Int
ern
at
i
onal
Journ
al of Ele
ctrical
an
d
Co
mput
er
En
gin
eeri
ng
(IJ
E
C
E)
Vo
l.
8
, No
.
6
,
Decem
ber
201
8
, p
p.
5195
~
5202
IS
S
N: 20
88
-
8708
,
DOI: 10
.11
591/
ij
ece
.
v8
i
6
.
pp
5195
-
52
02
5195
Journ
al h
om
e
page
:
http:
//
ia
es
core
.c
om/
journa
ls
/i
ndex.
ph
p/IJECE
Ingeni
ous
M
ethod
for
Co
nd
ucive
Handoff
Appl
iance
in
Cognitiv
e Radio
Network
s
J.
Josephin
e
Dh
iv
ya
,
M.
R
am
as
w
ami
Depa
rtment
o
f
C
om
pute
r
Applica
ti
ons,
Madur
ai Kamaraj
Univ
ersity
,
Indi
a
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Feb
6
, 2
01
8
Re
vised
Ju
l
2
3
,
201
8
Accepte
d
Aug 6
, 2
01
8
W
ire
le
ss
comm
unic
a
ti
ons
dep
lo
y
ed
in
the
cur
r
ent
epoc
h
claims
ce
ase
le
ss
conne
c
ti
on
among
it
s
users
the
reb
y
l
ea
d
ing
to
the
inve
stigation
of
Cognit
ive
Radi
o
Networks
(CRN)
which
e
nabl
es
to
m
ake
use
of
unal
lo
ca
t
ed
spec
trum
opti
m
al
l
y
and
p
rovide
s
uninterr
upte
d
conne
c
ti
o
n.
Est
abl
ishing
i
nte
rm
ina
bl
e
conne
c
ti
vi
t
y
dur
ing
the
handof
f
proc
ess
in
spe
ctrum
m
obil
ity
of
CRN
is
a
cha
l
le
nging
ta
sk
.
Thi
s
p
ape
r
el
u
ci
da
te
s
th
e
optim
iz
at
ion
of
han
doff
proc
ess
ca
rri
ed
out
in
CRN
b
y
in
cor
p
ora
ti
ng
an
in
te
l
l
ige
nt
m
et
hod
.
T
his
inc
lud
es
fuz
z
y
logic
whe
rei
n
the
h
andof
f
par
amete
rs
ar
e
p
roc
essed
the
r
eb
y
indi
cating
the
ne
ed
of
hand
off.
The
prof
fer
e
d
m
et
hod
al
so c
o
m
prises
of
a
par
t
of
gene
t
ic
al
gorit
hm
whi
ch
y
ie
lds
f
it
ness
v
al
ue
for
red
u
ci
n
g
the
h
andof
f
o
cc
urre
n
ce
s
and
enha
n
ci
ng
the
over
all
per
f
orm
anc
e
of
th
e
s
y
stem
is
pro
m
ote
d
using
cuc
koo
sea
r
ch
which
dec
id
es
the
m
obil
e
nod
e
from
which
the
handof
f
proc
ess
has
to
in
it
iate
base
d
on
t
he
prior
ity
g
ene
r
at
ed
.
Th
is
t
ec
hni
que
ensure
s
tha
t
d
ec
ision
is
t
ake
n
ah
ea
d
of
link
fai
lur
e
ra
the
r
tha
n
ran
g
e
fa
il
ur
e
which
ar
e
the
ke
y
poin
t
in
compari
son
to
th
e
exi
sting
s
y
ste
m
.
Result
s
obta
i
ned
through
the
sim
ula
t
ion
a
r
e
sati
sf
ac
tor
y
in te
rm
s of
dela
y
,
t
hroughput,
num
ber
of
f
ai
l
ed
handof
f
and
handof
fs
per
form
ed
in
compari
son
to
the
exi
sting
f
uzzy
b
ase
d
handof
f
pro
ce
ss
in
CRN.
Ke
yw
or
d:
Fit
ness value
Hand
off
Sp
ect
r
um
m
ob
il
it
y
Copyright
©
201
8
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights
reserv
ed
.
Corres
pond
in
g
Aut
h
or
:
J.
J
os
ep
hi
ne D
hiv
ya
,
Dep
a
rtm
ent o
f C
om
pu
te
r
A
pp
li
cat
ion
s,
Ma
durai Kam
araj
U
niv
e
rsity
,
Ma
durai
–
62
5021,
In
dia.
Em
a
il
:
j
os
ed
hi
vya@
gm
ail.co
m
1.
INTROD
U
CTION
The
m
ai
n
aim
of
CR
N
is
fa
ci
li
ta
t
ing
plian
t
wireless
com
m
un
ic
at
ion
by
util
iz
ing
the
avail
able
or
un
al
locat
e
d
spe
ct
ru
m
[1
]
.
Th
is
is
carried
out
by
al
lowing
the
un
li
cen
se
d
or
seco
ndar
y
us
er
to
acce
ss
the
sp
ect
r
um
tem
p
or
a
rily
wh
e
rin
the
pri
m
ary
u
ser
is
not
de
pl
oyed
[2
]
.
On
t
he
ar
rival
of
t
he
pri
m
ary
us
er
it
is
m
and
at
ory
for
the
un
li
ce
ns
e
d
us
e
r
to
vac
at
e
the
acce
ss
on
s
pectr
um
there
by
switc
hi
ng
ov
e
r
to
a
no
t
her
unocc
up
ie
d
ch
ann
el
to
pre
ve
nt
them
fr
om
har
m
fu
l
interfe
ren
ce
[
3].T
his
process
of
cha
ng
e
ov
e
r
of
sta
te
or
qu
it
ti
ng
the
c
ha
nn
el
is
ref
e
rred
to
as
ha
ndoff
[4
]
.
It
is
ess
en
ti
al
to
note
that
the
hand
off
proce
ss
s
houl
d
ta
ke
place
in
a
swi
ft
m
ann
er
without
any
hi
ndran
ce
to
e
nsur
e
op
ti
m
al
sp
ect
ru
m
han
dlin
g
that
is
achieved
by
analy
zi
ng
ha
nd
off
pe
rfor
m
ance
ind
ic
at
or
s
[5
]
.
