T
E
L
KO
M
NI
K
A
,
V
o
l.
1
4
,
N
o.
3
,
S
ept
em
ber
20
1
6
,
pp.
88
7
~
8
9
3
I
S
S
N
:
1
693
-
6
930
,
ac
c
r
edi
t
ed
A
b
y
D
IK
T
I,
D
e
c
r
e
e
N
o
:
58/
D
I
K
T
I
/
K
ep/
2013
D
O
I
:
10.
12928/
T
E
LK
O
M
N
I
K
A
.
v
1
4
i
3
.
3615
88
7
R
ec
ei
v
ed
M
ar
c
h
2
3
,
201
6
;
R
ev
i
s
ed
May
2
9
,
2
01
6
;
A
c
c
ept
e
d
J
une
1
0
,
20
1
6
A
d
a
p
t
i
v
e R
eso
u
r
ce
A
l
lo
ca
t
io
n
A
lg
o
r
it
h
m
i
n
W
ir
e
less
A
c
c
ess N
et
w
o
r
k
Z
h
a
n
j
u
n
L
i
u
,
Y
u
e S
h
en
*
,
Z
h
o
n
g
h
u
a
Y
u
,
F
e
n
g
x
i
e
Q
i
n
,
Q
i
a
n
b
i
n
C
h
e
n
C
hongq
i
ng K
ey
Labor
at
or
y
of
M
obi
l
e C
om
m
uni
c
at
i
on
T
ec
hn
ol
ogy
,
C
hongq
i
ng U
n
i
v
er
s
i
t
y
of
P
o
s
t
s
and T
el
e
c
o
m
m
u
ni
c
at
i
ons
(
C
Q
U
P
T
)
,
C
hongqi
ng
,
40
006
5,
C
h
i
na
*
C
or
r
es
po
ndi
ng a
ut
hor
,
e
-
m
a
i
l
:
ki
mj
j
p
cy
@
163
.c
o
m
A
b
st
r
act
W
i
r
el
e
s
s
net
w
or
k
s
t
at
e
v
ar
i
es
w
i
t
h
t
he
s
ur
r
oun
di
ng
env
i
r
on
m
ent
,
how
ev
er
,
t
he
e
x
i
s
t
i
n
g
r
es
our
c
e
al
l
o
c
at
i
on a
l
gor
i
t
hm
c
an
not
a
dapt
t
o t
he v
ar
y
i
ng
n
e
tw
o
r
k
s
ta
te
,
w
hi
c
h r
es
ul
t
s
t
o t
he
u
nder
ut
i
l
i
z
at
i
on
o
f
f
r
eque
nc
y
and pow
er
r
es
o
ur
c
e.
T
her
ef
or
e
,
i
n t
hi
s
pa
per
,
w
e pr
opos
e a
n ada
pt
i
v
e r
e
s
our
c
e al
l
oc
at
i
o
n
al
gor
i
t
hm
w
hi
c
h
c
an
ef
f
i
c
i
ent
l
y
adapt
t
o
t
he
v
ar
y
i
n
g
net
w
o
r
k
s
t
at
e
by
bui
l
di
ng
an
op
t
i
m
al
m
at
hem
at
i
c
a
l
m
odel
and t
hen c
han
gi
ng t
he
w
ei
ght
e
d v
a
l
ue of
t
he ob
j
ec
t
i
v
e f
unc
t
i
on
.
F
ur
t
h
er
m
or
e,
t
he o
pt
i
m
al
al
l
oc
at
i
o
n
of
s
ub
c
ar
r
i
er
a
nd pow
er
i
s
de
r
i
v
ed by
u
s
i
n
g t
he Lagr
a
nge du
al
dec
om
p
os
i
t
i
on
a
nd t
he s
u
bg
r
adi
en
t
m
et
hod
.
S
i
m
ul
at
i
on r
e
s
u
l
t
s
s
h
ow
t
h
at
t
he pr
opo
s
ed
al
gor
i
t
hm
c
a
n
adapt
i
v
e
l
y
al
l
oc
at
e
t
he r
es
our
c
e t
o t
h
e u
s
er
s
ac
c
or
di
n
g t
o
t
he
v
ar
y
i
n
g u
s
er
dens
i
t
y
w
hi
c
h r
ep
r
es
ent
s
t
he
n
et
w
or
k
s
t
at
e
.
K
e
y
w
o
r
d
s
:
w
i
r
e
l
es
s
a
c
c
es
s
n
et
w
or
k
,
net
w
or
k
e
nv
i
r
onm
ent
,
adapt
i
v
e
,
r
e
s
our
c
e a
l
l
o
c
at
i
on
,
l
agr
ang
e du
al
C
o
p
y
r
i
g
h
t
©
20
16 U
n
i
ver
si
t
a
s A
h
mad
D
ah
l
an
.
A
l
l
r
i
g
h
t
s r
eser
ved
.
1.
I
n
tr
o
d
u
c
ti
o
n
W
i
t
h t
he
de
v
e
l
opm
ent
of
w
i
r
el
es
s
m
obi
l
e c
om
m
uni
c
at
i
o
n s
uc
h
as
4G
an
d 5
G
,
t
he
num
ber
of
m
obi
l
e t
er
m
i
nal
i
s
i
nc
r
eas
i
ng gr
o
w
i
ng,
a
nd
t
h
e f
l
uc
t
ua
t
i
o
n of
us
er
dens
i
t
y
al
s
o
bec
om
es
hi
gher
,
e.
g
.
,
t
he
us
er
de
ns
i
t
y
of
c
om
m
er
c
i
al
ar
e
as
i
n
day
t
im
e
is
hi
gh
er
t
ha
n t
h
at
i
n
ev
e
ni
ng,
an
d t
he us
er
de
n
s
i
t
y
of
r
es
i
den
t
i
a
l
ar
eas
i
n
da
y
t
im
e
is
l
ow
er
t
han t
hat
i
n
e
v
e
ni
ng.
C
ons
eq
uen
t
l
y
,
t
he
c
h
ang
i
n
g
us
er
dens
i
t
y
w
i
l
l
af
f
ec
t
t
h
e
w
i
r
el
es
s
net
w
or
k
s
t
at
e
.
I
n
add
i
t
i
on,
t
he
i
nh
er
ent
c
har
ac
t
er
i
s
t
i
c
of
w
i
r
el
es
s
net
w
or
k
i
s
t
hat
t
he u
s
er
s
ar
e al
w
a
y
s
i
n m
ov
em
e
nt
.
T
her
ef
or
e,
an ef
f
ec
t
i
v
e
ada
pt
i
v
e
r
es
ou
r
c
e al
l
oc
at
i
on
i
s
nec
es
s
ar
y
f
or
t
he v
ar
y
i
ng
net
w
or
k
[1
].
E
x
i
s
t
i
n
g r
es
o
ur
c
e a
l
l
oc
at
i
on
al
gor
i
t
hm
s
c
an be
di
v
i
ded
i
nt
o
t
w
o c
at
egor
i
es
b
as
ed
on t
he
opt
i
m
i
z
at
i
o
n c
r
i
t
er
i
a
.
O
ne i
s
t
o
t
ak
e
t
he us
er
s
’
t
hr
oug
hp
ut
m
ax
i
m
i
z
at
i
o
n as
t
he o
bj
ec
t
i
v
e f
unc
t
i
on
[2
,
3]
,
and t
he ot
her
i
s
t
o t
ak
e
t
he
bas
e
s
t
at
i
o
n’
s
(
B
S
’
s
)
po
w
er
m
i
ni
m
i
z
at
i
o
n a
s
t
he obj
ec
t
i
v
e
f
unc
t
i
on
[4
,
5]
.
