TELKOM
NIKA Indonesia
n
Journal of
Electrical En
gineering
Vol. 12, No. 10, Octobe
r 20
14, pp. 7330
~ 733
6
DOI: 10.115
9
1
/telkomni
ka.
v
12i8.518
5
7330
Re
cei
v
ed
No
vem
ber 2
1
, 2013; Re
vi
sed
Jul
y
4, 2014;
Acce
pted Jul
y
28, 201
4
Model Construction and Simulation of Weighted Industrial
Cluster Complex Network
Lijun Wang*
1
,
Lei Wen
2
Dep
a
rtment of Econom
ics an
d Mana
gem
ent
, North Chin
a Electric Po
w
e
r Univers
i
t
y
, Ba
o
d
in
g, Chin
a;
*Corres
p
o
ndi
n
g
auth
o
r,
e-mai
l
: w
a
ng
lij
un_
03
25@
163.com, w
e
nl
ei
031
2@s
ohu.com
A
b
st
r
a
ct
In this p
aper,
w
e
construct a
w
e
ighted
in
du
strial cl
uster c
o
mpl
e
x netw
o
r
k
mo
de
l acc
o
rdin
g to
prefere
n
tial
att
a
ch
me
nt an
d
l
o
cal
w
o
r
l
d
me
chan
is
m, in
w
h
ich t
he
lo
gi
sti
cs a
m
o
unt of
busi
ness c
ont
act
betw
een on
e enterpr
ise an
d
others
is
def
ine
d
as
e
dge
w
e
ight. W
e
simulat
e
a
n
d
ana
ly
z
e
statist
i
c
character
i
stics of
the ind
u
stria
l
clust
e
r mo
del
such
as
de
gr
ee
distrib
u
tio
n
, po
int stren
g
th
, poi
nt stren
g
th-
poi
nt strength
correlati
on, a
n
d
relati
ons
hip
w
i
th degre
e
a
nd p
o
int stren
g
th. T
he si
mul
a
tion r
e
sults s
how
that the degr
e
e
and
poi
nt st
rength distrib
u
ti
on
are accord w
i
th
pow
er
law
distributi
on, th
e poi
nt strengt
h-
poi
nt strength
correlati
on is n
egativ
e, and th
e relati
ons
h
i
p
w
i
th degree
an
d poi
nt strengt
h is strongly l
i
n
ear.
Ke
y
w
ords
:
in
d
u
strial cl
uster, w
e
ighted c
o
mp
lex netw
o
rk, de
gree, po
int stre
ngth
Co
p
y
rig
h
t
©
2014 In
stitu
t
e o
f
Ad
van
ced
En
g
i
n
eerin
g and
Scien
ce. All
rig
h
t
s reser
ve
d
.
1. Introduc
tion
The
com
p
lex
netwo
rk theo
ry is an
impo
rt
ant ba
sis for
a lot of n
e
two
r
k
study i
n
re
al life,
with the
grad
ual d
eepe
nin
g
of the
com
p
lex net
wo
rk
study, the
we
ight differe
nce of
con
n
e
c
tion
betwe
en no
d
e
s in the net
work i
s
incre
a
sin
g
ly bei
ng
recogni
zed.
In the literatu
r
e [1] Barab
a
s
i
and Albe
rt m
entione
d that
weig
hted co
mplex networ
ks i
s
a
n
impo
rtant directio
n
of the com
p
l
e
x
netwo
rk
re
se
arch. In literature [2] S.
Boccale
tti, V. Latora, Y.
More
no, M.Chavez an
d D.-
U.Hwan
g de
scribe
d the e
ffects of wei
ghted net
wo
rk theo
ry for compl
e
x net
works. Wei
g
hted
netwo
rk ha
s
become
an i
m
porta
nt pa
rt
of the
fiel
d o
f
com
p
lex n
e
tworks. Ma
ny
actual
net
works
were re
se
arched a
s
a wei
ghted net
wo
rk, su
ch a
s
bi
ologi
cal net
works [3-4], so
cial net
wo
rks [5
-
6], economi
c
netwo
rks [7-8
], technology
netwo
rk
[9], th
e transpo
rt network[10] an
d so on.