On
e
good
so
l
ution
for
this
would
be
the
cho
ic
e
of
deci
din
g
th
e
appr
opriat
e
ch
ann
el
t
o
switc
h
ove
r
an
d
th
e
instant
at
w
hich
this
act
i
on
occ
urs;
al
so
choosi
ng
a
pe
rtinent
hand
off
sc
hem
e is m
and
at
or
y
[6
]
.
Am
on
g
al
l
the
avail
able
ty
pes
of
ha
ndoff
m
echan
ism
,
hybri
d
i
ntell
igent
ha
ndoff
m
et
ho
d
s
hows
to
be
qu
it
e
com
pro
m
i
sing
am
on
g
the
reacti
ve
a
nd
no
n
-
reacti
ve
hand
off
strat
eg
ie
s
[7
]
.
T
he
propose
d
m
et
ho
do
l
og
y
is on
e
of
t
he
hy
br
id m
et
ho
ds
wh
ic
h
inc
orp
orat
es f
uzzy ba
se
d
ge
netic
alg
ori
th
m
to
decide
the ch
a
nn
el
whic
h
is
al
read
y
avail
a
ble
by
pr
e
def
i
ned
r
ules.
F
ur
t
her
m
or
e
the
ha
ndoff
process
is
carrie
d
out
in
a
sm
oo
th
way
by
exten
ding
the
fitness
val
ue
obta
ined
t
hro
ugh
f
uzzy
ge
netic
al
go
rit
hm
an
d
us
i
ng
it
in
t
he
cuc
koo
sea
rch
[8]
op
ti
m
iz
ation
al
gorithm
.
The
r
est
of
this
pa
pe
r
descr
i
bes
t
he
m
e
thodo
l
og
y
of
f
uzzy
base
d
gen
et
ic
al
gor
it
h
m
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
8
, N
o.
6
,
Dece
m
ber
2
01
8
:
5195
-
5202
5196
coupled
with
cucko
o
searc
h
fo
r
achie
ving
qu
al
it
y
han
dof
f
proce
dure
al
ong
with
th
e
si
m
ulati
on
a
nd
th
e
analy
sis o
f han
doff p
e
rfo
rm
ance ind
ic
at
or
s
.
2.
RE
LATE
D
W
ORK
The
hybr
i
d
m
e
thodo
l
og
y of
fuzzy
with g
ene
ti
c
al
go
rithm
,
t
hough
fairly
ex
plored
c
onside
rab
le
works
can
be
f
ound
i
n
the
li
te
ratu
re.
In
W
a
nm
ai
and
Mi
ngch
ua
n
a
ver
ti
cal
ha
ndoff
decisi
on
al
gorithm
was
pr
opose
d
to
achieve
op
ti
m
u
m
han
dof
f
perform
ance
in
heteroge
neous
networ
ks
[
9]
.
The
aut
hors
in
[10]
pro
po
se
d
a
new
al
gorithm
fo
r
ha
ndoff
opti
m
izati
on
in
co
gn
it
ive
ra
dio
net
work
s b
y
segr
e
gat
ing
the W
R
AN
into
m
ulti
ple
c
el
ls.
An
al
gorithm
f
or
m
aking
deci
sion
base
d
on
m
ul
ti
ple
crit
eria
was
pr
opos
e
d
in
[
11]
f
or
c
hannel
sel
ect
io
n
a
nd
sp
ect
r
um
decisi
on
functi
on.
The
c
oncept
of
optim
iz
at
i
on
i
n
c
ogniti
ve
ra
dio
wit
h
cuc
koo
sea
rc
h
wa
s
introd
uced in
[
12
]
wh
ic
h fo
c
us o
n
e
ff
ic
ie
nt s
pectr
um
sen
sin
g
te
ch
nique.
The
disti
nct
ne
ss
of
our
w
ork
inclu
des
the
progressio
n
of
the
ha
ndoff
process
by
pr
e
dicti
ng
li
nk
fail
ur
e
a
head
and
in
co
rpo
rati
ng
cu
ck
oo
search
f
or
detect
ing
m
ob
il
e
n
odes
to
init
ia
te
the
hand
off
proces
s
tog
et
he
r wit
h f
uzzy ge
netic
s
yst
e
m
w
hich
is
an
i
niti
at
ive m
et
hodo
l
og
y
pro
po
s
ed
in
t
he
li
te
ratur
e
.
3.
P
R
OBL
EM
S
TATE
MENT
The
scarcit
y
of
rad
i
o
wa
ves
le
ads
to
co
ng
e
sti
on
pro
blem
s
in
wireless
com
m
un
ic
at
ion
and
this
is
ov
e
rc
o
m
e
by
t
he
e
vo
l
ution
of
co
gnit
ive
ra
di
o
net
works.T
he
s
pectr
um
m
ob
il
it
y
ph
as
e
involve
d
i
n
co
gn
it
iv
e
cy
cl
e
play
s
a
vital
ro
le
by
ensu
ri
ng
sm
oo
th
ha
ndoff
proc
ess.
The
existi
ng
m
et
hodo
l
og
ie
s
pro
posed
s
o
far
do
e
s
no
t
prov
i
de
se
a
m
le
ss
con
nect
ivit
y
and
cat
er
s
to
th
e
di
ff
e
re
nt
netw
ork
requirem
ents
hen
c
e
the
pro
posed
wor
k
fo
c
us
es
on
devi
sing
a
n
intel
li
gen
t m
et
ho
d w
hich
a
dm
inist
e
rs
a
nd f
ine
tu
ne
s the
hand
off pr
ocess
c
ar
ried
out by
m
aking
decisi
ons a
hea
d of
t
he
ch
a
ng
e
of am
bience of net
w
ork
.
4.
PROP
OSE
D SYSTE
M
The
pro
po
s
e
d
wor
k
c
om
pr
ise
s
of
de
vel
opin
g
a
c
ogniti
ve
m
ob
il
e
te
rm
inal
wh
ic
h
i
s
capa
ble
of
sensing
the
e
nvir
on
m
ent
an
d
est
ablishi
ng
c
onnecti
ons
bas
ed
on
the
a
vaila
bili
ty
of
the
pr
im
ary
us
er
(
PU)
there
by
reso
l
ve
s
the
pro
blem
s
caused
by
switc
hing
an
d
the
hi
gh
int
erf
e
re
nce
rates
occurri
ng
dur
ing
da
ta
transm
issi
on
[
13]
.Th
e
pr
im
ary
par
t
in
ou
r
work
is
us
i
ng
Gen
et
ic
Algorithm
(G
A
)
wi
th
cucko
o
sear
ch
for
hand
off
de
ci
sion
m
akin
g,
th
ereb
y
op
ti
m
izi
ng
f
uzzy
log
ic
m
e
m
ber
sh
ip
f
un
ct
io
ns
a
nd
wh
ic
h
is
ne
w
form
of
hybri
d
a
ppr
oac
h[1
4].
