Inter
national
J
our
nal
of
Electrical
and
Computer
Engineering
(IJECE)
V
ol.
10,
No.
4,
August
2020,
pp.
3854
3861
ISSN:
2088-8708,
DOI:
10.11591/ijece.v10i4.pp3854-3861
r
3854
Dynamic
r
esour
ce
allocation
f
or
opportunistic
softwar
e-defined
IoT
netw
orks:
stochastic
optimization
framew
ork
Sharhabeel
H.
Alnabelsi
1
,
Haythem
A.
Bany
Salameh
2
,
Zaid
M.
Albataineh
3
1
Computer
Eng.
Dept.,
F
aculty
of
Eng.
T
echnology
,
Al-Balqa
Applied
Uni
v
ersity
,
Jordan
1,2
Colle
ge
of
Engineering,
AL
Ain
Uni
v
ersity
,
AL
Ain,
United
Arab
Emirates
2,3
T
elecommunications
Eng.
Dept.,
Y
armouk
Uni
v
ersity
,
Jordan
Article
Inf
o
Article
history:
Recei
v
ed
No
v
9,
2019
Re
vised
Feb
11,
2020
Accepted
Feb
21,
2020
K
eyw
ords:
Cogniti
v
e
radio
netw
orks
Internet
of
things
Primary
users
Secondary
users
Stochastic
optimization
ABSTRA
CT
Se
v
eral
wireless
technologies
ha
v
e
recently
emer
ged
to
enable
ef
ficient
and
scalable
Internet-of-Things
(IoT)
netw
orking.
Cogniti
v
e
radio
(CR)
technology
,
enabled
by
softw
are-defined
radios,
is
considered
one
of
the
main
IoT
-enabling
technologies
that
can
pro
vide
opportunistic
wireless
access
to
a
la
r
ge
number
of
connected
IoT
de
vices.
An
important
challenge
in
this
domai
n
is
ho
w
to
dynamically
enable
IoT
transmis-
sions
while
achie
ving
ef
ficient
spectrum
usage
with
a
mini
mum
total
po
wer
consump-
tion
under
interference
and
traf
fic
demand
uncertainty
.
T
o
w
ard
this
end,
we
propose
a
dynamic
bandwidth/channel/po
wer
allocation
algorithm
that
aims
at
maximizing
the
o
v
erall
netw
ork’
s
throughput
while
selecting
the
set
of
po
wer
resulting
in
the
minimum
total
transmission
po
wer
.
This
problem
can
be
formulated
as
a
tw
o-stage
binary
linear
stochastic
programming.
Because
the
interference
o
v
er
dif
ferent
channels
is
a
contin-
uous
random
v
ariable
and
noting
that
the
interference
statistics
are
highly
correlated,
a
suboptim
al
sampling
solution
is
proposed.
Our
proposed
algorithm
is
an
adapti
v
e
algorithm
that
is
to
be
periodically
conducted
o
v
er
time
to
consider
the
changes
of
the
channel
and
interference
conditions.
Numerical
results
indicate
that
our
proposed
algorithm
significantly
increases
the
number
of
simultaneous
IoT
transmissions
com-
pared
to
a
typical
algorithm,
and
hence,
the
achie
v
ed
throughput
is
impro
v
ed.
Copyright
c
2020
Insitute
of
Advanced
Engineeering
and
Science
.
All
rights
r
eserved.
Corresponding
A
uthor:
Sharhabeel
H.
Alnabelsi,
Computer
Engineering
Dept.,
Al-Balqa
Applied
Uni
v
ersity
,
P
.O.
Box:
15008,
Amman
11134,
Jordan.
Email:
alnabsh1@bau.edu.jo
1.
INTR
ODUCTION
W
ith
the
e
xponential
gro
wth
of
Internet-of-things
(IoT)
applications
and
services,
it
is
e
xpected
that
more
than
50
billion
de
vices
will
be
connected
to
the
internet
by
2020.
IoT
netw
orking
connects
v
aried
wired
and
wireless
de
vices
and
systems.
The
enormous
number
of
connected
wireless
IoT
de
vices
significantly
increases
the
demand
for
more
spectrum
resources
and
ef
ficient
spectrum
utilization.
Softw
are-defined
net-
w
orking
enabled
by
cogniti
v
e
radio
(CR)
technology
is
considered
as
a
major
approach
to
impro
v
e
spectrum
utilization
and
pro
vide
wireless
access
to
a
lar
ge
number
of
connected
IoT
de
vices.
W
ireless
CR
technology
allo
ws
for
rapid
deplo
yment
of
scalable,
reliable
and
intelligent
IoT
netw
orking.
CR
technology
brings
intel-
ligence
right
to
the
edge
of
an
IoT
netw
ork.
The
intelligent
of
fered
by
the
CR
at
the
edge
nodes
pro
vides
a
complete
connecti
vity
stack
virtually
between
an
y
type
of
wireless
sensors
and
an
IoT
controller
.
J
ournal
homepage:
http://ijece
.iaescor
e
.com/inde
x.php/IJECE
Evaluation Warning : The document was created with Spire.PDF for Python.
Int
J
Elec
&
Comp
Eng
ISSN:
2088-8708
r
3855
(a)
Moti
v
ation
Due
to
the
f
act
that
the
demand
in
the
IoT
services
and
applications
is
e
xponentially
increasing,
CR
technology
allo
ws
to
use
underutilized
spectrum
using
dynamic
s
pectrum
access
technique.
In
a
CR
netw
ork
(CRN),
CR
users,
also
kno
wn
as
secondary
or
unlicensed
users,
are
a
w
are
about
licensed
spectrum
that
used
by
e
xisting
Primary
User
(PU)
netw
orks.
CR
users
can
opportunistically
access
the
licensed
spectrum
by
chang-
ing
their
transmission
parameters,
in
order
to
a
v
oid
af
fecti
ng
ongoing
PUs’
transmission.