I
n [
2]
,
t
he r
e
s
our
c
e al
l
oc
at
i
on m
ec
hani
s
m
w
i
t
h a t
w
o
-
s
t
ep s
ubo
pt
i
m
al
a
l
gor
i
t
hm
i
s
pr
opos
e
d,
w
hi
c
h j
o
i
nt
l
y
a
l
l
oc
at
es
s
ubc
ar
r
i
er
an
d
po
w
er
t
o
m
ax
i
m
i
z
e
t
h
e m
i
ni
m
u
m
r
at
e of
al
l
us
er
s
w
hi
l
e
i
m
pr
ov
i
n
g t
h
e
s
y
s
t
em
t
hr
oughp
ut
.
I
n [
3]
,
a
ad
apt
i
v
e r
es
o
ur
c
e a
l
l
oc
a
t
i
on
b
a
s
ed
on
O
F
D
M
A
f
is
h
s
w
ar
m
al
g
or
i
t
hm
ar
e s
t
udi
e
d
t
o
m
ax
i
m
i
z
e t
h
e s
y
s
t
em
t
hr
oug
hpu
t
.
T
he pr
opos
ed
m
e
c
hani
s
m
i
nc
l
udes
s
ubc
a
r
r
i
er
al
l
oc
at
i
on
bas
e
d
on
a
n
e
w
wa
y
an
d
po
w
er
a
l
l
oc
a
t
i
on
b
as
ed
o
n
t
he
i
m
pr
ov
ed
ada
pt
i
v
e
f
i
s
h
a
l
gor
i
t
hm
.
T
he
aut
hor
s
i
n
[
4]
pr
op
os
e
t
h
e
r
es
o
ur
c
e
al
l
oc
at
i
o
n
al
g
or
i
t
hm
i
nc
l
ud
es
t
hr
ee
s
t
eps
.
F
i
r
s
t
l
y
,
i
n
it
ia
li
z
e
t
he
s
ubc
ar
r
i
er
s
al
l
oc
at
i
o
n
ac
c
or
di
n
g
t
o
c
ha
nne
l
gai
n,
t
he
n,
s
el
ec
t
t
he m
os
t
s
ui
t
ab
l
e s
u
bc
ar
r
i
er
f
or
us
er
s
b
y
c
om
par
i
ng t
he
po
w
er
.
F
in
a
ll
y
,
a
s
s
ig
n
bi
t
s
f
or
s
ubc
ar
r
i
er
s
b
y
u
t
i
l
i
z
i
ng
t
he
gr
e
ed
y
w
at
er
-
f
il
lin
g
a
lg
o
r
it
h
m
.
I
n
[
5
]
,
t
h
e
aut
hor
s
us
e
H
ung
ar
i
a
n al
gor
i
t
hm
t
o i
ni
t
i
al
l
y
p
er
f
or
m
t
he s
ubc
ar
r
i
er
al
l
oc
at
i
on,
a
nd t
hen d
y
n
am
i
c
al
l
y
al
l
oc
at
es
bi
t
t
o s
ubc
ar
r
i
er
s
bas
ed
o
n t
he
c
han
nel
s
t
at
e
i
nf
or
m
at
i
on (
C
S
I
)
,
f
i
na
l
l
y
adj
us
t
t
he
po
w
er
of
s
ubc
ar
r
i
er
s
t
o m
i
ni
m
i
z
e t
h
e
t
r
ans
m
i
s
s
i
on po
w
er
.
E
v
en
t
hou
gh
t
he
abo
v
e
al
g
or
i
t
hm
s
c
an
i
m
pr
ov
e
t
he
s
y
s
t
em
per
f
or
m
anc
e
under
c
er
t
ai
n
s
i
t
uat
i
ons
,
how
ev
e
r
,
onc
e t
he r
es
our
c
e al
l
oc
at
i
on o
pt
i
m
i
z
at
i
on
m
odel
s
ar
e bui
l
t
,
t
hei
r
obj
ec
t
i
v
e
f
unc
t
i
on i
s
c
ons
t
an
t
no m
a
t
t
er
ho
w
t
h
e w
i
r
e
l
es
s
net
w
or
k
s
t
at
e c
hanges
.
i
.
e,
t
h
e
y
ar
e al
l
f
i
x
ed
r
es
our
c
e
a
l
l
oc
a
t
i
o
n
al
g
or
i
t
hm
.
H
ow
e
v
er
,
t
h
e
us
er
s
’
m
ov
e
m
ent
w
i
l
l
l
ead
t
o
t
h
e
v
ar
i
a
t
i
o
n
of
net
w
or
k
env
i
r
onm
ent
,
and
t
h
e
f
i
x
ed
r
es
o
ur
c
e
a
l
l
oc
at
i
o
n
is
n
o
t
opt
i
m
al
f
or
t
he
c
ha
ng
i
ng
w
ir
e
le
s
s
net
w
or
k
.
I
n
t
he c
as
e
of
hi
gh
us
er
dens
i
t
y
,
t
h
e
r
es
our
c
e
sca
r
ci
t
y
i
s
t
he
m
aj
or
pr
o
bl
em
,
and
t
he
t
ar
get
of
r
es
our
c
e
a
l
l
oc
at
i
o
n
i
s
t
o
m
ax
i
m
i
z
e
t
he
n
et
w
o
r
k
c
apac
i
t
y
.
H
o
w
e
v
er
,
i
n
t
h
e
c
as
e
of
l
o
w
us
er
dens
i
t
y
,
t
he
t
ar
get
bec
om
es
t
o
guar
ant
e
e
t
he
qua
l
i
t
y
of
s
er
v
i
c
e
(
Q
o
S
)
an
d
r
e
duc
e
t
h
e
t
ot
al
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
1
6
9
3
-
6
930
T
E
L
KO
M
NI
K
A
V
o
l.
1
4
,
N
o
.
3
,
S
ept
em
ber
201
6
:
88
7
–
8
9
3
888
BS
’
s
po
w
er
c
ons
um
pt
i
o
n
a
c
c
or
di
ng
t
o
t
he
S
h
ann
on
T
heor
y
.
A
s
t
he
ex
i
s
t
i
ng
r
es
o
ur
c
e
al
l
oc
at
i
o
n
al
g
or
i
t
hm
c
annot
d
y
n
am
i
c
al
l
y
v
ar
y
w
i
t
h
t
he v
ar
y
i
n
g
w
i
r
e
l
es
s
net
w
or
k
s
t
at
e
/
us
er
dens
i
t
y
.
T
her
ef
or
e,
i
n
t
hi
s
pap
er
,
a
n
ada
pt
i
v
e
r
es
our
c
e
a
l
l
oc
at
i
on
al
g
or
i
t
hm
i
s
pr
opos
ed,
w
hi
c
h
n
ot
o
nl
y
ca
n
adj
us
t
i
t
s
el
f
t
o
t
he
v
ar
y
i
ng
ne
t
w
or
k
env
i
r
o
nm
ent
,
but
al
s
o
c
an
ac
hi
ev
e
t
he
s
el
f
-
opt
i
m
i
z
at
i
o
n.
T
he pr
opos
ed
al
gor
i
t
hm
f
i
r
s
t
l
y
b
ui
l
ds
a m
at
hem
at
i
c
al
m
odel
b
y
c
om
bi
n
i
ng
w
i
t
h t
h
e v
ar
y
i
ng
c
har
ac
t
er
i
s
t
i
c
s
of
us
er
de
ns
i
t
y
,
an
d t
h
en t
he
opt
i
m
al
a
l
l
oc
at
i
on
of
s
ub
c
ar
r
i
er
s
and p
o
w
er
i
s
per
f
or
m
ed b
y
us
i
ng t
h
e L
agr
an
ge dua
l
dec
om
pos
i
t
i
on.
