I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
1
1
,
No
.
2
,
A
p
r
il
2
0
2
1
,
p
p
.
1
4
5
0
~1
4
5
9
I
SS
N:
2
0
8
8
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v
1
1
i
2
.
p
p
1
4
5
0
-
1
4
5
9
1450
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
ec
e.
ia
esco
r
e.
co
m
A co
m
prehe
nsiv
e
revi
ew
on hy
brid net
w
o
rk
t
ra
ff
ic p
rediction
m
o
del
J
in
m
ei
Sh
i
1
,
Yu
-
B
eng
L
ea
u
2
,
K
un
L
i
3
,
J
o
e
H
enry
O
bit
4
1
,
2
,4
F
a
c
u
l
ty
o
f
Co
m
p
u
ti
n
g
a
n
d
In
f
o
rm
a
ti
c
s,
Un
iv
e
rsiti
M
a
la
y
sia
S
a
b
a
h
,
Ko
ta Ki
n
a
b
a
lu
,
M
a
lay
sia
1
Ha
in
a
n
Vo
c
a
ti
o
n
a
l
Un
iv
e
rsity
o
f
S
c
ien
c
e
a
n
d
T
e
c
h
n
o
l
o
g
y
,
Ha
ik
o
u
Ha
in
a
n
,
C
h
in
a
3
Co
ll
e
g
e
o
f
En
g
in
e
e
rin
g
,
Bo
h
a
i
U
n
iv
e
rsity
,
Jin
z
h
o
u
,
C
h
in
a
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
A
p
r
2
2
,
2
0
2
0
R
ev
i
s
ed
Sep
1
0
,
2
0
2
0
A
cc
ep
ted
Oct
11
,
2
0
2
0
Ne
tw
o
rk
tra
ff
i
c
is
a
t
y
p
ica
l
n
o
n
li
n
e
a
r
ti
m
e
s
e
ries
.
A
s
su
c
h
,
trad
it
i
o
n
a
l
li
n
e
a
r
a
n
d
n
o
n
l
i
n
e
a
r
m
o
d
e
l
s
a
r
e
i
n
a
d
e
q
u
a
t
e
t
o
d
e
s
c
r
i
b
e
t
h
e
m
u
l
t
i
-
s
c
a
l
e
c
h
a
ra
c
teristics
o
f
traff
ic,
th
u
s
c
o
m
p
ro
m
isin
g
th
e
p
r
e
d
ictio
n
a
c
c
u
ra
c
y
.
T
h
e
re
f
o
re
,
th
e
re
se
a
rc
h
to
d
a
te
h
a
s
ten
d
e
d
t
o
f
o
c
u
s
o
n
h
y
b
rid
m
o
d
e
ls
ra
th
e
r
th
a
n
t
h
e
trad
it
i
o
n
a
l
li
n
e
a
r
a
n
d
n
o
n
-
li
n
e
a
r
o
n
e
s.
G
e
n
e
ra
ll
y
,
a
h
y
b
rid
m
o
d
e
l
a
d
o
p
ts
tw
o
o
r
m
o
re
m
e
th
o
d
s
a
s
c
o
m
b
in
e
d
m
o
d
e
ll
in
g
to
a
n
a
ly
z
e
a
n
d
th
e
n
p
re
d
ict
th
e
n
e
tw
o
rk
traff
ic.
Ag
a
in
st
th
is
b
a
c
k
d
ro
p
,
th
is
p
a
p
e
r
w
il
l
re
v
ie
w
p
a
st
re
s
e
a
rc
h
c
o
n
d
u
c
te
d
o
n
h
y
b
rid
n
e
tw
o
rk
tra
ff
ic
p
re
d
ictio
n
m
o
d
e
ls.
T
h
e
r
e
v
ie
w
c
o
n
c
lu
d
e
s
w
it
h
a
su
m
m
a
r
y
o
f
th
e
stre
n
g
th
s
a
n
d
li
m
it
a
ti
o
n
s
o
f
e
x
isti
n
g
h
y
b
rid
n
e
tw
o
rk
p
re
d
ictio
n
m
o
d
e
ls
w
h
ich
u
se
o
p
ti
m
iza
ti
o
n
a
n
d
d
e
c
o
m
p
o
siti
o
n
tec
h
n
iq
u
e
s,
re
sp
e
c
ti
v
e
l
y
.
T
h
e
se
t
w
o
tec
h
n
iq
u
e
s
h
a
v
e
b
e
e
n
id
e
n
ti
f
ied
a
s
m
a
jo
r
c
o
n
tri
b
u
ti
n
g
f
a
c
to
rs
in
c
o
n
stru
c
ti
n
g
a
m
o
re
a
c
c
u
ra
te
a
n
d
fa
st
re
sp
o
n
se
h
y
b
rid
n
e
tw
o
rk
tra
ff
ic p
re
d
ictio
n
.
K
ey
w
o
r
d
s
:
Dec
o
m
p
o
s
itio
n
tec
h
n
iq
u
e
H
y
b
r
id
m
o
d
el
Mo
d
e
d
ec
o
m
p
o
s
itio
n
P
ar
ticu
lar
s
w
ar
m
o
p
ti
m
iza
tio
n
Qu
a
n
tu
m
g
en
et
ic
T
r
af
f
ic
p
r
ed
ictio
n
W
av
elet
tr
an
s
f
o
r
m
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
:
Yu
-
B
en
g
L
ea
u
Facu
lt
y
o
f
C
o
m
p
u
tin
g
a
n
d
I
n
f
o
r
m
at
ics
Un
i
v
er
s
iti Ma
la
y
s
ia
Sab
ah
J
alan
UM
S,
8
8
4
0
0
,
Ko
ta
Kin
ab
alu
,
Sab
ah
,
Ma
la
y
s
ia
E
m
ail:
l
y
b
en
g
@
u
m
s
.
ed
u
.
m
y
1.
I
NT
RO
D
UCT
I
O
N
W
ith
t
h
e
h
u
g
e
tid
e
o
f
t
h
e
i
n
te
r
n
et
d
r
iv
i
n
g
t
h
e
r
ap
id
d
ev
e
lo
p
m
en
t
o
f
s
o
ciet
y
,
co
m
p
u
ter
n
et
w
o
r
k
h
a
s
s
in
ce
b
ec
o
m
e
an
i
m
p
o
r
ta
n
t
tech
n
ica
l
m
ea
n
s
o
f
th
e
i
n
f
o
r
m
a
tio
n
s
o
ciet
y
.
I
n
o
r
d
er
to
g
u
ar
an
tee
th
e
q
u
ali
t
y
o
f
n
et
w
o
r
k
s
er
v
ices
s
u
c
h
as
v
id
e
o
co
n
f
er
en
ci
n
g
,
o
n
li
n
e
g
a
m
i
n
g
an
d
th
e
li
k
e,
th
e
in
cr
e
m
e
n
t
o
f
n
et
w
o
r
k
tr
a
f
f
ics
is
n
ec
es
s
ar
y
[
1
,
2
]
.
Du
e
to
t
h
e
la
r
g
e
v
o
l
u
m
e
o
f
tr
af
f
ic
f
lo
w
i
n
an
d
f
lo
w
o
u
t
i
n
t
h
e
n
et
w
o
r
k
,
d
ata
is
b
ein
g
lea
k
ed
o
r
d
is
clo
s
ed
ev
er
y
d
a
y
[
3
-
5
]
.
I
t
is
h
ar
d
to
d
etec
t
t
h
e
ab
n
o
r
m
alitie
s
as
w
ell
a
s
p
r
o
p
o
s
e
p
r
ev
en
t
iv
e
r
e
m
ed
ies
to
m
i
n
i
m
ize
t
h
e
s
ec
u
r
it
y
r
i
s
k
s
i
n
ad
v
an
ce
[
6
,
7
]
.
C
o
n
s
eq
u
e
n
tl
y
,
n
et
w
o
r
k
f
ail
u
r
es
ca
u
s
ed
b
y
p
o
ten
tial
m
a
licio
u
s
in
tr
u
s
i
o
n
s
a
n
d
v
ir
u
s
i
n
v
asio
n
s
h
a
v
e
tr
i
g
g
er
ed
s
er
io
u
s
c
o
n
ce
r
n
s
a
m
o
n
g
t
h
e
n
et
w
o
r
k
m
a
n
a
g
e
m
e
n
t
an
d
m
o
n
ito
r
i
n
g
tea
m
[
8
-
1
0
]
.
Net
w
o
r
k
tr
af
f
ic
is
an
i
m
p
o
r
tan
t
p
ar
a
m
eter
to
ev
alu
ate
t
h
e
r
u
n
n
i
n
g
s
tate
o
f
a
n
et
w
o
r
k
.
I
t
i
s
f
o
u
n
d
to
b
e
a
n
o
n
li
n
ea
r
ti
m
e
s
er
ies
[
1
1
]
w
h
ic
h
h
a
s
t
h
e
c
h
ar
ac
ter
is
tic
s
o
f
t
i
m
e
-
v
ar
iab
ilit
y
,
lo
n
g
-
ter
m
co
r
r
elatio
n
,
s
e
lf
-
s
i
m
ilar
it
y
,
s
u
d
d
en
n
e
s
s
an
d
c
h
ao
s
[
1
2
]
.
T
h
er
ef
o
r
e,
a
m
o
r
e
a
cc
u
r
ate
a
n
d
f
a
s
t
r
esp
o
n
s
e
tr
a
f
f
ic
p
r
ed
ictio
n
m
o
d
el
i
s
m
u
c
h
d
e
s
i
r
e
d
t
o
e
n
s
u
r
e
a
s
a
f
e
a
n
d
h
e
a
l
t
h
y
n
e
tw
o
r
k
s
i
t
u
a
t
i
o
n
.
A
c
c
o
r
d
i
n
g
t
o
J
o
s
h
i
[
1
3
]
,
n
e
t
w
o
r
k
t
r
a
f
f
i
c
p
r
e
d
i
c
t
i
o
n
i
s
a
r
e
l
i
a
b
l
e
m
e
t
h
o
d
t
o
s
e
c
u
r
e
t
h
e
n
e
tw
o
r
k
c
o
m
m
u
n
i
c
a
t
i
o
n
i
n
a
n
e
t
w
o
r
k
m
a
n
a
g
e
m
e
n
t
a
n
d
m
o
n
i
to
r
in
g
s
y
s
te
m
.
I
t
i
s
a
p
r
o
ce
s
s
w
h
ic
h
an
al
y
s
e
s
th
e
ch
ar
ac
ter
is
tic
s
o
f
tr
af
f
ic
in
t
h
e
p
ast
an
d
p
r
esen
t,
g
en
er
ates
t
h
e
r
u
les
o
f
i
n
t
er
n
al
s
tr
u
ct
u
r
e
an
d
th
e
n
co
n
s
tr
u
ct
s
a
m
o
d
el
to
p
r
ed
ict
th
e
ch
ar
ac
ter
is
tics
a
n
d
tr
en
d
s
o
f
f
u
tu
r
e
tr
a
f
f
ic.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
co
mp
r
eh
en
s
ive
r
ev
iew
o
n
h
y
b
r
id
n
etw
o
r
k
tr
a
ffic p
r
ed
ictio
n
mo
d
el
(
Jin
mei
S
h
i
)
1451
A
h
y
b
r
id
m
o
d
el
w
as
f
ir
s
t
p
r
o
p
o
s
ed
b
y
B
ates
an
d
Gr
an
g
er
wh
o
in
te
g
r
ated
th
e
m
er
its
o
f
t
h
e
in
d
iv
id
u
al
m
o
d
el
s
[
1
4
]
.
Ho
w
e
v
er
,
Dick
i
n
s
o
n
co
n
f
ir
m
ed
t
h
at
th
e
v
ar
ia
n
ce
er
r
o
r
o
f
th
e
h
y
b
r
id
p
r
ed
ictin
g
m
o
d
el
is
le
s
s
th
an
a
n
y
o
f
t
h
e
s
in
g
le
m
o
d
el
s
[1
5
]
.
I
n
Yan
g
et
a
l
.
s
im
ilar
l
y
co
n
ten
d
ed
th
at
th
e
p
r
ed
ictio
n
er
r
o
r
o
f
a
h
y
b
r
id
m
o
d
el
is
lo
w
er
th
a
n
a
s
in
g
le
n
o
n
li
n
ea
r
m
o
d
el.
Giv
e
n
th
at,
it
is
f
ea
s
ib
le
to
u
s
e
h
y
b
r
id
m
o
d
els
to
p
r
ed
ict
th
e
n
et
w
o
r
k
tr
af
f
ic
[
1
6
]
.
Sin
ce
t
h
e
ea
r
lier
w
o
r
k
s
,
a
co
n
s
id
er
ab
le
a
m
o
u
n
t
o
f
liter
at
u
r
e
h
as
b
ee
n
p
u
b
li
s
h
ed
o
n
h
y
b
r
id
n
et
w
o
r
k
p
r
ed
ictio
n
m
o
d
els.
T
h
ese
s
tu
d
ies
g
e
n
er
all
y
ap
p
lied
th
e
o
p
ti
m
izatio
n
a
n
d
d
ec
o
m
p
o
s
itio
n
tech
n
iq
u
es i
n
t
h
e
h
y
b
r
id
m
o
d
el
to
p
r
o
d
u
ce
a
b
etter
p
r
e
d
ictio
n
ac
cu
r
ac
y
.
I
n
th
i
s
p
ap
er
,
th
e
r
ec
en
t
h
y
b
r
id
m
o
d
els
ar
e
co
m
p
r
e
h
en
s
iv
e
l
y
r
ev
ie
w
ed
.
T
h
e
p
ap
er
is
s
tr
u
ctu
r
ed
a
s
f
o
llo
w
s
.
Sectio
n
2
r
ev
ie
w
s
t
h
e
ap
p
licatio
n
o
f
o
p
ti
m
izatio
n
te
ch
n
iq
u
e
i
n
th
e
h
y
b
r
id
m
o
d
el
w
h
ile
th
e
u
tili
za
tio
n
o
f
d
ec
o
m
p
o
s
i
tio
n
tec
h
n
iq
u
es
i
n
th
e
h
y
b
r
id
m
o
d
el
w
il
l
b
e
in
v
esti
g
ated
i
n
th
e
f
o
llo
w
i
n
g
s
e
ctio
n
.
T
h
e
f
in
a
l
p
ar
t
o
f
th
e
p
ap
er
co
n
clu
d
es
w
it
h
a
s
u
m
m
ar
y
.
2.
O
P
T
I
M
I
Z
AT
I
O
N
T
E
CH
NI
Q
U
E
S
-
B
ASE
D
H
YB
RID M
O
DE
L
I
n
o
r
d
er
to
b
etter
d
escr
ib
e
th
e
tr
af
f
ic
ch
ar
ac
ter
is
tics
an
d
i
m
p
r
o
v
e
th
e
ac
cu
r
ac
y
o
f
t
h
e
p
r
ed
ictio
n
m
o
d
el,
r
esear
c
h
er
s
h
av
e
r
ec
e
n
tl
y
co
m
b
i
n
ed
v
ar
io
u
s
m
e
th
o
d
s
in
to
a
s
i
n
g
le
p
r
ed
ictio
n
m
o
d
el
an
d
o
p
tim
ize
it
w
ith
d
if
f
e
r
en
t
o
p
t
im
iz
a
t
i
o
n
t
ec
h
n
i
q
u
e
s
s
u
ch
a
s
p
a
r
ti
cu
l
a
r
s
w
ar
m
o
p
t
im
iz
a
t
i
o
n
(
PS
O
)
[
1
7
]
,
a
n
d
q
u
an
tu
m
g
en
e
ti
c
(
Q
G
)
[
1
8
,
19
]
am
o
n
g
o
t
h
e
r
s
.
2
.
1
.
P
SO
-
ba
s
ed
hy
brid
m
o
del
T
h
e
r
an
d
o
m
d
eter
m
i
n
atio
n
o
f
th
e
in
p
u
t
w
e
ig
h
t
s
an
d
h
id
d
en
b
iases
[
20
]
o
f
ex
tr
e
m
e
lear
n
i
n
g
m
ac
h
i
n
e
(
E
L
M
)
c
a
n
l
e
a
d
t
o
i
l
l
-
c
o
n
d
i
t
i
o
n
p
r
o
b
l
e
m
,
r
e
s
u
l
t
i
n
g
i
n
l
o
w
p
r
e
d
i
c
t
i
o
n
.
I
n
F
e
i
H
a
n
s
e
l
e
c
t
e
d
P
S
O
a
lg
o
r
it
h
m
w
ith
s
i
m
p
le
p
r
in
cip
le,
a
n
d
p
r
o
p
o
s
ed
th
e
A
PSO
-
E
L
M
h
y
b
r
id
m
o
d
el
to
ad
d
r
ess
th
e
d
r
a
w
b
ac
k
s
o
f
E
L
M.
He
ad
o
p
ted
ad
ap
tiv
e
alg
o
r
it
h
m
to
o
p
ti
m
iz
e
P
SO
w
h
ich
s
elec
t
s
t
h
e
i
n
p
u
t
w
e
ig
h
t
s
a
n
d
h
id
d
en
b
iases
.
T
h
en
h
e
u
s
ed
m
o
o
r
e
-
p
en
r
o
s
e
(
MP
)
g
en
er
alize
d
in
v
er
s
e
to
a
n
al
y
ticall
y
d
eter
m
in
e
th
e
o
u
tp
u
t
w
ei
g
h
ts
o
f
E
L
M.
I
n
o
r
d
er
to
o
b
tain
o
p
tim
a
l
p
ar
a
m
eter
s
o
f
E
L
M,
t
h
e
i
m
p
r
o
v
ed
P
SO
o
p
ti
m
ize
s
t
h
e
i
n
p
u
t
w
eig
h
t
s
a
n
d
h
id
d
en
b
iases
.
I
n
t
h
i
s
ca
s
e
,
th
e
m
o
d
el
n
o
t
o
n
l
y
o
b
tain
s
th
e
o
p
tim
al
r
o
o
t
m
ea
n
s
q
u
ar
e
er
r
o
r
(
R
MSE
)
,
b
u
t
it
also
o
b
tain
s
th
e
o
p
tim
al
o
u
tp
u
t
w
ei
g
h
t
n
o
r
m
.
T
h
is
d
ir
ec
tl
y
s
o
l
v
es
t
h
e
p
r
o
b
le
m
s
ca
u
s
ed
b
y
"
r
an
d
o
m
n
e
s
s
"
o
f
E
L
M
a
n
d
co
n
s
eq
u
en
tl
y
i
m
p
r
o
v
e
s
th
e
ac
cu
r
ac
y
o
f
t
h
e
p
r
ed
ictio
n
m
o
d
el.
I
t
s
h
o
u
ld
b
e
n
o
t
ed
th
at
h
i
s
p
ap
er
w
i
ll
o
n
l
y
f
o
cu
s
o
n
th
e
p
ar
a
m
eter
p
r
o
b
lem
o
f
E
L
M
w
h
ic
h
w
ill
a
f
f
ec
t
th
e
ac
c
u
r
ac
y
o
f
t
h
e
m
o
d
el
b
u
t
w
i
ll
ig
n
o
r
e
th
e
d
r
a
w
b
ac
k
s
o
f
lo
ca
l
o
p
ti
m
al
s
o
lu
tio
n
o
f
P
SO [
2
1
]
.
T
h
e
P
SO
alg
o
r
it
h
m
w
it
h
s
i
m
p
le
p
r
in
cip
le
a
n
d
f
e
w
p
ar
a
m
eter
s
ca
n
s
h
o
r
ten
t
h
e
tr
ai
n
i
n
g
s
p
ee
d
o
f
n
eu
r
al
n
et
w
o
r
k
,
w
h
ich
in
tu
r
n
i
m
p
r
o
v
es
th
e
co
n
v
er
g
en
ce
s
p
e
ed
o
f
th
e
m
o
d
el.
B
ased
o
n
t
h
i
s
id
ea
,
in
Yi
Ya
n
g
,
co
m
b
i
n
ed
th
e
t
h
r
ee
alg
o
r
ith
m
m
o
d
els
an
d
p
r
o
p
o
s
ed
a
n
e
w
h
y
b
r
id
m
et
h
o
d
w
h
ic
h
s
h
e
ca
ll
ed
SP
L
SS
VM
.
S
h
e
also
u
s
ed
P
SO
to
o
p
ti
m
ize
t
h
e
t
w
o
p
ar
a
m
e
ter
s
o
f
leas
t
s
q
u
ar
es
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
i
n
e
(
L
SS
V
M)
.
T
h
en
,
b
ased
o
n
th
e
s
ea
s
o
n
al
ad
j
u
s
t
m
en
t
(
S
A
)
a
n
d
L
SS
VM
,
s
h
e
r
ed
u
ce
d
th
e
s
ea
s
o
n
al
in
ter
f
e
r
en
ce
o
n
t
h
e
tr
a
f
f
ic
co
m
p
o
n
e
n
t
s
to
p
r
ed
ict
n
et
w
o
r
k
tr
a
f
f
ic
[
1
7
]
.
B
u
t
t
h
e
h
y
b
r
id
m
o
d
el
o
n
l
y
co
n
s
id
er
s
t
h
e
s
ea
s
o
n
al
ch
ar
ac
ter
i
s
tic
s
o
f
tr
af
f
ic
an
d
also
i
g
n
o
r
es t
h
e
p
r
o
b
lem
o
f
lo
ca
l o
p
ti
m
al
s
o
lu
t
io
n
o
f
P
SO.
