I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Art
if
icia
l In
t
ellig
ence
(
I
J
-
AI
)
Vo
l.
15
,
No
.
2
,
A
p
r
il
2
0
2
6
,
p
p
.
1
1
8
1
~
1
1
9
3
I
SS
N:
2
2
5
2
-
8
9
3
8
,
DOI
: 1
0
.
1
1
5
9
1
/ijai.v
15
.i
2
.
p
p
1
1
8
1
-
1
1
9
3
1181
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
a
i
.
ia
esco
r
e.
co
m
K
no
wledg
e gra
ph
-
ba
sed enha
nc
ed
v
irtual net
wo
rk e
mbedding
for 6
G
clo
ud da
ta
center
deploy
men
t
Sh
o
uro
k
Abdelra
him
1
,
Sa
my
G
ho
niem
y
2
,
M
o
ha
m
ed
Abo
rizk
a
1
1
C
o
l
l
e
g
e
o
f
C
o
m
p
u
t
i
n
g
a
n
d
I
n
f
o
r
m
a
t
i
o
n
Te
c
h
n
o
l
o
g
y
,
A
r
a
b
A
c
a
d
e
my
f
o
r
S
c
i
e
n
c
e
a
n
d
Te
c
h
n
o
l
o
g
y
,
C
a
i
r
o
,
Eg
y
p
t
2
D
e
p
a
r
t
me
n
t
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
E
n
g
i
n
e
e
r
i
n
g
,
F
a
c
u
l
t
y
o
f
C
o
m
p
u
t
e
r
S
c
i
e
n
c
e
a
n
d
E
n
g
i
n
e
e
r
i
n
g
,
G
a
l
a
l
a
U
n
i
v
e
r
si
t
y
,
C
a
i
r
o
,
E
g
y
p
t
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
May
24
,
2
0
2
5
R
ev
is
ed
J
an
19
,
2
0
2
6
Acc
ep
ted
Feb
6
,
2
0
2
6
Virtu
a
l
n
e
two
r
k
e
m
b
e
d
d
i
n
g
(V
NE)
is
th
e
e
ffe
c
ti
v
e
m
a
p
p
in
g
o
f
v
irt
u
a
l
n
e
two
rk
s
o
n
to
sh
a
re
d
p
h
y
sic
a
l
s
u
b
stra
te
n
e
two
r
k
s
wh
il
e
b
o
o
stin
g
re
so
u
rc
e
u
ti
li
z
a
ti
o
n
a
n
d
e
n
su
ri
n
g
q
u
a
li
t
y
o
f
se
rv
ice
(Qo
S
).
VN
E
is
a
re
a
l
c
h
a
ll
e
n
g
e
in
n
e
two
rk
v
irt
u
a
li
z
a
ti
o
n
,
e
sp
e
c
ia
ll
y
in
t
h
e
p
e
rsp
e
c
ti
v
e
o
f
6
G
-
e
n
a
b
led
d
a
tac
e
n
ters
,
wh
e
re
t
h
e
d
e
m
a
n
d
f
o
r
u
l
tra
-
lo
w l
a
ten
c
y
,
h
e
a
v
y
c
o
n
n
e
c
ti
v
it
y
,
a
n
d
d
y
n
a
m
ic
re
so
u
rc
e
a
ll
o
c
a
ti
o
n
is
v
it
a
l.
T
h
e
p
ro
p
o
se
d
s
o
lu
ti
o
n
e
n
a
b
les
th
e
a
b
il
it
y
to
i
n
fe
r
in
d
irec
t
p
a
th
s
fo
r
th
e
re
so
u
rc
e
s
p
re
d
ictio
n
tas
k
o
n
t
h
e
kno
wle
d
g
e
g
ra
p
h
(KG
)
b
y
m
a
k
in
g
imp
l
icit
m
e
a
n
in
g
fu
l
re
latio
n
s
a
m
o
n
g
th
e
e
n
ti
ti
e
s
th
a
t
c
o
m
p
o
se
th
e
re
so
u
rc
e
n
e
two
rk
.
T
h
e
sim
u
latio
n
re
su
lt
s
in
d
ica
ted
th
e
in
fe
re
n
c
e
m
e
c
h
a
n
ism
sig
n
ifi
c
a
n
tl
y
imp
r
o
v
e
s
e
fficie
n
c
y
a
n
d
a
d
a
p
tab
il
it
y
.
Th
is
lea
d
s
to
o
v
e
ra
ll
p
e
rfo
rm
a
n
c
e
g
a
in
s
in
term
s
o
f
ru
n
ti
m
e
sta
b
il
it
y
,
re
so
u
rc
e
u
ti
li
z
a
ti
o
n
,
a
n
d
e
n
e
r
g
y
sa
v
in
g
s
in
d
y
n
a
m
ic
6
G
sc
e
n
a
rio
s.
Th
e
e
x
p
e
rime
n
tal
re
su
lt
s
sh
o
we
d
t
h
a
t
th
e
p
ro
p
o
se
d
s
o
lu
t
io
n
p
r
o
v
i
d
e
d
a
2
4
.
9
%
re
d
u
c
ti
o
n
in
e
n
e
rg
y
c
o
n
su
m
p
t
io
n
f
o
r
sm
a
ll
-
siz
e
d
v
irt
u
a
l
n
e
two
r
k
re
q
u
e
sts
(VN
Rs
),
wh
il
e
m
a
in
tain
in
g
2
4
.
8
%
a
n
d
2
3
.
9
%
fo
r
m
e
d
i
u
m
a
n
d
la
rg
e
VN
Rs,
re
sp
e
c
ti
v
e
ly
,
wh
il
e
it
si
g
n
ifi
c
a
n
tl
y
d
e
c
re
a
se
d
th
e
d
e
lay
t
ime
c
o
m
p
a
re
d
to
th
e
re
su
lt
e
d
d
e
lay
u
sin
g
t
h
e
b
a
se
li
n
e
m
o
d
e
ls
su
c
h
a
s
a
sy
n
c
h
ro
n
o
u
s
a
d
v
a
n
tag
e
a
c
to
r
-
c
rit
ic
(A3
C)
+
g
ra
p
h
c
o
n
v
o
lu
ti
o
n
a
l
n
e
tw
o
rk
(
G
CN
)
.
Th
e
r
e
su
lt
s
a
lso
c
o
n
firme
d
th
a
t
t
h
e
in
te
g
ra
ti
o
n
o
f
th
e
in
fe
re
n
c
e
e
n
g
i
n
e
a
lg
o
r
it
h
m
with
th
e
e
m
b
e
d
d
in
g
p
ro
c
e
ss
re
su
lt
s
i
n
re
m
a
rk
a
b
le
re
d
u
c
ti
o
n
i
n
th
e
e
x
e
c
u
ti
o
n
ti
m
e
wh
il
e
p
re
se
rv
i
n
g
e
m
b
e
d
d
i
n
g
a
c
c
u
ra
c
y
.
K
ey
w
o
r
d
s
:
6
G
clo
u
d
d
atac
en
ter
Kn
o
wled
g
e
g
r
ap
h
Netwo
r
k
m
ap
p
i
n
g
Vir
tu
al
n
etwo
r
k
em
b
ed
d
i
n
g
V
i
r
t
u
a
l
n
e
t
w
o
r
k
i
n
f
e
r
e
n
c
e
p
r
e
d
i
c
a
t
e
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
:
Sh
o
u
r
o
k
Ab
d
elr
a
h
im
C
o
lleg
e
o
f
C
o
m
p
u
tin
g
an
d
I
n
f
o
r
m
atio
n
T
ec
h
n
o
l
o
g
y
,
Ar
ab
A
ca
d
em
y
f
o
r
Scien
ce
an
d
T
ec
h
n
o
lo
g
y
C
air
o
,
E
g
y
p
t
E
m
ail:
s
h
o
r
o
u
k
_
ab
d
el
r
ah
im
@
cic
-
ca
ir
o
.
co
m
1.
I
NT
RO
D
UCT
I
O
N
Ma
n
y
r
esear
ch
o
u
tco
m
es,
s
u
ch
as
in
[
1
]
–
[
4
]
co
n
f
ir
m
e
d
th
at
m
ap
p
in
g
v
ir
tu
al
n
etwo
r
k
r
eq
u
ests
(
VNRs
)
o
n
to
p
h
y
s
ical
in
f
r
astr
u
ctu
r
es
wh
ile
g
u
ar
an
teein
g
o
p
tim
al
r
eso
u
r
ce
u
tili
za
tio
n
an
d
s
ca
lab
ilit
y
is
a
r
ea
l
ch
allen
g
e
i
n
n
etwo
r
k
v
ir
tu
aliz
atio
n
.
T
h
is
ch
allen
g
e
ar
is
es
f
r
o
m
th
e
co
m
p
lex
ity
o
f
r
eso
u
r
c
e
allo
ca
tio
n
an
d
th
e
co
n
s
tr
ain
ts
o
f
n
etwo
r
k
to
p
o
l
o
g
y
[
5
]
.
Fro
m
an
ar
tific
ial
in
tellig
en
ce
(
AI
)
p
er
s
p
ec
tiv
e,
v
ir
tu
al
n
etwo
r
k
em
b
ed
d
in
g
(
VNE
)
ca
n
b
e
r
e
g
ar
d
ed
as
a
d
ec
is
io
n
-
m
a
k
in
g
a
n
d
co
m
b
in
ato
r
ial
o
p
tim
izatio
n
task
th
at
r
e
q
u
ir
es
r
ea
s
o
n
in
g
ab
o
u
t
m
u
ltip
le
o
b
je
ctiv
es
an
d
h
id
d
en
r
elatio
n
s
h
ip
s
am
o
n
g
n
etwo
r
k
elem
en
ts
.
R
ec
en
t
s
tu
d
ies
h
av
e
d
em
o
n
s
tr
ated
a
n
o
tab
le
ad
v
an
ce
m
en
t
in
th
e
d
e
p
lo
y
m
e
n
t
o
f
VNE
an
d
r
eso
u
r
ce
allo
ca
tio
n
t
ec
h
n
iq
u
es
u
s
in
g
AI
m
eth
o
d
s
s
u
ch
as
m
ac
h
in
e
lea
r
n
in
g
(
ML
)
,
d
ee
p
r
ein
f
o
r
ce
m
en
t
lear
n
in
g
(
DR
L
)
,
a
n
d
g
r
a
p
h
n
eu
r
al
n
etwo
r
k
s
(
GNN)
[
6
]
–
[
1
0
]
,
w
h
ich
e
n
h
an
ce
n
etwo
r
k
p
r
e
d
ictio
n
an
d
m
a
p
p
in
g
ac
cu
r
ac
y
to
co
p
e
with
t
h
e
lar
g
e
v
o
l
u
m
e
a
n
d
v
ar
iab
le
r
ate
o
f
r
eq
u
ests
in
6
G
-
en
ab
led
d
atac
en
ter
s
.