F
ur
t
her
m
or
e,
i
n or
der
t
o r
educ
e t
he
c
o
m
put
at
i
on c
om
pl
ex
i
t
y
,
w
e dec
om
pos
e t
he or
i
g
i
n
al
pr
obl
em
i
nt
o s
e
v
er
al
i
nd
epen
den
t
s
ub
-
pr
obl
em
s
.
S
i
m
ul
at
i
o
n r
es
u
l
t
s
s
ho
w
t
hat
t
he pr
o
pos
e
d al
g
or
i
t
hm
c
an adapt
i
v
el
y
al
l
oc
at
e t
he
r
es
our
c
e t
o t
he us
er
s
ac
c
or
di
n
g t
o t
he v
ar
y
i
ng us
er
dens
i
t
y
w
hi
c
h r
epr
es
e
nt
s
t
o t
he
w
ir
e
le
s
s
net
w
or
k
s
t
at
e
.
2.
S
yst
em
M
o
d
e
l
a
n
d
P
r
o
b
l
e
m
F
o
r
m
u
l
a
ti
o
n
W
e
c
ons
i
der
a
s
i
ng
l
e
-
c
el
l
s
c
enar
i
o
f
or
t
he
m
ul
t
i
us
er
c
el
l
ul
ar
net
w
or
k
w
i
t
h
a
B
S
[
6
].
W
e
as
s
u
m
e t
hat
t
he
t
ot
al
num
ber
of
s
ubc
ar
r
i
er
s
i
s
N
,
t
h
e t
ot
al
t
r
ans
m
i
t
po
w
er
of
B
S
i
s
P
t
ot
a
l
,
t
he
m
i
ni
m
u
m
r
equi
r
em
ent
and
m
a
x
i
m
u
m
l
i
m
i
t
of
us
e
r
r
at
e
i
s
R
m
in
and
R
ma
x
,
r
es
pec
t
i
v
e
l
y
.
K
us
er
s
ar
e
uni
f
or
m
l
y
d
i
s
t
r
i
b
ut
ed
ov
er
t
he
c
el
l
,
an
d
h
k,n
i
s
t
he
c
h
a
nne
l
ga
i
n
of
s
ubc
ar
r
i
er
n
b
et
w
e
e
n
B
S
an
d
us
er
k
.
Mor
eo
v
er
,
t
h
e
B
S
i
s
as
s
u
m
ed
t
o
be
ab
l
e
t
o
o
bt
ai
n
t
he
C
SI
f
eed
bac
k
[
7
]
,
wh
i
c
h
f
o
l
l
o
ws
t
he f
r
eque
nc
y
s
e
l
ec
t
i
v
e
f
adi
ng [
8
].
T
he t
r
ans
m
i
s
s
i
on r
at
e
of
us
er
k
on s
ubc
ar
r
i
er
n
is
R
k,n
,
and
i
t
c
an
be
r
epr
es
ent
ed
a
s
[
9
]:
2
,,
,2
2
||
l
o
g
(1
)
kn
kn
kn
P
h
RB
σ
=
+
(
1)
W
h
er
e
B
and
2
σ
ar
e s
ubc
ar
r
i
er
ban
d
w
i
dt
h a
nd a
d
di
t
i
v
e
w
hi
t
e G
aus
s
i
an no
i
s
e po
w
er
,
r
es
pec
t
i
v
el
y
.
T
he
t
hr
ough
put
of
eac
h us
e
r
i
s
r
epr
es
ent
ed
as
:
2
,,
,2
2
1
||
l
o
g
(1
)
N
kn
kn
k
kn
n
P
h
R
B
ρ
σ
=
=
+
∑
(
2)
W
h
er
e
,
1
k
n
ρ
=
denot
es
t
ha
t
s
ubc
a
r
r
i
er
n
i
s
ut
i
l
i
z
ed b
y
us
er
k
,
or
el
s
e
,
0
k
n
ρ
=
.
T
he t
ot
al
po
w
er
c
ons
um
pt
i
on i
s
:
,,
11
KN
kn
kn
kn
PP
ρ
=
=
=
∑∑
(
3)
A
s
des
c
r
i
be
d ear
l
i
er
,
w
h
en us
er
de
ns
i
t
y
i
s
h
ig
h
,
t
he
m
ai
n
obj
ec
t
i
v
e of
r
es
our
c
e
al
l
oc
at
i
on
i
s
t
o m
ax
i
m
i
z
e t
h
e us
er
s
’
t
hr
oug
hput
,
i
.
e.
:
2
,,
,2
2
11
||
ma
x
lo
g
(
1
)
KN
kn
kn
kn
kn
P
h
B
ρ
σ
=
=
+
∑∑
(
4)
W
h
en us
er
dens
i
t
y
i
s
lo
w
,
t
he
m
ai
n
obj
ec
t
i
v
e
b
ec
om
es
t
o
m
i
ni
m
i
z
e t
he B
S
’
s
p
ow
e
r
c
ons
um
pt
i
on,
i
.
e.
:
,,
11
mi
n
KN
kn
kn
kn
P
ρ
=
=
∑∑
(
5)
H
enc
e,
w
e
c
an f
or
m
ul
at
e t
he
abo
v
e
m
ul
t
i
p
l
e
obj
ec
t
i
v
e
opt
i
m
i
z
at
i
on
f
unc
t
i
o
n as
a
s
i
ngl
e
-
obj
ec
t
i
v
e o
pt
i
m
i
z
a
t
i
o
n f
unc
t
i
on:
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
K
A
I
S
S
N
:
1
693
-
6
930
A
da
pt
i
v
e
R
es
our
c
e
A
l
l
oc
at
i
on A
l
g
or
i
t
h
m
i
n W
i
r
el
es
s
A
c
c
es
s
N
et
w
or
k
(
Z
han
j
un Li
u
)
889
2
,,
,
2
,,
2
11
11
||
m
a
x
l
o
g
(1
)
(1
)
KN
KN
kn
kn
kn
kn
kn
kn
kn
P
h
BP
αρ
α
ρ
σ
=
=
=
=
+
−−
∑∑
∑∑
(6
)
2
,,
,
2
mi
n
2
1
||
.
.
l
o
g
(1
)
N
kn
kn
kn
n
P
h
st
B
R
ρ
σ
=
+≥
∑
(
7)
2
,,
,
2
ma
x
2
1
||
l
o
g
(1
)
N
kn
kn
kn
n
P
h
BR
ρ
σ
=
+≤
∑
(
8
)
,,
11
KN
k
n
k
n
t
ot
al
kn
PP
ρ
=
=
≤
∑∑
(
9)
,
{
0
,
1}
kn
ρ
∈
(
10)
,
1
1
K
kn
k
ρ
=
=
∑
(
11)
W
h
er
e
α
v
ar
i
es
w
i
t
h t
he n
um
ber
o
f
us
er
s
K
,
and i
t
d
es
c
r
i
bes
t
he i
m
por
t
anc
e of
net
w
or
k
per
f
or
m
anc
e
w
hen t
he us
er
’
s
dens
i
t
y
i
s
i
n
v
ar
i
at
i
on.