B
ased
o
n
th
e
s
a
m
e
d
esi
g
n
id
ea
,
W
eij
ie
Z
h
a
n
g
s
i
m
ilar
l
y
u
tili
s
ed
P
SO
alg
o
r
ith
m
w
i
th
s
i
m
p
l
e
p
r
in
cip
le
to
o
p
ti
m
ize
s
tr
u
ct
u
r
al
p
ar
a
m
et
er
s
o
f
R
B
F
n
eu
r
al
n
et
w
o
r
k
.
B
y
ad
j
u
s
ti
n
g
t
h
e
i
n
er
tia
w
ei
g
h
t
an
d
lear
n
i
n
g
f
ac
to
r
to
i
m
p
r
o
v
e
t
h
e
g
lo
b
al
s
ea
r
c
h
ab
ilit
y
in
t
h
e
g
lo
b
al
ex
tr
e
m
u
m
s
ea
r
ch
,
h
e
w
a
s
ab
le
to
s
o
l
v
e
a
n
d
av
o
id
lo
ca
l
o
p
tim
a
l
s
o
l
u
tio
n
o
f
P
SO.
T
h
en
h
e
o
p
ti
m
ized
th
e
f
o
u
r
p
ar
a
m
eter
s
o
f
th
e
R
B
F
s
o
as
to
o
b
tain
th
e
ac
c
u
r
ac
y
o
f
th
e
p
r
ed
ictio
n
m
o
d
el
[
2
2
]
.
L
am
en
tab
l
y
,
in
t
h
e
p
r
o
ce
s
s
o
f
o
b
tain
in
g
g
lo
b
al
o
p
tim
al
s
o
l
u
ti
o
n
s
,
R
B
F
h
a
s
to
o
m
an
y
p
ar
a
m
eter
s
to
b
e
o
p
ti
m
i
ze
d
,
an
d
th
i
s
w
il
l
i
n
cr
ea
s
e
t
h
e
ca
lcu
latio
n
s
ca
le,
th
e
tr
ain
i
n
g
ti
m
e
a
n
d
a
f
f
ec
t
t
h
e
co
n
v
er
g
e
n
ce
r
ate.
I
n
th
e
v
ie
w
th
a
t
PS
O
al
g
o
r
ith
m
h
as
a
s
i
m
p
le
p
r
in
cip
le
b
u
t
l
o
ca
l
o
p
tim
a
l
s
o
l
u
tio
n
p
r
o
b
le
m
,
He
et
a
l
.
in
2
0
1
6
,
in
tr
o
d
u
ce
d
Qu
an
t
u
m
n
o
n
-
g
a
te
to
r
ea
lize
m
u
tatio
n
o
p
er
atio
n
.
He
u
s
ed
p
ar
ticle
f
lig
h
t
p
ath
in
f
o
r
m
at
io
n
to
d
y
n
a
m
icall
y
u
p
d
ate
th
e
s
tat
u
s
o
f
q
u
an
t
u
m
b
it so
as to
av
o
i
d
lo
c
al
o
p
tim
izatio
n
d
r
a
w
b
ac
k
s
o
f
P
SO.
T
h
en
,
h
e
u
s
ed
I
P
SO
to
o
p
ti
m
ize
t
h
e
w
ei
g
h
t,
w
id
t
h
a
n
d
ce
n
ter
p
o
s
itio
n
p
ar
a
m
eter
s
o
f
r
ad
ial
b
asis
f
u
n
ct
io
n
(
R
B
F)
n
et
w
o
r
k
.
He
w
as
ab
le
to
r
ea
li
ze
o
p
ti
m
ized
p
ar
a
m
eter
s
n
e
u
r
al
n
et
w
o
r
k
an
d
e
s
tab
lis
h
s
el
f
-
ad
ap
tiv
e
P
SO
-
R
B
F
h
y
b
r
id
m
o
d
el.
As
s
u
ch
,
t
h
e
n
et
w
o
r
k
tr
af
f
ic
d
ata
w
it
h
n
o
n
li
n
ea
r
ch
ar
ac
ter
is
tic
s
ca
n
b
e
p
r
ed
icted
an
d
th
e
d
if
f
ic
u
lt
y
o
f
p
r
ed
ictio
n
is
r
ed
u
ce
d
s
ig
n
i
f
ica
n
tl
y
[
2
3
]
.
I
n
ev
ita
b
ly
,
h
e
s
o
l
v
ed
th
e
p
r
o
b
lem
o
f
P
SO,
y
et
n
e
g
lecte
d
th
e
co
m
p
le
x
ca
lc
u
lat
io
n
p
r
in
ci
p
le
o
f
Q
u
an
t
u
m
al
g
o
r
ith
m
.
C
o
n
s
id
er
in
g
th
e
d
i
f
f
ic
u
lt
y
o
f
ap
p
r
o
ac
h
in
g
t
h
e
g
lo
b
al
o
p
tim
a
l so
lu
tio
n
,
t
h
e
p
r
ed
ictio
n
ac
cu
r
ac
y
is
co
n
s
eq
u
e
n
tl
y
a
f
f
ec
ted
.
As
d
is
c
u
s
s
ed
ab
o
v
e,
t
h
ese
a
u
t
h
o
r
s
h
a
v
e
u
s
ed
d
if
f
er
en
t
m
e
th
o
d
s
to
s
o
lv
e
th
e
lo
ca
l
o
p
ti
m
al
s
o
l
u
tio
n
p
r
o
b
lem
o
f
P
SO.
Ho
w
ev
er
,
t
h
e
y
h
av
e
a
ll
i
g
n
o
r
ed
th
e
s
al
ien
t
f
ac
t
th
at
th
er
e
ar
e
to
o
m
a
n
y
p
ar
am
eter
s
o
f
R
B
F
o
p
tim
izatio
n
a
n
d
th
e
tr
ain
i
n
g
ti
m
e
is
to
o
lo
n
g
to
ap
p
r
o
x
im
ate
th
e
o
p
ti
m
a
l
s
o
lu
tio
n
.
I
n
ef
f
ec
t,
th
e
o
p
ti
m
al
s
o
lu
tio
n
o
f
P
SO
w
as
n
o
t e
f
f
ec
t
iv
el
y
s
o
lv
ed
,
t
h
u
s
af
f
ec
ti
n
g
th
e
co
n
v
er
g
e
n
ce
s
p
ee
d
an
d
ac
cu
r
ac
y
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
11
,
No
.
2
,
A
p
r
il 2
0
2
1
:
1
4
5
0
-
1459
1452
2
.
2
.
QG
-
ba
s
ed
hy
br
id
m
o
del
C
o
n
s
id
er
in
g
t
h
at
t
h
e
li
m
itatio
n
o
f
tr
ad
itio
n
a
l
n
et
w
o
r
k
tr
af
f
i
c
ti
m
e
s
er
ie
s
p
r
ed
ictio
n
m
o
d
el
an
d
th
e
p
r
o
b
lem
th
a
t
b
ac
k
p
r
o
p
ag
atio
n
(
B
P
)
n
eu
r
al
n
et
w
o
r
k
i
s
ea
s
y
to
g
et
i
n
to
lo
ca
l
s
o
lu
t
io
n
.
T
h
e
li
m
itatio
n
s
o
f
tr
ad
itio
n
al
n
et
w
o
r
k
tr
a
f
f
ic
ti
m
e
s
er
ies
p
r
ed
ictio
n
m
o
d
el
an
d
th
e
B
P
n
eu
r
al
n
et
w
o
r
k
h
av
e
b
ee
n
w
id
el
y
ac
k
n
o
w
led
g
ed
in
th
e
liter
at
u
r
e.
I
n
th
is
lig
h
t,
in
K
u
n
Z
h
an
g
in
tr
o
d
u
ce
d
th
e
Qu
a
n
t
u
m
a
l
g
o
r
ith
m
w
it
h
s
tr
o
n
g
g
lo
b
al
o
p
ti
m
izatio
n
ab
ilit
y
a
n
d
th
e
P
SO
al
g
o
r
it
h
m
w
i
th
s
i
m
p
le
p
r
in
c
ip
le.
He
ap
p
lied
th
e
co
m
b
in
a
tio
n
alg
o
r
ith
m
to
s
o
lv
e
t
h
e
g
r
ad
ien
t
ex
p
lo
s
io
n
p
r
o
b
le
m
o
f
B
P
,
an
d
h
e
t
h
en
p
r
o
p
o
s
ed
th
e
QP
SO
-
B
P
h
y
b
r
id
m
o
d
el.
T
o
elab
o
r
ate,
w
h
e
n
QP
SO
alg
o
r
ith
m
is
ap
p
lied
to
th
e
tr
ain
i
n
g
s
ta
g
e
o
f
p
r
ed
ictio
n
m
o
d
el,
a
s
et
o
f
w
e
ig
h
t
s
th
at
m
i
n
i
m
ize
t
h
e
er
r
o
r
f
u
n
c
tio
n
i
n
co
m
p
eti
tiv
e
ti
m
e
ca
n
b
e
o
b
tain
ed
.
T
h
e
w
ei
g
h
t
is
u
p
d
ated
g
r
ad
u
all
y
u
n
til
t
h
e
co
n
v
er
g
e
n
ce
cr
iter
io
n
is
s
ati
s
f
ied
.
Af
ter
t
h
at,
th
e
o
b
j
ec
ti
v
e
f
u
n
ctio
n
to
ac
h
ie
v
e
t
h
e
m
i
n
i
m
izatio
n
is
t
h
e
p
r
ed
ictio
n
er
r
o
r
f
u
n
ctio
n
s
o
as
to
i
m
p
r
o
v
e
t
h
e
p
r
ed
ictio
n
ac
cu
r
ac
y
o
f
th
e
m
o
d
el
[
2
4
]
.
I
n
ev
itab
l
y
,
in
th
e
p
r
o
ce
s
s
o
f
iter
ativ
e
tr
ain
i
n
g
,
t
h
e
co
m
p
u
tat
io
n
s
ca
le
o
f
Qu
a
n
tu
m
al
g
o
r
ith
m
is
v
er
y
h
ea
v
y
[
2
5
]
.
So
,
it
is
s
till
d
if
f
ic
u
lt to
ap
p
r
o
ac
h
th
e
g
lo
b
al
o
p
tim
izatio
n
a
n
d
o
b
tain
th
e
l
o
ca
l o
p
tim
a
l
s
o
lu
tio
n
.
I
n
v
ie
w
o
f
th
e
s
h
o
r
ta
g
e
o
f
B
P
n
eu
r
al
n
et
w
o
r
k
,
t
h
e
p
r
ed
ictio
n
er
r
o
r
an
d
j
itter
o
f
th
e
m
o
d
el
ar
e
ea
s
il
y
lar
g
e.
Hu
i
T
ian
also
u
s
ed
alg
o
r
ith
m
to
o
p
ti
m
ize
t
h
e
s
tr
u
ctu
r
e
o
f
B
P
n
et
w
o
r
k
b
ased
o
n
ef
f
i
cien
t
g
lo
b
al
s
ea
r
c
h
ca
p
ab
ilit
y
o
f
q
u
a
n
tu
m
g
e
n
etic
alg
o
r
ith
m
(
QG
A
)
.
T
h
en
,
h
e
ap
p
lied
w
av
e
let
tec
h
n
o
lo
g
y
(
W
T
)
to
d
ec
o
m
p
o
s
e
th
e
tr
a
f
f
ic
i
n
to
lo
w
f
r
eq
u
e
n
c
y
a
n
d
h
i
g
h
f
r
eq
u
en
c
y
d
ata.
He
s
u
b
s
eq
u
e
n
tl
y
p
r
o
p
o
s
ed
W
T
-
QGA
-
B
P
h
y
b
r
id
m
o
d
el
to
p
r
ed
ict
th
e
ch
a
o
s
o
f
n
et
w
o
r
k
tr
af
f
ic.
Ho
w
e
v
er
,
th
e
m
o
d
el
ig
n
o
r
ed
t
w
o
i
m
p
o
r
tan
t
f
a
ct
s
th
e
w
av
e
let
tech
n
o
lo
g
y
h
a
s
s
i
g
n
al
n
o
is
e
with
th
e
s
i
g
n
al,
a
n
d
t
h
e
Q
u
a
n
tu
m
m
ec
h
an
ical
ca
lc
u
latio
n
s
ca
l
e
is
to
o
lar
g
e.
T
h
es
e
is
s
u
es
w
il
l in
cr
ea
s
e
t
h
e
co
m
p
l
ex
it
y
o
f
t
h
e
h
y
b
r
id
m
o
d
el,
r
es
u
lti
n
g
i
n
lo
w
g
e
n
er
aliza
tio
n
p
er
f
o
r
m
an
ce
[
2
6
]
.
As
th
e
a
lg
o
r
it
h
m
d
e
v
elo
p
ed
an
d
ev
o
l
v
ed
,
r
esear
ch
er
s
f
o
u
n
d
th
at
alt
h
o
u
g
h
th
e
f
r
u
it
f
l
y
o
p
ti
m
izatio
n
alg
o
r
it
h
m
(
FO
A
)
ca
n
ea
s
il
y
f
all
in
to
th
e
p
r
o
b
le
m
o
f
lo
ca
l
o
p
tim
izatio
n
s
o
l
u
tio
n
,
it
s
till
h
as
t
h
e
s
tr
en
g
t
h
s
o
f
s
i
m
p
le
ca
lc
u
latio
n
a
n
d
co
d
in
g
co
n
v
en
ie
n
ce
.
I
n
Yin
g
Han
u
s
ed
Qu
an
t
u
m
m
ec
h
a
n
ic
s
th
eo
r
y
to
o
p
ti
m
ize
FO
A
.
T
h
en
h
e
u
s
ed
QFO
A
to
o
p
tim
ize
f
iv
e
i
m
p
o
r
tan
t
p
ar
a
m
et
er
s
o
f
ec
h
o
s
tate
n
et
w
o
r
k
(
E
SN)
an
d
p
r
o
p
o
s
ed
QFO
A
-
E
SN
h
y
b
r
id
m
o
d
el
to
p
r
o
v
id
e
m
o
d
el
ac
cu
r
ac
y
.
T
h
e
m
o
d
el
is
b
est
d
escr
ib
ed
as
f
o
llo
w
s
:
f
ir
s
t,
t
h
e
p
h
ase
-
s
p
ac
e
r
ec
o
n
s
tr
u
ctio
n
t
ec
h
n
o
lo
g
y
is
u
s
ed
to
r
ec
o
n
s
tr
u
ct
th
e
o
r
ig
i
n
al
n
et
w
o
r
k
tr
af
f
ic
d
ata
s
er
ie
s
;
af
ter
w
ar
d
s
,
t
h
e
E
SN
m
e
th
o
d
is
u
s
ed
to
b
u
ild
t
h
e
p
r
ed
ictio
n
m
o
d
el.
Me
an
w
h
ile
th
e
m
o
d
el
p
ar
am
eter
s
ar
e
o
p
tim
ized
b
y
th
e
Q
FO
A
.
Fi
n
all
y
,
t
h
e
o
p
ti
m
al
E
S
N
m
o
d
el
is
u
s
ed
f
o
r
m
u
lti
-
s
tep
p
r
ed
ictio
n
f
o
r
th
e
n
et
w
o
r
k
tr
af
f
ic
[
2
7
]
.
Ho
w
ev
er
,
th
e
m
o
d
el
h
as
to
o
m
a
n
y
o
p
ti
m
i
za
tio
n
p
ar
a
m
eter
s
an
d
th
e
c
o
m
p
u
tati
o
n
s
ca
le
o
f
Qu
a
n
tu
m
ca
n
b
ec
o
m
e
lar
g
er
.
Hen
ce
,
in
t
h
e
p
r
o
ce
s
s
o
f
tr
ai
n
in
g
,
th
e
s
e
w
i
ll
a
f
f
ec
t
th
e
co
n
v
er
g
en
ce
r
ate
a
n
d
ac
cu
r
ac
y
o
f
t
h
e
h
y
b
r
id
m
o
d
el.
B
o
th
au
t
h
o
r
s
(
i.e
.
,
n
a
m
es
)
e
v
id
en
tl
y
s
o
l
v
ed
th
e
l
i
m
itatio
n
s
o
f
n
e
u
r
al
n
et
w
o
r
k
.
Yet,
th
e
y
all
ig
n
o
r
ed
th
e
f
ac
t
th
at
al
th
o
u
g
h
t
h
e
Q
u
a
n
tu
m
alg
o
r
ith
m
ca
n
s
o
l
v
e
t
h
e
g
lo
b
al
o
p
ti
m
izatio
n
,
th
e
tr
ain
i
n
g
ti
m
e
o
f
th
e
m
o
d
el
d
u
e
to
th
e
lar
g
e
a
m
o
u
n
t o
f
co
m
p
u
tatio
n
a
n
d
co
m
p
le
x
ca
lc
u
latio
n
s
ca
le
i
s
in
cr
ea
s
ed
.
T
h
is
m
a
k
es
it
d
i
f
f
ic
u
lt
to
s
o
l
v
e
th
e
o
p
ti
m
al
s
o
lu
tio
n
w
h
ic
h
w
il
l
in
tu
r
n
af
f
ec
t
th
e
ac
c
u
r
ac
y
an
d
co
n
v
er
g
e
n
ce
r
ate
o
f
th
e
p
r
ed
ictio
n
m
o
d
el.
2
.
3
.
O
t
her
hy
brid
m
o
del
R
B
F
n
e
u
r
al
n
et
w
o
r
k
h
a
s
th
e
a
d
v
an
ta
g
e
o
f
g
lo
b
al
a
p
p
r
o
x
i
m
a
tio
n
to
n
o
n
l
in
ea
r
f
u
n
ctio
n
.
He
n
ce
,
it
ca
n
p
r
ed
ict
th
e
n
e
t
w
o
r
k
tr
a
f
f
ic
d
at
a
w
ith
n
o
n
li
n
ea
r
ch
ar
ac
ter
i
s
tic
s
.
B
ased
o
n
t
h
i
s
,
in
De
n
g
f
e
n
g
W
ei
in
tr
o
d
u
ce
d
t
h
e
g
r
av
it
y
s
ea
r
c
h
alg
o
r
it
h
m
(
GS
A
)
to
o
p
ti
m
ize
th
e
R
B
F
n
et
wo
r
k
s
tr
u
ct
u
r
e
an
d
i
m
p
r
o
v
e
t
h
e
co
n
v
er
g
e
n
c
e
r
ate
o
f
th
e
p
r
ed
ictio
n
m
o
d
el.
On
o
n
e
h
an
d
,
th
e
m
et
h
o
d
ca
n
o
p
ti
m
iz
e
th
e
p
ar
am
e
ter
s
s
u
ch
as
t
h
e
c
en
ter
ci
o
f
th
e
b
asic
f
u
n
ctio
n
s
o
f
h
id
d
en
u
n
it
s
,
w
i
d
th
r
i
an
d
n
et
w
o
r
k
co
n
n
ec
tio
n
w
eig
h
t
s
w
k
j
o
f
R
B
F.
On
t
h
e
o
th
er
,
th
e
f
itti
n
g
r
esu
lt
a
n
d
n
o
n
li
n
ea
r
ap
p
r
o
x
i
m
ati
o
n
ab
ilit
y
o
f
R
B
F
n
e
u
r
al
n
e
t
w
o
r
k
ar
e
b
etter
u
s
ed
to
o
b
tain
th
e
o
p
ti
m
al
n
e
u
r
al
n
et
w
o
r
k
p
r
ed
ictio
n
m
o
d
el.
I
n
th
e
iter
atio
n
p
r
o
ce
s
s
,
R
B
F
p
ar
a
m
eter
s
ar
e
la
m
e
n
tab
l
y
to
o
m
a
n
y
to
b
e
o
p
tim
ized
s
o
as to
o
b
tain
lo
ca
l o
p
tim
al
s
o
lu
tio
n
p
r
o
b
le
m
[
1
1
]
.
A
i
m
i
n
g
at
th
e
g
r
ad
ien
t
ex
p
lo
s
io
n
o
f
B
P
n
eu
r
al
n
et
w
o
r
k
an
d
th
e
lo
ca
l
o
p
tim
al
s
o
l
u
tio
n
o
f
l
o
n
g
s
h
o
r
t
-
ter
m
m
e
m
o
r
y
(
L
ST
M)
.
I
n
v
ie
w
o
f
th
e
g
r
ad
ien
t
e
x
p
lo
s
io
n
o
f
B
P
n
eu
r
al
n
et
w
o
r
k
a
n
d
th
e
lo
ca
l
o
p
ti
m
al
s
o
l
u
tio
n
o
f
lo
n
g
s
h
o
r
t
-
ter
m
m
e
m
o
r
y
(
L
ST
M)
,
A
zz
o
u
n
i,
in
2
0
1
7
,
p
r
o
p
o
s
ed
a
L
ST
M
-
R
NN
(
r
ec
u
r
r
en
t
n
eu
r
al
n
et
w
o
r
k
,
R
NN)
h
y
b
r
id
f
r
a
m
e
w
o
r
k
to
p
r
ed
ict
tr
af
f
ic
m
atr
i
x
.
B
y
v
al
id
atin
g
t
h
e
f
r
a
m
e
w
o
r
k
o
n
r
e
al
-
w
o
r
ld
d
ata
f
r
o
m
GE
A
NT
n
et
w
o
r
k
,
h
e
u
s
ed
th
e
s
lid
in
g
lear
n
in
g
w
i
n
d
o
w
m
et
h
o
d
to
s
o
lv
e
t
h
e
L
ST
M
n
e
u
r
al
n
et
w
o
r
k
li
m
ita
tio
n
s
.
T
h
e
n
h
e
co
m
b
in
e
d
it
w
ith
R
NN
n
e
u
r
al
n
et
w
o
r
k
to
ex
tr
ac
t
th
e
d
y
n
a
m
ic
ch
ar
ac
ter
is
tics
o
f
n
et
w
o
r
k
tr
af
f
ic
an
d
p
r
ed
ict
th
e
f
u
t
u
r
e
tr
af
f
ic.