Ho
wev
e
r
,
th
ese
s
o
lu
tio
n
s
o
f
ten
r
eq
u
ir
e
ex
ten
s
iv
e
tr
ain
in
g
d
ata
an
d
h
ig
h
co
m
p
u
tatio
n
al
r
eso
u
r
ce
s
,
wh
ich
lim
it
th
eir
p
r
ac
ticality
f
o
r
r
ea
l
-
tim
e
m
u
lti
-
ac
ce
s
s
s
er
v
ices
in
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
26
:
1
1
8
1
-
1
1
9
3
1182
d
y
n
am
ic
6
G
n
etwo
r
k
s
.
Oth
e
r
s
o
lu
tio
n
s
s
u
ch
as
d
y
n
am
ic
p
r
o
g
r
am
m
in
g
-
b
ased
alg
o
r
ith
m
[
1
1
]
,
r
ein
f
o
r
ce
m
en
t
lear
n
in
g
(
R
L
)
-
b
ased
d
y
n
am
ic
co
llab
o
r
ativ
e
m
u
lti
-
lay
er
VN
E
alg
o
r
ith
m
[
1
2
]
,
an
d
Ma
r
k
o
v
d
ec
is
io
n
p
r
o
ce
s
s
(
MD
P)
m
o
d
el
[
1
3
]
,
aim
ed
to
im
p
r
o
v
e
th
e
lo
a
d
b
alan
ci
n
g
a
n
d
o
p
tim
ized
r
eso
u
r
ce
allo
ca
t
io
n
b
y
co
o
r
d
in
atin
g
n
o
d
e
an
d
lin
k
m
ap
p
in
g
.
T
h
ey
f
ailed
to
r
ed
u
ce
co
m
p
u
tatio
n
a
l
co
m
p
lex
ity
d
u
e
to
th
e
lack
o
f
ad
ap
tiv
e
lear
n
in
g
lim
its
wh
ich
lim
it
th
eir
s
u
itab
ilit
y
in
6
G
d
y
n
am
ic
n
etwo
r
k
en
v
ir
o
n
m
en
ts
.
Z
h
a
n
g
et
a
l.
[
1
4
]
,
[
1
5
]
p
r
o
p
o
s
ed
a
co
m
b
in
in
g
h
o
r
izo
n
tal
f
ed
er
ated
an
d
RL
tech
n
iq
u
es
to
im
p
r
o
v
e
m
ap
p
in
g
ac
cu
r
ac
y
an
d
o
p
tim
ize
r
eso
u
r
ce
u
tili
za
tio
n
.
Ho
wev
er
,
th
ey
in
tr
o
d
u
ce
d
s
ig
n
if
ican
t
c
o
m
p
u
tatio
n
al
o
v
er
h
ea
d
d
u
e
to
th
e
u
s
e
o
f
b
r
ea
d
th
-
f
i
r
s
t
s
ea
r
ch
(
B
FS
)
in
lin
k
m
ap
p
in
g
.
Mo
r
e
r
ec
en
tly
,
k
n
o
wled
g
e
g
r
ap
h
s
(
KGs)
h
av
e
em
er
g
e
d
as a
p
r
o
m
is
in
g
s
o
lu
tio
n
f
o
r
in
tellig
en
t
VNE
o
p
tim
izatio
n
.
On
e
o
f
th
ese
s
o
lu
tio
n
s
is
t
h
e
p
r
o
p
o
s
ed
f
r
am
ewo
r
k
th
at
co
m
b
in
es
AI
-
b
ased
d
ec
is
io
n
m
ak
in
g
an
d
co
g
n
itiv
e
r
eso
u
r
ce
o
p
tim
izatio
n
al
g
o
r
i
th
m
s
f
o
r
6
G
n
etwo
r
k
s
[
1
6
]
,
b
u
t
h
ig
h
ly
co
m
p
lex
an
d
d
y
n
am
ic
n
etwo
r
k
s
lim
it th
eir
s
ca
lab
ilit
y
.
Als
o
,
Mitr
o
p
o
u
lo
u
et
a
l.
[
1
7
]
,
[
1
8
]
p
r
o
p
o
s
ed
th
e
u
s
e
o
f
s
tatic
KG
an
d
ML
to
d
etec
t
an
d
m
itig
ate
n
etwo
r
k
co
n
g
esti
o
n
an
d
an
o
m
aly
d
etec
tio
n
.
Ho
wev
er
,
li
m
itin
g
th
e
f
u
ll
ex
p
lo
itatio
n
o
f
laten
t
r
elatio
n
al
in
f
er
en
ce
s
.
Similar
ly
,
o
th
er
r
esear
ch
er
s
p
r
o
p
o
s
ed
g
r
ap
h
co
n
v
o
lu
tio
n
al
n
etwo
r
k
(
GC
N)
u
s
in
g
KG
f
o
r
VNE
.
T
h
eir
f
r
am
ewo
r
k
s
h
o
wed
lo
w
er
ef
f
icien
cy
in
r
ea
l
-
tim
e
im
p
lem
en
tatio
n
s
[
1
9
]
.
At
th
e
en
d
o
f
th
is
r
ev
iew,
an
in
ter
esti
n
g
wo
r
k
p
r
o
p
o
s
ed
i
n
[
2
0
]
p
r
esen
ted
a
c
o
m
b
in
atio
n
o
f
Pag
eRan
k
-
b
ased
h
eu
r
is
tic
VNE
an
d
B
ellm
an
-
Fo
r
d
alg
o
r
ith
m
to
r
a
n
k
n
o
d
es.
Alth
o
u
g
h
th
is
f
r
am
e
wo
r
k
im
p
r
o
v
es
th
e
p
er
f
o
r
m
an
ce
,
its
s
tatic
f
ea
tu
r
e
s
et
an
d
lack
o
f
m
ap
p
in
g
co
o
r
d
in
atio
n
im
p
ac
t
its
ad
ap
tab
ilit
y
in
d
y
n
am
ic
en
v
ir
o
n
m
e
n
ts
.
T
o
co
n
s
o
lid
ate
th
e
d
is
cu
s
s
io
n
,
T
ab
le
1
p
r
o
v
id
es
a
co
m
p
ar
ativ
e
o
v
e
r
v
iew
o
f
r
ep
r
esen
tativ
e
VNE
ap
p
r
o
ac
h
es,
in
clu
d
in
g
h
eu
r
is
tic,
DR
L
,
GNN,
f
ed
er
ated
lear
n
in
g
(
FL
)
,
an
d
KG
-
b
ased
m
eth
o
d
s
,
b
en
ch
m
a
r
k
ed
ag
ain
s
t
th
e
p
r
o
p
o
s
ed
d
y
n
am
ic
KG
f
r
am
ewo
r
k
.
T
h
is
s
u
m
m
ar
y
h
i
g
h
lig
h
ts
th
e
tr
ad
e
-
o
f
f
s
am
o
n
g
s
ca
lab
ilit
y
,
co
m
p
u
tatio
n
al
o
v
e
r
h
ea
d
,
ad
ap
tab
ilit
y
,
an
d
p
er
f
o
r
m
a
n
ce
m
etr
ics,
clar
if
y
in
g
wh
e
r
e
th
e
p
r
o
p
o
s
ed
s
o
lu
tio
n
ad
v
an
ce
s
th
e
s
tate
o
f
th
e
ar
t.
T
ab
le
1
.
C
o
m
p
a
r
ativ
e
an
aly
s
is
o
f
r
ep
r
esen
tativ
e
VNE
ap
p
r
o
a
ch
es f
o
r
6
G
d
atac
en
ter
s
A
p
p
r
o
a
c
h
R
e
p
r
e
se
n
t
a
t
i
v
e
w
o
r
k
s
K
e
y
f
e
a
t
u
r
e
s
S
c
a
l
a
b
i
l
i
t
y
&
a
d
a
p
t
a
b
i
l
i
t
y
O
v
e
r
h
e
a
d
&
l
i
m
i
t
a
t
i
o
n
s
R
e
p
o
r
t
e
d
me
t
r
i
c
s
H
e
u
r
i
s
t
i
c
VNE
-
N
R
[
2
0
]
N
o
d
e
r
a
n
k
i
n
g
;
sh
o
r
t
e
s
t
-
p
a
t
h
map
p
i
n
g
M
e
d
i
u
m
sca
l
a
b
i
l
i
t
y
;
l
o
w
a
d
a
p
t
a
b
i
l
i
t
y
Lo
w
o
v
e
r
h
e
a
d
;
st
a
t
i
c
f
e
a
t
u
r
e
s;
b
r
i
t
t
l
e
u
n
d
e
r
d
y
n
a
mi
c
s
I
mp
r
o
v
e
d
p
e
r
f
o
r
m
a
n
c
e
,
b
u
t
l
i
m
i
t
e
d
a
d
a
p
t
a
b
i
l
i
t
y
D
R
L
A
sy
n
c
h
r
o
n
o
u
s
a
d
v
a
n
t
a
g
e
a
c
t
o
r
-
c
r
i
t
i
c
(
A
3
C
)
+
G
C
N
[
2
1
]
;
D
e
e
p
V
i
N
E
[
2
2
]
En
d
-
to
-
e
n
d
n
o
d
e
/
l
i
n
k
p
o
l
i
c
y
l
e
a
r
n
i
n
g
M
e
d
i
u
m
sca
l
a
b
i
l
i
t
y
;
mo
d
e
r
a
t
e
a
d
a
p
t
a
b
i
l
i
t
y
H
i
g
h
o
v
e
r
h
e
a
d
;
l
o
n
g
c
o
n
v
e
r
g
e
n
c
e
;
sen
s
i
t
i
v
e
t
o
t
r
a
i
n
i
n
g
R
u
n
t
i
me
0
.
2
1
9
–
0
.
4
7
6
s;
a
c
c
e
p
t
a
n
c
e
8
2
–
8
7
%
@
1
1
0
0
-
t
i
m
e
u
n
i
t
s
G
N
N
+
R
L
G
C
N
+
R
L
[
2
3
]
;
G
r
a
p
h
V
i
N
E
[
2
4
]
G
C
N
e
n
c
o
d
e
r
s +
R
L
p
o
l
i
c
i
e
s
M
e
d
i
u
m
–
h
i
g
h
sca
l
a
b
i
l
i
t
y
;
mo
d
e
r
a
t
e
a
d
a
p
t
a
b
i
l
i
t
y
M
e
d
i
u
m
–
h
i
g
h
o
v
e
r
h
e
a
d
;
c
o
n
v
e
r
g
e
n
c
e
d
e
l
a
y
s
A
c
c
e
p
t
a
n
c
e
8
5
–
9
0
%
@
1
1
0
0
-
t
i
m
e
u
n
i
t
s
FL
H
o
r
i
z
o
n
t
a
l
F
L
-
V
N
E
[
1
5
]
C
r
o
ss
-
d
o
m
a
i
n
l
e
a
r
n
i
n
g
M
e
d
i
u
m
sca
l
a
b
i
l
i
t
y
;
mo
d
e
r
a
t
e
a
d
a
p
t
a
b
i
l
i
t
y
H
i
g
h
o
v
e
r
h
e
a
d
(
B
F
S
l
i
n
k
map
p
i
n
g
)
;
sca
l
a
b
i
l
i
t
y
i
ss
u
e
s
A
c
c
u
r
a
c
y
g
a
i
n
s
r
e
p
o
r
t
e
d
;
r
u
n
t
i
me
n
o
t
d
e
t
a
i
l
e
d
K
G
(
n
o
n
-
V
N
E)
[
1
7
]
–
[
1
9
]
S
e
ma
n
t
i
c
/
r
e
l
a
t
i
o
n
a
l
mo
d
e
l
i
n
g
M
e
d
i
u
m
sca
l
a
b
i
l
i
t
y
;
l
o
w
a
d
a
p
t
a
b
i
l
i
t
y
M
e
d
i
u
m
o
v
e
r
h
e
a
d
;
st
a
t
i
c
K
G
s
;
l
i
m
i
t
e
d
i
n
f
e
r
e
n
c
e
M
e
t
r
i
c
s
r
e
p
o
r
t
e
d
f
o
r
a
n
o
ma
l
y
/
l
i
n
k
p
r
e
d
i
c
t
i
o
n
P
r
o
p
o
se
d
Th
i
s
st
u
d
y
D
y
n
a
mi
c
K
G
+
h
i
d
d
e
n
i
n
f
e
r
e
n
c
e
e
n
g
i
n
e
H
i
g
h
sc
a
l
a
b
i
l
i
t
y
;
h
i
g
h
a
d
a
p
t
a
b
i
l
i
t
y
Lo
w
–
me
d
i
u
m
o
v
e
r
h
e
a
d
;
r
e
q
u
i
r
e
s
sema
n
t
i
c
e
n
r
i
c
h
me
n
t
R
u
n
t
i
me
0
.