E
qu
at
i
on (
7)
r
e
pr
es
ent
s
t
h
e m
i
ni
m
u
m
r
equi
r
em
ent
of
eac
h
us
er
’
s
ra
t
e
,
(8
)
r
epr
es
ent
s
t
he
m
ax
i
m
u
m
r
at
e
l
i
m
i
t
of
eac
h
us
er
,
(
9)
r
e
pr
es
ent
s
t
he
t
ot
al
t
r
ans
m
i
t
t
ed
po
w
er
l
i
m
i
t
of
t
he
B
S
,
(
10)
an
d (
11)
r
e
pr
es
ent
t
ha
t
a s
ubc
ar
r
i
er
c
ann
ot
b
e oc
c
upi
ed
b
y
m
or
e t
ha
n o
ne
us
er
at
t
he s
am
e
t
im
e
.
3
.
A
d
a
p
ti
v
e
R
e
s
o
u
r
c
e
A
l
l
o
c
a
ti
o
n
A
l
g
o
r
i
th
m
B
a
s
e
d
o
n
D
u
a
l
D
e
c
o
m
p
o
s
i
ti
o
n
I
f
an
opt
i
m
i
z
at
i
on
pr
ob
l
em
s
at
i
s
f
i
es
t
he
c
h
ar
ac
t
er
i
s
t
i
c
s
of
t
he
t
i
m
e
-
s
har
i
ng,
a
nd
t
h
en
t
h
e
or
i
gi
na
l
pr
obl
em
and i
t
s
d
ual
pr
ob
l
em
ha
v
e a z
er
o
dua
l
i
t
y
gap
[1
0
].
T
her
ef
or
e,
t
h
e
or
i
g
i
na
l
pr
obl
em
and t
he dual
pr
ob
l
em
hav
e t
he s
am
e op
t
im
a
l s
o
lu
t
io
n
.
E
qu
at
i
on
(
6
)
i
s
a non
-
c
on
v
ex
pr
obl
em
.
I
f
t
he
num
ber
o
f
s
ubc
ar
r
i
er
i
s
l
ar
ge
en
ou
gh,
t
he
opt
i
m
i
z
at
i
on
pr
ob
l
em
obv
i
o
us
l
y
s
a
t
is
f
ie
s
t
h
e
t
im
e
-
s
har
i
ng
c
ondi
t
i
o
n,
t
her
ef
or
e,
w
e
c
an t
ak
e ad
v
ant
age
of
t
h
e
Lagr
a
nge
du
al
dec
om
pos
i
t
i
on m
et
hod t
o
s
ol
v
e
t
he
opt
i
m
i
z
at
i
on pr
o
bl
em
w
her
e t
h
e s
o
l
ut
i
o
n i
s
pr
ogr
es
s
i
v
e
opt
i
m
al
[
1
1
].
T
her
ef
or
e,
i
n t
hi
s
s
ec
t
i
on,
t
he s
ol
ut
i
o
n of
E
quat
i
on (
6
)
i
s
der
i
v
e
d b
y
us
i
ng L
agr
an
g
e dua
l
dec
om
pos
i
t
i
on
m
et
hod.
L
et
λ
,
β
an
d
µ
be
Lagr
ang
e
m
ul
t
i
p
l
i
er
s
,
r
es
p
ec
t
i
v
e
l
y
.
T
her
ef
or
e,
t
h
e
Lagr
a
nge
dua
l
f
unc
t
i
on of
(
6
)
c
an be
ex
pr
es
s
ed
as
:
2
,,
,
2
,,
2
11
11
2
,,
,
2
mi
n
2
11
2
,,
ma
x
,
2
,
2
11
(
,,
)
m
a
x
(
,,
,,
)
||
m
a
x
{
[
l
o
g
(1
)
(1
)
]
||
(
l
o
g
(1
)
)
||
(
l
o
g
(1
)
)
(
KN
KN
kn
kn
kn
kn
kn
kn
kn
KN
kn
kn
k
kn
kn
KN
kn
kn
k
k
n
t
ot
al
k
n
k
k
n
gL
P
h
BP
P
h
B
R
P
h
R
B
u
P
P
µµ
αρ
α
ρ
σ
λ
ρ
σ
β
ρ
ρ
σ
=
=
=
=
=
=
=
=
=
=
+
−−
+
+−
+
−
+
+−
∑∑
∑∑
∑∑
∑∑
λβ
ρ
λβ
P
,
11
22
,,
,,
,2
,
2
22
11
2
,,
2
,
mi
n
ma
x
2
1
1
)}
||
||
m
a
x
{
[
l
o
g
(1
)
(1
)
l
o
g
(1
)
||
l
o
g
(1
)
]
}
KN
n
kn
KN
kn
kn
kn
kn
kn
kn
k
kn
K
K
kn
kn
k
k
n
k
k
t
ot
al
k
k
P
h
P
h
B
P
B
P
h
B
P
R
RP
ρα
α
β
σσ
λ
µ
λ
βµ
σ
=
=
=
=
=
=
=
+
−−
−
+
+
+
−−
+
+
∑∑
∑∑
∑∑
(
12)
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
1
6
9
3
-
6
930
T
E
L
KO
M
NI
K
A
V
o
l.
1
4
,
N
o
.
3
,
S
ept
em
ber
201
6
:
88
7
–
8
9
3
890
W
h
er
e:
1,
1
1,
2
1,
,1
,1
,
...
...
...
...
...
...
n
k
k
kn
ρρ
ρ
ρρ
ρ
ρ=
,
1,
1
1,
2
1,
,1
,
2
,
...
...
...
...
...
...
n
k
k
kn
pp
p
p
p
p
=
P
O
pt
i
m
i
z
at
i
on
pr
ob
l
em
of
dual
d
om
ai
n c
an b
e r
epr
es
e
nt
ed as
:
0
,
0
,
0
mi
n
(
,
,
)
g
µ
µ
≥
≥≥
λβ
λβ
(
13)
A
s
t
he
du
al
dom
ai
n pr
obl
e
m
(,
,
)
g
µ
λβ
i
s
l
i
ne
ar
f
or
λ
,
β
an
d
µ
,
t
her
ef
or
e,
i
t
b
ec
om
e
s
a c
onv
ex
opt
i
m
i
z
at
i
o
n pr
ob
l
em
,
and t
hen
w
e c
an us
e s
ubgr
a
di
e
nt
i
t
er
at
i
on m
et
hod
t
o guar
a
nt
ee
t
hat
t
he
(,
,
)
g
µ
λβ
c
an c
on
v
er
ge
t
o
m
i
ni
m
u
m
v
al
ue
.
M
or
eo
v
e
r
,
t
he L
agr
an
ge m
ul
t
i
pl
i
er
s
ar
e
upda
t
ed
bas
ed
on
t
he f
o
l
l
o
w
i
ng
E
q
uat
i
o
ns
[
12]
:
*2
,,
1*
,
2
mi
n
2
1
||
[
l
o
g
(1
)
]
k
N
kn
kn
i
ii
k
k
kn
n
P
h
yB
R
λ
λ
ρ
σ
+
=
=
−
+−
∑
(
14)
1
**
,,
1
()
N
i
i
i
t
ot
al
k
n
k
n
n
zP
P
µµ
ρ
+
=
=
−−
∑
(
15)
,
*2
,,
1*
ma
x
2
2
1
||
[
l
o
g
(1
)
]
k
n
N
kn
kn
i
ii
k
kk
n
P
h
R
B
β
βϖ
ρ
σ
+
=
=
−
−
+
∑
(
16)
W
h
er
e
*
,
kn
ρ
and
*
,
kn
P
ar
e
t
he
opt
i
m
a
l
s
ubc
ar
r
i
er
an
d
p
o
w
er
al
l
o
c
at
i
on
w
h
i
c
h
c
an
s
at
i
s
f
y
eq
uat
i
on
(
6)
,
r
es
pec
t
i
v
e
l
y
.
