
F
u
n
c
ti
o
n
m
i
n
imiz
a
ti
o
n
b
y
c
o
n
j
u
g
a
te
g
ra
d
ien
ts
,
”
T
h
e
Co
m
p
u
ter
J
o
u
r
n
a
l
,
v
o
l.
7
,
n
o
.
2
,
p
p
.
1
4
9
-
1
5
4
,
1
9
6
4
,
d
o
i:
1
0
.
1
0
9
3
/co
m
jn
l
/7
.
2
.
1
4
9
.
[5
]
R.
F
letc
h
e
r
,
“
Pra
c
ti
c
a
l
M
e
t
h
o
d
o
f
Op
ti
miza
ti
o
n
,
”
v
o
l.
I:
Un
c
o
n
stra
i
n
e
d
Op
ti
m
iza
ti
o
n
,
Wi
le
y
,
Ne
w
Yo
rk
,
NY
,
USA,
2
n
d
e
d
it
i
o
n
,
1
9
9
7
,
d
o
i:
1
0
.
1
0
0
2
/
9
7
8
1
1
1
8
7
2
3
2
0
3
.
[6
]
E.
P
o
lak
a
n
d
G
.
Rib
ière
,
“
No
te
S
u
r
la
c
o
n
v
e
r
g
e
n
c
e
d
e
d
ir
e
c
ti
o
n
s
c
o
n
j
u
g
a
tes
,
”
Rev
,
Rev
u
e
Fra
n
ç
a
ise
d
'In
fo
rm
a
t
iq
u
e
e
t
d
e
Rec
h
e
rc
h
e
Op
é
ra
ti
o
n
n
e
ll
e
,
v
o
l
.
3
,
n
o
.
1
6
,
p
p
.
3
5
-
4
3
,
1
9
6
9
,
[On
li
e
].
Av
a
il
a
b
le:
h
tt
p
:
//
ww
w.n
u
m
d
a
m
.
o
rg
/
it
e
m
?
id
=
M
2
AN
_
1
9
6
9
_
_
3
_
1
_
3
5
_
0
[7
]
B.
T.
P
o
l
y
a
k
,
“
T
h
e
c
o
n
j
u
g
a
te
g
ra
d
ien
t
m
e
th
o
d
in
e
x
trem
e
p
ro
b
lem
s
,
”
US
S
R
Co
mp
u
ta
ti
o
n
a
l
M
a
th
e
ma
t
ics
a
n
d
M
a
t
h
e
ma
ti
c
a
l
Ph
y
sic
s
,
v
o
l.
9
,
n
o
.
4
,
p
p
.
9
4
-
1
1
2
,
1
9
6
9
,
d
o
i:
1
0
.
1
0
1
6
/0
0
4
1
-
5
5
5
3
(
6
9
)
9
0
0
3
5
-
4
.
[8
]
Y.
Li
u
a
n
d
C.
S
t
o
re
y
,
“
Eff
icie
n
t
g
e
n
e
ra
li
z
e
d
c
o
n
ju
g
a
te
g
ra
d
ien
t
a
lg
o
rit
h
m
s.
I.
T
h
e
o
r
y
,
”
J
o
u
rn
a
l
o
f
Op
t
imiza
ti
o
n
T
h
e
o
ry
a
n
d
Ap
p
li
c
a
t
io
n
s
,
v
o
l.
6
9
,
p
p
.
1
2
9
-
1
3
7
,
1
9
9
1
,
d
o
i:
1
0
.
1
0
0
7
/
BF
0
0
9
4
0
4
6
4
.
[9
]
Y.
H.
Da
i
a
n
d
Y.
Yu
a
n
,
“
A n
o
n
li
n
e
a
r
c
o
n
j
u
g
a
te
g
ra
d
ien
t
with
a
str
o
n
g
g
lo
b
a
l
c
o
n
v
e
rg
e
n
c
e
p
ro
p
e
rty
,
”
S
IA
M
J
o
u
rn
a
l
o
n
O
p
ti
miza
t
io
n
,
v
o
l.
1
0
,
n
o
.
1
,
p
p
.
1
7
7
-
1
8
2
,
2
0
0
0
,
d
o
i
:
1
0
.
1
1
3
7
/S
1
0
5
2
6
2
3
4
9
7
3
1
8
9
9
2
.
[1
0
]
J.
Brz
il
a
i
a
n
d
J.
Bo
rwe
in
,
“
Two
-
p
o
i
n
t
ste
p
siz
e
g
ra
d
ien
t
m
e
th
o
d
s
,
”
IM
A
J
.
