I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lec
t
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
9
,
No
.
4
,
A
u
g
u
s
t
201
9
,
p
p
.
2
4
8
1
~2
4
9
0
I
SS
N:
2
0
8
8
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v9
i
4
.
p
p
2
4
8
1
-
2490
2481
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
Ada
pted
b
ra
nch
-
a
nd
-
b
o
und
a
lg
o
rith
m
u
sing
S
VM
w
ith
m
o
del
s
elec
ti
o
n
K
a
bb
a
j
M
o
ha
m
e
d M
us
t
a
ph
a
1
,
E
l
A
f
ia
Abdella
t
if
2
ENS
IA
S
,
M
o
h
a
m
e
d
V
Un
iv
e
rsit
y
,
Ra
b
a
t,
M
o
ro
c
c
o
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ma
y
1
,
2
0
1
8
R
ev
i
s
ed
J
an
5
,
2
0
1
9
A
cc
ep
ted
Ma
r
4
,
2
0
1
9
Bra
n
c
h
-
a
n
d
-
B
o
u
n
d
a
lg
o
rit
h
m
is
t
h
e
b
a
sis
f
o
r
th
e
m
a
jo
rit
y
o
f
so
lv
in
g
m
e
th
o
d
s
in
m
ix
e
d
in
teg
e
r
li
n
e
a
r
p
ro
g
ra
m
m
in
g
.
It
h
a
s
b
e
e
n
p
ro
v
in
g
it
s
e
ff
icie
n
c
y
in
d
if
fe
re
n
t
f
ield
s.
In
f
a
c
t,
it
c
re
a
tes
li
tt
le
b
y
li
tt
le
a
tree
o
f
n
o
d
e
s
b
y
a
d
o
p
ti
n
g
tw
o
stra
te
g
ies
.
T
h
e
se
stra
te
g
ies
a
re
v
a
riab
le
se
lec
ti
o
n
stra
teg
y
a
n
d
n
o
d
e
se
lec
ti
o
n
stra
teg
y
.
In
o
u
r
p
re
v
io
u
s
w
o
rk
,
we
e
x
p
e
rien
c
e
d
a
m
e
th
o
d
o
lo
g
y
o
f
lea
rn
in
g
b
ra
n
c
h
-
a
n
d
-
b
o
u
n
d
stra
te
g
ies
u
sin
g
re
g
re
ss
io
n
-
b
a
se
d
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
tw
ic
e
.
T
h
a
t
m
e
th
o
d
o
lo
g
y
a
ll
o
w
e
d
f
ir
stl
y
to
e
x
p
lo
it
in
f
o
rm
a
ti
o
n
f
ro
m
p
re
v
io
u
s
e
x
e
c
u
ti
o
n
s
o
f
Bra
n
c
h
-
a
n
d
-
Bo
u
n
d
a
lg
o
rit
h
m
o
n
o
th
e
r
in
sta
n
c
e
s.
S
e
c
o
n
d
ly
,
it
c
re
a
ted
in
f
o
rm
a
ti
o
n
c
h
a
n
n
e
l
b
e
tw
e
e
n
n
o
d
e
se
lec
ti
o
n
s
trate
g
y
a
n
d
v
a
riab
le
b
ra
n
c
h
in
g
stra
teg
y
.
A
n
d
th
ir
d
ly
,
it
g
a
v
e
g
o
o
d
re
su
lt
s
in
term
o
f
ru
n
n
in
g
ti
m
e
c
o
m
p
a
rin
g
to
sta
n
d
a
rd
Bra
n
c
h
-
a
n
d
-
B
o
u
n
d
a
lg
o
rit
h
m
.
In
th
is
w
o
rk
,
w
e
w
il
l
f
o
c
u
s
o
n
in
c
re
a
sin
g
S
VM
p
e
rf
o
rm
a
n
c
e
b
y
u
sin
g
c
ro
ss
v
a
li
d
a
ti
o
n
c
o
u
p
led
w
it
h
m
o
d
e
l
se
lec
ti
o
n
.
K
ey
w
o
r
d
s
:
No
d
e
s
elec
tio
n
s
tr
ate
g
y
Var
iab
le
b
r
an
ch
in
g
s
tr
ate
g
y
B
r
an
ch
an
d
b
o
u
n
d
SVM
C
r
o
s
s
v
a
lid
atio
n
M
o
d
el
s
elec
tio
n
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
:
K
ab
b
aj
Mo
h
a
m
ed
Mu
s
tap
h
a,
E
NSI
AS,
Un
i
v
er
s
i
t
y
Mo
h
a
m
e
d
V,
A
b
d
ellah
R
e
g
r
ag
u
i Str
ee
t,
Ma
d
in
at
AL
I
r
f
an
e,
R
a
b
at,
Mo
r
o
c
co
.
E
m
ail:
m
u
s
tap
h
a.
k
ab
b
aj
@
u
m
5
s
.
n
et.
m
a
1.
I
NT
RO
D
UCT
I
O
N
I
n
r
ea
l
lif
e,
MI
L
P
h
a
s
co
u
n
tles
s
ap
p
licatio
n
s
i
n
d
if
f
er
en
t
f
ield
s
li
k
e
lo
g
is
tic
s
,
f
in
an
ce
an
d
tr
an
s
p
o
r
tatio
n
.
A
v
er
y
co
m
m
o
n
s
o
l
u
tio
n
tech
n
iq
ue
o
f
MI
L
P
f
r
a
m
e
w
o
r
k
is
B
r
an
ch
-
a
n
d
-
B
o
u
n
d
.
I
t
co
n
ti
n
u
e
s
to
p
r
o
v
e
its
r
elev
an
ce
n
o
w
ad
a
y
s
.
B
r
an
ch
-
a
n
d
-
B
o
u
n
d
alg
o
r
ith
m
is
a
n
iter
ativ
e
al
g
o
r
ith
m
,
a
n
d
at
ea
ch
iter
atio
n
,
w
e
e
v
e
n
tu
a
ll
y
g
et
a
f
ea
s
ib
le
o
r
o
p
tim
al
s
o
lu
tio
n
o
f
a
n
i
n
iti
al
p
r
o
b
lem
.
C
o
n
cr
etel
y
,
t
h
e
al
g
o
r
ith
m
co
n
s
tr
u
ct
s
litt
le
b
y
litt
le
a
tr
ee
o
f
n
o
d
es,
w
h
er
e
ea
c
h
n
o
d
e
r
ep
r
esen
t
s
a
m
o
d
if
ied
v
er
s
io
n
o
f
t
h
e
o
r
i
g
in
a
l
p
r
o
b
le
m
.
T
h
e
co
n
s
tr
u
ct
io
n
o
f
c
h
ild
n
o
d
es
is
co
n
d
u
cted
b
y
a
v
ar
iab
le
b
r
an
ch
in
g
s
tr
ate
g
y
.
An
o
t
h
er
f
u
n
d
a
m
en
tal
ele
m
e
n
t
i
n
B
r
an
ch
-
a
n
d
-
B
o
u
n
d
alg
o
r
i
t
h
m
is
No
d
e
Selectio
n
Stra
te
g
y
th
at
ai
m
s
to
c
h
o
o
s
e
a
m
o
n
g
a
n
o
d
es
q
u
eu
e,
o
n
e
t
h
at
w
il
l sp
ee
d
u
p
th
e
s
ea
r
c
h
.
R
ec
en
t
l
y
,
s
o
m
e
w
o
r
k
s
h
as
b
e
en
tr
y
in
g
to
id
en
t
if
y
a
n
a
n
al
y
t
ic
ap
p
r
o
ac
h
th
at
d
ec
id
e
ab
o
u
t
s
tr
ate
g
ies
d
escr
ib
ed
ab
o
v
e,
g
iv
e
n
a
s
et
o
f
p
r
o
b
le
m
f
ea
tu
r
e
s
.
Au
th
o
r
s
u
s
e
li
k
el
y
m
ac
h
i
n
e
lear
n
i
n
g
tec
h
n
iq
u
es.
T
h
e
m
ai
n
r
e
m
ar
k
is
t
h
at
f
e
w
a
u
t
h
o
r
s
w
h
o
d
ea
l
w
it
h
n
o
d
e
s
elec
tio
n
s
tr
a
teg
y
,
a
n
d
if
s
o
,
th
e
y
d
id
n
o
t
u
s
e
m
ac
h
i
n
e
lear
n
i
n
g
f
r
a
m
e
w
o
r
k
.
Ou
r
co
n
tr
ib
u
tio
n
i
s
o
r
ien
ted
to
w
ar
d
s
lear
n
in
g
e
f
f
icien
t
b
r
an
ch
-
a
n
d
-
b
o
u
n
d
s
tr
ate
g
ies.
T
h
is
i
s
t
h
e
r
esu
lt
o
f
a
co
n
s
i
s
te
n
t
m
et
h
o
d
o
lo
g
y
b
eg
i
n
n
in
g
w
it
h
th
e
co
lle
ctio
n
o
f
th
e
d
ata
s
et,
an
d
en
d
i
n
g
w
it
h
t
h
e
test
o
f
th
e
f
in
al
h
y
p
o
t
h
esi
s
.
Mo
r
e
ex
p
licitl
y
,
w
e:
-
Def
i
n
e
f
ea
t
u
r
es
-
C
o
llect
d
ata
s
et
-
P
ick
th
e
o
p
ti
m
al
lear
n
in
g
m
o
d
el
-
L
ea
r
n
t
h
e
f
in
a
l
h
y
p
o
t
h
esi
s
w
it
h
th
e
c
h
o
s
e
n
m
o
d
el
-
I
m
p
le
m
e
n
t a
n
d
t
est th
e
f
i
n
al
h
y
p
o
t
h
esi
s
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
.
4
,
A
u
g
u
s
t 2
0
1
9
:
2
4
8
1
-
2490
2482
Nex
t,
w
e
ad
d
r
ess
r
e
s
ea
r
ch
p
ap
er
s
r
elati
v
e
to
o
u
r
w
o
r
k
.
T
o
d
o
s
o
,
w
e
d
iv
id
e
co
n
tr
ib
u
tio
n
s
in
f
i
v
e
s
u
b
-
s
ec
tio
n
s
r
elati
v
el
y
to
th
e
s
tr
at
eg
y
a
n
d
to
th
e
lear
n
i
n
g
tec
h
n
i
q
u
e
u
s
ed
.
Firstl
y
,
r
elati
v
el
y
to
v
ar
iab
le
b
r
an
ch
in
g
s
tr
ateg
y
,
au
t
h
o
r
s
in
[
1
]
,
lear
n
ed
a
f
u
n
ctio
n
to
b
e
w
it
h
ap
p
r
o
x
im
a
tiv
e
l
y
t
h
e
s
a
m
e
p
er
f
o
r
m
a
n
ce
as
s
tr
o
n
g
b
r
an
ch
i
n
g
i
n
ter
m
o
f
p
r
ec
is
io
n
,
an
d
in
th
e
s
a
m
e
t
i
m
e,
r
es
u
lt
s
in
a
g
ai
n
o
f
t
h
e
p
r
o
ce
s
s
in
g
ti
m
e.
I
n
th
i
s
s
a
m
e
ca
teg
o
r
y
,
w
e
ci
te
[
2
]
,
w
h
er
ein
a
u
t
h
o
r
s
in
f
er
co
n
s
i
s
te
n
t
d
ata
f
r
o
m
ap
p
l
y
i
n
g
an
alg
o
r
ith
m
f
o
r
d
etec
tin
g
cla
u
s
e
s
.
A
cla
u
s
e
b
ein
g
a
co
m
b
in
a
tio
n
o
f
b
in
ar
y
v
al
u
es
af
f
ec
ted
to
a
s
e
t
o
f
in
d
ex
ed
v
ar
iab
les,
th
at
i
f
it
h
ap
p
en
s
,
th
e
w
h
o
le
p
r
o
b
lem
w
o
u
ld
b
e
in
f
ea
s
ib
l
e.
I
n
ad
d
itio
n
,
th
at
a
lg
o
r
it
h
m
g
en
er
ate
s
m
in
i
m
a
l
clau
s
e
s
a
n
d
r
estar
ts
w
i
th
ac
ti
v
e
clau
s
e
s
(
“
t
h
o
s
e
t
h
at
ca
n
b
e
u
s
ed
i
n
f
at
h
o
m
i
n
g
ch
ild
n
o
d
e
s
”)
.
