I
AE
S In
t
er
na
t
io
na
l J
o
urna
l o
f
Ro
bo
t
ics a
nd
Aut
o
m
a
t
io
n
(
I
J
RA)
Vo
l.
14
,
No
.
4
,
Dec
em
b
er
20
25
,
p
p
.
5
9
1
~
6
0
1
I
SS
N:
2722
-
2
5
8
6
,
DOI
:
1
0
.
1
1
5
9
1
/i
jr
a
.
v
14
i
4
.
pp
5
9
1
-
601
591
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ij
r
a
.
ia
esco
r
e.
co
m
SH
IE
LD:
S
ecu
rit
y
bas
ed hybrid a
u
tono
mo
us deep le
a
rning
network
for lo
a
d
ba
la
ncing
in
clo
u
d
L
o
g
a
P
riy
a
da
rsh
ini
K
a
t
hirma
la
iy
a
n
,
Nit
hy
a
M
uthu
C
e
n
t
r
e
f
o
r
A
r
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
& D
a
t
a
S
c
i
e
n
c
e
(
C
A
I
D
S
)
,
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
,
V
i
n
a
y
a
k
a
M
i
ss
i
o
n
K
i
r
u
p
a
n
a
n
d
a
V
a
r
i
y
a
r
E
n
g
i
n
e
e
r
i
n
g
C
o
l
l
e
g
e
,
V
i
n
a
y
a
k
a
M
i
ssi
o
n
R
e
se
a
r
c
h
F
o
u
n
d
a
t
i
o
n
(
D
U
)
,
S
a
l
e
m,
Ta
m
i
l
N
a
d
u
,
I
n
d
i
a
Art
icle
I
nfo
AB
S
T
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Feb
6
,
2
0
2
5
R
ev
is
ed
Au
g
8
,
2
0
2
5
Acc
ep
ted
Oct
2
2
,
2
0
2
5
Lo
a
d
b
a
lan
c
in
g
in
th
e
in
te
rn
e
t
o
f
th
in
g
s
(
Io
T)
e
n
h
a
n
c
e
s
th
e
e
ffi
c
i
e
n
c
y
o
f
th
e
sy
st
e
m
b
y
d
y
n
a
m
ic
a
l
ly
a
ll
o
c
a
ti
n
g
ta
sk
s
a
c
ro
s
s
d
e
v
ic
e
s
a
n
d
c
lo
u
d
re
so
u
r
c
e
s.
Ho
w
e
v
e
r,
ta
sk
s
c
h
e
d
u
li
n
g
stru
g
g
les
with
u
n
p
re
d
ic
tab
l
e
t
a
sk
s,
s
c
a
lab
il
it
y
,
se
c
u
r
it
y
ri
sk
s
,
a
n
d
u
n
a
u
th
o
ri
z
e
d
a
c
c
e
s
s
c
o
n
t
ro
l.
To
o
v
e
r
c
o
m
e
th
e
s
e
li
m
it
a
ti
o
n
s
,
a
n
o
v
e
l
s
e
c
u
rit
y
-
b
a
s
e
d
h
y
b
rid
a
u
to
n
o
m
o
u
s
d
e
e
p
lea
rn
in
g
n
e
two
rk
fo
r
lo
a
d
b
a
l
a
n
c
in
g
in
c
lo
u
d
(S
HI
EL
D
)
fr
a
m
e
wo
rk
h
a
s
b
e
e
n
p
ro
p
o
se
d
fo
r
se
c
u
r
e
t
a
sk
s
c
h
e
d
u
li
n
g
in
c
lo
u
d
re
so
u
r
c
e
s.
In
it
ia
ll
y
,
th
e
d
a
t
a
re
c
e
iv
e
d
fro
m
th
e
Io
T
d
e
v
ic
e
s
i
s
p
a
s
s
e
d
u
n
d
e
r
c
e
rt
a
in
se
c
u
r
it
y
c
o
n
str
a
in
t
s
t
o
e
n
su
re
th
e
a
u
th
e
n
ti
c
it
y
o
f
th
e
d
a
t
a
.
Th
e
s
e
p
riv
a
c
y
-
p
re
s
e
rv
e
d
d
a
t
a
a
r
e
fe
d
t
o
th
e
t
a
sk
s
c
h
e
d
u
li
n
g
m
o
d
u
le
,
wh
i
c
h
i
s
e
m
p
lo
y
e
d
b
y
th
e
d
u
a
l
D
L
Ne
two
rk
to
g
e
n
e
ra
te
a
s
c
h
e
d
u
l
e
fo
r
r
e
so
u
r
c
e
m
a
n
a
g
e
m
e
n
t
.
F
in
a
ll
y
,
c
lo
u
d
r
e
so
u
rc
e
s
e
m
p
lo
y
o
p
ti
m
a
l
a
ll
o
c
a
t
io
n
o
f
t
a
sk
s
b
a
s
e
d
o
n
th
e
g
e
n
e
r
a
t
e
d
s
c
h
e
d
u
le
to
e
n
su
re
s
e
c
u
r
e
lo
a
d
b
a
l
a
n
c
in
g
.
Th
e
p
ro
p
o
s
e
d
fr
a
m
e
wo
rk
is
s
i
m
u
l
a
t
e
d
b
y
u
sin
g
Clo
u
d
S
i
m
u
l
a
to
r
7
G
(Cl
o
u
d
S
im7
G).
Th
e
S
HI
EL
D
fr
a
m
e
wo
rk
i
s
a
s
s
e
s
s
e
d
b
y
su
c
h
m
e
tri
c
s,
in
c
lu
d
in
g
a
c
c
u
r
a
c
y
,
r
e
c
a
l
l,
p
r
e
c
isio
n
,
F1
-
s
c
o
r
e
,
a
n
d
sp
e
c
ifi
c
i
ty
.
In
c
o
m
p
a
r
iso
n
,
t
h
e
p
ro
p
o
s
e
d
S
HIEL
D
f
r
a
m
e
wo
r
k
a
c
h
i
e
v
e
s
a
p
r
iv
a
c
y
o
v
e
rh
e
a
d
o
f
1
4
%
o
u
tp
e
rfo
r
m
s
th
e
e
x
i
stin
g
QO
D
A
-
L
B,
Be
st
-
KFF
,
SPSO
-
TCS
,
a
n
d
VMM
IS
D
t
e
c
h
n
iq
u
e
s
b
y
a
c
h
i
e
v
in
g
1
0
%
,
1
1
%
,
1
2
%
,
a
n
d
1
3
%
re
sp
e
c
ti
v
e
ly
.
K
ey
w
o
r
d
s
:
Gate
d
r
ec
u
r
r
e
n
t n
eu
r
al
n
etwo
r
k
I
n
ter
n
et
o
f
T
h
in
g
s
L
o
ad
b
alan
ci
n
g
R
ec
u
r
r
en
t n
eu
r
al
n
etwo
r
k
Secu
r
ity
co
n
s
tr
ain
ts
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
:
L
o
g
a
Priy
ad
a
r
s
h
in
i K
ath
ir
m
al
aiy
an
Vin
ay
a
k
a
Miss
i
o
n
Ki
r
u
p
an
an
d
a
V
ar
iy
ar
E
n
g
i
n
e
e
r
i
n
g
C
o
lle
g
e
,
V
in
a
y
a
k
a
Miss
io
n
R
ese
ar
c
h
Fo
u
n
d
a
ti
o
n
(
DU
)
Salem
,
T
am
il Na
d
u
,
I
n
d
ia
E
m
ail:
k
av
ip
r
iy
as1
3
@
g
m
ail.
c
o
m
1.
I
NT
RO
D
UCT
I
O
N
I
n
ter
n
et
o
f
t
h
in
g
s
(
I
o
T
)
an
d
clo
u
d
co
m
p
u
tin
g
(
C
C
)
h
av
e
r
ev
o
lu
tio
n
ized
d
ata
p
r
o
ce
s
s
in
g
an
d
r
eso
u
r
ce
m
an
ag
em
en
t
en
ab
lin
g
th
e
co
n
v
en
ien
t
o
p
er
atio
n
o
f
lin
k
ed
d
e
v
ices
[
1
]
,
[
2
]
.
R
ea
l
-
tim
e
p
r
o
ce
s
s
in
g
an
d
ef
f
icien
t
task
s
ch
ed
u
lin
g
a
r
e
n
ec
ess
ar
y
b
ec
au
s
e
o
f
th
e
e
n
o
r
m
o
u
s
v
o
lu
m
es
o
f
d
ata
g
en
e
r
ated
th
r
o
u
g
h
I
o
T
d
e
v
ices
[
3
]
,
[
4
]
.
Ad
ap
tiv
e
lo
ad
b
alan
cin
g
in
s
u
ch
s
y
s
tem
s
en
s
u
r
es
o
p
tim
al
r
eso
u
r
ce
u
tili
za
tio
n
an
d
p
r
e
v
en
ts
s
er
v
ice
in
ter
r
u
p
tio
n
s
[
5
]
,
[
6
]
.
Ho
wev
e
r
,
with
th
e
s
u
r
g
e
in
d
ata
ex
ch
an
g
e
an
d
task
s
ch
ed
u
lin
g
b
etwe
en
I
o
T
d
ev
ices a
n
d
clo
u
d
e
n
v
ir
o
n
m
en
ts
s
ec
u
r
ity
v
u
ln
er
ab
ilit
ies
h
av
e
escalate
d
m
ak
in
g
it
im
p
e
r
ativ
e
to
ad
o
p
t
s
ec
u
r
e
m
ec
h
an
is
m
s
th
at
s
af
eg
u
ar
d
task
e
x
ec
u
tio
n
wh
ile
m
ain
tain
in
g
s
y
s
tem
p
er
f
o
r
m
an
ce
[
7
]
,
[
8
]
.
Secu
r
e
task
s
ch
ed
u
lin
g
in
I
o
T
-
clo
u
d
s
y
s
tem
s
ty
p
ically
in
v
o
lv
es
th
e
im
p
lem
en
tatio
n
o
f
cr
y
p
to
g
r
ap
h
ic
tech
n
iq
u
es,
an
o
m
aly
d
etec
tio
n
m
ec
h
an
is
m
s
,
an
d
in
tellig
en
t
d
ec
is
io
n
-
m
a
k
in
g
f
r
am
ewo
r
k
s
[
9
]
,
[
1
0
]
.
Ad
d
itio
n
ally
,
l
ev
er
ag
in
g
m
ac
h
i
n
e
lear
n
in
g
(
ML
)
m
o
d
els
f
o
r
d
y
n
am
ic
th
r
ea
t
d
etec
tio
n
a
n
d
ad
ap
tiv
e
s
ch
ed
u
lin
g
f
u
r
th
e
r
en
h
an
ce
s
s
y
s
tem
r
esil
ien
ce
ag
ain
s
t
p
o
ten
tial
attac
k
s
[
1
1
]
,
[
1
2
]
.
Sev
er
al
e
x
is
tin
g
task
s
ch
ed
u
lin
g
tech
n
i
q
u
es
s
u
ch
as
h
eu
r
is
tic
-
b
ased
,
m
etah
eu
r
is
tic,
an
d
r
u
le
-
b
ased
alg
o
r
ith
m
s
h
av
e
b
ee
n
d
ev
elo
p
e
d
to
o
p
tim
ize
r
eso
u
r
c
e
allo
ca
tio
n
in
I
o
T
-
clo
u
d
s
y
s
tem
s
.