6
5
–
0
.
8
5
s
;
u
t
i
l
i
z
a
t
i
o
n
8
5
%;
e
n
e
r
g
y
−
2
4
.
9
%
;
a
c
c
e
p
t
a
n
c
e
9
4
%
@
9
0
0
-
t
i
m
e
u
n
i
t
s
Un
lik
e
p
r
ev
i
o
u
s
ly
r
ep
o
r
ted
KG
-
b
ased
m
eth
o
d
s
,
wh
ich
ar
e
ty
p
ically
s
tatic
an
d
s
tr
u
g
g
le
t
o
ca
p
tu
r
e
d
ee
p
er
r
elatio
n
al
f
ea
tu
r
es
as
g
r
ap
h
s
ize
in
cr
ea
s
es,
th
e
p
r
o
p
o
s
ed
f
r
am
ewo
r
k
em
p
lo
y
s
a
tim
e
-
ev
o
lv
in
g
,
s
em
an
tically
en
r
ich
ed
KG
in
t
eg
r
ated
with
a
h
i
d
d
en
-
r
elatio
n
in
f
e
r
en
ce
e
n
g
in
e
o
p
e
r
atin
g
d
ir
ec
tly
with
in
t
h
e
VNE
p
ip
elin
e.
T
h
is
in
teg
r
ati
o
n
en
ab
les
th
e
in
f
er
en
ce
o
f
in
d
ir
ec
t
o
r
d
is
tan
t
r
eso
u
r
ce
co
n
n
ec
tio
n
s
,
ac
ce
ler
ates
s
tab
ilizatio
n
o
f
th
e
ac
ce
p
tan
ce
r
atio
,
an
d
d
eliv
er
s
co
n
s
is
ten
t
en
er
g
y
r
ed
u
ctio
n
s
,
wh
ile
also
av
o
id
in
g
t
h
e
h
ea
v
y
tr
ain
in
g
o
v
er
h
ea
d
s
ass
o
ciate
d
with
DR
L
-
an
d
GNN
-
b
ased
ap
p
r
o
ac
h
es.
T
o
a
d
d
r
ess
th
e
s
e
lim
itatio
n
s
,
th
i
s
p
ap
er
p
r
esen
ts
an
in
tellig
en
t
VNE
f
r
am
ewo
r
k
b
ased
o
n
d
y
n
am
ic
KG
wh
ich
is
ab
le
to
d
y
n
am
ically
o
p
tim
ize
r
eso
u
r
ce
allo
ca
tio
n
a
n
d
im
p
r
o
v
e
th
e
laten
t
c
o
m
p
u
tatio
n
al
ca
p
ab
ilit
ies
b
y
in
f
er
r
in
g
h
id
d
en
r
elatio
n
s
h
ip
s
b
etwe
en
th
e
n
etwo
r
k
ass
ets an
d
r
eso
u
r
ce
s
.
T
h
e
p
r
o
p
o
s
ed
s
o
l
u
tio
n
o
f
f
er
s
a
n
o
v
el
m
ec
h
a
n
is
m
f
o
r
im
p
r
o
v
in
g
th
e
ef
f
icien
cy
an
d
s
ca
lab
ilit
y
o
f
V
NE
in
6
G
-
d
atac
en
ter
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
K
n
o
w
led
g
e
g
r
a
p
h
-
b
a
s
ed
en
h
a
n
ce
d
virt
u
a
l n
etw
o
r
k
emb
ed
d
i
n
g
fo
r
6
G
clo
u
d
…
(
S
h
o
u
r
o
k
A
b
d
elra
h
im
)
1183
2.
P
RO
P
O
SE
D
SYS
T
E
M
I
n
th
is
p
ap
e
r
,
th
e
p
r
o
p
o
s
ed
f
r
am
ewo
r
k
is
d
esig
n
e
d
b
ased
o
n
th
e
s
em
an
tically
e
n
r
ich
ed
KG
m
o
d
el
th
at
is
d
ev
elo
p
e
d
to
u
n
if
y
th
e
n
o
d
es
an
d
e
d
g
es
g
e
n
er
ati
o
n
with
tech
n
ical
s
p
ec
if
icatio
n
s
an
d
s
em
a
n
tics
d
escr
ip
tio
n
th
at
r
ep
r
esen
t
th
e
d
atac
en
ter
ass
ets.
I
n
th
e
p
r
o
p
o
s
ed
f
r
am
ewo
r
k
,
th
e
co
r
e
elem
en
t
is
th
e
KG
-
b
ased
n
o
v
el
i
n
f
er
en
ce
alg
o
r
ith
m
,
w
h
ich
is
d
esig
n
ed
b
ased
o
n
th
e
s
em
an
tically
e
n
r
ich
ed
K
G
m
o
d
el,
to
en
ab
le
ad
ap
tab
le
an
d
d
y
n
am
ic
VNE
,
an
d
to
r
ea
s
o
n
an
d
in
f
er
th
e
laten
t
r
elatio
n
s
h
ip
s
am
o
n
g
p
h
y
s
ical
in
f
r
astru
ctu
r
e
co
m
p
o
n
en
ts
in
th
e
s
u
b
s
tr
ate
n
etwo
r
k
.
T
h
e
p
r
o
p
o
s
ed
f
r
am
e
wo
r
k
also
in
clu
d
es
p
r
o
ac
tiv
e
r
eso
u
r
ce
allo
ca
tio
n
m
o
d
u
le,
f
a
u
lt
an
ticip
atio
n
p
r
o
ce
s
s
,
an
d
in
tellig
en
t
tr
af
f
ic
m
o
n
ito
r
in
g
an
d
m
an
a
g
em
en
t
ag
en
ts
,
as
d
ep
icted
in
Fig
u
r
e
1
.
Fig
u
r
e
1
.
T
h
e
p
r
o
p
o
s
ed
s
y
s
te
m
2
.
1
.
T
ec
hn
ica
l descript
io
n c
o
llect
io
n a
nd
v
a
lid
a
t
io
n pha
s
e
T
h
is
p
h
ase
is
r
esp
o
n
s
ib
le
f
o
r
co
llectin
g
an
d
v
alid
atin
g
d
etailed
tech
n
ical
d
escr
ip
tio
n
s
o
f
d
atac
en
ter
ass
ets,
in
clu
d
in
g
s
er
v
er
s
,
m
em
o
r
y
u
n
its
,
p
r
o
ce
s
s
o
r
s
,
an
d
s
to
r
ag
e
d
ev
ices.
T
h
e
co
llected
tec
h
n
ical
d
escr
ip
tio
n
s
u
n
d
er
g
o
p
r
ep
r
o
ce
s
s
in
g
s
tep
s
s
u
ch
as
d
ata
clea
n
in
g
,
n
o
r
m
aliza
tio
n
,
an
d
f
ea
tu
r
e
e
x
tr
ac
t
io
n
.
Fo
llo
win
g
th
e
co
llectio
n
o
f
tech
n
ical
d
escr
ip
tio
n
s
,
th
e
v
alid
atio
n
p
r
o
ce
s
s
is
in
itiated
to
en
s
u
r
e
th
at
th
e
co
l
lecte
d
in
f
o
r
m
atio
n
m
ee
ts
th
e
s
tan
d
ar
d
s
.
I
n
th
is
p
r
o
ce
s
s
we
ad
o
p
ted
an
d
ap
p
lie
d
a
ca
p
ac
ity
v
alid
atio
n
(
C
V)
r
u
le
g
iv
en
i
n
(
1
)
t
o
ass
es
s
wh
eth
er
th
e
av
ailab
le
r
e
s
o
u
r
ce
s
ar
e
s
u
f
f
icien
t to
m
ee
t
o
p
er
atio
n
al
d
em
an
d
s
.
=
(
)
×
100
(
1
)
r
ep
r
esen
ts
th
e
t
o
tal
av
ailab
le
r
eso
u
r
ce
s
an
d
d
en
o
tes
th
e
n
e
ce
s
s
ar
y
r
eso
u
r
ce
s
.
B
y
u
tili
zin
g
p
r
ed
ef
in
e
d
th
r
esh
o
ld
s
,
th
e
C
V
en
h
an
ce
s
lo
ad
b
alan
cin
g
,
r
ed
u
ce
s
co
m
p
u
tatio
n
al
b
o
ttlen
ec
k
s
,
an
d
g
u
ar
an
tees
ef
f
ec
tiv
e
r
eso
u
r
ce
m
an
ag
em
en
t in
th
e
f
o
llo
win
g
p
h
ases
.
D
u
r
i
n
g
s
e
m
a
n
ti
c
e
n
r
ic
h
m
e
n
t
s
t
e
p
,
ea
c
h
d
a
t
a
ce
n
t
e
r
ass
e
t
is
d
es
c
r
i
b
e
d
u
s
i
n
g
t
u
p
l
e
o
f
n
o
r
m
a
l
i
z
e
d
a
t
t
r
i
b
u
t
es
i
n
c
l
u
d
i
n
g
C
PU
c
a
p
ac
i
t
y
,
m
e
m
o
r
y
,
s
t
o
r
a
g
e
,
a
n
d
b
an
d
w
i
d
t
h
.
T
h
e
s
e
v
a
l
u
es
a
r
e
f
i
r
s
t
n
o
r
m
a
l
i
ze
d
t
o
a
[
0
,
1
]
r
a
n
g
e
.
N
e
x
t
,
e
n
t
i
t
y
l
i
n
k
in
g
i
s
p
e
r
f
o
r
m
e
d
b
y
m
a
p
p
i
n
g
e
a
c
h
a
t
t
r
i
b
u
t
e
t
o
a
n
o
n
t
o
l
o
g
y
t
er
m
r
e
p
r
e
s
e
n
t
i
n
g
i
ts
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
26
:
1
1
8
1
-
1
1
9
3
1184
s
e
m
a
n
t
ic
c
a
te
g
o
r
y
.