A
lt
h
o
u
g
h
h
i
s
w
o
r
k
m
an
ag
ed
to
s
o
l
v
e
t
h
e
L
ST
M’
s
li
m
ita
tio
n
u
s
i
n
g
t
h
e
s
lid
in
g
lear
n
i
n
g
w
i
n
d
o
w
m
e
th
o
d
,
th
e
to
tal
n
u
m
b
er
o
f
ti
m
e
s
lo
ts
b
ec
a
m
e
to
o
lar
g
e,
r
esu
lt
in
g
i
n
h
i
g
h
co
m
p
u
tatio
n
al
co
m
p
le
x
it
y
[
2
8
]
.
I
n
th
e
s
a
m
e
y
ea
r
,
b
ased
o
n
th
e
s
a
m
e
p
r
o
b
lem
o
f
L
ST
M
n
eu
r
al
n
et
w
o
r
k
,
Qin
z
h
e
n
g
Z
h
u
o
p
r
o
p
o
s
ed
a
m
o
d
el
o
f
n
e
u
r
al
n
et
w
o
r
k
w
h
i
ch
ca
n
b
e
u
s
ed
to
co
m
b
i
n
e
L
ST
M
w
it
h
d
ee
p
n
e
u
r
al
n
et
w
o
r
k
s
(
D
NN)
.
T
h
e
ai
m
w
a
s
to
s
o
lv
e
th
e
n
et
w
o
r
k
tr
a
f
f
ic
p
r
ed
ictio
n
o
f
a
u
to
co
r
r
elatio
n
n
o
n
li
n
ea
r
ti
m
e
s
er
ies
d
ata.
Au
to
-
co
r
r
elatio
n
co
ef
f
icie
n
t
i
s
th
e
n
ad
d
ed
to
t
h
e
m
o
d
el
to
i
m
p
r
o
v
e
th
e
ac
c
u
r
ac
y
o
f
t
h
e
p
r
ed
ictio
n
m
o
d
el
.
T
h
is
m
o
d
el
b
o
asts
h
ig
h
er
p
r
ec
is
io
n
w
h
e
n
co
m
p
ar
ed
to
th
e
o
t
h
er
tr
ad
itio
n
al
m
o
d
els.
A
f
ter
co
n
s
id
er
in
g
t
h
e
a
u
to
co
r
r
elatio
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
co
mp
r
eh
en
s
ive
r
ev
iew
o
n
h
y
b
r
id
n
etw
o
r
k
tr
a
ffic p
r
ed
ictio
n
mo
d
el
(
Jin
mei
S
h
i
)
1453
f
ea
t
u
r
es,
th
e
n
e
u
r
al
n
et
w
o
r
k
o
f
L
ST
M
an
d
DNN
h
as
c
lear
ad
v
an
ta
g
es
i
n
t
h
e
ac
c
u
r
ac
y
o
f
t
h
e
lar
g
e
g
r
an
u
lar
it
y
d
a
t
a
s
e
t
s
.
H
o
w
e
v
e
r
,
t
h
e
e
m
o
d
e
l
i
s
o
n
l
y
a
p
p
l
i
c
a
b
l
e
t
o
t
h
e
n
e
t
w
o
r
k
w
i
t
h
a
u
t
o
c
o
r
r
e
l
a
t
i
o
n
t
r
a
f
f
i
c
c
h
a
r
a
c
t
e
r
i
s
t
i
c
s
[
2
9
]
.
W
ith
t
h
e
r
ap
id
d
ev
elo
p
m
en
t
o
f
tec
h
n
o
lo
g
y
a
n
d
g
i
v
e
n
t
h
e
s
a
m
e
d
r
a
w
b
ac
k
s
o
f
L
ST
M
w
h
ic
h
ca
n
r
es
u
l
t
in
s
o
m
e
d
ev
iatio
n
o
f
t
h
e
p
r
ed
ictio
n
r
esu
lts
,
D
u
an
i
n
2
0
1
8
,
f
o
cu
s
ed
o
n
f
ilter
i
n
g
n
o
is
e
f
lo
w
d
ata
to
m
i
tig
a
te
th
e
d
ef
icien
c
ies
o
f
L
ST
M.
B
as
ed
o
n
th
e
id
ea
o
f
d
ec
o
m
p
o
s
itio
n
,
h
e
p
r
o
p
o
s
ed
th
e
s
ea
s
o
n
al
lo
ess
tr
en
d
d
ec
o
m
p
o
s
itio
n
(
ST
)
an
d
L
S
T
M
p
r
ed
ictio
n
m
o
d
el.
T
h
e
m
o
d
el
p
r
o
ce
s
s
m
et
h
o
d
is
aim
ed
at
d
ea
lin
g
w
it
h
p
er
io
d
ic
tr
af
f
ic
d
ata,
d
ec
o
m
p
o
s
in
g
tr
en
d
a
n
d
eli
m
i
n
ati
n
g
r
an
d
o
m
n
o
i
s
e.
B
u
t
th
e
h
y
b
r
id
m
o
d
el
is
l
i
m
ited
i
n
u
s
e
in
t
h
at
i
t
ca
n
o
n
l
y
ad
d
r
ess
th
e
p
er
io
d
ic
ch
ar
ac
ter
is
tics
o
f
tr
a
f
f
ic,
a
n
d
n
o
t
t
h
e
n
o
n
li
n
ea
r
m
u
lt
i
-
s
ca
le
ch
ar
ac
ter
is
tic
s
o
f
tr
af
f
ic
d
ata
[
3
0
]
.
As
is
e
v
id
en
t,
d
if
f
er
en
t
r
e
s
e
ar
ch
er
s
f
o
c
u
s
o
n
d
if
f
er
en
t
o
p
ti
m
izatio
n
alg
o
r
it
h
m
s
.
T
o
ad
d
r
ess
th
e
p
r
o
b
lem
o
f
"
p
r
em
at
u
r
it
y
"
o
f
FO
A
al
g
o
r
ith
m
,
Ha
n
o
p
ti
m
ized
t
h
e
E
SN
n
e
u
r
al
n
et
w
o
r
k
b
ased
o
n
t
h
e
Op
p
o
s
itio
n
-
B
ased
L
ea
r
n
in
g
m
ec
h
an
i
s
m
.
He
al
s
o
s
o
l
v
ed
FO
A's
d
ef
ec
t
in
o
r
d
er
to
r
ea
lize
m
u
lti
-
s
tep
p
r
ed
ictio
n
o
f
n
e
t
w
o
r
k
tr
a
f
f
ic.
He
f
ir
s
t
u
s
ed
th
e
p
h
a
s
e
s
p
ac
e
r
ec
o
n
s
tr
u
ct
io
n
tec
h
n
iq
u
e
to
r
ec
o
n
s
tr
u
ct
t
h
e
o
r
ig
i
n
al
n
et
w
o
r
k
f
lo
w
ti
m
e
s
er
ies
an
d
estab
li
s
h
th
e
m
o
d
el
b
ased
o
n
th
e
E
SN
m
eth
o
d
.
T
h
en
,
u
s
i
n
g
o
p
p
o
s
itio
n
-
b
ased
lear
n
i
n
g
m
ec
h
a
n
i
s
m
o
f
f
r
u
it
f
l
ie
s
o
p
tim
izatio
n
alg
o
r
ith
m
,
h
e
o
p
ti
m
i
ze
d
th
e
m
o
d
el
p
ar
am
eter
s
.
Fi
n
all
y
,
t
h
e
o
p
ti
m
ized
m
o
d
el
i
s
u
s
ed
to
r
ea
lize
th
e
m
u
lti
-
s
tep
p
r
ed
ictio
n
o
f
n
e
t
wo
r
k
tr
af
f
ic.
Ho
w
e
v
er
,
th
er
e
ar
e
f
o
u
r
p
ar
a
m
eter
s
o
f
E
SN
w
h
i
c
h
m
u
s
t
b
e
o
p
tim
ize
d
,
m
a
k
in
g
it
to
o
lar
g
e
to
ap
p
r
o
x
i
m
ate
t
h
e
o
p
ti
m
al
s
o
lu
tio
n
,
th
u
s
a
f
f
ec
tin
g
th
e
ac
cu
r
ac
y
[
3
1
]
.
I
n
W
en
q
u
a
n
X
u
,
p
o
s
ited
t
h
at
th
e
e
x
is
t
in
g
tr
a
f
f
ic
m
o
d
el
s
h
o
u
ld
f
o
c
u
s
o
n
f
in
d
i
n
g
p
ar
a
m
ete
r
s
s
u
c
h
a
s
th
e
w
e
ig
h
t
o
f
n
o
d
e
co
n
n
ec
tio
n
i
n
t
h
e
n
e
u
r
al
n
et
w
o
r
k
.
I
f
t
h
e
ap
p
r
o
p
r
iate
v
alu
e
ca
n
n
o
t
b
e
o
b
tai
n
ed
,
th
e
m
o
d
el
p
ar
am
eter
s
ea
r
ch
r
e
m
ai
n
s
in
t
h
e
lo
ca
l
o
p
ti
m
al,
h
e
n
ce
r
es
u
lt
in
g
i
n
a
co
m
p
r
o
m
is
ed
m
o
d
el
p
r
ec
is
io
n
.
D
u
e
to
th
is
,
t
h
e
a
u
th
o
r
u
s
ed
au
to
-
r
eg
r
ess
iv
e
(
AR
)
m
o
d
el
to
f
it
th
e
o
r
ig
in
al
d
ata
a
n
d
o
b
tain
th
e
AR
m
o
d
el
r
esid
u
al
s
b
et
w
ee
n
th
e
o
r
ig
i
n
al
d
ata
an
d
th
e
p
r
ed
icted
d
ata
o
f
th
e
AR
m
o
d
el.
T
h
e
r
esid
u
al
s
ar
e
r
eg
ar
d
ed
as
th
e
n
o
n
li
n
ea
r
co
m
p
o
n
e
n
t
a
n
d
ar
e
tak
e
n
as
in
p
u
ts
i
n
to
t
h
e
d
ee
p
b
elief
n
et
wo
r
k
(
DB
N)
m
o
d
el.
T
h
e
A
R
m
o
d
el
p
r
ed
ictio
n
an
d
th
e
o
u
tp
u
t
o
f
th
e
DB
N
m
o
d
el
ar
e
th
e
f
i
n
al
f
o
r
ec
asti
n
g
v
al
u
e
f
o
r
t
h
e
ti
m
e
s
er
ies.
I
n
e
v
itab
l
y
,
in
th
e
p
r
o
ce
s
s
o
f
s
u
b
s
ta
n
tial
tr
ai
n
i
n
g
s
an
d
r
esid
u
als,
t
h
e
m
o
d
el
h
a
s
to
b
e
co
n
s
tan
tl
y
ad
j
u
s
ted
b
y
t
h
e
co
ef
f
ici
en
t,
th
u
s
lead
i
n
g
to
a
n
i
n
c
r
e
a
s
e
i
n
t
h
e
c
a
l
c
u
l
a
t
i
o
n
s
c
a
l
e
a
n
d
t
i
m
e
[
3
2
]
.
F
o
r
a
b
e
t
t
e
r
u
n
d
e
r
s
t
a
n
d
i
n
g
o
f
t
h
e
a
p
p
l
i
c
a
t
i
o
n
o
f
t
h
e
o
p
t
i
m
i
z
a
t
i
o
n
t
e
c
h
n
i
q
u
e
i
n
t
h
e
h
y
b
r
i
d
m
o
d
e
l
i
n
t
e
r
m
s
o
f
i
t
s
s
t
r
e
n
g
t
h
s
a
n
d
l
im
i
t
a
t
i
o
n
s
,
a
c
o
m
p
a
r
i
s
o
n
is
p
r
esen
t
ed
in
T
a
b
le
1
in
th
e
A
p
p
en
d
i
x
.
I
n
a
n
u
ts
h
ell,
r
esear
ch
er
s
h
a
v
e
u
s
ed
d
if
f
er
en
t
o
p
ti
m
izatio
n
alg
o
r
ith
m
s
to
co
n
s
t
r
u
ct
tr
af
f
ic
m
o
d
els
w
h
ic
h
h
a
v
e
h
ig
h
er
p
er
f
o
r
m
a
n
ce
.
W
h
ile
t
h
e
y
h
a
v
e
s
u
f
f
ici
en
tl
y
co
n
s
id
er
ed
th
e
d
r
a
w
b
ac
k
s
o
f
s
i
n
g
le
n
e
u
r
al
n
et
w
o
r
k
,
s
o
m
e
li
m
ita
tio
n
s
r
e
m
ai
n
[
3
3
,
3
4
]
.
T
o
b
u
ild
b
etter
ac
cu
r
ate
m
o
d
els,
r
esear
c
h
er
s
ar
e
co
n
s
tan
tl
y
tr
y
i
n
g
o
u
t
n
e
w
tec
h
n
iq
u
e
s
an
d
m
et
h
o
d
s
.
W
ith
t
h
e
d
ev
elo
p
m
e
n
t
o
f
r
esear
ch
,
t
h
e
id
ea
b
ased
o
n
d
ec
o
m
p
o
s
itio
n
is
g
r
ad
u
all
y
in
tr
o
d
u
ce
d
in
to
th
e
p
r
ed
ictio
n
f
ield
o
f
tr
af
f
ic
ti
m
i
n
g
.
3.
DE
CO
M
P
O
SI
T
I
O
N
T
E
CH
NIQU
E
-
B
AS
E
D
H
YB
RID
M
O
DE
L
I
n
t
h
e
n
e
w
er
a
o
f
h
y
b
r
id
m
o
d
el
co
n
s
tr
u
ctio
n
,
r
esear
ch
er
s
in
tr
o
d
u
ce
ti
m
e
-
f
r
eq
u
e
n
c
y
an
a
l
y
s
i
s
i
n
to
tr
af
f
ic
la
w
a
n
al
y
s
i
s
an
d
ap
p
ly
t
h
e
s
i
g
n
al
an
al
y
s
i
s
th
e
o
r
y
to
tr
af
f
ic
ti
m
e
s
er
ies
a
n
al
y
s
i
s
.
T
h
er
ef
o
r
e,
d
ec
o
m
p
o
s
itio
n
tech
n
iq
u
e
s
ar
e
n
o
w
w
id
el
y
u
s
ed
i
n
h
y
b
r
id
m
o
d
el
s
,
m
ai
n
l
y
w
av
elet
tr
a
n
s
f
o
r
m
(
W
T
)
[
3
5
]
an
d
m
o
d
e
d
ec
o
m
p
o
s
itio
n
(
MD
)
s
u
ch
as
e
m
p
ir
ical
m
o
d
e
d
ec
o
m
p
o
s
itio
n
(
E
MD
)
[
3
6
]
,
en
s
em
b
le
e
m
p
ir
ical
m
o
d
e
d
ec
o
m
p
o
s
itio
n
(
E
E
MD
)
[
3
7
]
an
d
v
ar
iatio
n
al
m
o
d
e
d
ec
o
m
p
o
s
itio
n
(
VM
D)
[
3
8
]
.
T
h
ese
ti
m
e
s
er
ies
tr
af
f
ic
h
y
b
r
id
m
o
d
els ar
e
f
a
s
t b
ec
o
m
i
n
g
a
h
o
t to
p
ic
f
o
r
r
esear
ch
er
s
.
3
.
1
.
WT
-
ba
s
ed
hy
brid
m
o
de
l
T
h
e
n
et
w
o
r
k
tr
af
f
ic
h
a
s
th
e
ch
ar
ac
ter
is
tic
s
o
f
r
e
m
o
te
d
ep
en
d
en
ce
a
n
d
m
u
lti
f
r
ac
t
al,
r
e
n
d
er
in
g
t
h
e
s
in
g
le
n
eu
r
al
n
et
w
o
r
k
m
o
d
el
an
in
ad
eq
u
a
te
p
r
ed
ictio
n
to
o
l.
I
n
L
a
is
e
n
Nie
,
in
tr
o
d
u
ce
d
d
ec
o
m
p
o
s
itio
n
id
ea
a
n
d
u
s
ed
d
is
cr
ete
w
a
v
elet
tr
an
s
f
o
r
m
(
DW
T
)
to
d
iv
id
e
th
e
s
ig
n
a
l
in
to
lo
w
-
p
ass
a
n
d
h
ig
h
-
p
a
s
s
co
m
p
o
n
e
n
t
s
.
g
au
s
s
ia
n
m
o
d
el
(
GM
)
p
r
ed
icted
h
ig
h
-
p
ass
co
m
p
o
n
e
n
ts
a
n
d
d
ee
p
b
elief
n
e
t
w
o
r
k
(
DB
N)
m
o
d
el
p
r
ed
icted
lo
w
-
p
ass
co
m
p
o
n
e
n
ts
,
est
i
m
a
tin
g
t
h
e
p
ar
a
m
eter
s
o
f
t
h
e
Ga
u
s
s
ian
m
o
d
el
b
y
t
h
e
m
ax
i
m
u
m
li
k
eli
h
o
o
d
m
e
th
o
d
.
T
h
en
h
e
p
r
ed
icted
th
e
h
ig
h
-
p
ass
co
m
p
o
n
en
t
b
y
DW
T
-
DB
N
-
GM
h
y
b
r
id
m
o
d
el
[
3
9
]
.
B
ased
o
n
t
h
e
s
a
m
e
n
o
tio
n
,
L
aise
n
Nie
al
s
o
ad
o
p
ted
th
e
DW
T
m
et
h
o
d
to
d
ec
o
m
p
o
s
e
th
e
s
i
g
n
a
l.
T
h
e
au
th
o
r
u
s
ed
s
p
atio
te
m
p
o
r
al
co
m
p
r
es
s
i
v
e
s
en
s
i
n
g
(
S
C
S)
m
et
h
o
d
to
p
r
e
d
ict
h
ig
h
-
p
a
s
s
co
m
p
o
n
e
n
t
s
an
d
DB
N
m
o
d
el
to
p
r
e
d
ict
lo
w
-
p
as
s
co
m
p
o
n
e
n
t
s
.
He
s
u
b
s
eq
u
en
tl
y
p
r
o
p
o
s
ed
th
e
DW
T
-
DB
N
-
S
C
S
h
y
b
r
id
m
o
d
el
[
4
0
]
.
Usi
n
g
t
h
is
m
o
d
el,
t
h
e
d
ef
ec
t
o
f
s
in
g
le
n
eu
r
al
n
et
w
o
r
k
i
s
s
o
lv
ed
,
b
u
t
th
e
d
i
f
f
ic
u
lt
y
o
f
d
ec
o
m
p
o
s
i
tio
n
s
ca
le
o
f
w
av
ele
t
tr
an
s
f
o
r
m
(
W
T
)
[
4
1
]
is
o
v
er
lo
o
k
ed
,
w
h
ic
h
ca
n
t
h
en
af
f
ec
t th
e
ac
c
u
r
ac
y
o
f
t
h
e
h
y
b
r
id
m
o
d
el.
C
o
n
s
id
er
in
g
th
e
d
i
f
f
icu
l
t
y
o
f
a
cc
u
r
atel
y
p
r
ed
icti
n
g
co
m
p
le
x
n
et
w
o
r
k
tr
a
f
f
ic
d
ata
in
t
h
e
L
S
T
M
m
o
d
el,
Haip
en
g
L
u
et
a
l
.
i
n
tr
o
d
u
ce
d
w
a
v
elet
tr
an
s
f
o
r
m
(
W
T
)
to
co
n
s
tr
u
ct
W
T
-
L
ST
M
h
y
b
r
id
m
o
d
el
in
2
0
1
8
.
Firstl
y
,
th
e
tr
a
f
f
ic
i
s
d
ec
o
m
p
o
s
ed
to
an
ap
p
r
o
x
i
m
a
tio
n
s
eq
u
e
n
ce
a
n
d
s
e
v
er
al
d
etail
s
eq
u
e
n
ce
s
.
T
h
e
ap
p
r
o
x
im
at
io
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
11
,
No
.
2
,
A
p
r
il 2
0
2
1
:
1
4
5
0
-
1459
1454
s
eq
u
en
ce
co
n
tai
n
s
t
h
e
tr
e
n
d
an
d
c
y
clica
l
f
ea
t
u
r
es
o
f
tr
af
f
ic,
w
h
er
ea
s
t
h
e
d
etail
s
eq
u
en
ce
s
co
n
tain
t
h
e
d
etailed
in
f
o
r
m
atio
n
a
t
m
u
lt
is
ca
le.
T
h
en
th
e
ap
p
r
o
x
i
m
atio
n
s
eq
u
e
n
ce
is
u
s
ed
to
tr
ain
t
h
e
L
ST
M
n
et
w
o
r
k
,
w
h
ile
t
h
e
d
etail
s
eq
u
e
n
ce
s
ar
e
u
s
ed
to
co
n
s
tr
u
ct
th
e
e
m
p
ir
ical
d
eta
il
s
eq
u
en
ce
s
.
B
y
r
ec
o
n
s
tr
u
cti
n
g
s
eq
u
en
ce
w
it
h
p
r
ed
icted
a
p
p
r
o
x
i
m
atio
n
s
eq
u
en
ce
an
d
e
m
p
ir
ical
d
etail
s
eq
u
en
ce
s
,
th
e
p
r
ed
ictio
n
o
f
f
u
t
u
r
e
tr
af
f
ic
is
ac
q
u
ir
ed
.
