Indonesi
an
Journa
l
of El
ect
ri
cal Engineer
ing
an
d
Comp
ut
er
Scie
nce
Vo
l.
13
,
No.
3
,
Ma
rch
201
9
, p
p.
1
143
~
1
151
IS
S
N: 25
02
-
4752, DO
I:
10
.11
591/ijeecs
.v1
3
.i
3
.pp
1
143
-
1
151
1143
Journ
al h
om
e
page
:
http:
//
ia
es
core.c
om/j
ourn
als/i
ndex.
ph
p/ij
eecs
Concept
ual
m
obi
lity
m
odel of
v
ert
ical
h
an
dover
d
ecision
in
h
eteroge
n
eo
us
n
etworks
No
r
ak
m
ar
A
r
ba
in
1
,
Z
olida
h
Kasi
ran
2
1
Facul
t
y
of Elect
ric
a
l
Eng
ineeri
n
g,
Univer
si
ti T
ek
nologi
MA
RA (
UiTM),
Mal
a
y
si
a
2
Facul
t
y
of
Com
pute
r
Sc
ie
nc
e an
d
Mathe
m
atics,
Univer
siti
Te
kno
logi
MA
RA (Ui
TM),
Mal
a
y
sia
Art
ic
le
In
f
o
ABSTR
A
CT
Art
ic
le
history:
Re
cei
ved
Oct
10
, 201
8
Re
vised Dec
6
,
2018
Accepte
d Dec
20
, 201
8
In
het
ero
g
ene
o
us
net
work,
m
ai
nta
ini
ng
s
e
amless
conne
ctivity
ne
eds
exc
essive
eff
ort
s
from
var
ious
aspe
ct
s
such
as
net
work
availa
bi
li
t
y
and
m
obil
e
node
re
li
ability
.
Presen
tly
,
a
ve
rtica
l
handove
r
m
anagem
ent
is
a
pra
ctical
appr
oa
ch
in
fac
i
li
t
atin
g
the
service
c
onti
nuity
for
m
obil
e
users.
Man
y
res
e
arc
h
es
have
bee
n
conduc
te
d
in
thi
s
are
a
b
y
conside
rin
g
per
form
anc
e
improvem
ent
in
d
el
a
y
,
laten
c
y
,
a
nd
over
hea
d
.
Pr
ese
rving
th
e
Quali
t
y
of
Serv
ic
es
(QoS
)
base
d
on
user
m
obil
ity
and
pa
tt
ern
m
ovement
during
handover
dec
ision
h
as
bec
om
e
an
i
m
porta
nt
aspe
c
t
in
ver
t
ic
a
l
handove
r
m
ana
g
ement.
Th
is
pap
er
pre
sents
th
e
c
once
ptu
al
m
obil
i
t
y
m
odel
o
f
ver
tical
h
andov
er
decision
in
het
ero
g
ene
ous
net
work.
Hen
ce
,
sev
eral
rese
arc
h
es
in
v
ert
i
ca
l
h
andove
r
dec
ision
m
an
a
gement
has
be
e
n
rev
ie
wed
reg
ard
ing
th
e
issues
on
the
ver
ti
c
al
handove
r
dec
ision
al
go
rithm
s
such
as
RS
S
Based
Alg
orit
hm
,
MA
DM
Based
Algorithm
an
d
Inte
ll
ig
enc
e
Bas
ed
Algorit
hm
.
This
pape
r
highl
ig
hts
the
cur
ren
t
dec
ision
al
gor
it
hm
s
tha
t
int
egr
at
e
the
tr
adi
ti
on
al
m
et
ho
ds
with
intel
li
g
enc
e
al
go
rit
hm
for
be
tt
e
r
opti
m
iz
a
t
ion.
In
dec
ision
par
a
m
et
ers,
the
use
r
m
obil
ity
p
at
t
ern
ca
n
b
e
importanc
e
in
te
rm
s
of
d
ir
ec
t
ion
ran
dom
ness
and
m
obil
ity
sp
ee
d
.
Henc
e,
a
conc
e
ptua
l
m
obil
ity
-
a
ware
ness
m
odel
for
ver
ti
cal
ha
ndover
are
bee
n
proposed
i
n
ta
rg
et
ing
som
e
improvem
ent
of
handove
r
p
erf
o
r
m
anc
e.
Ke
yw
or
ds:
Heter
og
e
ne
ou
s
n
et
w
ork
MADM
Mob
il
it
y
m
anag
em
ent
Mob
il
it
y
p
at
te
r
n
Ver
ti
cal
h
a
ndover
Copyright
©
201
9
Instit
ut
e
o
f Ad
vanc
ed
Engi
n
ee
r
ing
and
S
cienc
e
.
Al
l
rights re
serv
ed.
Corres
pond
in
g
Aut
h
or
:
Nora
km
ar A
r
ba
in,
Faculty
of
Ele
c
tric
al
Engineer
ing
,
Un
i
ver
sit
i Te
knol
og
i M
ARA
(U
iTM
)
,
40450 S
hah A
l
a
m
, S
el
ango
r,
Ma
la
ysi
a.
Em
a
il
:
niez
m
ar
@g
m
ai
l.co
m
1.
INTROD
U
CTION
Re
centl
y,
m
ob
il
e
dev
ic
es
su
c
h
as
sm
artpho
nes,
la
ptops
an
d
ta
blets
ha
ve
been
e
xtensiv
e
ly
us
ed
wi
t
h
su
pp
or
t
of
a
dvanced
net
wor
k
te
chnolo
gies.
An
inter
net
ac
cess
an
d
c
omm
un
ic
at
ion
act
ivit
y
happen
at
anyt
i
m
e
and
any
wh
e
re.
Mob
il
it
y
m
a
y
aff
ect
t
he
c
om
m
un
ic
at
ion
connecti
vity
w
hen
the
m
ob
il
e
node
(M
N)
m
akes
changes
on
th
e
prefe
rr
e
d
net
work
durin
g
t
he
c
omm
un
ic
at
ion
process
.
Hen
ce
,
a
go
od
ap
proac
h
of
m
ob
il
i
ty
m
anag
em
ent
is
esse
ntial
in
preser
ving
t
he
Q
ualit
y
of
Ser
vice
(
QoS)
to
us
e
r.
M
obil
it
y
m
anag
em
ent
in
heter
og
e
ne
ou
s
wireless
netw
ork has
em
erg
ed
and r
e
searc
he
d ov
e
r
the
yea
r
s
[1
]
-
[
3]
. One of
t
he
im
po
rta
nt p
a
rts
in
m
ob
il
it
y
manag
em
ent
is
how
to
m
anag
e
the
handove
r
process
wh
e
n
the
m
ob
il
e
node
cha
nges
or
m
ov
e
away
from
cu
rr
e
nt
serv
ic
e
netw
ork
to
an
oth
e
r
ser
vice
netw
ork
with
out
disruptin
g
the
com
m
un
ic
at
ion
connecti
vity
.
Hand
ov
e
r
is
a
pr
oce
ss
w
hen
the
m
ob
il
e
no
de
m
ov
es
fro
m
on
e
wireles
s
cel
l
to
ano
th
er
and
requires
t
he
se
a
m
le
ss
con
nec
ti
on
with
t
he
nex
t
wireless
cel
l.
A
seam
less
co
nn
ect
ivit
y
[2]
or
“Al
wa
ys
Be
st
Connect
ed”
is
about
m
ai
ntaining
t
he
net
wor
k
co
nne
ct
ivit
y
of
al
l
r
unni
ng
app
li
cat
io
ns
on
the
m
ob
il
e
de
vice.
This
is
sti
ll
a
chall
eng
i
ng
ta
sk
to
facil
it
at
e
the
best
ap
proac
h
that
ca
n
s
upport
the
seam
le
s
s
co
nn
ect
ivit
y
wh
il
e
m
ai
ntaining
t
he
QoS
at
us
e
r
l
evels
[
3]
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
13
, N
o.
3
,
Ma
rc
h 201
9
:
1
143
–
1
151
1144
Hand
ov
e
r
m
a
nag
em
ent
in
wireless
netw
ork
ca
n
be
ca
rr
ie
d
out
ei
the
r
ho
rizo
ntall
y
or
ve
rtic
al
ly
.
Horizo
ntal
ha
ndove
r
ha
pp
e
ne
d
in
hom
og
e
ne
ou
s
e
nvir
on
m
ent
w
her
e
the
ne
twork
c
hangi
ng
process
occ
urred
within
the
sam
e
do
m
ai
n
of
ne
twork
,
su
c
h
as
the
m
ob
il
e
node
changed
the
pr
e
ferred
netw
ork
from
W
LA
N1
to
WL
AN2
with
in
the
sa
m
e
do
m
ai
n.