Ap
p
r
o
ac
h
es
s
u
ch
as
an
t
co
l
o
n
y
o
p
tim
iza
tio
n
(
AC
O)
,
an
d
f
ir
s
t
-
co
m
e
-
f
ir
s
t
-
s
er
v
e
(
FC
FS
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
14
,
No
.
4
,
Dec
em
b
er
20
25
:
5
9
1
-
6
0
1
592
Sch
ed
u
lin
g
f
o
cu
s
o
n
m
in
im
iz
in
g
ex
ec
u
tio
n
tim
e
an
d
e
n
er
g
y
co
n
s
u
m
p
tio
n
[
1
3
]
,
[
1
4
]
.
W
h
ile
th
ese
m
o
d
els
ad
d
r
ess
v
ar
io
u
s
p
er
f
o
r
m
an
ce
m
etr
ics
th
eir
lack
o
f
b
u
ilt
-
in
s
ec
u
r
ity
f
ea
tu
r
es
o
f
ten
ex
p
o
s
es
I
o
T
-
clo
u
d
s
y
s
tem
s
to
v
u
ln
er
ab
ilit
ies
[
1
5
]
,
[
1
6
]
.
E
n
h
an
cin
g
th
ese
tech
n
iq
u
es with
s
ec
u
r
e
lo
ad
-
b
alan
ci
n
g
m
ea
s
u
r
es is
a
cr
u
cial
s
tep
to
war
d
r
eliab
le
I
o
T
-
clo
u
d
in
te
g
r
atio
n
[
1
7
]
,
[
1
8
]
.
T
ask
s
ch
ed
u
lin
g
an
d
r
eso
u
r
c
e
allo
ca
tio
n
in
I
o
T
s
y
s
tem
s
h
av
e
d
r
awn
th
e
atten
tio
n
o
f
n
u
m
er
o
u
s
ac
ad
em
ics
in
r
ec
e
n
t
y
ea
r
s
.
T
h
is
s
ec
tio
n
f
o
c
u
s
es
o
n
d
y
n
am
ic
lo
ad
b
alan
cin
g
th
at
c
o
m
p
iles
th
e
r
elev
a
n
t
r
esear
ch
in
th
ese
f
ield
s
.
I
n
2
0
2
2
,
L
atch
o
u
m
i
an
d
Par
th
ib
a
n
[
1
9
]
s
u
g
g
ested
b
y
u
s
in
g
th
e
q
u
asi
-
o
p
p
o
s
itio
n
al
b
ased
lear
n
in
g
p
r
in
cip
le,
a
Q
ODA
-
L
B
tech
n
iq
u
e
ca
n
b
e
a
p
p
lied
to
b
o
o
s
t
th
e
d
r
ag
o
n
f
ly
alg
o
r
ith
m
'
s
(
DA)
ty
p
ical
co
n
v
er
g
en
ce
r
ate.
T
h
e
s
u
g
g
ested
QODA
-
L
B
tech
n
iq
u
e
co
m
p
u
tes
an
o
b
jectiv
e
f
u
n
ctio
n
b
ased
o
n
ex
ec
u
tio
n
tim
e
an
d
ex
ec
u
tio
n
co
s
t.
I
n
2
0
2
2
,
Fath
alla
et
a
l.
[
2
0
]
s
u
g
g
ested
an
Allo
ca
tio
n
o
f
r
eso
u
r
ce
s
f
o
r
C
C
s
y
s
tem
s
b
y
em
p
lo
y
in
g
a
m
u
lti
-
g
o
al
s
tr
ateg
y
.
Pre
em
p
tio
n
tim
es,
av
er
ag
e
waitin
g
tim
es
f
o
r
ad
v
an
ce
d
r
eser
v
atio
n
s
,
lease
p
r
e
-
em
p
ti
o
n
,
an
d
en
er
g
y
u
s
ag
e
ar
e
a
ll
d
ec
r
ea
s
ed
b
y
th
e
b
est
-
KFF
tech
n
iq
u
e
wh
ile
o
p
tim
izin
g
th
e
PMs'
u
s
e
o
f
r
eso
u
r
ce
s
.
I
n
2
0
2
3
,
Dev
i
et
a
l.
[
2
1
]
s
u
g
g
ested
Utilizin
g
a
d
ee
p
lo
ad
b
alan
ce
r
with
im
p
r
o
v
e
d
s
ec
u
r
ity
an
d
C
L
OUD
lo
ad
b
alan
cin
g
to
s
to
r
e
d
ata
f
o
r
th
e
I
o
T
.
T
h
e
s
u
g
g
e
s
ted
DL
B
ap
p
r
o
ac
h
ac
h
iev
es
b
etter
s
ca
lab
ilit
y
,
co
s
t
ef
f
icien
cy
,
an
d
en
h
an
ce
d
s
ec
u
r
ity
f
o
r
ef
f
icien
t
lo
ad
b
alan
cin
g
.
I
n
2
0
2
3
,
Sab
a
et
a
l.
[
2
2
]
s
u
g
g
ested
C
lo
u
d
-
ed
g
e
lo
ad
b
alan
cin
g
f
o
r
in
te
r
n
et
o
f
ev
er
y
t
h
in
g
(
I
o
E
)
ap
p
licatio
n
s
b
ased
o
n
d
is
tr
ib
u
ted
s
war
m
in
tellig
en
ce
.
T
h
e
s
u
g
g
ested
ap
p
r
o
ac
h
l
o
wer
s
n
etwo
r
k
co
s
t,
en
d
-
to
-
e
n
d
laten
cy
,
s
u
cc
ess
r
ate,
an
d
e
n
er
g
y
u
s
ag
e
b
y
2
0
%,
1
7
%,
an
d
1
9
%,
r
esp
ec
tiv
ely
.
I
n
2
0
2
3
,
A
g
h
ap
o
u
r
et
a
l.
[
2
3
]
s
u
g
g
ested
allo
ca
tin
g
r
eso
u
r
ce
s
an
d
u
n
lo
ad
in
g
task
s
th
r
o
u
g
h
d
ee
p
r
ein
f
o
r
ce
m
e
n
t
lear
n
in
g
(
DR
L
)
.
T
h
e
s
u
g
g
ested
m
eth
o
d
'
s
av
er
ag
e
p
o
we
r
co
n
s
u
m
p
tio
n
a
n
d
laten
cy
v
al
u
e
ar
e
9
2
%,
1
7
%,
an
d
1
2
%,
r
esp
ec
tiv
ely
.
I
n
2
0
2
4
,
Me
n
a
k
a
an
d
Ku
m
ar
[
2
4
]
s
u
g
g
ested
th
e
lo
ad
b
alan
cin
g
s
tr
ateg
y
in
C
C
is
s
u
p
p
o
r
tiv
e
p
ar
ticle
s
wa
r
m
o
p
tim
izatio
n
with
tim
e
-
c
o
n
s
cio
u
s
s
ch
ed
u
lin
g
(
SP
SO
-
T
C
S).
T
h
e
SP
SO
-
T
C
S
m
eth
o
d
u
s
es
T
C
S
an
d
SP
SO
to
s
h
o
r
ten
th
e
m
ak
e
-
s
p
a
n
ti
m
e
an
d
ac
h
ie
v
e
in
itial
l
o
ad
b
alan
ce
.
I
n
2
0
2
4
,
B
r
ah
m
am
an
d
An
an
d
[
2
5
]
s
u
g
g
ested
a
r
o
b
u
s
t
v
i
r
tu
al
m
a
ch
in
e
lo
ad
b
alan
cin
g
p
ar
a
d
ig
m
.
VM
m
ig
r
atio
n
ef
f
icien
cy
in
c
r
ea
s
ed
b
y
2
.
5
%
,
p
r
o
jecte
d
tim
e
l
o
wer
ed
b
y
9
.
5
%,
an
d
c
o
m
p
u
tatio
n
al
co
m
p
l
ex
ity
d
ec
r
ea
s
ed
b
y
8
.
3
%
with
th
e
aid
o
f
th
e
p
r
o
p
o
s
ed
VM
MI
SD.
T
h
er
ef
o
r
e
,
o
n
e
o
f
th
e
p
r
im
ar
y
s
h
o
r
tc
o
m
in
g
s
o
f
ex
is
tin
g
ap
p
r
o
ac
h
es
is
th
e
r
eq
u
ir
em
en
t
f
o
r
lo
ad
b
alan
cin
g
s
tr
ateg
ie
s
an
d
r
eso
u
r
ce
allo
ca
tio
n
t
h
at
co
n
s
id
er
b
o
t
h
task
r
eq
u
ir
em
e
n
ts
an
d
d
esti
n
atio
n
n
o
d
e
ca
p
ab
ilit
ies.
T
o
o
v
er
co
m
e
th
ese
is
s
u
es,
a
n
o
v
el
s
ec
u
r
ity
-
b
ased
h
y
b
r
id
au
t
o
n
o
m
o
u
s
d
ee
p
lear
n
in
g
n
etwo
r
k
f
o
r
lo
a
d
b
alan
cin
g
in
clo
u
d
(
SHI
E
L
D)
f
r
am
ewo
r
k
h
as
b
ee
n
p
r
o
p
o
s
e
d
to
d
e
v
elo
p
a
s
ec
u
r
ity
-
b
ased
task
s
ch
ed
u
lin
g
in
I
o
T
d
ev
ices
f
o
r
p
r
eser
v
in
g
t
h
e
s
en
s
itiv
e
in
f
o
r
m
atio
n
f
r
o
m
u
n
au
th
o
r
ize
d
ac
ce
s
s
an
d
m
itig
a
te
r
is
k
s
f
r
o
m
cy
b
er
th
r
ea
ts
am
o
n
g
th
e
n
etwo
r
k
.
T
h
e
k
ey
g
o
als
o
f
th
e
SHI
E
L
D
f
r
am
ewo
r
k
ar
e
d
escr
ib
ed
as
f
o
llo
ws:
T
h
e
aim
o
f
th
is
SHI
E
L
D
f
r
am
ewo
r
k
is
to
d
ev
elo
p
an
e
f
f
icien
t
f
r
am
ewo
r
k
f
o
r
r
o
b
u
s
t
lo
ad
b
alan
ci
n
g
in
I
o
T
en
v
ir
o
n
m
en
ts
b
y
in
co
r
p
o
r
atin
g
s
ec
u
r
ity
an
d
p
r
iv
ac
y
-
b
ased
m
ea
s
u
r
es
with
ad
v
an
ce
d
h
y
b
r
id
ized
d
u
al
d
ee
p
lear
n
in
g
(
DL
)
n
etwo
r
k
s
.
T
h
e
SHI
E
L
D
f
r
am
ewo
r
k
p
r
o
v
i
d
es
s
u
b
s
tan
tial
d
a
ta
s
ec
u
r
ity
b
y
s
ev
er
al
s
ec
u
r
it
y
co
n
ce
r
n
s
s
u
ch
as
d
ata
co
n
f
id
e
n
tiality
,
ac
ce
s
s
co
n
tr
o
l,
d
ata
in
te
g
r
ity
,
an
o
m
aly
d
etec
tio
n
,
an
d
d
ata
av
ailab
ilit
y
to
en
s
u
r
e
t
h
at
th
e
I
o
T
d
ata
is
r
esil
ien
t
to
th
r
ea
ts
.
T
h
e
h
y
b
r
id
ized
d
u
al
r
ec
u
r
r
e
n
t
n
eu
r
al
n
etwo
r
k
(
HDRNN)
p
r
o
ce
s
s
es
th
e
task
s
b
ased
o
n
p
r
i
o
r
ity
a
n
d
t
h
e
g
ate
d
r
ec
u
r
r
en
t
n
eu
r
al
n
etwo
r
k
(
G
R
U)
r
ec
o
g
n
izes
th
e
d
ep
e
n
d
en
cies
b
etwe
en
task
s
to
co
n
t
r
ib
u
te
t
o
a
p
r
ec
is
e
tas
k
s
ch
ed
u
le
.