I
n
p
ar
allel,
t
h
is
in
f
o
r
m
atio
n
i
s
u
s
ed
to
ch
o
o
s
e
b
r
an
ch
in
g
v
ar
iab
le
w
it
h
t
h
e
b
est
ef
f
ec
t.
Als
o
[
3
]
an
d
[
4
]
u
s
e
b
ac
k
tr
ac
k
i
n
g
t
o
i
m
p
r
o
v
e
b
r
an
ch
i
n
g
d
ec
is
io
n
s
.
B
esid
es,
th
er
e
ar
e
cla
s
s
ic
v
ar
i
ab
le
b
r
an
ch
i
n
g
r
u
les,
li
k
e
s
tr
o
n
g
b
r
an
c
h
i
n
g
,
p
s
eu
d
o
co
s
t
s
b
r
an
ch
i
n
g
,
h
y
b
r
id
b
r
an
c
h
i
n
g
,
r
el
iab
le
b
r
an
ch
i
n
g
,
i
n
f
er
en
ce
-
b
ased
b
r
an
c
h
in
g
[
5
]
.
No
te
th
at
r
eliab
ilit
y
b
r
an
ch
i
n
g
is
k
n
o
wn
to
b
e
th
e
b
est
b
r
an
ch
i
n
g
r
u
le
w
it
h
t
h
e
r
eliab
ilit
y
=
4
8
[
6
]
.
Fo
r
o
u
r
ex
p
er
i
m
e
n
ts
,
w
e
u
s
ed
:
=
8
.
T
h
e
r
eliab
ilit
y
p
ar
a
m
eter
is
f
ix
ed
to
s
to
p
t
h
e
ca
lc
u
latio
n
o
f
p
s
e
u
d
o
co
s
t
v
al
u
e
s
a
f
ter
atta
in
i
n
g
a
ce
r
tai
n
lev
el
o
f
t
h
e
b
r
an
ch
a
n
d
b
o
u
n
d
tr
ee
.
T
h
is
is
b
ec
au
s
e
p
s
e
u
d
o
co
s
t
v
alu
e
r
e
m
ai
n
ap
p
r
o
x
i
m
ati
v
el
y
co
n
s
ta
n
t
af
ter
ca
lcu
lati
n
g
it se
v
er
al
ti
m
es
f
o
r
a
d
eter
m
i
n
ed
v
ar
iab
le.
Seco
n
d
l
y
,
w
e
cite
f
r
o
m
n
o
d
e
s
elec
tio
n
s
tr
ateg
y
li
ter
atu
r
e
,
in
ad
d
itio
n
to
cla
s
s
ic
n
o
d
e
s
elec
tio
n
s
tr
ateg
ie
s
,
s
u
c
h
as
d
ep
th
-
f
ir
s
t
r
u
le,
b
r
ea
d
th
-
f
ir
s
t
r
u
le,
a
n
d
n
o
d
e
b
est
-
e
s
ti
m
ate
[
5
]
,
au
t
h
o
r
s
o
f
[
1
]
,
ex
tr
ac
ted
in
f
o
r
m
atio
n
f
r
o
m
MI
L
P
B
en
ch
m
ar
k
lib
r
ar
ies
b
y
u
s
i
n
g
s
p
ec
if
ic
al
g
o
r
ith
m
ca
lled
o
r
ac
le.
T
h
ir
d
l
y
,
co
n
ce
r
n
i
n
g
lear
n
in
g
a
lg
o
r
it
h
m
s
,
t
h
e
y
ar
e
u
s
ed
in
d
if
f
er
en
t
e
n
g
i
n
ee
r
i
n
g
f
ield
s
.
Alg
o
r
it
h
m
s
p
u
r
p
o
s
es
ar
e
clas
s
i
f
icatio
n
,
r
eg
r
ess
io
n
,
cl
u
s
ter
i
n
g
[
7
]
[
8
]
.
T
h
er
e
ar
e
alg
o
r
ith
m
s
t
h
at
te
n
d
to
d
o
w
el
l
i
n
p
r
ac
tice
m
o
r
e
th
a
n
o
t
h
er
s
[
9
]
[
1
0
]
.
I
n
th
e
s
a
m
e
co
n
te
x
t o
f
ap
p
l
y
in
g
lear
n
i
n
g
in
b
r
an
c
h
an
d
b
o
u
n
d
,
th
e
E
x
tr
aT
r
ee
s
is
ap
p
lied
in
[
1
1
]
.
Fo
u
r
th
l
y
,
w
h
e
n
lo
o
k
i
n
g
a
t
m
o
d
el
s
elec
tio
n
a
n
d
t
h
e
p
er
f
o
r
m
an
ce
o
f
al
g
o
r
it
h
m
s
,
t
h
er
e
ar
e
tech
n
iq
u
e
s
u
s
ed
to
tu
n
e
p
ar
a
m
eter
s
s
u
ch
as
F
u
zz
y
L
o
g
ic
co
n
tr
o
ller
f
o
r
A
n
t
C
o
lo
n
y
S
y
s
t
e
m
(
AC
S)
ep
s
ilo
n
p
ar
am
eter
[
1
2
]
.
A
ls
o
,
[
1
3
]
an
d
[
1
4
]
u
s
ed
Hid
d
en
Ma
r
k
o
v
Mo
d
el
(
HM
M)
al
g
o
r
it
h
m
t
o
tu
n
e
t
h
e
P
ar
ticle
S
w
ar
m
o
p
ti
m
izat
io
n
p
o
p
u
latio
n
s
ize
a
n
d
ac
ce
ler
atio
n
f
ac
to
r
s
p
ar
a
m
eter
s
.
B
es
id
es,
a
u
th
o
r
s
in
[
1
5
]
u
s
ed
HM
M
to
tu
n
e
th
e
in
er
tia
w
eig
h
t
p
a
r
a
m
eter
o
f
t
h
e
P
ar
ticle
S
w
ar
m
Op
ti
m
izatio
n
al
g
o
r
ith
m
.
Mo
r
eo
v
er
,
[
1
6
]
u
s
ed
Fu
zz
y
co
n
tr
o
ller
to
co
n
tr
o
l
Si
m
u
lated
An
n
ea
l
in
g
co
o
lin
g
la
w
,
[
1
7
]
an
d
[
1
8
]
u
s
ed
HM
M
to
tu
n
e
AC
S
ev
ap
o
r
atio
n
p
ar
a
m
eter
an
d
lo
c
al
p
h
er
o
m
o
n
e
d
ec
a
y
p
ar
a
m
ete
r
r
esp
ec
tiv
el
y
,
[
1
9
]
an
d
[
2
0
]
u
s
ed
HM
M
to
ad
ap
t
th
e
s
i
m
u
lated
an
n
ea
li
n
g
co
o
li
n
g
la
w
.
Fu
r
t
h
er
m
o
r
e,
[
1
4
]
u
s
ed
SVM
al
g
o
r
ith
m
to
p
r
ed
ict
th
e
p
er
f
o
r
m
an
ce
o
f
o
p
tim
izatio
n
p
r
o
b
le
m
s
.
Fi
n
all
y
,
au
th
o
r
s
i
n
[
2
0
]
u
s
ed
th
e
E
x
p
ec
tatio
n
-
Ma
x
i
m
izatio
n
alg
o
r
ith
m
to
lear
n
th
e
HM
M
al
g
o
r
ith
m
p
ar
a
m
eter
s
.
Fin
all
y
,
th
i
s
p
ap
er
is
t
h
e
co
n
ti
n
u
i
t
y
o
f
o
u
r
p
r
ev
io
u
s
p
ap
er
s
w
h
ic
h
d
ea
ls
w
it
h
th
e
lear
n
in
g
o
f
b
r
an
c
h
-
a
n
d
-
b
o
u
n
d
alg
o
r
ith
m
s
tr
ate
g
ies,
n
a
m
el
y
v
ar
iab
le
b
r
an
ch
in
g
s
tr
ate
g
y
an
d
n
o
d
e
s
elec
tio
n
s
tr
ateg
y
[
2
1
]
,
[
2
2
]
.
T
h
e
lear
n
in
g
al
g
o
r
ith
m
u
s
ed
w
as S
u
p
p
o
r
t V
ec
to
r
Ma
ch
in
e
(
S
V
M
).
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
r
e
ca
lls
s
o
m
e
b
asics
o
n
b
r
an
ch
-
a
n
d
-
b
o
u
n
d
alg
o
r
ith
m
a
n
d
S
VM
al
g
o
r
ith
m
w
i
th
p
ar
a
m
eter
t
u
n
in
g
.
I
n
s
ec
tio
n
3
,
w
e
p
r
ese
n
t
o
u
r
m
eth
o
d
o
lo
g
y
o
f
i
n
f
er
r
in
g
ef
f
icien
t
b
r
an
c
h
a
n
d
b
o
u
n
d
s
t
r
ateg
ies
a
n
d
e
x
p
er
i
m
e
n
tatio
n
co
n
f
i
g
u
r
atio
n
.
Sec
tio
n
s
4
is
d
ed
icate
d
to
r
esu
lt
s
.
Fin
all
y
,
w
e
co
n
cl
u
d
e
an
d
p
r
o
p
o
s
e
s
o
m
e
f
u
t
u
r
e
w
o
r
k
.
2.
B
RANC
H
-
AND
-
B
O
UND
AND
SVM
I
n
t
h
is
s
ec
tio
n
,
w
e
ar
e
f
ir
s
t
g
o
in
g
to
s
ee
a
n
o
v
er
v
ie
w
o
f
a
f
o
r
m
al
d
e
s
cr
ip
tio
n
o
f
b
r
an
c
h
-
an
d
-
b
o
u
n
d
s
tr
ateg
ie
s
an
d
p
r
esen
t
t
h
e
f
ea
t
u
r
es
u
s
ed
in
th
e
al
g
o
r
ith
m
.
Se
co
n
d
l
y
,
w
e
w
ill
in
v
e
s
ti
g
ate
S
VM
m
o
s
t
i
m
p
o
r
ta
n
t
ad
v
an
ta
g
es
w
i
th
a
r
e
m
a
in
d
er
o
f
lear
n
i
n
g
th
eo
r
y
.
2
.
1
.
B
ra
nch
-
a
nd
-
bo
un
d a
lg
o
rit
hm
B
r
an
ch
-
a
n
d
-
b
o
u
n
d
al
g
o
r
ith
m
is
o
u
tli
n
ed
i
n
th
is
s
ec
tio
n
.
W
e
f
ir
s
t
d
ef
i
n
e
u
s
e
f
u
l
n
o
tati
o
n
an
d
t
h
en
p
r
o
ce
ed
w
it
h
th
e
e
x
p
lan
a
tio
n
o
f
th
e
al
g
o
r
ith
m
s
tep
s
.
L
et
u
s
d
ef
in
e
a
g
en
er
al
MI
L
P
p
r
o
b
le
m
P
as f
o
llo
w
s
:
=
m
i
n
{
|
=
,
≥
0
,
N
on
-
ne
g
ati
v
e
v
ec
to
r
o
f
d
im
e
n
s
io
n
co
n
tain
i
n
g
at
lea
s
t o
n
e
i
n
teg
er
}
w
h
er
e
,
i
s
∗
d
i
m
en
s
io
n
m
atr
i
x
.
W
e
w
ill
u
s
e
also
d
ef
in
e
:
is
a
r
elax
ed
v
er
s
io
n
of
:
w
h
ic
h
is
=
{
|
=
,
≥
0
}
is
th
e
p
r
o
b
le
m
in
t
h
e
ℎ
iter
atio
n
w
h
ich
co
r
r
esp
o
n
d
s
to
a
n
o
d
e
in
b
r
an
ch
-
a
n
d
-
b
o
u
n
d
tr
ee
.
,
i
s
a
r
elax
ed
v
er
s
io
n
o
f
.
is
th
e
o
b
j
ec
tiv
e
v
al
u
e
o
f
ℎ
n
o
d
e.
(
∗
)
is
th
e
i
n
c
u
m
b
e
n
t p
o
in
t a
t iter
at
io
n
,
w
h
ic
h
m
ea
n
s
t
h
e
v
ec
to
r
th
at
lead
s
to
th
e
b
est
s
o
f
ar
.