Me
an
wh
il
e,
in
hete
roge
neous
net
work,
ve
rtic
al
handover
is
m
o
re
chall
eng
i
ng
as
the
m
ob
il
e
no
de
need
s
to
c
ha
ng
e
the
pr
e
fe
rr
e
d
netw
ork
f
ro
m
diff
e
ren
t
netw
ork
do
m
ai
n
an
d
te
chnolo
gies,
s
uch
a
s
W
iM
a
x,
W
iFi
,
UMTS/
LTE
an
d
oth
e
r
s.
Mo
reover
,
t
he
ha
ndove
r
m
anag
em
ent
pro
cedure
can
be
cl
assifi
ed
into
th
ree
ty
pes
w
hich
ar
e
h
ar
d,
soft
and
s
of
te
r
ha
ndov
e
r.
Hard
ha
ndover
is
refe
rr
e
d
to
br
ea
king
t
he
c
onnected
netw
ork
befo
re
m
ake
the
ne
xt
net
w
ork
c
onnecti
on
,
as
t
he
m
ob
il
e
node
only
co
nnect
ed
to
on
e
point
of
at
ta
ch
m
ent
(P
OA)
at
a
tim
e.
Me
anwhil
e,
soft
hand
ov
e
r
al
lowe
d
m
ob
il
e
no
de
t
o
co
nn
ect
to
two
POA
for
a
w
hi
le
un
ti
l
the
best
connecti
on
ob
ta
ine
d.
Howev
e
r,
s
of
te
r
hand
ov
e
r
[
4]
is
the
best
ap
proac
h
in
hand
ov
e
r
m
anag
em
ent
wh
ere
this
fast
and
s
m
oo
th
handov
er
can
m
ini
m
is
e
the
la
te
ncy
a
nd
pack
et
loss
durin
g
hand
ov
e
r
e
xec
ution.
In
gen
e
ra
l,
Han
do
ve
r
Ma
nage
me
nt
Archite
ct
ur
e
consi
sts
of
t
hree
par
ts
w
hi
ch
a
re
Han
do
ve
r Init
iati
on, Ha
ndov
er D
eci
sio
n
a
nd
Han
do
ve
r Ex
ecuti
on.
1.1.
Ha
n
dover
Init
iation
Durin
g
syst
em
disco
ver
y,
al
l
inform
at
ion
fr
om
m
ob
il
e
no
de
a
nd
netw
orks
are
c
ollec
te
d
su
c
h
as
batte
ry
powe
r,
netw
ork
ba
nd
width
a
nd
si
gnal
streng
t
h.
Q
oS
requirem
ents
are
al
so
ta
ken
into
co
ns
ide
ra
ti
on
as
an
in
put f
or the
n
e
xt h
a
ndove
r
phase.
1.2.
Ha
n
dover
Dec
isi
on
At
this
phase,
a
m
ob
il
e
node
needs
to
m
ake
a
ne
w
netw
ork
sel
ect
ion
beca
us
e
of
s
om
e
factor
s
su
c
h
a
s
rap
i
d
m
ob
il
e
no
de
m
ov
em
ent
and
stum
py
ne
twork
c
ov
e
ra
ge
.
Ma
inly
,
handove
r
decisi
on
sel
ect
s
new
pre
ferred
netw
ork
base
d
on
the
receive
d
sig
nal
stre
ng
t
h
an
d
highest
-
rankin
g
in
dicat
or
c
ollec
te
d
f
r
om
syst
e
m
discov
e
ry
process
.
So
m
e
decisi
on
al
gorithm
s
are
m
anipu
la
te
d
usi
ng
m
at
he
m
a
ti
cal
or
com
pu
ta
ti
on
al
f
or
m
ula
for
m
anag
in
g
t
he han
dove
r per
f
orm
ance li
ke
ov
erh
ea
d
a
nd
delay
.
1.3.
Ha
n
dover E
xe
cutio
n
Wh
e
n
the
sel
ect
ed
netw
ork
s
at
isfie
s
the
Q
oS
requirem
ents
of
m
ob
il
e
no
de,
the
ha
ndov
er
exec
uted
base
d
on
t
he
c
ertai
n
ha
ndove
r
co
ntr
oller.
T
he
hand
ov
e
r
m
anag
em
ent
co
nt
ro
ll
er
ca
n
be
ei
ther
at
m
ob
il
e
node
or
netw
ork
sid
e.
Hen
ce
,
f
our
ty
pes
of
hand
over
process
c
ontr
ol
are
Net
w
ork
Co
ntr
olled
HandOve
r
(NC
HO),
Mob
il
e
Co
ntr
olled
Hand
Over
(MC
HO),
Mob
il
e
-
Assiste
d
Hand
Ov
e
r
(MA
HO)
a
nd
Netw
ork
-
As
s
ist
ed
Hand
Ov
e
r (N
AHO)
[
2]
.
This
pa
per
pre
sents
the
c
on
c
eptual
m
ob
il
ity
m
od
el
of
ve
rtic
al
handover
decisi
on
i
n
he
te
rogen
e
ous
netw
ork
a
nd
orga
nized
a
s
f
ol
lows
.
Sect
io
n
2
disc
us
se
d
th
e
curre
nt
ve
r
ti
cal
hand
ov
e
r
m
anag
em
ent
i
nclu
ding
the
ver
ti
cal
ha
ndover
decisi
on
al
gorithm
s
analy
sis.
Sect
ion
3
pr
ese
nt
s
the
pro
pose
d
m
ob
il
it
y
-
awar
enes
s
m
od
el
du
rin
g
netw
ork
sel
ect
ion
.
Sect
io
n
4
enco
m
passes
the
co
nclusi
on
and
f
uture
rec
omm
end
at
io
ns
on
this
researc
h.
2.
R
ESE
A
R
CH MET
HO
D
2.1.
Vert
ic
al H
an
d
ov
er
D
eci
si
on
(VHD
) Alg
orithm
Seve
ral
li
te
ratur
es
ha
ve
pres
ented
a
n
over
vi
ew
of
ver
ti
cal
ha
ndover
deci
sion
strat
egies
in
diff
e
re
nt
cat
egories.
I
n
r
esearch
[
5]
,
th
e
auth
or
s
m
ake
the
com
par
is
on
betwee
n
ve
rtic
al
handove
r
decisi
on
strat
e
gies
in
five
cat
e
gories
su
c
h
a
s
Decisi
on
F
un
ct
io
n
(D
F
),
U
ser
Ce
ntric
(U
C
),
M
ulti
ple
Attri
bute
Decisi
on
(
MAD
)
,
Fu
zzy
L
og
ic
/
Neural
Net
wor
k
(
FL/N
N)
a
nd
Co
ntext
-
A
w
are
(C
A)
.
The
y
propose
d
a
new
ha
ndove
r
decisi
on
schem
e
con
ta
ins
tw
o
com
ponen
ts
wh
ic
h
ar
e
Fu
zzy
Lo
gic
Syst
e
m
(F
LS)
and
Netw
ork
Sele
ct
ion
(
us
in
g
A
HP
m
et
ho
d).
A
uthors
in
[6]
,
pr
opos
e
a
ve
rtic
al
m
ob
il
it
y
m
anag
em
ent
arc
hitec
ture
nam
e
d
as
C
onte
xt
-
Aw
a
re
Mob
il
it
y
Ma
nag
em
ent
Syste
m
(CAMMS)
w
hich
sup
port
the
cr
os
s
-
la
ye
r,
con
te
xt
-
awar
e
a
nd
se
a
m
le
ss
hand
ov
e
r
f
or
use
r
an
d
ser
vice
s.
They
desi
gn
four
m
ai
n
com
po
nen
ts
of
f
un
ct
io
nal
entit
ie
s
that
respon
s
ible
fo
r
con
te
xt
gatheri
ng
,
intel
li
gen
t
ha
ndover
de
ci
sion
-
m
akin
g,
accu
rate
ha
ndove
r
trig
ge
ring
an
d
post
-
ha
ndof
f
m
anag
em
ent.
J.
Má
r
qu
ez
-
B
arj
a
et
al
.
[7]
,
rev
ie
w
on
al
gorithm
s,
protoc
ols
an
d
t
oo
ls
i
n
ve
rtic
al
handove
r
m
anag
em
ent.
The
pa
pe
r
f
oc
us
on
veh
ic
ular
netw
ork
,
w
he
re
ve
hicle
as
a
node
that
ha
ve
seve
ral
co
nst
rained
su
c
h
as
top
ol
og
y
restrict
io
ns,
m
ob
il
it
y
pa
tt
ern
s,
po
wer
consum
ption
,
scal
abili
ty
,
rel
ia
bili
ty
and
sp
eed
.