This
w
ork
i
nvolv
e
s
th
e
us
a
ge
of
a
ne
w
m
ulti
crit
eria
kind
of
hybri
d
hand
off
str
at
egy
with
co
gn
it
ive ab
il
it
y
to
switc
h
base
d
on
net
work
c
onditi
on
s
and
a
vaila
bili
ty
of
channel
[
15
]
.
T
he
flo
w
of
ou
r
pro
po
se
d w
o
r
k i
s explai
ne
d b
y
m
eans of
Fi
g
ur
e
1
.
Fig
ure
1
.
Flo
w
of the
prop
os
e
d
m
od
el
The
pro
posed
work
ca
n
be
s
plit
into
t
wo
pa
rts.
T
he
first
par
t
c
om
pr
ise
s
of
dep
l
oying
the
F
uzzy
base
d
Ge
netic
Algorithm
(F
BGA
)
an
d
s
ub
se
qu
e
ntly
incorp
or
at
in
g
cuc
koo
search
f
or
opti
m
iz
ing
the
ou
t
pu
t
of
the
FBG
A.
FB
GA
gi
ves
th
e
f
it
ness
val
ue
th
ereb
y
yi
el
ding
the
sel
ect
ion
of
the
ap
pro
pr
ia
te
channel
to
s
witc
h
ov
e
r
base
d
on
the
pre
fer
e
nce
al
locat
ed.
The
proce
dure
i
nvol
ved
i
n
FBG
A
and
cu
ck
oo
se
arch
are
des
cri
bed
in
the foll
owin
g
s
ect
ion
.
4
.
1.
Fuz
z
y
Bas
ed Gene
tic
Algo
ri
th
m
A
FBG
A
is
a
hybri
d
f
or
m
of
GA
w
hich
m
a
kes
us
e
of
the
fu
zzy
lo
gic
hin
ge
d
proc
ed
ure.
The
m
ai
n
pur
po
se
of
usi
ng
fu
z
zy
logi
c
is
it
s
conv
enient
m
od
e
of
a
ddres
sin
g
pro
blem
s
that
are
unce
rtai
n
a
nd
unpredict
able.
The
pro
po
se
d
m
e
tho
d
inc
orp
or
at
es
f
uzz
y
log
ic
in
or
d
er
to
m
ake
decisi
on
s,
wh
ic
h
is
a
n
i
m
po
rtant
pa
rt
of
the
ha
nd
off
proces
s
a
nd
this
is
ac
hieved
by
m
aki
ng
use
of
f
uz
zy
log
ic
c
ontr
ollers.
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Inge
nious Met
hod
f
or
C
ondu
ci
ve Han
do
ff
A
pp
li
a
nce i
n
C
ogniti
ve R
ad
i
o N
et
works
(
J. J
os
e
ph
i
ne Dh
iv
ya
)
5197
Param
et
ers
that
play
vital
ro
le
in
han
do
ff
process
are
giv
en
as
in
put
to
f
uzzy
infer
e
nce
w
hich
inc
lud
es
m
ob
il
i
ty
,
delay
,
recei
ved
sig
na
l
stren
gth
a
nd
ba
ndwidt
h.
T
hese
i
nput
par
a
m
et
ers
are
the
n
gen
e
rated
to
if
the
n
el
se
ru
le
s
by
m
eans
of
t
he
trapez
oid
al
m
e
m
ber
sh
ip
f
unct
ion
s
resu
lt
in
g
i
n
kn
ow
i
ng
t
he
sta
tus
of
t
he
PU
that
include
s
the
ar
rival
an
d
de
part
ur
e
of
P
U
w
hich
is
ver
y
i
m
po
rta
nt
for
deci
ding
ei
ther
the
hand
off
proce
ss
ha
s
to o
cc
ur
or
not.
Th
e
rules f
or
gen
e
rati
ng
handoff are
prese
nt
ed
in
Table
1.
Table
1
.
R
ules
for Han
dof
f D
eci
sion
Distan
ce
Sp
eed
Delay
Han
d
o
ff
Pr
io
rity
Far
Fast
Low
Ver
y
Hig
h
Far
Av
erage
Low
Ver
y
Hig
h
Far
Slo
w
Low
Hig
h
Mediu
m
Fast
Mediu
m
MedHig
h
Mediu
m
Av
erage
Mediu
m
MedHig
h
Mediu
m
Slo
w
Mediu
m
MedHig
h
Clo
se
Fast
h
ig
h
Med
Clo
se
Av
erage
h
ig
h
MedLo
w
Clo
se
Slo
w
h
ig
h
MedLo
w
Far
Fast
Mediu
m
Ver
y
Hig
h
Far
Av
erage
Mediu
m
Hig
h
Far
Slo
w
Mediu
m
MedHig
h
Mediu
m
Fast
Low
MedHig
h
Mediu
m
Av
erage
Low
Med
Mediu
m
Slo
w
Low
MedLo
w
Clo
se
Fast
Mediu
m
MedLo
w
Clo
se
Av
erage
Mediu
m
Low
Clo
se
Slo
w
Mediu
m
Ver
y
Low
Far
Fast
h
ig
h
MedHig
h
Far
Av
erage
h
ig
h
MedHig
h
Far
Slo
w
h
ig
h
Med
Mediu
m
Fast
h
ig
h
MedLo
w
Mediu
m
Av
erage
h
ig
h
MedLo
w
Mediu
m
Slo
w
h
ig
h
MedLo
w
Clo
se
Fast
Low
Low
Clo
se
Av
erage
Low
Ver
y
Low
Clo
se
Slo
w
Low
Ver
y
lo
w
The
trapez
oida
l
m
e
m
ber
sh
ip
functi
ons
ar
e
us
ed
to
co
m
pu
te
values
for
input
par
a
m
et
es
and
in
gen
e
ral they
a
r
e g
ive
n by
µ
(
)
=
{
1
,
2
<
<
3
0
,
=
1
4
−
1
2
−
1
,
1
<
<
2
−
4
3
−
4
,
3
<
<
4
Fr
om
the
above
eq
uation
th
e
m
e
m
ber
sh
ip
fu
ncti
on
s
f
or
sp
eed,
distan
ce
and
delay
are
com
pu
te
d
separ
at
el
y.