∗
is
th
e
o
b
j
ec
tiv
e
f
u
n
ct
io
n
v
a
lu
e
o
n
(
∗
)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
d
a
p
ted
b
r
a
n
ch
-
a
n
d
-
b
o
u
n
d
a
l
g
o
r
ith
m
u
s
in
g
S
V
M w
ith
mo
d
e
l selec
tio
n
(
K
a
b
b
a
j Mo
h
a
med
Mu
s
ta
p
h
a
)
2483
B
r
ief
l
y
,
t
h
e
B
r
an
ch
-
a
n
d
-
b
o
u
n
d
alg
o
r
ith
m
,
in
th
e
ca
s
e
o
f
m
i
n
i
m
iz
atio
n
,
i
s
d
escr
ib
ed
as
f
o
llo
w
s
a
s
s
h
o
w
n
in
Alg
o
r
it
h
m
1
.
I
t
is
an
iter
ati
v
e
alg
o
r
it
h
m
,
a
n
d
in
ea
ch
iter
atio
n
,
w
e
h
a
v
e
at
least
th
r
ee
s
tep
s
w
h
ic
h
ar
e
:
Firstl
y
,
t
h
e
n
o
d
e
s
elec
tio
n
s
t
ep
ai
m
s
to
r
etr
iev
e
a
n
o
d
e
f
r
o
m
a
n
o
d
e
li
s
t
t
h
at
m
a
x
i
m
i
ze
s
s
o
m
e
cr
iter
io
n
.
T
h
is
latter
is
s
p
ec
if
ic
to
th
e
n
o
d
e
s
elec
tio
n
s
tr
ateg
y
.
Seco
n
d
l
y
,
an
d
o
n
ce
w
e
h
a
v
e
p
ick
ed
a
n
o
d
e
,
w
e
s
o
lv
e
its
r
ela
x
atio
n
,
b
y
a
n
al
g
o
r
ith
m
f
r
o
m
t
h
e
li
n
ea
r
p
r
o
g
r
a
m
m
in
g
f
r
a
m
e
w
o
r
k
s
u
c
h
as
s
i
m
p
lex
o
r
in
ter
io
r
p
o
in
ts
.
Dep
en
d
i
n
g
o
n
th
e
r
es
u
lt
s
,
w
e
d
i
s
ti
n
g
u
i
s
h
th
r
ee
ca
s
es.
T
h
e
f
ir
s
t
o
n
e
is
w
h
e
n
t
h
e
p
r
o
b
le
m
,
is
i
n
f
ea
s
ib
le
o
r
th
e
r
esu
lti
n
g
o
b
j
ec
tiv
e
f
u
n
ctio
n
v
al
u
e
is
g
r
e
ater
th
a
n
∗
.
C
o
n
s
eq
u
en
tl
y
,
t
h
e
c
u
r
r
en
t
i
ter
atio
n
is
ter
m
i
n
tated
.
T
h
e
s
ec
o
n
d
ca
s
e
is
w
h
e
n
th
e
s
o
lu
tio
n
is
i
n
te
g
er
an
d
<
∗
.
I
n
th
is
m
o
m
e
n
t,
w
e
u
p
d
ate
th
e
in
cu
m
b
e
n
t p
o
in
t
a
n
d
its
o
b
j
ec
t
iv
e
v
a
lu
e
∗
,
th
en
w
e
m
o
v
e
to
t
h
e
n
ex
t i
ter
atio
n
.
I
n
t
h
e
t
h
ir
d
ca
s
e,
w
h
e
n
n
o
n
e
o
f
th
e
co
n
d
if
io
n
s
m
en
tio
n
ed
b
ef
o
r
e
h
ap
p
en
s
,
w
e
p
er
f
o
r
m
v
ar
i
ab
le
b
r
an
ch
in
g
.
I
n
th
i
s
f
i
n
al
s
t
ep
,
w
e
m
u
s
t
s
elec
t
a
v
ar
iab
le
f
r
o
m
a
s
e
t
o
f
n
o
n
-
i
n
t
eg
er
v
ar
iab
les
r
e
lati
v
el
y
to
s
o
m
e
d
ef
in
ed
cr
iter
io
n
.
An
d
t
h
i
s
cr
iter
io
n
is
d
ef
i
n
ed
b
y
t
h
e
v
ar
iab
le
b
r
an
c
h
i
n
g
s
tr
at
eg
y
.
A
l
g
o
r
ith
m
1
.
B
r
an
ch
-
an
d
-
b
o
u
n
d
Alg
o
r
it
h
m
2
.
2
.
Su
pp
o
rt
Vec
t
o
r
M
a
chine
SVM
is
in
to
p
ten
m
ac
h
in
e
le
ar
n
in
g
alg
o
r
i
th
m
s
[
9
]
,
it
is
u
s
ed
f
o
r
b
o
th
class
if
ica
ti
o
n
an
d
r
eg
r
ess
i
o
n
.
I
t a
im
s
to
f
in
d
th
e
h
y
p
er
p
l
an
e
w
ith
th
e
b
est m
ar
g
in
.
T
h
e
b
est
is
d
em
o
n
s
tr
ate
d
t
o
b
e
th
e
la
r
g
e
o
n
e
d
if
f
e
r
en
t
iat
in
g
b
etw
ee
n
th
e
h
y
p
er
p
l
an
e
an
d
n
e
ar
est
d
ata
p
o
in
ts
ca
l
le
d
s
u
p
p
o
r
t
v
ec
t
o
r
s
.
2
.
2
.
1
.
Ca
s
e
o
f
L
in
ea
r
H
y
po
t
h
esis
s
e
t
f
o
r
SV
M
:
I
n
th
e
c
ase
o
f
r
eg
r
ess
io
n
,
an
d
esp
ec
i
ally
o
n
e
v
a
r
ian
t
o
f
SV
M
ca
l
le
d
-
SVM,
w
e
w
ill
p
r
es
en
t
n
ex
tly
,
th
e
c
ase
o
f
lin
e
a
r
h
y
p
o
th
esis
s
et.
L
e
t’
s
h
av
e
in
h
av
e
in
th
e
i
n
p
u
t,
tr
ain
in
g
d
at
a,
n
am
ely
(
,
)
,
0
≤
≤
T
h
e
o
u
t
p
u
t
o
f
th
e
alg
o
r
i
th
m
is
a
lin
e
ar
f
u
n
cti
o
n
:
(
)
=
+
,
w
ith
a
c
o
e
f
f
icien
t
v
ec
to
r
,
x
th
e
u
n
k
n
o
w
n
v
ec
t
o
r
an
d
b
a
c
o
n
s
tan
t
.
T
h
e
d
is
tan
c
e
b
etw
ee
n
a
h
y
p
er
p
l
an
e
o
f
e
q
u
a
ti
o
n
+
=
0
an
d
th
e
s
u
p
p
o
r
t
v
e
ct
o
r
s
,
is
1
|
|
|
|
.
C
o
n
s
eq
u
en
tly
,
m
ax
i
m
izin
g
th
e
m
ar
g
in
is
e
q
u
iv
alen
t
t
o
th
e
n
e
x
t o
p
t
im
izati
o
n
p
r
o
b
l
em
:
(
)
:
{
m
in
1
2
.
.
|
−
(
+
)
|
≤
,
∀
W
ith
,
b
ein
g
th
e
er
r
o
r
t
o
l
er
an
c
e
b
etw
ee
n
an
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
.
4
,
A
u
g
u
s
t 2
0
1
9
:
2
4
8
1
-
2490
2484
T
h
e
las
t p
r
o
b
lem
m
ig
h
t b
e
in
f
e
asib
le
.
A
n
d
t
o
a
d
d
m
o
r
e
ch
an
c
e
to
b
e
f
ea
s
i
b
l
e
,
w
e
ad
d
s
l
ac
k
v
ar
i
ab
les
to
th
e
p
r
o
b
lem
,
in
th
e
f
o
l
lo
w
in
g
w
a
y
:
(
)
:
{
m
i
n
(
)
=
1
2
+
∑
(
+
∗
)
=
1
.
.
−
(
+
)
≤
+
,
∀
(
+
)
−
≤
+
∗
,
∀
n
,
∗
≥
0
,
∀
w
ith
,
∗
ar
e
th
e
s
la
ck
v
a
r
i
ab
les
,
an
d
is
th
e
c
o
s
t
p
a
r
am
eter
u
s
ed
t
o
p
en
ali
ze
d
at
a
p
o
in
ts
o
u
t
s
id
e
th
e
m
ar
g
in
.
B
y
u
s
in
g
lag
r
ag
i
an
f
u
n
ctio
n
,
an
d
q
u
a
d
r
ati
c
o
p
tim
izati
o
n
o
r
o
th
e
r
r
es
o
lu
ti
o
n
m
eth
o
d
s
,
o
n
e
ca
n
p
r
o
v
e
th
at
th
e
s
o
lu
t
io
n
is
w
ith
th
e
f
o
r
m
o
f
:
(
)
=
∑
(
∗
−
)
=
1
+
w
ith
0
≤
,
∗
≤
,
∀
1
≤
≤
.
2
.
2
.
2
.
Ca
s
e
o
f
n
o
n
lin
e
a
r
h
y
p
o
t
hes
i
s
s
e
t
I
n
th
e
s
itu
at
io
n
,
w
h
er
e
w
e
c
an
n
o
t
f
in
d
a
h
y
p
er
p
l
an
e
c
o
n
t
ain
in
g
all
t
r
ain
in
g
in
s
tan
c
es,
o
n
e
m
ig
h
t
tr
an
s
f
o
r
m
th
e
s
p
a
ce
o
f
th
e
t
r
a
in
in
g
d
a
ta
to
an
o
th
e
r
,
in
s
u
ch
w
a
y
ca
n
b
e
c
o
m
p
r
is
e
d
in
o
n
e
h
y
p
er
p
l
an
e
o
n
th
e
n
e
w
s
p
ac
e.
T
o
d
o
th
is
tr
an
s
f
o
r
m
ati
o
n
,
o
n
e
c
an
u
s
e
th
e
w
ell
-
k
n
o
w
n
k
er
n
el
m
eth
o
d
s
.
I
n
f
ac
t
,
th
er
e
ar
e
in
lite
r
a
tu
r
e
d
if
f
er
en
t
k
e
r
n
els
u
s
ed
f
o
r
SV
M
,
s
u
ch
as
R
B
F
a
n
d
p
o
ly
n
o
m
ial.
F
o
r
th
e
r
est
,
w
e
w
ill
p
r
es
en
t
th
e
d
is
t
an
ce
c
alcu
l
ati
o
n
m
eth
o
d
f
o
r
th
e
R
B
F
k
er
n
el.
I
n
s
te
ad
o
f
u
s
in
g
th
e
s
t
an
d
ar
d
L
2
−
|
|
.
|
|
,
w
e
u
s
ed
th
e
n
o
r
m
ass
o
cia
te
d
w
ith
R
B
F
k
er
n
el
th
at
is
d
es
cr
ib
ed
as
f
o
ll
o
w
s
:
(
,
)
=
e
x
p
(
−
γ
|
|
x
−
x
|
|
2
)
w
ith
γ
,
is
th
e
g
am
m
a
p
a
r
am
ete
r
.
I
ts
g
eo
m
etr
ic
al
in
t
er
p
r
e
tat
io
n
i
s
,
w
h
en
th
e
g
am
m
a
p
ar
am
ete
r
h
as la
r
g
e
r
v
alu
e
s
,
th
e
h
y
p
er
p
lan
e
ass
o
c
iat
e
d
w
it
h
th
e
s
o
lu
ti
o
n
w
ill
h
av
e
m
o
r
e
in
clin
a
ti
o
n
s
to
c
o
n
ta
in
s
,
as
f
ar
as
p
o
s
s
i
b
l
e,
all
tr
a
in
in
g
d
at
a.
T
h
e
f
o
r
m
o
f
th
e
r
esu
ltin
g
t
ar
g
et
f
u
n
ct
io
n
,
w
ill
b
e
as
f
o
ll
o
w
s
:
(
)
=
∑
(
∗
−
)
(
,
)
=
1
+
I
n
th
is
p
a
p
e
r
,
w
e
w
ill
u
s
e
-
SVM
r
eg
r
ess
i
o
n
a
lg
o
r
ith
m
w
ith
th
e
R
B
F
k
e
r
n
el
tw
ice
f
o
r
le
a
r
n
in
g
n
o
d
e
s
ele
cti
o
n
s
tr
a
teg
y
an
d
v
ar
ia
b
l
e
b
r
an
ch
in
g
s
tr
ateg
y
r
esp
ec
t
iv
ely
.
2
.
3
.
L
ea
rning
o
f
v
a
ria
ble bra
nchin
g
s
t
ra
t
eg
y
a
nd
no
de
s
elec
t
i
o
n str
a
t
eg
y
C
o
n
ce
r
n
in
g
th
e
v
a
r
ia
b
l
e
b
r
a
n
ch
in
g
s
tr
a
teg
y
,
w
e
aim
in
th
is
p
a
p
er
to
im
itate
th
e
b
eh
av
io
r
o
f
th
e
r
el
ia
b
il
ity
b
r
an
ch
in
g
r
u
l
e.