Re
search
in
[
8]
cat
eg
or
iz
e
d
the
VHO
Al
gorithm
into
f
our
ty
pes
w
hi
ch
a
re
L
ocati
on
Ba
sed
Ha
ndove
r,
Mob
il
it
y
Ba
sed
Ha
ndover
,
P
olicy
Ba
sed
H
ando
ver
a
nd
L
earn
i
ng
Ba
se
d
Hand
ov
e
r.
Me
anwhil
e,
aut
hors
in
[
9]
cond
ucted
a
survey
of
ha
ndov
er
decisi
on
al
gorithm
s
fo
r
LT
E
-
A
fem
tocel
ls
.
They
cl
assifi
ed
the
H
O
Algo
rithm
into
fi
ve
as
pe
ct
wh
ic
h
ar
e
RSS,
S
pee
d,
I
nterf
e
re
nce
-
a
w
are,
C
os
t
-
funct
ion
a
nd
E
nerg
y
-
eff
ic
ie
nt.
Th
ey
al
so
highli
gh
te
d
on
the
nee
d
of
f
ut
ur
e
rese
arc
h
on
m
ulti
ple
-
m
ac
ro
cel
l
m
ulti
ple
-
fem
tocel
l
scenari
o
w
he
re
RSS
an
d
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Con
ce
ptual
mobili
ty
m
odel
of
vert
ic
al hand
ov
er d
eci
sio
n
i
n hetero
ge
neous netw
or
ks
(N
orakm
ar
Arbai
n)
1145
RSQ
as
m
ai
n
par
am
et
ers.
Y
an
et
al
.
in
[10]
,
s
umm
arize
the
ver
ti
cal
ha
ndover
decisi
on
al
gorithm
into
f
our
cat
egorie
s
w
hich
are
RSS
Ba
sed
V
H
D
Algo
rithm
s,
Ba
nd
w
idth
Ba
sed
VHD
Algo
rithm
s,
Cost
Functi
on
Ba
se
d
VHD
Algo
rith
m
s
and
Com
bin
at
ion
Ba
se
d
VHD
Algo
rith
m
s.
Twelve
V
HD
al
gorithm
s
ha
ve
bee
n
a
na
ly
ze
d
and
they
pr
es
ented
t
he
a
dvantages
an
d
di
sadv
a
ntages
of
eac
h
al
gorit
hm
.
T
he
cu
rrent
ve
rtic
al
ha
ndove
r
decisi
on
ca
n
be
cat
eg
ori
es
into
RSS
base
d
Algorithm
,
MADM
base
d
Algorithm
an
d
I
ntell
igence
base
d
Algorithm
as s
how
n
in
Fig
ure
1
.
Figure
1
.
V
e
rtic
al
hand
ov
e
r d
eci
sion
sc
hem
es
2.1.1.
RS
S
b
as
ed
Al
go
ri
th
m
Re
cei
ved
Si
gnal
Stren
gth
(RSS)
is
a
c
omm
on
hand
ov
e
r
decisi
on.
In
[
11
]
,
a
m
at
hem
at
ic
al
m
od
el
base
d
on
V
H
O
pr
e
dicti
on
appr
oach
ha
s
bee
n
re
view
by
c
onside
ring
s
om
e
par
a
m
et
ers
su
c
h
a
s
RSS,
UE
vel
ocity
,
l
oad
a
nd
c
os
t
pe
r
us
e
r
ba
ndw
i
dth
.
T
he
pro
pose
d
al
gorithm
has
bee
n
si
m
ulate
d
in
Ma
tl
ab
an
d
fo
ll
ows
the
Ja
ke’
s
m
od
el
.
E
valuati
on
of
th
e
networ
k
perf
or
m
ance
is
based
on
us
e
r
ve
locit
y
and
handove
r
nu
m
ber
s.
Handove
r
process
evaluate
d
in
W
i
Fi
and
W
i
Ma
x
acce
ss
ne
tworks
.
Au
t
ho
rs
in
[1
2
]
-
[
1
3
]
,
ev
al
uate
the
hand
ov
e
r
pe
rfor
m
ance
ba
sed
on
Re
cei
ve
d
Si
gnal
Stre
ngth
I
nd
ic
at
or
(
RSSI)
al
gorit
hm
fo
r
WLAN/
Ce
ll
ular
netw
ork
f
or
M
ob
il
e
V
oice
use
rs.
Ning
et
al
.
in
[1
2
]
,
pr
opose
d
ha
ndove
r
a
lgorit
hm
based
on
RS
S
an
d
Ma
rko
v
m
ob
il
i
ty
m
od
el
for
m
ini
m
iz
in
g
the
num
ber
of
hand
ov
e
rs.
2.1.2.
MAD
M
b
as
ed
A
lg
orit
hm
Mult
iple
Attribu
te
Decisi
on
Ma
kin
g
(M
AD
M
)
base
d
al
gorithm
is
a
popu
la
r
deci
sion
m
et
ho
d.
MADM
m
et
h
od
s
ca
n
facil
it
at
e
the
need
of
m
ulti
-
crit
e
ria
so
luti
on
f
or
a
vo
i
ding
inap
pro
pr
ia
te
ha
ndove
r
decisi
on
as
hi
gh
li
ghts
i
n
[5]
.
I
n
researc
h
[1
3
]
,
Ma
al
ou
l
et
al
.
propo
sed
a
n
ef
fici
ent
hand
ov
e
r
de
ci
sion
al
gorithm
based
on
M
A
DM
m
et
ho
d.
T
he
pa
per
highli
ghts
so
m
e
MADM
m
e
tho
ds
s
uch
as
Sim
ple
Additi
ve
Weig
hting
(SA
W)
,
W
ei
ght
Pr
od
uct
Me
thod
(
WPM),
Tec
hn
i
qu
e
f
or
Order
P
ref
e
ren
ce
by
Si
m
i
la
rity
to
Id
ea
l
So
luti
on
(T
OPSIS
),
G
rey
Re
la
ti
on
al
A
naly
sis
(
GRA
),
Di
sta
nce
to
I
dea
l
Alte
rn
a
ti
ve
(
DIA),
VIKO
R
an
d
ELECT
RE.
T
he
y
pr
op
os
e
d
a
Ra
nk
i
ng
Sc
he
m
e
based
on
pro
vid
e
d
Q
oS
a
lgorit
hm
wh
er
e
con
si
der
i
ng
con
te
xt
-
awar
e
ness
in
the
evaluati
on
par
am
et
ers
li
ke
cov
e
rag
e
are
a,
RSSI,
a
vaila
ble
band
width,
delay
,
respon
s
e
tim
e,
j
it
te
r,
secu
rity
,
us
er
pre
fer
e
nc
e
and
c
os
t.
Thi
s
researc
h
com
par
e
d
the
pr
opos
e
d
al
gorithm
with
so
m
e
M
AD
M
m
et
ho
ds
m
entio
n
ea
rlie
r.
Re
su
lt
sh
ows
so
m
e
i
m
pr
ov
em
ent
in
handove
r
pe
rfor
m
ance;
howev
e
r,
the
al
gori
thm
com
plexity
m
a
y i
ncr
ease t
he han
dove
r dela
y an
d ov
e
r
head.
In
[1
4
]
,
a
netw
ork
sel
ect
io
n
ba
sed
on
c
onte
xt
awar
e
ness
se
rv
ic
es
has
bee
n
pr
opos
e
d
by
the
auth
ors.
This
co
ntext
-
awar
e
ness
al
gorithm
co
m
bin
es
the
f
uzz
y
m
at
he
m
a
ti
c
s
evaluati
on
and
T
OPSIS
-
MADM
appr
oach.
Fro
m
their
si
m
ul
at
ion
res
ults,
fu
zzy
m
at
he
m
at
ic
s
evaluat
ion
has
pe
rfo
rm
ed
bette
r
re
ga
rd
s
to
netw
ork
sel
ect
ion
s
peed.
The
TOP
S
IS
-
MA
DM
ap
proac
h
su
pp
or
ts
well
i
n
util
it
y
per
for
m
ance
m
easurem
ent.
Au
t
hors
in
[1
5
]
com
par
ed
ha
ndover
pe
rform
ance
for
four
MADM
base
d
ver
ti
cal
ha
ndover
al
gorithm
nam
ely
ME
W
(Multi
pl
ic
at
ive
Expon
e
nt W
ei
ghti
ng),
SAW,
TO
PS
IS
and
GR
A.