S
pe
ed
m
e
m
ber
sh
ip
Functi
ons:
µ
(
)
=
{
1
,
0
<
<
5
0
,
=
0
10
−
0
5
−
0
,
0
<
<
5
−
10
7
−
10
,
7
<
<
10
µ
(
)
=
{
1
,
7
<
<
10
0
,
=
5
=
15
−
7
10
−
7
,
7
<
<
10
−
15
12
−
15
,
12
<
<
15
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
8
, N
o.
6
,
Dece
m
ber
2
01
8
:
5195
-
5202
5198
µ
ℎ
ℎ
(
)
=
{
1
,
15
<
<
20
0
,
=
12
25
−
15
12
−
15
,
12
<
<
15
−
25
20
−
25
,
20
<
<
25
Delay
m
e
m
ber
sh
ip
fun
ct
io
ns
:
µ
(
)
=
{
1
,
0
.
1
<
<
0
.
2
0
,
=
0
0
.
25
−
0
0
.
1
−
0
,
0
<
<
0
.
1
−
0
.
2
0
.
25
−
0
.
2
,
0
.
2
<
<
0
.
25
µ
(
)
=
{
1
,
0
.
3
<
<
0
.
4
0
,
=
0
.
2
0
.
5
−
0
.
2
0
.
3
−
0
.
2
,
0
.
2
<
<
0
.
3
−
0
.
5
0
.
4
−
0
.
5
,
0
.
4
<
<
0
.
5
µ
ℎ
ℎ
(
)
=
{
1
,
0
.
6
<
<
0
.
8
0
,
=
0
.
55
1
.
0
−
0
.
55
0
.
6
−
0
.
55
,
0
.
55
<
<
0
.
6
−
1
.
0
0
.
8
−
1
.
0
,
0
.
8
<
<
1
.
0
Dista
nce m
e
m
ber
s
hip f
unct
io
ns
:
µ
(
)
=
{
1
,
2
<
<
15
0
,
=
1
20
−
15
15
−
1
,
1
<
<
15
−
20
15
−
20
,
15
<
<
20
µ
(
)
=
{
1
,
12
<
<
50
0
,
=
10
60
−
10
12
−
10
,
10
<
<
12
−
60
50
−
60
,
50
<
<
60
µ
ℎ
ℎ
(
)
=
{
1
,
60
<
<
80
0
,
=
55
100
−
55
60
−
55
,
55
<
<
60
−
100
80
−
10
0
,
80
<
<
100
The
i
nf
e
ren
ce
ob
ta
ine
d
by
fuzzy
log
ic
is
part
ia
l,
no
t
acc
ur
a
te
at
al
l
conditi
on
s
an
d
t
her
e
is
a
nee
d
f
or
GA
to p
r
oduc
e
ap
pro
xim
a
te
resu
lt
s
w
hich
in
cl
ud
es
avail
abi
li
ty
of
cha
nnel
w
he
n
PU
an
d
s
econda
ry
us
e
r
m
ov
e
acro
s
s
the
net
work
t
her
e
by
enh
a
ncin
g
the
qual
it
y
of
ha
ndoff
proce
dure
.
The
G
A
ge
ts
te
rm
inate
d
by
i
te
rati
ng
the
al
gor
it
hm
ti
ll
the
su
it
able
fitness
value
(F
V
)
is
ge
ne
ra
te
d.
FV
in
t
he
pro
pose
d
wor
k
ref
e
rs
t
o
t
he
node
pr
e
dicta
bili
ty
wh
ic
h
sp
eci
fie
s
the
values
of
inp
ut
pa
ram
eter
s
associat
e
d
with
each
a
nd
ever
y
sin
gle
m
ob
il
e
node
i.e.
recei
ved
si
gn
al
stre
ng
t
h,
delay
et
c.,
in
dicat
in
g
th
e
m
ob
il
it
y
of
t
he
node
s
co
nnect
ed
to
the
ne
twork
.
The
ha
ndoff
pri
or
it
y
is
gen
e
rated
by
m
eans
o
f
the
tr
uth
ta
ble
pr
ese
nte
d
in
T
able
1.
The
val
ue
Dis
tan
ce
represe
nts
the
d
ist
ance f
r
om
m
ob
il
e
nodes
t
o
base s
ta
ti
on
,
Sp
ee
d
re
fer
s
to
the
m
ob
il
it
y
of n
odes
an
d
De
lay
are
ta
ken
into
co
nsi
der
at
io
n.
The
hand
off
decisi
on
is
ta
ke
n
bas
ed
on
the
tr
uth
ta
ble
(ta
ble
1)
co
ntainin
g
t
he
f
uzzy
ru
le
s.
With
res
pecti
ve
to
t
he
distance
var
ia
bl
e
the
m
e
m
ber
sh
ip
functi
ons
include
fa
r,
m
edium
,
cl
o
se
a
nd
t
he
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Inge
nious Met
hod
f
or
C
ondu
ci
ve Han
do
ff
A
pp
li
a
nce i
n
C
ogniti
ve R
ad
i
o N
et
works
(
J. J
os
e
ph
i
ne Dh
iv
ya
)
5199
range
of
val
ue
s
f
or
t
hese
a
re
0.1
25,
0.50
a
nd
0
res
pecti
ve
ly
.
In
ca
se
of
the
var
ia
ble
s
peed
the
m
e
m
ber
s
hip
functi
ons
are
s
low,
a
ver
a
ge,
f
ast
and
the
range
of
val
ues
ar
e
from
0
to
1.
The
delay
va
ri
able
has
m
e
m
b
ershi
p
functi
ons that i
nclu
des
l
ow,
m
edium
, h
ig
h
a
nd the
v
al
ues
a
r
e from
0
.12
5,0.50 an
d 0 res
pe
ct
ively
. Th
e
ha
ndoff
pr
i
or
it
y
is
set
t
o
ver
y
high
w
he
re
the
delay
is
low,
t
he
m
ob
il
e
nodes
are
fa
r
a
nd
the
sig
na
l
stren
gth
is
fa
st.