This
has
moti
v
ated
the
need
for
a
ne
w
spectrum
access
technology
that
introduced
in
CRNs,
such
that
the
spectrum
utilization
is
enhanced
without
af
fecting
the
PUs
operation.
A
CRN
is
dif
ferent
from
the
traditional
multi-channel
wire-
less
netw
orks.
Most
importantly
,
CRN
e
xperience
out-of-system
and
in-system
random
interference.
Another
characteristic
of
a
CRN
is
that
users
may
need
to
transmit
with
a
relati
v
ely
lo
w
signal
po
wer
,
with
po
wer
masks
constraints,
in
order
to
a
v
oid
causing
harmful
interference
to
the
PUs
[1].
On
the
other
hand,
the
appli-
cations
supported
by
the
IoT
de
vices
are
v
ery
di
v
erse,
requiring
heterogeneous
uncertain
bandwidth
and
rate
demands.
When
applying
CR
technology
in
IoT
netw
orks
(CRIoT
netw
orks),
these
peculiar
characteristics
call
for
ne
w
stochastic
channel
access
mechanism
that
can
ef
ficiently
utilize
the
a
v
ai
lable
spectrum
to
maximize
the
number
of
simultaneous
CRIoT
transmissions
with
minimum
total
transmission
po
wer
,
and
impro
v
ed
netw
ork
throughput[2–9].
(b)
Contrib
utions
Pre
vious
channel
as
signment
approaches
in
traditional
multi-channel
and
CR
wireless
netw
orks
were
designed
assuming
a
v
erage
interference
conditions,
fix
ed
channel
bandwidth
and
fix
ed
spectrum
demands
per
user
.
In
this
w
ork,
we
propose
a
stochastic
bandwidth/channel/po
wer
allocation
algorithm
that
impro
v
es
the
netw
ork
performance.
The
maximization
problem
can
be
established
as
a
tw
o-stage
stochastic
binary
linear
program.
It
is
w
orth
mentioning
that
the
interference
is
a
continuous
random
v
ariable
that
is
highly-correlated
o
v
er
time
[1,
10–15].
Thus,
our
optimization
problem
has
an
infinite
realizations.
Therefore,
solving
for
an
op-
timal
solution
is
impossible.
Instead,
we
propose
a
suboptimal
sampling
solution
that
e
xploits
the
interference’
s
correlation.
(c)
Or
g
anization
The
rest
of
this
paper
i
s
or
g
anized
as
follo
ws:
Section
2
presents
the
related
w
ork.
In
Section
3,
the
problem
model,
description
and
formulation
are
introduced.
Section
4
e
xplains
the
process
of
channels
assign-
ment
and
bandwidth
allocation
in
the
access
windo
w
.
Section
5
sho
ws
the
numerical
results
for
the
performance
of
our
proposed
scheme
compared
with
traditional
approaches.
Finally
,
Section
6
presents
conclusions.
2.
RELA
TED
W
ORK
Channels
assignment
in
CRNs
is
dif
ferent
from
the
traditional
netw
orks,
due
to
the
f
act
that
c
h
a
nn
e
ls
a
v
ailability
changes
o
v
er
time
due
to
licensed
users
acti
vities.
Moreo
v
er
,
CR
users
are
po
wer
constrained,
such
that
their
transmission
po
wer
should
not
e
xceed
a
certain
limit
to
a
v
oid
causing
harmful
interference
to
licensed
users.
Consequently
,
satisfying
CR
user
s
data
rate
demand
becomes
challenging.
Therefore,
we
are
moti
v
ated
in
this
w
ork
to
consider
these
f
actors
for
our
proposed
adapti
v
e
channels
assignment
technique.
F
or
concurrent
channels
assignment,
in
[16]
authors
proposed
a
scheme
that
allo
ws
a
group
of
CR
users
to
be
assigned
channels
instead
of
one
user
at
a
time,
also
the
y
assumed
channels
do
not
ha
v
e
a
fix
ed
bandwidth
as
a
practical
assumption,
therefore,
netw
ork
throughput
is
increased.
In
[16,
17],
a
guard
band
notion
is
introduced
between
idle
channel
blocks,
in
order
to
minimize
the
ef
fect
of
adjacent
interference
and
maximize
spectrum
ef
ficienc
y
,
such
that
in
[17],
the
number
of
required
guard
bands
are
reduced
when
grouping
idle
channels
as
one
block.
T
w
o
channels
assignment
methods
are
de
v
eloped
in
[18],
in
order
to
maximize
spectrum
ef
ficienc
y:
the
static
single-stage
method
when
a
centralized
spectrum
manager
does
not
e
xist,
the
second
method
is
an
adapti
v
e
tw
o-stage
technique
which
is
suitable
for
centralized
spectrum
manager
.
In
addition
to
the
uncertainty
of
the
channels,
the
authors
also
consider
tw
o
aspects
in
their
models:
the
f
act
of
adjacent
channels
interference
and
channels
bonding
and
aggre
g
ation.
In
addition
to
man
y
proposed
protocols
in
literature
that
aim
to
enhance
netw
ork
capacity
,
throughput
and
optimize
transmitted
po
wer
[1,
19,
20].
F
air
channels
ass
ignment
and
ener
gy
optimization
are
considered
in
[21].
CR
users
transmission
po
wer
should
be
controlled,
in
orde
r
to
a
v
oid
interference
with
neighbor
li-
censed
users
transmission
[22].
F
or
CR
Ad-Hoc
Netw
orks
(CRAHNs)
[23],
transmission
po
wer
control
and
spectrum
assignment
methods
are
de
v
eloped
to
enhance
netw
ork
capacity
.
Spectrum
assignment
method
is
Dynamic
r
esour
ce
allocation
for
opportunistic...
(Sharhabeel
H.