T
h
i
s
o
n
t
o
lo
g
y
-
b
a
s
e
d
l
a
b
e
l
i
n
g
e
n
s
u
r
e
s
t
h
a
t
h
e
t
e
r
o
g
e
n
e
o
u
s
r
es
o
u
r
c
es
a
r
e
r
e
p
r
e
s
e
n
t
e
d
i
n
a
u
n
i
f
i
e
d
s
c
h
e
m
a
w
it
h
i
n
KG
.
B
y
e
m
b
e
d
d
i
n
g
n
o
d
e
s
w
it
h
b
o
t
h
s
tr
u
c
t
u
r
a
l
i
d
e
n
ti
f
i
e
r
s
a
n
d
s
e
m
a
n
t
ic
t
a
g
s
,
K
G
c
a
p
t
u
r
es
n
o
t
o
n
l
y
r
e
s
o
u
r
c
e
q
u
a
n
t
i
t
i
es
b
u
t
a
l
s
o
t
h
e
i
r
c
o
n
t
e
x
t
u
al
m
e
a
n
i
n
g
,
w
h
i
c
h
i
s
c
r
i
t
ic
a
l
f
o
r
s
u
b
s
e
q
u
e
n
t
i
n
f
e
r
e
n
c
e
.
2
.
2
.
T
ra
ini
ng
a
nd
v
a
lid
a
t
ing
t
he
pro
po
s
ed
m
o
del
I
n
th
is
p
h
ase,
u
n
s
tr
u
ctu
r
ed
d
es
cr
ip
tio
n
s
ar
e
tr
an
s
f
o
r
m
ed
in
to
s
em
an
tically
s
tr
u
ctu
r
ed
KG
th
at
m
o
d
els
th
e
d
atac
en
ter
ass
ets.
T
o
ass
ess
th
e
co
r
r
ec
tn
ess
o
f
th
e
in
f
er
r
e
d
r
elatio
n
s
with
in
th
e
KG
,
t
h
e
m
o
d
el
is
v
alid
ated
u
s
in
g
s
tan
d
ar
d
class
if
icatio
n
m
etr
ics
in
clu
d
in
g
ac
c
u
r
ac
y
,
p
r
ec
is
io
n
,
r
ec
all,
an
d
F1
-
s
co
r
e
[
2
5
]
.
Acc
u
r
ac
y
g
i
v
en
in
(
2
)
,
m
ea
s
u
r
es
th
e
r
atio
o
f
co
r
r
ec
tly
class
if
ied
in
s
tan
c
es,
p
r
ec
is
io
n
a
n
d
r
ec
all
g
iv
e
n
in
(
3
)
,
e
v
alu
ate
p
r
ed
ictio
n
c
o
r
r
ec
tn
ess
an
d
d
etec
tio
n
ca
p
ab
ilit
y
,
r
esp
ec
tiv
ely
.
T
h
e
F1
-
s
co
r
e
g
iv
e
n
in
(
4
)
is
ad
o
p
ted
as
a
b
alan
ce
d
m
etr
ic
t
o
jo
in
tly
ac
c
o
u
n
t f
o
r
f
alse p
o
s
itiv
es a
n
d
f
al
s
e
n
eg
ativ
es.
=
+
+
+
+
(
2
)
=
+
an
d
=
+
(
3
)
1
−
=
2
×
×
+
(
4
)
2
.
3
.
G
ener
a
t
io
n a
nd
v
a
lid
a
t
i
o
n o
f
t
he
k
no
wledg
e
g
ra
ph
T
h
e
v
is
u
aliza
tio
n
o
f
s
em
an
tically
s
tr
u
ctu
r
ed
elem
en
ts
th
at
r
ep
r
esen
t
th
e
d
atac
en
ter
ass
ets,
an
d
th
e
in
ter
co
n
n
ec
tio
n
s
an
d
in
ter
ac
t
io
n
s
b
etwe
en
th
ese
ass
ets
a
n
d
p
r
o
v
id
e
en
h
an
ce
d
s
em
an
tic
r
ea
s
o
n
in
g
an
d
d
ec
is
io
n
-
m
ak
in
g
is
s
h
o
wn
in
Fig
u
r
e
2
.
T
h
e
s
u
f
f
icien
c
y
c
o
n
d
itio
n
(
S)
g
iv
e
n
b
y
(
5
)
is
u
s
ed
as
a
cr
itical
v
alid
atio
n
p
ar
a
m
eter
th
at
r
e
v
e
als wh
eth
er
th
e
s
elec
ted
r
eso
u
r
ce
s
ar
e
ca
p
ab
le
f
o
r
co
n
tin
u
in
g
s
y
s
tem
o
p
er
atio
n
s
.
=
∑
(
−
)
=
1
(
5
)
an
d
d
en
o
te
t
h
e
av
ailab
le
an
d
r
eq
u
ir
ed
r
eso
u
r
ce
s
at
n
o
d
e
i
,
r
e
s
p
ec
tiv
ely
,
an
d
n
is
th
e
ab
s
o
lu
te
n
u
m
b
e
r
o
f
n
o
d
es.
I
f
<
0
,
m
ea
n
s
th
e
s
y
s
tem
h
as r
eso
u
r
ce
s
r
estrictio
n
s
an
d
r
e
q
u
ir
es th
e
in
teg
r
atio
n
o
f
ad
d
iti
o
n
al
n
o
d
es.
Fig
u
r
e
2
.
Kn
o
wled
g
e
g
r
ap
h
co
n
s
tr
u
ctio
n
2
.
4
.
H
idd
en
re
la
t
io
ns
ph
a
s
e:
v
irt
ua
l net
wo
rk
e
m
bedd
ing
in dy
na
m
ic
env
iro
nm
ent
s
A
VNR
is
r
ep
r
esen
ted
as
V
N
R
i
=(
GV
i
,
t
a
,
t
e
)
,
wh
er
e
t
a
an
d
t
e
d
en
o
te
th
e
ar
r
iv
al
an
d
d
ep
a
r
tu
r
e
tim
es,
r
esp
ec
tiv
ely
.
T
h
e
VNE
p
r
o
c
ess
m
ap
s
th
e
v
ir
tu
al
n
etwo
r
k
GV
i
(N
V
,
L
V
)
o
n
to
a
s
u
b
g
r
ap
h
o
f
th
e
p
h
y
s
ical
s
u
b
s
tr
ate
n
etwo
r
k
GP
′
(
N
′
P
,
L′
P
)
th
at
s
atis
f
ie
s
th
e
r
eso
u
r
ce
co
n
s
tr
ain
ts
.
Fig
u
r
e
3
illu
s
tr
ates
an
ex
am
p
le
o
f
v
ir
tu
al
n
o
d
es
an
d
lin
k
s
em
b
ed
d
ed
o
n
to
th
e
p
h
y
s
ical
in
f
r
astru
ctu
r
e
[
2
3
]
,
wh
er
e
Fig
u
r
e
3
(
a
)
s
h
o
ws
th
e
co
n
d
itio
n
b
ef
o
r
e
VNRs
em
b
ed
d
in
g
an
d
Fig
u
r
e
3
(
b
)
s
h
o
ws th
e
co
n
d
itio
n
af
ter
VNRs
em
b
ed
d
in
g
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
K
n
o
w
led
g
e
g
r
a
p
h
-
b
a
s
ed
en
h
a
n
ce
d
virt
u
a
l n
etw
o
r
k
emb
ed
d
i
n
g
fo
r
6
G
clo
u
d
…
(
S
h
o
u
r
o
k
A
b
d
elra
h
im
)
1185
(
a)
(
b
)
Fig
u
r
e
3
.
Diag
r
a
m
o
f
th
e
p
h
y
s
ical
an
d
v
ir
tu
al
n
etwo
r
k
s
tr
u
ct
u
r
es: (
a)
b
ef
o
r
e
VNRs
em
b
ed
d
in
g
an
d
(
b
)
af
ter
VNRs
em
b
ed
d
in
g
I
n
p
r
ac
tical
6
G
d
atac
en
ter
en
v
ir
o
n
m
e
n
ts
,
v
ir
tu
al
n
etwo
r
k
s
ch
an
g
e
o
v
e
r
tim
e
d
u
e
to
d
y
n
am
ic
u
s
er
d
em
an
d
s
,
to
p
o
lo
g
y
r
ec
o
n
f
ig
u
r
atio
n
,
an
d
v
ar
y
in
g
r
eso
u
r
ce
r
eq
u
ir
em
en
ts
.
T
h
is
d
ef
in
es
th
e
d
y
n
am
ic
VNE
p
r
o
b
lem
,
wh
e
r
e
em
b
ed
d
in
g
d
ec
is
io
n
s
m
u
s
t
co
n
tin
u
o
u
s
ly
ad
ap
t
to
ch
an
g
es
in
v
ir
t
u
al
to
p
o
lo
g
y
an
d
r
eso
u
r
c
e
d
em
an
d
.
Fig
u
r
e
4
s
h
o
ws
r
ep
r
esen
tativ
e
s
tr
u
ctu
r
al
an
d
r
es
o
u
r
ce
c
h
an
g
es
ac
r
o
s
s
s
u
cc
ess
iv
e
tim
e
in
ter
v
als.
T
o
s
u
p
p
o
r
t
th
is
d
y
n
am
ic
b
eh
av
io
r
,
th
e
p
r
o
p
o
s
ed
f
r
am
ewo
r
k
in
teg
r
ates
h
id
d
e
n
-
r
elatio
n
in
f
er
en
ce
with
in
a
tim
e
-
ev
o
lv
in
g
KG
,
en
ab
lin
g
ad
ap
tiv
e
em
b
e
d
d
in
g
d
ec
is
io
n
s
with
o
u
t
r
ep
r
o
ce
s
s
in
g
th
e
e
n
tire
to
p
o
l
o
g
y
.
T
o
h
an
d
le
t
h
is
d
y
n
am
ic
b
eh
av
i
o
r
,
we
p
r
o
p
o
s
e
a
VNE
f
r
a
m
ewo
r
k
b
ased
o
n
h
id
d
en
r
elatio
n
in
f
er
en
ce
with
in
a
tim
e
-
ev
o
lv
in
g
KG.
L
aten
t
r
el
atio
n
s
h
ip
s
am
o
n
g
n
o
d
es
an
d
lin
k
s
as
s
h
o
wn
in
Fig
u
r
e
5
ar
e
in
f
er
r
e
d
u
s
in
g
s
tatis
t
ical
m
o
d
elin
g
an
d
lin
k
p
r
ed
ictio
n
,
allo
win
g
f
o
r
r
ea
l
-
tim
e
ad
ap
tatio
n
to
to
p
o
lo
g
y
m
o
d
if
icatio
n
s
an
d
f
lu
ctu
atin
g
r
eso
u
r
ce
d
e
m
an
d
s
.