At
this
co
ndit
ion
t
he
hand
off
is
i
niti
at
ed.
T
he
m
ai
n
functi
ona
li
t
ie
s
carried
out
in
FBG
A
i
nc
lud
es
init
ia
li
z
at
ion
of
popula
ti
on
by
real
valued
cod
i
ng
wh
ic
h
is
do
ne
by
the
sel
ect
ion
of
values
us
in
g
the
roulet
te
wh
eel
,
cal
culat
ion
of
fitness
value
s
by
m
eans
of
m
e
m
ber
sh
ip
f
un
ct
io
ns
an
d
hand
off
res
olut
ion
value
.
Upon
th
e
arr
ival
of
the
PU
wh
ic
h
is
determ
ined
by
the
f
uzzy
r
ul
es
the
oth
e
r
m
ob
il
e
no
de
s
occupyi
ng
the
cha
nn
el
sh
oul
d
vacate
to
av
oid
inter
f
eren
ce
a
nd
t
his
is
achieved
on
t
he
basis
of
us
i
ng
the
ac
cess
point
can
did
at
e
values
sto
red
i
n
th
e
hand
off
decisi
on
ta
ble
.
The
de
ci
sion
of
init
ia
li
zi
ng
t
he
ha
ndoff
is
ta
ken
purely
ba
sed
on
the access
poin
t cand
i
date
value ob
ta
ine
d by
fu
zzy
i
nf
e
ren
c
e en
gin
e a
nd t
he
h
a
ndoff res
olu
ti
on.
4
.
2.
Cu
c
koo
S
earch
The
cr
ux
in
c
hoos
i
ng
c
ucko
o
searc
h
(CS
)
[16]
is
it
s
si
m
pl
ic
it
y
and
s
m
oo
th
i
m
ple
m
entat
ion
in
com
par
ison
w
it
h
oth
e
r
m
eta
-
he
ur
ist
ic
al
gorithm
s.
In
c
on
te
xt
of
ou
r
pro
po
se
d
work,
cuc
koo
s
e
arch
is
inco
rpor
at
e
d
i
n
re
duci
ng
th
e
occ
urren
ce
of
ha
ndoffs
a
nd
sel
ect
ing
t
he
op
ti
m
al
neighbor
hood
be
fore
the
hand
off
proces
s
w
hich
is
the
basic
idea
be
hin
d
C
S
w
he
rein
the
cuc
koo
bir
d
sear
ches
f
or
ho
st
nest
to
la
y
eggs
[17].
F
V
obta
ined
th
rou
gh
F
BGA
m
echan
i
sm
is
giv
en
a
s
in
pu
t
t
o
CS
,
determ
ines
th
e
cases
wh
e
rei
n
the
switc
h
or
c
ha
nge
of
sta
te
of
m
ob
il
e
no
des
in
the
netw
ork
ta
kes
place
.
T
he
proce
dure
e
m
plo
y
ed
in
fin
e
-
tu
ne
d
CS inclu
des
t
he
foll
ow
i
ng steps:
Step
1:
Co
ns
id
er a ran
dom
p
opulati
on
of n h
os
t
nests as
xi
Step
2: A cuc
koo i
s
ob
ta
ine
d ran
dom
l
y by le
vy f
li
ght
beh
a
vi
or
process
i.
Step
3: T
he
fit
ness funct
io
n o
btained
b
y m
eans
of FBG
A
is
ta
ken
a
s Fi.
Ste
p 4: A ra
nd
om
n
est
is ch
ose
n
am
on
g t
he
host
nest j an
d
it
s f
it
ness
is cal
c
ulate
d
as
F
j
.
Step
5:
I
f
Fi>F
j
the
n j
is
r
e
pla
ced
by n
e
w sol
ution el
se
j
is t
he
s
olu
ti
on.
Step
6: A f
racti
on of t
he worst
n
est
is eli
m
inate
d
an
d ne
w ne
st are ide
ntifie
d by m
eans o
f
l
evy flig
ht sear
ch.
Step
7: T
he op
t
i
m
u
m
n
est
is ke
pt and
ste
p 2 i
s r
e
peated
f
or
m
axi
m
u
m
it
era
ti
on
s.
Step
8: T
he op
t
i
m
u
m
n
est
is obtai
ned fi
nally
.
Th
us
on
ce
ne
w
nest
gets
af
f
ixed
it
is
e
val
uated
by
m
eans
of
the
pr
e
viously
com
pu
te
d
FV
the
reb
y
fin
ding
the
i
de
a
l
nest
w
hich
tur
ns
out
to b
e
t
he
best
ch
oice o
f
determ
ining
w
hich
m
ob
il
e n
ode
s
houl
d
underg
o
switc
hing a
hea
d of l
ink fail
ur
e b
ase
d o
n
c
ha
nn
el
sta
te
and t
his im
plies a condit
ion o
f
m
i
nim
iz
ed
ha
ndoff.
5.
RESU
LT
S
AND DI
SCUS
S
ION
The
Netw
ork
si
m
ulator
N
S
2.31
to
ol
is
use
d
to
ca
rr
y
ou
t
the
sim
ulatio
n
f
or
the
pro
pose
d
w
ork.
Si
m
ulati
on
is
perform
ed
by
rangin
g
num
ber
of
m
ob
il
e
node
s,
i
niti
al
ly
the
nu
m
ber
of
m
ob
il
e
nodes
(M
N
)
e
m
plo
ye
d
is
20;
nu
m
ber
of
channels
util
iz
ed
is
set
to
10
and
the
m
axim
u
m
si
m
ula
tio
n
ti
m
e
is
s
et
to
200
seco
nd
s
.
O
nce
the
switc
hi
ng
ta
kes
place
the
sp
ect
r
um
m
an
ger
s
enses
the
channel
a
vaila
bili
ty
and
init
ia
te
s
th
e
nex
t
set
of
tra
ns
m
issi
on
.
Fig
ure
2
ex
plains
a
case
of
sim
u
la
ti
on
m
on
it
or
ed
at
tim
e
1.
21
9490
m
s
wh
erein
the
MN
11
en
ds
th
e
ha
ndoff
on
c
ha
nnel
1
a
nd
t
he
helpe
r
node
is
8
f
or
us
er
4
and
the
total
num
ber
of
pac
ke
ts
sent
is
9
an
d
this
proces
s
co
ntin
ue
s
ti
ll
MN
11
senses
for
c
ha
nn
el
a
nd
sta
rts
transm
issi
on
at
tim
e
1.