Ba
s
ed
on
their
sim
ulati
on
res
ult,
GRA
has
sli
gh
tl
y
hi
g
he
r
ba
ndwi
dth
an
d
lo
wer
delay
value
w
hen
c
oncer
ning
the
wei
gh
t
of
at
trib
utes
s
uch
as
band
width,
del
ay
,
j
it
te
r
a
nd
bi
t
error
rate
(B
ER).
In
[1
6
]
,
t
he
researc
h
pr
esented
the
ha
ndover
pe
rfo
r
m
anc
e
evaluati
on
on
seve
n
MA
DM
m
et
ho
ds
w
hich
are
SAW,
ME
W
,
T
OPSI
S,
GR
A,
E
LE
CTR
E,
V
IKO
R
and
WMC.
The
nu
m
erical
si
m
u
la
ti
on
is
desig
ne
d
in
MATL
A
B
so
ftwa
re
un
der
m
eans
of
diff
e
re
nt
app
li
cat
ion
scenari
os
s
uch
as
voic
e
,
data
and
cost
-
co
nst
rained
c
onnec
ti
on
.
T
he
ou
tc
om
e
sh
ow
s
t
ha
t
SAW
an
d
G
RA
ha
s
low
com
pu
ta
t
ion
al
com
plex
it
y
and
a
possible
so
luti
on
f
or
seam
less
ver
ti
cal
ha
ndover
.
Me
an
wh
il
e,
wh
e
n
c
onside
r
ing
t
he
ty
pes
of
ap
plica
ti
on
s
li
ke
vo
ic
e
a
nd
data,
VIK
O
R
an
d
ME
W
ar
e
the
be
st
MADM
decisi
on m
et
ho
ds
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
13
, N
o.
3
,
Ma
rc
h 201
9
:
1
143
–
1
151
1146
2.1.3.
Int
el
li
gence
b
as
ed
Algori
t
h
m
Trad
it
io
nal
R
SS
an
d
M
A
DM
base
d
al
gorithm
sti
l
l
hav
e
so
m
e
lim
it
a
ti
on
s
w
he
n
a
pp
li
ed
i
n
heter
og
e
ne
ou
s
networ
ks
.
Th
ese
m
et
ho
ds
c
reate
on
ly
sm
a
ll
diff
ere
nce
va
lues
in
ranki
ng
acc
ur
acy
[1
7
]
an
d
fixe
d
netw
ork
disco
ver
y
m
e
thod
[
18
]
.
He
nce,
com
bin
in
g
the
tradit
io
na
l
m
et
ho
ds
wi
th
intel
li
gen
ce
-
base
d
al
gorithm
m
ay
le
ads
to
bette
r
ha
ndover
pe
rfor
m
ance.
A
uthors
in
[
18
]
work
s
on
a
n
ty
pe
-
2
f
uzzy
log
ic
al
gorithm
that
sp
eci
fical
ly
focuse
d
on
ve
hic
ular
heter
ogen
eous
netw
orks.
Pin
k
et
al
.
in
[
1
9
]
pr
e
sente
d
a
f
uzzy
-
base
d
ve
rtic
al
hand
ov
e
r
de
ci
sion
al
gorith
m
wh
ic
h
ca
n
adap
ts
with
dev
ic
e
a
nd
netw
ork
cap
a
bili
ti
es.
The
resea
rch
has
bee
n
e
xpe
rim
ented
in
real
env
ir
on
m
ent
with
con
tr
ol
le
d
dev
ic
e
w
hi
ch
consi
der
i
ng
the
UMTS/
WLAN
netw
orks
.
T
he
eval
uation
f
ocused
on
Q
oS
an
d
res
ourc
e
co
nsum
ption
.
Re
su
lt
s
how
that
m
ob
il
i
ty
aff
ect
s
the
QoS
as
t
he
m
ob
il
e
node
cha
ng
es
the
path,
the
si
gn
a
l
stren
gth
m
ay
cha
ng
es
acco
r
dingly
.
In m
ini
m
isi
ng
the r
es
ource
consu
m
ption, the
algorit
hm
sli
gh
tl
y decreas
e t
he
m
axi
m
u
m
d
evice r
unt
i
m
e u
p.
Re
search
in
[2
0
]
pro
pose
d
a
new
hy
br
i
d
al
gorithm
te
chn
iqu
e
usi
ng
c
ombinati
on
of
A
BC
(Ant
Be
e
Colo
ny)
an
d
PSO
(
Partic
le
Sw
arm
Op
ti
m
iz
at
ion
in
the
process
of
sel
ect
ing
the
best
wireless
netw
ork.
Me
anwhil
e,
au
thors
in
[2
1
]
de
velo
ps
a
m
od
el
fo
r
hand
ov
e
r
decisi
on
al
go
rithm
by
app
ly
ing
hy
br
i
d
A
rtific
ia
l
Neural
Net
wor
k
(
A
NN).
B
oth
a
ppro
ac
hes
a
ble
to
reduce
t
he
c
os
t
an
d
pin
g
-
pong
ef
fects
in
ha
nd
ov
e
r.
Lu
o
e
t
al
.
[2
2
]
,
pro
posed
a
ne
w
ha
ndove
r
pre
dicti
on
al
gorithm
base
d
on
Hidd
en
Ma
r
kov
M
od
el
.
The
ex
pe
rim
ent
cond
ucted
f
or
wireless
net
wo
rk
in
offic
e
env
ir
onm
ent
by
con
sid
eri
ng
the
c
on
t
rol
le
d
sit
uation.
Hen
ce,
the
us
e
r’
s
m
obil
ity
in
pr
e
dicta
ble
an
d
RSS
value
is
acc
ur
a
te
ly
m
easur
ed
.
Sti
ll
,
the
research
has
li
m
it
ati
on
on
the
m
ob
il
ity
m
od
el
a
nd
em
issi
on
prob
a
bili
ty
need
to
know
and
le
ar
ns
on
each
ot
her
bette
r.
Ta
ble
1
s
how
the
su
m
m
ary of
curre
nt v
e
rtic
al
hando
ver
s
d
eci
s
ion
al
gorithm
s.
Table
1.
Su
m
m
ary o
f
Ver
ti
cal
Hand
ov
e
rs De
ci
sion
Algorith
m
s
RSS
MAD
M
Intellig
en
ce Algo
ri
th
m
s
Rev
iew pap
ers
Su
rvey
an
d
r
ev
ie
w
pap
ers on
m
o
b
ility
m
an
ag
e
m
en
t
[
5
]
-
[
1
0
]
,
[2
3
]
Bas
ic
-
RS
S
Receiv
ed
Sign
al St
reng
th
[
1
1
]
-
[1
2
]
,[
2
4
]
-
[
2
5
]
Ran
k
in
g
Sche
m
e
[1
3
]
ME
W
,
SA
W
,
TOP
SIS,
GRA
[1
5
]
SAW
,
M
E
W
,
TOP
SIS,
GRA,
EL
EC
TRE,
VI
KO
R,
W
MC
[1
6
]
NA
Fu
zzy
Log
ic (
FL
)
RSS, Fu
zz
y
[
2
6
]
Fu
zzy
[
2
7
]
MCDM
Fuzzy
-
AH
P
[
3
]
Fu
zzy
-
TOPS
IS
[1
4
]
Fu
zz
y
-
AHP
[
2
8
]
MCDMFuzz
y
-
TO
PSIS
[
2
9
]
Fu
zzy
MAD
M
[
3
0
]
MAD
M
–
GRA
[
3
1
]
Fu
zzy
-
MAD
M
[1
7
]
Ty
p
e
-
2
Fuz
zy
[
18
]
Fu
zzy
[
1
9
]
Artif
icial Neu
ral
N
etwo
rk
(ANN
)
NA
NA
ANN
[2
1
]
An
t Bee Co
lo
n
y
(
ABC
)
NA
NA
Clo
u
d
Assis
ted
Hand
o
v
er
(I
o
T)
–
ABC
[
3
2
]
Particle
Swa
m
Opt
i
m
izatio
n
(PSO)
NA
NA
PSO
-
ABC
[2
0
]
PSO
[
3
3
]
Mar
k
o
v
Ch
ain
MDP
[
3
4
]
NA
Hid
d
en
M
arko
v
[2
2
]
MI
H
& SDN
MI
H
[
3
5
]
VIKOR
[
3
6
]
NA
Note:
NA
–
Not
Applic
ab
le
Re
search
in
[
37]
fo
c
us
on
distribu
te
d
ha
ndov
e
r
an
d
m
ob
il
it
y
patte
rn
predict
ion
by
usi
ng
Ma
r
ko
v
theo
ry
an
d
sta
t
ist
ic
al
theor
y.