T
h
i
s
r
u
le
is
b
as
ed
o
n
s
tr
o
n
g
b
r
a
n
ch
in
g
,
w
h
ich
is
tim
e
c
o
n
s
u
m
in
g
.
B
y
an
d
l
a
r
g
e,
r
el
ia
b
il
ity
b
r
an
ch
in
g
u
s
es
an
u
n
r
eli
a
b
ili
ty
q
u
ali
ty
f
o
r
v
ar
i
a
b
l
e
p
s
eu
d
o
-
c
o
s
ts
v
a
lu
es.
Fo
r
th
i
s
r
ea
s
o
n
,
r
eli
ab
ilit
y
d
e
p
en
d
s
o
n
n
u
m
er
o
u
s
p
r
o
b
lem
f
ea
tu
r
es
.
T
h
ese
f
ea
tu
r
es
ar
e
to
b
e
class
if
ie
d
in
n
o
d
e
-
b
as
ed
f
e
atu
r
es
an
d
v
ar
ia
b
le
-
b
as
ed
f
ea
tu
r
es.
2
.
3
.
1
.
No
de
-
ba
s
ed
f
ea
t
ures
W
e
u
s
e
in
t
h
is
ca
te
g
o
r
y
f
ea
t
u
r
es b
elo
w
:
-
R
ed
u
ce
d
Ob
j
ec
tiv
e
v
al
u
es
g
ai
n
:
Δ
,
=
|
−
−
1
|
|
−
1
|
(
No
te
th
at
th
e
f
ea
tu
r
e
s
s
h
o
u
ld
b
e
in
d
ep
en
d
en
t o
f
t
h
e
p
r
o
b
le
m
s
ca
le)
-
Dep
th
in
b
r
an
c
h
an
d
b
o
u
n
d
tr
ee
s
tar
tin
g
f
r
o
m
ze
r
o
.
-
No
d
e
esti
m
ate.
-
L
P
Ob
j
ec
tiv
e
Valu
e
2
.
3
.
2
.
Va
ria
ble
-
ba
s
ed
f
ea
t
ures
I
n
th
e
s
a
m
e
t
h
i
n
k
i
n
g
li
n
e,
w
e
u
s
e:
-
P
s
eu
d
o
-
co
s
t v
al
u
e
-
T
h
e
p
o
s
itiv
e
r
ed
u
ce
d
co
s
t a
n
d
th
e
n
e
g
ati
v
e
r
ed
u
ce
d
co
s
ts
i.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:
2
0
8
8
-
8708
A
d
a
p
ted
b
r
a
n
ch
-
a
n
d
-
b
o
u
n
d
a
l
g
o
r
ith
m
u
s
in
g
S
V
M w
ith
mo
d
e
l selec
tio
n
(
K
a
b
b
a
j Mo
h
a
med
Mu
s
ta
p
h
a
)
2485
ma
x
(
|
,
|
∑
|
,
|
:
,
≤
0
,
0
)
an
d
ma
x
(
|
,
|
∑
|
,
|
:
,
≥
0
,
0
)
w
it
h
,
i
s
th
e
ℎ
co
m
p
o
n
en
t
v
alu
e
o
f
co
s
t
v
ec
to
r
o
f
iter
atio
n
.
T
h
ese
f
ea
tu
r
es
ai
m
to
p
r
esen
t
eith
er
in
m
i
n
i
m
izatio
n
o
r
m
a
x
i
m
izatio
n
p
r
o
b
lem
s
h
o
w
w
e
ap
p
r
o
ac
h
t
o
th
e
o
p
ti
m
al
s
o
l
u
tio
n
.
T
h
e
o
th
er
s
p
ec
if
icit
y
in
o
u
r
wo
r
k
b
ey
o
n
d
ch
a
n
g
es
in
f
ea
tu
r
e
s
b
ased
o
n
th
o
s
e
p
r
esen
ted
in
[
1
1
]
,
is
w
e
ad
d
th
e
v
a
lu
e
o
f
lear
n
ed
f
u
n
ct
io
n
r
ep
r
esen
ti
n
g
n
o
d
e
s
elec
tio
n
i
n
t
h
e
s
et
o
f
f
ea
tu
r
e
s
.
T
h
is
l
ast
p
o
in
t
is
j
u
s
ti
f
ied
in
th
e
f
o
llo
w
in
g
s
u
b
-
s
ec
tio
n
.
F
o
r
lear
n
in
g
n
o
d
e
s
elec
tio
n
s
tr
ateg
y
,
w
e
w
ill
i
m
itate
n
o
d
e
esti
m
ate
s
tr
ate
g
y
.
T
h
is
s
tr
ateg
y
is
t
h
e
d
ef
a
u
lt o
n
e
u
s
ed
in
S
C
I
P
s
o
lv
er
.
2
.
4
.
I
nte
ra
ct
io
n o
f
no
de
s
elec
t
io
n str
a
t
eg
y
a
nd
v
a
ria
ble selec
t
io
n str
a
t
eg
y
I
n
tu
itiv
e
ly
,
th
e
ch
o
ice
o
f
a
n
o
d
e,
b
y
a
n
o
d
e
s
el
ec
t
io
n
s
tr
a
teg
y
,
in
f
lu
en
ce
s
th
e
ch
o
ic
e
th
e
n
ex
t
b
r
an
ch
in
g
v
a
r
ia
b
l
e.
Fo
r
th
is
r
e
aso
n
,
w
e
d
esc
r
i
b
e
f
o
r
m
ally
th
e
v
a
r
ia
b
l
e
b
r
an
ch
in
g
s
t
r
at
eg
y
f
u
n
ctio
n
VB
in
f
u
n
ctio
n
o
f
a
co
m
b
in
at
io
n
o
f
NS
(
N
o
d
e
s
el
ec
ti
o
n
s
t
r
at
eg
y
f
u
n
cti
o
n
)
an
d
o
th
e
r
f
ea
tu
r
es
d
esc
r
i
b
e
d
b
el
o
w
:
(
)
=
∗
(
,
,
)
+
∑
∗
w
h
er
e
an
d
a
r
e
r
ea
l n
u
m
b
er
s
.
No
t
e
th
at
w
e
d
o
u
b
le
u
s
e
NS
f
e
atu
r
es
ad
d
m
o
r
e
p
r
e
ci
s
i
o
n
.
2
.
5
.
Ov
er
f
it
t
ing
a
nd
pa
ra
m
et
er
t
un
ing
I
n
th
i
s
s
u
b
-
s
ec
tio
n
,
w
e
w
ill
d
ef
in
e
o
v
er
f
i
ttin
g
,
w
h
ic
h
is
a
v
er
y
co
m
m
o
n
p
r
o
b
le
m
i
n
lear
n
i
n
g
tech
n
iq
u
es t
h
at
a
f
f
ec
ts
t
h
e
f
i
n
a
l p
er
f
o
r
m
a
n
ce
.
2
.
5
.
1
.
O
v
er
f
it
t
ing
A
lear
n
i
n
g
m
o
d
el
is
,
b
y
d
ef
i
n
i
tio
n
,
a
co
u
p
le
o
f
a
lear
n
i
n
g
al
g
o
r
ith
m
a
n
d
a
h
y
p
o
th
es
is
s
et.
A
lear
n
i
n
g
alg
o
r
ith
m
i
s
a
n
iter
ati
v
e
al
g
o
r
ith
m
t
h
at
s
ea
r
ch
es
th
e
b
est
h
y
p
o
th
esis
f
it
tin
g
t
h
e
tr
ai
n
i
n
g
d
ata.
T
h
is
h
y
p
o
th
es
i
s
is
in
cl
u
d
ed
in
t
h
e
h
y
p
o
th
e
s
i
s
s
et
c
h
o
s
en
i
n
itia
ll
y
.
A
v
er
y
co
m
m
o
n
p
r
o
b
le
m
en
co
u
n
t
er
ed
in
lear
n
in
g
i
s
o
v
er
f
itti
n
g
.
T
h
is
p
h
en
o
m
en
o
n
o
cc
u
r
s
w
h
en
t
h
e
lear
n
ed
h
y
p
o
th
esis
d
o
es
n
o
t
g
e
n
er
alize
w
ell
to
all
p
o
s
s
ib
le
v
alu
e
s
b
e
y
o
n
d
t
h
e
tr
ain
i
n
g
d
ata.
C
au
s
e
s
ar
e
n
u
m
b
er
o
f
d
ata
p
o
in
ts
,
n
o
is
e
a
n
d
tar
g
et
co
m
p
l
ex
it
y
[
7
]
.
T
h
e
ch
o
ice
o
f
lear
n
i
n
g
a
lg
o
r
it
h
m
s
co
u
ld
a
f
f
ec
t
t
h
e
n
o
i
s
e
b
y
af
f
ec
tin
g
eit
h
er
b
ias
o
r
v
ar
ia
n
ce
.
I
n
t
h
e
ca
s
e
o
f
SVM,
t
h
e
th
o
r
o
u
g
h
ch
o
ice
o
f
SVM
p
ar
a
m
e
ter
s
is
r
eq
u
ir
ed
to
p
r
ev
en
t
f
r
o
m
o
v
e
r
f
itti
n
g
.
T
h
e
R
B
F
Ker
n
el
SVR
al
g
o
r
ith
m
u
s
ed
in
th
i
s
w
o
r
k
h
as
t
w
o
p
ar
am
et
er
s
,
co
s
t
an
d
g
a
m
m
a.
C
o
s
t
d
ef
i
n
es
h
o
w
m
u
c
h
is
p
en
alize
d
m
i
s
clas
s
i
f
ied
ex
a
m
p
les
an
d
g
a
m
m
a
d
e
f
in
e
s
h
o
w
f
ar
t
h
e
i
n
f
lu
e
n
ce
o
f
a
s
i
n
g
le
tr
ain
i
n
g
ex
a
m
p
le
r
ea
ch
es.
As
k
n
o
w
n
s
m
al
l
co
s
t
an
d
lar
g
e
g
a
m
m
a,
g
i
v
e
h
ig
h
e
r
b
ias
an
d
lo
w
er
v
ar
ia
n
ce
.
I
n
ad
d
itio
n
,
lar
g
e
co
s
t
an
d
s
m
all
g
a
m
m
a
g
iv
e
lo
w
e
r
b
ias
an
d
lar
g
er
v
ar
ia
n
ce
.
C
o
n
s
eq
u
en
tl
y
,
w
e
s
h
o
u
ld
t
u
n
e
co
s
t
an
d
g
a
m
m
a
p
ar
am
eter
s
u
n
til
w
e
f
i
n
d
tr
ad
eo
f
f
v
al
u
es
to
m
i
n
i
m
ize
th
e
g
en
er
aliza
tio
n
er
r
o
r
.
On
e
w
a
y
to
tu
n
e
g
a
m
m
a
an
d
co
s
t p
ar
am
eter
s
is
to
u
s
e
cr
o
s
s
v
alid
atio
n
.
2
.
5
.
2
.
Cro
s
s
v
a
lid
a
t
io
n w
it
h
m
o
del
s
elec
t
io
n
B
ef
o
r
e
d
ef
in
i
n
g
cr
o
s
s
v
alid
ati
o
n
,
let
u
s
f
i
n
d
o
u
t
w
h
at
is
v
ali
d
atio
n
.
T
o
d
o
s
o
,
w
e
d
ef
in
e
s
o
m
e
u
s
e
f
u
l
n
o
tatio
n
:
th
e
d
ata
s
et
th
e
tr
ain
i
n
g
s
et
th
e
v
alid
atio
n
s
et
T
h
e
g
o
al
o
f
v
alid
atio
n
is
to
g
iv
e
an
e
s
ti
m
atio
n
o
f
t
h
e
g
e
n
e
r
aliza
tio
n
er
r
o
r
.
First,
it
d
iv
id
es
o
f
d
ata
p
o
in
ts
,
to
o
f
s
ize
−
an
d
o
f
s
ize
,
th
en
lear
n
s
th
e
tar
g
et
f
u
n
ctio
n
b
ased
o
n
.
Fin
all
y
,
w
e
ca
lc
u
late
er
r
o
r
o
f
th
e
tar
g
e
t
f
u
n
ctio
n
i
n
.