30
1000
m
s.
Fig
ure
3
e
xpla
ins
ye
t
ano
t
her
case
of
t
he
sim
ulati
on
w
he
rei
n
tran
sm
iss
ion
ta
kes
place
acr
os
s
MN v
aryi
ng
4
t
o
8,
6
to
7
an
d
s
o
on.
Packets
s
ent
durin
g
t
his
per
i
od
of
ti
m
e
is
from
18
to
100
pa
ckets
dro
pp
e
d
are
al
so
vi
sible
durin
g
the
sim
ulati
on
pe
rio
d.
This
is
again
fo
ll
owe
d
by
the
handoff,
sen
sing
an
d
tra
nsm
issi
on
ph
as
e
ti
ll
the
si
m
ulati
on
g
et
s
over
.
Figure
2. Ha
nd
off
te
m
inati
on
Fig
ure
3
.
S
witc
hing
of inter
fa
ce
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
8
, N
o.
6
,
Dece
m
ber
2
01
8
:
5195
-
5202
5200
Ba
sed
on
the
s
i
m
ulati
on
car
ried
ou
t
t
he
a
na
ly
sis
of
ha
ndoff
par
am
et
ers
i.e.
the
facto
rs
t
hat
ove
rall
determ
ines
the
qu
al
it
y
of
ha
ndoff
ca
n
be
m
on
it
ore
d
easi
ly
.
The
pa
ram
eter
s
ta
ken
for
con
si
der
at
io
n
in
cl
ud
e
thr
oughput,
de
la
y
and
num
ber
of
ha
ndoffs
occurre
d
an
d
nu
m
ber
of
fa
il
ed
hand
off.
These
are
disc
us
se
d
br
ie
fly
in
the
foll
ow
i
ng secti
on.
a.
Thro
ughput:
T
his
pa
ram
et
er
rep
rese
nts
the
s
uccess
fu
l
tra
nsm
issi
on
of
data
ov
e
r
the
gi
ven
per
io
d
of
ti
m
e
from
so
ur
ce
to
destinat
io
n
it
is
the
ben
c
hm
a
rk
for
deci
ding
the
ove
rall
effi
ci
ency
of
the
syst
e
m
.
Fig
ure
4.rep
rese
nts
the
am
ou
nt
of
data
tra
ns
m
i
ssion
occ
urre
d
du
rin
g
the
si
m
ulati
on
ti
m
e
with
X
a
xis
represe
nting sim
ula
ti
on
ti
m
e i
n
sec
onds an
d Y a
xis r
e
prese
nting t
he
am
ount of
data tra
nsm
itted in k
bp
s
.
b.
Delay
:
This
de
picts
t
he
ti
m
e
t
aken
f
or
t
ran
s
m
issi
on
of
bits
from
so
ur
ce
node
to
destinat
ion
node
.
Fig
ure
5.
represe
nts
the
ov
e
rall
del
ay
occurre
d
duri
ng
t
he
handoff
proce
ss
with
X
axis
r
epr
ese
ntin
g
th
e
si
m
ulati
on
tim
e in sec
onds an
d Y a
xis r
e
pres
enting t
he dela
y peri
od in
sec
o
nds.
c.
Nu
m
ber
of
ha
ndoffs:
The
m
axim
u
m
nu
m
be
r
of
ti
m
es
swit
chin
g
of
sta
te
s
occur
durin
g
da
ta
transm
issi
on
ref
e
rs
to
as
th
e
nu
m
ber
of
ha
ndoff
par
am
et
er.
This
s
houl
d
not
ta
ke
place
for
m
or
e
num
ber
of
ti
m
e
s.
Fig
ure
6
repre
sents
the
num
ber
of
hand
off
process
occ
ur
red
durin
g
the
entire
sim
ulatio
n
ti
m
e
with
X
axis
re
pr
ese
nti
ng
sim
ulati
on
tim
e
in
seco
nds
an
d
Y
a
xis
r
epr
ese
ntin
g
th
e
nu
m
ber
of
ha
ndoff
process
occurre
d
with
two
li
ne
s
re
pr
es
enting
t
he
exis
ti
ng
an
d
pro
posed
syst
em
.
Fr
om
the
si
m
ulatio
n
resu
lt
s
it
is
cl
ear
that
the
pro
po
se
d
syst
em
chan
ges
it
s
sta
te
s
m
ini
m
al
num
ber
of
ti
m
es
com
par
ed
to
the
existi
ng
syst
e
m
.
d.
Nu
m
ber
of
fail
ed
ha
ndoffs:
T
his
gauge
ex
hi
bits
the
capaci
t
y
of
a
netw
ork
i.e.
i
nd
ic
at
es
c
ases
w
he
rein
a
sta
te
in
a
chan
nel
durin
g
tr
ansm
issi
on
is
un
s
ucces
sf
ul.
Fig
ure
7.
re
presents
the
nu
m
ber
of
fail
ed
hand
offs
occ
urred
du
rin
g
the si
m
ulati
on
tim
e w
it
h
X
axis re
pr
ese
ntin
g
sim
ula
ti
on
tim
e
i
n
seco
nds and Y
axis r
e
presenti
ng the
num
ber
of f
ai
le
d ha
ndoff p
ro
ces
s.
Fr
om
the
analy
sis
of
the
pa
r
a
m
et
ers
inv
ol
ve
d
in
t
he
hand
off
proce
ss
ge
ner
at
e
d
an
d
th
e
com
par
iso
n
of
ha
ndoff
sc
hem
es
gen
erat
ed
by
dynam
ic
pro
gr
am
m
in
g
wit
h
bisect
i
on
(
DBA
)
al
gorithm
,
fu
zzy
base
d
gen
et
ic
an
d
t
he
pro
po
se
d
FB
GA
with
CS
in
Table
2,
evi
de
nt
that
the
res
ul
ts
are
far
bette
r
an
d
are
opti
m
iz
ed
com
p
ared
t
o
th
e existi
ng syst
e
m
s in
te
r
m
s o
f un
ce
rtai
n netw
ork
c
onditi
ons.
Figure
4. Th
r
ough
pu
t
Fig
ure
5
.
Dela
y
Figure
6. N
umber
of
hand
offs
Figure
7. N
umber
of
fail
ed ha
ndoffs
Evaluation Warning : The document was created with Spire.PDF for Python.