A
s
ec
u
r
ed
lo
ad
b
alan
cin
g
an
d
ef
f
icien
t
r
eso
u
r
ce
m
an
a
g
em
en
t
is
f
ac
ilit
ated
b
y
clo
u
d
r
eso
u
r
ce
s
b
ased
o
n
th
e
s
ch
e
d
u
le
g
en
e
r
at
ed
b
y
th
e
HDRNN
f
o
r
o
p
tim
a
l
task
ex
ec
u
tio
n
in
I
o
T
en
v
ir
o
n
m
e
n
ts
.
T
h
e
r
em
ain
in
g
s
ec
tio
n
s
o
f
th
e
r
esear
ch
is
g
iv
en
as f
o
llo
ws:
Sectio
n
2
p
r
esen
ts
th
e
SHI
E
L
D
Fra
m
ewo
r
k
.
Sectio
n
3
d
escr
ib
es
th
e
ex
p
er
im
en
tal
r
esu
lts
o
f
th
e
SHI
E
L
D
f
r
am
ewo
r
k
.
Sectio
n
4
p
r
esen
ts
th
e
d
is
cu
s
s
io
n
o
f
th
e
r
esear
ch
an
d
s
ec
tio
n
5
co
n
clu
d
es th
e
r
esear
ch
with
f
u
tu
r
e
en
h
a
n
ce
m
en
t.
2.
T
H
E
SH
I
E
L
D
M
E
T
H
O
DO
L
O
G
Y
I
n
th
is
s
ec
tio
n
,
a
n
o
v
el
Secu
r
ity
b
ased
h
y
b
r
id
au
to
n
o
m
o
u
s
d
ee
p
lear
n
in
g
n
etwo
r
k
f
o
r
lo
a
d
b
alan
cin
g
in
clo
u
d
(
SHI
E
L
D)
f
r
am
ew
o
r
k
h
as
b
ee
n
p
r
o
p
o
s
ed
to
d
ev
elo
p
a
p
r
iv
ac
y
-
b
ased
task
s
ch
e
d
u
lin
g
in
I
o
T
d
ev
ices
u
s
in
g
d
u
al
DL
n
etwo
r
k
s
.
I
n
itia
lly
,
r
aw
d
ata
s
u
ch
as task
d
ata,
r
eso
u
r
ce
d
ata,
task
co
n
s
tr
ain
ts
,
I
o
T
en
v
i
r
o
n
m
e
n
t
in
f
o
r
m
atio
n
,
an
d
h
is
to
r
ical
d
ata
ar
e
g
ath
er
ed
f
r
o
m
th
e
I
o
T
d
ev
ices
f
o
r
d
y
n
a
m
ic
task
s
ch
ed
u
lin
g
.
A
ze
r
o
-
tr
u
s
t
s
ec
u
r
ity
m
o
d
u
le
is
in
teg
r
ated
to
v
er
if
ies
ev
er
y
in
p
u
t
d
ata
th
r
o
u
g
h
r
ea
l
-
tim
e
ev
alu
atio
n
s
an
d
co
n
tex
t
-
awa
r
e
p
o
licies
th
at
d
etec
t
ad
v
er
s
ar
ial
attac
k
s
.
T
h
ese
v
er
if
ied
d
ata
ar
e
f
ed
to
s
ev
er
al
s
ec
u
r
ity
co
n
s
tr
ain
ts
s
u
ch
as
d
ata
co
n
f
id
en
tiality
,
ac
ce
s
s
co
n
tr
o
l
,
d
ata
in
te
g
r
ity
,
a
n
o
m
aly
d
ete
ctio
n
,
an
d
d
ata
av
ailab
ilit
y
f
o
r
d
ata
v
er
if
icatio
n
.
T
h
ese
v
er
if
ied
an
d
s
ec
u
r
ed
d
a
ta
alo
n
g
with
p
r
iv
ac
y
f
ea
tu
r
es
ar
e
f
ed
to
th
e
HDRNN
f
o
r
task
s
ch
ed
u
lin
g
.
T
h
e
HDRNN
g
en
er
ates
a
s
ch
ed
u
le
b
y
p
r
o
ce
s
s
in
g
th
e
task
s
b
ased
o
n
p
r
io
r
ity
th
r
o
u
g
h
R
NN
an
d
th
e
GR
U
n
etwo
r
k
r
ec
o
g
n
izes
th
e
d
ep
e
n
d
en
cies
b
etwe
en
task
s
.
T
h
e
SHAP
m
eth
o
d
p
r
o
ce
s
s
es
th
e
s
ch
ed
u
lin
g
d
ec
is
io
n
f
r
o
m
HDRNN
an
d
ass
ig
n
s
an
im
p
o
r
tan
ce
s
co
r
e
to
ea
c
h
task
a
cc
o
r
d
in
g
to
th
e
task
f
ea
tu
r
es
f
o
r
th
e
f
in
al
task
s
ch
ed
u
le.
L
I
ME
p
r
o
v
id
es
a
tas
k
s
p
ec
if
ic
e
x
p
lan
atio
n
an
d
e
n
s
u
r
es
th
at
t
h
e
h
i
g
h
-
p
r
io
r
ity
an
d
s
en
s
itiv
e
task
s
ar
e
h
an
d
led
co
r
r
ec
tly
with
o
u
t
p
r
i
o
r
itizin
g
th
e
u
n
au
th
o
r
ized
task
s
.
Fin
ally
,
th
e
clo
u
d
r
eso
u
r
c
es
allo
ca
te
th
e
ta
s
k
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
S
HI
E
LD:
S
ec
u
r
ity
b
a
s
ed
h
yb
r
id
a
u
to
n
o
mo
u
s
d
ee
p
lea
r
n
in
g
n
etw
o
r
k
…
(
Lo
g
a
P
r
iya
d
a
r
s
h
i
n
i Ka
th
ir
ma
la
iya
n
)
593
b
ased
o
n
th
e
s
ch
ed
u
le
g
en
e
r
ated
b
y
th
e
HDRNN
to
b
alan
ce
th
e
lo
ad
s
ec
u
r
ely
with
o
p
tim
al
r
eso
u
r
ce
u
tili
za
tio
n
.
T
h
e
b
l
o
ck
d
ia
g
r
am
o
f
th
e
d
ev
elo
p
e
d
SHI
E
L
D
m
o
d
el
is
r
ep
r
esen
ted
in
Fig
u
r
e
1
.
Fig
u
r
e
1
.
Pro
p
o
s
ed
SHI
E
L
D
f
r
am
ewo
r
k
2
.
1
.
Z
er
o
-
t
rus
t
m
ec
ha
nis
m
T
h
e
d
iv
er
s
e
ty
p
es
o
f
d
ata
ar
e
g
ath
er
ed
f
r
o
m
th
e
I
o
T
d
ev
ic
es
wh
ich
s
h
o
u
ld
co
n
tain
s
u
c
h
s
ec
u
r
ity
r
is
k
s
.
T
h
e
in
p
u
t
d
ata
f
r
o
m
th
e
I
o
T
d
e
v
ices
ar
e
ca
teg
o
r
ize
d
i
n
to
task
d
ata,
r
eso
u
r
ce
d
ata,
task
co
n
s
tr
ain
ts
,
I
o
T
co
n
tex
tu
al
d
ata,
task
f
ea
tu
r
es,
an
d
h
is
to
r
ical
d
ata
wh
ich
co
n
s
is
t
o
f
an
o
p
er
atio
n
al
en
v
ir
o
n
m
en
t
a
n
d
r
eso
u
r
ce
d
em
an
d
s
.
T
h
e
ze
r
o
-
tr
u
s
t
m
o
d
u
le
is
an
ad
v
an
ce
d
a
n
d
co
n
ti
n
u
o
u
s
s
ec
u
r
ity
m
ec
h
a
n
is
m
th
at
v
er
if
ies
ev
er
y
in
p
u
t
d
ata
th
r
o
u
g
h
r
ea
l
-
tim
e
ev
al
u
a
tio
n
s
an
d
co
n
tex
t
-
awa
r
e
p
o
lic
ies.
T
h
ese
ze
r
o
-
tr
u
s
t
r
o
b
u
s
t
s
ec
u
r
ity
m
ec
h
an
is
m
v
er
if
ies
th
e
d
ata
an
d
d
etec
ts
th
e
b
lack
-
b
o
x
ad
v
er
s
ar
ial
attac
k
s
in
clu
d
in
g
m
o
d
el
ev
asio
n
,
m
o
d
el
ex
tr
ac
tio
n
,
an
d
p
o
is
o
n
in
g
attem
p
ts
f
o
r
ef
f
icie
n
t
th
r
ea
t
m
itig
atio
n
also
to
en
s
u
r
e
s
ec
u
r
e
an
d
v
er
if
ie
d
lo
ad
b
alan
cin
g
am
o
n
g
th
e
clo
u
d
n
etwo
r
k
.
2.
2
.
P
riv
a
cy
ba
s
ed
f
ea
t
ures
T
h
e
in
p
u
t
d
ata
g
ath
er
ed
f
r
o
m
th
e
I
o
T
d
e
v
ices
n
ee
d
s
to
b
e
v
er
if
ied
if
th
e
task
'
s
s
ec
u
r
ity
le
v
el
is
less
th
an
th
at
o
f
th
e
u
n
ass
ig
n
ed
task
.
T
h
e
s
ec
u
r
e
d
task
s
ar
e
ex
tr
ac
ted
f
r
o
m
th
e
i
n
p
u
t
d
ata
b
y
v
alid
atin
g
ce
r
tai
n
s
ec
u
r
ity
co
n
s
tr
ain
ts
s
u
ch
as
ac
ce
s
s
co
n
tr
o
l,
co
n
f
id
en
tial
ity
,
an
o
m
aly
d
etec
tio
n
,
a
n
d
d
ata
av
ailab
ilit
y
.
Mo
r
eo
v
er
,
if
t
h
e
s
ec
u
r
ity
co
n
s
tr
ain
ts
ar
e
ev
alu
ated
a
n
d
th
e
o
b
tain
ed
s
o
lu
tio
n
is
in
f
ea
s
ib
le
th
en
th
e
task
is
r
ep
r
esen
ted
as a
n
u
n
s
ec
u
r
e
d
ta
s
k
wh
ich
will a
ls
o
b
e
f
ilter
ed
.
Priv
ac
y
-
b
ased
f
ilter
in
g
is
g
iv
e
n
in
Alg
o
r
ith
m
1
.
Alg
o
r
ith
m
1
: Pr
iv
ac
y
b
ased
f
ilter
in
g
Input: Security constraint of tasks
Output: Privacy Based Features
1. for
=
1
to N do
2.
for
=
1
to
|
|
do
3.
←
random (),
←
random ()
4.
if
(
(
(
)
)
<
(
)
)
do
5.
=
−
1
6.
end if
7.
end for
8.
while
(
(
)
>
)
do
9.
for each tc
∈
CP do
10.
.
[
.
]
←
(
)
,
.
[
.
]
←
(
)
11.
end for
12.
end while
13. end for
T
h
ese
s
ec
u
r
ed
d
ata
alo
n
g
with
ce
r
tain
p
r
iv
ac
y
f
ea
tu
r
es
ar
e
f
ed
to
h
y
b
r
id
ized
d
u
al
n
e
u
r
al
n
etwo
r
k
f
o
r
s
ec
u
r
e
task
s
ch
ed
u
lin
g
.