They
propose
d
an
al
gorithm
nam
ed
Patt
ern
Pr
e
dicti
on
an
d
Pas
sive
Ba
ndwi
dth
Ma
nag
em
ent
Algorithm
(3P
-
BM
A)
w
hich
ap
ply
Ma
r
kovian
Pr
e
dicti
on
sc
hem
e.
Ye
t,
this
researc
h
only
fo
c
us
e
d
on
guaran
te
es
t
he
se
rv
ic
e
c
on
ti
nuit
y
in
wirele
ss
c
el
lular
netw
or
ks
.
A
uthors
i
n
[38]
al
so
fo
c
use
d
on
m
ob
il
i
ty
pr
edi
ct
ion
f
or
wire
le
ss
cel
lular
ne
twork
,
in
op
t
i
m
izing
the
C
al
l
Ad
m
issi
on
Con
tr
ol
(C
A
C)
by
pro
po
se
d
t
he
I
n
-
a
dvance
Mul
ti
plexing
Ca
ll
Ad
m
issi
on
Co
ntr
ol
(IAM
-
C
AC)
sc
hem
e.
This
new
C
AC
schem
e
base
d
on
t
hr
es
ho
l
d
a
ppro
a
c
h
has
bee
n
pe
rfor
m
well
s
in
te
rm
s
of
Ca
ll
Bl
ock
in
g
P
r
obabili
ti
e
s/C
all
Dropp
i
ng
Pr
oba
bili
ti
es.
In
a
ddit
ion,
a
uthors
in
[39]
,
fo
c
us
e
d
on
li
nk
-
la
ye
r
inte
r
-
te
chnolo
gy
ha
ndovers
,
in
suppo
rting
t
he
lim
it
at
ion
s
of
us
e
r
m
ob
il
ity
rando
m
ness,
high
ha
ndov
er
ove
rh
ea
d,
op
ti
m
al
i
ty
req
uirem
ent.
The
pa
per
pro
po
se
d
a
ne
w
ha
ndover
de
ci
sion
m
aking
that
ex
plo
it
s
the
data
trac
es
of
u
ser
m
ob
il
it
y
in
real
m
ob
il
e
env
i
ronm
ents.
They
are
m
ining
the
data
f
r
om
Dar
t
m
ou
th
Coll
ege,
a
ca
m
pu
s
env
i
ron
m
ent
for
ass
um
ing
the
strong
re
gu
la
ri
ty
with
us
er
m
ob
il
i
ty
.
The
decisi
on
proc
e
ss
was
form
ulate
s
us
in
g
Ma
rko
v
decisi
on
process
(MDP)
a
nd
e
m
plo
ys
us
in
g
MADM
-
A
HP
appr
oach.
T
he
al
gorithm
has
bee
n
c
om
pared
with
ra
ndom
and
gr
ee
dy
al
gorithm
,
with
the
resu
lt
s
that
s
how
n
the
ne
w
al
gorithm
per
f
or
m
s
bette
r.
Howe
ver,
this
m
et
ho
d
requires
ex
te
nsi
ve
data t
r
aces
and sim
ulati
on
scena
rios.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Con
ce
ptual
mobili
ty
m
odel
of
vert
ic
al hand
ov
er d
eci
sio
n
i
n hetero
ge
neous netw
or
ks
(N
orakm
ar
Arbai
n)
1147
2.2 Pr
oposed
Metho
ds
The
pr
opos
e
d
scenari
os
f
or
ve
rtic
al
han
do
ve
r
are
sho
wn
i
n
Fig
ur
e
2.
A
few
sce
nar
i
os
will
be
set
up
for
colle
ct
ing
t
he
hum
an
m
ob
il
it
y
patte
rn
.
The
sam
ple
data
will
be
colle
ct
ed
from
sever
al
m
ob
il
e
nodes
in
ca
m
pu
s
net
work
e
nv
i
ronm
ent.
The
data
wi
ll
be
identif
ie
d
us
i
ng
Li
velab
netw
ork
m
e
asur
i
ng
s
ource
code
.
The
c
hosen
ac
cess
te
ch
no
l
ogie
s
will
be
a
WL
AN
(
W
i
Fi:
IEEE
802.1
1)
and
a
LT
E
ne
twork
.
T
he
de
sign
e
d
m
ob
il
i
ty
p
at
te
rn
m
od
el
w
il
l
be
inclu
ded in
verti
cal
h
an
dove
r
f
ram
ewo
r
k.
Figure
2
.
Pro
pose
d
sce
nar
i
os
for verti
cal
h
a
ndove
r
The
pr
opos
e
d
scenari
os
will
ref
e
r
to
resea
rc
h
in
[
40
]
as
s
how
n
in
Ta
ble
2.
The
c
ollec
te
d
data
will
be
m
easur
ed
an
d
analy
zed
f
or
res
pected
pa
r
a
m
et
ers
su
c
h
as
us
e
r’s
m
ove
m
ent
patte
rn,
sp
ee
d
a
nd
lo
cat
ion
.
Fr
om
this
real
data
traces,
a
pa
tt
ern
m
od
el
will
be
con
struct
and
us
e
d
in
la
tt
er
pr
op
os
e
d
al
gorithm
.
A
sa
m
ple
data
from
on
li
ne
data
base
na
m
ed
CR
A
W
D
AD
al
s
o
will
be
us
ed
as
ben
c
hm
ark
value
in
com
par
ison
w
it
h
the
real
data
trac
es
colle
ct
ed
i
n
Livel
a
b
Ne
tworks
.
I
n
a
ddit
ion
,
a
cr
owds
ourci
ng
a
pp
li
cat
io
n
na
m
ed
as
Op
e
nS
i
gn
al
w
il
l
be
use
d
f
or
c
ollec
ti
ng
a
nd
ret
rievin
g
netw
ork
pe
rfo
rm
ance
sta
ti
stics,
at
ta
r
get
l
ocati
on
resp
ect
ively
.
T
his
ap
plica
ti
on
can
rec
ords
t
he
pe
rfo
rm
ance
sta
ti
sti
cs
of
connecti
ons
be
tween
m
ob
il
e
nodes
with
the
e
Node
B
(LTE
)
a
nd W
L
AN
(
W
i
Fi)
.
It
al
s
o
pro
vide
s
an
ag
gregat
ed
sta
ti
sti
cs
of
netw
ork
c
over
age
that
pr
ese
nted
in
t
he
for
m
o
f
c
ove
rag
e
m
aps.
Table
2.
Scena
rios f
or
C
ollec
ti
ng
O
nline M
obil
it
y Pat
te
rn
Scen
arios
Netwo
rk Sele
ctio
n
Data Tech
n
o
lo
g
ies
Ap
p
licatio
n
s
Cas
e 1
LT
E
to
W
iFi
Gen
eral/Co
n
v
en
tio
n
al
W
eb
Brows
e
r,
Vo
ice,
SN
S
Cas
e 2
W
iFi
to
LT
E
Gen
eral/Co
n
v
en
tio
n
al
Cas
e 3
LT
E
to
W
iFi
Strea
m
in
g
Vid
eo
,
Live
TV
Cas
e 4
W
iFi to
LT
E
Strea
m
in
g
3.
RESU
LT
S
A
ND AN
ALYSIS
3.1.
Rand
om Mo
b
il
ity
M
od
el
and Al
go
ri
th
m
In
recent
ye
ars
,
seve
ral
resea
rch
es
be
gan
focusin
g
on
the
m
ob
il
i
ty
issue
as
m
ob
il
e
node
m
ay
hav
e
diff
e
re
nt
m
ob
il
it
y
sp
eed
an
d
m
ov
em
ent
directi
on.
He
nce
,
co
ns
ide
rin
g
the
ra
ndom
ness
of
m
ob
il
e
node
’s
m
ov
e
m
ents
and
va
riet
y
of
m
ob
il
i
ty
sp
eed
is
sti
ll
a
chall
eng
in
g
ta
sk
in
ver
ti
cal
han
do
ver
m
anag
em
ent.
Com
m
on
ly
,
m
ai
n
pa
ram
et
ers
in
ve
rtic
al
handove
r
decisi
on
are
bas
ed
on
netw
ork
c
onditi
on
s
s
uc
h
as
re
cei
ved
sign
al
str
e
ng
t
h
and
a
vaila
ble
ba
ndwidt
h.
Be
si
des,
at
the
us
er
side
(m
ob
il
e
no
de
),
se
ver
al
c
onditi
ons
al
so
bei
ng
consi
der
e
d
li
ke
m
ob
il
e
ener
gy
,
co
nnect
ion
c
os
t
an
d
us
e
r
pr
efere
nces.