In
t J
Elec
&
C
om
p
En
g
IS
S
N: 20
88
-
8708
Inge
nious Met
hod
f
or
C
ondu
ci
ve Han
do
ff
A
pp
li
a
nce i
n
C
ogniti
ve R
ad
i
o N
et
works
(
J. J
os
e
ph
i
ne Dh
iv
ya
)
5201
Table
2
.
C
om
par
iso
n of Ha
nd
off
Sc
hem
es
Han
d
o
ff
Par
a
m
et
er
s
Tr
an
s
m
iss
io
n
ti
m
e
in
sec
Han
d
o
ff
bas
ed
on
DBA
Han
d
o
ff
b
ased
on
FGA
Prop
o
sed
Han
d
o
ff
bas
ed
on
h
y
b
rid FBGA
with
CS
Av
erage Del
ay
in
sec
1
2
3
4
5
6
7
8
9
10
12
16
19
21
23
26
28
31
32
38
8
12
15
17
21
24
25
27
30
35
4
7
9
15
16
19
21
23
25
29
Av
erage T
h
rou
g
h
p
u
t
in
kbps
1
2
3
4
5
6
7
8
9
10
1892
1750
1627
1582
1488
1201
1057
937
893
742
2174
2058
1964
1846
1702
1600
1592
1458
1387
1272
3512
3201
3085
2840
2792
2502
2359
2075
1957
1810
Nu
m
b
e
r
o
f
Hand
o
ff
s
1
2
3
4
5
6
7
8
9
10
24
26
31
34
38
42
44
49
52
55
18
19
25
29
32
34
37
39
42
45
14
16
22
27
29
31
33
34
37
40
Nu
m
b
e
r
o
f
Failed
Han
d
o
ff
s
1
2
3
4
5
6
7
8
9
10
10
13
15
19
22
26
28
31
34
37
7
10
13
16
18
19
22
24
27
29
4
7
9
12
15
16
17
19
21
22
6.
CONCL
US
I
O
N
In
t
his
pa
per,
the
ha
ndoff
process
a
par
t
of
the
s
pectr
um
m
ob
il
it
y
a
chall
eng
in
g
ta
sk
in
CR
N
is
addresse
d
a
nd
analy
sed
by
inco
r
porati
ng
hy
br
id
m
et
ho
dolog
y
of
f
uzzy
base
d
gen
et
ic
al
gorithm
and
it
is
op
ti
m
ise
d
us
ing
cuc
koo
sear
ch.
T
he
si
m
ulati
on
res
ults
sh
ow
that
the
pro
posed
m
et
ho
d
is
reli
able
one
a
nd
ens
ur
es
that
th
e
ha
ndoff
proc
ess
is
ca
rr
ie
d
ou
t
i
n
a
sm
oo
th
way
wit
h
m
i
nim
a
l
back
l
ogs
in
c
om
par
iso
n
wit
h
the ex
ist
in
g
sy
stem
.
REFERE
NCE
S
[1]
Herná
ndez,
C.
;
Salga
do,
C
.
;
Ló
pez
,
H
.
;
Rodríg
uez
-
Col
ina
,
E.
,
“
Multi
var
ia
b
le
al
gorit
hm
for
d
y
nami
c
cha
nn
el
sele
c
ti
on
in
cogn
it
ive
rad
io
n
et
wo
rks,”
EURA
S
IP
Journal
on
Wire
le
ss
Comm
unic
ati
ons
and
Net
wo
rking
,
vol
.
2015,
no.
1
,
pp
.
1
–
17
,
2015.
[2]
Krishan
Kum
ar,
Arun
Praka
sh
and
Raj
ee
v
Tri
p
a
thi
,
“
Spect
rum
Handoff
Scheme
with
ultiple
A
tt
ributes
Dec
i
sio
n
Making
for
Optimal
Networ
k
Sele
ction
in
Cognit
ive
R
a
dio
Networks”,
Digit
al
Com
municat
ions
an
d
Net
works
,
Janua
r
y
2017.
[3]
Pradip
Vara
de,Akanksha
W
aba
le,
“
Throughput
Maximiza
ti
o
n
of
Cognit
ive
rad
io
m
ult
i
rel
a
y
n
et
work
with
int
erf
ere
n
ce
m
an
age
m
ent
”
,
Int
er
nati
onal
Journal
of
Elec
tric
al
an
d
Computer
Eng
ine
ering
(
IJE
C
E
)
,
Vol.
8,
No.
4,
Augus
t
2018,
pp
.
2230
-
2238.
[4]
Prithi
viRaj
A,
K
rishnamurth
y
K
,
“
Fuzz
y
logi
c
bas
ed
de
ci
sion maki
ng
al
gori
thm to
opti
m
iz
e
the ha
n
doff
per
form
ance
in
Hetn
et
s”,
Ci
rcuit
s and
S
ystem
s
756
-
3
777,
Se
pte
m
ber
2016
.
[4]
Prithi
viRaj
A,
Krishnam
urth
y
K,
“
Fuzz
y
logic
base
d
decisio
n
m
aki
ng
al
gor
it
hm
to
opti
m
iz
e
the
h
andof
f
per
form
anc
e
in
Hetne
ts”
,
C
ircuits and
Syst
ems
7
56
-
3777,
Sept
e
m
ber
2016.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2088
-
8708
In
t J
Elec
&
C
om
p
En
g,
V
ol.
8
, N
o.
6
,
Dece
m
ber
2
01
8
:
5195
-
5202
5202
[5]
Pree
tha
m
C.
S
,
Prasad
M.S.G
et
al
,
“
Perform
anc
e
anal
y
s
is
of
coope
rative
h
y
bri
d
Cognit
ive
rad
io
net
work
wit
h
var
ious
dive
rsi
t
y
te
chn
ique
s”
,
Int
ernati
onal
Journal
of
E
lectric
a
l
a
nd
Computer
En
gine
ering
(
IJE
C
E)
,
Vol.
6
,
No.5
,
Octobe
r
2016
,
p
p
2125
-
2133.
[6]
Tsai
,
K.
,
L
iu,
H.
and
Li
u
,
Y
,
“
Us
ing
Fuzz
y
Logic
to
Redu
ce
Pin
g
Pong
Handove
r
Eff
ect
in
LT
E
Networks
”,
Sof
t
Computing
,
20
,
1683
-
1694,
201
6.