Ho
w
e
v
er
,
th
e
li
m
ita
tio
n
s
o
f
W
T
a
r
e
s
i
m
ilar
l
y
ig
n
o
r
ed
a
n
d
th
e
p
r
o
b
lem
o
f
lo
ca
l
o
p
ti
m
al
s
o
l
u
tio
n
o
f
L
ST
M
r
e
m
ain
s
u
n
s
o
l
v
ed
[
4
2
]
.
Giv
e
n
t
h
at
t
h
e
d
ec
o
m
p
o
s
itio
n
s
ca
le
o
f
W
T
tech
n
iq
u
e
i
s
d
if
f
i
cu
lt,
Ma
d
an
e
t
a
l
.
u
s
ed
in
v
er
s
e
d
is
cr
ete
w
a
v
elet
tr
an
s
f
o
r
m
(
iDW
T
)
tech
n
o
lo
g
y
to
d
ec
o
m
p
o
s
e
t
h
e
t
r
af
f
ic
in
to
d
etails
a
n
d
ap
p
r
o
x
i
m
ate
co
m
p
o
n
e
n
ts
.
T
h
r
o
u
g
h
t
h
e
iDW
T
r
ec
o
n
s
tr
u
ctio
n
,
t
h
e
s
eq
u
en
ce
is
r
ec
o
n
s
tr
u
cted
to
o
b
tain
a
n
e
w
ti
m
e
s
er
ie
s
.
T
h
en
t
h
e
s
elec
ted
A
R
I
M
A
m
o
d
el
i
s
u
s
e
d
to
p
r
ed
ict
th
e
lo
w
co
m
p
o
n
en
t,
an
d
t
h
e
R
NN
n
e
u
r
al
n
e
t
w
o
r
k
i
s
u
s
ed
to
p
r
ed
ict
th
e
h
ig
h
co
m
p
o
n
e
n
t,
r
esp
ec
ti
v
el
y
.
T
h
e
p
r
o
p
o
s
ed
h
y
b
r
id
m
o
d
el
is
a
ti
m
e
s
er
ies
w
h
ic
h
ca
n
b
e
u
s
ed
to
p
r
ed
ict
th
e
f
u
t
u
r
e
tr
af
f
ic
tr
en
d
s
i
n
a
co
m
p
u
ter
n
et
w
o
r
k
.
T
h
e
m
o
d
el
h
a
s
s
u
f
f
icie
n
tl
y
s
o
lv
ed
t
h
e
h
ar
d
er
p
r
o
b
le
m
o
f
W
T
’
s
d
ec
o
m
p
o
s
itio
n
s
ca
le.
H
o
w
e
v
er
,
th
e
co
m
p
le
x
it
y
o
f
ca
lcu
latio
n
i
n
R
NN
is
i
g
n
o
r
ed
,
h
en
ce
th
e
ac
cu
r
ac
y
o
f
p
r
ed
ictio
n
m
o
d
el
ca
n
b
e
q
u
esti
o
n
ab
le
[
4
3
]
.
3
.
2
.
MD
-
ba
s
ed
hy
brid
m
o
de
l
T
h
e
co
n
s
tr
ain
ts
o
f
W
T
tech
n
o
lo
g
y
an
d
s
i
n
g
le
n
e
u
r
al
n
et
w
o
r
k
in
cl
u
d
e
th
e
ten
d
e
n
cies
o
f
f
a
llin
g
in
t
o
lo
ca
l
m
i
n
i
m
u
m
,
a
n
d
o
v
er
f
itt
in
g
.
T
h
e
s
elec
tio
n
o
f
n
et
w
o
r
k
s
tr
u
ct
u
r
e
is
a
ls
o
to
o
d
ep
en
d
en
t
o
n
e
x
p
er
ien
ce
.
T
h
ese
li
m
i
tatio
n
s
d
ir
ec
tl
y
a
f
f
e
ct
th
e
r
eliab
ili
t
y
o
f
n
eu
r
al
n
et
w
o
r
k
s
f
o
r
ti
m
e
s
er
ies
p
r
ed
ictio
n
an
d
m
o
d
elli
n
g
.
I
n
T
ian
Z
h
o
n
g
d
a
f
ir
s
t
p
r
o
p
o
s
ed
to
d
ec
o
m
p
o
s
e
tr
a
f
f
ic
i
n
to
s
ta
b
le
d
ata
s
ig
n
als
o
f
d
i
f
f
er
e
n
t
c
h
ar
ac
ter
is
tic
s
ca
le
s
b
ased
o
n
E
m
p
ir
ical
m
o
d
e
d
ec
o
m
p
o
s
i
tio
n
(
E
MD
)
tech
n
o
lo
g
y
.
T
h
e
co
m
p
o
n
en
ts
af
ter
d
ec
o
m
p
o
s
in
g
r
e
m
o
v
e
th
e
lo
n
g
co
r
r
elatio
n
an
d
t
h
e
d
if
f
er
en
t
y
et
p
r
o
m
i
n
e
n
t
lo
ca
l
ch
ar
a
cter
is
tics
o
f
t
i
m
e
s
er
ies
w
h
ich
ca
n
in
t
u
r
n
r
ed
u
ce
th
e
n
o
n
-
s
tatio
n
ar
y
o
f
ti
m
e
s
er
ies.
He
th
e
n
p
r
o
p
o
s
ed
th
e
E
MD
-
E
L
M
h
y
b
r
id
m
o
d
el
w
it
h
th
e
i
n
co
r
p
o
r
atio
n
o
f
t
h
e
E
L
M
n
e
u
r
a
l
n
e
tw
o
r
k
[
4
4
]
.
U
n
f
o
r
t
u
n
a
t
e
l
y
,
E
M
D
i
s
p
r
o
n
e
t
o
m
o
d
e
a
l
i
a
s
i
n
g
a
n
d
e
n
d
p
o
i
n
t
e
f
f
e
c
t
p
r
o
b
l
e
m
s
[
4
5
]
d
u
r
in
g
th
e
d
ec
o
m
p
o
s
itio
n
p
r
o
ce
s
s
w
h
ich
ca
n
ev
e
n
t
u
all
y
co
m
p
r
o
m
i
s
e
th
e
p
r
ed
ictio
n
ac
cu
r
a
c
y
.
I
n
v
ie
w
o
f
t
h
e
tr
af
f
ic
lo
n
g
a
n
d
s
h
o
r
t
co
r
r
elatio
n
,
C
h
e
n
in
tr
o
d
u
ce
d
th
e
E
MD
-
P
SO
-
SVM
h
y
b
r
id
m
o
d
el
b
ased
o
n
e
m
p
ir
ical
m
o
d
e
d
ec
o
m
p
o
s
i
tio
n
,
p
ar
ticle
s
w
ar
m
o
p
ti
m
izatio
n
an
d
s
u
p
p
o
r
t
v
ec
to
r
m
ac
h
in
e.
First,
t
h
e
E
MD
s
u
s
ed
to
eli
m
i
n
ate
th
e
i
n
f
lu
e
n
ce
o
f
tr
a
f
f
ic
n
o
is
e
s
i
g
n
a
ls
.
T
h
en
p
ar
ticle
s
w
ar
m
o
p
ti
m
izatio
n
al
g
o
r
ith
m
is
u
s
ed
to
o
p
ti
m
ize
th
e
p
ar
a
m
eter
s
o
f
SV
M.
T
h
e
ef
f
ec
ti
v
e
n
es
s
o
f
t
h
e
p
r
ese
n
ted
m
et
h
o
d
is
ex
a
m
i
n
ed
b
y
ev
alu
a
tin
g
it
w
it
h
d
if
f
er
en
t
m
e
th
o
d
s
in
cl
u
d
in
g
b
asic
SVM
a
n
d
E
MD
.
Fin
all
y
,
SVM
i
s
u
s
ed
f
o
r
m
o
d
el
tr
ain
i
n
g
an
d
f
itti
n
g
tr
af
f
ic
m
o
d
el
[
4
6
]
.
W
h
ile
th
i
s
m
o
d
el
ca
n
i
m
p
r
o
v
e
th
e
ac
c
u
r
ac
y
o
f
n
et
w
o
r
k
t
r
af
f
ic
p
r
ed
ictio
n
,
it
ig
n
o
r
es
t
h
e
m
o
d
el
alia
s
in
g
p
r
o
b
lem
an
d
en
d
p
o
in
t
ef
f
ec
t
s
o
f
E
MD
;
t
h
e
d
e
f
in
iten
e
s
s
o
f
m
o
d
el
p
r
ed
ictio
n
i
s
s
u
b
s
eq
u
en
t
l
y
a
f
f
ec
te
d.
T
o
ad
d
r
ess
th
e
li
m
itat
io
n
o
f
E
MD
,
W
an
w
ei
Hu
a
n
g
in
tr
o
d
u
ce
d
en
s
e
m
b
le
e
m
p
ir
i
ca
l
m
o
d
e
d
ec
o
m
p
o
s
itio
n
(
E
E
MD
)
tech
n
o
lo
g
y
an
d
q
u
a
n
t
u
m
n
eu
r
al
n
et
w
o
r
k
al
g
o
r
ith
m
to
co
n
s
tr
u
ct
th
e
Q
NN
-
E
E
MD
h
y
b
r
id
m
o
d
el.
T
h
e
E
E
MD
tec
h
n
iq
u
e
i
s
u
s
ed
to
d
ec
o
m
p
o
s
e
t
h
e
ti
m
e
s
er
ies
i
n
to
I
MF
to
r
e
m
o
v
e
m
o
d
al
alia
s
i
n
g
an
d
r
ed
u
n
d
a
n
c
y
.
T
h
en
h
e
u
s
e
d
QNN
to
p
r
o
ce
s
s
t
h
e
d
ec
o
m
p
o
s
ed
I
MF
an
d
o
p
ti
m
ize
t
h
e
p
ar
am
eter
s
o
f
th
e
m
o
d
el
s
o
th
a
t
th
e
co
n
v
er
g
e
n
c
e
s
p
ee
d
o
f
th
e
h
y
b
r
id
m
o
d
el
i
s
i
m
p
r
o
v
ed
[
4
7
]
.
Ho
w
ev
er
,
t
h
e
m
o
d
el
ig
n
o
r
es
t
h
e
i
m
p
ac
t
o
f
to
o
lar
g
e
co
m
p
u
tati
o
n
s
ca
le
o
f
Q
u
an
tu
m
al
g
o
r
ith
m
m
ec
h
a
n
ic
s
.
I
n
ad
d
itio
n
,
th
e
E
E
MD
d
ep
en
d
en
c
e
o
n
a
m
p
lit
u
d
e
an
d
n
u
m
b
er
o
f
e
x
p
er
ien
ce
s
[
4
8
]
w
ill a
f
f
ec
t t
h
e
ac
cu
r
ac
y
o
f
t
h
e
p
r
ed
ictio
n
.
Du
e
to
th
e
d
ef
icie
n
cie
s
o
f
E
MD
an
d
E
E
MD
,
L
in
a
P
an
ar
g
u
ed
t
h
at
E
SN
ca
n
ea
s
il
y
s
u
f
f
er
f
r
o
m
t
h
e
in
f
lu
e
n
ce
s
o
f
i
n
itial
r
a
n
d
o
m
w
ei
g
h
ts
.
S
h
e
f
ir
s
t
in
tr
o
d
u
ce
d
th
e
co
n
ce
p
t
o
f
v
ar
iatio
n
al
m
o
d
e
d
ec
o
m
p
o
s
itio
n
(
VM
D)
to
o
v
er
co
m
e
th
e
p
r
o
b
le
m
s
o
f
E
MD
an
d
E
E
MD
an
d
ef
f
ec
tiv
el
y
d
ec
o
m
p
o
s
e
th
e
tr
af
f
ic.
Usi
n
g
B
at
A
l
g
o
r
ith
m
(
B
A
)
alg
o
r
ith
m
to
i
m
p
r
o
v
e
an
d
o
p
ti
m
ize
E
S
N
p
ar
a
m
eter
s
,
s
h
e
th
e
n
p
r
o
p
o
s
ed
th
e
VM
D
-
BA
-
E
SN
n
et
w
o
r
k
tr
a
f
f
ic
h
y
b
r
id
p
r
ed
ictio
n
m
o
d
el.
I
n
th
e
p
r
o
ce
s
s
o
f
t
h
e
d
ec
o
m
p
o
s
itio
n
,
VM
D
is
u
til
ized
to
d
ec
o
m
p
o
s
e
th
e
o
r
ig
i
n
al
i
n
ter
n
et
tr
a
f
f
ic
s
e
r
ies
in
to
s
e
v
er
al
b
an
d
-
li
m
ited
in
tr
in
s
ic
m
o
d
e
f
u
n
ctio
n
s
(
B
L
I
MF
s
)
.
I
n
ev
i
tab
l
y
,
d
ec
o
m
p
o
s
itio
n
la
y
er
s
w
ill b
e
an
i
m
p
o
r
tan
t f
ac
to
r
to
d
eter
m
in
e
th
e
ac
cu
r
ac
y
o
f
p
r
ed
ictio
n
[
4
9
]
.
Giv
e
n
th
e
s
tr
o
n
g
n
o
n
-
s
tatio
n
ar
y
a
n
d
h
ig
h
co
m
p
lex
it
y
o
f
t
h
e
c
h
ao
tic
t
i
m
e
s
er
ie
s
,
it
is
d
if
f
ic
u
lt
to
d
i
r
e
c
t
l
y
a
n
a
l
y
s
e
a
n
d
p
r
e
d
i
c
t
b
y
ju
s
t
d
e
p
e
n
d
i
n
g
o
n
a
s
i
n
g
l
e
m
o
d
e
l
.
H
e
n
c
e
,
i
n
X
i
n
g
h
a
n
X
u
a
p
p
l
i
e
d
a
tw
o
-
l
a
y
e
r
d
e
c
o
m
p
o
s
i
t
i
o
n
a
p
p
r
o
a
c
h
a
n
d
o
p
t
i
m
i
z
e
d
B
P
n
e
u
r
a
l
n
e
tw
o
r
k
.
T
h
e
h
y
b
r
i
d
m
o
d
e
l
a
i
m
s
t
o
o
b
t
a
i
n
c
o
m
p
r
e
h
e
n
s
i
v
e
i
n
f
o
r
m
a
t
i
o
n
o
f
t
h
e
c
h
a
o
t
i
c
t
i
m
e
s
e
r
i
e
s
w
h
i
c
h
i
s
c
o
m
p
o
s
e
d
o
f
c
o
m
p
l
e
t
e
e
n
s
e
m
b
l
e
e
m
p
i
r
i
c
a
l
m
o
d
e
d
ec
o
m
p
o
s
itio
n
w
it
h
ad
ap
tiv
e
n
o
is
e
(
C
E
E
M
DAN)
an
d
v
ar
iatio
n
al
m
o
d
e
d
ec
o
m
p
o
s
itio
n
.
T
h
e
VM
D
al
g
o
r
ith
m
is
u
s
ed
f
o
r
f
u
r
t
h
er
d
ec
o
m
p
o
s
itio
n
o
f
th
e
h
ig
h
f
r
eq
u
e
n
c
y
s
u
b
s
eq
u
e
n
ce
o
b
tain
ed
b
y
C
E
E
MD
AN,
af
ter
wh
ich
th
e
p
r
ed
ictio
n
p
er
f
o
r
m
a
n
ce
is
s
i
g
n
if
ican
t
l
y
i
m
p
r
o
v
ed
.
T
h
en
t
h
e
B
P
NN
o
p
ti
m
ized
b
y
a
f
ir
ef
l
y
al
g
o
r
it
h
m
(
FA
)
is
u
t
ilized
f
o
r
p
r
e
d
ictio
n
.
T
h
e
h
y
b
r
id
m
o
d
el
f
u
ll
y
co
n
s
id
er
s
t
h
e
i
m
p
o
r
tan
ce
o
f
d
ec
o
m
p
o
s
itio
n
s
ig
n
al
s
.
Ho
w
e
v
er
,
it
i
g
n
o
r
es
th
e
p
er
f
o
r
m
an
ce
o
f
VM
D
d
et
er
m
in
ed
b
y
t
h
e
n
u
m
b
er
o
f
d
e
co
m
p
o
s
i
tio
n
la
y
er
s
,
w
h
ich
is
l
ik
el
y
to
ca
u
s
e
o
v
er
-
d
ec
o
m
p
o
s
itio
n
o
r
u
n
d
er
-
d
ec
o
m
p
o
s
itio
n
a
n
d
ca
n
a
f
f
ec
t th
e
a
cc
u
r
ac
y
o
f
t
h
e
m
o
d
el
[
5
0
]
.
I
n
v
ie
w
o
f
t
h
e
ex
ten
s
i
v
e
ap
p
licatio
n
o
f
d
ec
o
m
p
o
s
i
tio
n
tec
h
n
iq
u
e
i
n
n
et
w
o
r
k
t
r
a
f
f
ic
p
r
ed
ictio
n
an
d
atte
m
p
ts
to
i
m
p
r
o
v
e
t
h
e
p
r
e
d
ictio
n
ac
c
u
r
ac
y
o
f
n
o
n
li
n
ea
r
n
o
n
-
s
ta
tio
n
ar
y
tr
a
f
f
ic
d
ata,
Yin
g
Han
,
et
a
l
.
,
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
co
mp
r
eh
en
s
ive
r
ev
iew
o
n
h
y
b
r
id
n
etw
o
r
k
tr
a
ffic p
r
ed
ictio
n
mo
d
el
(
Jin
mei
S
h
i
)
1455
p
r
o
p
o
s
ed
a
I
FOA
-
E
SN
co
m
b
i
n
ed
p
r
ed
ictio
n
m
o
d
el.
T
h
is
m
o
d
el
is
b
ased
o
n
VM
D
b
y
L
e
v
y
f
l
ig
h
t
f
u
n
ct
io
n
an
d
clo
u
d
g
e
n
er
ato
r
.
First,
V
MD
is
u
s
ed
to
d
ec
o
m
p
o
s
e
t
h
e
o
r
ig
in
al
n
et
w
o
r
k
tr
af
f
ic
d
ata
i
n
to
s
e
v
er
al
s
u
b
s
ets.
T
h
en
,
m
u
ltip
le
s
u
b
r
eser
v
o
ir
s
ar
e
b
u
ilt
af
ter
p
er
f
o
r
m
i
n
g
t
h
e
p
h
ase
s
p
ac
e
r
ec
o
n
s
tr
u
c
tio
n
(
P
SR
)
o
f
ea
ch
d
ata
s
u
b
s
et.
Fin
a
ll
y
,
t
h
e
tr
ai
n
i
n
g
s
e
t
is
u
s
ed
to
tr
ain
t
h
e
p
r
ed
ictio
n
m
o
d
el.
T
h
is
m
ec
h
an
i
s
m
s
o
l
v
es
t
h
e
p
r
o
b
le
m
o
f
VM
D
r
eq
u
ir
in
g
a
ce
r
tain
n
u
m
b
er
o
f
p
r
e
-
s
et
p
a
tter
n
s
a
n
d
iter
atio
n
f
ac
to
r
s
,
w
h
ich
ca
n
n
o
t
b
e
d
eter
m
in
ed
b
y
s
u
b
j
ec
tiv
e
e
x
p
er
ien
ce
.
U
n
f
o
r
t
u
n
ate
l
y
,
t
h
e
s
y
n
c
h
r
o
n
o
u
s
o
p
ti
m
izatio
n
in
s
id
e
an
d
o
u
ts
id
e
o
f
t
h
e
m
u
lt
ip
le
s
u
b
r
eser
v
o
ir
s
n
ec
es
s
itate
s
lo
n
g
er
ca
lcu
latio
n
ti
m
e.
T
h
is
ca
n
n
e
g
ativ
el
y
a
f
f
ec
t
th
e
tr
ai
n
in
g
t
i
m
e
o
f
th
e
m
o
d
el
an
d
th
e
co
n
v
er
g
e
n
ce
s
p
ee
d
as
w
e
ll
as
t
h
e
p
er
f
o
r
m
a
n
ce
o
f
th
e
m
o
d
el
m
ec
h
a
n
is
m
.
T
h
u
s
,
th
e
b
ig
g
er
ca
lc
u
latio
n
s
ca
le
r
e
m
ai
n
s
an
u
n
s
o
lv
ed
s
c
ien
ti
f
ic
co
n
u
n
d
r
u
m
[
5
1
]
.
T
h
e
s
tr
en
g
th
s
an
d
li
m
itatio
n
s
o
f
th
e
d
ec
o
m
p
o
s
itio
n
tech
n
iq
u
e
i
n
th
e
h
y
b
r
id
n
et
w
o
r
k
tr
af
f
ic
p
r
ed
ictio
n
m
o
d
el
ar
e
s
u
m
m
ar
ized
in
T
ab
le
2
in
th
e
A
p
p
en
d
i
x
.
4.
CO
NCLU
SI
O
N
I
n
co
n
clu
s
io
n
,
o
p
ti
m
izat
io
n
an
d
d
ec
o
m
p
o
s
itio
n
ar
e
t
w
o
i
m
p
o
r
ta
n
t
p
r
o
ce
s
s
es
in
a
h
y
b
r
id
n
et
w
o
r
k
tr
af
f
ic
p
r
ed
ictio
n
m
o
d
el
i
n
en
s
u
r
i
n
g
a
h
i
g
h
er
p
r
ed
ictio
n
ac
cu
r
ac
y
a
n
d
f
a
s
ter
co
n
v
er
g
e
n
ce
s
p
ee
d
.
T
h
is
p
ap
er
f
o
u
n
d
th
at
P
SO
as
w
e
ll
as
o
th
er
o
p
tim
izat
io
n
al
g
o
r
ith
m
ca
n
g
en
er
all
y
id
en
tify
n
et
w
o
r
k
tr
a
f
f
ic
ti
m
e
s
eq
u
e
n
ce
s
b
etter
g
i
v
en
it
s
s
tr
en
g
th
s
o
f
s
i
m
p
le
p
r
in
c
ip
le,
s
m
all
ca
lc
u
la
tio
n
s
ca
le,
f
ast
co
n
v
er
g
e
n
ce
s
p
ee
d
an
d
s
o
f
o
r
th
.