Nu
me
ric
a
l
An
a
lys
is
,
v
o
l.
8
,
n
o
.
1
,
p
p
.
141
-
1
4
8
,
1
9
8
8
,
d
o
i
:
1
0
.
1
0
9
3
/i
m
a
n
u
m
/8
.
1
.
1
4
1
.
[1
1
]
E.
Bir
g
in
a
n
d
J.
M
a
rti
n
e
z
,
“
A
sp
e
c
tral
c
o
n
ju
g
a
te
g
ra
d
ien
t
m
e
th
o
d
fo
r
u
n
c
o
n
stra
in
e
d
o
p
ti
m
i
z
a
ti
o
n
,
”
A
p
p
li
e
d
M
a
t
h
e
ma
ti
c
s Op
ti
miz
a
ti
o
n
,
v
o
l.
4
3
,
p
p
.
1
1
7
-
1
2
8
,
2
0
0
1
,
d
o
i
:
1
0
.
1
0
0
7
/s0
0
2
4
5
-
0
0
1
-
0
0
0
3
-
0
.
[1
2
]
N.
An
d
rie,
“
S
c
a
led
c
o
n
j
u
g
a
te
g
ra
d
ien
t
a
l
g
o
rit
h
m
fo
r
u
n
c
o
n
stra
i
n
e
d
o
p
t
imiz
a
ti
o
n
,
”
C
o
mp
u
ta
t
io
n
a
l
O
p
ti
miza
ti
o
n
a
n
d
Ap
p
li
c
a
ti
o
n
,
v
o
l.
3
8
,
p
p
.
4
0
1
-
4
1
6
,
2
0
0
7
,
d
o
i
:
1
0
.
1
0
0
7
/s1
0
5
8
9
-
0
0
7
-
9
0
5
5
-
7
.
[1
3
]
F
.
F
a
rv
a
n
e
h
,
A.
Ke
y
v
a
n
,
“
A
m
o
d
ifi
e
d
s
p
e
c
tral
c
o
n
ju
g
a
te
g
ra
d
ien
t
m
e
th
o
d
wi
th
g
lo
b
a
l
c
o
n
v
e
rg
e
n
c
e
,
”
J
o
u
rn
a
l
o
f
Op
ti
miza
ti
o
n
T
h
e
o
ry
a
n
d
A
p
p
l
ica
ti
o
n
,
v
o
l.
1
8
2
,
p
p
.
6
6
7
-
6
9
0
,
2
0
1
9
,
doi
:
1
0
.
1
0
0
7
/s
1
0
9
5
7
-
0
1
9
-
0
1
5
2
7
-
6
.
[1
4
]
H.
Ah
m
e
d
a
n
d
G
h
a
d
a
M
.
Al
-
Na
e
m
i
,
“
A
m
o
d
ifi
e
d
Da
i
-
Yu
a
n
c
o
n
ju
g
a
te
g
ra
d
ien
t
m
e
th
o
d
s
a
n
d
it
s
g
l
o
b
a
l
c
o
n
v
e
rg
e
n
c
e
,
”
Ira
q
i
J
o
u
r
n
a
l
o
f
S
c
ien
c
e
,
v
o
l
.
5
3
,
n
o
.
3
,
p
p
.
6
2
0
-
6
2
8
,
2
0
1
2
.
[1
5
]
G
.
M
.
Al
-
Na
e
m
i.
“
A
G
lo
b
a
l
Co
n
v
e
rg
e
n
c
e
o
f
S
p
e
c
tral
Co
n
j
u
g
a
te
G
ra
d
ien
t
M
e
th
o
d
f
o
r
Larg
e
S
c
a
le
Op
ti
m
iza
ti
o
n
,
”
J
o
u
rn
a
l
o
f
E
d
u
c
a
t
io
n
a
n
d
S
c
ien
c
e
,
v
o
l.
2
7
,
n
o
.
3
,
p
p
.
1
4
3
-
1
6
2
,
2
0
1
8
,
d
o
i:
1
0
.
3
3
8
9
9
/ed
u
sj.2
0
1
8
.
1
5
9
3
2
3
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
d
o
n
esian
J
E
lec
E
n
g
&
C
o
m
p
Sci
I
SS
N:
2502
-
4
7
5
2
N
ew s
ca
led
a
lg
o
r
ith
m
fo
r
n
o
n
-
lin
ea
r
co
n
ju
g
a
te
g
r
a
d
ien
ts
in
u
n
co
n
s
tr
a
in
ed
…
(
G
h
a
d
a
M.