T
he
num
ber
of
i
t
er
at
i
ons
i
s
i
.
i
k
y
,
i
k
ϖ
and
i
z
ar
e
t
he
i
t
er
a
t
i
o
n
s
t
e
p
l
eng
t
h
of
Lagr
a
nge m
ul
t
i
pl
i
er
s
,
r
es
pe
c
t
i
v
e
l
y
.
1
1
1
l
i
m
0,
;
l
i
m
0,
;
l
i
m
0,
i
i
ii
i
i
kk
k
k
i
ii
i
i
i
y
y
zz
ϖ
ϖ
∞
∞
∞
→∞
→∞
→∞
=
=
=
=
=
∞=
=
∞
=
=
∞
∑∑
∑
(
17)
I
f
t
he
s
t
ep
l
e
ngt
h
s
at
i
s
f
i
es
t
he
c
r
i
t
er
i
a
i
n
(
17)
,
t
h
e
s
ub
gr
adi
ent
m
et
hod
c
an
c
o
nv
er
ge
t
o
t
he o
pt
i
m
al
d
ual
s
ol
u
t
i
o
n.
T
he opt
i
m
i
z
at
i
on
v
ar
i
ab
l
es
ρ
and
P
need
t
o b
e de
t
er
m
i
ned un
der
t
h
e c
as
e of
gi
v
e
n
dua
l
v
ar
i
abl
es
λ
,
β
and
µ
f
or
t
he
s
ol
ut
i
on
of
t
he
dua
l
f
unc
t
i
on
(,
,
)
g
µ
λβ
.
I
n
or
der
t
o
r
e
duc
e
t
he c
om
put
at
i
on c
om
pl
ex
i
t
y
,
t
he
du
al
pr
ob
l
em
i
s
dec
o
m
pos
ed i
nt
o
N
i
nde
pen
den
t
opt
i
m
i
z
at
i
o
n
pr
obl
em
.
mi
n
ma
x
1
1
1
(,
,
)
(,
,
)
N
K
K
n
k
k
t
ot
al
n
k
k
g
g
R
RP
µ
µλ
β
µ
=
=
=
=
−+
+
∑
∑∑
λβ
λβ
(
18)
2
,,
,
2
,
,
2
11
22
,,
,,
,
2,
2,
,
22
11
1
||
(
,
)
m
a
x
[
l
o
g
(1
)
(1
)
||
||
l
o
g
(1
)
l
o
g
(1
)
]
KK
kn
kn
n
kn
kn
kn
kk
KK
K
kn
kn
kn
kn
kn
k
kn
k
kn
kn
kk
k
P
h
gB
P
P
h
P
h
B
BP
µ
ρα
ρ
α
σ
ρ
λ
ρ
β
ρ
µ
σσ
=
=
=
=
=
=
+
−−
+
+−
+−
∑∑
∑∑
∑
λ
(
19)
T
he
E
qua
t
i
o
n (
19)
c
a
n b
e
t
r
ans
f
er
r
ed i
nt
o t
h
e b
el
o
w
f
or
m
at
s
b
y
us
i
n
g
K
ar
s
h
-
K
uhn
-
T
uc
k
er
c
ondi
t
i
on
[
13
]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
K
A
I
S
S
N
:
1
693
-
6
930
A
da
pt
i
v
e
R
es
our
c
e
A
l
l
oc
at
i
on A
l
g
or
i
t
h
m
i
n W
i
r
el
es
s
A
c
c
es
s
N
et
w
or
k
(
Z
han
j
un Li
u
)
891
2
,,
,
22
,
,,
|
|(
)
(,
)
(1
)
,
l
n(
2)
(
|
|
)
kn
kn
k
k
n
kn
kn
kn
kn
Bh
g
kn
P
P
h
ρ
α
λ
β
µ
ρ
αµ
σ
+
−
∂
=
−
−+
∀
∂
+
λ
(
20)
2
,,
2,
2
,
||
(,
)
l
o
g
(1
)
(
)
(1
)
,
kn
kn
n
k
k
kn
kn
P
h
g
B
P
kn
µ
α
λ
β
α
µ
ρ
σ
∂
=
+
+
−
−
−+
∀
∂
λ
(
21)
T
he opt
i
m
al
po
w
er
al
l
oc
at
i
on c
a
n be
obt
a
i
n
ed u
nde
r
t
he c
ond
i
t
i
ons
of
t
he
sp
e
ci
f
i
c
s
ubc
ar
r
i
er
a
l
l
oc
a
t
i
o
n b
y
us
i
ng of
K
K
T
c
ondi
t
i
on
:
2
,
2
,
()
[
]
,
l
n(
2)
(
1
)
||
kk
kn
kn
B
P
kn
h
α
λ
β
σ
αµ
+
+
−
=
−
∀
−
+
(
22)
W
h
er
e
[
]
A
+
r
epr
es
ent
(
)
m
a
x
0,
A
,
and (
2
3)
i
s
der
i
v
e
d b
y
s
u
bs
t
i
t
u
t
i
n
g (
22)
t
o (
1
9)
:
2
2
,
2
22
1
,
|
|(
)
()
(
,
)
m
a
x
(
)
[l
o
g
(
)]
(
1
)[
]
l
n(
2)
(
1
)
l
n(
2)
(
1
)
|
|
kn
k
k
kk
n
kk
kK
kn
Bh
B
gB
k
h
α
λ
β
α
λ
β
σ
µ
α
λ
β
α
µ
αµ
σ
αµ
+
+
≤≤
+
−
+
−
=
+
−
−−
+
−
∀
−
+
−
+
λ
(
23)
F
r
o
m
t
he ab
ov
e a
nal
y
s
es
,
w
e c
an
get
t
he
opt
i
m
al
al
l
o
c
at
i
on
r
ul
e of
s
ubc
ar
r
i
er
:
2
2
,
*
2
22
,
|
|(
)
()
a
rg
m
a
x
{
(
)
[l
o
g
(
)]
(
1
)[
]
}
l
n(
2)
(
1
)
l
n(
2)
(
1
)
|
|
kn
k
k
kk
kk
k
kn
Bh
B
k
B
h
α
λ
β
α
λ
β
σ
α
λ
β
α
µ
αµ
σ
αµ
+
+
+
−
+
−
=
+
−
−−
+
−
−+
−+
*
,
1
kn
ρ
=
(
24)
T
he
opt
i
m
al
po
w
er
a
l
l
oc
a
t
i
on
r
ul
es
:
t
h
e
B
S
al
l
oc
at
es
po
w
er
on
t
he
c
or
r
es
pond
i
ng
s
ubc
ar
r
i
er
of
us
er
s
b
y
f
ol
l
o
w
i
n
g eq
uat
i
on (
2
2)
,
w
he
n t
he a
l
l
oc
at
i
on of
s
ubc
ar
r
i
er
i
s
al
r
e
ad
y
f
i
ni
s
hed.
W
e
f
i
nal
l
y
obt
ai
n
t
he
opt
i
m
al
al
l
oc
at
i
o
n
r
ul
es
of
s
u
bc
ar
r
i
er
an
d
po
w
er
b
y
d
e
r
i
v
i
ng
E
qu
at
i
on (
6)
,
w
h
i
c
h c
an
ad
apt
t
o t
h
e us
er
dens
i
t
y
.
T
he al
l
oc
at
i
on
r
ul
es
of
t
he
pr
op
os
ed a
l
g
or
i
t
hm
ar
e s
um
m
ar
i
z
ed
as
:
S
t
ep
1:
I
n
it
ia
li
z
e
0
k
y
,
0
z
,
0
k
ϖ
,
0
k
λ
,
0
µ
,
0
k
β
,
,
kn
∀
w
i
t
h r
and
om
non
-
nega
t
i
v
e nu
m
ber
s
.