Alnabelsi)
Evaluation Warning : The document was created with Spire.PDF for Python.
3856
r
ISSN:
2088-8708
presented
in
[24]
and
solv
ed
using
a
learning
technique,
also
an
adapti
v
e
po
wer
allocation
method
is
solv
ed
as
an
optimization
problem.
The
h
a
rmful
interference
reduction
to
licensed
users
is
studied
in
[25],
also
using
the
deep-reinforcement
learning
technique
[26],
mobile
CR
users
empo
wered
to
change
their
ph
ysical
location
when
jamming
is
high.
Researchers
ha
v
e
studied
netw
ork
connecti
vity
in
CRNs,
especially
,
it
is
essential
in
routing
stability
.
Noting
that
its
connecti
vity
is
dif
ferent
from
traditional
netw
orks,
since
the
licensed
spectrum
a
v
ailability
changes
o
v
er
time.
In
[27],
links
are
established
in
a
w
ay
that
minimizes
interference
and
enhance
connecti
vity
de
gree.
Also,
authors
in
[28]
proposed
some
CR
transcei
v
ers
to
be
maintain
the
lo
west
threshold
for
connecti
vity
.
F
or
routing
protection
in
terms
of
connecti
vity
,
a
resilient
method
is
introduced
in
[29–31].
Also,
CR
users
pack
ets
reco
v
ery
due
to
primary
users
acti
vity
is
studied
in
[32].
3.
MODELS,
PR
OBLEM
DESCRIPTION
AND
FORMULA
TIONS
3.1.
Netw
ork
model
In
this
w
ork,
we
will
consider
the
scenari
o
of
a
single-hop
opportunistic
wireless
cogniti
v
e
(unli-
censed)
radio
netw
ork
(CRN)
that
tries
to
e
xploit
spectrum
holes
in
the
presence
of
dif
ferent
(le
g
ac
y)
primary
radio
netw
orks
with
channels.
CR
user
acts
as
a
secondary
user
by
continuously
scanning
the
frequenc
y
spec-
trum
and
identifying
underutilized
channels
to
e
xploit
opportunistic
access.
The
CRN
comprises
a
collection
of
single-hop
users
between
which
requests
for
pack
et
trans
mission
arise.
Each
CR
user
can
transm
it
o
v
er
one
of
the
M
a
v
ailable
channels.
This
can
be
seen
as
M
possible
links.
Due
to
the
nature
of
wireless
CRNs,
a
channel
(link),
which
is
occupied
by
a
CR
user
,
cannot
be
allocated
to
other
CR
users
in
its
one-hop
communication
range.
Furthermore,
each
channel
link
e
xperiences
a
random
primary
netw
ork
interference
conditions,
and
each
CR
user
has
a
random
demand
data
rate.
T
o
satisfy
a
gi
v
en
demand,
a
bandwidth
must
be
allocated
for
each
channel.
Because
of
the
radio
capability
restrictions,
the
maximum
bandwidth
(
B
)
that
can
be
used
o
v
er
the
v
arious
channels
is
constrained.
Therefore,
the
opti
mization
problem
is
to
determine
channel
bandwidths
that
maximizing
the
o
v
er
all
netw
ork
throughput
(bandwidth
utilization)
while
selecting
the
set
of
po
wer
resulting
in
the
minimum
total
transmission
po
wer
.
This
problem
lends
itself
to
a
natural
tw
o-stage
stochastic
inte
ger
linear
programming.
That
is,
the
maximum
bandwidth,
which
can
be
used
by
CRN,
must
be
allocated
to
the
v
arious
channels
before
the
rate
demand
and
the
interference
conditions
can
be
kno
wn.
Once
B
has
been
allocated
to
dif
ferent
channel
s,
CR
requests
can
be
serv
ed
in
a
manner
that
allo
ws
ef
ficient
spectrum
use
with
minimum
total
po
wer
consumption.
The
optimization
of
bandwidth/channel/po
wer
is
an
adapti
v
e
algorithm
that
is
to
be
periodically
conducted
o
v
er
time
to
account
for
the
changes
for
the
channel
and
the
primary
netw
ork
interference
conditions.
The
distrib
utions
of
the
rate
demands
and
the
interference
po
wer
are
dynamically
updated
based
on
localized
spectrum
and
control
information
observ
ed
o
v
er
the
pre
vious
transmissions
time.
3.2.
Assumptions
and
feasibility
conditions
Before
formulating
our
optimization
problem,
we
first
state
our
assumptions
and
feasibility
conditions.
(a)
There
are
tw
o
sets:
i
2
I
:
channels,
and
j
2
J
:
CR
users.
(b)
The
rate
demand
e
d
j
is
a
discrete
uniform
random
v
ariables,
8
j
2
J
.
(c)
Each
CR
user
maintains
an
K-entry
historical
-data
table.
The
i
th
entry
in
the
table
consists
of
one
fields
indicating
the
pre
viously
observ
ed
interference
o
v
er
the
i
th
access
windo
w
(A
W)
time.
(d)
The
rate
demand
(
e
d
j
)
and
the
interference
(
e
P
(
i
)
I
;
8
i
2
I
)
are
independent
random
v
ariables.
(e)
The
interference
(
e
P
(
i
)
I
;
8
i
2
I
)
is
a
continuous
positi
v
e
random
v
ariable
with
unkno
wn
distrib
ution.
(f)
The
interference
at
dif
ferent
channels
is
independent
and
identically
distrib
uted
(iid).
T
o
ensure
a
feasible
spectrum
sharing,
we
introduce
these
constraints:
(a)
At
most
one
channel
can
be
assigned
for
one
transmission.
(b)
A
channel
cannot
be
assigned
for
more
than
one
transmission.
(c)
Rate
demand
constra
int:
the
data
rate
pro
vided
by
a
channel
should
be
greater
than
the
rate
demand
of
the
request
that
associated
with
that
channel.