2
.
2
.
T
a
s
k
s
cheduli
ng
v
ia
H
DRNN
T
h
ese
v
er
if
ied
d
ata
alo
n
g
wit
h
p
r
iv
ac
y
f
ea
tu
r
es
ar
e
f
ed
to
th
e
HDRNN
wh
ich
co
m
b
in
es
r
ec
u
r
r
en
t
n
eu
r
al
n
etwo
r
k
an
d
g
ate
d
r
e
cu
r
r
en
t
n
e
u
r
al
n
etwo
r
k
f
o
r
t
ask
s
ch
ed
u
lin
g
.
T
h
e
SHI
E
L
D
m
eth
o
d
o
lo
g
y
f
o
r
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
14
,
No
.
4
,
Dec
em
b
er
20
25
:
5
9
1
-
6
0
1
594
s
ch
ed
u
lin
g
th
e
s
ec
u
r
e
d
d
ata
u
tili
zin
g
R
NN,
Den
s
e,
an
d
GR
U
lay
er
s
.
T
h
e
ar
ch
itectu
r
al
c
o
n
f
ig
u
r
atio
n
o
f
th
e
HDRNN i
s
s
h
o
wn
in
Fig
u
r
e
2
.
T
h
e
ℎ
an
d
d
eter
m
in
es th
e
h
id
d
en
s
tate
an
d
in
p
u
t
.
A
h
id
d
e
n
s
tate
h
_
t
an
d
a
n
in
p
u
t
x
_
t f
r
o
m
th
e
p
r
e
v
io
u
s
tim
e
s
tep
ar
e
u
s
ed
b
y
th
e
R
NN
to
cr
e
ate
a
n
ew
ℎ
+
1
an
d
an
o
u
tp
u
t
at
ea
ch
tim
e
s
tep
t.
E
q
u
atio
n
s
(
1
)
th
r
o
u
g
h
(
2
)
lis
t th
e
s
tep
s
.
ℎ
+
1
=
(
ℎ
∗
+
ℎℎ
∗
ℎ
+
)
(
1
)
=
(
ℎ
∗
ℎ
+
1
+
)
(
2
)
H
er
e
,
ℎ
,
ℎℎ
,
an
d
ℎ
en
s
u
r
es th
e
i
n
p
u
t
to
h
id
d
e
n
,
h
id
d
en
to
h
id
d
e
n
,
a
n
d
h
id
d
en
to
o
u
tp
u
t w
eig
h
t m
a
tr
ices
ar
e
d
is
tin
g
u
is
h
ed
b
y
th
ei
r
clas
s
if
icatio
n
.
I
n
th
is
ca
s
e,
an
d
ar
e
b
iass
ed
ter
m
s
.
T
h
e
h
id
d
e
n
la
y
er
em
p
lo
y
s
,
an
d
th
e
o
u
tp
u
t la
y
er
u
s
es
.
I
t is co
n
n
ec
ted
to
th
e
d
en
s
e
lay
er
b
y
ea
ch
n
eu
r
o
n
f
r
o
m
th
e
p
r
ec
e
d
in
g
lev
els.
Fig
u
r
e
2
.
Ar
c
h
itectu
r
e
o
f
HDRNN
E
q
u
atio
n
(
3
)
d
escr
ib
es
th
e
R
e
L
U
o
f
th
e
d
en
s
e
lay
e
r
,
a
n
d
th
e
(
4
)
d
is
p
lay
s
th
e
p
r
ed
icted
o
u
tco
m
e
o
f
th
e
d
en
s
e
lay
er
.
W
h
er
e
is
b
ia
s
ed
,
is
th
e
in
p
u
t
o
f
th
e
d
en
s
e
lay
er
,
an
d
ℎ
is
th
e
we
ig
h
t
m
atr
ix
.
T
h
e
g
atin
g
m
eth
o
d
allo
ws th
e
GR
U
an
d
R
NN
to
s
elec
t u
p
d
ate
o
r
ig
n
o
r
e
d
ata
f
r
o
m
th
e
p
r
ec
ed
in
g
tim
e
s
tep
.
(
)
=
(
0
,
)
(
3
)
ℎ
=
(
ℎ
∗
+
)
(
4
)
T
h
e
r
eset
a
n
d
u
p
d
ate
g
ates
a
r
e
th
e
two
g
ates
in
th
e
GR
U
th
at
r
eg
u
late
th
e
f
lo
w
o
f
i
n
f
o
r
m
atio
n
.
T
h
e
GR
U
'
s
o
u
tp
u
t
is
tr
an
s
f
o
r
m
ed
in
to
a
p
r
o
b
ab
ilit
y
d
is
tr
ib
u
tio
n
ac
r
o
s
s
all
p
o
ten
tial
o
u
tp
u
ts
u
s
in
g
th
e
So
f
tMa
x
ac
tiv
atio
n
f
u
n
ctio
n
.
E
q
u
atio
n
(
5
)
–
(
8
)
p
r
o
v
id
es
a
m
ath
em
at
ical
r
ep
r
esen
tatio
n
o
f
th
e
GR
U
o
p
er
atio
n
s
.
,
ℎ
,
,
ℎ
,
ℎ
,
an
d
ℎℎ
r
ep
r
esen
ts
th
e
weig
h
t
m
atr
ices
in
in
p
u
t
-
u
p
d
ate,
in
p
u
t
-
r
eset,
in
p
u
t
-
h
i
d
d
en
,
an
d
h
id
d
en
-
h
id
d
e
n
co
n
n
ec
tio
n
s
.
H
er
e,
,
,
,
an
d
ℎ
ar
e
th
e
b
iased
ter
m
s
.
=
(
∗
+
ℎ
∗
ℎ
+
)
(
5
)
=
(
∗
+
ℎ
∗
ℎ
+
)
(
6
)
ℎ
′
=
ℎ
(
ℎ
∗
+
ℎℎ
∗
(
∗
ℎ
)
+
ℎ
)
(
7
)
ℎ
=
∗
ℎ
+
(
1
−
)
∗
ℎ
′
(
8
)
T
h
e
HDRNN
p
r
o
ce
s
s
es
th
e
v
er
if
ied
I
o
T
task
to
g
en
er
ate
th
e
s
ch
ed
u
lin
g
d
ec
is
io
n
b
ased
o
n
th
e
task
p
r
io
r
ity
an
d
task
d
ep
en
d
en
cy
.
T
h
e
SHAP
m
eth
o
d
p
r
o
ce
s
s
es
th
e
s
ch
ed
u
lin
g
d
ec
is
io
n
f
r
o
m
HDRNN
an
d
ass
ig
n
s
an
im
p
o
r
ta
n
ce
s
co
r
e
t
o
ea
ch
in
p
u
t
f
ea
tu
r
e
f
o
r
th
e
f
i
n
al
task
s
ch
ed
u
le
ac
co
r
d
in
g
to
th
e
u
r
g
en
c
y
o
f
th
e
task
,
task
s
ize,
task
d
ea
d
lin
e,
an
d
r
eso
u
r
ce
av
ailab
ilit
y
.
L
I
ME
p
r
o
v
id
es
a
task
-
s
p
ec
if
ic
ex
p
lan
atio
n
to
t
h
e
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
S
HI
E
LD:
S
ec
u
r
ity
b
a
s
ed
h
yb
r
id
a
u
to
n
o
mo
u
s
d
ee
p
lea
r
n
in
g
n
etw
o
r
k
…
(
Lo
g
a
P
r
iya
d
a
r
s
h
i
n
i Ka
th
ir
ma
la
iya
n
)
595
ad
m
in
s
to
u
n
d
er
s
tan
d
th
e
e
x
a
ct
r
ea
s
o
n
s
b
eh
in
d
t
h
e
d
ec
is
io
n
f
o
r
a
p
ar
ticu
lar
task
s
ch
ed
u
lin
g
.
LIME
en
s
u
r
es
th
at
th
e
h
ig
h
-
p
r
io
r
ity
an
d
s
en
s
itiv
e
task
s
wer
e
h
an
d
led
co
r
r
ec
tly
with
o
u
t
p
r
io
r
itizin
g
th
e
u
n
au
th
o
r
ized
task
s
.
Pru
n
in
g
is
a
s
tr
ateg
y
th
at
r
ed
u
ce
s
th
e
s
ize
an
d
co
m
p
u
tatio
n
al
co
m
p
lex
ity
o
f
th
e
m
o
d
el
with
o
u
t
s
ig
n
if
ican
tly
af
f
ec
tin
g
its
p
er
f
o
r
m
a
n
ce
,
in
cr
ea
s
in
g
th
e
HDRNN
's
ef
f
icac
y
an
d
ef
f
icien
c
y
.
Pru
n
in
g
is
th
e
p
r
o
ce
s
s
o
f
elim
in
atin
g
n
etwo
r
k
elem
e
n
ts
th
at
h
av
e
th
e
least
ef
f
ec
t
o
n
th
e
n
etwo
r
k
'
s
o
u
tp
u
t,
s
u
ch
as
we
ig
h
ts
an
d
n
eu
r
o
n
s
.
T
h
is
ap
p
r
o
ac
h
p
r
o
d
u
ce
s
a
co
m
p
ac
t a
n
d
ef
f
ec
tiv
e
m
o
d
el
th
at
en
ab
les q
u
ick
er
in
f
e
r
en
ce
tim
e
s
an
d
lo
wer
en
er
g
y
u
s
e.
T
o
m
in
im
ize
th
e
lo
s
s
f
u
n
ctio
n
,
th
e
g
r
a
d
ien
t
d
escen
t
is
u
s
ed
to
u
p
d
ate
th
e
weig
h
ts
th
r
o
u
g
h
o
u
t
th
e
b
ac
k
p
r
o
p
ag
atio
n
tr
ain
in
g
.
B
ec
au
s
e
q
u
an
tizatio
n
is
n
ec
ess
a
r
y
f
o
r
b
o
th
m
o
d
el
s
ize
r
ed
u
c
tio
n
an
d
i
n
f
er
e
n
ce
s
p
ee
d
,
it
ca
n
b
e
u
s
ed
in
s
itu
at
io
n
s
with
lim
ited
r
eso
u
r
ce
s
.
On
e
im
p
o
r
ta
n
t
m
eth
o
d
f
o
r
c
o
m
p
r
ess
in
g
m
o
d
els
is
q
u
an
tizatio
n
,
wh
ich
lo
wer
s
s
t
o
r
ag
e
n
ee
d
s
an
d
im
p
r
o
v
es
co
m
p
u
tatio
n
e
f
f
icien
cy
.
Qu
an
tiz
atio
n
,
a
tech
n
iq
u
e
th
at
in
cr
ea
s
es
co
m
p
u
tin
g
s
p
ee
d
an
d
m
em
o
r
y
u
s
e
with
o
u
t
s
u
b
s
tan
tially
af
f
ec
tin
g
m
o
d
el
p
e
r
f
o
r
m
a
n
ce
,
in
v
o
lv
es
s
im
p
lify
in
g
th
e
m
o
d
el
an
d
r
e
d
u
cin
g
th
e
p
r
ec
is
io
n
o
f
th
e
weig
h
ts
an
d
ac
tiv
atio
n
s
.
T
h
e
HDRNN
g
en
er
ates
a
s
ch
ed
u
le
b
y
p
r
o
ce
s
s
in
g
th
e
task
s
b
ased
o
n
p
r
io
r
ity
th
r
o
u
g
h
R
NN
an
d
th
e
GR
U
n
etw
o
r
k
r
ec
o
g
n
izes
th
e
d
ep
en
d
e
n
cies
b
etwe
en
task
s
.