Tab
le
3
s
how
the
par
am
et
er
m
et
rics
for
Ra
ndom
Mob
il
it
y
Para
m
et
er
s
in
Mo
bili
ty
-
awar
e
ness
of
Ver
ti
cal
Ha
ndov
e
r
Ma
na
ge
m
ent
fo
r
LTE
-
W
L
A
N
netw
orks. M
ea
nwhile
, a blo
ck
d
ia
gram
in
Fig
ur
e
3 sh
ow the
Fu
zzy
-
T
opsis
with M
ob
il
it
y
-
awar
e
ness
m
od
el
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
13
, N
o.
3
,
Ma
rc
h 201
9
:
1
143
–
1
151
1148
Table
3.
Ra
ndom
Mob
il
ity Para
m
et
ers
in Mo
bili
ty
-
awar
e
ne
ss of
Ver
ti
cal
Hand
ov
e
r
Ma
na
gem
ent
Figure
3
.
F
uzz
y
-
TOPSIS
bl
oc
k diag
ram
w
it
h
ra
ndom
m
ob
il
it
y
-
awar
ene
ss
3.2.
Net
w
or
k Per
f
orm
an
ce
and
Availabil
it
y
Figure
4
s
ho
w
the
net
wor
k
perform
ance
sta
ti
sti
cs,
at
ta
rg
et
ed
locat
ion
res
pecti
vel
y
base
d
on
a
crow
ds
our
ci
ng
ap
plica
ti
on
nam
ed
as
Op
e
nSi
gn
al
.
T
his
a
pp
li
cat
io
n
ca
n
rec
ord
the
pe
rfor
m
ance
sta
ti
sti
cs
of
connecti
ons
be
tween m
ob
il
e
nodes wit
h t
he e
NodeB
(LTE
)
and
WL
AN (
W
i
Fi).
I
t al
s
o p
rovides
an ag
gregate
d
sta
ti
sti
cs o
f
ne
twork
c
ov
e
rag
e
that prese
nted
in the f
or
m
o
f c
ov
e
ra
ge
m
aps.
Figure
4
.
Net
w
ork per
f
or
m
ance sp
ee
d
te
st
hi
story a
nd cov
e
rag
e
m
aps
fro
m
O
pen
Sig
nal
Apps
3.3.
Concept
ua
l
Model
of
M
ob
il
ity
-
aw
areness
Vert
ic
al H
an
d
ov
er
Man
age
ment
Re
la
ti
vely
,
a
pro
po
se
d
ne
w
ve
rtic
al
ha
ndov
e
r
decisi
on
that
e
xp
l
oits
m
ob
il
it
y
-
awar
enes
s
m
od
el
sh
oul
d
c
on
si
de
r
the
m
ob
il
e n
ode’s
spee
d an
d patt
ern ra
ndom
ness
as
s
how
n i
n
Fig
ure
5.
Para
m
eter
Ind
icatio
n
Value
Para
m
eter
Ind
icatio
n
Value
Receiv
ed
Sign
al St
reng
th
Rss
MT
Mov
e Out
MT
m
o
Av
ailab
le
Ban
d
wid
th
bd
Sp
eed
s(n
)
MT
Mov
e I
n
MT
m
i
Direction
s
d
(n)
Input
Para
m
et
ers
Fuzzi
fi
er
MADM
-
TOPS
IS
Defuz
zi
f
ie
r
Mobili
t
y
-
aw
are
n
ess
Handove
r
De
ci
si
on
Mobili
t
y
Speed
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Con
ce
ptual
mobili
ty
m
odel
of
vert
ic
al hand
ov
er d
eci
sio
n
i
n hetero
ge
neous netw
or
ks
(N
orakm
ar
Arbai
n)
1149
Figure
5
.
Mo
bili
ty
-
awar
eness
ver
ti
cal
h
a
ndover
m
anag
em
e
nt
4.
CONCL
US
I
O
N
This
pa
per
pre
sents
t
he
c
on
c
eptual
m
ob
il
it
y
m
od
el
of
ve
rtic
al
hand
ove
r
decisi
on
in
he
te
rogen
e
ous
netw
ork.
He
nc
e,
m
ai
ntaining
net
work
c
onne
ct
ivit
y
in
het
eroge
neous
ne
twork
re
quires
ast
oundin
g
ef
fort
in
academ
ia
an
d
i
ndus
trie
s
. Rapi
d
m
ob
il
it
y has
beco
m
e a b
ig
ge
r
c
halle
ng
e
in faci
li
ta
ti
ng
the
serv
ic
e c
onti
nu
it
y t
o
m
ob
il
e
us
ers.
Fu
rt
her
m
or
e,
va
riet
y
of
netw
ork
acce
ss
su
c
h
as
W
iFi
,
W
i
MAX,
3GPP
a
nd
LT
E
al
so
c
reates
chao
ti
c
m
ob
il
e
env
ir
onm
ents
wh
e
n
the
handove
r
proces
s
no
t
bein
g
co
nducte
d
preci
se
ly
.
Hen
ce,
t
his
pap
e
r
rev
ie
ws
seve
ra
l
cur
re
nt
resear
ches
on
ve
rtic
al
han
do
ver
m
a
nag
em
ent
archi
te
ct
ur
e.
In
a
ddit
ion
,
the
li
te
r
at
ur
es
on
ver
ti
cal
handove
r
decisi
on
al
go
rithm
al
so
being
a
naly
se
and
cat
e
gorise
d
into
th
ree
se
ct
ion
s
w
hich
a
re
RSS
Ba
sed
Al
gorit
hm
,
MADM
Ba
sed
Algorit
hm
and
In
te
ll
igence
Ba
se
d
Algorithm
.
Most
c
urren
t
re
searche
s
fo
c
us
e
d
on
c
om
bin
ing
t
he
RSS
a
nd
MADM
a
ppr
oac
hes
with
intel
li
gen
ce
al
go
rithm
li
ke
Fu
zz
y
log
ic
,
to
enh
a
nce
th
e
handove
r
pe
rfor
m
ance
su
c
h
as
ov
e
r
head
and
delay
.
F
r
om
the
relat
ed
rev
ie
ws
,
this
pap
e
r
pro
po
se
d
a
m
ob
il
it
y
-
awar
e
ne
ss
ver
ti
cal
handove
r
m
anag
em
ent
that
encom
passes
m
ob
ilit
y
sp
eed
and
patte
rn
as
an
im
po
rtan
t
at
tribu
te
duri
ng
t
he
ha
ndov
er
decisi
on
pro
cess.
He
nce,
w
e
hope
the
fu
t
ur
e
ver
ti
cal
ha
ndover
m
anag
em
ent can p
r
ovide m
ore seam
le
ss n
et
work c
onnecti
vity
.
ACKN
OWLE
DGE
MENTS
This
researc
h
was
fina
ncial
ly
su
pport
ed
by
the
Mi
nistry
of
Higher
E
ducat
ion
(M
O
H
E)
thr
ough
Fund
am
ental
Resea
rch G
ra
nt
Schem
e (F
RG
S)
, 6
00
-
RM
I/F
RGS
5/3(0
006/
2016)
.
Evaluation Warning : The document was created with Spire.PDF for Python.
IS
S
N
:
2502
-
4752
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci,
Vo
l.
13
, N
o.
3
,
Ma
rc
h 201
9
:
1
143
–
1
151
1150
REFERE
NCE
S
[1]
N.
Om
heni
,
et
al
.
,
“
Sm
art
m
obil
it
y
m
ana
gement
i
n
5G
het
ero
geneous
net
works
,
”
IET
Net
works
,
vol
/i
ss
ue:
7
(
3
)
,
pp
.
119
–
128,
2018
.
[2]
A.
Gani,
et
al
.
,
“
A
rev
ie
w
on
int
erworki
ng
an
d
m
obil
ity
t
ec
h
nique
s
for
sea
m
le
ss
connect
iv
ity
in
m
obil
e
cl
o
ud
computing,
”
J
.
Net
w.
Comput.
Appl
.
, v
o
l. 43, p
p.
84
–
102
,
2014
.
[3]
M.
Driss
i,
et
a
l
.
,
“
A
Multi
-
Crit
er
ia
Dec
ision
Fra
m
ework
for
netw
ork
sele
ction
o
ver
LTE
and
W
LAN,”
Eng
.
Ap
pl.
Arti
f. Intell.
,
vol
.
66,
pp.
113
–
127
,
2017
.
[4]
N.
K.
Panigr
ah
y
and
S.
C
.
Ghos
h,
“
Anal
y
zi
ng
t
he
Eff
ec
t
of
Sof
t
Handove
r
on
Handove
r
Lo
ad
Condit
ion,”
I
EEE
Tr
ans.
Ve
h.
Tec
hnol.
,
vol
/
issue:
67
(
4
)
,
pp
.
