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.
9
,
No
.
3
,
J
u
n
e
201
9
,
p
p
.
1861
~
1
8
7
0
I
SS
N:
2088
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v
9
i
3
.
p
p
1
8
6
1
-
1870
1861
J
o
ur
na
l
ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
co
r
e
.
co
m/
jo
u
r
n
a
ls
/in
d
ex
.
p
h
p
/
I
JE
C
E
Bla
ck
-
bo
x
m
o
deli
ng
of nonlin
ea
r sy
ste
m
using
evo
luti
o
na
ry
neura
l NA
R
X
m
o
del
Ng
uy
en
Ng
o
c
So
n
,
Ng
uy
en
Duy
K
ha
nh
, T
ra
n M
inh
Chi
nh
F
a
c
u
lt
y
o
f
El
e
c
tro
n
ics
T
e
c
h
n
o
lo
g
y
,
In
d
u
strial
Un
iv
e
rsity
o
f
Ho
Ch
i
M
in
h
Cit
y
,
V
iet
Na
m
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ma
y
1
2
,
2
0
1
8
R
ev
i
s
ed
Dec
1
8
,
2
0
1
8
A
cc
ep
ted
Dec
2
9
,
2
0
1
8
No
n
li
n
e
a
r
s
y
ste
m
s
w
it
h
u
n
c
e
rtain
ty
a
n
d
d
istu
rb
a
n
c
e
a
re
v
e
r
y
d
if
f
icu
lt
to
m
o
d
e
l
u
sin
g
m
a
th
e
m
a
ti
c
a
p
p
ro
a
c
h
.
T
h
e
re
f
o
re
,
a
b
lac
k
-
b
o
x
m
o
d
e
li
n
g
a
p
p
ro
a
c
h
w
it
h
o
u
t
a
n
y
p
rio
r
k
n
o
w
led
g
e
is
n
e
c
e
ss
a
r
y
.
T
h
e
re
a
re
so
m
e
m
o
d
e
li
n
g
a
p
p
ro
a
c
h
e
s
h
a
v
e
b
e
e
n
u
se
d
to
d
e
v
e
lo
p
a
b
lac
k
b
o
x
m
o
d
e
l
su
c
h
a
s
f
u
z
z
y
lo
g
ic,
n
e
u
ra
l
n
e
tw
o
rk
,
a
n
d
e
v
o
lu
t
io
n
a
lg
o
rit
h
m
s.
In
th
is
p
a
p
e
r,
a
n
e
v
o
lu
ti
o
n
a
ry
n
e
u
ra
l
n
e
tw
o
rk
b
y
c
o
m
b
in
in
g
a
n
e
u
ra
l
n
e
tw
o
rk
a
n
d
a
m
o
d
if
ied
d
if
fe
re
n
ti
a
l
e
v
o
lu
ti
o
n
a
lg
o
rit
h
m
is
a
p
p
li
e
d
to
m
o
d
e
l
a
n
o
n
li
n
e
a
r
sy
ste
m
.
T
h
e
f
e
a
sib
il
it
y
a
n
d
e
ffe
c
ti
v
e
n
e
ss
o
f
th
e
p
ro
p
o
se
d
m
o
d
e
li
n
g
a
re
te
ste
d
o
n
a
p
iez
o
e
lec
tri
c
a
c
tu
a
to
r
S
IS
O
sy
s
tem
a
n
d
a
n
e
x
p
e
ri
m
e
n
tal
q
u
a
d
r
u
p
le
tan
k
M
IM
O sy
ste
m
.
K
ey
w
o
r
d
s
:
Dif
f
er
en
t
ial
ev
o
l
u
tio
n
E
v
o
lu
tio
n
ar
y
n
eu
r
al
n
et
w
o
r
k
s
No
n
li
n
ea
r
s
y
s
te
m
id
en
ti
f
icatio
n
P
iezo
elec
tr
ic
ac
tu
ato
r
Qu
ad
r
u
p
le
tan
k
s
y
s
te
m
Co
p
y
rig
h
t
©
2
0
1
9
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts
re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
Ng
u
y
e
n
N
g
o
c
So
n
,
Facu
lt
y
o
f
E
lectr
o
n
ics T
ec
h
n
o
lo
g
y
,
I
n
d
u
s
tr
ial
U
n
i
v
er
s
i
t
y
o
f
Ho
C
h
i M
in
h
C
it
y
,
1
2
Ng
u
y
e
n
Va
n
B
ao
Stre
et,
4
W
ar
d
,
Go
Vap
Dis
tr
ict,
Ho
C
h
i M
in
h
C
it
y
,
Viet
Na
m
.
E
m
ail:
n
g
u
y
en
n
g
o
cso
n
@
i
u
h
.
e
d
u
.
v
n
1.
I
NT
RO
D
UCT
I
O
N
Ma
th
e
m
atica
l
m
o
d
elin
g
o
f
s
y
s
te
m
s
i
s
a
v
er
y
co
m
m
o
n
m
et
h
o
d
o
lo
g
y
in
e
n
g
i
n
ee
r
i
n
g
.
I
t
is
u
s
ed
b
o
th
as
a
m
ea
n
s
f
o
r
ac
h
ie
v
in
g
d
ee
p
er
k
n
o
w
led
g
e
ab
o
u
t
a
s
y
s
te
m
a
n
d
as
a
b
asis
f
o
r
s
i
m
u
la
tio
n
s
o
r
f
o
r
th
e
d
es
ig
n
o
f
co
n
tr
o
ller
s
.
Ho
w
e
v
er
,
in
p
r
ac
tice,
it
is
n
o
t
ea
s
y
t
o
o
b
tain
an
ac
cu
r
ate
m
at
h
e
m
atica
l
m
o
d
el
o
f
a
n
o
n
lin
ea
r
s
y
s
te
m
b
ec
au
s
e
o
f
a
lack
o
f
k
n
o
w
led
g
e
o
f
t
h
e
s
y
s
te
m
an
d
d
i
s
tu
r
b
an
ce
i
m
p
ac
t
to
s
y
s
te
m
m
a
y
b
e
s
til
l
u
n
k
n
o
w
n
.
I
n
th
ese
ca
s
e
s
,
s
y
s
te
m
id
e
n
ti
f
icatio
n
ca
n
b
e
a
w
a
y
o
f
s
o
l
v
i
n
g
t
h
e
m
o
d
elin
g
p
r
o
b
lem
.
S
y
s
te
m
id
en
ti
f
icatio
n
d
ea
ls
w
it
h
th
e
p
r
o
b
le
m
o
f
h
o
w
to
esti
m
ate
a
m
o
d
el
o
f
a
s
y
s
te
m
f
r
o
m
m
ea
s
u
r
ed
in
p
u
t
a
n
d
o
u
tp
u
t
s
i
g
n
a
ls
.
B
lack
-
B
o
x
n
o
n
lin
ea
r
m
o
d
eli
n
g
ap
p
r
o
ac
h
es
i
n
cl
u
d
e
n
o
n
-
p
ar
a
m
etr
ic
m
et
h
o
d
s
,
s
u
c
h
as
n
e
u
r
al
n
et
w
o
r
k
s
,
f
u
zz
y
lo
g
ic,
g
e
n
etic
al
g
o
r
ith
m
etc.
,
[
1
]
w
h
ic
h
d
o
n
o
t r
eq
u
ir
e
a
p
r
io
r
i
k
n
o
w
led
g
e.
R
ec
en
t
l
y
,
t
h
e
m
u
lti
la
y
er
f
ee
d
-
f
o
r
w
ar
d
n
e
u
r
al
n
et
w
o
r
k
(
M
L
P
)
h
as
b
ee
n
w
id
el
y
u
s
ed
in
n
o
n
li
n
ea
r
s
y
s
te
m
id
e
n
ti
f
icatio
n
.
A
s
w
e
k
n
o
w
,
t
h
e
p
er
f
o
r
m
an
ce
o
f
a
n
e
u
r
al
n
et
w
o
r
k
is
d
ep
en
d
en
t
o
n
th
e
tr
ain
in
g
p
r
o
ce
s
s
.
A
p
o
p
u
lar
tr
ai
n
i
n
g
a
lg
o
r
it
h
m
is
t
h
e
b
ac
k
-
p
r
o
p
ag
atio
n
.
Ho
wev
er
,
th
e
b
ac
k
-
p
r
o
p
ag
atio
n
(
B
P
)
alg
o
r
ith
m
o
n
l
y
p
er
f
o
r
m
a
lo
ca
l
s
ea
r
ch
ar
o
u
n
d
th
e
in
itia
l
v
al
u
es
an
d
p
r
o
v
id
e
lo
ca
l
o
p
tim
izatio
n
s
[
2
]
,
[
3
]
.
T
h
er
ef
o
r
e
th
e
m
e
ta
-
h
eu
r
i
s
tic
al
g
o
r
it
h
m
s
b
ased
tr
a
in
i
n
g
p
r
o
ce
d
u
r
es
ar
e
co
n
s
id
er
ed
as
p
r
o
m
is
i
n
g
alter
n
ati
v
es.
T
h
e
m
eta
-
h
e
u
r
is
t
ic
alg
o
r
ith
m
s
g
e
n
er
ate
g
lo
b
al
o
p
ti
m
u
m
b
ec
au
s
e
o
f
t
h
eir
ab
ilit
y
to
s
ea
r
ch
t
h
e
g
lo
b
al
s
o
lu
t
io
n
s
p
ac
e
an
d
av
o
id
i
n
g
f
alli
n
g
in
to
a
lo
ca
ll
y
o
p
tim
a
l
s
o
lu
tio
n
.
Fo
r
ex
a
m
p
le,
a
g
e
n
etic
alg
o
r
it
h
m
[
4
]
-
[
6
]
is
u
s
ed
f
o
r
tu
r
n
in
g
th
e
s
tr
u
ct
u
r
e
an
d
p
ar
a
m
eter
s
o
f
n
e
u
r
al
n
e
t
w
o
r
k
s
.
P
ap
er
s
[
7
]
-
[
9
]
in
tr
o
d
u
ce
d
th
e
p
ar
ticle
s
w
ar
m
o
p
tim
izatio
n
(
P
SO)
to
tr
ain
th
e
n
e
u
r
al
n
et
w
o
r
k
m
o
d
el.
Valia
n
et
a
l
.
[
1
0
]
p
r
o
p
o
s
ed
th
e
i
m
p
r
o
v
ed
cu
ck
o
o
s
ea
r
ch
alg
o
r
ith
m
f
o
r
tr
ain
i
n
g
f
ee
d
-
f
o
r
w
ar
d
n
e
u
r
al
n
et
w
o
r
k
f
o
r
t
w
o
b
en
c
h
m
ar
k
c
lass
i
f
icatio
n
p
r
o
b
le
m
s
.
A
l
th
o
u
g
h
t
h
ese
p
r
o
p
o
s
ed
m
et
h
o
d
s
o
b
tain
ed
g
o
o
d
r
esu
lt
s
,
t
w
o
c
h
alle
n
g
e
s
t
h
at
n
ee
d
to
b
e
f
u
r
th
er
i
m
p
r
o
v
ed
in
t
r
ai
n
i
n
g
n
e
u
r
al
n
et
w
o
r
k
ar
e
h
o
w
to
f
i
n
d
th
e
g
lo
b
al
o
p
tim
al
s
o
l
u
tio
n
a
n
d
h
o
w
to
ac
h
ie
v
e
a
f
as
t c
o
n
v
er
g
e
n
ce
s
p
ee
d
.
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.
9
,
No
.
3
,
J
u
n
e
201
9
:
1
8
6
1
-
1
8
7
0
1862
Am
o
n
g
m
eta
-
h
eu
r
i
s
tic
al
g
o
r
ith
m
s
,
t
h
e
d
i
f
f
er
e
n
tial
e
v
o
l
u
ti
o
n
(
DE
)
tech
n
iq
u
e
w
as
r
e
g
a
r
d
ed
as
a
p
r
o
m
i
s
in
g
s
to
c
h
asti
c
g
lo
b
al
o
p
ti
m
izatio
n
alg
o
r
ith
m
.