I
n
r
ea
l
-
tim
e
o
p
e
r
atio
n
,
u
p
d
at
es
to
th
e
v
ir
tu
al
n
etwo
r
k
ar
e
m
an
ag
ed
th
r
o
u
g
h
a
s
lid
in
g
o
b
s
er
v
atio
n
win
d
o
w
T
.
At
th
e
b
eg
in
n
in
g
o
f
ea
ch
win
d
o
w,
n
ew
ev
en
ts
ar
e
d
etec
ted
,
an
d
th
e
KG
is
u
p
d
ated
ac
co
r
d
in
g
ly
.
E
ac
h
m
o
d
if
icatio
n
tr
ig
g
e
r
s
r
e
-
in
f
er
en
ce
f
o
r
o
n
ly
th
e
c
h
an
g
ed
s
u
b
g
r
ap
h
,
r
ath
e
r
th
a
n
r
e
-
p
r
o
ce
s
s
in
g
th
e
e
n
tire
to
p
o
lo
g
y
,
wh
ic
h
r
ed
u
ce
s
o
v
e
r
h
ea
d
.
E
lig
ib
ilit
y
o
f
in
f
er
r
ed
lin
k
s
is
th
en
v
alid
ated
ag
ain
s
t
th
e
p
r
e
d
ef
in
e
d
th
r
esh
o
ld
s
γ
(
s
em
an
tic
af
f
in
ity
)
an
d
t
hr
es
ho
l
d
(
r
eso
u
r
ce
f
ea
s
ib
ilit
y
)
.
I
f
b
o
th
co
n
d
itio
n
s
ar
e
s
atis
f
ied
,
th
e
r
elatio
n
is
ad
d
ed
to
th
e
ca
n
d
id
ate
em
b
ed
d
in
g
s
et;
o
th
er
wis
e,
it
i
s
d
is
ca
r
d
ed
.
T
h
is
in
cr
em
en
tal
u
p
d
ate
m
ec
h
an
is
m
en
ab
les
th
e
s
y
s
te
m
to
ad
ap
t
q
u
ick
l
y
to
to
p
o
lo
g
y
an
d
d
em
a
n
d
v
ar
iatio
n
s
with
o
u
t
r
estar
tin
g
t
h
e
em
b
ed
d
in
g
p
r
o
ce
s
s
.
Fig
u
r
e
4
.
E
x
am
p
le
illu
s
tr
atin
g
th
e
o
p
er
atio
n
o
f
d
y
n
am
ic
VNE
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
26
:
1
1
8
1
-
1
1
9
3
1186
Fig
u
r
e
5
.
Hid
d
en
r
elatio
n
in
k
n
o
wled
g
e
g
r
ap
h
2
.
4
.
1
.
K
no
wledg
e
-
ba
s
ed
la
t
e
nt
re
la
t
io
ns
m
o
delin
g
T
h
e
p
r
o
p
o
s
ed
s
y
s
tem
in
tr
o
d
u
ce
s
an
in
f
er
en
ce
-
d
r
iv
e
n
f
r
a
m
ewo
r
k
t
h
at
lev
er
a
g
es
a
s
e
m
an
tically
s
tr
u
ctu
r
ed
KG
to
id
en
tify
lat
en
t
n
o
d
e
d
ep
e
n
d
en
cies
a
n
d
e
n
ab
les
d
y
n
am
ic
r
eso
u
r
ce
-
awa
r
e
d
ec
is
io
n
-
m
ak
in
g
th
r
o
u
g
h
a
b
in
a
r
y
in
d
icato
r
f
u
n
ctio
n
∈
,
as
d
ef
in
ed
in
(
6
)
,
th
a
t
q
u
an
tifie
s
th
e
ex
is
ten
ce
o
f
i
m
p
licit
r
elatio
n
s
b
etwe
en
n
o
d
es.
∈
=
{
1
,
(
,
)
=
1
0
,
othe
r
w
ise
(
6
)
T
h
e
s
u
b
s
eq
u
en
t
m
ath
em
atic
al
ex
p
r
ess
io
n
s
in
(
7
)
to
(
1
4
)
s
y
s
tem
atica
lly
o
u
tlin
e
t
h
e
co
r
e
elig
i
b
ilit
y
r
eq
u
ir
em
e
n
ts
an
d
o
p
tim
izatio
n
cr
iter
ia
th
at
u
n
d
er
p
in
th
e
p
r
o
p
o
s
ed
in
f
er
en
ce
m
ec
h
an
is
m
.
∅
(
,
)
=
(
(
)
(
)
)
(
7
)
(
,
)
=
{
1
,
if
∅
(
,
)
≥
a
n
d
pd
(
,
)
≤
t
hr
es
ho
l
d
0
,
othe
r
w
ise
(
8
)
HC
C
(
)
=
{
1
,
∃
∈
{
}
s
uc
h
tha
t
(
,
)
=
1
0
,
othe
r
wi
s
e
(
9
)
(
)
=
∑
(
(
,
)
=
1
∧
∅
(
,
)
≥
)
∈
\
{
}
(
1
0
)
(
)
=
∑
(
(
,
)
=
0
∧
pd
(
,
)
≤
t
hr
es
ho
l
d
)
∈
\
{
}
(
1
1
)
(
)
+
(
)
=
|
|
−
1
⇒
∈
(
1
2
)
Ω
=
∑
∑
(
,
)
∈
\
{
}
∈
(
1
3
)
M
a
ximi
ze
Ω
s
ubj
e
c
t
to
HC
C
(
)
=
1
,
∀
∈
(
1
4
)
I
n
(
7
)
s
h
o
ws
th
e
s
em
an
tic
a
f
f
i
n
ity
b
etwe
en
an
y
p
air
o
f
n
o
d
e
s
is
q
u
an
tifie
d
u
s
in
g
a
s
im
ilar
ity
f
u
n
ctio
n
∅
(
,
)
.
T
h
e
o
u
tp
u
t
∅
(
,
)
∈
[
0
,
1
]
d
en
o
tes
th
e
p
r
o
b
ab
ilis
tic
co
n
f
id
en
ce
o
f
a
late
n
t
s
em
an
tic
lin
k
ag
e
.
I
n
(
8
)
in
d
icate
s
th
e
leg
itima
cy
o
f
s
u
ch
in
f
er
r
e
d
r
elatio
n
s
.
T
o
r
e
g
u
late
p
ar
ticip
atio
n
in
th
e
em
b
e
d
d
in
g
p
r
o
ce
s
s
,
we
d
ef
in
e
a
co
n
n
ec
tiv
ity
c
o
n
s
tr
ai
n
t
HC
C
(
)
in
(
9
)
th
at
ev
alu
ates
th
e
ex
is
ten
ce
o
f
at
least
o
n
e
v
alid
i
n
f
er
r
ed
lin
k
f
o
r
a
n
o
d
e
.
I
n
(
1
0
)
d
en
o
tes
t
h
e
s
tr
o
n
g
r
elatio
n
al
in
ten
s
ity
o
f
a
n
o
d
e
as
(
)
,
r
ep
r
esen
tin
g
th
e
ca
r
d
in
ality
o
f
s
em
an
tically
r
o
b
u
s
t
in
f
er
e
n
ce
s
it
h
o
ld
s
with
o
th
er
s
in
t
h
e
g
r
ap
h
.
C
o
n
v
er
s
ely
,
th
e
wea
k
in
f
er
en
ce
d
eg
r
ee
(
)
,
is
d
en
o
ted
i
n
(
1
1
)
en
u
m
er
at
es
to
p
o
lo
g
ically
v
alid
y
et
s
em
an
tically
m
ar
g
in
al
c
o
n
n
ec
ti
o
n
s
.
A
n
o
d
e
u
q
u
alif
ies
f
o
r
in
clu
s
io
n
i
n
th
e
f
in
al
em
b
ed
d
in
g
s
et
f
th
e
a
g
g
r
eg
ate
o
f
its
s
tr
o
n
g
an
d
wea
k
in
f
er
en
ce
s
s
p
an
s
th
e
en
tire
ty
o
f
t
h
e
g
r
a
p
h
m
in
u
s
its
elf
,
as
s
tated
in
(
1
2
)
.
I
n
(
1
3
)
s
h
o
ws
th
e
g
lo
b
al
in
f
e
r
en
ti
al
co
h
er
en
ce
o
f
th
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
K
n
o
w
led
g
e
g
r
a
p
h
-
b
a
s
ed
en
h
a
n
ce
d
virt
u
a
l n
etw
o
r
k
emb
ed
d
i
n
g
fo
r
6
G
clo
u
d
…
(
S
h
o
u
r
o
k
A
b
d
elra
h
im
)
1187
s
u
b
g
r
ap
h
;
we
d
ef
in
e
th
e
in
f
er
en
ce
p
o
wer
m
etr
ic
Ω
as
th
e
s
u
m
m
atio
n
o
f
all
co
n
f
ir
m
e
d
in
f
er
en
ce
s
ac
r
o
s
s
th
e
s
elec
ted
n
o
d
e
s
et.
I
n
(
1
4
)
in
d
i
ca
te
th
e
o
v
er
ar
c
h
in
g
o
p
tim
izatio
n
task
aim
s
to
m
ax
im
ize
th
e
in
f
er
en
ce
u
tili
ty
Ω
co
n
s
tr
ain
ed
b
y
th
e
co
n
n
ec
tiv
i
ty
co
n
d
itio
n
HC
C
(
)
=
1
f
o
r
ev
er
y
ca
n
d
i
d
ate
n
o
d
e
∈
.
T
h
e
p
r
o
ce
d
u
r
al
s
tep
s
ar
e
o
u
tlin
ed
in
Alg
o
r
ith
m
1
,
an
d
its
f
lo
wch
ar
t r
e
p
r
ese
n
tatio
n
is
illu
s
tr
ated
in
Fig
u
r
e
6
.
Alg
o
r
ith
m
1
.