T
h
is
latter
er
r
o
r
is
p
r
o
v
en
an
esti
m
atio
n
o
f
t
h
e
g
en
er
aliza
tio
n
er
r
o
r
.
T
h
e
F
ig
u
r
e
1
r
e
p
r
esen
ts
w
h
at
is
d
escr
ib
ed
ab
o
v
e
.
Fi
g
u
r
e
1
.
Valid
atio
n
m
et
h
o
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
.
4
,
A
u
g
u
s
t 2
0
1
9
:
2
4
8
1
-
2490
2486
T
h
e
er
r
o
r
is
a
g
o
o
d
esti
m
ate
o
f
g
en
er
aliza
tio
n
er
r
o
r
b
u
t
it
is
n
o
t
to
o
p
r
ec
is
e.
T
o
im
p
r
o
v
e
th
e
p
r
ec
is
io
n
,
o
th
er
tech
n
iq
u
e
s
r
ep
o
s
e
o
n
v
alid
atio
n
li
k
e
cr
o
s
s
v
al
id
atio
n
.
W
ith
o
u
t
th
e
l
o
s
s
o
f
g
e
n
er
alit
y
,
w
e
p
r
ese
n
t n
e
x
t 1
0
-
f
o
ld
cr
o
s
s
v
alid
atio
n
p
r
o
ce
s
s
.
L
et
’
s
p
ar
titi
o
n
to
1
,
2
to
10
.
W
e
u
s
e
v
alid
atio
n
p
r
o
ce
s
s
te
n
ti
m
es
f
o
r
,
=
w
h
er
e
1
≤
≤
10
an
d
,
=
\
.
I
n
th
e
o
u
tp
u
t,
w
e
h
a
v
e
10
er
r
o
r
s
1
to
10
.
T
h
en
w
e
ca
lc
u
lat
e
cr
o
s
s
v
alid
atio
n
er
r
o
r
d
en
o
ted
b
y
w
h
ic
h
i
s
t
h
e
m
ea
n
v
alid
atio
n
er
r
o
r
s
.
T
h
e
cr
o
s
s
v
a
lid
atio
n
er
r
o
r
is
m
o
r
e
p
r
ec
is
e
th
at
v
al
id
atio
n
er
r
o
r
.
W
e
r
esu
m
e
t
h
is
p
r
o
ce
s
s
i
n
th
e
F
ig
u
r
e
2.
Fig
u
r
e
2
.
C
r
o
s
s
v
alid
atio
n
f
o
r
a
s
p
ec
if
ic
lea
r
n
in
g
alg
o
r
it
h
m
No
w
t
h
at
w
e
h
a
v
e
p
r
ese
n
ted
cr
o
s
s
v
alid
atio
n
,
let
u
s
lo
o
k
f
o
r
w
ar
d
m
o
d
el
s
elec
tio
n
,
t
h
at
u
s
ed
i
n
t
h
i
s
p
ap
er
to
tu
n
e
p
ar
am
eter
s
o
f
g
a
m
m
a
an
d
co
s
t.
Fo
r
∗
d
if
f
er
e
n
t
co
m
b
i
n
atio
n
s
o
f
co
s
t
a
n
d
g
a
m
m
a,
let
’
s
n
o
te
a
co
u
p
le
(
,
)
w
ith
1
≤
≤
an
d
1
≤
≤
.
A
s
m
e
n
t
io
n
ed
in
t
h
e
F
ig
u
r
e
3
,
cr
o
s
s
v
alid
atio
n
is
ex
ec
u
ted
m
u
lt
ip
le
ti
m
es
w
it
h
d
if
f
er
en
t
p
ar
a
m
eter
co
n
f
ig
u
r
atio
n
.
As
r
esu
lt,
w
e
g
et
er
r
o
r
s
1
,
1
to
,
.
I
n
th
e
e
n
d
,
w
e
h
av
e
t
h
e
co
n
f
i
g
u
r
atio
n
t
h
at
h
a
v
e
t
h
e
lo
w
er
er
r
o
r
.
Fig
u
r
e
1
.
C
r
o
s
s
v
alid
atio
n
f
o
r
m
o
d
el
s
elec
tio
n
3.
RE
S
E
ARCH
M
E
T
H
O
D
I
n
th
is
s
ec
t
io
n
,
w
e
o
u
tlin
e
t
h
e
m
et
h
o
d
o
lo
g
y
,
s
tep
b
y
s
tep
,
o
f
lear
n
i
n
g
t
h
e
n
o
d
e
s
elec
tio
n
s
tr
ateg
y
NS
an
d
v
ar
iab
le
b
r
an
ch
i
n
g
s
tr
ate
g
y
VB
u
s
i
n
g
p
ar
a
m
ete
r
tu
n
i
n
g
.
T
h
en
,
w
e
p
r
esen
t th
e
e
x
p
er
i
m
en
t c
o
n
f
i
g
u
r
atio
n
.
3
.
1
.
Co
llect
ing
D
a
t
a
s
et
s
W
e
u
s
e
th
e
MI
P
L
I
B
2
0
1
0
lib
r
ar
y
a
s
i
n
s
ta
n
ce
s
to
w
h
ic
h
w
e
ap
p
ly
t
h
e
B
r
an
c
h
-
a
n
d
-
B
o
u
n
d
alg
o
r
ith
m
f
ea
t
u
r
ed
b
y
r
eliab
ilit
y
b
r
an
c
h
i
n
g
r
u
le
a
n
d
b
est e
s
ti
m
ate
s
elec
tio
n
r
u
le.
T
h
e
n
w
e
e
x
tr
a
ct
i
n
f
o
r
m
atio
n
o
f
f
ea
t
u
r
es
d
escr
ib
ed
b
ef
o
r
e.
No
te
th
at
th
e
b
est es
ti
m
ate
s
elec
tio
n
r
u
le
i
s
t
h
e
d
ef
a
u
lt
o
n
e
is
v
ar
io
u
s
o
p
t
i
m
izatio
n
to
o
ls
li
k
e
SC
I
P
.
Her
e
is
th
e
p
s
eu
d
o
-
co
d
e
o
f
th
e
d
ata
co
llectio
n
s
tep
as
s
h
o
w
n
i
n
A
l
g
o
r
ith
m
2
.
Fo
r
I
in
s
tan
c
e
in
M
I
P
L
I
B
2
0
1
0
So
lv
e
I
b
y
b
r
an
ch
-
an
d
-
b
o
u
n
d
r
u
le
d
b
y
r
elia
b
i
lity
b
r
an
ch
in
g
an
d
n
o
d
e
est
im
ate
s
ele
cti
o
n
r
u
le
an
d
co
l
le
ct
f
e
atu
r
e
v
alu
es
.
R
etu
r
n
th
e
lis
t
o
f
f
ea
tu
r
es
v
alu
es
A
l
g
o
r
ith
m
2
.
Data
C
o
llectio
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
d
a
p
ted
b
r
a
n
ch
-
a
n
d
-
b
o
u
n
d
a
l
g
o
r
ith
m
u
s
in
g
S
V
M w
ith
mo
d
e
l selec
tio
n
(
K
a
b
b
a
j Mo
h
a
med
Mu
s
ta
p
h
a
)
2487
3
.
2
.
L
ea
rning
NS a
nd
VB
First,
w
e
d
iv
id
e
co
lle
cted
d
ata
in
to
t
w
o
s
ets,
o
n
e
u
s
ed
f
o
r
tr
ain
i
n
g
a
n
d
v
alid
atio
n
an
d
th
e
o
th
er
f
o
r
test
.
B
y
ap
p
l
y
in
g
t
w
ice
cr
o
s
s
v
alid
atio
n
b
ased
m
o
d
el
s
elec
tio
n
d
escr
ip
ted
ab
o
v
e,
w
e
f
ir
s
tl
y
lear
n
t
h
e
s
co
r
e
f
u
n
ctio
n
o
f
e
v
er
y
n
o
d
e
in
t
h
e
n
o
d
es
q
u
eu
e.
So
th
e
n
o
d
e
h
a
v
in
g
t
h
e
b
e
s
t
s
co
r
e
(
)
w
i
ll
b
e
ch
o
s
en
in
b
r
an
ch
-
a
n
d
-
b
o
u
n
d
.
Seco
n
d
l
y
,
w
e
w
i
ll lea
r
n
t
h
e
s
co
r
e
f
u
n
ct
io
n
VB
o
f
v
ar
iab
le
b
r
an
ch
in
g
s
el
ec
tio
n
.
3
.
3
.
E
x
peri
m
e
nts
I
n
t
h
is
s
u
b
-
s
ec
tio
n
,
w
e
p
r
ese
n
t
t
h
e
lab
-
tes
t
u
s
ed
to
e
x
p
er
im
en
t
o
u
r
m
et
h
o
d
o
lo
g
y
an
d
w
e
g
i
v
e
t
h
e
p
s
eu
d
o
-
co
d
es u
s
ed
f
o
r
test
s
.
3
.
3
.
1
.
E
x
peri
m
e
nta
t
io
n c
o
nfig
ura
t
i
o
n
W
e
u
s
ed
SC
I
P
3
.
2
.
1
f
o
r
th
e
r
a
is
o
n
t
h
at
is
t
h
e
b
est
in
o
p
en
-
s
o
u
r
ce
an
d
f
r
ee
to
o
ls
[
2
3
]
.
Mo
r
eo
v
er
,
f
o
r
SVM
alg
o
r
it
h
m
a
n
d
m
o
d
el
s
elec
tio
n
,
w
e
u
s
ed
th
e
p
ac
k
a
g
e
e1
0
7
1
[
2
3
]
o
f
th
e
lan
g
u
ag
e
R
3
.
2
.
5
k
n
o
w
n
to
b
e
a
m
o
n
g
th
e
m
o
s
t p
er
f
o
r
m
an
t la
n
g
u
a
g
es i
n
i
m
p
le
m
e
n
tatio
n
o
f
SVM
alg
o
r
it
h
m
s
[
1
3
]
.
T
h
e
co
s
t
r
an
g
e
u
s
ed
is
{
10
−
4
,
10
−
3
,
…
,
10
5
}
an
d
t
h
e
g
a
m
m
a
r
an
g
e
is
{
2
−
8
,
2
−
7
,
…
,
2
1
}
.
T
h
ese
r
an
g
es
co
v
er
to
o
s
m
all
a
n
d
to
o
h
ig
h
v
alu
e
s
o
f
co
s
t
an
d
g
a
m
m
a.
T
h
e
OS
u
s
ed
is
Deb
ian
7
3
2
b
its
,
8
Go
R
A
M,
I
n
tel
2
.
4
0
GHz
P
r
o
ce
s
s
o
r
.
W
e
u
s
e
f
o
r
MI
L
P
i
n
s
ta
n
ce
s
t
h
e
b
e
n
ch
m
ar
k
s
e
t o
f
MI
P
L
I
B
2
0
1
0
[
2
8
]
to
co
llect
d
ata,
v
alid
it,
an
d
to
tes
t
r
es
u
lti
n
g
m
o
d
els.
F
o
r
tr
ain
in
g
a
n
d
v
a
lid
atio
n
s
et,
w
e
to
o
k
t
h
e
f
o
ll
o
w
i
n
g
i
n
s
ta
n
ce
s
as
s
h
o
w
n
in
F
ig
u
r
e
4
.