T
h
e
DE
alg
o
r
ith
m
w
a
s
f
ir
s
t
s
tu
d
ied
an
d
p
u
b
lis
h
ed
b
y
Sto
r
n
R
.
a
n
d
P
r
ice
K.
V.
[
1
1
]
.
I
ts
ad
v
an
ta
g
es
ar
e
as
f
o
llo
w
s
:
t
h
e
s
i
m
p
le
an
d
s
tr
ai
g
h
t
f
o
r
w
ar
d
f
ea
t
u
r
es
i
n
in
s
ta
llatio
n
,
b
etter
p
er
f
o
r
m
a
n
c
e,
f
e
w
er
p
ar
a
m
eter
s
i
n
v
o
l
v
ed
,
an
d
lo
w
co
m
p
lex
i
t
y
.
O
v
er
r
ec
en
t
y
ea
r
s
,
th
e
D
E
tech
n
iq
u
e
h
ad
b
ee
n
ap
p
lied
t
o
lear
n
a
n
eu
r
al
m
o
d
el
v
ia
o
p
ti
m
izi
n
g
r
ea
l
an
d
co
n
s
tr
ai
n
e
d
in
teg
er
w
ei
g
h
ti
n
g
v
alu
e
s
[
1
2
]
-
[
1
5
]
.
No
n
eth
e
less
,
DE
tec
h
n
iq
u
e
g
et
s
t
h
e
s
lo
w
co
n
v
er
g
e
n
ce
s
p
ee
d
f
o
r
h
ig
h
-
s
ized
o
p
ti
m
izat
io
n
task
s
.
Fu
r
t
h
er
m
o
r
e,
DE
tec
h
n
iq
u
e
o
f
te
n
s
ee
m
s
ea
s
y
to
f
all
in
to
a
s
ta
g
n
a
n
t
s
itu
a
tio
n
t
h
at
is
a
s
tate
t
h
at
t
h
e
o
p
tim
u
m
s
ea
r
ch
i
n
g
p
r
o
ce
d
u
r
e
s
ta
g
n
a
tes
b
e
f
o
r
e
o
b
tain
i
n
g
a
g
lo
b
al
o
p
ti
m
u
m
s
o
l
u
tio
n
.
So
m
e
s
u
g
g
esti
o
n
s
o
n
h
o
w
to
i
m
p
r
o
v
e
t
h
e
DE
p
er
f
o
r
m
an
ce
o
n
t
h
e
tr
ain
i
n
g
n
eu
r
al
n
et
w
o
r
k
h
av
e
b
ee
n
s
tu
d
ied
[
16
]
-
[
2
0
]
.
P
ap
er
[
2
1
]
p
r
o
p
o
s
ed
th
e
h
y
b
r
id
m
o
d
i
f
ied
d
if
f
er
en
tial
ev
o
l
u
tio
n
p
lu
s
b
a
ck
-
p
r
o
p
ag
atio
n
al
g
o
r
ith
m
(
M
DE
-
B
P
)
f
o
r
tr
ain
i
n
g
n
eu
r
al
n
et
w
o
r
k
m
o
d
el.
T
h
e
p
er
f
o
r
m
an
ce
a
n
d
e
f
f
ic
ien
c
y
o
f
t
h
e
p
r
o
p
o
s
ed
MD
E
-
B
P
tr
ain
in
g
m
eth
o
d
ar
e
test
ed
o
n
id
en
t
if
y
i
n
g
s
o
m
e
b
en
c
h
m
a
r
k
n
o
n
li
n
ea
r
s
y
s
te
m
s
.
Ho
w
e
v
er
,
p
ap
er
[
2
1
]
o
n
l
y
p
er
f
o
r
m
f
o
r
SISO
n
o
n
li
n
ea
r
id
en
ti
f
icatio
n
s
y
s
te
m
.
T
h
er
ef
o
r
e,
th
is
p
ap
er
co
n
ti
n
u
o
u
s
l
y
s
t
u
d
ies
th
e
p
r
o
p
o
s
ed
MD
E
-
B
P
tr
ain
i
n
g
a
lg
o
r
it
h
m
i
n
[
2
1
]
to
tr
ain
a
n
eu
r
al
n
e
t
w
o
r
k
f
o
r
a
g
e
n
er
al
n
o
n
lin
ea
r
id
e
n
ti
f
icatio
n
an
d
it
s
ap
p
r
o
ac
h
to
m
o
d
el
a
p
iezo
elec
tr
ic
ac
tu
ato
r
SISO
s
y
s
te
m
a
n
d
an
ex
p
er
i
m
e
n
tal
q
u
ad
r
u
p
le
tan
k
MI
MO
s
y
s
te
m
.
T
h
ese
s
y
s
te
m
s
ar
e
s
u
itab
le
to
te
s
t
th
e
e
f
f
ec
ti
v
en
e
s
s
o
f
th
e
p
r
o
p
o
s
ed
m
et
h
o
d
b
ec
au
s
e
o
f
it
s
h
y
s
ter
esi
s
b
eh
a
v
io
r
a
n
d
h
ig
h
l
y
n
o
n
li
n
ea
r
it
y
ch
ar
ac
ter
is
ti
c.
T
h
e
r
est
o
f
t
h
is
p
ap
er
is
o
r
g
a
n
ized
as
f
o
llo
w
s
.
Sectio
n
2
i
n
tr
o
d
u
ce
s
th
e
p
r
o
p
o
s
ed
MD
E
-
B
P
tr
ain
in
g
n
eu
r
al
N
AR
X
m
o
d
el
f
o
r
MI
MO
n
o
n
li
n
ea
r
s
y
s
te
m
id
en
ti
f
icat
io
n
.
Sectio
n
3
p
r
esen
t
s
th
e
r
es
u
lts
a
n
d
d
is
cu
s
s
io
n
.
W
h
er
e
S
ec
tio
n
3
.
1
in
tr
o
d
u
ce
s
to
m
o
d
el
a
p
iezo
ele
ctr
ic
ac
tu
ato
r
SISO
s
y
s
te
m
.
Sec
tio
n
3
.
2
in
tr
o
d
u
ce
s
a
n
ex
p
er
i
m
e
n
tal
q
u
ad
r
u
p
le
tan
k
s
et
u
p
an
d
th
e
id
en
ti
f
i
ca
tio
n
r
esu
lt
s
.
Fi
n
all
y
,
S
e
ctio
n
4
co
n
tain
s
s
o
m
e
co
n
cl
u
s
io
n
s
.
2.
RE
S
E
ARCH
M
E
T
H
O
D
E
v
o
lu
tio
n
ar
y
n
eu
r
al
N
AR
X
(
E
NN)
m
o
d
el
is
p
r
o
p
o
s
ed
i
n
[
2
1
]
b
y
co
n
n
ec
ti
n
g
th
e
Mu
lti
-
L
a
y
er
P
er
ce
p
tr
o
n
Neu
r
al
Net
w
o
r
k
s
(
ML
P
)
m
o
d
el
an
d
N
AR
X
s
tr
u
ct
u
r
e.
T
h
e
w
eig
h
t
v
al
u
e
o
f
E
NN
m
o
d
el
is
o
p
tim
ized
b
y
a
h
y
b
r
id
izi
n
g
m
o
d
if
ied
d
if
f
er
en
t
ial
e
v
o
lu
tio
n
a
n
d
b
ac
k
p
r
o
p
ag
atio
n
tr
ain
i
n
g
(
MD
E
-
B
P
)
alg
o
r
ith
m
s
.
B
ased
o
n
t
h
is
co
m
b
in
atio
n
,
t
h
e
E
NN
m
o
d
el
o
b
ta
in
s
a
s
tr
o
n
g
l
y
ap
p
r
o
x
i
m
at
in
g
f
ea
tu
r
e
o
f
t
h
e
M
L
P
s
ch
e
m
e
a
n
d
a
p
o
w
er
f
u
l
p
r
ed
ictiv
e
f
ea
t
u
r
e
o
f
th
e
N
A
R
X
s
tr
u
ctu
r
e.
T
h
e
E
NN
m
o
d
el
s
tr
u
c
tu
r
e,
w
h
ich
is
u
s
ed
f
o
r
n
o
n
li
n
ea
r
MI
MO
s
y
s
te
m
i
d
en
tific
atio
n
,
is
i
llu
s
tr
ated
in
Fig
u
r
e
1
.
1
z
1
(
k
1
)
u
nb
z
M
D
E
-
B
P
a
l
g
o
r
i
t
h
m
N
o
n
l
i
n
e
a
r
M
I
M
O
s
y
s
t
e
m
1
u
n
u
1
y
m
y
1
ˆ
y
ˆ
m
y
M
L
P
-
N
A
R
X
e
r
r
o
r
e
r
r
o
r
1
(
k
)
u
nb
1
z
(
k
1
)
n
u
nb
z
(
k
)
n
u
nb
1
z
(
k
1
)
m
y
na
z
(
k
)
m
y
na
1
z
1
(
k
1
)
y
na
z
1
(
k
)
y
na
Fig
u
r
e
1
.
T
h
e
E
NN
m
o
d
el
f
o
r
a
n
o
n
li
n
ea
r
s
y
s
te
m
id
en
t
if
ica
ti
o
n
W
h
er
e,
u
(
k)
an
d
y
(
k)
r
ep
r
ese
n
t
th
e
i
n
p
u
t
a
n
d
o
u
tp
u
t
s
i
g
n
al
s
o
f
n
o
n
l
in
ea
r
MI
MO
s
y
s
te
m
;
na
an
d
nb
ar
e
o
r
d
er
o
f
o
u
tp
u
t
y
(
k
)
a
n
d
in
p
u
t
u
(
k
)
,
r
esp
ec
ti
v
el
y
;
q
is
t
h
e
b
ac
k
w
ar
d
ti
m
e
s
h
if
t
o
p
er
ato
r
d
ef
in
ed
as
q
-
1
u
(
k
)
=
u
(
k
-
1
)
.
v
jn
is
th
e
w
ei
g
h
t o
f
i
n
p
u
t la
y
er
,
v
j
0
is
t
h
e
w
ei
g
h
t o
f
b
ias in
p
u
t la
y
er
,
w
1
j
i
s
t
h
e
w
ei
g
h
t o
f
h
id
d
en
la
y
er
,
w
10
is
t
h
e
w
ei
g
h
t
o
f
b
ias
h
id
d
en
la
y
er
;
f
j
(
:
)
is
s
i
g
m
o
id
f
u
n
ctio
n
at
h
id
d
en
la
y
e
r
,
F
(
:
)
i
s
li
n
ea
r
ac
ti
v
atio
n
fu
n
ctio
n
at
o
u
tp
u
t
la
y
er
;
t
h
e
d
i
m
e
n
s
io
n
n
u
m
b
er
D
i
s
th
e
s
u
m
o
f
t
h
e
n
u
m
b
er
o
f
w
eig
h
t
s
a
n
d
th
r
es
h
o
ld
s
o
f
t
h
e
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:
2088
-
8708
B
la
ck
-
b
o
x
m
o
d
elin
g
o
f n
o
n
lin
ea
r
s
ystem
u
s
in
g
ev
o
lu
tio
n
a
r
y
n
eu
r
a
l NAR
X
mo
d
el
(
N
g
u
ye
n
N
g
o
c
S
o
n
)
1863
n
eu
r
al
N
AR
X
m
o
d
el;
12
(
k
)
,
u
,
.
.
.
,
u
n
u
u
k
k
k
is
th
e
in
p
u
t
s
i
g
n
al
v
ec
to
r
an
d
12
(
k
)
,
y
,
.