(d)
The
CR-to-PU
spectrum
mask:
the
maximum
allo
w
able
transmission
po
wer
of
CR
us
ers
must
be
con-
strained
by
a
po
wer
mask,
such
that
the
CR
users
will
not
cause
unacceptable
interference
to
primary
users.
(e)
The
signal
to
interference
noise
ratio
(SINR)
at
a
CR
user
should
be
greater
than
the
minimum
required
threshold
at
the
selected
channel.
Int
J
Elec
&
Comp
Eng,
V
ol.
10,
No.
4,
August
2020
:
3854
–
3861
Evaluation Warning : The document was created with Spire.PDF for Python.
Int
J
Elec
&
Comp
Eng
ISSN:
2088-8708
r
3857
3.3.
Pr
oblem
f
ormulation
The
problem
of
bandwidth/channel/po
wer
allocation
can
be
formulated
as
a
tw
o-stage
stochastic
pro-
gramming.
In
the
first
stage,
the
maximum
bandwidth,
which
can
be
used
by
C
RN,
is
allocated
to
the
v
arious
channels
before
the
rate
demand
and
the
interference
conditions
can
be
kno
wn.
Then,
in
the
second
stage,
we
allocate/select
channels/po
wers
to
dif
ferent
CR
users
such
as
the
total
po
wer
consumption
is
minimized.
T
o
formulate
the
problem,
we
introduce
the
indices,
data,
random
v
ariables,
and
decision
v
ariables:
(a)
Sets
(indices):
i
2
I
:
channels,
and
j
2
J
:
CR
users.
(b)
Data:
B
is
the
total
bandwidth
that
can
be
allocated
to
the
v
arious
channels,
M
is
the
number
of
channels
that
are
to
be
considered,
N
is
the
total
number
of
users,
P
(
i
)
th
is
t
he
thermal
noise
po
wer
at
the
i
th
channel,
and
is
the
minimum
required
signal-to-Interference-and-noise-ratio.
(c)
Random
v
ariables:
=
(
e
d
j
;
e
P
(
i
)
I
)
:
t
he
random
v
ariables
that
represent
the
demand
and
the
interference
at
v
arious
channels
and
dif
ferent
users.
(d)
Random
Data:
P
(
i
)
j
=
(
P
(
i
)
th
+
e
P
(
i
)
I
)
:
the
required
transmit
po
wer
o
v
er
the
i
th
channel
for
the
j
th
user
o
v
er
the
v
arious
channels.
(e)
Decision
v
ariables:
X
i
:
is
the
amount
of
capacity
to
be
assigned
to
the
i
th
channel.
(
i
)
j
:
is
channel
assignment
indicator
that
is
gi
v
en
by:
(
i
)
j
=
1
;
if
channel
j
is
assigned
to
the
i
th
transmission;
0
;
otherwise.
(1)
W
ith
these
notations,
the
general-recourse
model
for
bandwidth/channel/po
wer
problem
is
gi
v
en
as:
max
X
i
E
[
h
(
X
;
)]
M
X
i
=1
X
i
B
X
i
0
i
2
I
(2)
where
h
(
X
;
)
represents
the
channel
utilization
when
the
demand
for
service
and
the
interference
are
gi
v
en.
This
function
is
represented
by
the
optimal
v
alue
function
of
a
second-stage
program.
Based
on
the
abo
v
e
notation,
the
second
stage
problem
can
be
formulated
as
follo
ws:
max
P
M
i
=1
P
N
j
=1
(
i
)
j
P
M
i
=1
P
N
j
=1
(
i
)
j
P
(
i
)
j
P
j
(
i
)
j
1
;
i
2
I
P
i
(
i
)
j
1
;
j
2
J
P
i
P
j
(
i
)
j
M
(
i
)
j
2
f
0
;
1
g
;
i
2
I
;
j
2
J
X
i
log
2
1
+
P
(
i
)
j
e
P
(
i
)
I
+
P
(
i
)
th
e
d
j
(
(
i
)
j
1)
;
i
2
I
;
j
2
J
(3)
where
is
a
v
ery
lar
ge
number
.
Clearly
,
the
formulation
in
(3)
is
a
tw
o-stage
stochastic
binary
linear
program.
3.4.
Suboptimal
sampling
pr
oblem
f
ormulation
Since
the
interference
o
v
er
dif
ferent
channels
is
a
continuous
random
v
ariable,
the
problem
instance
as
described
abo
v
e
has
an
infinite
number
of
scenarios.
Therefore,
a
solution
with
a
deterministic
equi
v
alent
is
not
possible.
Ho
we
v
er
,
by
noting
that,
the
interference
conditions
measured
at
a
certain
channel
are
highly
correlated.
Thus,
the
K
most
recent
observ
ed
interference
scenarios
are
considered
to
find
a
suboptimal
solu-
tion.
T
o
account
for
the
dynamic
(random)
changes
in
the
interference
conditions
,
our
optimization
program
is
an
adapti
v
e
algorithm
that
is
to
be
periodically
conducted
o
v
er
time
(Access
windo
w).
No
w
,
gi
v
en
the
K-entry
interference
table
and
considering
the
constrained
listed
abo
v
e,
the
deterministic
equi
v
alent
for
one
scenario
!
(
!
is
one
realization)
can
be
formulated
as
follo
ws:
Dynamic
r
esour
ce
allocation
for
opportunistic...
(Sharhabeel
H.
Alnabelsi)
Evaluation Warning : The document was created with Spire.PDF for Python.
3858
r
ISSN:
2088-8708
(4)
max
M
X
i
=1
N
X
j
=1
(
i
)
j
!
M
X
i
=1
N
X
j
=1
(
i
)
j
!
P
(
i
)
j
!
s:t:
M
X
i
=1
X
i
B
X
j
(
i
)
j
!