3612
–
3624,
2018
.
[5]
M.
Kass
ar,
et
a
l
.
,
“
An
over
vi
e
w
of
ver
ti
c
al
ha
ndover
de
ci
sion
strategie
s
in
h
e
te
roge
n
eous
wir
el
ess
ne
tworks,”
Comput.
Comm
un.
,
vol
/i
ss
ue
:
31
(
10
)
,
pp
.
2607
–
2
620,
2008
.
[6]
S.
Ferna
ndes
a
nd
A.
Karm
ou
ch,
“
Vert
ic
a
l
Mobili
t
y
Man
a
gement
Archi
tectur
es
in
W
irel
ess
Networks:
A
Com
pre
hensive
Surve
y
and
Futu
re
Dire
ctions,”
I
EE
E
Comm
un.
S
urv.
Tutor
ial
s
,
v
ol
/i
ss
ue:
14
(
1
)
,
p
p.
45
–
63
,
2012
.
[7]
J.
M
.
B
arj
a
,
et
al
,
“
An
over
vi
e
w
of
ver
t
ical
h
andove
r
techniq
ues:
Algori
thms
,
proto
col
s
and
tool
s,
”
Comput
.
Comm
un.
,
vol
/
issue:
34
(
8
)
,
pp
.
9
85
–
997,
2011
.
[8]
G.
Gódor,
et
al
.
,
“
A
survey
of
h
andove
r
m
ana
g
e
m
ent
in
LTE
-
ba
sed
m
ult
i
-
ti
e
r
fe
m
toc
el
l
n
et
work
s:
Requi
rement
s,
cha
l
le
nges
and
s
olut
ions,
”
Comp
ut.
N
et
works
,
vo
l.
76
,
pp
.
17
–
41
,
2015.
[9]
D.
Xena
kis
,
e
t
a
l
.
,
“
Mobili
t
y
m
a
nage
m
ent
for
fe
m
toc
el
ls
in
L
TE
-
adva
nc
ed:
K
e
y
aspe
ct
s
and
sur
v
e
y
of
h
andove
r
dec
ision
al
gor
it
h
m
s,”
IEEE
Com
mun.
Surv.
Tutor
ial
s
,
vo
l
/i
ss
ue:
16
(
1
)
,
pp
.
64
–
91
,
2014.
[10]
X.
Yan,
et
al
.
,
“
A
surve
y
of
ver
ti
c
al
handove
r
dec
ision
al
g
orit
h
m
s
in
Fourth
G
ene
ra
ti
on
hetero
gene
ous
wire
le
s
s
net
works
,
”
Comput.
N
et
works
,
v
ol
/i
ss
ue:
54
(
11
)
,
pp.
1848
–
1863
,
2010.
[11]
S.
B.
Johns
on,
et
al
.
,
“
An
Optimize
d
Algori
thm
for
Vert
ical
Hand
off
in
Hete
roge
n
eous
W
ire
le
ss
Networks,”
IE
EE
Conf.
In
f. Commun.
Techno
l.
,
pp
.
1206
–
1210,
201
3.
[12]
Z.
Ning,
et
a
l
.
,
“
Markov
-
base
d
v
ert
i
ca
l
h
andof
f
d
ec
ision
al
gori
th
m
s
in
het
ero
gen
eous
wire
le
ss
ne
tworks,”
Comput.
El
e
ct
r.
Eng.
,
vo
l
/i
ss
ue:
40
(
2
)
,
pp
.
456
–
472,
2014.
[13]
S.
Maa
loul,
e
t
a
l.
,
“
An
eff
ic
i
ent
handove
r
d
ec
isi
on
m
aki
ng
for
h
et
ero
g
ene
ous
wi
rel
ess
conn
ec
t
ivit
y
m
an
age
m
ent,
”
2013
21st
Int
.
C
onf.
So
ft
ware
,
T
el
e
commun.
Co
mput.
Ne
tworks
-
(
Soft
COM 2013)
,
pp.
1
–
8
,
201
3.
[14]
L.
U.
O.
L
ij
u
an,
et
a
l
.
,
“
Network
Sel
ec
t
io
n
Based
on
C
onte
xt
-
Aw
are
n
e
s
s
Servic
es,
”
China
Comm
un.
,
pp.
50
–
56
,
2013
.
[15]
E.
S
.
Nava
rro
a
nd
V.
W
.
W
ong
,
“
Com
par
ison
bet
wee
n
Vert
i
cal
Handoff
Dec
isio
n
Algorit
hm
s
fo
r
Hete
rog
ene
ous
W
ire
le
ss
Networ
k,
”
Ve
h
.
Te
chnol.
Conf
.
,
vol
.
2
,
p
p.
947
–
951
,
200
6.
[16]
E.
S
.
Nava
rro
,
et
al
.
,
“
Eva
lu
at
i
on
of
Vert
ic
a
l
Handoff
Dec
isi
on
Algoright
m
s
Based
on
MA
DM
Methods
for
Hete
roge
n
eous
W
ire
le
ss
Networ
ks,”
J
.
Appl. Res
.
Techno
l.
,
vo
l. 1
0,
pp
.
534
–
548
,
2012.
[17]
H.
Jin,
et
a
l
.
,
“
I
FM
ADM:
An
eff
ic
i
ent
n
et
work
s
el
e
ct
ion
al
gor
it
h
m
in
int
egr
ated
het
ero
g
ene
ous
w
ire
l
ess
net
work,
”
2011
4th
I
EEE
I
nt.
Con
f. B
roadb
and
Net
w
.
Mu
lt
i
med.
Te
chnol
.
,
p
p.
511
–
516
,
201
1.
[18]
B.
Ma
and
X.
Li
ao,
“
Speed
-
ad
a
pti
ve
ver
ti
c
al
ha
ndoff
al
gorit
hm
base
d
on
fuz
z
y
logi
c
in
vehicul
ar
het
ero
g
ene
ou
s
net
works
,
”
2012
9th
In
t. Conf
.
F
uzz
y
Syst. Knowl.
Discov
.
,
pp
.
37
1
–
375,
2012
.
[19]
M.
Pink,
e
t
al
.
,
“
Towa
rds
a
sea
m
le
ss
m
obil
ity
s
olut
ion
for
the
r
ea
l
world
:
Hand
over
decision,
”
2012
Int.
S
ymp.
Wirel
.
Comm
un.
Syst.
,
pp
.
651
–
6
55,
2012
.
[20]
S.
Goudarz
i,
e
t
al
.
,
“
ABC
-
PS
O
for
ver
ti
c
al
hand
over
in
het
e
roge
neous
wire
le
ss
net
works
,
”
Neuro
computi
ng
,
vol
.
256,
pp
.
63
–
81
,
2017.
[21]
A.
M.
Aibinu
,
e
t
al
.
,
“
Deve
lop
m
ent
of
h
y
brid
art
if
ic
i
al
int
e
ll
ig
ent
b
ase
d
hando
ver
de
ci
sion
al
g
orit
hm
,
”
Eng.
S
ci
.
Tec
hnol. an
In
t.
J.
,
vol
/i
ss
ue:
20
(
2
)
,
pp
.
381
–
390
,
2017.
[22]
Y.
Luo,
et
al
.
,
“
Handove
r
pre
diction
for
wire
le
ss
net
works
in
off
ic
e
environm
ents
using
Hidden
Markov
Model,
”
2013
IFI
P
Wire
l
.
Days
,
pp.
1
–
7,
2013.
[23]
J.
W
u
and
P.
Fan,
“
A
Survey
on
High
Mobili
t
y
W
ir
eless
Com
m
unic
at
ions:
Chal
l
enge
s,
Opportunit
ie
s
a
nd
Soluti
ons,”
IEEE
A
ccess
,
vol
.
4
,
pp.
450
–
476,
20
16.
[24]
T.
Ali
and
M.
S
aqui
b,
“
Perform
anc
e
Eva
lu
at
ion
of
W
LAN/Cel
lul
ar
Media
Ac
c
ess
for
Mobile
Voice
Us
ers
under
Random Mobili
t
y
Mod
el
s,”
IE
E
E
Tr
ans.
Wirel. Com
mun.
,
vol
/
issue:
10
(
10
)
,
pp
.
3241
–
3255,
201
1.
[25]
T.
Ali
and
M
.
Saquib,
“
Anal
y
ti
c
al
fr
amework
for
W
LAN
-
ce
ll
ul
a
r
voice
handov
e
r
evalua
t
ion,”
IE
EE
Tr
ans.
Mob
.
Comput.
,
vo
l
/i
ss
ue:
12
(
3
)
,
pp
.
44
7
–
460,
2013
.
[26]
X.