3
0
n
2
0
b
8
.
m
p
s
a
cc
-
tig
h
t5
.
mp
s
a
flo
w
4
0
b
.
mp
s
a
ir
0
4
.
mp
s
a
p
p
1
-
2
.
mp
s
a
s
h
6
0
8
g
p
ia
-
3
co
l.
mp
s
b
a
b
5
.
mp
s
b
ea
s
leyC3
.
mp
s
b
iella
1
.
mp
s
b
ien
s
t2
.
mp
s
b
in
ka
r
1
0
_
1
.
mp
s
b
ley
_
xl1
.
mp
s
b
n
a
tt3
5
0
.
mp
s
co
r
e2
5
3
6
-
6
9
1
.
mp
s
co
v
1
0
7
5
.
mp
s
csch
ed
0
1
0
.
m
p
s
d
a
n
o
in
t.m
p
s
d
f
n
-
g
w
in
-
UUM.
mp
s
eil3
3
-
2
.
m
p
s
eilB
1
0
1
.
mp
s
en
lig
h
t1
3
.
m
p
s
en
lig
h
t1
4
.
mp
s
ex
9
.
mp
s
g
la
s
s
4
.
mp
s
g
mu
-
35
-
4
0
.
mp
s
iis
-
bupa
-
co
v.
mp
s
i
is
-
p
ima
-
co
v.
mp
s
lectsch
ed
-
4
-
o
b
j.m
p
s
m1
0
0
n
5
0
0
k4
r
1
.
mp
s
ma
cro
p
h
a
g
e.
m
p
s
mcsch
ed
.
m
p
s
mik
-
250
-
1
-
1
0
0
-
1
.
mp
s
min
e
-
166
-
5
.
mp
s
min
e
-
90
-
1
0
.
mp
s
n
3
d
iv3
6
.
mp
s
n
4
-
3
.
mp
s
n
e
o
s
-
1
1
0
9
8
2
4
.
mp
s
n
e
o
s
-
1
3
3
7
3
0
7
.
m
p
s
n
eo
s
-
1
3
9
6
1
2
5
.
mp
s
n
eo
s
1
3
.
mp
s
n
eo
s
-
1
6
0
1
9
3
6
.
mp
s
n
eo
s
1
8
.
mp
s
n
eo
s
-
6
8
6
1
9
0
.
m
p
s
n
eo
s
-
8
4
9
7
0
2
.
mp
s
n
eo
s
-
9
1
6
7
9
2
.
mp
s
n
eo
s
-
9
3
4
2
7
8
.
mp
s
n
et1
2
.
m
p
s
Fig
u
r
e
4
.
T
r
ain
in
g
an
d
v
a
lid
ati
o
n
s
ets
T
h
ese
in
s
tan
ce
s
ab
o
u
t
ten
s
o
f
t
h
o
u
s
an
d
s
o
f
r
o
w
s
a
n
d
co
lu
m
n
s
.
T
h
e
to
tal
d
escr
ip
tio
n
i
s
av
a
ilab
le
in
[
28]
.
C
o
n
ce
r
n
i
n
g
t
h
e
v
al
id
atio
n
s
et,
i
t
co
n
tai
n
s
ap
p
r
o
x
im
ati
v
el
y
f
if
th
o
f
n
u
m
b
er
o
f
m
en
tio
n
ed
i
n
s
ta
n
ce
s
ab
o
v
e
[
7
]
.
Fin
all
y
,
th
e
n
o
d
e
li
m
it
is
f
i
x
ed
to
f
iv
e
h
u
n
d
r
ed
n
o
d
es
an
d
r
u
n
n
in
g
ti
m
e
li
m
it
is
f
i
x
ed
to
s
ix
-
h
u
n
d
r
ed
s
ec
o
n
d
s
.
3
.
3
.
2
.
P
s
eudo
-
co
des
I
n
t
h
e
s
o
lv
i
n
g
p
r
o
ce
s
s
,
th
e
al
g
o
r
ith
m
a
s
i
t
i
s
i
m
p
le
m
en
ted
in
S
C
I
P
is
e
x
ec
u
ti
n
g
n
u
m
er
o
u
s
e
v
e
n
t
co
d
es
r
elate
d
to
s
o
m
e
e
v
en
ts
.
Nex
t,
w
e
w
il
l
d
escr
ib
e
th
ese
e
v
en
t
s
,
an
d
g
i
v
e
th
e
p
s
e
u
d
o
-
co
d
e
r
elativ
e
to
ea
ch
o
n
e.
B
esid
es,
m
ai
n
ev
e
n
ts
m
o
d
if
ied
ar
e
r
esp
ec
tiv
el
y
:
3
.
3
.
3
.
No
de
s
elec
t
io
n e
v
ent
T
h
is
ev
e
n
t
o
cc
u
r
s
w
h
e
n
t
h
e
al
g
o
r
ith
m
i
s
i
n
t
h
e
p
h
ase
o
f
s
ele
ctin
g
t
h
e
n
ex
t
n
o
d
e
to
s
o
lv
e.
T
h
e
cr
iter
i
a
o
f
s
elec
tio
n
is
d
eter
m
i
n
ed
b
y
th
e
s
tr
ateg
y
i
m
p
le
m
e
n
ted
.
No
te
th
at
th
is
e
v
en
t
co
d
e
is
also
ex
ec
u
ted
ev
e
n
in
t
h
e
r
o
o
t
n
o
d
e
s
elec
tio
n
.
I
n
th
is
ev
en
t,
w
e
i
m
p
le
m
e
n
t
t
h
e
n
o
d
e
s
elec
tio
n
r
u
le
s
co
r
e
f
u
n
ct
io
n
NS
th
at
is
alr
ea
d
y
estab
lis
h
ed
.
Mo
r
eo
v
er
,
it
ca
lcu
lates
t
h
e
s
co
r
e
f
o
r
ea
ch
n
o
d
e
in
th
e
n
o
d
e
lis
t.
Fin
all
y
,
it
r
etu
r
n
s
th
e
n
o
d
e
w
it
h
th
e
m
a
x
i
m
u
m
s
co
r
e.
T
h
e
p
s
eu
d
o
-
co
d
e
is
th
e
Alg
o
r
it
h
m
3
.
Fo
r
e
ac
h
le
a
p
n
C
alcu
la
t
e
NS
(
n
)
R
etu
r
n
th
e
lea
p
w
ith
m
ax
i
m
u
m
NS
(
n
)
A
l
g
o
r
ith
m
3
.
No
d
e
s
elec
tio
n
e
v
en
t p
s
e
u
d
o
-
co
d
e
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
.
4
,
A
u
g
u
s
t 2
0
1
9
:
2
4
8
1
-
2490
2488
3
.
3
.
4
.
Va
ria
ble bra
nchin
g
ev
ent
T
h
is
ev
e
n
t
o
cc
u
r
s
w
h
e
n
t
h
e
a
lg
o
r
ith
m
i
s
i
n
t
h
e
p
h
a
s
e
o
f
s
elec
tin
g
a
b
r
an
c
h
i
n
g
v
ar
iab
le
o
f
a
n
o
d
e
alr
ea
d
y
s
o
lv
ed
a
n
d
h
ad
g
a
v
e
a
n
o
n
-
i
n
te
g
r
al
v
al
u
e.
I
n
t
h
is
ev
en
t,
w
e
i
m
p
le
m
e
n
t
th
e
v
ar
i
ab
le
s
elec
tio
n
r
u
le
s
co
r
e
f
u
n
ctio
n
VB
.
T
h
e
in
p
u
t
o
f
th
i
s
f
u
n
ctio
n
is
t
h
e
v
a
lu
e
s
o
f
th
e
n
o
d
e
-
b
ased
f
ea
tu
r
es
a
n
d
th
e
v
ar
iab
le
-
b
ased
f
ea
t
u
r
es.
T
h
e
n
o
d
e
-
b
ased
f
ea
t
u
r
es
ar
e
r
elate
d
to
a
f
ix
ed
n
o
d
e
an
d
th
e
y
ar
e
ca
lc
u
lated
f
o
r
t
h
e
cu
r
r
en
t
n
o
d
e
as
a
f
ir
s
t
s
tep
.
B
esid
es,
w
e
ca
lcu
l
ate,
th
e
v
al
u
e
o
f
N
S
o
n
t
h
e
c
u
r
r
en
t
n
o
d
e.
T
h
en
w
e
ca
lc
u
la
te
th
e
v
al
u
es
o
f
th
e
v
ar
iab
le
-
b
ased
f
ea
t
u
r
es
a
n
d
co
n
s
eq
u
e
n
tl
y
t
h
e
v
a
lu
e
o
f
VB
f
o
r
ea
ch
v
ar
iab
le.
I
n
t
h
e
o
u
tp
u
t
,
w
e
w
i
ll
r
et
u
r
n
t
h
e
v
ar
iab
le
w
i
th
t
h
e
m
a
x
i
m
u
m
s
co
r
e
VB.
Fin
all
y
,
w
e
b
r
an
c
h
o
n
th
e
s
elec
ted
v
ar
iab
le
a
n
d
cr
ea
ted
t
w
o
r
elate
d
ch
ild
r
en
n
o
d
es.
T
h
e
p
s
eu
d
o
-
co
d
e
is
as
Alg
o
r
it
h
m
4
.
C
alcu
la
te
th
e
v
alu
e
o
f
th
e
n
o
d
e
-
b
ase
d
f
ea
tu
r
es
C
al
cu
lat
e
th
e
v
alu
e
o
f
NS
r
el
ativ
e
t
o
th
e
p
r
es
en
t n
o
d
e.
F
o
r
e
ac
h
b
r
an
ch
in
g
v
ar
ia
b
le
ca
n
d
id
ate
C
alcu
la
te
v
alu
es
o
f
v
a
r
ia
b
l
e
-
b
a
s
ed
f
ea
tu
r
es
C
alcu
la
te
VB
in
t
er
m
s
o
f
ca
l
cu
late
d
f
ea
tu
r
es
.
R
etu
r
n
th
e
m
ax
o
f
VB
an
d
r
e
lativ
e
v
a
r
ia
b
le
C
r
ea
t
e
tw
o
ch
il
d
r
en
n
o
d
es
r
e
ly
in
g
u
p
o
n
th
e
ch
o
s
en
v
ar
ia
b
l
e
C
alcu
lat
e
p
o
s
s
ib
le
n
o
d
e
-
b
as
ed
f
ea
tu
r
es
v
al
u
es o
f
t
h
e
t
w
o
c
h
ild
r
en
A
l
g
o
r
ith
m
4
.
Var
iab
le
b
r
an
ch
i
n
g
e
v
en
t p
s
eu
d
o
-
co
d
e
3
.
3
.
5
.
No
de
So
lv
ed
ev
ent
T
h
is
ev
e
n
t
o
cc
u
r
s
w
h
e
n
t
h
e
a
lg
o
r
ith
m
i
s
t
h
e
s
tate
o
f
lea
v
i
n
g
t
h
e
n
o
d
e
alr
ea
d
y
s
o
l
v
ed
.
W
e
u
s
e
t
h
i
s
ev
en
t
to
ca
lc
u
late
th
e
v
alu
e
s
o
f
L
P
Ob
j
ec
tiv
e
Valu
e
o
f
th
e
cu
r
r
en
t
n
o
d
e,
an
d
th
e
r
ed
u
ce
d
o
b
j
ec
tiv
e
v
alu
es
g
ain
.
T
h
e
p
s
eu
d
o
-
co
d
e
is
t
h
e
A
l
g
o
r
ith
m
5
.
Get
th
e
L
P O
b
j
e
ctiv
e
V
alu
e
o
f
th
e
p
r
es
en
t n
o
d
e
.
C
alcu
la
te
th
e
r
e
d
u
ce
d
o
b
ject
iv
e
v
alu
es g
a
in
A
l
g
o
r
ith
m
5
.
No
d
e
s
o
lv
ed
ev
e
n
t p
s
eu
d
o
-
co
de
4.
RE
SU
L
T
S
W
e
p
r
esen
t
in
th
is
s
ec
tio
n
,
a
co
m
p
ar
is
o
n
b
et
w
ee
n
alg
o
r
it
h
m
s
r
esu
lted
f
r
o
m
o
u
r
ap
p
r
o
ac
h
es
an
d
s
tan
d
ar
d
b
r
an
c
h
-
a
n
d
-
b
o
u
n
d
al
g
o
r
ith
m
.
T
h
e
co
m
p
ar
i
s
o
n
is
d
o
n
e
i
n
ter
m
o
f
R
u
n
n
in
g
Time
,
Du
a
l
B
o
u
n
d
an
d
N
u
mb
er
o
f
S
o
lved
n
o
d
es.
T
h
e
d
u
al
b
o
u
n
d
b
ein
g
a
q
u
an
t
it
y
co
n
v
er
g
i
n
g
to
t
h
e
o
p
ti
m
al
s
o
lu
tio
n
if
i
t
ex
i
s
ts
.
T
h
e
g
r
ea
ter
v
alu
e
o
f
d
u
a
l b
o
u
n
d
is
th
e
b
est o
n
e.
T
o
g
et
to
t
h
is
co
m
p
ar
is
o
n
,
w
e
d
id
th
r
ee
d
if
f
er
en
t
s
o
l
v
i
n
g
co
n
f
ig
u
r
atio
n
s
o
n
te
s
t
s
et.
T
h
e
f
i
r
s
t
is
d
o
n
e
b
y
s
tan
d
ar
d
b
r
an
ch
-
a
n
d
-
b
o
u
n
d
(
SB
B
)
alg
o
r
ith
m
r
u
led
b
y
r
eliab
ilit
y
p
s
e
u
d
o
-
co
s
t
b
r
an
c
h
in
g
r
u
le
an
d
b
est
esti
m
ate
n
o
d
e
s
elec
tio
n
r
u
le.