A
l
-
N
a
emi
)
1595
[1
6
]
A.
S
.
A
h
m
e
d
,
H.
M
.
K
h
u
d
u
r
a
n
d
M
.
S
.
Na
jmu
ld
e
e
n
,
“
A
n
e
w
p
a
ra
m
e
ter
in
th
re
e
-
term
s
c
o
n
ju
g
a
te
g
ra
d
ien
t
a
lg
o
rit
h
m
s
fo
r
u
n
c
o
n
stra
i
n
e
d
o
p
t
imiz
a
ti
o
n
,
”
I
n
d
o
n
e
sia
n
J
o
u
rn
a
l
o
f
El
e
c
trica
l
En
g
in
e
e
rin
g
a
n
d
C
o
mp
u
ter
S
c
ien
c
e
(IJ
EE
CS
)
,
v
o
l
.
2
3
,
n
o
.
1
,
p
p
.
3
3
8
-
3
4
4
,
Ju
l
y
2
0
2
1
,
d
o
i:
1
0
.
1
1
5
9
1
/i
je
e
c
s.v
2
3
.
i
1
.
p
p
3
3
8
-
3
4
4
.
[1
7
]
J.
K.
L
iu
,
Y.
M
.
F
e
n
g
a
n
d
L.
M
.
Zo
u
,
“
A
sp
e
c
tral
c
o
n
ju
g
a
t
e
g
ra
d
ien
t
m
e
th
o
d
fo
r
so
l
v
in
g
larg
e
-
sc
a
le
a
n
d
u
n
c
o
n
stra
in
e
d
o
p
ti
m
iza
ti
o
n
,
”
El
se
v
ier
L
td
.
,
v
o
l
.
7
7
,
p
p
.
7
3
1
-
7
3
9
,
2
0
1
8
,
d
o
i
:
1
0
.
1
0
1
6
/
j.
c
a
m
wa
.
2
0
1
8
.
1
0
.
0
0
2
.
[1
8
]
M.
Da
wa
h
d
e
h
,
I.
M
.
S
u
laim
a
n
,
M
.
Riv
a
ie
a
n
d
M
.
M
a
m
a
,
“
A
Ne
w S
p
e
c
tral
Co
n
ju
g
a
te
G
ra
d
ien
t
M
e
t
h
o
d
wit
h
S
tro
n
g
Wo
lfe
-
P
o
we
ll
Li
n
e
S
e
a
rc
h
,
”
In
te
rn
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
Eme
rg
i
n
g
T
re
n
d
s
i
n
En
g
i
n
e
e
rin
g
Res
e
a
rc
h
,
v
o
l.
8
,
n
o
.
2
,
F
e
b
ru
a
ry
2
0
2
0
,
d
o
i
:
1
0
.
3
0
5
3
4
/
ij
e
t
e
r/2
0
2
0
/
2
5
8
2
2
0
2
0
.
[1
9
]
A.
H.
S
h
e
e
k
o
o
G
h
a
d
a
a
n
d
M
.
Al
-
Na
e
m
i
,
“
G
lo
b
a
l
c
o
n
v
e
rg
e
n
c
e
Co
n
d
it
i
o
n
f
o
r
a
Ne
w
S
p
e
c
tral
Co
n
j
u
g
a
te
G
ra
d
ien
t
M
e
th
o
d
f
o
r
Larg
e
-
S
c
a
le
Op
ti
m
iz
a
ti
o
n
,
”
J
o
u
r
n
a
l
o
f
P
h
y
sic
s
c
o
n
fer
e
n
c
e
se
rie
s
IOP
Pu
b
li
sh
i
n
g
,
v
o
l
.
1
8
7
9
,
n
o
.
2
0
2
1
,
2
0
2
1
,
d
o
i:
1
0
.
1
0
8
8
/
1
7
4
2
-
6
5
9
6
/1
8
7
9
/
3
/0
3
2
0
0
1
.
[2
0
]
N.
S
.
M
o
h
a
m
e
d
,
“
G
lo
b
a
l
Co
n
v
e
r
g
e
n
c
e
o
f
a
Ne
w
Co
e
fficie
n
t
Co
n
ju
g
a
te
G
ra
d
ien
t
M
e
th
o
d
,
”
In
d
o
n
e
s
ia
n
J
o
u
rn
a
l
o
f
El
e
c
trica
l
En
g
in
e
e
rin
g
a
n
d
C
o
mp
u
ter
S
c
ien
c
e
(IJ
EE
CS
)
,
v
o
l.