S.
and
J.
Z.
L
.
Zha
ng
,
L
.
Ge
,
“
Fuzz
y
Log
ic
b
a
sed
Vert
ical
Han
dover
Algorit
hm
,
”
017
26th
Wire
le
ss
and
Optic
al
Comm
unic
ati
on
Confe
renc
e
(
WOCC)
,
2017.
[27]
J.
M.
Kang
,
et
al
.
,
“
Autonom
ic
per
sonal
ized
h
andove
r
d
ecisio
ns
for
m
obil
e
services
in
heter
ogene
ous
wire
les
s
net
works
,
”
Comput.
N
et
works
,
v
ol
/i
ss
ue:
55
(
7
)
,
p
p.
1520
–
1
532
,
2
011.
[28]
R.
K.
Go
y
a
l,
e
t
al
.
,
“
Th
e
uti
l
i
t
y
b
ase
d
non
-
linear
fuz
z
y
AH
P
opti
m
iz
at
ion
m
odel
for
net
work
sele
ction
in
het
ero
g
ene
ous w
ire
l
ess ne
tworks,
”
App
l. Sof
t
Comput.
J.
,
2016
.
[29]
Y.
S.
Hus
sein,
et
al
.
,
“
A
novel
c
el
l
-
se
le
c
ti
on
optim
iz
at
ion
handov
er
for
long
-
te
rm
evol
uti
on
(
LT
E)
m
ac
roc
el
lusi
n
g
fuz
z
y
TOPS
IS,”
Comput.
Comm
un.
,
vol. 73, pp. 2
2
–
33,
2016
.
[30]
A.
B
.
Zi
neb
,
e
t
al
.
,
“
An
Enha
nce
d
Ver
ti
c
al
Handove
r
Base
d
on
Fuzz
y
In
fer
ence
MA
DM
Approac
h
for
Hete
roge
n
eous
Networks,”
Arab.
J
.
S
ci. Eng.
,
v
ol
/i
ss
ue:
42
(
8
)
,
p
p.
3263
–
3274
,
2
017.
[31]
M.
Mansouri
a
nd
C.
Le
gh
ris,
“
A
bat
te
r
y
l
eve
l
awa
re
MA
DM
combinat
ion
fo
r
the
ver
tical
h
andove
r
de
ci
sio
n
m
aki
ng,
”
2017
1
3th
Int
.
W
irel
.
C
omm
un.
Mob.
C
omput.
Conf
.
IW
CMC
2017
,
pp
.
1448
–
1452,
201
7.
[32]
D.
Li
,
et
al
.
,
“
A
cl
oud
-
assisted
handove
r
op
ti
m
i
za
t
ion
strateg
y
f
or
m
obil
e
nodes
in
industri
a
l
wire
le
ss
net
works
,
”
Comput.
Ne
tworks
,
vol
.
128
,
201
7.
Evaluation Warning : The document was created with Spire.PDF for Python.
Ind
on
esi
a
n
J
E
le
c Eng &
Co
m
p
Sci
IS
S
N:
25
02
-
4752
Con
ce
ptual
mobili
ty
m
odel
of
vert
ic
al hand
ov
er d
eci
sio
n
i
n hetero
ge
neous netw
or
ks
(N
orakm
ar
Arbai
n)
1151
[33]
K.
Ahuja,
et
al
.
,
“
Parti
cl
e
sw
arm
opti
m
iz
at
ion
base
d
net
work
select
ion
in
heter
ogene
ous
wire
les
s
envi
ronm
ent
,
”
Opt.
-
Int. J.
Lig
ht
E
le
c
tron Opt
.
,
vol
/
issue:
125
(
1
)
,
pp
.
214
–
219
,
2014.
[34]
S.
Za
ng
,
e
t
al
.
,
“
Mobili
t
y
h
andov
er
opti
m
i
za
t
ion
i
n
m
il
li
m
et
e
r
wa
ve
he
te
roge
n
eou
s
net
works
,
”
in
1
7th
Inte
rnat
ional
Symposium on
C
omm
unic
ati
ons
and
Information T
ec
hnologies,
IS
CIT 2017
,
pp
.
1
–
6,
2017
.
[35]
W
.
Mansouri,
et
al
.
,
“
Cross
lay
e
r
arc
hitect
u
re
wi
th
int
egr
at
ed
MI
H
in
het
ero
geneous
wire
le
ss
net
works
,
”
Comput.
Net
works
,
vo
l. 1
27,
pp
.
126
–
137
,
2017.
[36]
X.
Li
,
et
al
.
,
“
A nove
l
opti
m
ized vertical
h
andover fra
m
ework
for
sea
m
le
ss
net
working
int
egr
at
ion in
c
y
ber
-
ena
bl
ed
s
y
stems
,
”
F
utur.
Gene
r.
Comput.
Syst.
,
vo
l. 79, p
p.
417
–
430
,
201
6.
[37]
P.
Fazi
o,
et
al
.
,
“
A
distri
bute
d
hand
-
over
m
ana
g
ement
and
pa
tt
er
n
pre
dic
t
ion
a
lg
orit
hm
for
wire
l
ess
net
works
W
i
th
m
obil
e
hosts,”
9
th
Inte
rnationa
l
Wirel
ess
Comm
unic
ati
ons
and
Mobil
e
Computi
ng
Confe
ren
c
e,
I
WCMC
2013
,
pp.
294
–
298
,
2013
.
[38]
P.
Fazi
o,
e
t
al
.
,
“
Mobili
t
y
pr
edi
c
ti
on
in
wire
le
ss
ce
l
lul
ar
n
et
work
s
for
the
opti
m
izati
on
of
Ca
ll
Adm
ission
Control
sche
m
es,
”
2014
IEE
E
27th
Can.
Conf.
El
e
ct
r.
Co
mput.
Eng
.
,
pp.
1
–
5,
2014
.
[39]
Y.
Zhu,
et
al
.
,
“
Expl
oit
ing
m
o
bil
ity
p
at
t
ern
s
f
or
int
er
-
te
chno
l
og
y
handov
er
i
n
m
obil
e
envi
ro
nm
ent
s,”
Comput.
Comm
un.
,
vol
/
issue:
36
(
2
)
,
pp
.
2
03
–
210,
2013
.
[40]
M.
La
hb
y
,
e
t
al
.
,
“
An
enha
nce
d
-
TOPS
IS
base
d
n
et
work
sele
c
ti
on
te
chni
qu
e
for
ne
xt
gene
ra
ti
on
wi
rel
ess
net
works
,
”
Int
.
Con
f. Tel
e
co
mm
un.
,
pp.
1
–
5,
2013.
BIOGR
AP
HI
ES OF
A
UTH
ORS
Norakm
ar
Arbai
n
@
Sulai
m
an
r
ec
e
ive
d
h
er
B.
S
c.
d
egr
ee
in
Inf
orm
at
ion
S
y
st
e
m
Engi
nee
ring
(Com
pute
r
Scie
n
ce
)
in
2005
,
her
M.Sc.
in
Re
al
T
i
m
e
Software
En
gine
er
ing
in
200
9,
while
she
i
s
cur
ren
t
l
y
pursue
d
her
Ph.D.
at
t
he
Facult
y
of
C
om
pute
r
Scie
nc
e
and
Math
ematics,
at
Univer
sit
i
Te
knolog
i
MA
R
A,
Malay
sia
.
Norakm
ar
is
al
so
a
le
c
ture
r
a
t
Facult
y
of
El
e
ct
ri
cal
Engi
ne
eri
ng
of
Univer
siti
T
ekn
ologi
MA
RA.
Her
cur
ren
t
res
ea
rch
in
te
r
ests
are
in
the
ar
eas
of
m
obil
e
net
working,
c
lou
d
computing and
software
engi
n
e
eri
ng.
Dr.
Zol
ida
h
Kas
ira
n,
r
ec
e
ive
d
h
er
B.
Sc.
deg
ree
in
Com
pute
r
Inform
a
ti
on
S
y
ste
m
in
1991,
her
M.Sc.
in
Inform
at
ion
T
ec
hnolog
y
in
2001
and
Ph.D.
in
Inform
at
ion
T
ec
hnolog
y
,
at
UiTM
in
2012.
She
is
cur
r
ent
l
y
a
seni
or
l
e
ct
ure
r
at
Facult
y
of
Com
pute
r
Sc
ie
nc
e
and
Mathem
at
ic
s,
Ui
TM.
He
r
cur
ren
t
res
ea
rch
in
te
r
ests
are
rese
a
rch
in
m
obil
e
net
works
and
computing;
and
web
appl
i
ca
t
ion.
Evaluation Warning : The document was created with Spire.PDF for Python.