S
te
p
2
:
O
bt
a
i
n o
pt
i
m
al
s
u
bc
ar
r
i
er
al
l
oc
at
i
o
n b
y
us
i
n
g t
he c
ur
r
ent
Lagr
a
ng
e m
ul
t
i
pl
i
er
ac
c
or
di
ng
t
o
equ
at
i
on (
2
4)
.
St
e
p
3:
O
bt
ai
n op
t
i
m
al
po
w
er
al
l
oc
at
i
on
on
eac
h s
u
bc
a
r
r
i
er
ac
c
or
di
n
g t
o
equ
at
i
on (
22)
.
S
t
ep
4:
U
pdat
e L
agr
an
ge
m
ul
t
i
pl
i
er
ac
c
or
di
ng t
o eq
ua
t
i
ons
(
1
4)
,
(
15)
a
nd (
1
6)
.
S
te
p
5
:
Go
bac
k
t
o s
t
ep2
u
nt
i
l
c
on
v
er
g
enc
e.
4
.
S
i
m
u
l
a
ti
o
n
A
n
al
ysi
s
I
n t
hi
s
s
ec
t
i
o
n,
w
e d
es
c
r
i
be
t
he s
i
m
ul
at
i
o
n par
am
et
er
s
and p
er
f
or
m
anc
e of
t
he pr
opos
ed
al
g
or
i
t
hm
.
T
he s
i
m
ul
at
i
on
par
am
et
er
s
ar
e s
ho
w
n
i
n
T
abl
e 1.
F
i
gur
e
1 i
s
t
he c
a
pac
i
t
y
c
o
m
par
i
s
i
on am
ong al
gor
i
m
1,
al
gor
i
t
hm
2 and t
he pr
opos
ed a
l
g
or
i
t
hm
,
w
h
er
ei
n
al
g
or
i
t
hm
1 onl
y
t
ak
es
t
he
m
ax
i
m
i
z
at
i
o
n of
s
y
s
t
em
t
hr
oughput
as
t
he o
b
j
ec
t
i
v
e f
unc
t
i
o
n [
14]
,
a
nd a
l
gor
i
t
hm
2 onl
y
c
ons
i
der
s
t
he
m
i
ni
m
i
z
at
i
on
of
B
S
’
s
po
w
er
c
ons
um
pt
i
o
n
as
t
he
obj
ec
t
i
v
e
f
unc
t
i
o
n
[
15]
.
F
r
om
t
he
F
i
gur
e
1,
w
e c
an
obs
er
v
e
t
hat
t
he c
ur
v
e
of
t
hr
oug
hpu
t
appr
o
ac
hes
t
o a
l
gor
i
t
hm
2
w
hen
t
he
us
er
dens
i
t
y
i
s
l
o
w
,
bec
aus
e
i
n t
hi
s
c
as
e,
t
he
w
i
r
e
l
es
s
r
es
o
ur
c
e i
s
v
er
y
s
uf
f
i
c
i
ent
,
a
nd
t
he m
ai
n t
ar
get
of
r
es
our
c
e a
l
l
oc
a
t
i
o
n
i
s
t
o
m
i
ni
m
i
z
e
t
h
e
B
S
’
s
p
o
w
er
c
ons
um
pt
i
on
w
hi
c
h
i
s
abl
e
t
o s
a
t
i
s
f
y
t
he
us
er
r
at
e r
equi
r
em
ent
ac
c
or
di
n
g t
o t
he
S
ha
nno
n
th
e
o
r
y
, i
.e
.,
C
h
an
nel
c
ap
ac
i
t
y
i
s
not
o
nl
y
depe
nde
nt
on t
h
e ban
d
w
i
d
t
h,
but
a
l
s
o
t
he t
r
ans
m
i
s
s
i
on
po
w
er
.
H
o
w
e
v
er
,
w
i
t
h t
he i
nc
r
eas
e of
t
he n
um
ber
of
us
er
s
,
t
he
w
i
r
el
es
s
r
es
our
c
e b
ec
om
es
s
hor
t
ag
e,
a
nd
t
he
m
ai
n t
ar
g
et
bec
om
es
t
o
m
a
x
im
i
z
e t
h
e s
y
s
t
em
t
hr
oughpu
t
,
th
a
t i
s
to
s
a
y
,
i
n or
d
er
t
o
sa
t
i
sf
y
t
he
us
er
r
at
e r
equ
i
r
em
ent
,
it
’
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SSN
:
1
6
9
3
-
6
930
T
E
L
KO
M
NI
K
A
V
o
l.
1
4
,
N
o
.
3
,
S
ept
em
ber
201
6
:
88
7
–
8
9
3
892
c
annot
t
ak
e t
he m
i
ni
m
i
z
at
i
on of
B
S
’
s
po
w
er
c
ons
um
pt
i
o
n as
t
he
obj
ec
t
i
v
e
f
unc
t
i
on
aga
i
n,
and
c
ons
eque
nt
l
y
,
t
he c
ur
v
e
ap
pr
oac
hes
t
o
a
l
gor
i
t
hm
1
.
F
i
gur
e 2
i
s
t
h
e B
S
’
s
po
w
er
c
ons
u
m
pt
i
on of
al
g
or
i
t
hm
1,
al
g
or
i
t
hm
2
and
t
he
pr
o
pos
ed
al
g
or
i
t
hm
.
F
r
o
m
F
i
gur
e 2,
w
e c
an s
ee
t
hat
w
h
en t
h
e
num
ber
o
f
us
er
s
i
s
l
ow
,
t
he obj
ec
t
i
v
e of
r
es
our
c
e al
l
oc
at
i
o
n
i
s
t
o s
at
i
s
f
y
t
h
e m
i
ni
m
u
m
r
at
e r
equi
r
em
ent
of
eac
h
us
er
a
n
d m
i
ni
m
i
z
e
t
he
BS
’
s
po
w
er
c
ons
um
pt
i
on
,
w
her
eas
t
h
e
m
ax
i
m
i
z
at
i
o
n of
s
y
s
t
em
t
hr
oughput
i
s
not
t
he m
ai
n f
ac
t
or
w
h
ic
h
w
e
s
h
oul
d
c
on
s
i
d
er
.
T
her
ef
or
e,
t
he
c
ur
v
e
of
pow
er
c
ons
um
pt
i
on
appr
oac
h
es
t
o
al
g
or
i
t
hm
2,
w
h
i
c
h ac
hi
ev
es
t
he o
pt
i
m
i
z
at
i
on
of
po
w
er
c
ons
um
pt
i
on
.
W
i
t
h t
he i
nc
r
eas
e of
u
s
er
dens
i
t
y
,
t
h
e
r
es
our
c
e
s
hor
t
age
b
ec
om
e
s
s
er
i
ous
,
and
t
he
n
t
he
m
ax
i
m
i
z
at
i
o
n
of
s
y
s
t
em
t
hr
oughput
b
ec
om
e
s
t
he m
ai
n obj
ec
t
i
v
e,
w
h
i
c
h l
ea
ds
t
o m
or
e pow
er
c
ons
um
pt
i
on t
o s
at
i
s
f
y
t
he r
at
e
r
equi
r
em
ent
bas
ed
on t
he
S
h
ann
on
t
h
eor
y
.
T
her
ef
or
e,
t
he c
ur
v
e of
po
w
er
c
ons
um
pt
i
on
appr
o
ac
hes
t
o
al
g
or
i
t
hm
1,
w
h
i
c
h
i
nc
r
eas
e
t
h
e
p
o
w
er
c
o
ns
um
pt
i
on.