Fin
ally
,
th
e
cl
o
u
d
r
eso
u
r
ce
s
al
lo
ca
te
th
e
task
b
ased
o
n
th
e
s
ch
ed
u
le
g
en
er
ate
d
b
y
th
e
HDRNN to
b
alan
ce
th
e
lo
ad
s
ec
u
r
ely
with
o
p
tim
al
r
eso
u
r
ce
u
tili
za
tio
n
.
3.
RE
SU
L
T
AND
DI
SCUS
SI
O
N
T
h
e
ex
p
e
r
im
en
tal
ass
ess
m
en
t
o
f
th
e
SHI
E
L
D
a
p
p
r
o
ac
h
is
d
i
s
cu
s
s
ed
in
th
is
s
ec
tio
n
.
T
h
e
P
C
u
s
ed
f
o
r
th
e
an
aly
s
is
is
th
e
W
in
d
o
ws 1
0
Pro
with
an
I
n
tel
(
R
)
C
o
r
e
(
T
M)
i5
C
PU,
clo
ck
in
g
at
3
.
2
0
GHz
an
d
with
8
GB
o
f
R
AM
.
T
h
e
SHI
E
L
D
ap
p
r
o
a
ch
is
s
im
u
lated
b
y
u
s
in
g
C
lo
u
d
Simu
lato
r
(
C
lo
u
d
Sim7
G)
.
3
.
1
.
Co
m
pa
ra
t
iv
e
a
na
ly
s
is
T
h
e
SHI
E
L
D
m
o
d
el
is
an
aly
ze
d
with
s
ev
er
al
p
r
ev
io
u
s
ly
d
ev
elo
p
ed
m
o
d
els
an
d
d
is
p
lay
s
ex
p
er
im
en
tal
r
esu
lts
.
T
h
e
ac
c
u
r
ac
y
,
r
ec
all,
p
r
ec
is
io
n
,
F1
-
s
co
r
e,
a
n
d
s
p
ec
if
icity
o
f
th
e
S
HI
E
L
D
s
tr
ateg
y
ar
e
co
n
tr
asted
with
ex
is
tin
g
m
eth
o
d
s
in
Fig
u
r
e
3
(
a)
.
T
h
e
SHI
E
L
D
s
tr
ateg
y
o
u
tp
er
f
o
r
m
s
th
e
QODA
-
L
B
,
B
est
-
KFF,
SP
SO
-
T
C
S,
an
d
VM
MI
SD
m
o
d
els
an
d
ac
h
ie
v
es
an
ac
cu
r
ac
y
o
f
8
5
%,
9
0
%,
8
8
%,
8
7
%,
an
d
9
3
.
5
%
r
esp
ec
tiv
ely
.
T
h
e
c
o
m
p
ar
is
o
n
o
f
p
o
wer
co
n
s
u
m
p
tio
n
f
o
r
th
e
SHI
E
L
D
f
r
a
m
ewo
r
k
is
r
ep
r
esen
ted
in
Fig
u
r
e
3
(
b
)
.
T
h
e
SHI
E
L
D
f
r
a
m
ewo
r
k
ac
h
ie
v
es
th
e
lo
west
p
o
wer
co
n
s
u
m
p
tio
n
o
f
1
8
.
2
W
to
2
4
.
3
W
ac
r
o
s
s
all
task
r
an
g
es
f
r
o
m
2
0
0
0
to
1
0
,
0
0
0
.
T
h
is
in
d
icate
s
th
e
SHI
E
L
D
f
r
am
ewo
r
k
’
s
s
u
p
e
r
io
r
en
er
g
y
e
f
f
icien
cy
b
y
m
ain
tain
in
g
u
p
to
4
1
%
lo
wer
co
n
s
u
m
p
tio
n
at
h
ig
h
er
l
o
ad
s
ag
ain
s
t
th
e
QODA
-
L
B
,
B
est
-
KFF,
SP
SO
-
T
C
S,
an
d
VM
MI
SD te
ch
n
iq
u
es.
(
a)
(
b
)
Fig
u
r
e
3
.
An
al
y
s
is
o
f
(
a)
p
e
r
f
o
r
m
an
ce
an
d
(
b
)
p
o
we
r
co
n
s
u
m
p
tio
n
o
f
SHI
E
L
D
f
r
am
ewo
r
k
T
h
e
SHI
E
L
D
f
r
am
ewo
r
k
c
o
n
s
is
ten
tly
o
u
tp
er
f
o
r
m
s
ex
is
tin
g
t
ec
h
n
iq
u
es
i
n
s
ca
lab
ilit
y
ac
r
o
s
s
all
task
s
is
r
ep
r
esen
ted
in
Fig
u
r
e
4
.
I
t
a
ch
iev
es
s
ca
lab
ilit
y
r
ates
o
f
8
9
%,
8
7
%,
8
5
%,
an
d
8
3
%
f
o
r
ta
s
k
s
1
0
0
,
2
0
0
,
3
0
0
,
an
d
4
0
0
,
r
esp
ec
tiv
ely
.
T
h
e
SH
I
E
L
D
f
r
am
ewo
r
k
o
u
tp
e
r
f
o
r
m
s
ex
is
tin
g
QODA
-
L
B
,
B
est
-
KF
F,
SP
S
O
-
T
C
S,
an
d
VM
MI
SD
ap
p
r
o
ac
h
es
b
y
m
a
in
tain
in
g
th
e
h
ig
h
est
s
co
r
es
f
r
o
m
9
3
%
at
2
0
0
0
task
s
to
8
6
%
at
1
0
,
0
0
0
task
s
r
esp
ec
tiv
ely
.
T
h
e
SHI
E
L
D
f
r
a
m
ewo
r
k
s
h
o
ws
th
e
s
u
p
er
io
r
a
b
ilit
y
to
h
an
d
le
in
cr
ea
s
in
g
task
lo
ad
s
ef
f
icien
tly
,
with
u
p
to
2
0
.
3
%
b
etter
s
ca
lab
ilit
y
at
h
i
g
h
er
task
r
a
n
g
es.
T
h
e
p
r
o
p
o
s
ed
SHI
E
L
D
f
r
am
ew
o
r
k
o
u
tp
e
r
f
o
r
m
s
th
e
ex
is
tin
g
tech
n
iq
u
es
in
r
eso
u
r
c
e
u
tili
za
tio
n
ac
r
o
s
s
all
task
s
is
s
h
o
wn
i
n
Fig
u
r
e
5
.
Fo
r
task
s
1
0
0
,
2
0
0
,
3
0
0
,
an
d
4
0
0
,
it
ac
h
iev
es
u
tili
za
tio
n
r
ates
o
f
8
9
%,
8
7
%,
8
5
%,
an
d
8
3
%
r
esp
ec
tiv
ely
.
I
n
co
m
p
ar
is
o
n
,
QODA
-
L
B
ac
h
ie
v
es
7
0
%,
B
est
-
KFF
a
ch
ie
v
es
7
2
%
,
S
PS
O
-
T
C
S
ac
h
i
e
v
es
7
1
%
,
a
n
d
V
MM
I
SD
ac
h
i
e
v
es
7
4
%
r
es
p
e
cti
v
e
ly
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
14
,
No
.
4
,
Dec
em
b
er
20
25
:
5
9
1
-
6
0
1
596
Fig
u
r
e
4
.
An
al
y
s
is
o
f
s
ca
lab
ilit
y
Fig
u
r
e
5
.
An
al
y
s
is
o
f
r
eso
u
r
ce
u
tili
za
tio
n
T
h
e
p
r
o
p
o
s
ed
f
r
am
ewo
r
k
i
n
c
u
r
s
s
lig
h
tly
h
ig
h
er
p
r
iv
ac
y
o
v
er
h
ea
d
s
co
m
p
ar
ed
to
ex
is
tin
g
tech
n
iq
u
es
ac
r
o
s
s
all
task
s
wh
ich
ar
e
illu
s
tr
ated
in
Fig
u
r
e
6
.
Fo
r
task
s
1
0
0
,
2
0
0
,
3
0
0
,
a
n
d
4
0
0
it
o
b
tain
s
o
v
er
h
ea
d
s
o
f
1
4
%,
1
6
%,
1
8
%,
an
d
2
0
%,
r
e
s
p
ec
tiv
ely
.
Fig
u
r
e
7
d
e
p
icts
th
e
p
er
f
o
r
m
an
ce
ass
ess
m
en
t
o
f
th
e
SHI
E
L
D
m
o
d
el
with
ex
is
tin
g
Hy
b
r
id
DQN
-
PPO
-
GNN
-
R
L
,
HAFed
R
L
,
an
d
T
F
-
DDRL
tech
n
iq
u
es.
T
h
e
p
r
o
p
o
s
ed
SHI
E
L
D
f
r
am
ewo
r
k
ac
h
iev
es d
ec
en
tr
ali
ze
d
s
ec
u
r
ity
o
f
9
7
%,
ad
a
p
tiv
e
task
s
ch
ed
u
lin
g
o
f
9
2
%,
an
d
s
eq
u
en
ce
p
r
ed
ictio
n
o
f
9
4
% wh
ich
o
u
tp
e
r
f
o
r
m
s
th
e
ex
is
tin
g
tech
n
iq
u
es r
esp
ec
tiv
e
ly
.
Fig
u
r
e
8
s
h
o
ws
th
e
co
m
p
ar
is
o
n
o
f
co
n
s
u
m
p
tio
n
o
f
p
o
we
r
b
y
th
e
E
d
g
e
AI
Dev
ices
f
o
r
en
er
g
y
ef
f
icien
t
lo
ad
b
ala
n
cin
g
.
T
h
e
E
SP
3
2
d
ev
ice
attain
s
th
e
lo
w
est
p
o
wer
co
n
s
u
m
p
tio
n
o
f
0
.
0
8
W
,
0
.
2
5
W
,
an
d
0
.
4
W
f
o
r
id
le
p
o
we
r
,
av
er
ag
e
p
o
wer
,
an
d
m
a
x
im
u
m
p
o
wer
f
o
r
SHI
E
L
D
s
ch
ed
u
lin
g
r
esp
e
ctiv
ely
.
I
n
co
n
tr
ast
,
J
etso
n
Nan
o
co
n
s
u
m
es
h
ig
h
p
o
wer
o
f
2
.
6
W
,
5
.
1
W
,
5
.
6
W
f
o
r
id
le
p
o
wer
,
av
e
r
ag
e
p
o
wer
,
an
d
m
a
x
im
u
m
p
o
wer
wh
ich
is
s
u
itab
le
f
o
r
p
o
wer
-
r
ich
AI
-
in
ten
s
iv
e
task
s
.
I
n
r
ea
l
-
wo
r
ld
d
ep
l
o
y
m
en
t,
t
h
e
SHI
E
L
D
f
r
am
ewo
r
k
will p
r
o
v
id
e
ef
f
icien
t p
er
f
o
r
m
an
ce
u
n
d
er
h
eter
o
g
en
e
o
u
s
I
o
T
d
ev
ices f
o
r
r
o
b
u
s
t a
n
d
s
ca
lab
l
e
lo
ad
b
alan
ci
n
g
.
T
h
e
co
m
p
ar
is
o
n
o
n
laten
c
y
is
d
ep
icted
in
Fig
u
r
e
9
.