1
;
i
2
I
X
i
(
i
)
j
!
1
;
j
2
J
X
i
X
j
(
i
)
j
!
M
(
i
)
j
!
2
f
0
;
1
g
;
i
2
I
;
j
2
J
X
i
log
2
1
+
P
(
i
)
j
!
P
(
i
)
I
!
+
P
(
i
)
th
!
d
!
j
(
(
i
)
j
1)
;
8
i
2
I
;
8
j
2
J
X
i
0
;
i
2
I
(5)
4.
HIST
ORICAL
SAMPLING/
A
CCESS
WINDO
W
At
the
be
ginning
of
an
A
W
and
gi
v
en
the
interference
or
demand
conditions
o
v
er
the
pre
vious
A
W
,
the
maximum
bandwidth,
which
can
be
used
by
CRN,
i
s
allocated
to
the
v
arious
channels,
this
conducted
in
the
first
stage.
This
can
be
achie
v
ed
by
solving
the
deterministic
equi
v
alent
for
the
K-historical
samples.
In
the
second
stage,
where
the
interference
and
rate
demands
are
realized,
we
allocate/select
channels/po
wers
to
di
f
ferent
CR
users
such
as
the
total
po
wer
consumption
is
minimized.
During
the
current
A
W
time,
the
interference
conditions
and
rate
demands
are
recorded.
Then
the
abo
v
e
process
is
repeated
o
v
er
a
nd
o
v
er
for
e
v
ery
A
W
time.
T
o
illustrate
this
mechanism,
we
consider
a
CRN
scenario
as
sho
wn
in
Figure
1,
where
6
CR
pairs
content
to
access
3
dif
ferent
channels.
Figure
2
sho
ws
the
associated
timing
diagram
for
decisions
or
stages
of
our
optimization
problem.
1
2
3
4
5
6
Figure
1.
Cogniti
v
e
radios
links
Update B
andwidth
Allocation
(1
st
Stage)
Update B
andwidth
Allocation
(1
st
Stage)
Multiple Channel
Assignment
(2
nd
Stage)
Multiple Channel
Assignment
(2
nd
Stage)
t
Figure
2.
Optimization
timing
diagram
5.
NUMERICAL
RESUL
TS
W
e
illustrate
the
pre
viously
discussed
optimization
process
with
a
numerical
e
xample.
W
e
compare
the
performance
of
our
proposed
scheme
to
that
of
traditional
schemes
such
as
the
static
allocation
[1],
weighted
a
v
erage
schemes
[10]
and
optimal
solution.
The
static
assignment
is
based
on
pro
viding
a
fix
ed-bandwidth
per
channel
irrespecti
v
e
of
the
user’
s
demand.
The
weighted
a
v
erage
attempt
at
pro
viding
v
ariable
bandwidth
depends
on
the
a
v
erage
users’
demand,
rather
than
the
actual
demand.
The
optimal
solution
is
f
o
und
using
a
brute-force
method
that
requires
an
e
xhausti
v
e
search
o
v
er
a
lar
ge
state
space
that
increases
e
xponentially
with
number
of
channel
and
number
of
CR
users.
W
e
consi
d
e
r
3
primary
users
netw
orks
(
M
=
3
)
and
4
CRN
links.
Suppose
that
the
A
W
consists
of
4
periods,
K
=
4
.
W
e
set
=
3
,
P
(
i
)
th
=
0
:
001
;
8
i
,
and
B
=
30
Mbps.
Int
J
Elec
&
Comp
Eng,
V
ol.
10,
No.
4,
August
2020
:
3854
–
3861
Evaluation Warning : The document was created with Spire.PDF for Python.
Int
J
Elec
&
Comp
Eng
ISSN:
2088-8708
r
3859
At
a
be
ginning
of
an
A
W
,
assume
that
the
recorded
interference
!
e
P
(
i
)
I
:
i
=
1
;
2
;
3
and
the
rate
demand
d
j
:
j
=
1
;
2
;
3
are
gi
v
en
by:
e
P
(1)
I
=
f
0
:
25
;
0
:
1
;
0
:
15
;
0
:
25
g
,
and
e
d
j
=
f
5
;
10
;
11
;
12
g
.
e
P
(2)
I
=
f
0
:
5
;
0
:
45
;
0
:
35
;
0
:
35
g
,
and
e
d
j
=
f
5
;
6
;
8
;
7
g
.
e
P
(3)
I
=
f
0
:
3
;
0
:
2
;
0
:
15
;
0
:
15
g
,
and
e
d
j
=
f
10
;
10
;
10
;
10
g
.
Also
assume
that
the
interference
o
v
er
the
ne
xt
A
W
is
gi
v
en
by:
e
P
(1)
I
=
f
0
:
3
;
0
:
15
;
0
:
2
;
0
:
29
g
,
and
e
d
j
=
f
8
;
7
;
5
;
10
g
.
e
P
(2)
I
=
f
0
:
48
;
0
:
46
;
0
:
33
;
0
:
31
g
,
and
e
d
j
=
f
6
;
8
;
5
;
5
g
.
e
P
(3)
I
=
f
0
:
27
;
0
:
24
;
0
:
11
;
0
:
25
g
,
and
e
d
j
=
f
7
;
5
;
10
;
8
g
.
The
reported
results
are
a
v
eraged
o
v
er
100
e
xperiments.
Figure
3
sho
ws
the
details
of
the
tw
o
stages
of
the
proposed
channel
optimization
process.
The
outcome
of
this
process
is
sho
wn
in
Figure
4.
This
figure
sho
ws
that
our
stochastic
scheme
significantly
impro
v
es
netw
ork
throughput.
This
impro
v
ement
is
attrib
uted
to
the
the
proper
bandwidth/channel
assignment
algorithm.