T
h
en
f
o
r
th
e
s
ec
o
n
d
an
d
th
ir
d
,
w
e
u
s
ed
o
u
r
alg
o
r
ith
m
s
w
it
h
SVR
w
it
h
o
u
t
m
o
d
el
s
elec
tio
n
(
A
B
B
)
an
d
w
it
h
m
o
d
el
s
elec
tio
n
(
A
B
B
+M
S).
T
h
e
r
esu
lt
s
ar
e
d
etailed
in
th
e
T
ab
le
1
.
T
ab
le
1
.
R
esu
lts
o
f
ex
p
er
i
m
e
n
tatio
n
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
d
a
p
ted
b
r
a
n
ch
-
a
n
d
-
b
o
u
n
d
a
l
g
o
r
ith
m
u
s
in
g
S
V
M w
ith
mo
d
e
l selec
tio
n
(
K
a
b
b
a
j Mo
h
a
med
Mu
s
ta
p
h
a
)
2489
I
n
t
h
is
test
,
w
e
h
ad
ten
i
n
s
ta
n
ce
s
f
r
o
m
MI
P
L
I
B
2
0
1
0
.
T
h
ese
i
n
s
ta
n
ce
s
h
a
v
e
t
h
r
ee
d
i
f
f
er
en
t
t
y
p
es.
T
h
e
f
ir
s
t
o
n
e
is
m
ix
ed
in
teg
er
p
r
o
g
r
am
(
MI
P
)
th
at
r
eg
r
o
u
p
s
in
teg
er
an
d
co
n
ti
n
u
o
u
s
v
ar
ia
b
les.
T
h
e
s
ec
o
n
d
is
m
i
x
ed
b
in
ar
y
p
r
o
g
r
a
m
(
MB
P
)
,
w
h
ich
in
cl
u
d
es
b
o
t
h
co
n
ti
n
u
o
u
s
an
d
b
i
n
ar
y
v
ar
iab
les.
An
d
th
e
f
i
n
al
o
n
e
is
B
in
ar
y
p
r
o
g
r
a
m
(
B
P
)
th
at
co
n
tain
s
e
x
cl
u
s
i
v
el
y
b
i
n
ar
y
v
ar
iab
l
es.
T
h
ese
r
esu
lt
s
s
h
o
w
u
p
t
h
at
o
u
r
ap
p
r
o
ac
h
es
g
i
v
e
eq
u
i
v
alen
t
if
n
o
t
b
etter
d
u
a
l
b
o
u
n
d
co
m
p
ar
i
n
g
to
s
tan
d
ar
d
b
r
an
ch
-
a
n
d
-
b
o
u
n
d
i
n
ter
m
o
f
d
u
a
l
b
o
u
n
d
in
8
0
%
o
f
ca
s
e
s
e
x
ce
p
t
f
r
o
m
o
p
m
2
-
z7
-
s
2
an
d
r
a
n
1
6
x
1
6
in
s
ta
n
ce
s
.
An
o
t
h
er
i
m
p
o
r
tan
t
r
esu
lt,
i
s
th
a
t
o
u
r
last
ap
p
r
o
ac
h
g
i
v
es
eq
u
iv
ale
n
t
o
r
b
etter
r
u
n
n
i
n
g
t
i
m
e
co
m
p
ar
i
n
g
o
u
r
last
ap
p
r
o
ac
h
in
8
0
%
o
f
ca
s
es.
A
l
s
o
,
w
h
e
n
co
m
p
ar
i
n
g
it
to
t
h
e
s
ta
n
d
ar
d
b
r
an
ch
an
d
b
o
u
n
d
alg
o
r
ith
m
r
u
led
b
y
r
eliab
ilit
y
b
r
an
ch
i
n
g
a
n
d
n
o
d
e
b
est
es
i
m
ate
r
u
le,
o
u
r
ap
p
r
o
ac
h
g
i
v
es
b
etter
o
r
eq
u
iv
ale
n
t
r
esu
lt i
n
ab
o
u
t h
al
f
o
f
to
tal
i
n
s
tan
ce
s
.
W
e
n
o
ticed
th
at
th
er
e
i
s
a
n
e
m
p
ir
ical
r
elatio
n
b
et
w
ee
n
th
e
p
er
f
o
r
m
a
n
ce
o
f
d
u
al
b
o
u
n
d
a
n
d
th
e
n
u
m
b
er
o
f
co
n
s
tr
ain
ts
o
f
th
e
p
r
o
b
le
m
f
r
o
m
t
h
e
o
n
e
h
an
d
,
an
d
a
r
elatio
n
b
et
w
ee
n
th
e
p
er
f
o
r
m
a
n
ce
o
f
r
u
n
n
i
n
g
ti
m
e
an
d
th
e
n
u
m
b
er
o
f
v
ar
iab
les
f
r
o
m
t
h
e
o
th
er
h
a
n
d
.
T
o
c
o
n
cr
etize
th
ese
las
t p
o
in
ts
,
w
e
p
lo
t t
h
e
s
e
in
Fi
g
u
r
e
5
.
Fig
u
r
e
5
.
I
n
cr
ea
s
e
o
r
d
ec
r
ea
s
e
o
f
d
u
al
b
o
u
n
d
an
d
r
u
n
n
i
n
g
ti
m
e
r
esp
ec
tiv
el
y
T
h
e
lef
t
-
h
a
n
d
f
i
g
u
r
e
s
h
o
w
s
t
h
at
in
s
ta
n
ce
s
w
it
h
less
t
h
at
5
0
0
0
co
n
s
tr
ain
ts
g
av
e
b
etter
d
u
a
l
b
o
u
d
f
o
r
o
u
r
ap
p
r
o
ac
h
es c
o
m
p
ar
in
g
to
s
tan
d
ar
d
b
r
an
ch
an
d
b
o
u
n
d
.
As a
m
atter
o
f
f
ac
t,
t
h
e
o
p
m
2
-
z7
-
s
2
i
n
s
tan
ce
,
w
h
ic
h
is
r
ep
r
esen
ted
b
y
t
h
e
is
o
lated
p
o
in
t
in
th
e
d
o
w
n
-
r
ig
n
t
s
id
e
h
as
ap
p
r
o
x
im
a
tiv
e
l
y
3
1
0
0
0
v
ar
iab
les.
C
o
n
ce
r
n
i
n
g
th
e
r
ig
h
t
-
h
an
d
f
i
g
u
r
e,
it
s
h
o
ws
th
at
in
s
tan
ce
s
w
it
h
m
o
r
e
th
an
2
5
0
0
v
ar
iab
les,
in
cr
ea
s
ed
th
e
p
er
f
o
r
m
a
n
ce
o
f
r
u
n
n
i
n
g
ti
m
e,
w
h
en
r
eso
l
v
ed
b
y
o
u
r
ap
p
r
o
ac
h
es,
esp
ec
iall
y
f
o
r
n
s
1
2
0
8
4
0
0
,
n
s
1
6
8
8
3
4
7
an
d
r
ail5
0
7
in
s
tan
ce
s
.
5.
C
O
NCLU
SI
O
N
I
n
th
i
s
p
ap
er
,
w
e
ad
d
p
ar
a
m
et
er
tu
n
i
n
g
to
in
f
er
b
etter
co
n
f
i
g
u
r
atio
n
o
f
SVM.
Sa
y
in
g
t
h
i
s
,
w
e
u
s
ed
−
SVM
r
eg
r
ess
io
n
lear
n
in
g
alg
o
r
ith
m
k
n
o
w
n
f
o
r
h
is
h
i
g
h
a
cc
u
r
ac
y
to
lear
n
b
r
an
ch
-
a
n
d
-
b
o
u
n
d
alg
o
r
ith
m
n
o
d
e
s
elec
tio
n
a
n
d
v
ar
iab
le
s
b
r
an
ch
i
n
g
s
tr
ate
g
ies.
T
h
ese
c
h
o
ices
lead
to
b
etter
r
esu
lt
s
co
m
p
ar
in
g
to
r
eliab
ilit
y
p
s
eu
d
o
co
s
t
r
u
le
an
d
b
e
s
t
es
ti
m
ate
s
elec
tio
n
r
u
le,
w
h
ich
ar
e
k
n
o
w
n
t
o
b
e
f
r
o
m
th
e
b
est
i
n
liter
at
u
r
e.
I
n
p
er
s
p
ec
tiv
es,
w
e
w
il
l
w
o
r
k
o
n
eli
m
in
a
tin
g
n
o
i
s
e
in
d
ata,
co
m
p
ar
e
w
it
h
d
if
f
er
en
t
lear
n
in
g
alg
o
r
ith
m
s
a
v
ailab
l
e
in
liter
at
u
r
e.
RE
F
E
R
E
NC
E
S
[1
]
He
H
.
,
e
t
a
l
.
,
“
L
e
a
rn
in
g
to
se
a
r
c
h
in
b
ra
n
c
h
a
n
d
b
o
u
n
d
a
lg
o
rit
h
m
s
,
”
Ad
v
a
n
c
e
s
in
n
e
u
ra
l
in
f
o
rm
a
ti
o
n
p
ro
c
e
ss
in
g
sy
ste
ms
,
p
p
.
3
2
9
3
-
3
3
0
1
,
2
0
1
4
.
[2
]
M
o
ll
R
.
,
e
t
a
l
.
,
“
L
e
a
rn
in
g
in
sta
n
c
e
-
in
d
e
p
e
n
d
e
n
t
v
a
lu
e
f
u
n
c
ti
o
n
s
to
e
n
h
a
n
c
e
lo
c
a
l
se
a
rc
h
,
”
Ad
v
a
n
c
e
s
in
Ne
u
ra
l
In
fo
rm
a
t
io
n
Pro
c
e
ss
in
g
S
y
ste
ms
,
p
p
.
1
0
1
7
-
1
0
2
3
,
1
9
9
9
.
[3
]
Ka
rz
a
n
F
.
K
.
,
e
t
a
l
.
,
“
In
f
o
r
m
a
ti
o
n
-
b
a
se
d
b
ra
n
c
h
i
n
g
sc
h
e
m
e
s
f
o
r
b
in
a
ry
li
n
e
a
r
m
ix
e
d
in
teg
e
r
p
ro
b
lem
s
,
”
M
a
th
e
ma
ti
c
a
l
Pro
g
ra
mm
in
g
Co
m
p
u
t
a
ti
o
n
,
v
o
l/
issu
e
:
1
(4
)
,
p
p
.
2
4
9
-
93
,
2
0
0
9
.
[4
]
Da
v
e
y
B
.
,
e
t
a
l
.
,
“
Eff
icie
n
t
in
telli
g
e
n
t
b
a
c
k
trac
k
in
g
u
sin
g
li
n
e
a
r
p
r
o
g
ra
m
m
in
g
,
”
INFORM
S
J
o
u
rn
a
l
o
n
C
o
mp
u
ti
n
g
,
vo
l/
issu
e
:
1
4
(4
)
,
p
p
.
3
7
3
-
86
,
2
0
0
2
.
[5
]
Ch
e
n
D
.
S
.
,
e
t
a
l
.
,
“
A
p
p
li
e
d
i
n
teg
e
r
p
ro
g
ra
m
m
in
g
:
m
o
d
e
li
n
g
a
n
d
so
lu
ti
o
n
,
”
Jo
h
n
W
il
e
y
&
S
o
n
s
,
2
0
1
1
.
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
.
4
,
A
u
g
u
s
t 2
0
1
9
:
2
4
8
1
-
2490
2490
[6
]
A
c
h
terb
e
rg
T
.
,
e
t
a
l
.
,
“
Bra
n
c
h
in
g
ru
les
re
v
isit
e
d
,
”
Op
e
ra
ti
o
n
s R
e
se
a
rc
h
L
e
tt
e
rs
,
v
o
l/
iss
u
e
:
3
3
(1
)
,
p
p
.
42
-
54
,
2
0
0
5
.
[7
]
A
b
u
-
Mo
sta
f
a
Y
.
S
.
,
e
t
a
l
.
,
“
L
e
a
rn
in
g
f
ro
m
d
a
ta
,
”
Ne
w
Yo
rk
,
NY
,
USA
,
A
M
L
Bo
o
k
,
2
0
1
2
.
[8
]
C
.
Ru
d
in
,
“
P
re
d
icti
o
n
:
M
a
c
h
in
e
L
e
a
rn
in
g
a
n
d
S
tatisti
c
s
,”
2
0
1
2
.
[9
]
X
.