T
h
e
S
HI
E
L
D
f
r
am
ewo
r
k
co
n
s
is
ten
tly
ac
h
iev
es
lo
wer
d
elay
co
m
p
ar
e
d
to
ex
i
s
tin
g
tech
n
iq
u
es
s
u
ch
as
QO
DA
-
L
B
,
B
e
s
t
-
KFF,
SP
SO
-
T
C
S,
an
d
VM
MI
S
D
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
S
HI
E
LD:
S
ec
u
r
ity
b
a
s
ed
h
yb
r
id
a
u
to
n
o
mo
u
s
d
ee
p
lea
r
n
in
g
n
etw
o
r
k
…
(
Lo
g
a
P
r
iya
d
a
r
s
h
i
n
i Ka
th
ir
ma
la
iya
n
)
597
ac
r
o
s
s
all
task
s
.
Fo
r
task
s
1
0
0
,
2
0
0
,
3
0
0
,
a
n
d
4
0
0
,
it
ac
h
iev
es
d
elay
s
o
f
9
5
m
s
,
1
0
0
m
s
,
1
1
0
m
s
,
an
d
1
2
0
m
s
,
r
esp
ec
tiv
ely
.
T
h
e
SHI
E
L
D
f
r
am
ewo
r
k
ac
h
iev
es
th
e
lo
west
laten
cy
o
f
9
7
m
s
f
o
r
2
0
0
0
t
ask
s
to
2
4
9
m
s
f
o
r
1
0
,
0
0
0
task
s
f
o
r
h
an
d
li
n
g
h
ig
h
task
lo
a
d
s
.
I
n
r
ea
l
-
w
o
r
l
d
d
e
p
lo
y
m
e
n
t,
th
e
ef
f
icien
c
y
o
f
th
e
SHI
E
L
D
f
r
am
ewo
r
k
is
n
o
t v
ar
ied
d
u
r
in
g
n
etwo
r
k
laten
cy
f
lu
ctu
atio
n
s
f
o
r
h
ig
h
-
s
p
ee
d
an
d
tim
e
-
ef
f
ici
en
t lo
ad
b
alan
cin
g
.
T
h
e
SHI
E
L
D
f
r
am
ewo
r
k
ac
h
i
ev
es
a
s
u
p
er
io
r
c
o
m
p
u
tatio
n
al
ef
f
icien
cy
o
f
8
3
.
9
to
7
4
.
3
at
2
0
0
0
to
1
0
,
0
0
0
task
s
wh
ich
o
u
tp
e
r
f
o
r
m
s
th
e
ex
is
tin
g
QODA
-
L
B
,
B
est
-
KF
F,
SP
S
O
-
T
C
S,
an
d
VM
MI
SD
s
tr
ateg
ies.
I
n
co
m
p
a
r
is
o
n
,
th
e
SHI
E
L
D
f
r
am
ewo
r
k
g
ain
s
a
c
o
m
p
u
tatio
n
al
ef
f
icien
c
y
o
f
2
0
%
u
n
d
er
lar
g
e
wo
r
k
l
o
ad
s
with
e
x
is
tin
g
tech
n
iq
u
es.
T
h
e
p
r
o
p
o
s
ed
SHI
E
L
D
f
r
am
ewo
r
k
d
em
o
n
s
tr
ates
h
ig
h
er
th
r
o
u
g
h
p
u
t
co
m
p
ar
ed
to
th
e
ex
is
tin
g
tech
n
iq
u
es
ac
r
o
s
s
all
task
s
i
s
s
h
o
wn
in
Fig
u
r
e
1
0
.
Fo
r
task
s
1
0
0
,
2
0
0
,
3
0
0
,
an
d
4
0
0
it
a
ch
iev
es
th
r
o
u
g
h
p
u
t
r
ates
o
f
3
0
0
,
2
9
0
,
2
8
0
,
an
d
2
7
0
task
s
r
esp
ec
tiv
ely
.
T
ab
le
1
p
r
esen
ts
th
e
ex
p
er
im
en
tal
o
u
tco
m
es
o
f
th
e
SHI
E
L
D
f
r
am
ewo
r
k
.
Fig
u
r
e
6
.
An
al
y
s
is
o
f
p
r
iv
ac
y
o
v
er
h
ea
d
Fig
u
r
e
8
.
Po
wer
c
o
n
s
u
m
p
tio
n
o
f
ed
g
e
AI
d
ev
ices
Fig
u
r
e
7
.
Per
f
o
r
m
an
c
e
co
m
p
ar
is
o
n
o
f
AI
m
o
d
els
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
14
,
No
.
4
,
Dec
em
b
er
20
25
:
5
9
1
-
6
0
1
598
Fig
u
r
e
9
.
An
al
y
s
is
o
f
l
aten
cy
Fig
u
r
e
1
0
.
An
aly
s
is
o
f
t
h
r
o
u
g
h
p
u
t
Evaluation Warning : The document was created with Spire.PDF for Python.
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
I
SS
N:
2722
-
2
5
8
6
S
HI
E
LD:
S
ec
u
r
ity
b
a
s
ed
h
yb
r
id
a
u
to
n
o
mo
u
s
d
ee
p
lea
r
n
in
g
n
etw
o
r
k
…
(
Lo
g
a
P
r
iya
d
a
r
s
h
i
n
i Ka
th
ir
ma
la
iya
n
)
599
T
ab
le
1
.
E
x
p
er
im
en
tal
r
esu
lt c
o
m
p
ar
is
o
n
o
n
v
a
r
io
u
s
m
etr
ics
M
e
t
r
i
c
S
H
I
ELD
QODA
-
LB
B
e
st
-
K
F
F
SPSO
-
TC
S
V
M
M
I
S
D
A
c
c
u
r
a
c
y
(
%)
9
3
.
5
%
8
5
%
9
0
%
8
8
%
8
7
%
S
c
a
l
a
b
i
l
i
t
y
(
%)
8
6
%
7
4
%
7
6
%
7
5
%
7
8
%
R
e
s
o
u
r
c
e
u
t
i
l
i
z
a
t
i
o
n
(
%)
8
3
%
7
0
%
7
2
%
7
1
%
7
4
%
P
r
i
v
a
c
y
o
v
e
r
h
e
a
d
(
%)
2
0
%
1
6
%
1
7
%
1
8
%
1
9
%
La
t
e
n
c
y
(
ms)
2
4
9
m
s
3
4
7
m
s
3
3
2
m
s
3
1
4
m
s
2
9
3
m
s
Th
r
o
u
g
h
p
u
t
(
M
b
p
s)
2
7
0
M
b
p
s
2
2
0
M
b
p
s
2
4
0
M
b
p
s
2
3
5
M
b
p
s
2
5
0
M
b
p
s
P
o
w
e
r
c
o
n
su
mp
t
i
o
n
(
W
a
t
t
s)
2
4
.
3
W
a
t
t
s
4
1
.
2
W
a
t
t
s
3
9
.
1
W
a
t
t
s
3
6
.
7
W
a
t
t
s
3
4
.
9
W
a
t
t
s
C
o
m
p
u
t
a
t
i
o
n
a
l
e
f
f
i
c
i
e
n
c
y
7
4
.
3
61
6
4
.
5
63
6
6
.
2
4.
DIS
CU
SS
I
O
N
I
n
th
is
r
esear
ch
,
a
n
o
v
el
SHI
E
L
D
f
r
am
ewo
r
k
was
p
r
o
p
o
s
ed
f
o
r
s
ec
u
r
e
task
s
ch
e
d
u
lin
g
in
clo
u
d
r
eso
u
r
ce
s
.
T
h
e
SHI
E
L
D
f
r
a
m
ewo
r
k
is
ass
es
s
ed
an
d
co
m
p
ar
ed
a
g
ain
s
t
s
ev
er
al
p
ar
am
eter
s
in
clu
d
in
g
s
ca
lab
ilit
y
,
r
eso
u
r
ce
u
tili
za
tio
n
,
p
r
iv
ac
y
o
v
er
h
ea
d
,
laten
c
y
,
th
r
o
u
g
h
p
u
t,
a
n
d
p
o
wer
co
n
s
u
m
p
tio
n
r
esp
ec
tiv
ely
.
T
h
e
co
m
p
ar
is
o
n
o
f
s
ca
lab
ilit
y
f
o
r
th
e
SHI
E
L
D
f
r
am
ewo
r
k
is
s
h
o
wn
in
Fig
u
r
e
4
.
I
n
lar
g
e
s
ca
le
I
o
T
wo
r
k
lo
ad
s
,
T
h
e
SHI
E
L
D
f
r
am
ewo
r
k
o
u
tp
er
f
o
r
m
s
e
x
is
tin
g
QODA
-
L
B
,
B
est
-
KF
F,
SP
SO
-
T
C
S,
an
d
V
MM
I
SD
ap
p
r
o
ac
h
es
b
y
m
ai
n
tain
in
g
th
e
h
ig
h
est
s
co
r
es
f
r
o
m
9
3
%
at
2
0
0
0
t
ask
s
to
8
6
%
at
1
0
,
0
0
0
task
s
r
esp
ec
tiv
ely
.
T
h
e
co
m
p
ar
is
o
n
o
f
laten
c
y
f
o
r
th
e
SHI
E
L
D
f
r
am
ewo
r
k
is
s
h
o
w
n
in
Fig
u
r
e
9
.
Du
r
in
g
lar
g
e
I
o
T
wo
r
k
l
o
ad
s
,
th
e
SHI
E
L
D
f
r
am
ewo
r
k
a
ch
iev
es
th
e
lo
west
laten
cy
o
f
9
7
m
s
f
o
r
2
,
0
0
0
task
s
to
2
4
9
m
s
f
o
r
1
0
,
0
0
0
task
s
f
o
r
h
an
d
lin
g
h
ig
h
task
lo
ad
s
.
Fig
u
r
e
7
d
e
p
icts
th
e
p
er
f
o
r
m
an
ce
a
s
s
es
s
m
en
t
o
f
th
e
SHI
E
L
D
m
o
d
el
with
p
r
ev
i
o
u
s
ly
d
ev
elo
p
e
d
AI
m
o
d
els.
T
h
e
SHI
E
L
D
m
eth
o
d
ac
h
iev
es
d
ec
e
n
t
r
alize
d
s
ec
u
r
ity
o
f
9
7
%,
ad
ap
ti
v
e
task
s
ch
ed
u
lin
g
o
f
9
2
%,
an
d
s
eq
u
en
ce
p
r
ed
i
ctio
n
o
f
9
4
%
wh
ich
o
u
tp
er
f
o
r
m
s
th
e
ex
is
tin
g
Hy
b
r
id
DQN
-
PPO
-
GNN
-
R
L
,
HAFed
R
L
,
an
d
T
F
-
DD
R
L
tec
h
n
iq
u
es
r
esp
ec
tiv
ely
.
T
h
e
co
m
p
ar
is
o
n
o
f
p
o
wer
co
n
s
u
m
p
ti
o
n
f
o
r
th
e
SHI
E
L
D
f
r
am
ewo
r
k
is
r
ep
r
esen
ted
in
Fig
u
r
e
3
(
b
)
.
At
h
ig
h
er
wo
r
k
lo
ad
s
,
th
e
SHI
E
L
D
f
r
am
ewo
r
k
ac
h
iev
es
th
e
lo
west
p
o
wer
c
o
n
s
u
m
p
tio
n
o
f
1
8
.
2
W
to
2
4
.
3
W
ac
r
o
s
s
all
task
r
an
g
es
f
r
o
m
2
,
0
0
0
to
1
0
,
0
0
0
ag
ain
s
t
th
e
ex
is
tin
g
QODA
-
L
B
,
B
e
s
t
-
KFF,
S
PS
O
-
T
C
S,
an
d
VM
MI
SD
tech
n
iq
u
es
r
esp
ec
tiv
ely
.