T
h
e
p
a
p
e
r
also
co
n
f
ir
m
ed
th
at
th
e
d
ec
o
m
p
o
s
i
tio
n
tec
h
n
iq
u
e
is
an
ef
f
ec
t
iv
e
m
et
h
o
d
to
d
ea
l
w
it
h
n
o
n
-
lin
ea
r
it
y
an
d
n
o
n
-
s
tatio
n
ar
it
y
o
f
d
ata
a
s
it
p
r
o
v
id
es
a
m
o
d
elli
n
g
id
ea
b
ased
o
n
t
h
e
ti
m
e
f
r
eq
u
e
n
c
y
an
al
y
s
is
f
o
r
tr
af
f
ic
an
al
y
s
is
.
E
s
p
ec
iall
y
,
VM
D
ca
n
o
v
er
co
m
e
t
h
e
m
u
ltire
s
o
l
u
tio
n
an
d
d
ec
o
m
p
o
s
itio
n
s
ca
le
p
r
o
b
lem
i
n
W
T
,
s
o
lv
e
th
e
p
r
o
b
le
m
o
f
m
o
d
e
a
lias
i
n
g
an
d
w
h
ite
n
o
is
e
a
m
p
lit
u
d
e
i
n
E
MD
an
d
E
E
MD
d
ec
o
m
p
o
s
iti
o
n
tech
n
iq
u
es.
T
h
e
r
e
v
i
ew
h
a
s
,
t
o
s
o
m
e
e
x
t
e
n
t
,
h
e
l
p
e
d
e
n
h
a
n
c
e
o
u
r
u
n
d
e
r
s
t
a
n
d
i
n
g
o
f
t
h
e
i
m
p
o
r
t
a
n
c
e
o
f
o
p
t
i
m
i
z
a
t
i
o
n
a
n
d
d
e
c
o
m
p
o
s
i
t
i
o
n
t
e
c
h
n
i
q
u
e
s
i
n
a
h
y
b
r
i
d
n
e
t
w
o
r
k
p
r
e
d
i
c
t
i
o
n
m
o
d
e
l
.
T
h
e
p
a
r
a
m
e
t
e
r
o
p
t
i
m
i
z
a
t
i
o
n
o
f
d
e
c
o
m
p
o
s
i
t
i
o
n
t
e
c
h
n
i
q
u
e
a
n
d
o
p
t
i
m
i
z
a
t
i
o
n
a
l
g
o
r
i
t
h
m
i
s
t
h
e
k
e
y
p
r
o
c
e
s
s
t
o
d
e
t
e
r
m
i
n
e
t
h
e
p
r
e
d
i
c
t
i
o
n
a
c
c
u
r
a
c
y
a
n
d
c
o
n
v
e
r
g
e
n
c
e
r
a
t
e
.
F
u
t
u
r
e
r
e
s
e
a
r
c
h
s
h
o
u
l
d
t
h
e
r
e
f
o
r
e
c
o
n
c
e
n
t
r
a
t
e
o
n
t
h
e
i
n
v
e
s
t
i
g
a
t
i
o
n
o
f
h
o
w
t
o
s
i
m
p
l
i
f
y
t
h
e
o
p
t
i
m
i
z
a
t
i
o
n
a
l
g
o
r
i
t
h
m
w
i
t
h
f
e
w
e
r
p
a
r
a
m
e
t
e
r
s
,
s
h
o
r
t
e
n
t
h
e
c
o
n
v
e
r
g
e
n
c
e
s
p
e
e
d
a
n
d
i
m
p
r
o
v
e
t
h
e
d
ec
o
m
p
o
s
itio
n
e
f
f
ec
ts
to
s
u
b
s
eq
u
e
n
tl
y
en
h
a
n
ce
t
h
e
n
et
w
o
r
k
tr
af
f
ic
p
r
ed
ictio
n
ac
cu
r
ac
y
.
AP
P
E
NDI
X
T
ab
le
1
.
A
p
p
licatio
n
o
f
o
p
ti
m
i
za
tio
n
tech
n
iq
u
e
i
n
h
y
b
r
id
m
o
d
el
No
.
Y
e
a
r
A
u
t
h
o
r
O
r
i
g
i
n
a
l
M
o
d
e
l
H
y
b
r
i
d
M
o
d
e
l
S
t
r
e
n
g
t
h
L
i
mi
t
a
t
i
o
n
1
2
0
1
3
F
e
i
H
a
n
,
e
t
al
.
,
[
2
1
]
P
a
r
t
i
c
l
e
sw
a
r
m
o
p
t
i
m
i
z
a
t
i
o
n
A
P
S
O
-
E
L
M
O
p
t
i
mi
z
e
s i
n
p
u
t
w
e
i
g
h
t
a
n
d
d
e
v
i
a
t
i
o
n
o
f
EL
M
b
a
se
d
o
n
t
h
e
a
d
a
p
t
i
v
e
P
S
O
a
l
g
o
r
i
t
h
m
.
T
h
e
p
r
o
b
l
e
m o
f
l
o
c
a
l
o
p
t
i
mal
so
l
u
t
i
o
n
o
f
P
S
O
i
s s
t
i
l
l
n
o
t
so
l
v
e
d
a
n
d
m
a
y
a
f
f
e
c
t
t
h
e
a
c
c
u
r
a
c
y
o
f
t
h
e
mo
d
e
l
.
Ex
t
r
e
me
l
e
a
r
n
i
n
g
mac
h
i
n
e
2
2
0
1
3
K
u
n
Z
h
a
n
g
,
e
t
a
l
.
,
[
2
4
]
P
a
r
t
i
c
l
e
sw
a
r
m
o
p
t
i
m
i
z
a
t
i
o
n
Q
P
S
O
-
BP
S
o
l
v
e
s B
P
g
r
a
d
i
e
n
t
e
x
p
l
o
si
o
n
b
a
se
d
o
n
t
h
e
Q
u
a
n
t
u
m
a
l
g
o
r
i
t
h
m
a
n
d
P
S
O
.
T
h
e
q
u
a
n
t
u
m
a
l
g
o
r
i
t
h
m
i
s
d
i
f
f
i
c
u
l
t
t
o
c
a
l
c
u
l
a
t
e
;
l
e
a
sy
t
o
f
a
l
l
i
n
t
o
t
h
e
l
o
c
a
l
o
p
t
i
m
a
l
so
l
u
t
i
o
n
.
Q
u
a
n
t
u
m
B
P
n
e
u
r
a
l
n
e
t
w
o
r
k
3
2
0
1
4
Y
i
Y
a
n
g
,
e
t
al
.
,
[
1
7
]
S
e
a
so
n
a
l
t
r
a
n
sf
o
r
m
SA
-
PSO
-
L
S
S
V
M
S
e
q
u
e
n
c
e
e
l
i
mi
n
a
t
i
o
n
b
y
S
A
r
e
d
u
c
e
s t
h
e
i
n
t
e
r
f
e
r
e
n
c
e
o
f
se
a
so
n
s o
n
c
o
mp
o
n
e
n
t
s
a
n
d
o
p
t
i
m
i
z
e
s t
w
o
p
a
r
a
me
t
e
r
s o
f
L
S
S
V
M
b
a
se
d
o
n
t
h
e
P
S
O
.
O
n
l
y
c
o
n
si
d
e
r
s t
h
e
se
a
so
n
a
l
c
h
a
r
a
c
t
e
r
i
st
i
c
s o
f
t
r
a
f
f
i
c
b
u
t
i
g
n
o
r
e
s t
h
e
P
S
O
l
o
c
a
l
o
p
t
i
m
a
l
so
l
u
t
i
o
n
.
P
a
r
t
i
c
l
e
sw
a
r
m
o
p
t
i
m
i
z
a
t
i
o
n
L
e
a
st
sq
u
a
r
e
s su
p
p
o
r
t
v
e
c
t
o
r
mac
h
i
n
e
4
2
0
1
6
D
e
n
g
F
e
n
g
W
e
i
[
1
1
]
G
r
a
v
i
t
y
se
a
r
c
h
a
l
g
o
r
i
t
h
m
I
G
S
A
-
RBF
I
mp
r
o
v
e
s t
h
e
sp
e
e
d
se
l
e
c
t
i
o
n
f
o
r
mu
l
a
b
a
se
d
o
n
t
h
e
G
S
A
a
n
d
o
p
t
i
m
i
z
e
s t
h
r
e
e
p
a
r
a
me
t
e
r
s o
f
R
B
F
.
G
S
A
l
a
c
k
s
t
h
e
o
r
e
t
i
c
a
l
g
u
i
d
a
n
c
e
;
R
B
F
o
p
t
i
mi
z
a
t
i
o
n
p
a
r
a
me
t
e
r
s a
r
e
t
o
o
m
a
n
y
;
e
a
sy
t
o
f
a
l
l
i
n
t
o
t
h
e
l
o
c
a
l
o
p
t
i
mal
so
l
u
t
i
o
n
p
r
o
b
l
e
m.
R
a
d
i
a
l
b
a
s
i
s f
u
n
c
t
i
o
n
n
e
u
r
a
l
n
e
t
w
o
r
k
P
a
r
t
i
c
l
e
sw
a
r
m
o
p
t
i
m
i
z
a
t
i
o
n
5
2
0
1
6
T
a
o
H
e
,
e
t
al
.
,
[
2
3
]
R
a
d
i
a
l
b
a
s
i
s f
u
n
c
t
i
o
n
n
e
u
r
a
l
n
e
t
w
o
r
k
PSO
-
RBF
A
v
o
i
d
s l
o
c
a
l
o
p
t
i
mi
z
a
t
i
o
n
d
r
a
w
b
a
c
k
o
f
P
S
O
w
i
t
h
Q
u
a
n
t
u
m
b
i
t
a
n
d
o
p
t
i
m
i
z
e
s t
h
e
w
e
i
g
h
t
,
w
i
d
t
h
a
n
d
c
e
n
t
e
r
p
o
si
t
i
o
n
o
f
R
B
F
n
e
t
w
o
r
k
b
a
se
d
o
n
t
h
e
I
P
S
O
.
P
S
O
i
t
e
r
a
t
i
v
e
l
y
o
p
t
i
mi
z
e
s t
h
e
t
h
r
e
e
p
a
r
a
me
t
e
r
p
r
o
c
e
sse
s o
f
R
B
F
,
w
h
i
c
h
m
a
k
e
s
i
t
d
i
f
f
i
c
u
l
t
t
o
a
p
p
r
o
a
c
h
t
h
e
g
l
o
b
a
l
o
p
t
i
m
a
l
.
P
a
r
t
i
c
l
e
sw
a
r
m
o
p
t
i
m
i
z
a
t
i
o
n
B
l
a
c
k
h
o
l
e
6
2
0
1
7
A
b
d
e
l
h
a
d
i
A
z
z
o
u
n
i
,
e
t
al
.
,
[
2
8
]
R
e
c
u
r
r
e
n
t
n
e
u
r
a
l
n
e
t
w
o
r
k
R
N
N
-
L
S
T
M
S
o
l
v
e
s
L
S
T
M
p
r
o
b
l
e
m b
y
sl
i
d
i
n
g
l
e
a
r
n
i
n
g
w
i
n
d
o
w
a
n
d
e
x
t
r
a
c
t
s
t
h
e
d
y
n
a
mi
c
c
h
a
r
a
c
t
e
r
i
st
i
c
s o
f
t
r
a
f
f
i
c
b
a
se
d
o
n
t
h
e
R
N
N
.
W
h
i
l
e
so
l
v
i
n
g
t
h
e
L
S
T
M
p
r
o
b
l
e
m,
t
h
e
a
c
c
u
r
a
c
y
o
f
p
r
e
d
i
c
t
i
o
n
i
s i
mp
r
o
v
e
d
,
b
u
t
t
h
e
c
o
m
p
l
e
x
i
t
y
o
f
c
a
l
c
u
l
a
t
i
o
n
i
s
n
e
g
l
e
c
t
e
d
.
L
o
n
g
sh
o
r
t
-
t
e
r
m me
mo
r
y
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
11
,
No
.
2
,
A
p
r
il 2
0
2
1
:
1
4
5
0
-
1459
1456
T
ab
le
1
.
A
p
p
licatio
n
o
f
o
p
ti
m
i
za
tio
n
tech
n
iq
u
e
i
n
h
y
b
r
id
m
o
d
el
(
co
n
tin
u
e)
N
o
.
Y
e
a
r
A
u
t
h
o
r
O
r
i
g
i
n
a
l
M
o
d
e
l
H
y
b
r
i
d
M
o
d
e
l
S
t
r
e
n
g
t
h
L
i
mi
t
a
t
i
o
n
7
2
0
1
7
Q
i
n
z
h
e
n
g
Z
h
u
o
,
e
t
a
l
.
.
[
2
9
]
D
e
e
p
l
e
a
r
n
i
n
g
n
e
u
r
a
l
n
e
t
w
o
r
k
L
S
T
M
-
DNN
S
o
l
v
e
s
L
S
T
M
p
r
o
b
l
e
m a
n
d
f
u
l
l
y
c
o
n
si
d
e
r
s t
h
e
a
u
t
o
c
o
r
r
e
l
a
t
i
o
n
a
n
d
t
i
m
i
n
g
c
h
a
r
a
c
t
e
r
i
st
i
c
s o
f
t
r
a
f
f
i
c
.
I
t
i
s a
p
p
l
i
c
a
b
l
e
t
o
n
e
t
w
o
r
k
s
w
i
t
h
a
u
t
o
c
o
r
r
e
l
a
t
i
o
n
c
h
a
r
a
c
t
e
r
i
st
i
c
s.
L
o
n
g
sh
o
r
t
-
t
e
r
m
me
mo
r
y
8
2
0
1
7
Y
i
n
g
H
a
n
,
e
t
al
.
,
[
2
7
]
Ec
h
o
S
t
a
t
e
N
e
t
w
o
r
k
Q
F
O
A
-
ESN
S
o
l
v
e
s t
h
e
l
o
c
a
l
o
p
t
i
mi
z
a
t
i
o
n
so
l
u
t
i
o
n
o
f
F
O
A
b
a
se
d
o
n
t
h
e
q
u
a
n
t
u
m me
c
h
a
n
i
c
s a
n
d
o
p
t
i
m
i
z
e
s fi
v
e
i
mp
o
r
t
a
n
t
p
a
r
a
me
t
e
r
s o
f
ESN
w
i
t
h
Q
F
O
A
.
H
a
s t
o
o
ma
n
y
o
p
t
i
m
i
z
a
t
i
o
n
p
a
r
a
me
t
e
r
s a
n
d
t
o
o
m
u
c
h
c
o
mp
u
t
a
t
i
o
n
,
w
h
i
c
h
a
f
f
e
c
t
s
t
h
e
c
o
n
v
e
r
g
e
n
c
e
r
a
t
e
.
F
r
u
i
t
o
p
t
i
m
i
z
a
t
i
o
n
a
l
g
o
r
i
t
h
m
Q
u
a
n
t
u
n
9
2
0
1
7
Y
i
n
g
H
a
n
[
3
1
]
Ec
h
o
s
t
a
t
e
n
e
t
w
o
r
k
O
B
L
-
F
O
A
-
ESN
O
p
t
i
mi
z
e
s F
O
A
b
a
se
d
o
n
t
h
e
O
B
L
a
n
d
t
h
e
n
o
p
t
i
mi
z
e
s fo
u
r
p
a
r
a
me
t
e
r
s o
f
ESN
mo
d
e
l
t
o
p
r
o
v
i
d
e
mo
d
e
l
a
c
c
u
r
a
c
y
.
T
h
e
o
p
t
i
m
i
z
a
t
i
o
n
p
a
r
a
me
t
e
r
s
a
r
e
t
o
o
man
y
;
t
h
e
c
a
l
c
u
l
a
t
i
o
n
i
s l
a
r
g
e
,
t
h
e
s
t
r
u
c
t
u
r
e
i
s
c
o
mp
l
e
x
w
i
l
l
i
n
t
u
r
n
a
f
f
e
c
t
t
h
e
c
o
n
v
e
r
g
e
n
c
e
sp
e
e
d
.
F
r
u
i
t
f
l
y
o
p
t
i
mi
z
a
t
i
o
n
A
l
g
o
r
i
t
h
m
10
2
0
1
8
W
e
i
j
i
e
Z
h
a
n
g
,
e
t
a
l
.
,
[
2
2
]
R
a
d
i
a
l
b
a
s
i
s f
u
n
c
t
i
o
n
I
P
S
O
-
RBF
O
p
t
i
mi
z
e
s PS
O
b
y
u
si
n
g
mu
t
a
t
i
o
n
me
t
h
o
d
t
o
a
v
o
i
d
l
o
c
a
l
mi
n
i
m
u
m
p
r
o
b
l
e
m a
n
d
t
h
e
n
o
p
t
i
m
i
z
e
s t
h
r
e
e
p
a
r
a
me
t
e
r
s o
f
R
B
F
.
I
t
i
s d
i
f
f
i
c
u
l
t
t
o
a
p
p
r
o
a
c
h
t
h
e
g
l
o
b
a
l
o
p
t
i
m
a
l
so
l
u
t
i
o
n
d
u
e
t
o
t
o
o
man
y
p
a
r
a
me
t
e
r
s w
h
e
n
u
si
n
g
I
P
S
O
t
o
o
p
t
i
mi
z
e
R
B
F
.
P
a
r
t
i
c
l
e
sw
a
r
m
o
p
t
i
m
i
z
a
t
i
o
n
A
d
j
u
st
i
n
g
i
n
e
r
t
i
a
w
e
i
g
h
t
a
n
d
l
e
a
r
n
i
n
g
f
a
c
t
o
r
11
2
0
1
8
H
u
i
T
i
a
n
,
e
t
al
.
,
[
2
6
]
Q
u
a
n
t
u
m g
e
n
e
t
i
c
a
l
g
o
r
i
t
h
m
WT
-
Q
G
A
-
BP
D
e
t
e
r
mi
n
e
s e
mb
e
d
d
e
d
d
i
me
n
si
o
n
a
n
d
a
sso
c
i
a
t
e
d
d
i
me
n
si
o
n
b
a
se
d
o
n
t
h
e
C
-
C
a
n
d
G
-
P
a
l
g
o
r
i
t
h
m;
r
e
c
o
n
st
r
u
c
t
s
a
n
d
o
p
t
i
mi
z
e
s B
P
b
a
se
d
o
n
t
h
e
W
T
.
W
T
c
a
n
e
a
s
i
l
y
c
a
u
se
si
g
n
a
l
n
o
i
se
i
n
t
e
r
f
e
r
e
n
c
e
,
a
n
d
d
e
c
o
mp
o
si
t
i
o
n
sc
a
l
e
i
s
d
i
f
f
i
c
u
l
t
;
n
e
w
mo
d
e
l
i
s t
o
o
c
o
mp
l
e
x
a
n
d
l
o
w
e
r
s t
h
e
g
e
n
e
r
a
l
i
z
a
t
i
o
n
p
e
r
f
o
r
man
c
e
.
W
a
v
e
l
e
t
t
r
a
n
sf
o
r
m
B
P
n
e
u
r
a
l
n
e
t
w
o
r
k
12
2
0
1
8
Q
i
D
u
a
n
,
e
t
al
.
,
[
3
0
]
L
o
n
g
sh
o
r
t
-
t
e
r
m me
mo
r
y
S
L
T
-
L
S
T
M
T
h
e
n
o
n
-
st
a
t
i
o
n
a
r
y
,
c
o
r
r
e
l
a
t
e
d
a
n
d
p
e
r
i
o
d
i
c
t
r
a
f
f
i
c
d
a
t
a
a
r
e
p
r
o
c
e
sse
d
b
y
S
L
T
t
o
d
e
c
o
mp
o
se
t
h
e
t
r
e
n
d
a
n
d
e
l
i
mi
n
a
t
e
t
h
e
r
a
n
d
o
m n
o
i
se
.
O
n
l
y
p
a
y
s a
t
t
e
n
t
i
o
n
t
o
t
h
e
d
a
t
a
c
h
a
r
a
c
t
e
r
i
st
i
c
s o
f
n
e
t
w
o
r
k
t
r
a
f
f
i
c
,
w
h
i
l
e
i
g
n
o
r
i
n
g
n
o
n
l
i
n
e
a
r
mu
l
t
i
-
sca
l
e
c
h
a
r
a
c
t
e
r
i
st
i
c
s o
f
t
r
a
f
f
i
c
.
S
e
a
so
n
a
l
l
o
e
ss t
r
e
n
d
13
2
0
1
8
W
e
n
q
u
a
n
X
u
e
t
a
l
.
,
[
3
2
]
D
e
e
p
l
e
a
r
n
i
n
g
n
e
u
r
a
l
n
e
t
w
o
r
k
D
B
N
-
A
R
/
A
R
I
M
A
C
o
mp
u
t
e
s re
si
d
u
a
l
s
t
r
a
f
f
i
c
d
a
t
a
b
a
se
d
o
n
t
h
e
A
R
a
n
d
A
R
I
M
A
;
a
d
j
u
s
t
s
t
h
e
mo
d
e
l
c
o
e
f
f
i
c
i
e
n
t
s
a
f
t
e
r
t
r
a
i
n
i
n
g
t
h
e
r
e
si
d
u
a
l
s.