A
l
l
o
cat
e
b
an
d
w
i
d
t
h
fo
r
ch
an
n
el
s
1
,
2
,
an
d
3
A
s
s
i
g
n
ch
an
n
el
s
t
o
CR
u
s
er
s
1
,
2
,
an
d
3
Figure
3.
Example
that
illustrates
the
optimization
process
in
a
dynamic
CRN.
0
1
2
3
4
5
6
7
8
9
1
0
1
1
1
2
1
2
3
Th
r
ou
gh
p
ut
(
p
ac
k
e
t
s/A
W)
T
im
e
in
t
e
r
m
s
of
A
W
S
tati
c
A
l
l
o
c
a
ti
o
n
W
e
i
g
h
ted
Av
e
rage
S
to
c
h
a
st
i
c
S
c
h
e
me
Op
ti
m
a
l
B
o
u
n
d
Figure
4.
Comparison
of
dif
ferent
allocation
schemes.
6.
CONCLUSIONS
In
this
paper
,
we
propose
a
no
v
el
stochastic
bandwidth/channel/po
wer
allocation.
Our
proposed
scheme
maximizes
the
CRN
throughput
through
a
proper
bandwidth/channel
allocation
process
while
in
the
same
time
minimizes
the
total
po
wer
consumption.
W
e
proposed
a
tw
o-stage
stochastic
bandwidth
and
channel
assignment
scheme
that
dynamically
e
xploits
the
correlation
between
the
interference
conditions
and
the
rate
demands
to
maximize
the
o
v
erall
netw
ork
throughput.
Compared
to
traditional
bandwidth/channel
allocation
schemes,
numerical
results
sho
wed
that
our
proposed
scheme
re
v
eals
significant
performance
impro
v
ement
in
the
o
v
erall
achie
v
ed
netw
ork
throughput.
REFERENCES
[1]
H.
Ban
y
Salameh,
M.
Krunz,
O.
Y
ounis,
“Cooperati
v
e
adapti
v
e
spectrum
sharing
in
cogniti
v
e
radio
net-
w
orks,
”
IEEE/A
CM
T
r
ans.
on
Networking(T
ON)
,
v
ol.18,
no.4,
pp.
1181-1194,
2010.
[2]
M.
Bani
Hani,
H.
Ban
y
Salameh,
Y
.
Jararweh,
and
A.
Bousselham,
”T
raf
fic-a
w
are
self-coe
xistence
man-
agement
in
IEEE
802.22
WRAN
systems,
”
In
2013
7th
IEEE
GCC
Confer
ence
and
Exhibition
,
pp.
507-
510,
2013.
[3]
A.
Ranjan
and
G.
Somani,
”Access
control
and
authentication
in
the
internet
of
things
en
vironment,
”
Computer
Communications
and
Networks
,
pp.
283-305,
2016.
[4]
H.
Al-Mahdi,
F
.
Y
asser
,
”Design
and
analysis
of
routing
protocol
for
cognit
i
v
e
radio
ad
hoc
netw
orks
in
Heterogeneous
En
vironment,
”
International
J
ournal
of
Electrical
and
Computer
Engineering
(IJECE)
,
v
ol.
9,
no.
1,
pp.
341-351,
2019.
[5]
A.
Khan,
M.
Rehmani,
and
A.
Rachedi,
”Cogniti
v
e-radio-based
internet
of
things:
Applications,
architec-
tures,
spectrum
related
functionalities,
and
future
research
directions”,
IEEE
W
ir
eless
Communications
,
V
ol.
24,
No.3,
pp.
17-25,
2017.
Dynamic
r
esour
ce
allocation
for
opportunistic...
(Sharhabeel
H.
Alnabelsi)
Evaluation Warning : The document was created with Spire.PDF for Python.
3860
r
ISSN:
2088-8708
[6]
H.
Ban
y
Salameh,
S.
Almajali,
M.
A
yyash,
and
H.
Elg
ala,
”Security-a
w
are
channel
assignment
in
IoT-
based
cogniti
v
e
radio
netw
orks
for
time-critical
applications,
”
F
ourth
International
Confer
ence
on
Soft-
war
e
Defined
Systems
(SDS)
,
pp.
43-47,
V
alencia,
2017.
[7]
S.
Razmi,
N.
P
arhizg
ar
,
”Adapti
v
e
resources
assignment
in
OFDM-based
cogniti
v
e
radio
systems,
”
Inter
-
national
J
ournal
of
Electrical
and
Computer
Engineering
(IJECE)
,
v
ol.
9,
no.
3,
pp.
1935-1943,
2019.
[8]
P
.
V
arade,
A.
W
abale,
R.
Y
erram,
and
R.
Jaisw
al,
”Throughput
Maximization
of
Cogniti
v
e
Radio
Multi
Relay
Netw
ork
wit
h
Interference
Management,
”
International
J
ournal
of
Electrical
and
Computer
Engi-
neering
(IJECE)
,
v
ol.
8,
no.
4,
pp.
2230-2238,
2018.
[9]
S.
Lakhal,
Z.
Guennoun,
”Equity-based
free
channels
assignment
for
secondary
us
ers
in
a
cogniti
v
e
radio
netw
ork,
”
International
J
ournal
of
Electrical
and
Computer
Engineering
(IJECE)
,
v
ol.
9,
no.
3,
pp.
2057-
2063,
2019.
[10]
H.
Ban
y
Salame
h
,
”Ef
ficient
Resource
Allocation
for
Multi-cell
Heterogeneous
Cogniti
v
e
Netw
orks
W
ith
V
arying
Spectrum
A
v
ailability
,
”
IEEE
T
r
ansactions
on
V
ehicular
T
ec
hnolo
gy
,
v
ol.
65,
no.
8,
pp.
6628-
6635,
2016.
[11]
H.
Ban
y
Salameh,
H.
Kasasbeh
and
B.