T
h
e
c
o
m
p
ar
is
o
n
o
f
p
o
wer
co
n
s
u
m
p
tio
n
o
f
E
d
g
e
AI
De
v
ices
f
o
r
en
er
g
y
ef
f
icien
t
lo
a
d
b
alan
cin
g
is
d
ep
icted
in
Fi
g
u
r
e
8
.
T
h
e
E
SP
3
2
d
ev
ice
attain
s
th
e
lo
west
p
o
w
er
co
n
s
u
m
p
tio
n
o
f
0
.
0
8
W
,
0
.
2
5
W
,
an
d
0
.
4
W
f
o
r
id
le
p
o
wer
,
av
er
ag
e
p
o
wer
,
an
d
m
a
x
im
u
m
p
o
we
r
f
o
r
SHI
E
L
D
r
esp
ec
tiv
ely
.
I
n
r
ea
l
-
wo
r
ld
d
ep
l
o
y
m
en
t,
th
e
SHI
E
L
D
m
o
d
el
will
p
r
o
v
i
d
e
s
u
p
er
io
r
p
e
r
f
o
r
m
an
ce
u
n
d
er
h
eter
o
g
en
eo
u
s
I
o
T
d
ev
ices f
o
r
r
o
b
u
s
t a
n
d
s
ca
lab
le
lo
a
d
b
alan
c
in
g
.
5.
CO
NCLU
SI
O
N
I
n
th
is
p
a
p
er
,
a
n
o
v
el
SHI
E
L
D
f
r
am
ewo
r
k
is
d
e
v
elo
p
ed
f
o
r
s
ec
u
r
e
task
s
ch
e
d
u
lin
g
in
I
o
T
d
ev
ices
u
s
in
g
h
y
b
r
i
d
DL
n
etwo
r
k
s
.
T
h
e
SHI
E
L
D
f
r
am
ewo
r
k
is
s
im
u
lated
b
y
u
s
in
g
C
lo
u
d
Sim
7
G.
A
co
m
p
ar
ativ
e
an
aly
s
is
was
co
n
d
u
cted
with
th
e
ex
is
tin
g
tech
n
iq
u
es
s
u
ch
as
r
eso
u
r
ce
u
tili
za
tio
n
,
laten
cy
,
th
r
o
u
g
h
p
u
t,
p
r
iv
ac
y
o
v
er
h
ea
d
,
an
d
s
ca
lab
ilit
y
.
I
n
co
m
p
ar
is
o
n
,
th
e
SHI
E
L
D
f
r
am
ewo
r
k
en
s
u
r
es
a
p
r
iv
ac
y
o
v
er
h
ea
d
o
f
1
4
%,
wh
er
ea
s
th
e
QODA
-
L
B
,
B
e
s
t
-
KFF,
SP
SO
-
T
C
S,
an
d
VM
M
I
SD
tech
n
iq
u
es
ac
h
iev
e
o
v
e
r
h
ea
d
s
o
f
1
0
%,
1
1
%,
1
2
%,
an
d
1
3
%,
r
esp
ec
tiv
ely
.
I
n
th
e
f
u
tu
r
e
,
th
e
SHI
E
L
D
f
r
am
ewo
r
k
will
in
teg
r
ate
an
ad
v
an
ce
d
b
lo
c
k
ch
ain
m
eth
o
d
o
l
o
g
y
t
o
h
in
d
er
th
e
m
o
s
t
v
u
ln
er
ab
le
a
d
v
er
s
ar
ial
att
ac
k
s
d
u
r
in
g
lo
ad
b
alan
cin
g
.
T
h
is
in
teg
r
atio
n
will
en
h
an
ce
th
e
s
ec
u
r
ity
an
d
p
r
i
v
ac
y
o
f
I
o
T
-
clo
u
d
in
ter
ac
tio
n
s
f
o
r
s
ec
u
r
e
lo
a
d
b
alan
ci
n
g
.
Als
o
,
th
e
SHI
E
L
D
f
r
am
ewo
r
k
will
b
e
d
ep
lo
y
ed
in
Go
o
g
le
C
lo
u
d
,
a
r
ea
l
-
wo
r
l
d
clo
u
d
p
latf
o
r
m
t
h
at
d
y
n
am
ically
ad
ju
s
ts
task
s
ch
ed
u
lin
g
b
ased
o
n
c
h
an
g
i
n
g
I
o
T
en
v
ir
o
n
m
e
n
ts
an
d
r
eso
u
r
c
e
av
ailab
ilit
y
.
ACK
NO
WL
E
DG
M
E
N
T
S
T
h
e
au
t
h
o
r
wo
u
ld
lik
e
to
ex
p
r
ess
h
is
h
ea
r
tf
elt
g
r
atitu
d
e
t
o
th
e
s
u
p
er
v
is
o
r
f
o
r
h
is
g
u
i
d
an
ce
a
n
d
u
n
wav
er
in
g
s
u
p
p
o
r
t d
u
r
in
g
t
h
is
r
esear
ch
f
o
r
h
is
g
u
id
an
ce
an
d
s
u
p
p
o
r
t.
F
UNDING
I
NF
O
R
M
A
T
I
O
N
No
t a
p
p
licab
le.
AUTHO
R
CO
NT
RI
B
UT
I
O
NS ST
A
T
E
M
E
N
T
T
h
is
jo
u
r
n
al
u
s
es
th
e
C
o
n
tr
ib
u
to
r
R
o
les
T
ax
o
n
o
m
y
(
C
R
ed
iT)
to
r
ec
o
g
n
ize
in
d
iv
id
u
al
au
th
o
r
co
n
tr
ib
u
tio
n
s
,
r
ed
u
ce
au
th
o
r
s
h
ip
d
is
p
u
tes,
an
d
f
ac
ilit
ate
co
llab
o
r
atio
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
7
2
2
-
2
5
8
6
I
AE
S
I
n
t
J
R
o
b
&
A
u
to
m
,
Vo
l
.
14
,
No
.
4
,
Dec
em
b
er
20
25
:
5
9
1
-
6
0
1
600
Na
m
e
o
f
Aut
ho
r
C
M
So
Va
Fo
I
R
D
O
E
Vi
Su
P
Fu
L
o
g
a
Priy
ad
ar
s
h
in
i
Kath
ir
m
alaiy
an
✓
✓
✓
✓
✓
✓
✓
Nith
y
a
Mu
th
u
✓
✓
✓
✓
✓
✓
C
:
C
o
n
c
e
p
t
u
a
l
i
z
a
t
i
o
n
M
:
M
e
t
h
o
d
o
l
o
g
y
So
:
So
f
t
w
a
r
e
Va
:
Va
l
i
d
a
t
i
o
n
Fo
:
Fo
r
mal
a
n
a
l
y
s
i
s
I
:
I
n
v
e
s
t
i
g
a
t
i
o
n
R
:
R
e
so
u
r
c
e
s
D
:
D
a
t
a
C
u
r
a
t
i
o
n
O
:
W
r
i
t
i
n
g
-
O
r
i
g
i
n
a
l
D
r
a
f
t
E
:
W
r
i
t
i
n
g
-
R
e
v
i
e
w
&
E
d
i
t
i
n
g
Vi
:
Vi
su
a
l
i
z
a
t
i
o
n
Su
:
Su
p
e
r
v
i
s
i
o
n
P
:
P
r
o
j
e
c
t
a
d
mi
n
i
st
r
a
t
i
o
n
Fu
:
Fu
n
d
i
n
g
a
c
q
u
i
si
t
i
o
n
CO
NF
L
I
C
T
O
F
I
N
T
E
R
E
S
T
ST
A
T
E
M
E
NT
T
h
e
au
t
h
o
r
s
d
ec
lar
e
th
at
th
e
y
h
av
e
n
o
k
n
o
wn
c
o
m
p
etin
g
f
in
an
cial
in
ter
ests
o
r
p
er
s
o
n
al
r
el
atio
n
s
h
ip
s
th
at
co
u
ld
h
av
e
ap
p
ea
r
ed
t
o
in
f
lu
en
ce
th
e
wo
r
k
r
e
p
o
r
te
d
in
t
h
is
p
ap
er
.
I
NF
O
RM
E
D
CO
NS
E
N
T
I
ce
r
tify
th
at
I
h
a
v
e
ex
p
lain
e
d
th
e
n
atu
r
e
an
d
p
u
r
p
o
s
e
o
f
th
i
s
s
tu
d
y
to
th
e
ab
o
v
e
-
n
am
ed
i
n
d
iv
id
u
al,
an
d
I
h
av
e
d
is
cu
s
s
ed
th
e
p
o
te
n
tial
b
en
ef
its
o
f
th
is
s
tu
d
y
p
ar
ticip
atio
n
.
T
h
e
q
u
esti
o
n
s
th
e
i
n
d
iv
id
u
al
h
ad
ab
o
u
t
th
is
s
tu
d
y
h
av
e
b
ee
n
an
s
wer
ed
,
an
d
we
will a
lway
s
b
e
av
aila
b
le
to
ad
d
r
ess
f
u
tu
r
e
q
u
esti
o
n
s
.
E
T
H
I
CAL AP
P
RO
V
AL
My
r
esear
ch
g
u
id
e
r
e
v
iewe
d
a
n
d
eth
ically
ap
p
r
o
v
ed
th
is
m
a
n
u
s
cr
ip
t f
o
r
p
u
b
lis
h
in
g
in
th
is
j
o
u
r
n
al.
DATA AV
AI
L
AB
I
L
I
T
Y
Data
s
h
ar
in
g
is
n
o
t
ap
p
licab
l
e
to
th
is
ar
ticle
as
n
o
d
atasets
wer
e
g
en
er
ated
o
r
an
aly
ze
d
d
u
r
in
g
th
e
cu
r
r
en
t stu
d
y
.
RE
F
E
R
E
NC
E
S
[
1
]
M
.
S
h
u
a
i
b
e
t
a
l
.
,
“
A
n
o
p
t
i
mi
z
e
d
,
d
y
n
a
m
i
c
,
a
n
d
e
f
f
i
c
i
e
n
t
l
o
a
d
-
b
a
l
a
n
c
i
n
g
f
r
a
mew
o
r
k
f
o
r
r
e
so
u
r
c
e
ma
n
a
g
e
m
e
n
t
i
n
t
h
e
i
n
t
e
r
n
e
t
o
f
t
h
i
n
g
s (I
o
T)
e
n
v
i
r
o
n
m
e
n
t
,
”
E
l
e
c
t
r
o
n
i
c
s
,
v
o
l
.
1
2
,
n
o
.
5
,
p
.
1
1
0
4
,
F
e
b
.
2
0
2
3
,
d
o
i
:
1
0
.
3
3
9
0
/
e
l
e
c
t
r
o
n
i
c
s
1
2
0
5
1
1
0
4
.
[
2
]
N
.
M
a
z
u
m
d
a
r
,
A
.
N
a
g
,
a
n
d
J.
P
.
S
i
n
g
h
,
“
Tr
u
st
-
b
a
s
e
d
l
o
a
d
-
o
f
f
l
o
a
d
i
n
g
p
r
o
t
o
c
o
l
t
o
r
e
d
u
c
e
serv
i
c
e
d
e
l
a
y
s
i
n
f
o
g
-
c
o
mp
u
t
i
n
g
-
e
mp
o
w
e
r
e
d
I
o
T,
”
C
o
m
p
u
t
e
r
s &
El
e
c
t
ri
c
a
l
En
g
i
n
e
e
ri
n
g
,
v
o
l
.
9
3
,
p
.
1
0
7
2
2
3
,
Ju
l
.