C
o
n
st
a
n
t
l
y
a
d
j
u
st
s
b
y
t
h
e
r
e
si
d
u
a
l
s
a
n
d
c
o
e
f
f
i
c
i
e
n
t
,
a
n
d
t
h
e
c
a
l
c
u
l
a
t
i
o
n
sc
a
l
e
a
n
d
t
i
me
w
i
l
l
c
o
n
t
i
n
u
e
t
o
i
n
c
r
e
a
se
.
D
e
e
p
b
e
l
i
e
f
n
e
t
w
o
r
k
A
R
/
A
R
I
M
A
mo
d
e
l
T
ab
le
2
.
A
p
p
licatio
n
o
f
d
ec
o
m
p
o
s
itio
n
tech
n
iq
u
e
i
n
h
y
b
r
id
m
o
d
el
N
o
.
Y
e
a
r
A
u
t
h
o
r
O
r
i
g
i
n
a
l
M
o
d
e
l
H
y
b
r
i
d
M
o
d
e
l
S
t
r
e
n
g
t
h
L
i
mi
t
a
t
i
o
n
1
2
0
1
7
L
a
i
se
n
N
i
e
,
e
t
a
l
.
,
[
39
]
D
i
scre
t
e
w
a
v
e
l
e
t
t
r
a
n
sf
o
r
m
D
W
T
-
D
B
N
-
GM
I
t
i
s d
i
v
i
d
e
d
i
n
t
o
l
o
w
a
n
d
h
i
g
h
-
p
a
ss
c
o
mp
o
n
e
n
t
s
b
y
D
W
T
;
G
a
u
ssi
a
n
mo
d
e
l
p
r
e
d
i
c
t
s
h
i
g
h
-
p
a
ss
c
o
mp
o
n
e
n
t
,
w
h
i
l
e
D
B
N
me
t
h
o
d
p
r
e
d
i
c
t
s l
o
w
-
p
a
ss c
o
mp
o
n
e
n
t
.
I
g
n
o
r
e
s
t
h
e
d
i
f
f
i
c
u
l
t
y
o
f
w
a
v
e
l
e
t
d
e
c
o
mp
o
si
t
i
o
n
sc
a
l
e
a
n
d
t
h
e
c
o
mp
l
e
x
i
t
y
o
f
t
h
e
n
e
w
mo
d
e
l
st
r
u
c
t
u
r
e
,
l
e
a
d
i
n
g
t
o
l
o
w
g
e
n
e
r
a
l
i
z
a
t
i
o
n
p
e
r
f
o
r
man
c
e
.
G
a
u
ssi
o
n
mo
d
e
l
D
e
e
p
b
e
l
i
e
f
n
e
t
w
o
r
k
2
2
0
1
7
L
i
n
a
P
a
n
,
e
t
a
l
.
,
[
4
9
]
Ec
h
o
s
t
a
t
e
n
e
t
w
o
r
k
V
M
D
-
BA
-
ESN
D
e
c
o
mp
o
se
s
t
h
e
c
o
mp
l
e
x
se
q
u
e
n
c
e
i
n
t
o
se
v
e
r
a
l
si
mp
l
e
c
o
mp
o
n
e
n
t
s
b
y
V
M
D
;
g
l
o
b
a
l
o
p
t
i
m
i
z
a
t
i
o
n
i
s
u
se
d
t
o
d
e
t
e
r
mi
n
e
t
h
e
i
n
i
t
i
a
l
w
e
i
g
h
t
o
f
ESN
a
n
d
r
e
d
u
c
e
s
t
h
e
i
n
f
l
u
e
n
c
e
.
D
e
c
o
mp
o
si
t
i
o
n
l
a
y
e
r
s a
f
f
e
c
t
t
h
e
d
e
c
o
mp
o
si
t
i
o
n
e
f
f
e
c
t
a
n
d
u
l
t
i
m
a
t
e
l
y
a
f
f
e
c
t
t
h
e
a
c
c
u
r
a
c
y
o
f
p
r
e
d
i
c
t
i
o
n
.
V
a
r
i
a
t
i
o
n
a
l
mo
d
e
d
e
c
o
mp
o
si
t
i
o
n
B
a
t
a
l
g
o
r
i
t
h
m
3
2
0
1
8
L
a
i
se
n
N
i
e
,
e
t
a
l
.
,
[
40
]
D
i
scre
t
e
w
a
v
e
l
e
t
t
r
a
n
sf
o
r
m
D
W
T
-
D
B
N
-
S
C
S
I
t
i
s d
i
v
i
d
e
d
i
n
t
o
l
o
w
a
n
d
h
i
g
h
-
p
a
ss
c
o
mp
o
n
e
n
t
s
b
y
D
W
T
;
D
B
N
mo
d
e
l
p
r
e
d
i
c
t
s l
o
w
-
p
a
ss c
o
mp
o
n
e
n
t
,
w
h
e
r
e
a
s
S
C
S
me
t
h
o
d
p
r
e
d
i
c
t
s
h
i
g
h
-
p
a
ss
c
o
m
p
o
n
e
n
t
.
I
g
n
o
r
e
s
t
h
e
d
i
f
f
i
c
u
l
t
y
o
f
w
a
v
e
l
e
t
d
e
c
o
mp
o
si
t
i
o
n
sc
a
l
e
a
n
d
t
h
e
c
o
mp
l
e
x
i
t
y
o
f
t
h
e
n
e
w
mo
d
e
l
st
r
u
c
t
u
r
e
,
l
e
a
d
i
n
g
t
o
l
o
w
g
e
n
e
r
a
l
i
z
a
t
i
o
n
p
e
r
f
o
r
man
c
e
.
D
e
e
p
b
e
l
i
e
f
n
e
t
w
o
r
k
S
p
a
t
i
o
t
e
mp
o
r
a
l
c
o
mp
r
e
ssi
v
e
se
n
si
n
g
4
2
0
1
8
H
a
i
p
e
n
g
Lu
,
e
t
a
l
.
,
[
42
]
L
o
n
g
sh
o
r
t
-
t
e
r
m me
mo
r
y
WT
-
L
S
T
M
D
e
c
o
mp
o
se
s
t
h
e
t
r
a
f
f
i
c
b
y
W
T
a
n
d
p
r
e
d
i
c
t
s
b
a
se
d
o
n
t
h
e
W
T
-
L
S
T
M
h
y
b
r
i
d
mo
d
e
l
.
T
h
e
sca
l
e
d
i
f
f
i
c
u
l
t
y
o
f
W
T
a
n
d
t
h
e
l
o
c
a
l
o
p
t
i
mal
so
l
u
t
i
o
n
o
f
L
S
T
M
a
r
e
n
o
t
w
e
l
l
d
e
a
l
t
w
i
t
h
.
W
a
v
e
l
e
t
t
r
a
n
sf
o
r
m
5
2
0
1
8
R
i
s
h
a
b
h
M
a
d
a
n
,
e
t
al
.
,
[
43
]
D
i
scre
t
e
w
a
v
e
l
e
t
t
r
a
n
sf
o
r
m
D
W
T
-
A
R
I
M
A
-
R
N
N
D
e
c
o
mp
o
se
s
r
e
c
o
n
st
r
u
c
t
i
o
n
t
r
a
f
f
i
c
i
n
t
o
d
e
t
a
i
l
s a
n
d
a
p
p
r
o
x
i
mat
e
s
c
o
mp
o
n
e
n
t
s
b
y
i
D
W
T
;
A
R
I
M
A
mo
d
e
l
i
s u
se
d
t
o
p
r
e
d
i
c
t
t
h
e
l
o
w
c
o
mp
o
n
e
n
t
a
n
d
R
N
N
i
s
u
se
d
t
o
p
r
e
d
i
c
t
t
h
e
h
i
g
h
c
o
m
p
o
n
e
n
t
.
T
h
e
si
g
n
a
l
d
e
c
o
m
p
o
si
t
i
o
n
a
n
d
r
e
c
o
n
st
r
u
c
t
i
o
n
i
s
su
e
s
a
r
e
su
f
f
i
c
i
e
n
t
l
y
a
d
d
r
e
sse
d
,
b
u
t
t
h
e
c
a
l
c
u
l
a
t
i
o
n
o
f
R
N
N
i
s
d
i
f
f
i
c
u
l
t
a
n
d
t
h
e
W
T
d
e
c
o
mp
o
si
t
i
o
n
sca
l
e
i
s c
o
m
p
l
e
x
.
A
R
/
A
R
I
M
A
mo
d
e
l
R
e
c
u
r
r
e
n
t
n
e
u
r
a
l
n
e
t
w
o
r
k
6
2
0
1
8
T
i
a
n
Z
h
o
n
g
d
a
[
44
]
Emp
i
r
i
c
a
l
mo
d
e
d
e
c
o
mp
o
si
t
i
o
n
EM
D
-
EL
M
D
e
c
o
mp
o
se
s
t
h
e
t
r
a
f
f
i
c
i
n
t
o
st
a
b
l
e
d
a
t
a
s
i
g
n
a
l
s w
i
t
h
d
i
f
f
e
r
e
n
t
c
h
a
r
a
c
t
e
r
i
st
i
c
sc
a
l
e
s
b
y
EM
D
;
c
o
mb
i
n
e
s
t
h
e
a
d
v
a
n
t
a
g
e
s o
f
EL
M
's
f
a
st
l
e
a
r
n
i
n
g
s
p
e
e
d
t
o
i
mp
r
o
v
e
t
h
e
a
c
c
u
r
a
c
y
o
f
n
e
t
w
o
r
k
t
r
a
f
f
i
c
p
r
e
d
i
c
t
i
o
n
.
T
h
e
E
L
M
r
a
n
d
o
m i
n
p
u
t
w
e
i
g
h
t
s
d
e
v
i
a
t
i
o
n
s
a
n
d
mo
d
e
a
l
i
a
s
i
n
g
o
f
EM
D
c
a
n
i
n
v
e
r
se
l
y
i
mp
a
c
t
t
h
e
a
c
c
u
r
a
c
y
o
f
t
h
e
mo
d
e
l
.
Ex
t
r
e
me
l
e
a
r
n
i
n
g
mac
h
i
n
e
S
e
a
so
n
a
l
l
o
e
ss
t
r
e
n
d
d
e
c
o
mp
o
si
t
i
o
n
p
r
o
c
e
ss
m
e
t
h
o
d
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
co
mp
r
eh
en
s
ive
r
ev
iew
o
n
h
y
b
r
id
n
etw
o
r
k
tr
a
ffic p
r
ed
ictio
n
mo
d
el
(
Jin
mei
S
h
i
)
1457
T
ab
le
2
.
A
p
p
licatio
n
o
f
d
ec
o
m
p
o
s
itio
n
tech
n
iq
u
e
i
n
h
y
b
r
id
m
o
d
el
(
co
n
tin
u
e)
N
o
.
Y
e
a
r
A
u
t
h
o
r
O
r
i
g
i
n
a
l
M
o
d
e
l
H
y
b
r
i
d
M
o
d
e
l
S
t
r
e
n
g
t
h
L
i
mi
t
a
t
i
o
n
7
2
0
1
8
W
a
n
w
e
i
H
u
a
n
g
,
e
t
a
l
.
,
[
47
]
En
se
mb
l
e
e
mp
i
r
i
c
a
l
mo
d
e
d
e
c
o
mp
o
si
t
i
o
n
QNN
-
EEM
D
D
e
c
o
mp
o
si
t
i
o
n
t
i
me
se
r
i
e
s i
n
t
o
I
M
F
b
y
EEM
D
so
a
s
t
o
r
e
mo
v
e
mo
d
a
l
a
l
i
a
si
n
g
a
n
d
o
p
t
i
m
i
z
e
p
a
r
a
me
t
e
r
s
o
f
Q
N
N
mo
d
e
l
t
o
a
v
o
i
d
l
o
c
a
l
o
p
t
i
m
a
l
so
l
u
t
i
o
n
.
F
o
c
u
s
es
o
n
so
l
v
i
n
g
t
h
e
p
r
o
b
l
e
m
o
f
si
g
n
a
l
d
e
c
o
mp
o
si
t
i
o
n
,
b
u
t
h
o
w
t
o
o
p
t
i
mi
z
e
t
h
e
EE
M
D
d
e
p
e
n
d
e
n
t
a
mp
l
i
t
u
d
e
a
n
d
e
x
p
e
r
i
e
n
c
e
i
s
s
t
i
l
l
a
c
h
a
l
l
e
n
g
e
A
r
t
i
f
i
c
i
a
l
n
e
u
r
a
l
n
e
t
w
o
r
k
Q
u
a
n
t
u
n
8
2
0
1
9
W
e
n
b
o
C
h
e
n
,
e
t
a
l
.
[
46
]
Emp
i
r
i
c
a
l
mo
d
e
d
e
c
o
mp
o
si
t
i
o
n
EM
D
-
PSO
-
S
V
M
E
l
i
mi
n
a
t
e
s
n
o
i
se
f
r
o
m d
a
t
a
b
a
se
d
on
t
h
e
E
M
D
,
a
n
d
o
p
t
i
mi
z
e
s
S
V
M
b
a
se
d
o
n
t
h
e
P
S
O
.
T
h
e
i
n
f
l
u
e
n
c
e
o
f
mo
d
a
l
a
l
i
a
s
i
n
g
i
n
EM
D
o
n
mo
d
e
l
a
c
c
u
r
a
c
y
i
s
i
g
n
o
r
e
d
.
S
u
p
p
o
r
t
v
e
c
t
o
r
ma
c
h
i
n
e
P
a
r
t
i
c
l
e
sw
a
r
m
o
p
t
i
m
i
z
a
t
i
o
n
9
2
0
1
9
X
i
n
g
h
a
n
Xu
,
e
t
a
l
.
,
[
5
0
]
B
P
n
e
u
r
a
l
n
e
t
w
o
r
k
C
EE
M
D
-
V
M
D
-
FA
-
B
P
N
N
I
mp
r
o
v
e
s
EEM
D
d
e
f
e
c
t
s b
y
i
n
c
r
e
a
si
n
g
a
d
a
p
t
i
v
e
w
h
i
t
e
n
o
i
se
a
mp
l
i
t
u
d
e
a
n
d
f
o
r
m
s
a
t
w
o
-
st
a
g
e
d
e
c
o
mp
o
si
t
i
o
n
w
i
t
h
V
M
D
t
e
c
h
n
o
l
o
g
y
;
o
p
t
i
mi
z
e
s
t
h
e
t
h
r
e
sh
o
l
d
a
n
d
w
e
i
g
h
t
o
f
B
P
n
e
u
r
a
l
n
o
d
e
b
a
se
d
o
n
F
A
t
o
i
mp
r
o
v
e
t
h
e
a
b
i
l
i
t
y
o
f
f
u
n
c
t
i
o
n
a
p
p
r
o
x
i
mat
i
o
n
t
o
n
e
u
r
a
l
n
e
t
w
o
r
k
.
I
g
n
o
r
es
t
h
e
n
u
mb
e
r
o
f
V
M
D
d
e
c
o
mp
o
si
t
i
o
n
l
a
y
e
r
s
w
h
i
c
h
c
a
n
i
n
f
l
u
e
n
c
e
t
h
e
d
e
c
o
mp
o
si
t
i
o
n
e
f
f
e
c
t
,
t
h
u
s re
d
u
c
i
n
g
t
h
e
p
r
e
d
i
c
t
i
o
n
a
c
c
u
r
a
c
y
o
f
t
h
e
mo
d
e
l
.
V
a
r
i
a
t
i
o
n
a
l
mo
d
e
d
e
c
o
mp
o
si
t
i
o
n
En
se
mb
l
e
e
mp
i
r
i
c
a
l
mo
d
e
d
e
c
o
mp
o
si
t
i
o
n
F
i
r
e
f
l
y
a
l
g
o
r
i
t
h
m
10
2
0
1
9
Y
i
n
g
H
a
n
,
e
t
al
.
,
[
51
]
V
a
r
i
a
t
i
o
n
a
l
mo
d
e
d
e
c
o
mp
o
si
t
i
o
n
V
M
D
-
PSR
-
I
F
O
A
-
ESN
I
mp
r
o
v
e
s
F
O
A
w
i
t
h
t
h
e
c
l
o
u
d
mo
d
e
l
a
n
d
t
h
e
l
e
v
y
f
l
i
g
h
t
f
u
n
c
t
i
o
n
t
o
o
p
t
i
mi
z
e
V
M
D
mo
d
e
s a
n
d
i
t
e
r
a
t
i
v
e
f
a
c
t
o
r
;
t
h
e
s
u
b
-
mo
d
e
s
a
f
t
e
r
V
M
D
d
e
c
o
mp
o
si
t
i
o
n
a
r
e
r
e
c
o
n
st
r
u
c
t
e
d
b
y
P
S
R
;
ESN
mu
l
t
i
p
l
e
su
b
r
e
se
r
v
o
i
r
s
i
s
a
u
t
o
mat
i
c
a
l
l
y
b
u
i
l
t
a
c
c
o
r
d
i
n
g
t
o
o
p
t
i
m
i
z
e
d
d
e
c
o
m
p
o
si
t
i
o
n
r
e
su
l
t
s
by
V
M
D
.
T
h
e
n
e
w
c
o
mb
i
n
e
d
mo
d
e
l
n
e
e
d
s
l
o
n
g
e
r
c
a
l
c
u
l
a
t
i
o
n
t
i
me
,
w
h
i
c
h
i
s
d
u
e
t
o
sy
n
c
h
r
o
n
o
u
s o
p
t
i
m
i
z
a
t
i
o
n
i
n
s
i
d
e
a
n
d
o
u
t
si
d
e
o
f
t
h
e
mu
l
t
i
p
l
e
su
b
r
e
se
r
v
o
i
r
s
.
T
h
i
s w
i
l
l
i
mp
a
c
t
t
h
e
t
r
a
i
n
i
n
g
t
i
me
o
f
t
h
e
mo
d
e
l
a
n
d
a
f
f
e
c
t
t
h
e
c
o
n
v
e
r
g
e
n
c
e
sp
e
e
d
a
n
d
t
h
e
p
e
r
f
o
r
man
c
e
o
f
t
h
e
mo
d
e
l
p
r
e
d
i
c
t
i
o
n
.
P
h
a
se
s
p
a
c
e
r
e
c
o
n
st
r
u
c
t
i
o
n
C
l
o
u
d
g
e
n
e
r
a
t
o
r
F
r
u
i
t
f
l
y
o
p
t
i
mi
z
a
t
i
o
n
a
l
g
o
r
i
t
h
m
Ec
h
o
s
t
a
t
e
n
e
t
w
o
r
k
ACK
NO
WL
E
D
G
E
M
E
NT
S
T
h
e
au
th
o
r
s
w
o
u
ld
li
k
e
to
t
h
a
n
k
Fac
u
lt
y
o
f
C
o
m
p
u
ti
n
g
an
d
I
n
f
o
r
m
atics,
Un
i
v
er
s
i
t
y
Ma
la
y
s
ia
Sab
ah
an
d
also
Hain
a
n
Vo
ca
tio
n
a
l
Un
i
v
er
s
it
y
o
f
Sc
ien
ce
a
n
d
T
e
ch
n
o
lo
g
y
u
n
d
er
2
0
1
9
A
ca
d
em
ic
C
o
o
p
er
atio
n
an
d
C
o
llab
o
r
ativ
e
E
d
u
ca
tio
n
P
r
o
j
e
ct
o
f
C
h
in
a
Mi
n
i
s
tr
y
o
f
E
d
u
ca
t
io
n
(
No
.
2
0
1
9
0
1
0
1
2
0
0
8
)
.
RE
F
E
R
E
NC
E
S
[1
]
Jitt
a
w
iri
y
a
n
u
k
o
o
n
C.
,
“
A
p
p
ro
x
ima
ti
o
n
o
f
re
g
re
ss
io
n
-
b
a
se
d
f
a
u
lt
m
i
n
i
m
iza
ti
o
n
f
o
r
n
e
t
w
o
rk
tra
ff
i
c
,
”
T
EL
KOM
NIKA
T
e
lec
o
mm
u
n
ica
ti
o
n
,
Co
m
p
u
t
in
g
,
El
e
c
tro
n
ics
a
n
d
C
o
n
tro
l
,
v
o
l.
1
8
,
n
o
.
4
,
p
p
.
1
8
0
2
-
1
8
0
8
,
2
0
2
0
.
[2
]
Ch
a
k
ra
v
a
rth
y
V
.
D
.
a
n
d
Am
u
th
a
B.
,
“
P
a
t
h
b
a
se
d
l
o
a
d
b
a
la
n
c
in
g
f
o
r
d
a
ta ce
n
ter n
e
tw
o
rk
s u
sin
g
S
DN
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
C
o
mp
u
t
e
r
En
g
i
n
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
9
,
n
o
.
4
,
p
p
.
3
2
7
9
-
3
2
8
5
,
2
0
1
9
.
[3
]
Kh
in
S
w
e
Yin
a
n
d
M
a
y
Ay
e
Kh
in
e
,
“
Op
ti
m
a
l
re
m
o
te
a
c
c
e
ss
tro
jan
s
d
e
te
c
ti
o
n
b
a
se
d
o
n
n
e
t
w
o
rk
b
e
h
a
v
io
r,
”
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
E
lec
trica
l
a
n
d
C
o
mp
u
ter
En
g
in
e
e
rin
g
(
IJ
ECE
)
, v
o
l.
9
,
n
o
.
3
,
p
p
.
2
1
7
7
-
2
1
8
4
,
2
0
1
9
[4
]
A
l
g
e
lal
Z
.
M
.
,
G
h
a
n
iald
h
a
h
e
r
E
.
A
.
,
A
b
d
u
l
-
W
a
d
o
o
d
D
.
N.