Harb,
”A
Batch-Based
MA
C
Design
W
ith
Simultaneous
As-
signment
Decisions
for
Impro
v
ed
Throughput
in
Guard-Band-Constrained
Cogniti
v
e
Netw
orks,
”
IEEE
T
r
ansactions
on
Communications
,
v
ol.
64,
no.
3,
pp.
1143-1152,
March
2016.
[12]
A.
Doulat,
A.
Al
Abed
Al
Aziz,
M.
Al-A
yyoub,
Y
.
Jararweh,
H.
Ban
y
Salameh
and
A.
A.
Khreishah,
”Softw
are
defined
frame
w
ork
for
multi-cell
Cogniti
v
e
Radio
Netw
orks,
”
IEEE
10th
Inter
na
t
ional
Con-
fer
ence
on
W
ir
eless
and
Mobile
Computing
,
Networking
and
Communications
(W
iMob)
,
Larnaca,
pp.
513-518,
2014.
[13]
H.
Ban
y
Salameh,
O.
Badarneh,
”Opportunistic
medium
access
control
for
maximizing
pack
et
deli
v
ery
rate
in
dynamic
access
netw
orks,
”
J
ournal
of
Network
and
Computer
Applications
,
V
olume
36,
Issue
1,
pp.
523-532,
2013.
[14]
M.
Labib,
S.
Ha,
W
.
Saad,
and
J.
H.
Reed,
”A
colonel
blotto
g
ame
for
anti-jamming
in
the
internet
of
things,
”
in
the
IEEE
Global
Communications
Confer
ence
(GLOBECOM)
,
pp.
1-6,
Dec
2015.
[15]
N.
Namv
ar
,
W
.
Saad,
N.
Bahadori,
and
B.
K
elle
y
,
”Jamming
in
the
internet
of
things:
A
g
ame-theoretic
perspecti
v
e,
”
in
IEEE
Global
Communications
Confer
ence
(GLOBECOM)
,
Dec
2016,
pp.
1-6.
[16]
H.
Ban
y
Salameh,
H.
Kasasbeh
and
B.
Harb,
“
An
Opportunistic
Guard-band-a
w
are
Channel
Assignment:
A
batch-based
Approach”,
Pr
oc.
of
the
IEEE
WCNC’16
Confer
ence
,
Qatar
,
pp.
1-6,
April
2016.
[17]
G.
S.
Uyanik,
M.
J.
Abdel-Rahman,
and
M.
Krunz,
”Optimal
guardband-a
w
are
channel
assignment
with
bonding
and
aggre
g
ation
in
multi-channel
systems,
”
in
IEEE
Pr
oceedings
of
GLOBECOM
Confer
ence
,
pp.
4769-4774,
Dec.
2013.
[18]
M.
A.
Rahman,
M.
Krunz,
“Stochastic
guard-band-a
w
are
channel
assignment
with
bonding
and
aggre
g
a-
tion
for
DSA
netw
orks,
”
IEEE
T
r
ans.
W
ir
eless
Commun.
,
v
ol.
14,
no.
7,
pp.
3888-3898,
Jul.
2015.
[19]
M.
I
.
B.
Shahi
d,
J.
Kamruzzaman,
M.
R.
Hassan,
“Modeling
multi-user
spect
rum
allocation
for
cogniti
v
e
radio
netw
orks,
”
Comput.
Electr
.
Eng
.
,
v
ol.
52,
pp.
266-283,
May
2016.
[20]
S.
Qureshi,
S.
Ahmad,
A.
Ikram,
N.
Hasan.
”Joint
ener
gy
and
throughput
based
multi-channel
assignment
in
cogniti
v
e
radio
sensor
netw
ork,
”
In
2016
IEEE
3r
d
International
Symposium
on
T
elecommunication
T
ec
hnolo
gies
(ISTT)
,
pp.
65-69,
2016.
[21]
Z.-H.
W
ei
and
B.-J.
Hu,
”A
f
air
multi-channel
assignment
algorithm
with
practical
implementation
in
distrib
uted
cogniti
v
e
radio
netw
orks,
”
IEEE
Access
,
v
ol.
6,
pp.
14255–14267,
2018.
[22]
H.
Ban
y
Salameh,
M.
Krunz,
D.
Manzi,
“Spectrum
Bonding
and
Aggre
g
ation
with
Guard-band-a
w
areness
in
Cogniti
v
e
Radio
Netw
orks,
”
IEEE
T
r
ansaction
on
Mobile
Computing
,
v
ol.13,
no.3,
pp.569-581,
March
2014.
[23]
J.
Chen,
S.
Ping,
J.
Jia,
Y
.
Deng,
M.
Dohler
,
and
H.
Aghv
ami,
“Cross-layer
optimization
for
spectrum
aggre
g
ation-based
cogniti
v
e
radio
ad-hoc
netw
orks,
”
in
Pr
oc.
IEEE
GLOBECOM
,
pp.
1-6,
Dec.
2017.
[24]
M.
Ghorbel,
B.
Hamdaoui,
R.
Hamdi,
M.
Guizani,
and
M.
NoroozOliaee,
“Distrib
uted
dynamic
spec-
trum
access
with
adapti
v
e
po
wer
allocation:
Ener
gy
ef
ficienc
y
and
cross
layer
a
w
areness,
”
in
Pr
oc.
IEEE
INFOCOM
,
pp.
694-699,
Apr
.
2014.
[25]
L.
Lei,
and
C.
Chig
an,
”A
V
irtual
MIMO
based
anti-jamming
strate
gy
for
cogniti
v
e
radio
netw
orks,
”
In
2016
IEEE
International
Confer
ence
on
Communications
(ICC)
,
pp.
1-6,
2016.
Int
J
Elec
&
Comp
Eng,
V
ol.
10,
No.
4,
August
2020
:
3854
–
3861
Evaluation Warning : The document was created with Spire.PDF for Python.