.
.
,
y
m
y
y
k
k
k
is
th
e
o
u
tp
u
t
s
i
g
n
al
v
ec
to
r
o
f
n
o
n
li
n
ea
r
MI
MO
s
y
s
te
m
.
W
e
d
eter
m
i
n
e
t
h
e
p
r
ed
ictio
n
o
u
tp
u
t a
s
f
o
llo
w
s
:
̂
(
)
(
∑
(
∑
(
)
)
)
(
1
)
W
h
er
e,
(
)
k
j
is
th
e
r
eg
r
e
s
s
io
n
v
ec
to
r
an
d
q
is
th
e
w
ei
g
h
t v
ec
to
r
,
d
ef
in
ed
b
y
1
1
,
.
.
.
,
,
1
,
.
.
.
,
,
,
.
.
.
,
1
,
.
.
.
,
,
1
,
.
.
.
,
T
T
mm
n
nn
y
k
y
k
n
a
y
k
y
k
n
a
k
k
k
u
k
u
k
n
b
u
k
u
k
n
b
(
2
)
[
]
,
-
(
3
)
T
h
e
m
o
d
i
f
ied
d
if
f
er
en
tial
e
v
o
l
u
tio
n
tr
ai
n
i
n
g
a
lg
o
r
it
h
m
i
s
u
s
e
d
to
o
p
tim
all
y
g
e
n
er
ate
t
h
e
v
a
lu
e
o
f
t
h
e
w
ei
g
h
t
o
f
t
h
e
n
e
u
r
al
N
AR
X
m
o
d
el.
T
h
e
o
u
tp
u
t
o
f
t
h
e
n
e
u
r
al
NARX
m
o
d
el
is
a
p
r
ed
ictio
n
f
u
n
ct
io
n
(
)
ˆ
yk
q
o
f
th
e
s
y
n
ap
tic
w
e
ig
h
t
v
ec
to
r
an
d
r
eg
r
ess
io
n
v
ec
to
r
k
.
A
l
g
o
r
ith
m
1
.
T
h
e
p
r
o
p
o
s
ed
M
DE
-
B
P
tr
ain
in
g
alg
o
r
it
h
m
1.
B
e
g
i
n
2.
G
e
n
e
r
a
t
e
r
a
n
d
o
ml
y
i
n
i
t
i
a
l
i
z
e
d
a
p
o
p
u
l
a
t
i
o
n
⃗
⃗
⃗
⃗
⃗
⃗
[
]
3.
Ev
a
l
u
a
t
e
t
h
e
f
i
t
n
e
ss
of
e
a
c
h
i
n
d
i
v
i
d
u
a
l
i
n
t
h
e
p
o
p
u
l
a
t
i
o
n
4.
Fo
r
G
=
1
t
o
G
E
N
do
5.
Fo
r
i
=
1
t
o
N
P
do
6.
j
r
a
nd
=
r
a
n
d
i
n
t
(
1
,
D
)
7.
F
=
ra
n
d
[
0
:
4
;
1
:
0
],
C
R
=
r
a
n
d
[0
:
7
;
1
:
0
]
8.
Fo
r
j
=
1
t
o
D
do
9.
I
f
r
a
n
d
[
0
,
1
]
<
CR
o
r
j
=
=
j
ra
n
d
t
h
e
n
1
0
.
If
r
a
n
d
[0
,
1
] >
t
h
r
e
s
h
o
l
d
t
h
e
n
1
1
.
S
e
l
e
c
t
r
a
n
d
o
ml
y
*
+
1
2
.
(
)
1
3
.
E
l
se
1
4
.
S
e
l
e
c
t
r
a
n
d
o
ml
y
*
+
1
5
.
(
)
1
6
.
E
n
d
i
f
1
7
.
E
l
s
e
1
8
.
1
9
.
E
n
d
i
f
2
0
.
E
n
d
f
o
r
2
1
.
I
f
(
⃗
⃗
)
(
⃗
⃗
)
t
h
e
n
2
2
.
⃗
⃗
⃗
⃗
2
3
.
E
l
s
e
2
4
.
⃗
⃗
⃗
⃗
2
5
.
E
n
d
i
f
2
6
.
E
n
d
f
o
r
2
7
.
U
se
⃗
⃗
a
s
a
n
i
n
i
t
i
a
l
p
o
i
n
t
f
o
r
B
P
a
l
g
o
r
i
t
h
m
2
8
.
R
e
p
l
a
c
e
⃗
⃗
b
y
bp
o
b
t
a
i
n
i
n
g
f
r
o
m
B
P
a
l
g
o
r
i
t
h
m
2
9
.
E
n
d
f
o
r
3
0
.
E
n
d
I
n
th
e
tr
ain
in
g
p
r
o
ce
s
s
,
b
o
t
h
th
e
in
p
u
t
s
ig
n
al
u
(
k
)
a
n
d
t
h
e
o
u
tp
u
t
s
i
g
n
al
y
(
k
)
o
f
n
o
n
li
n
e
ar
MI
MO
s
y
s
te
m
ar
e
k
n
o
w
n
a
n
d
t
h
e
s
y
n
ap
tic
w
ei
g
h
t
v
ec
to
r
is
ad
ap
ted
to
o
b
tain
ap
p
r
o
p
r
iate
f
u
n
c
tio
n
al
m
ap
p
in
g
s
f
r
o
m
t
h
e
i
n
p
u
t
u
(
k
)
to
t
h
e
o
u
tp
u
t
y
(
k)
.
Ge
n
er
all
y
,
th
e
ad
ap
tati
o
n
p
r
o
ce
s
s
ca
n
b
e
ca
r
r
ied
o
u
t
b
y
m
i
n
i
m
izi
n
g
th
e
n
et
w
o
r
k
er
r
o
r
f
u
n
ctio
n
E
N
w
h
ich
is
b
ased
o
n
th
e
i
n
tr
o
d
u
cti
o
n
o
f
a
m
ea
s
u
r
e
o
f
clo
s
e
n
ess
in
t
er
m
s
o
f
a
m
ea
n
s
q
u
ar
e
er
r
o
r
(
MSE
)
cr
iter
i
o
n
.
(
)
∑
,
(
)
̂
(
)
-
,
(
)
̂
(
)
-
(
4
)
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.
9
,
No
.
3
,
J
u
n
e
201
9
:
1
8
6
1
-
1
8
7
0
1864
W
h
er
e
th
e
tr
ain
in
g
d
ata
N
Z
is
s
p
ec
if
ied
b
y
*
,
(
)
(
)
-
+
.
T
h
e
o
p
tim
izatio
n
g
o
al
is
to
m
i
n
i
m
ize
t
h
e
o
b
j
ec
tiv
e
f
u
n
ct
io
n
E
N
b
y
o
p
ti
m
izi
n
g
t
h
e
v
alu
e
s
o
f
t
h
e
n
et
w
o
r
k
w
ei
g
h
ts
[
]
12
,
,
.
.
.
.
,
T
D
w
w
w
q
=
b
y
u
s
i
n
g
t
h
e
M
DE
-
B
P
alg
o
r
ith
m
w
h
ic
h
i
s
d
escr
ib
ed
in
d
etails
a
t
p
ap
er
[
2
1
]
.
T
h
e
w
o
r
k
i
n
g
s
tep
s
i
n
v
o
lv
ed
in
e
m
p
lo
y
i
n
g
MD
E
-
B
P
tr
ain
i
n
g
al
g
o
r
ith
m
as
Alg
o
r
it
h
m
1
.
3.
RE
SU
L
T
S
A
ND
AN
AL
Y
SI
S
A
ll
o
f
t
h
e
m
o
d
eli
n
g
s
is
p
er
f
o
r
m
ed
b
y
Ma
tlab
v
er
s
io
n
2
0
1
3
b
o
n
an
I
n
tel
C
o
r
e
i3
co
m
p
u
ter
w
i
th
a
clo
ck
r
ate
o
f
2
.
5
3
GHz
an
d
2
.
0
0
GB
o
f
R
A
M
.
3
.
1
.
M
o
delin
g
t
he
hy
s
t
er
esis
beha
v
io
r
o
f
t
he
p
iezo
elec
t
ric
a
ct
ua
t
o
r
Mic
r
o
p
o
s
itio
n
in
g
s
ta
g
es
u
s
i
n
g
p
iezo
elec
tr
i
c
ac
tu
ato
r
ar
e
w
id
el
y
u
s
ed
in
a
v
ar
ie
t
y
o
f
ap
p
licatio
n
s
s
u
c
h
as
m
icr
o
g
r
ip
p
er
r
o
b
o
t
[
2
2
]
,
en
er
g
y
h
ar
v
e
s
ti
n
g
[
2
3
]
,
M
E
Ms
[
2
4
]
an
d
s
o
o
n
.
T
h
e
p
ie
zo
elec
tr
ic
ac
tu
ato
r
p
o
s
s
ess
es
a
h
i
g
h
er
d
is
p
lace
m
e
n
t
ac
c
u
r
ac
y
,
lar
g
er
g
e
n
er
atio
n
f
o
r
ce
an
d
h
i
g
h
er
r
esp
o
n
s
e
s
p
e
ed
th
a
n
o
th
er
t
y
p
e
s
o
f
ac
tu
a
to
r
s
.
Ho
w
e
v
er
,
t
h
e
n
o
n
lin
ea
r
it
ies
h
y
s
ter
esi
s
o
f
th
e
p
iezo
elec
tr
ic
ac
tu
ato
r
ca
n
g
r
ea
tl
y
d
eg
r
ad
e
t
h
e
p
o
s
itio
n
in
g
ac
c
u
r
ac
y
o
f
m
icr
o
p
o
s
itio
n
in
g
s
y
s
te
m
s
[
2
5
]
,
[
2
6
]
.
T
h
er
ef
o
r
e,
m
o
d
eli
n
g
p
r
ec
is
el
y
an
d
id
en
tify
i
n
g
p
ar
am
eter
e
f
f
ec
t
iv
el
y
f
o
r
p
i
ez
o
elec
tr
ic
ac
tu
ato
r
s
till
r
e
m
ai
n
a
ch
allen
g
i
n
g
p
r
o
b
le
m
n
o
w
ad
a
y
s
.
T
o
d
escr
ib
e
th
e
h
y
s
ter
esi
s
b
eh
av
io
r
o
f
th
e
p
iezo
elec
tr
ic
ac
tu
ato
r
,
m
a
n
y
m
ath
e
m
atica
l
m
o
d
el
s
h
a
v
e
b
ee
n
p
r
o
p
o
s
ed
s
u
ch
as
P
r
an
d
tl
-
I
s
h
lin
s
k
ii,
P
r
eisach
m
o
d
el,
B
o
u
c
-
W
en
m
o
d
el,
KP
m
o
d
el,
Du
h
e
m
m
o
d
el
an
d
s
o
o
n
[
2
7
]
.
Ho
w
e
v
er
,
th
e
d
if
f
ic
u
lt
ies i
n
d
eter
m
i
n
i
n
g
t
h
e
p
ar
am
eter
s
o
f
th
e
h
y
s
ter
es
is
m
o
d
el
lim
it its
ap
p
licatio
n
i
n
p
r
ac
tice.
Fig
u
r
e
2
.
Selectio
n
o
f
tr
ai
n
i
n
g
d
ata
f
o
r
esti
m
at
in
g
an
d
v
alid
at
in
g
p
u
r
p
o
s
es
T
o
o
v
er
co
m
e
t
h
is
p
r
o
b
le
m
,
p
a
p
er
[
2
5
]
p
r
o
p
o
s
ed
a
p
ar
am
eter
id
en
ti
f
icatio
n
m
et
h
o
d
b
ased
o
n
ar
tific
ial
n
eu
r
al
n
et
w
o
r
k
s
f
o
r
th
e
D
u
h
e
m
h
y
s
ter
e
s
is
m
o
d
el.