Enterpri
se
s i
n
the indu
st
rial cl
uste
r n
e
tw
ork e
s
tab
lish the a
ppropriate
co
nn
ection
s
according to
busine
s
s contact
s
. Logi
stics amou
nt between e
n
terpri
se
s can
distingui
sh
the
st
ren
g
t
h
of
b
u
sin
e
s
s
co
nt
act
s
,
it
sh
ow
s t
h
e
co-o
pe
ration
betwe
en the
ente
r
prises,
and
can
reflect the
strength of dep
ende
ncy bet
wee
n
the ent
erp
r
ise and o
t
her ente
r
p
r
ises. In this pa
per
we defin
e the
logistics am
ount betw
een
enterp
r
ises i
n
the indu
stri
al clu
s
ter a
s
the edg
e wei
g
ht,
build
a
weig
h
t
ed ind
u
stri
al
clu
s
ter compl
e
x network m
odel, a
nd
ana
lyze
statistics ch
aracte
risti
c
s
of the wei
ght
ed ind
u
stri
al
clu
s
ter
com
p
l
e
x netwo
rk
such
as
deg
re
e, point stren
g
th, relation
ship
of degre
e
an
d point stre
ng
th by simulation.
Figure 1. The
Industrial
Clu
s
ter
Netwo
r
k
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Model Con
s
truction a
nd Sim
u
lation of Weighted Ind
u
strial Clu
s
ter
Com
p
lex… (Lij
un Wa
ng)
7331
Figure 1 exp
r
esse
s ind
u
st
rial clu
s
te
r n
e
tw
ork with
30 nod
es. T
he rou
n
d
s
re
pre
s
ent
upstream
ent
erp
r
ises, th
e
squ
a
re
s
re
prese
n
t mid
s
tre
a
m ente
r
p
r
ises
and
the tri
angle
s
rep
r
e
s
ent
downstream enterp
r
i
s
e
s
.
There
a
r
e 10 numbe
rs
respec
tively of t
he three
kind
s of e
n
terpri
ses i
n
the figure. Th
e left is hybrid and the rig
h
t
is layered b
a
se
d on no
de
types.
2. Model Co
nstru
c
tion o
f
Weighted Industrial Clu
s
ter Com
p
lex
Net
w
o
r
k
2.1. Definitio
n
of
w
e
igh
t
e
d
industrial
cluster c
o
mplex net
w
o
r
k
(1) Nod
e
defi
n
ition
The nod
e
d
e
f
inition
of we
ighted
i
ndu
st
rial clu
s
ter complex network
i
s
v
i
=(no,
Cla
s
s,
Orde
r)
, in
which
no
stan
da
rds fo
r the n
o
de num
ber,
Cla
s
s
stan
da
rds fo
r the n
o
de cl
ass, in the
indu
strial
clu
s
ter all
node
s are divided
into
upstre
a
m, middle
and do
wn
stream ente
r
p
r
i
s
e
s
.
Orde
r
sta
nda
rds fo
r the bu
sine
ss order i
n
the indu
st
ri
al clu
s
ter. Th
e ord
e
r of mi
ddle ente
r
p
r
ise is
highe
r th
an
upstream
ent
erp
r
ise, an
d
lowe
r th
an d
o
wn
stre
am e
n
terp
rise.
When a
lo
w
order
node is
con
n
e
cted to a hig
h
orde
r nod
e, the directio
n sho
u
ld be fro
m
the low ord
e
r nod
e to high
orde
r no
de.
(2) Edg
e
defi
n
ition
The ed
ge d
e
f
inition of wei
ghted
compl
e
x network i
s
e=[(v
i
,v
j
),w
ij
],
whi
c
h represents the
busi
n
e
ss
co
ntact of the
enterpri
s
e
v
i
an
d
v
j
in
the ind
u
stri
al clu
s
te
r ne
twork. Weigh
t
w
ij
rep
r
e
s
ent
s th
e logi
stics a
m
ount b
e
twe
en the
ente
r
p
r
ise
v
i
an
d
v
j
, and
w
ij
is
we
ight of the
ed
ge
from
v
i
to
v
j
.
It is equal to
w
ji
which is wei
ght of the edg
e from
v
j
to
v
i
.
(3) Net
definit
ion
The net d
e
fin
i
tion of wei
g
h
t
ed indu
strial
clu
s
ter
com
p
lex network is
G
=
(V,E).
V is the
node
s set
of
indu
strial clu
s
ter ente
r
pri
s
s,
an
d
E is the ed
ge
s
set
whi
c
h
pre
s
e
n
t the bu
sin
e
s
s
contract a
m
o
ng the ente
r
p
r
ise
nod
es of
weighte
d
ind
u
strial clu
s
ter.
Weighted a
d
jacen
c
y
matrix
W
is used to
repre
s
e
n
t the weighte
d
n
e
twork.
w
ij
prese
n
ts logi
stics am
ount of
the node i and
node j. If there is no conn
e
c
tion bet
wee
n
the node
i
a
nd nod
e
j
, the weight
w
ij
=0
.