I
n
f
e
r
en
ce
e
n
g
in
e
f
o
r
h
i
d
d
en
p
ath
s
p
r
e
d
ictio
n
I
n
p
u
t: G(
V,
E
)
,
r
eso
u
r
ce
r
eq
u
i
r
em
en
ts
R
,
o
b
s
er
v
atio
n
win
d
o
w
T
1
: I
n
itialize
VS ←
∅
; n
←
|
V
|
;
C
←
n
; m
p
d
(
G)
←
n
–
1
2
: f
o
r
ea
c
h
tim
e
s
tep
t
∈
T
d
o
3
:
Mo
n
ito
r
VNR ev
en
ts
(
ar
r
iv
al,
u
p
d
ate,
d
eletio
n
)
4
:
Up
d
ate
G
b
y
m
o
d
if
y
in
g
a
f
f
ec
ted
n
o
d
es/ed
g
es
5
:
Ma
r
k
ch
an
g
ed
n
o
d
es f
o
r
r
e
-
in
f
er
e
n
ce
6
:
wh
ile
∃
u
n
p
r
o
ce
s
s
ed
u
∈
V
d
o
7
:
if
m
p
d
(
G)
<
R
_
th
r
esh
o
ld
th
en
8
:
Per
f
o
r
m
c
o
m
m
u
n
ity
d
etec
tio
n
C
D(
G)
9
:
Select
u
wh
er
e
s
tate(
u
)
=
0
∧
p
d
(
u
)
≥
R
_
th
r
esh
o
ld
1
0
:
wh
ile
s
(
u
)
+
e(
u
)
=
n
−
1
d
o
1
1
:
VS ←
VS
∪
{u
}
1
2
:
u
←
n
ex
t
v
alid
u
n
p
r
o
ce
s
s
ed
v
er
tex
1
3
:
en
d
wh
ile
1
4
:
m
p
d
(
G)
←
0
1
5
:
f
o
r
ea
ch
u
∈
V
d
o
1
6
:
Select
v
wh
er
e
p
d
(
u
,
v
)
≥
R
_
th
r
esh
o
ld
1
7
:
Pr
o
b
e
c
o
n
n
ec
tio
n
s
tr
en
g
th
f
(
u
,
v
)
1
8
:
i
f
f
(
u
,
v
)
=
∅
th
en
1
9
:
e(
u
)
++
,
e
(
v
)
++
; s
tate(
u
)
--
,
s
tate(
v
)
--
2
0
:
else
2
1
:
s
(
u
)
++
,
s
(
v
)
++
2
2
:
i
f
HC
C
(
u
)
=
0
o
r
H
C
C
(
v
)
=
0
th
en
2
3
:
r
et
u
r
n
∅
2
4
:
i
f
s
(
u
)
+
e
(
u
)
=
n
−
1
th
en
VS ←
VS
∪
{u
}
2
5
:
i
f
s
(
v
)
+
e
(
v
)
=
n
−
1
th
en
VS ←
VS
∪
{v
}
2
6
:
en
d
f
o
r
2
7
:
e
n
d
wh
ile
2
8
: e
n
d
f
o
r
2
9
: r
etu
r
n
VS w
ith
ass
o
ciate
d
in
f
er
en
ce
m
etr
ics
T
h
e
co
m
p
u
tatio
n
al
co
m
p
lex
it
y
o
f
th
e
alg
o
r
ith
m
ca
n
b
e
a
n
aly
ze
d
as
f
o
llo
ws:
i)
m
o
n
ito
r
in
g
an
d
u
p
d
atin
g
t
h
e
KG
ac
r
o
s
s
|
|
n
o
d
es
r
eq
u
ir
es
(
|
|
)
;
ii)
co
m
m
u
n
i
ty
d
etec
tio
n
(
lin
e
8
)
is
b
o
u
n
d
ed
b
y
(
|
|
|
|
)
;
an
d
iii)
af
f
in
ity
s
co
r
in
g
an
d
v
alid
atio
n
(
lin
es
1
6
–
2
5
)
r
e
q
u
ir
e
(
|
|
²
)
in
th
e
wo
r
s
t
ca
s
e.
Ov
er
all,
th
e
to
tal
r
u
n
tim
e
p
e
r
o
b
s
er
v
atio
n
is
ap
p
r
o
x
im
at
ely
(
|
|
²
+
|
|
|
|
)
.
W
h
ich
is
tr
ac
tab
le
f
o
r
m
e
d
iu
m
-
to
lar
g
e
-
s
ca
le
to
p
o
lo
g
ies wh
en
c
o
m
p
a
r
ed
with
DR
L
tr
ain
in
g
lo
o
p
s
.
A
cr
itical
ch
allen
g
e
in
d
y
n
a
m
ic
en
v
ir
o
n
m
e
n
ts
is
co
n
ce
p
t
d
r
if
t,
wh
er
e
th
e
s
tatis
tical
p
r
o
p
er
ties
o
f
VNRs
ev
o
lv
e
o
v
er
tim
e.
T
o
a
d
d
r
ess
th
is
,
th
e
f
r
am
ewo
r
k
co
n
tin
u
o
u
s
ly
r
e
-
v
alid
ates
s
em
an
tic
s
im
ilar
ity
v
alu
e
s
∅
(
,
)
(
6
)
an
d
th
r
esh
o
l
d
s
(
7
)
at
ea
ch
o
b
s
er
v
atio
n
win
d
o
w.
I
f
p
r
ev
i
o
u
s
ly
in
f
er
r
e
d
r
elatio
n
s
n
o
lo
n
g
er
m
ee
t
th
e
s
em
an
tic
o
r
r
eso
u
r
ce
th
r
esh
o
l
d
s
,
th
ey
ar
e
p
r
u
n
ed
f
r
o
m
th
e
KG,
en
s
u
r
in
g
th
at
e
m
b
ed
d
in
g
r
em
ain
s
co
n
s
is
ten
t
with
cu
r
r
en
t
n
etwo
r
k
c
o
n
d
itio
n
s
.
2
.
5
.
O
pti
m
iza
t
io
n
a
nd
decisi
o
n
-
m
a
k
ing
T
h
is
p
h
ase
ev
alu
ates
th
e
ef
f
ec
tiv
en
ess
o
f
th
e
p
r
o
p
o
s
ed
in
f
er
en
ce
-
d
r
iv
e
n
VNE
f
r
am
ewo
r
k
u
s
in
g
r
u
n
tim
e
ef
f
icien
cy
,
r
eso
u
r
ce
u
tili
za
tio
n
,
lo
ad
b
alan
cin
g
,
an
d
f
au
lt
r
ec
o
v
er
y
m
etr
ics.
T
h
ese
in
d
icato
r
s
q
u
a
n
tify
th
e
im
p
ac
t o
f
h
id
d
e
n
r
elatio
n
i
n
f
er
en
ce
o
n
em
b
ed
d
i
n
g
p
e
r
f
o
r
m
an
ce
u
n
d
er
d
y
n
am
ic
c
o
n
s
tr
ain
ts
.
i)
R
u
n
tim
e
ac
ce
ler
atio
n
:
r
u
n
tim
e
is
m
ea
s
u
r
ed
as
th
e
tim
e
r
eq
u
ir
ed
to
em
b
ed
a
VNR
in
th
e
s
u
b
s
tr
ate
n
etwo
r
k
,
as g
iv
e
n
in
(
1
5
)
.
=
(
1
5
)
wh
er
e
is
th
e
tim
e
r
eq
u
ir
ed
to
g
en
er
ate
an
d
co
r
r
ec
tly
em
b
e
d
a
VNR in
th
e
s
u
b
s
tr
ate
n
etwo
r
k
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
26
:
1
1
8
1
-
1
1
9
3
1188
ii)
R
eso
u
r
ce
u
tili
za
tio
n
ef
f
icien
c
y
:
th
is
m
etr
ic
is
in
ten
d
e
d
to
ass
es
s
h
o
w
well
th
e
in
f
er
e
n
c
e
-
b
ased
laten
t
r
elatio
n
d
etec
tio
n
alg
o
r
ith
m
(
Alg
o
r
ith
m
1
)
f
in
d
s
h
id
d
e
n
lin
k
s
b
etwe
en
n
etwo
r
k
item
s
.
R
eso
u
r
ce
u
tili
za
tio
n
r
ef
lects th
e
ef
f
ec
tiv
en
ess
o
f
in
f
er
r
e
d
r
elatio
n
s
in
g
u
id
in
g
n
o
d
e
a
n
d
p
ath
s
elec
tio
n
.
iii)
Dec
is
io
n
v
alid
atio
n
:
th
e
lo
ad
b
alan
ce
in
d
ex
(
L
B
I
)
as
in
(
1
6
)
an
d
f
au
lt
r
ec
o
v
e
r
y
ef
f
icien
c
y
as
in
(
1
7
)
ar
e
u
s
ed
to
ev
alu
ate
wo
r
k
lo
a
d
d
is
tr
ib
u
tio
n
a
n
d
s
y
s
tem
r
esil
ien
ce
.
=
1
−
(
1
6
)
=
+
(
1
7
)
iv
)
Sy
s
tem
ev
alu
atio
n
p
h
ase:
in
(
1
8
)
is
u
s
ed
t
o
ev
alu
ate
th
e
o
v
er
all
ef
f
ec
tiv
e
n
ess
an
d
s
ca
lab
ilit
y
o
f
th
e
p
r
o
p
o
s
ed
s
y
s
tem
.
A
h
ig
h
er
ac
ce
p
tan
ce
r
atio
m
ea
n
s
b
etter
lo
ad
d
is
tr
ib
u
tio
n
,
b
etter
e
m
b
ed
d
in
g
,
an
d
in
cr
ea
s
ed
f
lex
ib
ilit
y
in
r
esp
o
n
s
e
to
ch
an
g
in
g
n
etwo
r
k
d
em
a
n
d
s
.
=
l
im
→
∞
∑
=
0
∑
=
0
(
1
8
)
Fig
u
r
e
6
.
Flo
wch
ar
t
o
f
Alg
o
r
ith
m
1
illu
s
tr
atin
g
th
e
m
ain
p
r
o
ce
d
u
r
al
s
tep
s
3.
SI
M
UL
A
T
I
O
N
SE
T
UP
Simu
latio
n
s
wer
e
co
n
d
u
cte
d
u
s
in
g
Min
in
et
[
2
6
]
an
d
ONOS
[
2
7
]
u
n
d
er
v
ar
y
in
g
s
u
b
s
tr
ate
s
izes,
r
eso
u
r
ce
ca
p
ac
ities
,
an
d
VN
R
ar
r
iv
al
p
atter
n
s
to
ev
alu
at
e
s
ca
lab
ilit
y
an
d
p
e
r
f
o
r
m
an
c
e.
T
h
e
s
im
u
latio
n
ex
p
er
im
en
ts
h
a
v
e
b
e
en
c
o
n
d
u
cted
u
s
in
g
v
ar
iety
o
f
C
PUs
r
an
g
in
g
f
r
o
m
5
to
5
0
with
u
p
to
1
0
0
p
r
o
ce
s
s
in
g
n
o
d
es,
an
d
5
0
0
co
m
m
u
n
icatio
n
lin
k
s
with
b
an
d
wid
th
r
a
n
g
in
g
f
r
o
m
1
0
to
1
0
0
M
b
p
s
.
T
h
e
VNRs
v
ar
ied
f
r
o
m
2
to
5
0
v
ir
t
u
al
n
o
d
es,
an
d
d
if
f
er
en
t
ar
r
iv
al
r
ate,
t
o
test
s
ca
l
ab
ilit
y
an
d
th
e
alg
o
r
ith
m
’
s
ef
f
icien
cy
in
ter
m
s
o
f
r
u
n
tim
e
an
d
r
eso
u
r
ce
u
tili
za
tio
n
in
6
G
d
atac
e
n
ter
s
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J Ar
tif
I
n
tell
I
SS
N:
2252
-
8
9
3
8
K
n
o
w
led
g
e
g
r
a
p
h
-
b
a
s
ed
en
h
a
n
ce
d
virt
u
a
l n
etw
o
r
k
emb
ed
d
i
n
g
fo
r
6
G
clo
u
d
…
(
S
h
o
u
r
o
k
A
b
d
elra
h
im
)
1189
3
.
1
.