T
h
e
m
at
h
e
m
atica
l
m
o
d
el
o
f
t
h
e
p
iezo
elec
tr
ic
ac
tu
ato
r
b
ased
Du
h
e
m
m
o
d
el
in
[
2
5
]
ca
n
b
e
d
escr
ib
ed
as
m
y
t
n
y
t
k
y
t
k
d
u
t
w
t
w
t
u
t
c
u
t
w
t
b
u
t
(
5
)
w
h
er
e
m
is
th
e
eq
u
iv
a
len
t
m
a
s
s
;
n
is
t
h
e
eq
u
iv
ale
n
t
d
a
m
p
i
n
g
;
k
is
t
h
e
eq
u
i
v
ale
n
t
s
ti
f
f
n
e
s
s
;
u
(
t)
is
t
h
e
i
n
p
u
t
v
o
ltag
e;
y
(
t
)
is
th
e
o
u
tp
u
t
d
is
p
lace
m
en
t
;
w
(
t)
is
t
h
e
h
y
s
ter
e
tic
s
tate
v
ar
iab
le;
p
ar
am
eter
s
α
,
b
,
c
c
o
n
tr
o
ls
th
e
s
h
ap
es a
n
d
a
m
p
l
itu
d
e
s
o
f
t
h
e
h
y
s
ter
esi
s
lo
o
p
; a
n
d
d
is
th
e
e
f
f
ec
ti
v
e
p
iezo
elec
tr
ic
co
ef
f
icie
n
t o
f
th
e
ac
tu
a
to
r
.
I
n
th
is
p
ap
er
,
th
e
E
NN
m
o
d
el
tr
ain
ed
MD
E
-
B
P
alg
o
r
ith
m
is
p
r
o
p
o
s
ed
t
o
id
en
tify
t
h
e
n
o
n
li
n
ea
r
ities
D
u
h
e
m
h
y
s
ter
esi
s
o
f
t
h
e
p
iezo
elec
tr
ic
ac
tu
ato
r
.
Fir
s
tl
y
,
th
e
ab
o
v
e
s
y
s
te
m
(
5
)
is
i
m
p
le
m
en
ted
in
M
A
T
L
A
B
.
T
h
e
tr
ain
in
g
d
ata
s
et
an
d
v
ali
d
atin
g
d
ata
s
et
ar
e
co
llected
f
o
r
m
s
y
s
te
m
(
5
)
an
d
s
h
o
w
n
in
Fi
g
u
r
e
2
.
Se
co
n
d
l
y
,
t
h
e
E
NN
s
tr
u
c
tu
r
e
i
s
s
elec
ted
b
y
li
n
k
i
n
g
th
e
3
-
la
y
er
ML
P
w
it
h
S1
h
id
d
en
n
o
d
es
an
d
th
e
2
nd
o
r
d
e
r
NARX
s
tr
u
ct
u
r
e,
is
p
r
esen
ted
w
it
h
th
e
r
eg
r
es
s
io
n
an
d
p
r
ed
ictio
n
f
u
n
ctio
n
s
,
r
esp
ec
tiv
el
y
.
Selecti
n
g
p
ar
a
m
eter
S1
=
5
,
NP =
6
2
an
d
GE
N
=
1
0
0
0
0
.
0
1
2
3
4
5
-5
0
5
D
a
t
a
s
e
t
f
o
r
t
r
a
i
n
i
n
g
i
n
p
u
t
d
a
t
a
0
1
2
3
4
5
-5
0
5
t
i
m
e
(
s
e
c
)
o
u
t
p
u
t
d
a
t
a
0
1
2
3
4
5
-4
-2
0
2
4
D
a
t
a
s
e
t
f
o
r
v
a
l
i
d
a
t
i
o
n
0
1
2
3
4
5
-4
-2
0
2
4
t
i
m
e
(
s
e
c
)
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:
2088
-
8708
B
la
ck
-
b
o
x
m
o
d
elin
g
o
f n
o
n
lin
ea
r
s
ystem
u
s
in
g
ev
o
lu
tio
n
a
r
y
n
eu
r
a
l NAR
X
mo
d
el
(
N
g
u
ye
n
N
g
o
c
S
o
n
)
1865
Fig
u
r
e
3
.
C
o
n
v
er
g
en
ce
r
ate
i
n
tr
ain
i
n
g
Fig
u
r
e
4
.
I
d
en
tif
icatio
n
p
er
f
o
r
m
an
ce
Fin
all
y
,
t
h
e
p
er
f
o
r
m
an
ce
id
en
t
if
icatio
n
u
s
i
n
g
th
e
p
r
o
p
o
s
ed
E
NN
m
o
d
el
i
s
co
n
d
u
c
ted
.
Fig
u
r
e
3
s
h
o
ws
th
e
co
n
v
er
g
e
n
ce
r
ates
o
f
MD
E
-
BP
-
E
NN
in
t
h
e
tr
ai
n
in
g
p
r
o
ce
s
s
.
Fi
g
u
r
e
4
d
e
m
o
n
s
tr
ates
t
h
e
er
r
o
r
s
b
etw
ee
n
th
e
p
r
ed
icted
o
u
tp
u
t
o
f
t
h
e
n
eu
r
a
l
n
et
w
o
r
k
an
d
t
h
e
f
o
r
m
u
la
o
u
tp
u
t
o
f
D
u
h
e
m
m
o
d
el.
T
h
e
r
esu
lt
s
i
n
d
icate
t
h
e
p
r
o
p
o
s
ed
m
et
h
o
d
ca
n
b
e
u
s
ed
to
id
en
tify
an
d
p
r
ed
ict
th
e
h
y
s
t
er
esis
b
eh
a
v
io
r
o
f
th
e
p
iezo
ele
ctr
ic
ac
tu
ato
r
.
3.
2
.
M
o
delin
g
a
qu
a
drup
le
t
a
nk
s
y
s
t
e
m
I
n
p
r
ac
tice,
a
co
m
m
o
n
tan
k
s
y
s
te
m
co
n
s
i
s
t
s
o
f
ta
n
k
s
,
p
u
m
p
s
,
p
ip
in
g
,
v
a
lv
e
s
,
liq
u
id
le
v
e
l
s
e
n
s
o
r
s
.
Du
r
in
g
o
p
er
atio
n
,
d
ep
en
d
i
n
g
o
n
th
e
tech
n
o
lo
g
y
r
eq
u
ir
e
m
e
n
ts
an
d
lo
ad
r
eq
u
ir
e
m
en
ts
,
t
h
e
v
alv
e
s
ca
n
c
h
an
g
e
th
e
ap
er
tu
r
e,
o
r
th
e
liq
u
id
t
y
p
e
in
t
h
e
ta
n
k
s
ca
n
c
h
an
g
e.
T
h
i
s
lead
s
to
a
n
o
n
li
n
ea
r
c
h
ar
ac
te
r
is
tic
o
f
th
e
s
y
s
te
m
th
at
ca
n
v
ar
y
.
Mo
r
eo
v
er
,
th
e
p
ar
a
m
eter
s
o
f
t
h
e
liq
u
id
tan
k
s
y
s
te
m
s
u
c
h
as t
h
e
s
ize
an
d
d
is
c
h
ar
g
e
co
ef
f
icie
n
t o
f
th
e
v
al
v
es,
t
h
e
p
o
w
er
f
ac
to
r
o
f
th
e
p
u
m
p
in
s
o
m
e
ca
s
e
s
m
a
y
b
e
s
till
u
n
k
n
o
w
n
.
T
h
er
ef
o
r
e,
it
is
n
o
t
ea
s
y
to
o
b
tain
an
ac
cu
r
ate
t
h
eo
r
etica
l
m
o
d
el
o
f
a
q
u
ad
r
atu
r
e
tan
k
s
y
s
te
m
.
T
o
o
v
er
co
m
e
t
h
e
d
is
ad
v
an
ta
g
es
o
f
th
e
t
h
eo
r
etica
l
m
o
d
eli
n
g
m
et
h
o
d
,
s
o
m
e
id
en
t
if
ica
tio
n
ap
p
r
o
ac
h
es
h
av
e
b
ee
n
i
n
v
e
s
ti
g
ated
.
P
ap
e
r
[
2
8
]
p
r
o
p
o
s
ed
an
ad
ap
tiv
e
f
u
zz
y
m
o
d
el
p
r
ed
ictiv
e
co
n
tr
o
l
b
ased
o
n
th
e
an
t
co
lo
n
y
o
p
ti
m
izatio
n
(
A
C
O)
f
o
r
n
o
n
li
n
ea
r
p
r
o
ce
s
s
s
y
s
te
m
.
P
ap
er
[
2
9
]
u
s
ed
a
r
ad
ial
b
asis
f
u
n
c
tio
n
n
e
u
r
al
n
et
w
o
r
k
a
s
th
e
p
r
ed
ictio
n
f
u
n
ctio
n
d
u
e
to
its
ap
p
r
o
x
i
m
at
io
n
ab
ilit
y
.
T
h
e
n
,
a
n
o
n
li
n
ea
r
m
o
d
el
p
r
ed
ictiv
e
co
n
tr
o
l
m
e
th
o
d
w
a
s
p
r
o
p
o
s
ed
to
g
u
ar
a
n
tee
t
h
e
s
y
s
te
m
s
tab
i
lit
y
an
d
co
m
p
en
s
a
te
t
h
e
n
et
wo
r
k
-
i
n
d
u
ce
d
d
ela
y
s
an
d
p
ac
k
et
d
r
o
p
o
u
ts
.
I
p
lik
ci
[
3
0
]
p
r
o
p
o
s
ed
a
s
u
p
p
o
r
t
v
ec
t
o
r
m
ac
h
i
n
e
(
S
VM
)
ap
p
r
o
ac
h
to
th
e
g
e
n
er
alize
d
p
r
ed
ictiv
e
co
n
tr
o
l
(
GP
C
)
o
f
a
n
e
x
p
er
i
m
e
n
tal
t
h
r
ee
-
tan
k
s
y
s
te
m
.
O
n
e
ad
v
a
n
ta
g
e
o
f
t
h
e
b
l
ac
k
-
b
o
x
m
o
d
eli
n
g
ap
p
r
o
ac
h
is
th
at
it
i
s
u
n
n
ec
e
s
s
ar
y
to
p
r
ec
is
el
y
a
n
al
y
ze
it
s
in
ter
n
al
m
ec
h
a
n
i
s
m
o
r
th
e
r
elatio
n
s
h
ip
a
m
o
n
g
p
h
y
s
ical
v
ar
iab
les.
I
n
th
i
s
p
a
p
er
,
th
e
E
NN
m
o
d
el
tr
ain
ed
MD
E
-
B
P
alg
o
r
ith
m
i
s
p
r
o
p
o
s
ed
to
id
en
ti
f
y
t
h
e
b
lack
-
b
o
x
m
o
d
elin
g
o
f
a
q
u
ad
r
u
p
le
tan
k
s
y
s
te
m
.
3
.
2
.
1
.
P
hy
s
ica
l
m
o
del
T
h
e
q
u
ad
r
u
p
le
tan
k
s
y
s
te
m
wh
ich
is
d
escr
ib
ed
in
F
ig
u
r
e
5
is
a
n
o
n
lin
ea
r
p
la
n
t
w
it
h
s
tr
o
n
g
co
u
p
li
n
g
an
d
lar
g
e
t
i
m
e
d
ela
y
.
T
h
e
q
u
a
d
r
u
p
le
tan
k
p
lan
t
ca
n
b
e
s
ee
n
as
a
s
y
s
te
m
w
it
h
t
w
o
i
n
p
u
t
s
,
t
h
e
v
o
ltag
e
s
o
f
th
e
p
u
m
p
s
,
a
n
d
t
w
o
o
u
tp
u
ts
,
t
h
e
wate
r
lev
els o
f
t
h
e
lo
w
er
tan
k
s
.