,
“
Bo
t
n
e
t
d
e
tec
ti
o
n
u
si
n
g
e
n
se
m
b
le
c
las
si
fiers
o
f
n
e
t
w
o
rk
f
lo
w
,
”
In
ter
n
a
ti
o
n
a
l
J
o
u
rn
a
l
o
f
El
e
c
trica
l
a
n
d
Co
m
p
u
ter
E
n
g
i
n
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
1
0
,
n
o
.
3
,
p
p
.
2
5
4
3
-
2
5
5
0
,
2
0
2
0
.
[5
]
Bo
u
k
h
a
lf
a
A
.
,
A
b
d
e
ll
a
o
u
i
A
.
,
Hm
in
a
N
.
,
e
t
a
l.
,
“
L
S
T
M
d
e
e
p
lea
rn
in
g
m
e
th
o
d
f
o
r
n
e
tw
o
rk
in
tru
sio
n
d
e
tec
ti
o
n
sy
s
tem
,”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
E
l
e
c
t
r
ic
a
l
a
n
d
C
o
m
p
u
t
e
r
E
n
g
i
n
e
e
ri
n
g
(
I
J
E
C
E
)
,
v
o
l
.
1
0
,
n
o
.
3
,
p
p
.
3
3
1
5
-
3
3
2
2
,
2
0
2
0
.
[6
]
Ly
se
n
k
o
S
.
,
Bo
b
ro
v
n
ik
o
v
a
K
.
,
M
a
ti
u
k
h
S
.
,
e
t
a
l.
,
“
De
tec
ti
o
n
o
f
th
e
b
o
tn
e
ts'
lo
w
-
ra
te
D
Do
S
a
tt
a
c
k
s
b
a
se
d
o
n
se
lf
-
s
i
m
i
l
a
r
i
t
y
,”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
E
l
e
c
t
r
i
c
a
l
a
n
d
C
o
m
p
u
t
e
r
E
n
g
i
n
e
e
r
i
n
g
(
I
J
E
C
E
)
,
v
o
l
.
1
0
,
n
o
.
4
,
p
p
.
3
6
5
1
-
3
6
5
9
,
2020
.
[7
]
S
h
e
tt
y
N
.
P
.
,
S
h
e
tt
y
J
.
,
Na
ru
la
R
.
,
e
t
a
l.
,
“
Co
m
p
a
riso
n
stu
d
y
o
f
m
a
c
h
in
e
lea
rn
in
g
c
la
ss
if
ier
s
to
d
e
tec
t
a
n
o
m
a
li
e
s,”
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
E
lec
trica
l
a
n
d
C
o
mp
u
ter
En
g
in
e
e
rin
g
(
IJ
ECE
)
,
v
o
l.
1
0
,
n
o
.
5
,
p
p
.
5
4
4
5
-
5
4
5
2
,
2
0
2
0
.
[8
]
Ko
n
g
,
W
.
,
Do
n
g
,
Z.
Y.,
Jia
,
Y.
,
Hill
,
D.
J.
,
X
u
,
Y.,
Zh
a
n
g
,
Y.
,
“
S
h
o
rt
-
T
e
rm
Re
sid
e
n
ti
a
l
L
o
a
d
F
o
re
c
a
stin
g
Ba
se
d
o
n
L
S
T
M
Re
c
u
rre
n
t
Ne
u
ra
l
Ne
t
w
o
rk
,
”
IEE
E
T
r
a
n
sa
c
ti
o
n
s
o
n
S
m
a
rt
G
rid
,
v
o
l.
1
0
,
n
o
.
1
,
p
p
.
8
4
1
-
8
5
1
,
2
0
1
9
.
[9
]
Ch
e
n
,
K.,
Ya
n
,
Z.
J.,
a
n
d
Hu
o
,
Q.,
“
A
c
o
n
tex
t
-
se
n
siti
v
e
-
c
h
u
n
k
BP
T
T
a
p
p
ro
a
c
h
to
trai
n
i
n
g
d
e
e
p
L
S
T
M
//
B
L
S
T
M
re
c
u
rre
n
t
n
e
u
ra
l
n
e
tw
o
rk
s
f
o
r
o
ffli
n
e
h
a
n
d
w
rit
in
g
re
c
o
g
n
it
io
n
,
”
2
0
1
5
1
3
t
h
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Do
c
u
me
n
t
An
a
lys
is
a
n
d
Rec
o
g
n
it
i
o
n
(
ICDAR)
,
T
u
n
is
,
2
0
1
5
,
p
p
.
4
11
-
4
1
5
.
[1
0
]
W
h
e
e
b
A
H.
,
“
P
e
rf
o
r
m
a
n
c
e
e
v
a
lu
a
ti
o
n
o
f
UD
P
,
DCCP,
S
CT
P
a
n
d
T
F
RC
f
o
r
d
iff
e
re
n
t
tr
a
ff
ic
f
lo
w
in
w
ired
n
e
t
w
o
r
k
s
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
E
l
e
c
t
r
i
c
a
l
a
n
d
C
o
m
p
u
t
e
r
E
n
g
i
n
e
e
r
i
n
g
(
I
J
E
C
E
)
,
v
o
l
.
7
,
n
o
.
6
,
p
p
.
3552
-
3557
,
2
0
1
7
.
[1
1
]
W
e
i,
D.,
“
Ne
tw
o
rk
tra
ffi
c
p
re
d
ictio
n
b
a
se
d
o
n
RB
F
n
e
u
ra
l
n
e
tw
o
rk
o
p
ti
m
ize
d
b
y
i
m
p
ro
v
e
d
g
ra
v
it
a
ti
o
n
se
a
rc
h
a
lg
o
rit
h
m
,
”
Ne
u
ra
l
Co
mp
u
ti
n
g
a
n
d
Ap
p
li
c
a
ti
o
n
s
,
v
o
l.
2
8
,
n
o
.
8
,
p
p
.
2
3
0
3
-
2
3
1
2
,
2
0
1
7
.
[1
2
]
L
i,
Q.,
a
n
d
L
in
,
R.
-
C.
,
“
A
N
e
w
A
p
p
ro
a
c
h
f
o
r
Ch
a
o
ti
c
T
i
m
e
S
e
ri
e
s
P
re
d
ictio
n
Us
in
g
Re
c
u
rre
n
t
Ne
u
ra
l
Ne
t
w
o
rk
,
”
M
a
th
e
ma
ti
c
a
l
Pro
b
lem
s in
E
n
g
in
e
e
rin
g
,
v
o
l.
2
0
1
6
,
2
0
1
6
,
p
p
.
1
-
9
.
[1
3
]
Jo
sh
i,
M
.
R.
,
a
n
d
Ha
d
i,
T
.
H.,
“
A
Re
v
ie
w
o
f
Ne
t
w
o
rk
T
ra
ff
ic
A
n
a
l
y
sis
a
n
d
P
re
d
icti
o
n
T
e
c
h
n
iq
u
e
s,
”
a
rXiv:1
5
0
7
.
0
5
7
2
2
,
2
0
1
5
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
11
,
No
.
2
,
A
p
r
il 2
0
2
1
:
1
4
5
0
-
1459
1458
[1
4
]
Ba
tes
J
.
M
.
a
n
d
G
r
a
n
g
e
r
C
.
W
.
,
“
T
h
e
c
o
m
b
in
a
ti
o
n
o
f
f
o
re
c
a
sts,”
Op
e
ra
t
io
n
a
l
Res
e
a
rc
h
Qu
a
rt
e
rly
,
v
o
l.
2
0
,
n
o
.
4
,
p
p
.
4
5
1
-
4
6
8
,
1
9
6
9
.
[1
5
]
D
i
c
k
i
n
s
o
n
J
.
P.
,
“
S
o
m
e
c
o
m
m
e
n
ts
o
n
t
h
e
c
o
m
b
i
n
a
t
i
o
n
o
f
f
o
re
c
a
s
t
s
,
”
O
p
e
r
a
t
i
o
n
a
l
R
e
s
e
a
r
c
h
q
u
a
r
t
e
r
l
y
,
v
o
l
.
2
6
,
n
o
.
1
,
p
p
.
2
0
5
-
2
10
,
1
9
7
5
.
[1
6
]
Ya
n
g
,
Y.,
Ch
e
n
,
Y.,
L
i
,
C.
,
G
u
i,
X
.
,
L
i,
L
.
,
“
Ne
t
w
o
rk
T
r
a
f
f
i
c
P
re
d
ictio
n
Ba
se
d
o
n
L
S
S
V
M
Op
ti
m
ize
d
b
y
P
S
O,”
2
0
1
4
IEE
E
1
1
th
In
tl
Co
n
f
o
n
Ub
i
q
u
it
o
u
s
I
n
telli
g
e
n
c
e
a
n
d
Co
m
p
u
t
i
n
g
a
n
d
2
0
1
4
IEE
E
1
1
th
I
n
tl
C
o
n
f
o
n
A
u
t
o
n
o
mic
a
n
d
T
ru
ste
d
C
o
mp
u
ti
n
g
a
n
d
2
0
1
4
IEE
E
1
4
th
I
n
tl
Co
n
f
o
n
S
c
a
la
b
le
Co
m
p
u
t
in
g
a
n
d
Co
mm
u
n
i
c
a
ti
o
n
s
a
n
d
Its
Asso
c
ia
ted
W
o
rk
sh
o
p
s
,
Ba
li
,
2
0
1
4
,
p
p
.
8
2
9
-
8
3
4
.
[1
7
]
Eb
e
r
h
a
rt,
R
.
,
a
n
d
Ke
n
n
e
d
y
,
J.,
“
A
Ne
w
Op
ti
m
ize
r
Us
in
g
P
a
rti
c
l
e
S
w
a
r
m
T
h
e
o
r
y
,
”
M
HS
'
9
5
.
Pr
o
c
e
e
d
in
g
s
o
f
t
h
e
S
ixth
I
n
ter
n
a
ti
o
n
a
l
S
y
mp
o
siu
m
o
n
M
icr
o
M
a
c
h
in
e
a
n
d
Hu
ma
n
S
c
ien
c
e
,
Na
g
o
y
a
,
Ja
p
a
n
,
1
9
9
5
,
p
p
.
3
9
-
43.
[1
8
]
Ho
ll
a
n
d
J
.
H.
,
“
G
e
n
e
ti
c
a
lg
o
rit
h
m
s
a
n
d
t
h
e
o
p
ti
m
a
l
a
ll
o
c
a
ti
o
n
o
f
tri
a
ls,”
S
IAM
J
o
u
rn
a
l
o
n
C
o
mp
u
ti
n
g
,
v
o
l.
2
,
n
o
.
2
,
p
p
.
88
-
1
0
5
,
1
9
7
3
.
[1
9
]
Na
r
a
y
a
n
a
n
,
A
.
,
a
n
d
M
o
o
re
,
M
.
,
“
Qu
a
n
t
u
m
-
in
sp
ired
g
e
n
e
ti
c
a
l
g
o
rit
h
m
s
,
”
Pro
c
e
e
d
in
g
s
o
f
IEE
E
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Evo
l
u
ti
o
n
a
ry
Co
m
p
u
t
a
ti
o
n
,
Na
g
o
y
a
,
Ja
p
a
n
,
1
9
9
6
,
p
p
.
61
-
66.
[2
0
]
L
a
n
,
Y.,
S
o
h
,
Y.
C.
,
a
n
d
Hu
a
n
g
,
G
.
Bin
,
“
Tw
o
-
sta
g
e
e
x
tre
m
e
lea
rn
in
g
m
a
c
h
in
e
f
o
r
re
g
re
ss
io
n
,
”
Ne
u
ro
c
o
mp
u
ti
n
g
,
vo
l.
7
3
,
n
o
.
1
6
-
1
8
,
p
p
.
3
0
2
8
-
3
0
3
8
,
2
0
1
0
.
[2
1
]
Ha
n
,
F
.
,
Ya
o
,
H.,
a
n
d
L
in
g
,
Q.,
“
A
n
i
m
p
ro
v
e
d
e
v
o
lu
ti
o
n
a
ry
e
x
tr
e
m
e
lea
rn
in
g
m
a
c
h
in
e
b
a
se
d
o
n
p
a
rti
c
le
sw
a
r
m
o
p
ti
m
i
z
a
ti
o
n
,
”
Ne
u
r
o
c
o
mp
u
ti
n
g
,
v
o
l.
1
1
6
,
p
p
.
8
7
-
93
,
2
0
1
3
.
[2
2
]
Zh
a
n
g
,
W
.
,
a
n
d
W
e
i,
D.
,
“
P
re
d
ictio
n
f
o
r
n
e
tw
o
rk
tra
ff
ic
o
f
r
a
d
ial
b
a
sis
f
u
n
c
ti
o
n
n
e
u
ra
l
n
e
tw
o
rk
m
o
d
e
l
b
a
se
d
o
n
i
m
p
r
o
v
e
d
p
a
r
t
i
c
l
e
s
w
a
r
m
o
p
t
i
m
i
z
a
t
i
o
n
a
l
g
o
r
i
t
h
m
,
”
N
e
u
r
a
l
C
o
m
p
u
t
i
n
g
a
n
d
A
p
p
l
i
c
a
t
i
o
n
s
,
v
o
l
.
2
9
,
n
o
.
4
,
p
p
.
1
1
4
3
-
1
1
5
2
,
2
0
1
8
.
[2
3
]
He
,
T
.
,
Da
n
,
T
.
,
W
e
i,
Y.,
L
i,
H.,
Ch
e
n
,
X
.
,
Qin
,
G
.
,
“
P
a
rti
c
le
sw
a
rm
o
p
ti
m
iz
a
ti
o
n
RBF
n
e
u
ra
l
n
e
t
w
o
rk
m
o
d
e
l
f
o
r
in
tern
e
t
traff
ic
p
re
d
ictio
n
,”
2
0
1
6
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
In
t
e
ll
ig
e
n
t
T
ra
n
sp
o
rta
ti
o
n
,
Bi
g
Da
t
a
&
S
ma
rt
Cit
y
(
ICIT
BS
)
,
Ch
a
n
g
sh
a
,
2
0
1
6
,
p
p
.
5
2
1
-
5
2
4
.
[2
4
]
Zh
a
n
g
,
K.,
L
ian
g
,
L
.
,
a
n
d
Hu
a
n
g
,
Y.
,
“
A
Ne
t
w
o
rk
T
r
a
ff
ic
P
re
d
icti
o
n
M
o
d
e
l
Ba
se
d
o
n
Qu
a
n
tu
m
In
sp
ired
P
S
O
a
n
d
Ne
u
ra
l
Ne
tw
o
rk
,”
2
0
1
3
S
ixth
I
n
ter
n
a
ti
o
n
a
l
S
y
mp
o
siu
m
o
n
Co
mp
u
ta
ti
o
n
a
l
In
tel
li
g
e
n
c
e
a
n
d
De
sig
n
,
Ha
n
g
z
h
o
u
,
2
0
1
3
,
p
p
.
2
1
9
-
2
2
2
.
[2
5
]
Na
y
a
k
,
S
.
,
Na
y
a
k
,
S
.
,
a
n
d
S
in
g
h
,
P
.
J.
P
.
,
“
An
In
tro
d
u
c
ti
o
n
to
Q
u
a
n
tu
m
Ne
u
ra
l
Co
m
p
u
ti
n
g
,
”
J
o
u
rn
a
l
o
f
Gl
o
b
a
l
Res
e
a
rc
h
in
C
o
mp
u
ter
S
c
ien
c
e
,
v
o
l.
2
,
n
o
.
8
,
p
p
.
5
0
-
55
,
2
0
1
1
.
[2
6
]
T
ian
,
H.,
Z
h
o
u
,
X
.
,
a
n
d
L
iu
,
J.,
“
A
h
y
b
rid
n
e
tw
o
rk
traff
ic
p
re
d
ictio
n
m
o
d
e
l
b
a
se
d
o
n
o
p
ti
m
ize
d
n
e
u
ra
l
n
e
tw
o
rk
.
P
a
ra
ll
e
l
a
n
d
Distri
b
u
ted
C
o
m
p
u
ti
n
g
,
”
2
0
1
7
1
8
t
h
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
Pa
ra
ll
e
l
a
n
d
Distrib
u
t
e
d
Co
mp
u
t
in
g
,
Ap
p
li
c
a
ti
o
n
s
a
n
d
T
e
c
h
n
o
l
o
g
ies
(
PDCAT
)
,
T
a
ip
e
i
,
2
0
1
7
,
p
p
.
2
8
4
-
2
8
7
.
[2
7
]
Ha
n
,
Y.,
Jin
g
,
Y.,
a
n
d
L
i,
K.
,
“
M
u
lt
i
-
ste
p
p
re
d
ictio
n
f
o
r
th
e
n
e
tw
o
rk
tra
ff
ic
b
a
se
d
o
n
e
c
h
o
sta
te
n
e
tw
o
rk
o
p
ti
m
i
z
e
d
b
y
q
u
a
n
tu
m
-
b
e
h
a
v
e
d
f
ru
it
f
l
y
o
p
ti
m
iza
ti
o
n
a
lg
o
rit
h
m
,
”
2
0
1
7
2
9
th
C
h
in
e
se
Co
n
tro
l
a
n
d
De
c
is
io
n
C
o
n
fer
e
n
c
e
(
CCDC
)
,
Ch
o
n
g
q
in
g
,
2
0
1
7
,
p
p
.
2
2
7
0
-
2
2
7
4
.
[2
8
]
A
z
z
o
u
n
i
A
.
,
P
u
j
o
ll
e
G
.
,
“
A
lo
n
g
s
h
o
rt
-
term
m
e
m
o
r
y
re
c
u
rre
n
t
n
e
u
r
a
l
n
e
tw
o
rk
f
ra
m
e
w
o
rk
f
o
r
n
e
t
w
o
rk
tra
ff
ic
m
a
tri
x
p
re
d
ictio
n
,
”
a
rXiv p
re
p
ri
n
t
a
rXiv:
1
7
0
5
.
0
5
6
9
0
,
2
0
1
7
.
[2
9
]
Zh
u
o
,
Q.,
L
i,
Q.,
Ya
n
,
H
.
,
Qi,
Y.
,
“
L
o
n
g
sh
o
rt
-
term
m
e
m
o
r
y
n
e
u
ra
l
n
e
tw
o
rk
f
o
r
n
e
tw
o
rk
tra
ff
ic
p
r
e
d
ictio
n
,
”
2
0
1
7
1
2
t
h
In
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
In
telli
g
e
n
t
S
y
ste
ms
a
n
d
Kn
o
wled
g
e
En
g
in
e
e
rin
g
(
IS
KE)
,
Na
n
ji
n
g
,
2
0
1
8
,
p
p
.
1
-
6.
[3
0
]
Du
a
n
,
Q.,
W
e
i,
X
.
,
G
a
o
,
Y.,
Zh
o
u
,
F
.
,
“
Ba
se
S
tatio
n
T
ra
ff
i
c
P
re
d
ict
io
n
b
a
se
d
o
n
S
T
L
-
L
S
T
M
Ne
t
w
o
r
k
s,”
2
0
1
8
2
4
t
h
Asia
-
Pa
c
if
ic Co
n
fer
e
n
c
e
o
n
C
o
m
mu
n
ica
t
io
n
s (
AP
CC)
,
Nin
g
b
o
,
Ch
in
a
,
2
0
1
8
,
p
p
.
4
0
7
-
4
1
2
.
[3
1
]
Ha
n
,
Y.,
a
n
d
L
i,
K.,
“
I
m
p
ro
v
e
d
F
OA
-
ES
N
m
e
th
o
d
u
si
n
g
o
p
p
o
sit
io
n
-
b
a
se
d
lea
rn
in
g
m
e
c
h
a
n
ism
fo
r
th
e
n
e
tw
o
rk
traff
ic p
re
d
ictio
n
w
it
h
m
u
lt
ip
le st
e
p
s,”
2
0
1
7
C
h
i
n
e
se
Au
to
m
a
ti
o
n
C
o
n
g
re
ss
(
CAC),
Jin
a
n
,
2
0
1
7
,
p
p
.
7
1
8
3
-
7
1
8
7
.
[3
2
]
X
u
,
W
.
,
P
e
n
g
,
H.,
Zen
g
,
X.,
Zh
o
u
,
F
.
,
T
ian
,
X
.
,
P
e
n
g
,
X
.
,
“
A
h
y
b
rid
m
o
d
e
ll
in
g
m
e
th
o
d
f
o
r
ti
m
e
se
ries
f
o
re
c
a
stin
g
b
a
se
d
o
n
a
li
n
e
a
r
re
g
re
ss
io
n
m
o
d
e
l
a
n
d
d
e
e
p
lea
rn
i
n
g
,
”
Ap
p
li
e
d
I
n
te
ll
ig
e
n
c
e
,
v
o
l
.
4
9
,
p
p
.
3
0
0
2
-
3
0
1
5
,
2
0
1
9
.
[3
3
]
G
a
o
X
.
,
S
u
n
L
.
,
S
u
n
D.
,
“
A
n
e
n
h
a
n
c
e
d
p
a
rti
c
le
s
w
a
r
m
o
p
ti
m
iza
ti
o
n
a
lg
o
r
it
h
m
,
”
In
fo
rm
a
ti
o
n
T
e
c
h
n
o
l
o
g
y
J
o
u
rn
a
l
,
v
o
l.
8
,
n
o
.
8
,
p
p
.
1
2
6
3
-
1
2
6
8
,
2
0
0
9
.
[3
4
]
El
ro
b
y
M
.
M
.
H
.
,
M
e
k
h
a
m
e
r
S
.
F
.
,
T
a
laa
t
H
.
E
.
A
.
,
e
t
a
l.