W
u
,
e
t
a
l
.
,
“
T
o
p
1
0
a
lg
o
rit
h
m
s
in
d
a
ta
m
in
in
g
,”
K
n
o
wled
g
e
a
n
d
in
fo
rm
a
t
io
n
sy
ste
ms
,
v
o
l
/
i
ssu
e
:
14
(
1
),
p
p
.
1
-
37
,
2
0
0
8
.
[1
0
]
B.
L
a
n
tz,
“
M
a
c
h
in
e
lea
rn
i
n
g
w
it
h
R
,”
P
a
c
k
t
P
u
b
l
ish
i
n
g
L
td
,
2
0
1
5
.
[1
1
]
A
l
v
a
re
z
A
.
M
.
,
e
t
a
l
.
,
“
A
m
a
c
h
in
e
lea
rn
in
g
-
b
a
se
d
a
p
p
r
o
x
im
a
ti
o
n
o
f
stro
n
g
b
ra
n
c
h
i
n
g
,
”
INFOR
M
S
J
o
u
rn
a
l
o
n
Co
mp
u
t
in
g
,
v
o
l/
issu
e
:
2
9
(1
)
,
p
p
.
185
-
95
,
2
0
1
7
.
[1
2
]
A
o
u
n
O
.
,
e
t
a
l
.
,
“
In
v
e
stig
a
ti
o
n
o
f
h
id
d
e
n
m
a
rk
o
v
m
o
d
e
l
f
o
r
th
e
tu
n
in
g
o
f
m
e
tah
e
u
risti
c
s
in
a
irl
in
e
sc
h
e
d
u
li
n
g
p
ro
b
lem
s
,
”
IFA
C
-
Pa
p
e
rs
On
L
in
e
,
v
o
l/
issu
e
:
4
9
(3
)
,
p
p
.
3
4
7
-
52
,
2
0
1
6
.
[1
3
]
E
.
Af
ia
A
.
,
e
t
a
l
.
,
“
Hid
d
e
n
m
a
r
k
o
v
m
o
d
e
l
c
o
n
tro
l
o
f
in
e
rti
a
we
ig
h
t
a
d
a
p
tati
o
n
f
o
r
P
a
rti
c
le
sw
a
r
m
o
p
ti
m
iza
ti
o
n
,
”
IF
AC
-
Pa
p
e
rs
On
L
in
e
,
v
o
l/
issu
e
:
5
0
(1
)
,
p
p
.
9
9
9
7
-
1
0
0
0
2
,
2
0
1
7
.
[1
4
]
E
.
Af
ia
A
.
a
n
d
S
a
rh
a
n
i
M
.
,
“
P
e
rf
o
r
m
a
n
c
e
p
re
d
ictio
n
u
sin
g
su
p
p
o
rt
v
e
c
to
r
m
a
c
h
in
e
f
o
r
th
e
c
o
n
f
ig
u
ra
t
io
n
o
f
o
p
ti
m
iza
ti
o
n
a
lg
o
rit
h
m
s
,
”
Clo
u
d
Co
mp
u
t
in
g
T
e
c
h
n
o
l
o
g
ies
a
n
d
A
p
p
li
c
a
ti
o
n
s
(
Clo
u
d
T
e
c
h
)
,
2
0
1
7
3
r
d
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
,
p
p
.
1
-
7
,
2
0
1
7
.
[1
5
]
E
.
Af
ia
A
.
,
e
t
a
l
.
,
“
F
u
z
z
y
lo
g
ic
c
o
n
tr
o
ll
e
r
f
o
r
a
n
a
d
a
p
t
iv
e
Hu
a
n
g
c
o
o
li
n
g
o
f
si
m
u
late
d
a
n
n
e
a
li
n
g
,
”
Pro
c
e
e
d
in
g
s
o
f
th
e
2
n
d
in
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
o
n
B
ig
D
a
ta
,
Clo
u
d
a
n
d
Ap
p
li
c
a
ti
o
n
s
,
p
p
.
6
4
,
2
0
1
7
.
[1
6
]
Bo
u
z
b
it
a
S
.
,
e
t
a
l
.
,
“
A
n
o
v
e
l
b
a
se
d
Hid
d
e
n
M
a
rk
o
v
M
o
d
e
l
a
p
p
r
o
a
c
h
f
o
r
c
o
n
tro
l
li
n
g
t
h
e
A
CS
-
TS
P
e
v
a
p
o
ra
ti
o
n
p
a
ra
m
e
ter
,
”
M
u
lt
ime
d
ia
Co
m
p
u
ti
n
g
a
n
d
S
y
ste
ms
(
ICM
CS
),
2
0
1
6
5
t
h
In
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
,
p
p
.
6
3
3
-
6
3
8
,
2
0
1
6
.
[1
7
]
L
a
lao
u
i
M
.
,
e
t
a
l
.
,
“
Hid
d
e
n
M
a
r
k
o
v
M
o
d
e
l
f
o
r
a
se
l
f
-
le
a
rn
in
g
o
f
S
im
u
late
d
A
n
n
e
a
li
n
g
c
o
o
li
n
g
law
,
”
M
u
lt
ime
d
ia
Co
mp
u
t
in
g
a
n
d
S
y
ste
ms
(
ICM
CS
)
,
2
0
1
6
5
t
h
I
n
ter
n
a
t
io
n
a
l
Co
n
fer
e
n
c
e
,
p
p
.
5
5
8
-
5
6
3
,
2
0
1
6
.
[1
8
]
L
a
lao
u
i
M
.
,
e
t
a
l
.
,
“
A
se
lf
-
a
d
a
p
ti
v
e
v
e
r
y
fa
st
si
m
u
late
d
a
n
n
e
a
li
n
g
b
a
se
d
o
n
Hid
d
e
n
M
a
rk
o
v
m
o
d
e
l
,
”
Clo
u
d
Co
mp
u
t
in
g
T
e
c
h
n
o
l
o
g
ies
a
n
d
A
p
p
li
c
a
ti
o
n
s (
Clo
u
d
T
e
c
h
),
2
0
1
7
3
rd
I
n
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
,
p
p
.
1
-
8
,
2
0
1
7
.
[1
9
]
E
.
Af
ia
A
.
,
e
t
a
l
.
,
“
T
h
e
E
ff
e
c
t
o
f
Up
d
a
ti
n
g
th
e
L
o
c
a
l
P
h
e
ro
m
o
n
e
o
n
A
CS
P
e
rf
o
rm
a
n
c
e
u
sin
g
F
u
z
z
y
L
o
g
i
c
,
”
In
ter
n
a
t
io
n
a
l
J
o
u
rn
a
l
o
f
E
lec
trica
l
a
n
d
C
o
mp
u
ter
En
g
in
e
e
rin
g
(
IJ
ECE
)
,
v
o
l/
issu
e
:
7
(4
)
,
p
p
.
2
1
6
1
-
8
,
2
0
1
7
.
[2
0
]
Bo
u
z
b
it
a
S
.
,
e
t
a
l
.
,
“
Dy
n
a
m
i
c
a
d
a
p
tatio
n
o
f
th
e
A
CS
-
T
S
P
lo
c
a
l
p
h
e
ro
m
o
n
e
d
e
c
a
y
p
a
ra
m
e
ter
b
a
s
e
d
o
n
t
h
e
Hid
d
e
n
M
a
rk
o
v
M
o
d
e
l
,
”
Clo
u
d
Co
m
p
u
ti
n
g
T
e
c
h
n
o
lo
g
ies
a
n
d
A
p
p
li
c
a
ti
o
n
s
(
Clo
u
d
T
e
c
h
),
2
0
1
6
2
n
d
In
ter
n
a
ti
o
n
a
l
Co
n
fer
e
n
c
e
,
pp
.
3
4
4
-
3
4
9
,
2
0
1
6
.
[2
1
]
Ka
b
b
a
j
M
.
M
.
a
n
d
E
.
A
f
ia
A
.
,
“
T
o
wa
rd
s
lea
rn
in
g
in
teg
ra
l
stra
teg
y
o
f
b
ra
n
c
h
a
n
d
b
o
u
n
d
,
”
M
u
lt
im
e
d
ia
Co
m
p
u
t
in
g
a
n
d
S
y
ste
ms
(
ICM
CS
),
2
0
1
6
5
th
I
n
ter
n
a
ti
o
n
a
l
C
o
n
fer
e
n
c
e
,
p
p
.
6
2
1
-
626
,
2
0
1
6
.
[2
2
]
E
.
Af
ia
A
.
a
n
d
Ka
b
b
a
j
M
.
M.
,
“
S
u
p
e
rv
ise
d
lea
rn
in
g
in
Br
a
n
c
h
-
a
n
d
-
c
u
t
stra
teg
ies
,
”
Pro
c
e
e
d
in
g
s
o
f
th
e
2
n
d
in
ter
n
a
t
io
n
a
l
Co
n
fer
e
n
c
e
o
n
B
ig
Da
ta
,
C
lo
u
d
a
n
d
A
p
p
l
ica
ti
o
n
s
,
p
p
.
1
1
4
,
2
0
1
7
.
[2
3
]
h
tt
p
:
//
S
c
ip
.
z
ib
.
d
e
.
[2
4
]
h
tt
p
:
//
M
i
p
li
b
.
z
ib
.
d
e
.
[2
5
]
Da
v
id
M
.
,
“
S
u
p
p
o
rt
V
e
c
to
r
M
a
c
h
in
e
s:
T
h
e
In
ter
f
a
c
e
to
li
b
sv
m
in
P
a
c
k
a
g
e
e
1
0
7
1
,
”
Da
v
id
.
M
e
y
e
r
@
R
-
Pro
jec
t.
o
rg
.
2
0
1
7
.
B
I
O
G
RAP
H
I
E
S
O
F
AUTH
O
RS
M
.
K
a
b
b
a
j
M
o
h
a
m
e
d
M
u
sta
p
h
a
is
a
P
HD
stu
d
e
n
t
in
a
t
Na
ti
o
n
a
l
S
c
h
o
o
l
o
f
Co
m
p
u
ter S
c
ien
c
e
a
n
d
S
y
st
e
m
s
A
n
a
l
y
sis
(EN
S
I
A
S
),
Ra
b
a
t,
M
o
ro
c
c
o
.
He
o
b
tain
e
d
M
.
E
n
g
.
in
2
0
1
3
i
n
Co
m
p
u
ter
S
c
ien
c
e
f
ro
m
N
a
ti
o
n
a
l
S
c
h
o
o
l
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
S
y
ste
m
s
A
n
a
l
y
sis.
Re
se
a
r
c
h
a
re
a
s
o
f
in
tere
st
a
r
e
M
a
c
h
in
e
L
e
a
rn
in
g
,
Co
m
b
in
a
to
rial
Op
ti
m
iza
ti
o
n
,
S
t
o
c
h
a
stic
P
r
o
g
ra
m
m
in
g
a
n
d
F
e
a
tu
re
S
e
lec
ti
o
n
.
Abd
e
ll
a
tif
El
Afia
is
a
n
A
ss
o
c
iat
e
P
r
o
f
e
ss
o
r
a
t
Na
ti
o
n
a
l
S
c
h
o
o
l
o
f
Co
m
p
u
ter
S
c
ien
c
e
a
n
d
S
y
ste
m
s
A
n
a
l
y
si
s
(ENS
IA
S
),
Ra
b
a
t,
M
o
r
o
c
c
o
.
He
re
c
e
iv
e
d
h
is
M
.
S
c
.
d
e
g
re
e
s
in
A
p
p
li
e
d
M
a
th
e
m
a
ti
c
s
f
ro
m
Un
iv
e
rsit
y
o
f
S
h
e
rb
ro
o
k
.
He
o
b
ta
in
e
d
h
is
P
h
.
D.
in
1
9
9
9
in
Op
e
ra
ti
o
n
Re
se
a
rc
h
f
ro
m
Un
iv
e
rsit
y
o
f
S
h
e
rb
r
o
o
k
,
Ca
n
a
d
a
.
Re
se
a
rc
h
a
r
e
a
s
o
f
in
tere
st
a
re
M
a
th
e
m
a
ti
c
a
l
P
r
o
g
ra
m
m
in
g
(S
to
c
h
a
stic
a
n
d
d
e
term
in
isti
c
),
M
e
tah
e
u
risti
c
s,
Re
c
o
m
m
e
n
d
a
ti
o
n
S
y
ste
m
s a
n
d
M
a
c
h
in
e
L
e
a
rn
in
g
.
Evaluation Warning : The document was created with Spire.PDF for Python.