W
h
er
e
L
i
is
th
e
liq
u
id
lev
el
i
n
tan
k
i
;
D
ti
i
s
th
e
o
u
tlet
cr
o
s
s
-
s
ec
tio
n
al
ar
ea
o
f
ta
n
k
i
;
D
oi
i
s
th
e
cr
o
s
s
-
s
ec
tio
n
al
ar
ea
o
f
ta
n
k
i
;
i
is
t
h
e
p
o
r
tio
n
o
f
th
e
f
lo
w
in
to
th
e
u
p
p
er
tan
k
f
r
o
m
p
u
m
p
i
;
u
j
is
t
h
e
s
p
ee
d
s
ett
in
g
o
f
p
u
m
p
j
,
w
ith
t
h
e
co
r
r
esp
o
n
d
in
g
g
ai
n
k
j
.
A
m
a
th
e
m
atica
l
m
o
d
el
b
ased
o
n
p
h
y
s
ical
d
ata
to
g
eth
er
w
it
h
it
s
lin
ea
r
izatio
n
ar
o
u
n
d
an
o
p
er
atin
g
p
o
in
t i
s
d
er
iv
ed
in
J
o
h
a
n
s
s
o
n
[
3
1
]
.
0
2000
4000
6000
8000
10000
10
-3
10
-2
10
-1
10
0
10
1
G
e
n
e
r
a
t
i
o
n
M
S
E
f
o
r
t
r
a
i
n
i
n
g
0
1
2
3
4
5
-4
-2
0
2
4
o
u
t
p
u
t
0
1
2
3
4
5
-
0
.
4
-
0
.
2
0
0
.
2
t
i
m
e
(
s
e
c
)
e
r
r
o
r
a
c
t
u
a
l
p
r
e
d
i
c
t
e
d
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.
9
,
No
.
3
,
J
u
n
e
201
9
:
1
8
6
1
-
1
8
7
0
1866
1
2
3
4
P
u
m
p
2
P
u
m
p
1
Fig
u
r
e
5
.
Sch
e
m
atic
o
f
t
h
e
q
u
a
d
r
u
p
le
-
tan
k
s
y
s
te
m
3
.
2
.
2
.
E
x
peri
m
e
nta
l qua
dru
ple t
a
nk
s
et
up
I
n
th
is
p
ar
t,
th
e
e
x
p
er
i
m
en
ta
l
q
u
ad
r
u
p
le
tan
k
s
y
s
te
m
h
a
s
b
ee
n
s
t
u
d
ied
.
Fig
u
r
e
6
d
es
cr
ib
es
th
e
d
iag
r
a
m
b
lo
ck
o
f
t
h
e
e
x
p
er
i
m
en
tal
q
u
ad
r
u
p
le
tan
k
s
y
s
te
m
.
T
ab
le
1
d
escr
ib
es
in
d
etail
t
h
e
p
h
y
s
ical
p
ar
a
m
eter
s
of
th
e
ta
n
k
m
o
d
el
s
y
s
te
m
.
T
ab
le
1
.
T
h
e
P
a
r
am
eter
s
o
f
th
e
Qu
ad
r
u
p
le
-
ta
n
k
S
y
s
te
m
P
a
r
a
me
t
e
r
s
V
a
l
u
e
U
n
i
t
T
h
e
o
u
t
l
e
t
c
r
o
ss
-
se
c
t
i
o
n
a
l
a
r
e
a
o
f
t
h
e
t
a
n
k
(
1
2
3
4
,
,
,
t
t
t
t
D
D
D
D
)
4
.
6
cm
T
h
e
c
r
o
ss
-
se
c
t
i
o
n
a
l
a
r
e
a
o
f
t
a
n
k
1
a
n
d
3
(
1
o
3
,
o
DD
)
0
.
8
cm
T
h
e
c
r
o
ss
-
se
c
t
i
o
n
a
l
a
r
e
a
o
f
t
a
n
k
2
a
n
d
4
(
2
o
4
,
o
DD
)
0
.
6
5
cm
T
h
e
p
o
r
t
i
o
n
o
f
t
h
e
f
l
o
w
i
n
t
o
t
h
e
u
p
p
e
r
t
a
n
k
f
r
o
m p
u
m
p
1
,
1
0
.
7
------
T
h
e
p
o
r
t
i
o
n
o
f
t
h
e
f
l
o
w
i
n
t
o
t
h
e
u
p
p
e
r
t
a
n
k
f
r
o
m p
u
m
p
2
,
2
0
.
6
5
------
T
h
e
c
o
r
r
e
sp
o
n
d
i
n
g
g
a
i
n
o
f
P
u
mp
,
p
k
6
.
9
3
c
m
/
s
/
V
T
h
e
sp
e
e
d
se
t
t
i
n
g
o
f
t
h
e
p
u
m
p
,
(
12
,
uu
)
------
V
T
h
e
p
u
m
p
u
s
ed
in
t
h
e
q
u
ad
r
u
p
le
-
tan
k
s
y
s
te
m
i
s
an
O
g
i
h
ar
a
en
g
i
n
e
w
it
h
a
2
4
V
s
u
p
p
l
y
a
n
d
a
p
u
m
p
f
lo
w
r
ate
o
f
1
0
liter
s
p
er
m
i
n
u
te.
I
n
f
ac
t,
t
h
e
m
o
to
r
s
p
u
m
p
wate
r
in
to
t
h
e
tan
k
i
n
th
e
o
p
er
atin
g
r
an
g
e
o
f
3
0
%
-
7
0
%
(
V
m
ax
)
.
T
h
e
Fre
escale
S
e
m
ico
n
d
u
cto
r
's
Stra
i
n
Ga
u
g
e
MP
X1
0
is
u
s
ed
to
m
ea
s
u
r
e
wate
r
lev
els
in
ta
n
k
s
.
Deta
ils
o
f
t
h
e
eq
u
ip
m
en
t p
ar
am
eter
s
o
f
t
h
e
q
u
ad
r
u
p
le
-
ta
n
k
s
y
s
te
m
ar
e
d
escr
ib
ed
in
T
ab
le
2
.
A
m
p
l
i
f
i
e
r
NI
-
P
C
I
6
2
2
1
P
r
e
s
s
u
r
e
s
e
n
s
o
r
D
A
C
M
a
t
l
a
b
/
S
i
m
u
l
i
n
k
A
D
C
Fig
u
r
e
6
.
Diag
r
a
m
o
f
th
e
e
x
p
er
i
m
en
tal
q
u
ad
r
u
p
le
tan
k
s
y
s
te
m
T
ab
le
2
.
T
h
e
p
ar
am
eter
s
o
f
P
u
m
p
a
n
d
MP
X1
0
Eq
u
i
p
me
n
t
P
a
r
a
me
t
e
r
s
P
u
mp
-
S
u
p
p
l
y
v
o
l
t
a
g
e
:
2
4
V
DC
-
F
l
o
w
r
a
t
e
:
1
0
l
i
t
e
r
s
/
m
i
n
u
t
e
-
P
o
w
e
r
c
o
n
su
mp
t
i
o
n
:
2
6
W
M
P
X
1
0
se
n
so
r
-
Ty
p
e
o
f
M
P
X
1
0
D
P
C
A
S
E
3
4
4
C
.
-
S
u
p
p
l
y
v
o
l
t
a
g
e
:
3
-
6
VDC
-
S
e
n
si
t
y
∆
V
/
∆P
:
3
.
5
mV
/
k
P
a
-
R
a
n
g
e
:
0
-
1
0
k
P
a
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:
2088
-
8708
B
la
ck
-
b
o
x
m
o
d
elin
g
o
f n
o
n
lin
ea
r
s
ystem
u
s
in
g
ev
o
lu
tio
n
a
r
y
n
eu
r
a
l NAR
X
mo
d
el
(
N
g
u
ye
n
N
g
o
c
S
o
n
)
1867
3
.
2
.
3
.
I
dentif
ica
t
io
n inv
er
s
e
m
o
de
l r
esu
lt
s
I
n
f
ac
t,
t
h
e
in
v
er
s
e
p
r
ed
ictiv
e
m
o
d
el
p
la
y
s
an
i
m
p
o
r
tan
t
r
o
le
in
th
e
d
esi
g
n
o
f
m
o
d
el
-
b
ased
co
n
tr
o
l.
I
n
th
is
ar
ticle,
th
e
i
n
v
er
s
e
m
o
d
e
l
o
f
th
e
q
u
ad
r
u
p
le
-
ta
n
k
s
y
s
te
m
i
s
id
en
ti
f
ie
d
b
y
u
s
i
n
g
t
h
e
E
NN
m
o
d
el
.
T
h
e
p
r
o
ce
d
u
r
e
co
n
s
is
ts
o
f
f
o
u
r
b
asi
c
s
tep
s
as f
o
llo
w
s
:
Firstl
y
,
th
e
ex
p
er
i
m
en
tal
in
p
u
t
-
o
u
tp
u
t
d
ata
s
et,
ap
p
lied
f
o
r
tr
ain
i
n
g
a
n
d
v
alid
ati
n
g
p
u
r
p
o
s
es
o
f
t
h
e
E
NN
b
ased
MD
E
-
B
P
tr
ain
in
g
alg
o
r
it
h
m
,
i
s
co
llected
f
r
o
m
th
e
q
u
ad
r
u
p
le
-
ta
n
k
s
y
s
te
m
.
T
h
e
in
p
u
t
d
ata
s
et
co
n
s
is
ts
o
f
th
e
i
n
p
u
t
s
ig
n
al
as
th
e
v
o
lta
g
e
f
o
r
th
e
p
u
m
p
m
o
t
o
r
an
d
th
e
o
u
tp
u
t
s
ig
n
al
o
f
ta
n
k
w
a
ter
lev
el
2
an
d
tan
k
4
co
llected
f
r
o
m
t
h
e
e
x
p
er
i
m
en
tal
m
o
d
el
d
escr
ib
ed
in
F
ig
u
r
e
7
.
W
h
ic
h
,
1
()
uk
an
d
2
()
uk
ar
e
th
e
v
o
ltag
e
s
ig
n
al
f
o
r
th
e
p
u
m
p
1
an
d
p
u
m
p
2
.
2
()
Lk
an
d
4
()
Lk
is
th
e
w
ater
le
v
el
i
n
tan
k
2
an
d
ta
n
k
4
.
T
h
e
d
ataset
w
i
t
h
th
e
i
n
p
u
t
v
o
lta
g
e
r
atio
h
as
a
c
h
an
g
e
i
n
a
m
p
lit
u
d
e
f
r
o
m
0
to
1
an
d
th
e
c
h
a
n
g
e
f
r
eq
u
e
n
c
y
d
escr
ib
ed
in
Fi
g
u
r
e
7
w
il
l b
e
u
s
ed
to
es
ti
m
ate
a
n
d
e
v
a
lu
a
te
t
h
e
MD
E
-
E
NN
p
r
ed
ictio
n
m
o
d
el.
W
h
er
e,
Fi
g
u
r
e
7
(
a)
is
u
s
ed
to
e
s
ti
m
ate
th
e
m
o
d
el,
Fi
g
u
r
e
7
(
b
)
is
u
s
ed
to
ev
alu
ate
t
h
e
m
o
d
el.
Fig
u
r
e
7
.
E
x
p
er
i
m
en
tal
d
ataset
f
o
r
tr
ain
in
g
a
n
d
v
alid
atio
n
o
f
th
e
E
NN
m
o
d
el
Seco
n
d
l
y
,
t
h
e
E
NN
s
tr
u
ct
u
r
e
is
s
elec
ted
b
y
li
n
k
in
g
t
h
e
3
-
la
y
er
ed
ML
P
m
o
d
el
w
ith
t
h
e
1
st
o
r
d
e
r
NARX
s
tr
u
ctu
r
e
an
d
d
escr
ib
ed
in
Fig
u
r
e
8
.