1
1
,
n
o
.
3
,
p
p
.
1
1
8
8
-
1
1
9
3
,
2
0
1
8
,
d
oi
:
1
0
.
1
1
5
9
1
/i
jee
c
s.v
1
1
.
i
3
.
p
p
1
1
8
8
-
1
1
9
3
.
[2
1
]
Z.
Wei,
S
.
Ya
o
a
n
d
L
.
Li
u
,
“
Th
e
c
o
n
v
e
r
g
e
n
c
e
p
r
o
p
e
rti
e
s
o
f
so
m
e
n
e
w
c
o
n
j
u
g
a
te
g
ra
d
ie
n
t
m
e
th
o
d
,
”
A
p
p
l.
M
a
t
h
.
Co
mp
u
te
,
v
o
l.
1
8
3
,
n
o
.
2
,
p
p
.
1
3
4
1
-
1
3
5
0
,
2
0
0
6
,
d
o
i
:
1
0
.
1
0
1
6
/
j.
a
m
c
.
2
0
0
6
.
0
5
.
1
5
0
.
[2
2
]
F.
N.
Ja
rd
o
w
a
n
d
G
.
M
.
Al
-
Na
e
m
i
,
“
A
Ne
w
Hy
b
rid
Co
n
ju
g
a
te
G
ra
d
ien
t
Alg
o
rit
h
m
fo
r
Un
c
o
n
stra
i
n
e
d
Op
ti
m
iza
ti
o
n
with
I
n
e
x
a
c
t
l
in
e
se
a
rc
h
,
”
In
d
o
n
e
sia
n
J
o
u
r
n
a
l
o
f
E
lec
trica
l
E
n
g
in
e
e
rin
g
a
n
d
C
o
mp
u
ter
S
c
ien
c
e
(IJ
E
ECS
)
,
v
o
l.
2
0
,
n
o
.
2
,
p
p
.
9
3
9
-
9
4
7
,
N
o
v
e
m
b
e
r
2
0
2
0
,
d
o
i
:
1
0
.
1
1
5
9
1
/i
jee
c
s.v
2
0
.
i
2
.
p
p
9
3
9
-
9
4
7
.
[2
3
]
G
.
Zo
u
ten
d
ij
k
,
“
No
n
li
n
e
a
r
p
ro
g
ra
m
m
in
g
,
c
o
m
p
u
tati
o
n
a
l
m
e
th
o
d
s
,
”
In
teg
e
r
a
n
d
No
n
li
n
e
a
r
Pr
o
g
ra
m
min
g
,
v
o
l.
1
4
3
,
p
p
.
3
7
-
8
6
,
1
9
7
0
.
[2
4
]
I.
Bo
n
g
a
rtz,
A.
R.
Co
n
n
,
N.
G
o
u
ld
a
n
d
P
.
L
.
T
o
i
n
t
,
“
CUTE:
Co
n
s
train
e
d
a
n
d
u
n
c
o
n
stra
in
e
d
tes
ti
n
g
e
n
v
ir
o
n
m
e
n
t
,
”
ACM
T
r
a
n
sa
c
ti
o
n
s
o
n
M
a
th
e
ma
ti
c
a
l
S
o
ft
wa
re
,
v
o
l.
2
1
,
n
o
.
1
,
p
p
.
1
2
3
-
1
6
0
,
1
9
9
5
,
d
o
i
:
1
0
.
1
1
4
5
/
2
0
0
9
7
9
.
2
0
1
0
4
3
.
[2
5
]
N.
An
d
re
i,
"
An
Un
c
o
n
stra
in
e
d
O
p
ti
m
iza
ti
o
n
Tes
t
F
u
n
c
ti
o
n
s
C
o
ll
e
c
ti
o
n
,
"
A
d
v
.
M
o
d
e
l.
Op
ti
miz
a
ti
o
n
,
v
o
l.
1
0
,
n
o
.
1
,
p
p
.
1
4
7
-
1
6
1
,
2
0
0
8
.
[On
li
n
e
]
.
Av
a
il
a
b
le:
h
t
tp
:/
/www
.
a
p
m
a
th
.
sp
b
u
.
r
u
/cn
sa
/p
d
f/o
b
z
o
r/.
p
d
f
Evaluation Warning : The document was created with Spire.PDF for Python.