Da
t
a
s
et
s
chem
a
a
nd
prepro
ce
s
s
ing
Data
ce
n
ter
ass
ets
wer
e
r
ep
r
esen
ted
as
a
s
em
an
tically
en
r
ich
ed
KG
with
n
o
r
m
alize
d
n
o
d
e
an
d
lin
k
attr
ib
u
tes.
T
h
e
p
r
ep
r
o
ce
s
s
in
g
in
clu
d
es
:
i
)
t
y
p
e
h
ar
m
o
n
iza
tio
n
(
SI
u
n
its
)
,
ii
)
m
in
–
m
ax
n
o
r
m
aliza
tio
n
o
f
n
u
m
er
ical
attr
ib
u
tes
to
[
0
,
1
]
,
iii
)
d
ed
u
p
licatio
n
o
f
ex
ac
t/n
ea
r
-
d
u
p
licate
ass
ets,
iv
)
I
D
ca
n
o
n
icaliza
tio
n
f
o
r
en
tity
lin
k
in
g
,
v
)
o
u
tlier
clip
p
in
g
at
1
s
t/9
9
th
p
e
r
ce
n
tiles
,
an
d
vi
)
o
n
to
lo
g
y
tag
g
in
g
(
e.
g
.
,
C
o
m
p
u
te.
C
PU,
Netwo
r
k
.
B
an
d
wid
th
)
f
o
r
s
em
a
n
tic
en
r
ich
m
en
t
.
T
h
is
o
n
to
l
o
g
y
tag
g
in
g
was
u
s
ed
in
s
im
ilar
ity
f
u
n
ctio
n
(
,
)
.
3
.
2
.
H
y
perpa
ra
m
e
t
er
t
a
ble
All
r
elev
an
t
ex
p
er
im
en
tal
p
a
r
am
eter
s
an
d
s
y
m
b
o
ls
u
s
ed
in
th
is
s
tu
d
y
ar
e
s
u
m
m
ar
is
ed
in
T
ab
le
2
.
T
h
ese
in
clu
d
e
t
h
e
,
t
hr
es
ho
l
d
,
,
s
ee
d
s
K
,
v
ir
tu
al
n
etwo
r
k
s
ize,
s
u
b
s
tr
at
e
s
ize,
C
PU
u
n
its
,
b
an
d
wid
th
lim
its
,
co
m
m
u
n
ity
d
etec
tio
n
,
an
d
c
o
n
v
er
g
en
ce
g
u
a
r
d
.
T
h
ese
p
ar
am
eter
s
d
eter
m
in
e
th
e
co
n
tr
o
l
b
eh
av
io
r
o
f
th
e
alg
o
r
ith
m
(
e
.
g
.
ea
r
l
y
ter
m
in
atio
n
an
d
t
hr
es
h
o
l
d
)
as
well
as
th
e
s
im
u
latio
n
en
v
ir
o
n
m
en
t
a
n
d
r
eso
u
r
ce
s
co
n
s
tr
ain
ts
(
e.
g
.
C
PU,
b
an
d
wi
d
th
,
an
d
v
ir
tu
al
n
etwo
r
k
s
ize
o
f
th
e
n
o
d
es).
I
n
o
r
d
er
t
o
en
s
u
r
e
a
f
air
co
m
p
a
r
is
o
n
o
f
th
e
tr
ials
an
d
t
h
e
r
ep
r
o
d
u
cib
ilit
y
o
f
th
e
r
esu
lts
,
th
e
tr
ials
s
h
all
b
e
s
et
u
p
an
d
clea
r
ly
r
ep
o
r
ted
.
T
ab
le
2
.
Simu
latio
n
p
ar
am
eter
s
an
d
h
y
p
er
p
ar
a
m
eter
s
ettin
g
s
S
y
mb
o
l
/
n
a
me
M
e
a
n
i
n
g
V
a
l
u
e
/
r
a
n
g
e
S
e
ma
n
t
i
c
a
f
f
i
n
i
t
y
t
h
r
e
sh
o
l
d
i
n
(
6
a
n
d
7)
0
.
6
5
(
d
e
f
a
u
l
t
)
,
sw
e
p
t
0
.
5
–
0
.
8
th
r
e
s
h
o
l
d
R
e
s
o
u
r
c
e
/
f
e
a
s
i
b
i
l
i
t
y
c
u
t
o
f
f
(
p
a
t
h
/
c
a
p
a
c
i
t
y
)
0
.
6
o
f
r
e
s
i
d
u
a
l
c
a
p
a
c
i
t
y
(
l
i
n
k
+
n
o
d
e
)
O
b
serv
a
t
i
o
n
w
i
n
d
o
w
(
r
e
-
i
n
f
e
r
e
n
c
e
p
e
r
i
o
d
)
50
-
t
i
m
e
u
n
i
t
s
S
e
e
d
s
I
n
d
e
p
e
n
d
e
n
t
r
u
n
s
p
e
r
sce
n
a
r
i
o
10
V
i
r
t
u
a
l
n
e
t
w
o
r
k
s
i
z
e
V
N
R
v
i
r
t
u
a
l
n
o
d
e
s
p
e
r
r
e
q
u
e
st
2
–
3
2
(
a
l
so
t
e
st
e
d
2
–
5
0
i
n
s
t
r
e
ss)
S
u
b
s
t
r
a
t
e
s
i
z
e
P
h
y
s
i
c
a
l
n
o
d
e
s/
l
i
n
k
s
u
p
t
o
1
0
0
n
o
d
e
s/
5
0
0
l
i
n
k
s
C
P
U
u
n
i
t
s
P
e
r
n
o
d
e
c
a
p
a
c
i
t
y
(
n
o
r
mal
i
z
e
d
)
5
–
5
0
u
n
i
t
s
B
a
n
d
w
i
d
t
h
l
i
mi
t
s
P
e
r
l
i
n
k
c
a
p
a
c
i
t
y
10
–
1
0
0
M
b
p
s
C
o
mm
u
n
i
t
y
d
e
t
e
c
t
i
o
n
Tr
i
g
g
e
r
& me
t
h
o
d
W
h
e
n
a
v
g
.
(
)
<
th
r
e
s
h
o
l
d
Ea
r
l
y
-
st
o
p
C
o
n
v
e
r
g
e
n
c
e
g
u
a
r
d
N
o
c
h
a
n
g
e
i
n
a
c
c
e
p
t
a
n
c
e
f
o
r
3
w
i
n
d
o
w
s
4.
RE
SU
L
T
S AN
D
D
I
SCU
SS
I
O
N
T
h
e
p
er
f
o
r
m
an
ce
o
f
th
e
VN
E
f
r
am
ewo
r
k
was
ev
alu
ated
th
r
o
u
g
h
a
r
an
g
e
o
f
s
ce
n
ar
i
o
s
.
I
t
wa
s
ass
es
s
ed
u
s
in
g
k
ey
p
er
f
o
r
m
a
n
ce
m
etr
ics,
i
n
clu
d
in
g
r
u
n
ti
m
e,
n
o
d
es
an
d
lin
k
s
u
tili
za
tio
n
,
l
o
ad
b
alan
cin
g
,
en
er
g
y
ef
f
icien
c
y
,
a
n
d
ac
ce
p
t
an
ce
r
atio
.
T
o
v
alid
ate
th
e
r
esu
lts
,
o
u
r
r
esu
lts
co
m
p
a
r
ed
with
th
e
r
esu
lts
o
f
s
tate
-
of
-
th
e
-
ar
t a
lg
o
r
ith
m
s
,
in
c
lu
d
in
g
GC
N+
R
L
[
2
3
]
,
Gr
a
p
h
ViNE
[
2
4
]
,
A3
C
+G
C
N
[
2
1
]
,
a
n
d
Dee
p
ViNE
[
2
2
]
.
4
.
1
.
Runtim
e
a
cc
eler
a
t
io
n
T
ab
le
3
r
ep
o
r
ts
r
u
n
tim
e
r
esu
lts
u
n
d
er
v
ar
y
in
g
c
o
n
d
itio
n
s
.
Un
less
o
th
er
wis
e
s
tated
,
all
r
ep
o
r
ted
r
esu
lts
r
ep
r
esen
t
th
e
av
er
a
g
e
o
f
1
0
r
u
n
s
,
with
s
tan
d
ar
d
d
e
v
iatio
n
in
d
icate
d
.
Sig
n
if
ican
c
e
was
test
ed
u
s
in
g
p
air
ed
t
-
test
s
at
a
9
5
%
co
n
f
i
d
en
ce
lev
el
(
p
<0
.
0
5
)
.
C
o
m
p
ar
ed
to
A3
C
+G
C
N
[
2
1
]
wh
er
e
th
ey
p
r
esen
ted
a
r
u
n
tim
e
o
f
0
.
2
1
9
±
0
.
0
0
6
s
ec
o
n
d
s
at
an
ar
r
iv
al
r
ate
o
f
4
to
2
0
VNRs
p
er
1
0
0
-
tim
e
u
n
its
at
r
eso
u
r
ce
r
eq
u
est
r
ate
o
f
0
.
2
2
7
±
0
.
0
0
7
s
ec
o
n
d
s
,
an
d
0
.
4
7
6
±
0
.
0
0
8
s
ec
o
n
d
s
,
at
n
o
d
e
s
ize
ex
p
an
s
io
n
f
r
o
m
2
to
3
2
n
o
d
es,
th
e
p
r
o
p
o
s
ed
s
o
lu
tio
n
s
h
o
wed
a
co
n
s
is
ten
t
i
m
p
r
o
v
e
m
en
ts
in
th
e
s
ca
lab
ilit
y
an
d
r
u
n
tim
e
s
tab
ilit
y
u
n
d
er
d
if
f
er
en
t
d
atac
en
ter
n
etwo
r
k
co
m
p
lex
ities
.
T
ab
le
3
.
T
h
e
av
er
a
g
e
r
u
n
tim
e
in
a
s
ec
o
n
d
Te
st
t
y
p
e
Te
st
d
e
t
a
i
l
s
A
v
e
r
a
g
e
r
u
n
t
i
me
(
s
e
c
o
n
d
s)
A
r
r
i
v
a
l
r
a
t
e
V
a
r
y
i
n
g
a
r
r
i
v
a
l
r
a
t
e
s
f
r
o
m
4
t
o
2
0
V
N
R
s
p
e
r
1
0
0
-
t
i
me
u
n
i
t
s
0
.
6
5
0
±
0
.
0
1
4
R
e
s
o
u
r
c
e
r
e
q
u
e
s
t
V
a
r
y
i
n
g
C
P
U
a
n
d
b
a
n
d
w
i
d
t
h
d
e
ma
n
d
f
r
o
m 0
t
o
3
0
u
n
i
t
s
0
.
7
5
0
±
0
.
0
2
1
N
o
d
e
si
z
e
e
x
p
a
n
s
i
o
n
I
n
c
r
e
a
si
n
g
v
i
r
t
u
a
l
n
o
d
e
si
z
e
f
r
o
m
2
t
o
3
2
n
o
d
e
s
0
.
8
5
0
±
0
.
0
1
7
4
.
2
.
Reso
urce
utiliza
t
io
n
Fig
u
r
e
7
illu
s
tr
ates
th
e
r
eso
u
r
ce
u
tili
za
tio
n
ac
h
iev
ed
b
y
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
.