I
n
th
is
ca
s
e,
th
e
n
u
m
b
er
o
f
h
id
d
en
n
e
u
r
o
n
s
is
S1
=
1
2
,
p
o
p
u
latio
n
s
ize
NP
=
1
2
0
,
n
u
m
b
er
o
f
g
e
n
er
atio
n
s
o
f
tr
ai
n
ed
GE
N
=
7
0
0
,
th
e
co
ef
f
icie
n
t
o
f
lear
n
in
g
B
P
0
.
0
0
0
0
0
1
an
d
n
u
m
b
er
o
f
iter
atio
n
s
f
o
r
B
P
iter
3
=
5
.
e
1
ˆ
uk
2
(
k
1
)
L
1
()
uk
I
n
v
e
r
s
e
N
N
A
R
X
m
o
d
e
l
M
D
E
-
B
P
1
z
e
1
z
1
z
1
z
2
ˆ
uk
2
()
uk
4
(
k
1
)
L
Q
u
a
d
r
u
p
l
e
t
a
n
k
s
y
s
t
e
m
Fig
u
r
e
8
.
T
h
e
in
v
er
s
e
m
o
d
el
o
f
th
e
q
u
ad
r
u
p
le
-
ta
n
k
s
y
s
te
m
0
500
1000
1500
2000
0
0
.
5
1
(
a
)
u1e
0
500
1000
1500
2000
0
10
20
L2e
0
500
1000
1500
2000
0
0
.
5
1
u2e
0
500
1000
1500
2000
0
10
t
i
m
e
(
s
a
m
p
l
e
s
)
L4e
0
500
1000
1500
2000
0
0
.
5
1
(
b
)
u1v
0
500
1000
1500
2000
15
20
L2v
0
500
1000
1500
2000
0
0
.
5
1
u2v
0
500
1000
1500
2000
10
15
t
i
m
e
(
s
a
m
p
l
e
s
)
L4v
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.
9
,
No
.
3
,
J
u
n
e
201
9
:
1
8
6
1
-
1
8
7
0
1868
Fin
all
y
,
t
h
e
e
s
ti
m
atio
n
an
d
v
alid
atio
n
p
r
o
ce
s
s
ar
e
co
n
d
u
ct
ed
to
id
en
tify
t
h
e
E
N
N
m
o
d
e
l.
Fig
u
r
e
9
s
h
o
w
s
t
h
e
p
er
f
o
r
m
an
ce
o
f
M
DE
-
BP
-
E
NN
b
ased
o
n
th
e
v
alu
es
o
f
MSE
in
th
e
tr
ain
i
n
g
p
r
o
ce
s
s
.
Fig
u
r
e
1
0
s
h
o
w
s
t
h
e
p
er
f
o
r
m
a
n
ce
o
f
M
DE
-
BP
-
E
NN
b
ased
o
n
t
h
e
a
v
er
ag
e
v
a
lu
e
s
o
f
MSE
o
n
t
h
e
v
alid
atio
n
p
r
o
ce
s
s
.
B
ased
o
n
th
e
ab
o
v
e
r
es
u
lt
s
,
we
s
ee
th
at
th
e
E
N
N
p
r
ed
ictio
n
m
o
d
el
w
it
h
w
ei
g
h
t
in
g
s
o
p
ti
m
ized
b
y
t
h
e
MD
E
-
B
P
alg
o
r
ith
m
h
as
b
ee
n
id
en
t
if
ied
an
d
p
r
ed
icted
f
air
l
y
ac
cu
r
atel
y
th
e
i
n
v
er
s
e
d
y
n
a
m
ic
p
r
o
p
er
ties
o
f
th
e
q
u
ad
r
u
p
le
-
ta
n
k
s
y
s
te
m
.
Fig
u
r
e
9
.
T
h
e
MSE
v
alu
e
o
n
t
h
e
tr
ai
n
i
n
g
p
r
o
ce
s
s
Fig
u
r
e
1
0
.
T
h
e
MSE
v
al
u
e
o
n
th
e
v
al
id
atio
n
p
r
o
ce
s
s
4.
CO
NCLU
SI
O
N
T
h
is
p
ap
er
in
tr
o
d
u
ce
s
h
o
w
to
m
o
d
eli
n
g
a
n
o
n
l
in
ea
r
SI
SO
a
n
d
MI
MO
s
y
s
te
m
b
ased
o
n
i
n
p
u
t
-
o
u
tp
u
t
d
ata
u
s
i
n
g
t
h
e
p
r
o
p
o
s
ed
E
v
o
lu
tio
n
ar
y
Neu
r
al
N
AR
X
(
E
NN
)
m
o
d
el.
T
h
e
E
NN
s
ch
e
m
e
i
s
u
s
ed
b
y
co
n
n
ec
ti
n
g
th
e
Mu
l
ti
-
L
a
y
er
P
er
ce
p
tr
o
n
N
eu
r
al
Net
w
o
r
k
s
(
M
L
P
)
m
o
d
el
an
d
NARX
s
tr
u
ctu
r
e.
T
h
e
w
eig
h
t
v
al
u
e
o
f
E
NN
m
o
d
el
is
o
p
ti
m
ized
b
y
h
y
b
r
id
i
zin
g
a
m
o
d
if
ied
d
if
f
er
en
tial
e
v
o
lu
tio
n
an
d
b
ac
k
p
r
o
p
ag
atio
n
t
r
ain
in
g
(
M
DE
-
B
P
)
alg
o
r
ith
m
s
.
T
h
e
ef
f
ec
ti
v
e
n
ess
o
f
E
NN
m
o
d
el
is
test
ed
o
n
m
o
d
elin
g
a
p
iezo
elec
tr
ic
ac
tu
at
o
r
an
d
a
q
u
ad
r
u
p
le
tan
k
s
y
s
te
m
.
T
h
e
r
esu
lt
s
p
r
o
v
e
th
at
th
e
E
NN
m
o
d
el
h
as
t
h
e
ab
ilit
y
to
lear
n
th
e
d
y
n
a
m
ic
s
m
o
d
el
to
r
ed
u
ce
th
e
tr
ac
k
in
g
er
r
o
r
to
n
ea
r
l
y
ze
r
o
.
T
h
e
r
esu
lts
w
ill
b
e
u
s
ed
to
in
itialize
th
e
d
esig
n
o
f
th
e
m
o
d
el
-
b
ased
co
n
tr
o
ller
f
o
r
n
o
n
li
n
ea
r
s
y
s
te
m
s
i
n
t
h
e
f
u
t
u
r
e
r
esear
ch
.
ACK
NO
WL
E
D
G
E
M
E
NT
S
T
h
e
s
tu
d
y
w
as
s
u
p
p
o
r
ted
b
y
Scien
ce
an
d
T
ec
h
n
o
lo
g
y
I
n
cu
b
ato
r
Yo
u
th
P
r
o
g
r
am
,
m
a
n
ag
ed
b
y
th
e
C
en
ter
f
o
r
Scie
n
ce
a
n
d
T
ec
h
n
o
lo
g
y
Dev
e
lo
p
m
en
t,
Ho
C
h
i
Min
h
C
o
m
m
u
n
i
s
t
Yo
u
th
U
n
io
n
,
t
h
e
co
n
tr
ac
t
n
u
m
b
er
is
"
1
3
/2
0
1
8
/HĐ
-
KHC
N
-
V
Ư
"
.
0
500
1000
1500
2000
0
0
.
5
1
L
2
(
s
o
l
i
d
)
a
n
d
L
2
_
h
a
t
(
d
a
s
h
e
d
)
o
u
t
p
u
t
0
500
1000
1500
2000
-
0
.
5
0
0
.
5
t
i
m
e
(
s
a
m
p
l
e
s
)
p
r
e
d
i
c
t
i
o
n
e
r
r
o
r
0
500
1000
1500
2000
0
0
.
5
1
L
4
(
s
o
l
i
d
)
a
n
d
L
4
_
h
a
t
(
d
a
s
h
e
d
)
0
500
1000
1500
2000
-
0
.
5
0
0
.
5
t
i
m
e
(
s
a
m
p
l
e
s
)
0
500
1000
1500
2000
0
0
.
5
1
L
2
(
s
o
l
i
d
)
a
n
d
L
2
_
h
a
t
(
d
a
s
h
e
d
)
o
u
t
p
u
t
0
500
1000
1500
2000
-
0
.
5
0
0
.
5
t
i
m
e
(
s
a
m
p
l
e
s
)
p
r
e
d
i
c
t
i
o
n
e
r
r
o
r
0
500
1000
1500
2000
0
0
.
5
1
L
4
(
s
o
l
i
d
)
a
n
d
L
4
_
h
a
t
(
d
a
s
h
e
d
)
0
500
1000
1500
2000
-
0
.
5
0
0
.
5
t
i
m
e
(
s
a
m
p
l
e
s
)
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:
2088
-
8708
B
la
ck
-
b
o
x
m
o
d
elin
g
o
f n
o
n
lin
ea
r
s
ystem
u
s
in
g
ev
o
lu
tio
n
a
r
y
n
eu
r
a
l NAR
X
mo
d
el
(
N
g
u
ye
n
N
g
o
c
S
o
n
)
1869
RE
F
E
R
E
NC
E
S
[1
]
L
ju
n
g
,
L
e
n
n
a
rt
,
"
S
y
st
e
m
id
e
n
ti
f
ic
a
ti
o
n
,
"
S
ig
n
a
l
a
n
a
lys
is
a
n
d
p
re
d
ic
ti
o
n
,
p
p
.
1
6
3
-
1
7
3
,
1
9
9
8
.
[2
]
W
a
n
g
,
L
in
,
Yi
Zen
g
,
a
n
d
T
a
o
Ch
e
n
,
"
Ba
c
k
p
ro
p
a
g
a
ti
o
n
n
e
u
ra
l
n
e
tw
o
rk
w
it
h
a
d
a
p
ti
v
e
d
iff
e
r
e
n
ti
a
l
e
v
o
lu
ti
o
n
a
lg
o
rit
h
m
f
o
r
ti
m
e
se
ries
f
o
re
c
a
stin
g
.
"
Exp
e
rt S
y
ste
ms
wit
h
A
p
p
l
ica
t
io
n
s
,
v
o
l
4
2
(2
)
,
p
p
.
8
5
5
-
8
6
3
,
2
0
1
5
.
[3
]
P
i
o
tr
o
w
sk
i,
A
d
a
m
P
.
,
"
Diff
e
re
n
ti
a
l
e
v
o
lu
ti
o
n
a
lg
o
rit
h
m
s
a
p
p
l
ied
to
n
e
u
ra
l
n
e
tw
o
rk
train
in
g
su
ffe
r
f
ro
m
sta
g
n
a
ti
o
n
.
"
Ap
p
li
e
d
S
o
ft
C
o
mp
u
ti
n
g
,
v
o
l.
2
1
,
p
p
.
3
8
2
-
4
0
6
,
2
0
1
4
.
[4
]
Yu
,
F
e
n
g
,
a
n
d
X
iao
z
h
o
n
g
X
u
,
"
A
sh
o
rt
-
term
lo
a
d
f
o
re
c
a
stin
g
m
o
d
e
l
o
f
n
a
tu
ra
l
g
a
s
b
a
se
d
o
n
o
p
t
im
iz
e
d
g
e
n
e
ti
c
a
lg
o
rit
h
m
a
n
d
im
p
ro
v
e
d
B
P
n
e
u
ra
l
n
e
tw
o
rk
,
"
Ap
p
li
e
d
En
e
rg
y
,
v
o
l.
1
3
4
,
p
p
.
1
0
2
-
1
1
3
,
2
0
1
4
.
[5
]
L
e
u
n
g
,
F
ra
n
k
Hu
n
g
-
F
a
t,
e
t
a
l
.