Int
J
Elec
&
Comp
Eng
ISSN:
2088-8708
r
3861
[26]
G.
Han,
L.
Xiao,
and
H.
Poor
,
”T
w
o-dimensional
anti-jamming
communication
based
on
deep
reinforce-
ment
learning,
”
In
2017
IEEE
International
Confer
ence
on
Acoustics,
Speec
h
and
Signal
Pr
ocessing
(ICASSP)
,
pp.
2087-2091,
2017.
[27]
A.
Ralhan,
R.
Y
ada
v
,
and
R.
Misra,
”Minimum
Interference
T
opology
Control
in
Cogniti
v
e
Radio
Net-
w
orks
through
Channel
Assignment,
”
In
2018
International
Confer
ence
on
Advances
in
Computing
,
Com-
munications
and
Informatics
(ICA
CCI)
,
pp.
1418-1423,
2018.
[28]
R.
Irwin,
A.
MacK
enzie,
L.
DaSilv
a,
”Resource-minimized
channel
assignment
for
multi-transcei
v
er
cog-
niti
v
e
radio
netw
orks,
”
IEEE
J
ournal
on
Selected
Ar
eas
in
Communications
,
v
ol.
31,
no.
3,
pp.
442-450,
2013.
[29]
P
.
Tseng,
W
.
Chung,
P
.
Hsiu,
”Minimum
interference
topology
construction
for
rob
ust
multi-hop
cogniti
v
e
radio
netw
orks,
”
In
2013
IEEE
W
ir
eless
Communications
and
Networking
Confer
ence
(WCNC)
,
pp.
101-
105,
2013.
[30]
S.
H.
Alnabelsi
and
A.
E.
Kamal,
”Res
ilient
multicast
routing
in
CRNs
using
a
multilayer
h
yper
-graph
approach,
”
IEEE
International
Confer
ence
on
Communications
(ICC)
,
Budapest,
pp.
2910-2915,
2013.
[31]
S.
H.
Alnabelsi,
”Finding
an
Immuned
P
ath
ag
ainst
Single
Primary
User
Acti
vity
in
Cogniti
v
e
Radio
Net-
w
orks,
”
International
J
ournal
on
Communications
Antenna
and
Pr
opa
gation
(IRECAP)
,
Praise
W
orth
y
Prize,
Italy
,
2017.
[32]
S.
H.
Alnabelsi,
and
A.
E.
Kamal,
”Interference-based
pack
et
reco
v
ery
for
ener
gy
sa
ving
in
Cogniti
v
e
Radio
Netw
orks,
”
IEEE
International
Confer
ence
on
Communications
(ICC)
,
Otta
w
a,
pp.
5978-5982,
2012.
BIOGRAPHIES
OF
A
UTHORS
Dr
.
Sharhabeel
H.
Alnabelsi
is
an
associate
professor
at
Computer
and
Netw
orks
Eng.
Dept.
at
Al-Balqa
Applied
Uni
v
ersity
,
Jordan.
He
is
also
with
the
computer
Eng.
Dept.
at
Al
Ain
Uni
v
ersity
,
U
AE.
He
recei
v
ed
his
Ph.D.
in
computer
engineering
from
Io
w
a
State
Uni
v
ersity
,
USA,
2012.
He
recei
v
ed
his
M.Sc.
in
computer
engineeri
ng
from
The
Uni
v
ersity
of
Alabama
in
Huntsville,
USA,
2007.
His
research
interests
include
Cogniti
v
e
Radio
Netw
orks,
W
ireless
Sensor
Netw
orks.
Email:
alnabsh1@bau.edu.jo,
sharhabeel.alnabelsi@aau.ac.ae
Pr
of
.
Haythem
A.
Bany
Salameh
is
a
Professor
of
Netw
orks
Communication
Engineering
with
Al
Ain
Uni
v
ersity
,
U
AE.
He
recei
v
ed
the
Ph.D.
de
gree
in
electrical
and
computer
engineering
from
the
Uni
v
ersity
of
Arizona,
USA,
2009.
He
is
also
i
n
a
sabbatical
lea
v
e
from
Y
armouk
Uni
v
ersity
,
Jordan.
In
August
2009,
he
joi
ned
YU,
after
a
brief
postdoctoral
position
with
the
Uni
v
ersity
of
Arizona.
His
research
interests
include
optical
communication
technology
and
wireless
netw
orking.
In
the
summer
of
2008,
he
w
as
a
member
of
the
R&D
Long-T
erm
Ev
olution
De
v
elopment
Group,
Q
U
AL-
COMM,
Inc.,
San
Die
go,
CA,
USA.
He
is
an
IEEE
Senior
Member
class
of
2016.
Email:
haythem@yu.edu.jo,
haythem.ban
ysalameh@aau.ac.ae,
Dr
.
Zaid
M.
Albataineh
is
an
associate
professor
at
Y
armouk
Uni
v
ersity
,
Jordan.
He
recei
v
ed
his
Ph.D.
in
Electrical
and
Computer
Eng.
from
Michig
an
State
Uni
v
ersity
,
USA,
2014,
and
his
M.Sc.
de
gree
in
the
communication
and
electronic
engineering
from
the
Jordan
Uni
v
ersity
of
Science
and
T
echnology
,
Jordan,
2009.
His
research
interests
include
Blind
Source
Separation,
Independent
Com-
ponent
analysis,
Nonne
g
at
i
v
e
matrix
F
actorization,
W
ireless
Com
munication,
DSP
Implementation,
VLSI,
Analog
Inte
grated
Circuit
and
RF
Inte
grated
Circuit.
Email:
zaid.bataineh@yu.edu.jo
Dynamic
r
esour
ce
allocation
for
opportunistic...
(Sharhabeel
H.
Alnabelsi)
Evaluation Warning : The document was created with Spire.PDF for Python.