[7]
Potdar,
S.M.;
Pa
ti
l,
K.P
.
,
"Effici
ent
spec
trum
ha
ndoff
in
CR
netw
ork
base
d
on
m
obil
ity
,
QoS
a
nd
priori
t
y
usin
g
fuz
z
y
log
ic
and
neur
al
n
et
work,"
Conte
mpor
ary
Computing
(
IC3
)
,
2013
S
ix
th
Int
ernati
onal
Conf
ere
nce
on
,
vo
l
.
,
no.
,
pp.
53
,
58,
8
-
10
Aug.
2013
.
[8]
Prudhvi
Raj
M
et
ti,
K.
Rushen
dra
Babu
,
Sum
it
Kum
ar,
“
Spe
ct
rum
Handoff
Mec
hani
sm
in
Cognit
ive
Radio
Networks
using
Fuzz
y
Logic”
In
te
rnational
Journal
of
Scientific
&
Engi
nee
ring
Re
search
,
Volu
m
e
5,
Iss
ue
10,
Octobe
r
-
2014.
[9]
W
anmai
Yuan,
Mingchua
n
Yan
g,
“
Im
prove
d
cuc
koo
sea
rch
algorithm
for
spec
trum
sensing
i
n
sparse
sate
l
li
t
e
cogni
ti
v
e
s
y
stem
s”
IEEE
84
th
V
eh
ic
ular Te
chnol
o
gy
con
fe
ren
ce
2
016.
[10]
Ali
Cal
han
,
Cela
l
Ceka
n
“
An
Op
ti
m
um
Vert
ic
al
Handoff
Dec
isio
n
Algorit
hm
Based
on
Adapti
ve
Fuzz
y
Logi
c
and
Gene
tic
Algor
it
h
m
”,
Wire
le
ss
P
e
rs
Comm
un
(2012)
64:647
–
664.
[11]
Miguel
Tube
rq
uia
,
Cesar
Her
nande
z
,
“
New
Approac
hes
in
cogni
ti
v
e
rad
ios
using
evol
uti
o
nar
y
al
gori
thms
”
,
Inte
rnational
jou
rnal
of El
e
ct
ri
ca
l
and
Comput
er
Engi
ne
ering
(
IJ
ECE
)
,
Vol.
8
,
No
.
3
,
June
2018,
p
p.
1636
-
1646
.
[12]
Kale
Sandik
ar
R.
S,
Vijay
M.
W
adha
i,
“
New
Algorit
hm
for
Handoff
Optimiza
t
ion
in
Cognit
ive
Radi
o
Ne
tworks
using Fuzzy
logic
and
Artif
ic
i
al
Neura
l
N
et
work
”,
El
sev
ie
r
Publ
i
ca
t
ions,2
013.
[13]
Salga
do.
C
,
Hern
ande
z
.
C,
Mol
ina.V,
“
Inte
l
li
gen
t
al
gorit
hm
for
sp
ec
trum
m
obil
ity
in
cognitive
wi
rel
ess
net
works
”
,
Proce
dia
Computer
Sc
ie
nc
e
83
,
2
78
–
283
,
2016
.
[14]
Ekt
a
Dua
,
Vikr
ant
Gulati,
“
En
hanc
ement
in
Spect
rum
Sensin
g
Algorit
hm
s
for
Cognit
ive
R
ad
ios
”
Inte
rnation
al
Journal
of
Adv
a
nce
d
Re
search
i
n
El
ec
troni
cs
and
Comm
unic
ati
on
Engi
nee
ring
(
IJA
RE
CE)
Volum
e
5,
Iss
ue
8,
Augus
t
2016.
[15]
hang
Trung
Nguy
en
,
Dieu
Ngoc
Vo,
“
The
appl
i
c
at
ion
of
one
ran
k
cuc
koo
sea
rch
al
gorit
hm
for
solvingecono
m
ic
loa
d
d
ispat
ch
pr
oble
m
s”
Applied
Soft Computing
37,
763
–
773
,
201
5.
[16]
W
.
Ahm
ed,
J.
Gao
and
H.
Surawee
ra
,
“
Anal
y
sis
of
cogni
t
iv
e
rad
io
spec
tru
m
ac
ce
ss
with
opti
m
al
cha
nne
l
rese
rva
t
ion”,
I
E
EE
Tr
ansacti
ons
on
Wire
le
ss
Co
mm
unic
ati
ons
,
v
ol.
8
,
no
.
9
,
pp
.
4
488
-
44
91,
2014.
[17]
Yanhong
Feng,
Ke
Jia,
and
Yicha
o
He
,
“
An
Im
prove
d
H
y
brid
Enc
od
ing
Cuckoo
Sear
ch
Algorit
hm
for
0
-
1
Knapsac
k
Probl
ems
”,
Hindawi
Publ
ishing
Cor
poration
Computati
onal
In
te
l
li
g
enc
e
and
N
euroscie
nc
e
Volum
e
2014.
BIOGR
AP
H
I
ES
OF
A
UTH
ORS
J.Jos
ephi
ne
Dhiv
y
a
r
ec
e
ive
d
the
MCA
degr
ee
fr
om
Madura
i
Kam
ara
j
Univer
sit
y
and
is
cur
ren
tly
a
Resea
r
ch
Schola
r
in
Madura
i
Kam
ara
j
Uni
ver
sit
y
.
Her
area
s
of
int
ere
st
in
cl
ude
W
ireles
Com
m
unic
at
ions a
nd
Network
in
g
Dr.
M.
R
amasw
ami
(M.Sc,
M.Phil
,
M
.
C.
A)
was
awa
rd
ed
the
Phd
degr
ee
in
Com
puter
Applic
a
ti
ons
fro
m
Madura
i
Kam
ara
j
Univer
sit
y
.
He
is
cur
r
ent
l
y
working
as
a
Profess
or
in
the
Depa
rtment
of
Com
pute
r
Applic
ations
in
Madura
i
Kam
ara
j
Uni
ver
sit
y
and
an
a
ct
iv
e
rese
ar
ch
supervisor
and
he
has
publi
shed
about
45
pape
rs
in
var
ious
rep
ute
d
journ
al
s
at
nat
ion
al
and
int
ern
at
ion
al
le
v
el
.
Evaluation Warning : The document was created with Spire.PDF for Python.