,
"
T
u
n
in
g
o
f
th
e
stru
c
tu
re
a
n
d
p
a
ra
m
e
ters
o
f
a
n
e
u
ra
l
n
e
tw
o
rk
u
sin
g
a
n
im
p
ro
v
e
d
g
e
n
e
ti
c
a
lg
o
rit
h
m
,
"
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
Ne
u
ra
l
n
e
two
rk
s
,
v
o
l
.
1
4
(1
),
p
p
.
79
-
88
,
2
0
0
3
.
[6
]
A
s
a
d
i,
Eh
sa
n
,
e
t
a
l
.
"
M
u
lt
i
-
o
b
jec
ti
v
e
o
p
ti
m
iza
ti
o
n
f
o
r
b
u
il
d
in
g
re
tro
f
it
:
A
m
o
d
e
l
u
sin
g
g
e
n
e
ti
c
a
lg
o
rit
h
m
a
n
d
a
rti
f
icia
l
n
e
u
ra
l
n
e
tw
o
rk
a
n
d
a
n
a
p
p
li
c
a
ti
o
n
,
"
E
n
e
rg
y
a
n
d
Bu
il
d
in
g
s
,
v
o
l.
8
1
,
p
p
.
4
4
4
-
4
5
6
,
2
0
1
4
.
[7
]
S
o
n
g
,
Yi
n
g
,
Zen
g
q
ia
n
g
Ch
e
n
,
a
n
d
Zh
u
z
h
i
Yu
a
n
,
"
Ne
w
c
h
a
o
ti
c
P
S
O
-
b
a
se
d
n
e
u
ra
l
n
e
tw
o
rk
p
re
d
ict
iv
e
c
o
n
tro
l
f
o
r
n
o
n
li
n
e
a
r
p
ro
c
e
ss
,
"
IEE
E
tra
n
sa
c
t
io
n
s
o
n
n
e
u
ra
l
n
e
tw
o
rk
s
,
v
o
l.
1
8
(
2
),
p
p
.
5
9
5
-
6
0
1
,
2
0
0
7
.
[8
]
Ba
sh
ir
Z.
A
.
,
a
n
d
M
.
E.
E
l
-
Ha
wa
r
y
,
"
A
p
p
l
y
in
g
wa
v
e
lets
to
sh
o
rt
-
term
lo
a
d
f
o
re
c
a
stin
g
u
sin
g
P
S
O
-
b
a
se
d
n
e
u
r
a
l
n
e
tw
o
rk
s
,
"
IEE
E
tra
n
sa
c
ti
o
n
s
o
n
p
o
we
r sy
ste
ms
,
v
o
l.
2
4
(1
)
,
p
p
.
20
-
27
,
2
0
0
9
.
[9
]
G
o
rd
a
n
,
Be
h
ro
u
z
,
e
t
a
l
,
"
P
re
d
icti
o
n
o
f
se
is
m
ic
slo
p
e
sta
b
il
it
y
th
ro
u
g
h
c
o
m
b
in
a
ti
o
n
o
f
p
a
rti
c
le
sw
a
r
m
o
p
ti
m
iz
a
ti
o
n
a
n
d
n
e
u
ra
l
n
e
tw
o
rk
,
"
En
g
i
n
e
e
rin
g
wit
h
C
o
mp
u
ter
s
,
3
2
(1
),
p
p
.
85
-
97
,
2
0
1
6
.
[1
0
]
V
a
li
a
n
,
Eh
sa
n
,
S
h
a
h
ra
m
M
o
h
a
n
n
a
,
a
n
d
S
a
e
e
d
T
a
v
a
k
o
li
,
"
Im
p
ro
v
e
d
c
u
c
k
o
o
se
a
rc
h
a
lg
o
rit
h
m
f
o
r
fe
e
d
f
o
rw
a
rd
n
e
u
ra
l
n
e
tw
o
rk
train
in
g
,
"
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
Arti
fi
c
ia
l
In
telli
g
e
n
c
e
&
Ap
p
li
c
a
ti
o
n
s
,
v
o
l.
2
(3
)
,
p
p
.
36
-
43
,
2
0
1
1
.
[1
1
]
S
to
rn
,
Ra
in
e
r,
a
n
d
Ke
n
n
e
th
P
rice
,
"
Di
ff
e
re
n
ti
a
l
e
v
o
lu
ti
o
n
–
a
sim
p
le
a
n
d
e
ff
icie
n
t
h
e
u
risti
c
f
o
r
g
lo
b
a
l
o
p
ti
m
iza
ti
o
n
o
v
e
r
c
o
n
ti
n
u
o
u
s sp
a
c
e
s
,
"
J
o
u
rn
a
l
o
f
g
lo
b
a
l
o
p
ti
miza
ti
o
n
,
v
o
l.
1
1
(4
),
p
p
.
3
4
1
-
3
5
9
,
1
9
9
7
.
[1
2
]
Ilo
n
e
n
,
Ja
rm
o
,
Jo
n
i
-
Krist
ian
Ka
m
a
ra
in
e
n
,
a
n
d
Jo
u
n
i
L
a
m
p
in
e
n
,
"
D
iff
e
re
n
ti
a
l
e
v
o
lu
ti
o
n
train
in
g
a
lg
o
rit
h
m
f
o
r
f
e
e
d
-
f
o
r
w
a
rd
n
e
u
ra
l
n
e
tw
o
rk
s
,
"
Ne
u
ra
l
Pro
c
e
ss
in
g
L
e
tt
e
rs
,
v
o
l
.
1
7
(1
),
p
p
.
93
-
1
0
5
,
2
0
0
3
.
[1
3
]
Ba
ştü
rk
,
A
lp
e
r,
a
n
d
En
is
G
ü
n
a
y
,
"
E
ff
icie
n
t
e
d
g
e
d
e
tec
ti
o
n
in
d
ig
it
a
l
im
a
g
e
s
u
sin
g
a
c
e
ll
u
lar
n
e
u
ra
l
n
e
tw
o
rk
o
p
ti
m
ize
d
b
y
d
iff
e
re
n
ti
a
l
e
v
o
lu
ti
o
n
a
lg
o
rit
h
m
,
"
Exp
e
rt S
y
ste
ms
wit
h
Ap
p
li
c
a
ti
o
n
s
,
v
o
l.
3
6
(
2
),
p
p
.
2
6
4
5
-
2
6
5
0
,
2
0
0
9
.
[1
4
]
S
u
b
u
d
h
i,
Bid
y
a
d
h
a
r,
a
n
d
De
b
a
sh
ish
a
Je
n
a
,
"
A
d
iffere
n
ti
a
l
e
v
o
lu
ti
o
n
b
a
se
d
n
e
u
ra
l
n
e
tw
o
rk
a
p
p
ro
a
c
h
to
n
o
n
li
n
e
a
r
s
y
ste
m
id
e
n
ti
f
ica
ti
o
n
,
"
Ap
p
li
e
d
S
o
ft
Co
mp
u
ti
n
g
,
v
o
l.
1
1
(
1
),
p
p
.
8
6
1
-
871
,
2
0
1
1
.
[1
5
]
Ch
a
u
h
a
n
,
Nik
u
n
j,
V
a
d
lam
a
n
i
Ra
v
i,
a
n
d
D.
Ka
rth
ik
Ch
a
n
d
ra
,
"
Di
ff
e
re
n
ti
a
l
e
v
o
lu
ti
o
n
train
e
d
w
a
v
e
let
n
e
u
ra
l
n
e
tw
o
rk
s:
A
p
p
li
c
a
ti
o
n
to
b
a
n
k
ru
p
tcy
p
re
d
ictio
n
in
b
a
n
k
s
,
"
E
x
p
e
rt
S
y
ste
ms
wit
h
Ap
p
li
c
a
ti
o
n
s
,
v
o
l.
3
6
(4
)
,
pp.
76
59
-
7
6
6
5
,
2
0
0
9
.
[1
6
]
W
a
n
g
,
L
in
,
Yi
Zen
g
,
a
n
d
T
a
o
Ch
e
n
,
"
Ba
c
k
p
ro
p
a
g
a
ti
o
n
n
e
u
ra
l
n
e
tw
o
rk
w
it
h
a
d
a
p
ti
v
e
d
iff
e
r
e
n
ti
a
l
e
v
o
lu
ti
o
n
a
lg
o
rit
h
m
f
o
r
ti
m
e
se
ries
f
o
re
c
a
stin
g
,
"
Exp
e
rt S
y
ste
ms
wit
h
A
p
p
l
ica
t
io
n
s
v
o
l
.
4
2
(2
)
,
p
p
.
8
5
5
-
8
6
3
,
2
0
1
5
.
[1
7
]
L
iu
,
W
e
i,
e
t
a
l
.
,
"
Ne
u
ra
l
n
e
tw
o
r
k
b
a
se
d
o
n
se
lf
-
a
d
a
p
ti
v
e
d
if
f
e
r
e
n
ti
a
l
e
v
o
lu
t
io
n
f
o
r
u
lt
ra
-
sh
o
rt
-
ter
m
p
o
we
r
lo
a
d
f
o
re
c
a
stin
g
,
"
In
ter
n
a
ti
o
n
a
l
c
o
n
fer
e
n
c
e
o
n
i
n
telli
g
e
n
t
c
o
m
p
u
t
in
g
.
S
p
rin
g
e
r,
Ch
a
m
,
2
0
1
4
.
[1
8
]
Ch
e
n
,
C
h
e
n
g
-
Hu
n
g
,
Ch
e
n
g
-
Jia
n
L
in
,
a
n
d
Ch
i
n
-
T
e
n
g
L
in
,
"
No
n
li
n
e
a
r
sy
ste
m
c
o
n
tro
l
u
sin
g
a
d
a
p
ti
v
e
n
e
u
ra
l
f
u
z
z
y
n
e
tw
o
rk
s
b
a
se
d
o
n
a
m
o
d
if
ied
d
iff
e
r
e
n
ti
a
l
e
v
o
lu
ti
o
n
.
"
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
S
y
ste
ms
,
M
a
n
,
a
n
d
Cy
b
e
rn
e
ti
c
s,
Pa
r
t
C
(
Ap
p
li
c
a
ti
o
n
s
a
n
d
Rev
iews
)
,
v
o
l.
3
9
(4
),
p
p
.
4
5
9
-
4
7
3
,
2
0
0
9
,
[1
9
]
S
lo
w
ik
,
A
d
a
m
,
"
A
p
p
li
c
a
ti
o
n
o
f
a
n
a
d
a
p
ti
v
e
d
if
f
e
r
e
n
ti
a
l
e
v
o
lu
ti
o
n
a
lg
o
rit
h
m
w
it
h
m
u
lt
ip
le
tri
a
l
v
e
c
to
rs
to
a
rti
f
icia
l
n
e
u
ra
l
n
e
tw
o
rk
train
i
n
g
,
"
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
I
n
d
u
stria
l
El
e
c
tro
n
ics
,
5
8
(8
),
p
p
.
3
1
6
0
-
3
1
6
7
,
2
0
1
1
.
[2
0
]
Ku
m
a
r,
Jiten
d
ra
,
a
n
d
A
sh
u
to
s
h
Ku
m
a
r
S
in
g
h
,
"
W
o
rk
lo
a
d
p
re
d
ictio
n
in
c
l
o
u
d
u
si
n
g
a
rti
f
icia
l
n
e
u
ra
l
n
e
tw
o
rk
a
n
d
a
d
a
p
ti
v
e
d
if
f
e
r
e
n
ti
a
l
e
v
o
lu
ti
o
n
,
"
F
u
tu
re
Ge
n
e
ra
ti
o
n
Co
mp
u
ter
S
y
ste
ms
,
v
o
l.
8
1
,
p
p
.
41
-
52
,
2
0
1
8
.