I
n
c
onc
l
us
i
on
,
t
h
e
pr
opos
e
d
al
gor
i
t
hm
c
an
be a
dapt
i
v
e t
o t
h
e v
ar
y
i
ng
of
us
er
dens
i
t
y
as
w
el
l
as
t
he
c
ha
ngi
ng
net
w
or
k
env
i
r
o
nm
ent
.
F
i
gur
e
1
.
N
um
ber
of
U
se
r
s v
s.
C
a
p
a
ci
t
y
F
i
gur
e
2
.
N
um
ber
of
U
se
r
s v
s.
P
o
w
e
r
T
abl
e 1
.
S
i
m
ul
at
i
on
P
ar
am
et
er
s
P
ar
a
m
et
er
s
V
al
ues
S
y
s
t
e
m
B
andw
i
dt
h
2
M
Hz
N
um
ber
o
f
s
ubc
ar
r
i
er
s
64
Us
e
r
’
s
l
ow
es
t
r
at
e
17 bi
t
/
s
Us
e
r
’
s
m
ax
i
m
al
r
at
e
30 bi
t
/
s
T
ot
al
pow
er
46 db
m
W
hi
t
e
G
au
s
s
i
an noi
s
e pow
er
10
-
2
W
C
hannel
m
odel
R
ay
l
ei
gh
f
adi
ng c
hannel
N
um
ber
o
f
us
e
r
s
K
1~
30
α
K
/
30
5
.
C
o
n
c
l
u
s
i
o
n
A
s
t
he t
r
ad
i
t
i
ona
l
r
es
our
c
e al
l
oc
a
t
i
o
n c
ann
ot
ada
pt
i
v
e
l
y
v
ar
y
w
i
t
h t
h
e
w
i
r
el
es
s
c
o
m
m
uni
c
at
i
on en
v
i
r
o
nm
e
nt
.
T
her
ef
or
e,
i
n t
h
i
s
pa
per
,
w
e pr
o
pos
e a
n ada
pt
i
v
e r
es
our
c
e
al
l
oc
at
i
on
al
gor
i
t
hm
t
o ad
apt
t
o t
he
d
y
n
am
i
c
net
w
o
r
k
env
i
r
onm
ent
.
F
i
r
s
t
l
y
,
a
n op
t
i
m
i
z
at
i
on
m
at
he
m
at
i
c
al
m
odel
i
s
bui
l
t
ac
c
or
di
n
g t
o t
he
v
ar
y
i
ng
n
et
w
or
k
env
i
r
onm
ent
.
T
hen,
t
he
w
ei
g
ht
s
of
obj
ec
t
i
v
e f
unc
t
i
on
d
y
n
am
i
c
al
l
y
v
ar
y
w
i
t
h t
hi
s
k
i
nd
of
v
ar
i
at
i
on,
w
hi
c
h
l
e
ads
t
o t
he a
dapt
at
i
on.
F
ur
t
her
m
or
e,
t
he L
agr
an
ge
dua
l
m
et
hod
a
nd s
u
bgr
a
di
ent
m
et
hod
i
s
ut
i
l
i
z
ed
t
o
g
et
t
h
e
op
t
i
m
al
s
ol
ut
i
on of
t
he o
pt
i
m
i
z
at
i
o
n m
odel
.
F
i
na
l
l
y
,
i
n or
der
t
o v
er
i
f
y
t
h
e al
gor
i
t
hm
p
er
f
or
m
anc
e,
w
e
as
s
u
m
e
t
hat
us
er
de
ns
i
t
y
i
s
ut
i
l
i
z
ed
t
o
r
e
pr
es
ent
t
he
v
ar
y
i
ng
n
et
w
or
k
env
i
r
onm
ent
,
and
t
he
t
o
t
a
l
w
i
r
el
es
s
r
es
our
c
e i
s
f
i
x
ed.
T
he r
es
our
c
e
s
hor
t
ag
e i
s
t
he
m
os
t
c
hal
l
eng
es
at
pr
e
s
ent
w
h
en
t
he
us
er
dens
i
t
y
i
s
h
i
g
her
,
s
o
t
he t
ar
g
et
of
r
es
our
c
e al
l
o
c
at
i
on
bec
om
es
t
o
m
ax
i
m
i
z
e t
h
e s
y
s
t
e
m
t
hr
oug
hpu
t
.
H
o
w
e
v
er
,
t
he
r
es
our
c
e i
s
s
uf
f
i
c
i
ent
w
h
en
t
he us
er
de
ns
i
t
y
i
s
l
o
w
er
,
an
d t
h
e t
ar
get
of
0
5
10
15
20
25
30
0
100
200
300
400
500
600
700
N
u
m
ber
of
u
s
er
s
C
apac
i
t
y
(
bi
t
/
s
)
T
h
e pr
opos
ed a
l
gor
i
t
h
m
al
g
o
r
i
t
hm
1
al
g
o
r
i
t
hm
2
0
5
10
15
20
25
30
0
5
10
15
20
25
30
35
40
45
N
u
m
ber
of
u
s
er
s
P
o
w
e
r (W
)
T
h
e pr
opos
ed a
l
gor
i
t
h
m
al
g
o
r
i
t
hm
1
al
g
o
r
i
t
hm
2
Evaluation Warning : The document was created with Spire.PDF for Python.
T
E
L
KO
M
NI
K
A
I
S
S
N
:
1
693
-
6
930
A
da
pt
i
v
e
R
es
our
c
e
A
l
l
oc
at
i
on A
l
g
or
i
t
h
m
i
n W
i
r
el
es
s
A
c
c
es
s
N
et
w
or
k
(
Z
han
j
un Li
u
)
893
r
es
our
c
e al
l
oc
at
i
on
b
ec
om
es
t
o s
at
i
s
f
y
t
he Q
o
S
r
eq
ues
t
of
eac
h us
er
and
l
o
w
er
t
he po
w
er
c
ons
um
pt
i
on.
S
i
m
ul
at
i
o
n r
e
s
ul
t
s
s
ho
w
t
h
at
,
i
n t
he
c
as
e
of
v
ar
y
i
ng
us
er
dens
i
t
y
,
w
e
c
an
ob
t
ai
n
t
he a
dapt
i
v
e r
es
our
c
e a
l
l
o
c
at
i
on
b
y
c
h
ang
i
n
g t
he
w
ei
g
ht
s
of
obj
ec
t
i
v
e f
unc
t
i
o
n,
an
d i
t
’
s
al
s
o
v
er
i
f
y
t
he
v
a
l
i
di
t
y
of
t
he
pr
opos
e
d
a
l
gor
i
t
hm
.
W
e
bel
i
e
v
e
t
hat
our
pr
opos
e
d
al
gor
i
t
hm
i
s
pr
om
i
s
i
ng f
or
t
he f
ut
ur
e 5G
m
obi
l
e c
om
m
uni
c
at
i
on s
y
s
t
em
.
A
c
k
n
o
w
l
e
d
g
e
m
e
n
ts
T
hi
s
w
or
k
i
s
s
upp
or
t
ed
b
y
863
pr
oj
ec
t
(
2
014
A
A
01A
701)
and
t
h
e S
c
i
enc
e an
d
T
ec
hnol
og
y
R
es
e
ar
c
h P
r
oj
ec
t
of
C
hongq
i
n
g Mu
ni
c
i
pal
E
d
uc
at
i
o
n C
om
m
i
s
s
i
on
of
C
hi
na
(
K
J
1205
10)
.
R
ef
er
en
ces
[1
]
Li
u J
i
ay
ong.
R
es
ear
c
h on A
d
apt
i
v
e R
es
o
ur
c
e A
l
l
oc
at
i
on P
r
obl
e
m
i
n O
F
D
M
S
y
s
t
em
.