,
“
P
o
p
u
lat
io
n
b
a
se
d
o
p
ti
m
iza
ti
o
n
a
lg
o
rit
h
m
s
i
m
p
ro
v
e
m
e
n
t
us
i
n
g
t
h
e
p
r
e
d
i
c
t
i
v
e
p
a
r
t
i
c
le
s
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
E
l
e
c
t
r
i
c
a
l
a
n
d
C
o
m
p
u
t
e
r
E
n
g
i
n
e
e
r
i
n
g
(
IJ
E
C
E)
,
v
o
l
.
1
0
,
n
o
.
3
,
p
p
.
3
2
6
1
-
3
2
7
4
,
2
0
2
0
.
[3
5
]
Ried
i,
R.
H.,
Rib
e
ir
o
,
V.
J.,
a
n
d
Ba
ra
n
iu
k
,
R.
G
.
,
“
A
M
u
lt
if
ra
c
t
a
l
W
a
v
e
let
M
o
d
e
l
w
it
h
A
p
p
li
c
a
ti
o
n
t
o
Ne
tw
o
rk
T
ra
ff
ic,
"
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
I
n
fo
rm
a
t
io
n
T
h
e
o
ry
,
v
o
l.
4
5
,
n
o
.
3
,
p
p
.
9
9
2
-
1
0
1
8
,
1
9
9
9
.
[3
6
]
Hu
a
n
g
,
N.
E
.
,
S
h
e
n
,
Z.
,
L
o
n
g
,
S
.
R
.
,
W
u
,
M
.
C.
,
S
h
ih
,
H.
H.,
Ye
n
,
N
.
,
L
iu
,
H.
H.
,
“
T
h
e
e
m
p
iri
c
a
l
m
o
d
e
d
e
c
o
m
p
o
siti
o
n
a
n
d
th
e
Hilb
e
rt
s
p
e
c
tru
m
f
o
r
n
o
n
li
n
e
a
r
a
n
d
n
o
n
-
s
tatio
n
a
ry
ti
m
e
se
ries
a
n
a
l
y
sis,”
R
o
y
a
l
S
o
c
iety
o
f
L
o
n
d
o
n
Pr
o
c
e
e
d
in
g
s
S
e
rie
s A
,
v
o
l
.
4
5
4
,
n
o
.
1
,
p
p
.
9
0
3
-
9
9
5
,
1
9
9
8
.
[3
7
]
Z
h
a
n
g
,
J.,
Ya
n
,
R.
,
G
a
o
,
R.
X
.
,
F
e
n
g
,
Z.
,
“
P
e
rf
o
rm
a
n
c
e
e
n
h
a
n
c
e
m
e
n
t
o
f
e
n
se
m
b
le
e
m
p
iri
c
a
l
m
o
d
e
d
e
c
o
m
p
o
siti
o
n
,
”
M
e
c
h
a
n
ica
l
S
y
ste
ms
a
n
d
S
i
g
n
a
l
P
ro
c
e
ss
in
g
,
v
o
l.
2
4
,
n
o
.
7
,
p
p
.
2
1
0
4
-
2
1
2
3
,
2
0
1
0
.
[3
8
]
Dra
g
o
m
ir
e
tsk
i
y
,
K.
a
n
d
Z
o
ss
o
,
D.,
“
V
a
riati
o
n
a
l
m
o
d
e
d
e
c
o
m
p
o
s
it
io
n
,
”
IEE
E
T
ra
n
sa
c
t
i
o
n
s
o
n
S
i
g
n
a
l
Pro
c
e
ss
in
g
,
v
o
l.
6
2
,
n
o
.
3
,
p
p
.
5
3
1
-
5
4
4
,
2
0
1
4
.
[3
9
]
Nie
,
L
.
,
Jia
n
g
,
D.,
Y
u
,
S
.
,
S
o
n
g
,
H.
,
“
Ne
tw
o
rk
T
r
a
ff
ic
P
re
d
icti
o
n
Ba
se
d
o
n
De
e
p
Be
li
e
f
Ne
t
w
o
rk
in
W
irele
ss
M
e
sh
Ba
c
k
b
o
n
e
Ne
t
w
o
rk
s,”
2
0
1
7
IEE
E
W
ire
les
s
Co
mm
u
n
ica
ti
o
n
s
a
n
d
Ne
two
rk
in
g
Co
n
fer
e
n
c
e
(
W
CNC)
,
S
a
n
F
ra
n
c
isc
o
,
CA
,
2
0
1
7
,
p
p
.
1
-
5
.
[4
0
]
Nie
,
L
.
,
W
a
n
g
,
X
.
,
W
a
n
,
L
.
,
S
o
n
g
,
H.,
Jia
n
g
,
D.
,
“
Ne
tw
o
rk
T
ra
ff
i
c
P
re
d
ictio
n
Ba
se
d
o
n
De
e
p
Be
li
e
f
Ne
t
w
o
rk
a
n
d
S
p
a
ti
o
tem
p
o
ra
l
C
o
m
p
re
ss
iv
e
S
e
n
sin
g
in
W
irele
ss
M
e
sh
Ba
c
k
b
o
n
e
Ne
tw
o
rk
s,”
W
ire
les
s
Co
mm
u
n
ica
ti
o
n
s
a
n
d
M
o
b
il
e
Co
m
p
u
ti
n
g
,
v
o
l.
2
0
1
8
,
2
0
1
8
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
co
mp
r
eh
en
s
ive
r
ev
iew
o
n
h
y
b
r
id
n
etw
o
r
k
tr
a
ffic p
r
ed
ictio
n
mo
d
el
(
Jin
mei
S
h
i
)
1459
[4
1
]
P
e
n
g
,
Z
.
K.,
T
se
,
P
.
W
.
,
a
n
d
C
h
u
,
F
.
L
.
,
“
A
n
im
p
ro
v
e
d
Hilb
e
rt
-
Hu
a
n
g
tran
sf
o
rm
a
n
d
it
s
a
p
p
l
ica
ti
o
n
i
n
v
ib
ra
ti
o
n
sig
n
a
l
a
n
a
l
y
sis,”
J
o
u
rn
a
l
o
f
S
o
u
n
d
a
n
d
Vi
b
r
a
ti
o
n
,
v
o
l.
2
8
6
,
n
o
.
1
-
2
,
p
p
.
1
8
7
-
2
0
5
,
2
0
0
5
.
[4
2
]
L
u
,
H.,
a
n
d
Ya
n
g
,
F
.
,
“
A
Ne
t
w
o
rk
T
r
a
ff
ic
P
re
d
ictio
n
M
o
d
e
l
Ba
se
d
o
n
W
a
v
e
let
T
r
a
n
sf
o
r
m
a
ti
o
n
a
n
d
L
S
T
M
Ne
tw
o
rk
,
”
2
0
1
8
IEE
E
9
th
I
n
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
S
o
ft
wa
r
e
En
g
i
n
e
e
rin
g
a
n
d
S
e
rv
ice
S
c
ien
c
e
(
ICS
ES
S
)
,
Be
ij
in
g
,
Ch
i
n
a
,
2
0
1
8
,
p
p
.
1
-
4.
[4
3
]
M
a
d
a
n
,
R.
,
a
n
d
S
a
ra
th
i,
M
.
P
.
,
“
P
re
d
ictin
g
Co
m
p
u
ter
Ne
tw
o
rk
T
r
a
ff
ic:
A
T
i
m
e
S
e
ri
e
s
F
o
re
c
a
stin
g
A
p
p
ro
a
c
h
u
sin
g
DWT
,
A
RIM
A
a
n
d
RNN
,
”
2
0
1
8
El
e
v
e
n
th
I
n
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
C
o
n
tem
p
o
ra
ry
C
o
mp
u
ti
n
g
(
IC3
)
,
No
i
d
a
,
2
0
0
8
,
p
p
.
1
-
5
.
[4
4
]
T
ian
,
Z.
,
M
a
o
,
C
.
,
W
a
n
g
,
G
.
,
a
n
d
Re
n
,
Y.,
“
A
p
p
ro
a
c
h
f
o
r
ti
m
e
se
ri
e
s
p
re
d
ictio
n
b
a
se
d
o
n
e
m
p
iri
c
a
l
m
o
d
e
d
e
c
o
m
p
o
siti
o
n
a
n
d
e
x
trem
e
lea
rn
in
g
m
a
c
h
in
e
,
”
2018
C
h
in
e
se
Co
n
tro
l
a
n
d
De
c
isio
n
Co
n
fer
e
n
c
e
(
C
CDC)
,
S
h
e
n
y
a
n
g
,
2
0
1
8
,
p
p
.
3
1
1
9
-
3
1
2
3
.
[4
5
]
De
e
ri
n
g
,
R.
,
a
n
d
Ka
ise
r,
J.
F
.
,
“
T
h
e
u
se
o
f
a
m
a
sk
in
g
sig
n
a
l
to
im
p
ro
v
e
Em
p
iri
c
a
l
M
o
d
e
De
c
o
m
p
o
siti
o
n
,
”
Pro
c
e
e
d
in
g
s.
(
ICAS
S
P
'
0
5
).
IEE
E
In
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
Aco
u
stics
,
S
p
e
e
c
h
,
a
n
d
S
ig
n
a
l
P
ro
c
e
ss
in
g
,
2
0
0
5
,
P
h
il
a
d
e
l
p
h
ia,
P
A
,
v
o
l.
4
,
p
p
.
iv
/
4
8
5
-
iv
/4
8
8
,
2
0
0
5
.
[4
6
]
Ch
e
n
,
W
.
,
S
h
a
n
g
,
Z.
,
a
n
d
C
h
e
n
,
Y.,
“
A
No
v
e
l
H
y
b
rid
Ne
t
w
o
rk
T
ra
ff
ic
P
re
d
ictio
n
A
p
p
ro
a
c
h
Ba
s
e
d
o
n
S
u
p
p
o
r
t
V
e
c
to
r
M
a
c
h
in
e
s,”
J
o
u
r
n
a
l
o
f
Co
mp
u
ter
Ne
two
rk
s a
n
d
C
o
mm
u
n
ica
ti
o
n
s,
v
o
l
.
2
0
1
9
,
2
0
1
9
.
[4
7
]
Hu
a
n
g
,
W
.
,
Zh
a
n
g
,
J
.
,
L
ian
g
,
S
.
,
a
n
d
S
u
n
,
H.,
“
Ba
c
k
b
o
n
e
Ne
tw
o
rk
T
ra
ff
i
c
P
re
d
icti
o
n
Ba
se
d
o
n
M
o
d
ifi
e
d
EE
M
D
a
n
d
Qu
a
n
tu
m
Ne
u
ra
l
Ne
tw
o
rk
,
”
W
ire
les
s P
e
r
so
n
a
l
C
o
mm
u
n
c
a
ti
o
n
s,
v
o
l.
9
9
,
p
p
.
1
5
6
9
-
1
5
8
8
,
2
0
1
8
.
[4
8
]
Du
,
S
.
,
L
iu
,
T
.
,
Hu
a
n
g
,
D.
,
L
i,
G.,
“
An
Op
ti
m
a
l
En
se
m
b
le
E
m
p
iri
c
a
l
M
o
d
e
De
c
o
m
p
o
siti
o
n
M
e
th
o
d
f
o
r
V
i
b
ra
ti
o
n
S
ig
n
a
l
De
c
o
m
p
o
siti
o
n
,
”
J
o
u
rn
a
l
o
f
Vi
b
ra
ti
o
n
a
n
d
Aco
u
stics
,
v
o
l.
3
9
,
n
o
.
3
,
p
p
.
0
3
1
0
0
3
(1
)
-
0
2
1
0
0
3
(1
8
)
,
2
0
1
6
.
[4
9
]
P
a
n
,
L
.
,
Ch
e
n
g
,
J.,
L
i,
H.,
Zh
a
n
g
,
Y.,
Ch
e
n
,
X
.
,
“
A
n
Im
p
ro
v
e
d
Ech
o
S
tate
Ne
tw
o
rk
Ba
se
d
o
n
Va
riatio
n
a
l
M
o
d
e
De
c
o
m
p
o
siti
o
n
a
n
d
Ba
t
O
p
ti
m
iz
a
ti
o
n
f
o
r
In
ter
n
e
t
T
ra
ff
ic
F
o
re
c
a
s
ti
n
g
,
”
2
0
1
7
I
EE
E
1
7
t
h
In
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
o
n
C
o
mm
u
n
ica
ti
o
n
T
e
c
h
n
o
lo
g
y
(
ICCT
)
,
Ch
e
n
g
d
u
,
2
0
1
7
,
p
p
.
4
1
7
-
4
2
1
.
[5
0
]
X
u
,
X
.
,
a
n
d
Re
n
,
W
.
,
“
S
S
sy
m
m
e
tr
y
A
H
y
b
rid
M
o
d
e
l
Ba
se
d
o
n
a
T
w
o
-
L
a
y
e
r
De
c
o
m
p
o
siti
o
n
A
p
p
ro
a
c
h
a
n
d
a
n
Op
ti
m
ize
d
Ne
u
ra
l
Ne
t
w
o
rk
f
o
r
C
h
a
o
ti
c
T
im
e
S
e
ries
P
re
d
ictio
n
,
”
J
o
u
rn
a
l
S
y
mm
e
try
,
v
o
l.
1
1
,
n
o
.
5
,
p
.
6
1
0
,
2
0
1
9
[5
1
]
Ha
n
,
Jin
g
,
L
i.
,
a
n
d
Dim
iro
v
sk
i,
G
.
M
.
,
“
Ne
t
w
o
rk
T
r
a
ff
ic
P
re
d
icti
o
n
Us
in
g
V
a
riatio
n
a
l
M
o
d
e
De
c
o
m
p
o
siti
o
n
a
n
d
M
u
lt
i
-
re
se
rv
o
irs
Ech
o
S
tate
Ne
tw
o
rk
,
”
I
EE
E
Acc
e
ss
,
v
o
l.
7
,
p
p
.
1
3
8
3
6
4
-
1
3
8
3
7
7
,
2
0
1
9
.
B
I
O
G
RAP
H
I
E
S
O
F
AUTH
O
RS
J
in
m
e
i
S
h
i
P
h
D
S
tu
d
e
n
t,
Co
m
p
u
ter
S
c
ien
c
e
,
F
a
c
u
lt
y
o
f
Co
m
p
u
ti
n
g
a
n
d
In
f
o
rm
a
ti
c
s,
Un
iv
e
rsit
y
M
a
la
y
sia
S
a
b
a
h
.
S
h
e
h
a
s
b
e
e
n
e
n
g
a
g
e
d
in
re
se
a
rc
h
a
n
d
tea
c
h
in
g
f
o
r
1
1
y
e
a
rs
a
n
d
h
a
s
a
l
w
a
y
s
b
e
e
n
in
th
e
f
o
re
f
ro
n
t
o
f
sc
ien
ti
f
ic
re
s
e
a
rc
h
.
S
h
e
p
lay
s
a
n
a
c
ti
v
e
ro
le
in
th
e
c
o
n
stru
c
ti
o
n
o
f
c
o
m
p
u
ter
sc
ien
c
e
a
n
d
p
e
rso
n
n
e
l
train
i
n
g
,
a
n
d
h
a
s
g
o
o
d
p
ra
c
ti
c
a
l
e
x
p
e
rien
c
e
a
n
d
a
c
a
d
e
m
ic
f
o
u
n
d
a
ti
o
n
.
S
h
e
m
a
in
ly
re
se
a
rc
h
e
s
n
e
t
w
o
rk
tra
ff
i
c
p
re
d
ictio
n
,
a
lg
o
rit
h
m
a
n
a
l
y
sis
a
n
d
so
f
twa
re
.
In
re
c
e
n
t
y
e
a
rs,
sh
e
h
a
s
b
e
e
n
p
u
b
li
sh
e
d
1
0
p
a
p
e
rs,
3
sp
o
n
s
o
re
d
p
ro
jec
ts,
3
b
o
o
k
,
a
n
d
1
0
a
w
a
rd
s
a
b
o
v
e
th
e
p
ro
v
i
n
c
ial
lev
e
l.
Yu
-
B
e
n
g
Le
a
u
is
a
se
n
io
r
lec
t
u
re
r
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
t
t
h
e
F
a
c
u
lt
y
o
f
Co
m
p
u
ti
n
g
a
n
d
In
f
o
rm
a
ti
c
s,
Un
iv
e
rsiti
M
a
lay
sia
S
a
b
a
h
in
M
a
lay
si
a
.
He
re
c
e
iv
e
d
h
is
B.
S
c
d
e
g
re
e
(M
u
lt
im
e
d
ia
T
e
c
h
n
o
lo
g
y
)
f
ro
m
Un
iv
e
rsiti
M
a
l
a
y
sia
S
a
b
a
h
,
M
.
S
C
d
e
g
re
e
(In
f
o
r
m
a
ti
o
n
S
e
c
u
rit
y
)
f
ro
m
Un
iv
e
r
siti
T
e
k
n
o
lo
g
i
M
a
lay
sia
a
n
d
P
h
D
d
e
g
re
e
(In
tern
e
t
In
f
ra
stru
c
tu
re
s
S
e
c
u
rit
y
)
f
ro
m
Un
iv
e
rsiti
S
a
in
s
M
a
la
y
sia
.
His
c
u
rre
n
t
re
se
a
rc
h
in
tere
sts
a
re
in
tru
sio
n
a
lert
d
e
te
c
ti
o
n
a
n
d
p
re
d
ictio
n
,
n
e
tw
o
rk
se
c
u
rit
y
situ
a
ti
o
n
a
w
a
r
e
n
e
ss
,
IP
v
6
se
c
u
rit
y
,
In
tern
e
t
o
f
T
h
in
g
s
(Io
T
)
a
n
d
In
f
o
r
m
a
ti
o
n
Ce
n
tri
c
Ne
tw
o
rk
(ICN).
K
u
n
Li
wa
s
re
c
e
iv
e
d
th
e
B.
S
c
.
d
e
g
re
e
f
ro
m
S
h
a
n
d
o
n
g
Un
iv
e
rsit
y
o
f
S
c
ien
c
e
a
n
d
T
e
c
h
n
o
lo
g
y
in
2
0
0
5
,
a
n
d
re
c
e
iv
e
d
h
is
M
.
S
c
.
d
e
g
re
e
f
ro
m
L
iao
n
in
g
T
e
c
h
n
ica
l
Un
i
v
e
rsit
y
in
2
0
0
8
,
a
n
d
re
c
e
iv
e
d
His
P
h
.
D.
d
e
g
re
e
s
f
ro
m
No
rth
e
a
ste
rn
Un
iv
e
rsity
S
h
e
n
y
a
n
g
o
f
Ch
in
a
in
2
0
1
3
.
He
is
c
u
rre
n
tl
y
a
n
A
s
so
c
iate
P
ro
f
e
ss
o
r
in
Co
ll
e
g
e
o
f
En
g
in
e
e
rin
g
,
Bo
h
a
i
Un
iv
e
rsity
.
His
c
u
rre
n
t
re
se
a
rc
h
w
o
rk
s
f
o
c
u
s
o
n
c
o
m
p
lex
in
d
u
strial
p
ro
c
e
ss
m
o
d
e
li
n
g
,
in
telli
g
e
n
t
o
p
ti
m
a
l
c
o
n
tr
o
l,
m
a
c
h
in
e
lea
rn
in
g
,
a
n
d
th
e
ir
a
p
p
li
c
a
ti
o
n
.
J
o
e
H
e
n
r
y
O
b
it
is
a
n
As
so
c
iate
P
ro
f
e
ss
o
r
o
f
Co
m
p
u
ter
S
c
ien
c
e
,
d
e
p
a
rtm
e
n
t
o
f
D
a
ta
S
c
ien
c
e
a
t
Un
iv
e
rsiti
M
a
la
y
sia
S
a
b
a
h
.
His
m
a
in
re
se
a
r
c
h
in
tere
st
li
e
s
a
t
th
e
i
n
terf
a
c
e
o
f
Op
e
ra
ti
o
n
a
l
Re
se
a
rc
h
a
n
d
Co
m
p
u
ter
S
c
ien
c
e
.
In
p
a
rti
c
u
lar,
th
e
e
x
p
lo
ra
ti
o
n
a
n
d
d
e
v
e
lo
p
m
e
n
t
o
f
in
n
o
v
a
ti
v
e
Op
e
r
a
ti
o
n
a
l
Re
se
a
rc
h
,
A
rti
f
icia
l
In
telli
g
e
n
c
e
,
a
n
d
Distrib
u
ted
A
rti
f
icia
l
In
telli
g
e
n
c
e
m
o
d
e
ls
a
n
d
m
e
th
o
d
o
l
o
g
ies
f
o
r
a
u
to
m
a
ti
c
a
ll
y
p
ro
d
u
c
in
g
h
ig
h
q
u
a
li
ty
so
lu
ti
o
n
s
t
o
a
w
id
e
ra
n
g
e
o
f
re
a
l
w
o
rld
c
o
m
b
in
a
to
rial
o
p
ti
m
iza
ti
o
n
a
n
d
sc
h
e
d
u
li
n
g
p
ro
b
lem
s.
Dr.
Jo
e
Ob
tain
e
d
h
is
P
h
D
in
Co
m
p
u
ter
S
c
ien
c
e
f
ro
m
th
e
S
c
h
o
o
l
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
t
th
e
Un
iv
e
rsity
o
f
No
tt
in
g
h
a
m
.
His
P
h
D
t
h
e
sis
is
d
e
v
e
lo
p
in
g
a
No
v
e
l
M
e
ta
-
h
e
u
risti
c
,
Hy
p
e
r
-
h
e
u
risti
c
a
n
d
C
o
o
p
e
ra
ti
v
e
S
e
a
rc
h
.
Evaluation Warning : The document was created with Spire.PDF for Python.