[2
1
]
Ng
u
y
e
n
,
S
o
n
Ng
o
c
,
V
in
h
Ho
‐
Hu
u
,
a
n
d
A
n
h
P
h
a
m
Hu
y
Ho
,
"
A
n
e
u
ra
l
d
if
fe
re
n
ti
a
l
e
v
o
lu
ti
o
n
id
e
n
ti
f
ica
ti
o
n
a
p
p
ro
a
c
h
to
n
o
n
li
n
e
a
r
sy
ste
m
s an
d
m
o
d
e
ll
in
g
o
f
sh
a
p
e
m
e
m
o
r
y
a
ll
o
y
a
c
tu
a
to
r
,
"
Asia
n
J
o
u
rn
a
l
o
f
Co
n
tro
l
,
2
0
1
8
.
[2
2
]
Na
h
,
S
.
K.,
a
n
d
Z.
W
.
Zh
o
n
g
,
"
A
m
icro
g
rip
p
e
r
u
si
n
g
p
iez
o
e
lec
tri
c
a
c
tu
a
ti
o
n
f
o
r
m
icro
-
o
b
jec
t
m
a
n
ip
u
lati
o
n
,
"
S
e
n
so
rs
a
n
d
Actu
a
to
rs
A:
P
h
y
sic
a
l
,
v
o
l
.
1
3
3
(
1
),
p
p
.
218
-
2
2
4
,
2
0
0
7
.
[2
3
]
Bo
w
e
n
,
C.
R.
,
e
t
a
l
.
,
"
P
iez
o
e
lec
tri
c
a
n
d
f
e
rro
e
lec
tri
c
m
a
teria
ls
a
n
d
str
u
c
tu
re
s
f
o
r
e
n
e
rg
y
h
a
rv
e
stin
g
a
p
p
li
c
a
ti
o
n
s
,
"
En
e
rg
y
&
En
v
iro
n
me
n
ta
l
S
c
ien
c
e
,
v
o
l.
7
(1
)
,
p
p
.
25
-
44
,
2
0
1
4
.
[2
4
]
Ng
u
y
e
n
,
Cu
o
n
g
H.,
Ulrik
Ha
n
k
e
,
a
n
d
Ei
n
a
r
Ha
lv
o
rse
n
,
"
A
c
tu
a
ti
o
n
o
f
P
iez
o
e
lec
tri
c
L
a
y
e
r
e
d
Be
a
m
s
w
it
h
d
3
1
a
n
d
d
3
3
Co
u
p
l
in
g
,
"
IEE
E
T
ra
n
s
a
c
ti
o
n
s o
n
Ultr
a
so
n
ics
,
Fer
ro
e
lec
trics
,
a
n
d
Fre
q
u
e
n
c
y
Co
n
tr
o
l
,
2
0
1
8
.
[2
5
]
W
a
n
g
,
Ge
n
g
,
a
n
d
G
u
o
q
ian
g
Ch
e
n
,
"
Id
e
n
ti
f
ica
ti
o
n
o
f
p
iez
o
e
lec
tri
c
h
y
ste
re
sis
b
y
a
n
o
v
e
l
Du
h
e
m
m
o
d
e
l
b
a
se
d
n
e
u
ra
l
n
e
tw
o
rk
,
"
S
e
n
so
rs
a
n
d
Actu
a
to
rs
A:
Ph
y
sic
a
l
,
v
o
l.
2
6
4
,
p
p
.
2
8
2
-
2
8
8
,
2
0
1
7
.
[2
6
]
M
a
o
,
Xu
e
f
e
i,
e
t
a
l
.
,
"
A
H
y
b
rid
F
e
e
d
f
o
rwa
rd
-
F
e
e
d
b
a
c
k
H
y
ste
re
sis
Co
m
p
e
n
sa
to
r
in
P
iez
o
e
lec
tri
c
A
c
tu
a
to
rs
Ba
se
d
o
n
L
e
a
st
-
S
q
u
a
re
s
S
u
p
p
o
rt
V
e
c
to
r
M
a
c
h
in
e
,
"
IEE
E
T
r
a
n
sa
c
ti
o
n
s
o
n
I
n
d
u
stri
a
l
El
e
c
tro
n
ics
,
v
o
l
.
6
5
(7
),
p
p
.
5
7
0
4
-
5
7
1
1
,
2
0
1
8
.
[2
7
]
Kra
sn
o
se
l'
sk
ii
,
M
a
rk
A
.
,
a
n
d
Ale
k
se
i
V
.
P
o
k
ro
v
sk
ii
,
"
S
y
ste
m
s
w
it
h
h
y
ste
re
sis
,
"
S
p
rin
g
e
r
S
c
ien
c
e
&
Bu
sin
e
ss
M
e
d
ia
,
2
0
1
2
.
[2
8
]
Bo
u
o
u
d
e
n
S
.
,
M
o
h
a
m
m
e
d
Ch
a
d
li
,
a
n
d
Ha
m
id
Re
z
a
Ka
ri
m
i
,
"
A
n
a
n
t
c
o
lo
n
y
o
p
ti
m
iza
ti
o
n
-
b
a
se
d
f
u
z
z
y
p
re
d
ictiv
e
c
o
n
tro
l
a
p
p
ro
a
c
h
f
o
r
n
o
n
li
n
e
a
r
p
r
o
c
e
ss
e
s
,
"
In
fo
rm
a
ti
o
n
S
c
ien
c
e
s
,
v
o
l.
2
9
9
,
p
p
.
1
4
3
-
1
5
8
,
2
0
1
5
.
[2
9
]
W
a
n
g
,
T
o
n
g
,
Hu
ij
u
n
G
a
o
,
a
n
d
Ji
a
n
b
in
Qiu
,
"
A
c
o
m
b
in
e
d
a
d
a
p
ti
v
e
n
e
u
ra
l
n
e
tw
o
rk
a
n
d
n
o
n
li
n
e
a
r
m
o
d
e
l
p
re
d
ictiv
e
c
o
n
tro
l
f
o
r
m
u
lt
irate
n
e
tw
o
rk
e
d
in
d
u
str
ial
p
r
o
c
e
ss
c
o
n
tro
l
,
"
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
Ne
u
r
a
l
Ne
two
rk
s
a
n
d
L
e
a
rn
i
n
g
S
y
ste
ms
,
v
o
l.
2
7
(2
),
p
p
.
4
1
6
-
4
2
5
,
2
0
1
6
.
[3
0
]
Ip
li
k
c
i,
S
e
rd
a
r
,
"
A
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
-
b
a
se
d
c
o
n
tro
l
a
p
p
li
c
a
ti
o
n
to
th
e
e
x
p
e
rim
e
n
tal
th
re
e
-
tan
k
s
y
ste
m
,
"
IS
A
tran
sa
c
ti
o
n
s,
v
o
l
.
4
9
(3
)
,
p
p
.
3
7
6
-
3
8
6
,
2
0
1
0
.
[3
1
]
Jo
h
a
n
ss
o
n
,
Ka
rl
He
n
rik
,
"
T
h
e
q
u
a
d
r
u
p
le
-
tan
k
p
ro
c
e
ss
:
A
m
u
lt
iv
a
riab
le
lab
o
ra
to
ry
p
ro
c
e
ss
w
it
h
a
n
a
d
j
u
sta
b
le
z
e
ro
,
"
IEE
E
T
ra
n
sa
c
ti
o
n
s
o
n
c
o
n
t
ro
l
sy
ste
ms
tec
h
n
o
l
o
g
y
,
v
o
l
.
8
(
3
),
p
p
.
4
5
6
-
465
,
2
0
0
0
.
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.
9
,
No
.
3
,
J
u
n
e
201
9
:
1
8
6
1
-
1
8
7
0
1870
B
I
O
G
RAP
H
I
E
S O
F
AUTH
O
RS
S
o
n
Ng
u
y
e
n
re
c
e
iv
e
d
h
is
M
.
S
c
.
a
n
d
Ph
.
D.
d
e
g
re
e
s
in
th
e
F
a
c
u
lt
y
o
f
El
e
c
tri
c
a
l
a
n
d
El
e
c
tro
n
ics
En
g
in
e
e
rin
g
(
F
EE
E
)
f
ro
m
Ho
Ch
i
M
i
n
h
Cit
y
Un
iv
e
rsit
y
o
f
T
e
c
h
n
o
lo
g
y
in
2
0
1
2
a
n
d
2
0
1
7
,
re
sp
e
c
ti
v
e
l
y
.
He
is
c
u
rre
n
tl
y
a
V
ice
-
De
a
n
o
f
th
e
F
a
c
u
lt
y
o
f
El
e
c
tro
n
ics
T
e
c
h
n
o
l
o
g
y
,
In
d
u
strial
Un
iv
e
rsit
y
o
f
Ho
Ch
i
M
in
h
Cit
y
,
V
iet
Na
m
.
His
c
u
rre
n
t
re
s
e
a
rc
h
in
tere
sts
in
c
lu
d
e
in
telli
g
e
n
t
c
o
n
tr
o
l,
r
o
b
o
ti
c
s,
i
d
e
n
ti
f
ica
ti
o
n
o
f
n
o
n
li
n
e
a
r
s
y
ste
m
s,
a
n
d
t
h
e
in
ter
n
e
t
o
f
th
i
n
g
s.
K
h
a
n
h
Ng
u
y
e
n
Du
y
r
e
c
e
iv
e
d
h
is
M
.
S
c
d
e
g
re
e
in
th
e
F
a
c
u
lt
y
o
f
T
e
l
e
c
o
m
m
u
n
ica
ti
o
n
s
f
ro
m
P
o
sts
a
n
d
T
e
lec
o
m
m
u
n
ica
ti
o
n
s
In
stit
u
t
e
o
f
Tec
h
n
o
lo
g
y
in
2
0
1
4
.
He
is
c
u
rre
n
tl
y
a
Ch
a
ir
m
a
n
o
f
S
tu
d
e
n
t
S
c
ien
c
e
Re
se
a
rc
h
Clu
b
o
f
F
a
c
u
lt
y
o
f
El
e
c
tro
n
ics
T
e
c
h
n
o
l
o
g
y
,
In
d
u
strial
Un
iv
e
rsity
o
f
Ho
Ch
i
M
in
h
Cit
y
,
V
iet
Na
m
.
His
re
se
a
r
c
h
in
t
e
re
sts
in
c
lu
d
e
in
tel
li
g
e
n
t
sy
ste
m
s
,
m
a
c
h
in
e
lea
rn
in
g
,
i
d
e
n
ti
f
ica
ti
o
n
s
y
ste
m
s,
a
n
d
t
h
e
in
ter
n
e
t
o
f
th
i
n
g
.
Tr
a
n
M
in
h
C
h
i
n
h
g
ra
d
u
a
ted
i
n
2
0
1
1
a
n
d
re
c
e
iv
e
d
Ph
.
D.
d
e
g
r
e
e
s
in
2
0
1
6
f
ro
m
T
a
m
b
o
v
S
tad
e
T
e
c
h
n
ica
l
Un
iv
e
rsit
y
,
Ru
ss
ia.
F
ro
m
9
/2
0
1
6
,
h
e
w
o
rk
s
a
s
a
te
a
c
h
e
r
in
t
h
e
F
a
c
u
lt
y
o
f
El
e
c
tro
n
ics
T
e
c
h
n
o
lo
g
y
,
In
d
u
strial
Un
iv
e
rsity
o
f
Ho
Ch
i
M
in
h
Cit
y
,
V
iet
Na
m
.
His res
e
a
r
c
h
in
tere
sts a
re
m
o
d
e
rn
c
o
n
tro
l,
i
n
telli
g
e
n
t
sy
ste
m
s,
a
rti
f
i
c
ial
ly
in
telli
g
e
n
ce
,
m
a
c
h
in
e
lea
rn
i
n
g
.
Evaluation Warning : The document was created with Spire.PDF for Python.