T
h
ese
r
e
s
u
lts
h
av
e
b
ee
n
co
m
p
a
r
ed
with
th
e
r
esu
l
ts
o
f
lately
p
u
b
lis
h
ed
r
esear
ch
s
u
ch
as
in
[
2
1
]
,
[
2
2
]
,
[
2
4
]
,
wh
er
e
th
ey
ac
h
ie
v
ed
0
.
7
9
6
±
0
.
0
0
1
,
0
.
8
0
0
±
0
.
0
0
1
,
an
d
0
.
8
0
0
±
0
.
0
0
1
n
o
d
e
u
tili
za
tio
n
,
r
esp
ec
tiv
ely
,
an
d
0
.
7
5
0
±
0
.
0
0
1
,
0
.
8
5
0
±
0
.
0
0
1
,
an
d
0
.
4
0
0
±
0
.
0
0
1
lin
k
u
tili
za
tio
n
,
r
esp
ec
tiv
el
y
,
as
s
u
m
m
a
r
ized
in
Fig
u
r
e
7
.
T
ak
i
n
g
a
ll
to
g
eth
er
,
th
ese
co
m
p
ar
ativ
e
in
s
ig
h
ts
h
i
g
h
lig
h
t
th
e
e
f
f
icien
cy
o
f
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
,
co
n
f
ir
m
th
at
th
e
p
r
o
p
o
s
ed
s
o
lu
tio
n
o
u
tp
er
f
o
r
m
s
th
e
ex
is
tin
g
s
o
lu
tio
n
s
in
ter
m
s
o
f
r
eso
u
r
ce
n
o
d
e
u
tili
za
tio
n
,
an
d
th
e
li
n
k
u
tili
za
tio
n
was
co
m
p
ar
ab
le
t
o
[
2
4
]
0
.
8
0
0
±
0
.
0
0
1
,
with
n
o
s
tatis
tically
s
ig
n
if
ican
t
d
if
f
er
en
ce
p
≈
0
.
9
8
.
T
h
ese
f
in
d
in
g
s
co
n
f
ir
m
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
2
5
2
-
8
9
3
8
I
n
t J Ar
tif
I
n
tell
,
Vo
l.
15
,
No
.
2
,
Ap
r
il
20
26
:
1
1
8
1
-
1
1
9
3
1190
th
at
th
e
p
r
o
p
o
s
ed
s
y
s
tem
im
p
r
o
v
es
b
o
th
n
o
d
e
an
d
lin
k
u
tili
za
tio
n
,
o
f
f
er
in
g
s
ig
n
if
ican
t
g
ain
s
ac
r
o
s
s
m
o
s
t
b
aselin
es wh
ile
p
er
f
o
r
m
in
g
co
m
p
etitiv
ely
with
[
2
4
]
in
lin
k
u
tili
za
tio
n
.
Fig
u
r
e
7
.
No
d
es a
n
d
lin
k
s
u
tili
za
tio
n
r
esu
lts
o
f
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
co
m
p
ar
e
d
with
th
at
in
[
2
1
]
,
[
2
2
]
,
[
2
4
]
4
.
3
.
L
o
a
d ba
la
ncing
a
nd
s
y
s
t
em
s
t
a
bil
it
y
T
h
e
p
r
o
p
o
s
ed
al
g
o
r
ith
m
ac
h
i
ev
ed
a
n
o
d
e
r
eso
u
r
ce
b
alan
c
e
s
co
r
e
ex
ce
ed
i
n
g
0
.
9
0
5
±
0
.
0
0
4
wh
ile
it
s
u
cc
ee
d
ed
in
m
ain
tain
in
g
lin
k
lo
ad
im
b
alan
ce
r
atio
o
f
0
.
2
5
0
±
0
.
0
0
3
.
T
h
e
c
o
m
p
ar
is
o
n
with
[
2
1
]
as su
m
m
ar
ized
in
T
ab
le
4
s
h
o
ws
th
at
th
e
lo
ad
im
b
alan
ce
r
atio
ef
f
icien
cy
o
f
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
ex
ce
e
d
ed
th
at
in
[
2
1
]
by
ap
p
r
o
x
im
ately
0
.
2
3
w
h
er
e
th
ey
attain
ed
a
b
alan
ce
s
co
r
e
o
f
0
.
6
7
0
±
0
.
0
0
3
.
Me
a
n
wh
ile,
t
h
e
co
m
p
ar
is
o
n
also
s
h
o
wed
th
at
t
h
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
s
u
cc
ess
f
u
lly
k
ep
t
a
lo
wer
lin
k
lo
ad
im
b
alan
ce
r
atio
b
y
ap
p
r
o
x
im
ately
0
.
1
3
c
o
m
p
ar
e
d
with
th
at
in
[
2
1
]
wh
er
e
th
ey
d
e
m
o
n
s
tr
ated
a
h
ig
h
er
im
b
alan
ce
r
atio
o
f
0
.
3
8
1
±
0
.
0
0
2
.
T
ab
le
4
.
C
o
m
p
a
r
is
o
n
o
f
n
o
d
e
r
eso
u
r
ce
b
alan
ce
s
co
r
e
an
d
lin
k
im
b
alan
ce
r
atio
with
b
aselin
e
[
2
1
]
M
e
t
r
i
c
P
r
o
p
o
se
d
h
i
d
d
e
n
a
l
g
o
r
i
t
h
m
A
3
C
+
G
C
V
N
o
d
e
r
e
s
o
u
r
c
e
b
a
l
a
n
c
e
s
c
o
r
e
>
0
.
9
0
5
±
0
.
0
0
4
0
.
6
7
0
±
0
.
0
0
3
Li
n
k
l
o
a
d
i
mb
a
l
a
n
c
e
r
a
t
i
o
0
.
2
5
0
±
0
.
0
0
3
0
.
3
8
1
±
0
.
0
0
2
4
.
4
.
E
nerg
y
ef
f
iciency
a
nd
co
m
pu
t
a
t
io
na
l
o
v
er
hea
d r
educt
io
n
A
s
et
o
f
s
im
u
latio
n
s
ce
n
ar
io
s
h
av
e
b
ee
n
co
n
d
u
cte
d
to
ev
alu
ate
th
e
en
er
g
y
co
n
s
u
m
p
tio
n
o
f
th
e
VNE
p
r
o
ce
s
s
u
n
d
e
r
d
if
f
er
en
t
VNR
s
izes.
T
h
e
s
im
u
latio
n
r
esu
lt
s
s
h
o
wed
th
at
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
r
ed
u
ce
d
en
er
g
y
c
o
n
s
u
m
p
tio
n
b
y
2
4
.
9
0
±
0
.
0
1
,
2
4
.
8
0
±
0
.
0
1
,
a
n
d
2
3
.
9
0
±
0
.
0
1
f
o
r
s
m
all,
m
e
d
iu
m
,
an
d
lar
g
e
VNRs
,
r
esp
ec
tiv
ely
,
co
m
p
ar
ed
to
th
at
attain
ed
wh
en
t
h
e
s
im
u
la
tio
n
is
p
er
f
o
r
m
ed
u
s
in
g
th
e
b
aselin
e.
T
ab
le
5
h
ig
h
lig
h
ts
th
e
r
ed
u
ctio
n
o
f
e
n
er
g
y
c
o
n
s
u
m
p
tio
n
u
s
in
g
th
e
p
r
o
p
o
s
ed
h
id
d
e
n
in
f
er
e
n
ce
a
lg
o
r
ith
m
co
m
p
ar
ed
with
th
at
p
r
esen
ted
i
n
[
2
8
]
,
w
h
er
e
th
e
y
s
h
o
wed
a
r
ed
u
ctio
n
in
e
n
er
g
y
c
o
n
s
u
m
p
tio
n
2
4
.
8
0
±
0
.
0
1
,
2
4
.
8
0
±
0
.
0
1
,
an
d
2
3
.
9
0
±
0
.
0
1
f
o
r
s
m
all,
m
e
d
iu
m
,
a
n
d
la
r
g
e
VNRs
,
r
esp
e
ctiv
ely
,
co
m
p
ar
ed
to
t
h
e
b
ase
lin
e.
Ho
wev
er
,
th
e
co
m
p
ar
ativ
e
r
esu
lts
in
T
ab
le
5
s
h
o
w
a
s
lig
h
t
im
p
r
o
v
em
e
n
t
o
f
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
o
v
er
th
at
in
[
2
8
]
,
th
e
p
r
o
p
o
s
ed
alg
o
r
ith
m
d
em
o
n
s
t
r
ates
b
etter
r
eso
u
r
ce
o
r
ch
estra
tio
n
,
co
n
s
is
ten
cy
,
s
u
s
tain
ab
ilit
y
,
s
tab
ilit
y
,
an
d
lo
wer
v
ar
ian
ce
d
u
r
in
g
t
h
e
em
b
ed
d
in
g
p
r
o
ce
s
s
ac
r
o
s
s
v
ar
y
in
g
VNR
s
izes,
co
m
p
ar
ed
with
th
e
en
er
g
y
ef
f
icien
t
alg
o
r
ith
m
in
[
2
8
]
th
at
s
h
o
ws n
o
tab
le
f
lu
ctu
atio
n
s
an
d
h
i
g
h
er
co
m
p
u
tatio
n
al
o
v
er
h
ea
d
s
.
T
ab
le
5
.
E
n
er
g
y
co
n
s
u
m
p
tio
n
co
m
p
ar
is
o
n
with
t
h
at
in
[
2
8
]
A
l
g
o
r
i
t
h
m
S
mal
l
-
s
i
z
e
d
V
N
R
s
M
e
d
i
u
m
-
si
z
e
d
V
N
R
s
La
r
g
e
-
si
z
e
d
V
N
R
s
P
r
o
p
o
se
d
h
i
d
d
e
n
i
n
f
e
r
e
n
c
e
a
l
g
o
r
i
t
h
m
2
4
.
9
0
±
0
.
0
1
2
4
.
8
0
±
0
.
0
1
2
3
.
9
0
±
0
.
0
1
En
e
r
g
y
-
e
f
f
i
c
i
e
n
t
a
l
g
o
r
i
t
h
m
[
2
8
]
2
4
.
8
0
±
0
.
0
1
2
4
.
8
0
±
0
.
0
1
2
3
.
9
0
±
0
.
0
1
4
.
5
.
Acc
ept
a
nce
r
a
t
io
f
o
r
v
ir
t
ua
l net
wo
rk
re
qu
ests
T
h
e
last
s
et
o
f
s
im
u
latio
n
s
ce
n
ar
io
s
h
av
e
b
ee
n
d
ev
o
ted
to
ev
alu
atin
g
th
e
m
o
d
el
ac
ce
p
tan
ce
r
atio
o
v
e
r
tim
e
u
n
d
er
d
if
f
er
e
n
t
a
n
d
v
ar
y
in
g
VNR
s
izes
an
d
co
n
d
itio
n
s
.
Fig
u
r
e
8
illu
s
tr
ates
th
is
r
esu
lt
co
m
p
ar
ed
with
Evaluation Warning : The document was created with Spire.PDF for Python.