2
0
2
1
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
c
o
m
p
e
l
e
c
e
n
g
.
2
0
2
1
.
1
0
7
2
2
3
.
[
3
]
G
.
Th
a
h
n
i
y
a
t
h
a
n
d
M
.
Ja
y
a
p
r
a
sa
d
,
“
S
e
c
u
r
e
a
n
d
l
o
a
d
b
a
l
a
n
c
e
d
r
o
u
t
i
n
g
mo
d
e
l
f
o
r
w
i
r
e
l
e
ss
se
n
so
r
n
e
t
w
o
r
k
s,
”
J
o
u
r
n
a
l
o
f
K
i
n
g
S
a
u
d
U
n
i
v
e
rsi
t
y
-
C
o
m
p
u
t
e
r
a
n
d
I
n
f
o
rm
a
t
i
o
n
S
c
i
e
n
c
e
s
,
v
o
l
.
3
4
,
n
o
.
7
,
p
p
.
4
2
0
9
–
4
2
1
8
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
1
6
/
j
.
j
k
s
u
c
i
.
2
0
2
0
.
1
0
.
0
1
2
.
[
4
]
D
.
K
.
S
a
h
,
T.
N
.
N
g
u
y
e
n
,
K
.
C
e
n
g
i
z
,
B
.
D
u
mb
a
,
a
n
d
V
.
K
u
m
a
r
,
“
L
o
a
d
-
b
a
l
a
n
c
e
sc
h
e
d
u
l
i
n
g
f
o
r
i
n
t
e
l
l
i
g
e
n
t
se
n
so
r
s
d
e
p
l
o
y
me
n
t
i
n
i
n
d
u
st
r
i
a
l
i
n
t
e
r
n
e
t
o
f
t
h
i
n
g
s,
”
C
l
u
st
e
r
C
o
m
p
u
t
i
n
g
,
v
o
l
.
2
5
,
n
o
.
3
,
p
p
.
1
7
1
5
–
1
7
2
7
,
2
0
2
2
,
d
o
i
:
1
0
.
1
0
0
7
/
s
1
0
5
8
6
-
021
-
0
3
3
1
6
-
1.
[
5
]
N
.
A
g
r
a
w
a
l
,
“
D
y
n
a
mi
c
l
o
a
d
b
a
l
a
n
c
i
n
g
a
ss
i
st
e
d
o
p
t
i
mi
z
e
d
a
c
c
e
ss
c
o
n
t
r
o
l
mec
h
a
n
i
sm
f
o
r
e
d
g
e
-
f
o
g
-
c
l
o
u
d
n
e
t
w
o
r
k
i
n
I
n
t
e
r
n
e
t
o
f
Th
i
n
g
s
e
n
v
i
r
o
n
me
n
t
,
”
C
o
n
c
u
rre
n
c
y
a
n
d
C
o
m
p
u
t
a
t
i
o
n
:
Pr
a
c
t
i
c
e
a
n
d
Ex
p
e
ri
e
n
c
e
,
v
o
l
.
3
3
,
n
o
.
2
1
,
p
.
6
4
4
0
,
2
0
2
1
,
d
o
i
:
1
0
.
1
0
0
2
/
c
p
e
.
6
4
4
0
.
[
6
]
S
.
F
u
g
k
e
a
w
,
R
.
P
r
a
sa
d
G
u
p
t
a
,
a
n
d
K
.
W
o
r
a
p
a
l
u
k
,
“
S
e
c
u
r
e
a
n
d
f
i
n
e
-
g
r
a
i
n
e
d
a
c
c
e
ss
c
o
n
t
r
o
l
w
i
t
h
o
p
t
i
mi
z
e
d
r
e
v
o
c
a
t
i
o
n
f
o
r
o
u
t
s
o
u
r
c
e
d
I
o
T
EH
R
s
w
i
t
h
a
d
a
p
t
i
v
e
l
o
a
d
-
s
h
a
r
i
n
g
i
n
f
o
g
-
a
ssi
s
t
e
d
c
l
o
u
d
e
n
v
i
r
o
n
m
e
n
t
,
”
I
EEE
A
c
c
e
ss
,
v
o
l
.
1
2
,
p
p
.
8
2
7
5
3
–
8
2
7
6
8
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
A
C
C
ESS
.
2
0
2
4
.
3
4
1
2
7
5
4
.
[
7
]
V
.
V
e
e
r
a
m
a
c
h
a
n
e
n
i
,
“
En
h
a
n
c
i
n
g
I
o
T
sec
u
r
i
t
y
w
i
t
h
d
y
n
a
m
i
c
r
e
p
u
t
a
t
i
o
n
-
b
a
s
e
d
t
r
u
s
t
ma
n
a
g
e
me
n
t
,
”
R
e
c
e
n
t
I
n
n
o
v
a
t
i
o
n
s
i
n
W
i
re
l
e
ss
N
e
t
w
o
rk
S
e
c
u
ri
t
y
,
v
o
l
.
7
,
n
o
.
1
,
p
p
.
2
7
–
4
4
,
2
0
2
5
.
[
8
]
K
.
D
h
a
n
u
s
h
k
o
d
i
,
R
.
K
u
mar,
P
.
M
i
t
t
a
l
,
S
.
S
.
D
a
s,
N
.
N
.
S
.
S
u
r
y
a
v
e
n
u
,
a
n
d
K
.
V
e
n
k
a
t
a
r
a
m
a
n
i
,
“
En
h
a
n
c
i
n
g
r
e
s
o
u
r
c
e
u
t
i
l
i
z
a
t
i
o
n
a
n
d
p
r
i
v
a
c
y
i
n
I
o
T
d
a
t
a
p
l
a
c
e
me
n
t
t
h
r
o
u
g
h
f
u
z
z
y
l
o
g
i
c
a
n
d
P
S
O
o
p
t
i
m
i
z
a
t
i
o
n
,
”
C
l
u
s
t
e
r
C
o
m
p
u
t
i
n
g
,
v
o
l
.
2
7
,
n
o
.
9
,
p
p
.
1
2
6
0
3
–
1
2
6
2
6
,
2
0
2
4
,
d
o
i
:
1
0
.
1
0
0
7
/
s1
0
5
8
6
-
0
2
4
-
0
4
5
4
2
-
z.
[
9
]
M
.
Jes
i
,
A
.
A
p
p
a
t
h
u
r
a
i
,
M
.
K
u
mar
a
n
,
a
n
d
A
.
K
u
m
a
r
,
“
L
o
a
d
b
a
l
a
n
c
i
n
g
i
n
c
l
o
u
d
c
o
mp
u
t
i
n
g
v
i
a
ma
y
f
l
y
o
p
t
i
mi
z
a
t
i
o
n
a
l
g
o
r
i
t
h
m,
”
Re
v
u
e
R
o
u
m
a
i
n
e
d
e
s
S
c
i
e
n
c
e
s
T
e
c
h
n
i
q
u
e
s
S
e
ri
e
El
e
c
t
ro
t
e
c
h
n
i
q
u
e
e
t
En
e
r
g
e
t
i
q
u
e
,
v
o
l
.
6
9
,
n
o
.
1
,
p
p
.
7
9
–
8
4
,
2
0
2
4
,
d
o
i
:
1
0
.
5
9
2
7
7
/
R
R
S
T
-
EE.
2
0
2
4
.
1
.
1
4
.
[
1
0
]
D
.
C
h
a
mp
l
a
,
S
.
G
.
A
l
i
,
B
.
M
u
t
h
u
k
u
mar,
a
n
d
M
.
M
.
S
i
t
h
i
k
,
“
C
-
A
V
P
S
O
:
d
y
n
a
mi
c
l
o
a
d
b
a
l
a
n
c
i
n
g
u
si
n
g
A
f
r
i
c
a
n
v
u
l
t
u
r
e
p
a
r
t
i
c
l
e
sw
a
r
m o
p
t
i
mi
z
a
t
i
o
n
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
D
a
t
a
S
c
i
e
n
c
e
a
n
d
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
1
,
n
o
.
2
,
p
p
.
1
–
9
,
2
0
2
3
.
[
1
1
]
G
.
S
r
e
e
t
h
a
a
n
d
G
.
S
a
n
t
h
i
y
a
,
“
S
A
F
E
-
A
C
I
D
:
a
n
o
v
e
l
sec
u
r
i
t
y
a
ssess
me
n
t
f
r
a
mew
o
r
k
f
o
r
I
o
T
i
n
t
r
u
si
o
n
d
e
t
e
c
t
i
o
n
v
i
a
d
e
e
p
l
e
a
r
n
i
n
g
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
D
a
t
a
S
c
i
e
n
c
e
a
n
d
Art
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
,
v
o
l
.
2
,
n
o
.
1
,
p
p
.
2
0
–
2
6
,
2
0
2
4
.
[
1
2
]
S
.
R
.
A
d
d
u
l
a
e
t
a
l
.
,
“
D
y
n
a
mi
c
l
o
a
d
b
a
l
a
n
c
i
n
g
i
n
c
l
o
u
d
c
o
mp
u
t
i
n
g
u
s
i
n
g
h
y
b
r
i
d
k
o
o
k
a
b
u
r
r
a
-
p
e
l
i
c
a
n
o
p
t
i
mi
z
a
t
i
o
n
a
l
g
o
r
i
t
h
ms
,
”
2
0
2
4
I
n
t
e
r
n
a
t
i
o
n
a
l
C
o
n
f
e
r
e
n
c
e
o
n
A
u
g
m
e
n
t
e
d
R
e
a
l
i
t
y
,
I
n
t
e
l
l
i
g
e
n
t
S
y
st
e
m
s,
a
n
d
I
n
d
u
st
r
i
a
l
Au
t
o
m
a
t
i
o
n
,
ARII
A
2
0
2
4
,
v
o
l
.
2
,
n
o
.
4
,
p
p
.
98
–
1
0
4
,
2
0
2
4
,
d
o
i
:
1
0
.
1
1
0
9
/
A
R
I
I
A
6
3
3
4
5
.
2
0
2
4
.
1
1
0
5
1
8
9
3
.
[
1
3
]
C
.
S
.
S
i
n
g
h
,
S
.
N
a
a
z
,
a
n
d
G
.
S
a
r
a
n
y
a
,
“
I
o
T
c
e
n
t
r
i
c
d
a
t
a
p
r
o
t
e
c
t
i
o
n
u
si
n
g
d
e
e
p
l
e
a
r
n
i
n
g
t
e
c
h
n
i
q
u
e
f
o
r
p
r
e
serv
i
n
g
se
c
u
r
i
t
y
a
n
d
p
r
i
v
a
c
y
i
n
c
l
o
u
d
,
”
I
n
t
e
r
n
a
t
i
o
n
a
l
J
o
u
r
n
a
l
o
f
D
a
t
a
S
c
i
e
n
c
e
a
n
d
Ar
t
i
f
i
c
i
a
l
I
n
t
e
l
l
i
g
e
n
c
e
(
I
J
D
S
AI
)
,
2
0
2
4
.
[
1
4
]
A
.
A
l
w
h
e
l
a
t
,
O
.
A
.
F
a
d
a
r
e
,
F
.
A
l
-
T
u
r
j
man
,
a
n
d
L
.
I
b
r
a
h
i
m
,
“
A
d
a
p
t
i
v
e
l
o
a
d
b
a
l
a
n
c
i
n
g
i
n
c
l
o
u
d
c
o
mp
u
t
i
n
g
u
si
n
g
d
e
e
p
d
e
t
e
r
m
i
n
i
st
i
c
Evaluation Warning : The document was created with Spire.PDF for Python.