U
n
i
ve
r
s
i
t
y o
f
E
l
ec
t
r
oni
c
S
c
i
en
c
e a
nd T
ec
h
n
ol
ogy
of
C
hi
na
.
20
08.
[2
]
Z
hang
X
i
aox
i
a,
X
uem
i
n
S
h
er
m
an
S
h
en,
X
i
e
Li
ang
l
i
an
g.
J
oi
nt
S
u
bc
ar
r
i
er
a
nd
P
ow
er
A
l
l
oc
a
t
i
on
f
o
r
C
ooper
at
i
v
e C
om
m
uni
c
at
i
on
s
i
n LT
E
-
A
dv
anc
ed N
e
t
w
or
k
s
.
I
E
E
E
T
r
an
s
ac
t
io
n
s
on W
i
r
e
l
es
s
C
om
m
uni
c
at
i
ons
.
20
14
;
1
3
(2
):
658
-
6
68.
[3
]
W
a
n
g Z
h
ao,
L
i
Y
oum
i
n
g,
C
h
en B
i
n.
O
F
D
M
A
A
dapt
i
v
e R
e
s
our
c
e A
l
l
oc
at
i
o
n B
as
ed on
F
i
s
h S
w
ar
m
A
l
gor
i
t
hm
.
A
c
t
a P
h
y
s
ic
a
S
in
ic
a
.
2013
;
62(
12)
:
1
-
7.
[4
]
C
ao
S
haol
o
ng,
Z
h
ou
Li
,
Li
B
i
n.
R
es
our
c
e
A
l
l
oc
at
i
on
A
l
gor
i
t
hm
f
or
M
I
M
O
-
O
F
D
M
S
y
s
t
em
B
as
ed
o
n
M
i
ni
m
um
T
r
ans
m
i
t
t
i
ng P
ow
er
.
C
om
put
er
S
y
s
t
em
s
&
A
ppl
i
c
at
i
ons
.
201
4
;
23(
5)
:
19
2
-
1
95.
[5
]
T
i
an Y
i
l
an,
X
u B
oqi
ng.
S
ubc
a
r
r
i
er
and B
i
t
A
l
l
o
c
at
i
on A
l
g
or
i
t
hm
f
or
M
u
l
ti
-
us
er
O
F
D
M
S
y
s
t
em
.
R
a
dio
C
om
m
uni
c
at
i
ons
T
e
c
hno
l
og
y
.
2012
;
9
(4
):
35
-
37
.
[6
]
T
an Z
hengl
i
n,
D
eng X
i
aoc
h
en
g.
R
es
our
c
e A
l
l
oc
a
t
i
on A
l
gor
i
t
hm
f
or
S
i
ng
l
e C
el
l
D
ow
nl
i
n
k
o
f
M
ul
t
i
us
er
O
F
D
M S
y
s
t
e
m
.
I
nf
or
m
at
i
on
T
e
c
hno
l
og
y
.
2
012
;
8
(7
):
92
-
95
.
[7
]
J
ang
J
,
Lee K
B
.
T
r
ans
m
i
t
P
ow
er
A
dapt
a
t
i
on f
or
M
ul
t
i
us
er
O
F
D
M
S
y
s
t
em
s
.
I
E
E
E
J
our
nal
o
n
S
el
ec
t
e
d
A
r
eas
C
om
m
uni
c
at
i
on
s
.
2
003
;
21(
2)
:
171
-
178.
[8
]
K
M
A
w
an,
A
H
A
bdul
l
a
h,
K
H
us
s
a
i
n.
A
ddi
t
i
o
nal
R
es
our
c
e A
l
l
oc
at
i
on f
or
I
m
pr
ov
i
ng F
a
i
r
ne
s
s
i
n
W
i
M
A
X
.
T
EL
KO
M
N
I
KA
T
el
ec
om
m
uni
c
at
i
on
,
C
om
put
i
ng,
E
l
ec
t
r
oni
c
s
and C
on
t
r
ol
.
201
4
;
1
2(
2)
:
455
-
464.
[9
]
N
M
A
dr
i
ans
y
a
h,
M
A
s
v
i
al
,
B
B
udi
ar
dj
o.
M
odi
f
i
e
d G
r
ee
dy
P
hy
s
i
c
a
l
L
i
n
k
S
c
hed
ul
i
ng A
l
g
or
i
t
h
m
f
or
I
m
pr
ov
i
n
g W
i
r
el
e
s
s
M
es
h
N
et
w
or
k
P
er
f
or
m
anc
e
.
T
el
k
om
ni
k
a T
el
ec
om
m
uni
c
at
i
on,
C
om
put
i
n
g
,
E
l
ec
t
r
oni
c
s
and
C
o
nt
r
o
l
.
20
15
;
13(
1)
:
202
-
210
.
[1
0
]
Y
u
W
,
Li
u R
.
D
ual
M
et
hod
s
f
or
N
onc
onv
ex
S
pec
t
r
um
O
pt
i
m
i
z
a
t
i
on o
f
M
ul
t
i
c
ar
r
i
er
S
y
s
t
e
m
s
.
I
EEE
T
r
ans
ac
t
i
on on
C
om
m
uni
c
at
i
o
ns
.
2
006
;
54(
7)
:
13
10
-
1
322.
[1
1
]
Z
hang
C
ui
z
hi
,
C
he
n S
h
um
i
n
,
Y
u Q
i
ang,
Li
ang
S
hu
c
he
n
g,
X
u Y
uanx
i
n.
P
ow
er
A
l
l
o
c
at
i
on
an
d
S
ubc
ar
r
i
er
P
ai
r
i
n
g
f
or
A
F
-
O
F
D
M
B
as
ed
C
o
gni
t
i
v
e
R
adi
o
S
y
s
t
em
s
.
J
o
u
r
n
al
of
Z
he
j
i
an
g
U
n
i
v
er
s
it
y
(
E
ngi
ne
er
i
n
g S
c
i
en
c
e)
.
2011
;
62(
12)
:
225
9
-
22
64.
[1
2
]
B
oy
d S
,
X
i
ao L
,
M
ut
apc
i
c
A
.
S
ubgr
ad
i
en
t
M
et
hods
[
E
B
/
O
L
]
.
2003.
[1
3
]
B
oy
d S
,
Va
ndenb
er
gh
e L.
C
o
nv
ex
O
pt
i
m
al
.
C
am
br
i
d
ge
,
UK
:
C
am
br
i
d
ge U
ni
v
er
s
i
t
y
P
r
es
s
.
2004:
2
15
-
247.
[1
4
]
D
i
ng B
i
n.
T
he S
ub
-
o
pt
i
m
al
P
ow
er
A
l
l
oc
at
i
o
n A
d
apt
i
v
e
A
l
g
or
i
t
h
m
i
n O
F
D
M
T
ec
hnol
ogy
.
C
om
put
e
r
K
now
l
ed
ge a
nd T
e
c
h
nol
o
gy
.
2
015
;
2
0(
12)
:
22
3
-
224
.
[1
5
]
F
u
C
hun
s
he
ng,
F
u
B
i
l
i
n
,
C
he
n
Ler
,
S
he
F
eng
.
P
ow
er
M
i
ni
m
i
z
at
i
o
n
A
l
gor
i
t
hm
f
or
Mu
l
t
i
-
u
s
e
r
MI
MO
-
O
FD
M
S
y
s
te
m
.
V
i
de
o E
ng
i
ne
e
r
i
ng
.
2009
;
33(
S
1)
:
11
9
-
12
1.
Evaluation Warning : The document was created with Spire.PDF for Python.