2.2. Model Constr
uction
of Weigh
t
e
d
Industrial Cl
uster
Compl
ex Ne
t
w
o
r
k
The m
odel
constructio
n
o
f
weig
hted in
dustri
a
l clu
s
ter com
p
lex
n
e
twork proce
s
s
is
a
s
follows
:
(1) Sta
r
t with
a small
num
ber
m
0
of nodes
and
sma
ll numbe
r
e
0
of edge
s. Th
e initial
node
s are ra
ndom allo
cat
ed with nod
e
attribution, orde
r, and ed
ge weig
ht. Point stren
g
th of
these n
ode
s i
s
equ
al to the sum of the e
dge weight.
(2) A
dd a
ne
w no
de
j
at e
v
ery time ste
p
t
, set the
node
attributi
on in
clud
e cl
ass an
d
orde
r, then
set the edg
e
numbe
r m
a
nd its lo
cal
worl
d [11]
A
j
acco
rding
n
ode attri
butio
n as
follows
:
a)
If the new no
de
j
i
s
a
up
stream
ente
r
pri
s
e, the
n
m
=
3, the lo
cal
world i
s
all
mid
d
le
enterp
r
i
s
e
s
a
nd lower o
r
de
r upst
r
eam e
n
terp
rises;
b)
If the new no
de
j
i
s
a
midd
le ente
r
pri
s
e,
then
m
= 5, the
local worl
d
is all
u
p
stre
am
enterp
r
i
s
e
s
, all down
s
tre
a
m
enterp
r
ises, and other o
r
der mid
d
le en
terpri
se
s;
c)
If the new node
j
is a do
wn
strea
m
en
terpri
se, then
m
= 3, the local
worl
d is all
middle ente
r
p
r
ise
s
;
(3) In the
lo
cal
wo
rld
of nod
e
j
, usi
ng a preferential a
ttachm
ent with probability
()
local
ji
defined by:
1
()
j
loc
a
l
i
l
lA
j
is
s
Selec
t
m
of node
s to conne
ct with
the ne
w nod
e
j
acco
rdin
g to the co
nne
ction
prob
ability. Randomly a
s
si
gn weight
w
ij
of edge
s b
e
twee
n sele
cte
d
old n
ode
i
and
n
ode
j
, a
nd
cal
c
ulate the
point stre
ngth
of all nodes.
(4)
Repe
at 2), 3) until the
numbe
r of no
de ent
erpri
s
e
s
in the weig
hted netwo
rk rea
c
he
s
to N.
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 10, Octobe
r 2014: 733
0
– 7336
7332
3. Simulation of Weigh
t
e
d
Industrial
Cluster Com
p
lex Ne
t
w
o
r
k
In the si
mul
a
tion of
wei
ghted in
du
strial clu
s
te
r
complex n
e
twork mo
del, the initial
indu
strial
clu
s
ter net
work
is
set
10
of
node
s, eve
r
y
ne
w u
p
st
rea
m
, middle
an
d do
wn
strea
m
enterp
r
i
s
e ta
ke
s 3,
5, 3
e
dge
s
whi
c
h
conne
ct ol
d n
ode
s. Th
e n
u
mbe
r
s of th
e three
kin
d
s o
f
node
s which
eventually join the ind
u
s
trial
clu
s
ter netwo
rk
are set 48
0, 800 an
d 12
80
respe
c
tively. Give the e
d
g
e
wei
ght of th
e ne
w n
ode
with a
ran
d
o
m
intege
r ran
ge from
1 to
3, the
old node
s co
nne
cted to each ed
ge get
the edge we
ight, and the increa
se valu
e of node poi
nt
stren
g
th is th
e new e
dge
weig
ht.
3.1. Point Strength
Distri
bution of
We
ighted Indus
trial Cluster
Complex
Netw
o
r
k
In the weight
ed indu
strial
clu
s
ter
compl
e
x
network m
odel, point strength is d
e
fin
ed by:
i
ii
j
jN
s
w
N
i
is th
e nei
g
hbors
colle
cti
on of the n
o
d
e
i
. The p
o
int
stre
ngth ta
kes
con
s
id
er
with both
the num
ber
of neigh
bors and
weig
ht betwe
en th
e nod
e
i
a
n
d
its n
e
ighb
ors,
whi
c
h i
s
a
comp
re
hen
si
ve reflection
of the node'
s
local info
rmat
ion.
Figure 2.
Point Strength Di
stributio
n
As sho
w
in
Figure 2,
ent
erp
r
ises nod
e poi
nt
stren
g
th di
stributi
on i
s
a
c
cord
with th
e
power law di
stribution. M
o
st point
stre
ngt
h of the
e
n
terprise
node
i
s
small,
and
a f
e
w
ente
r
pri
s
e
s
point stre
ngth
is very large.
In the evolution of indu
stri
al clu
s
ter net
work
p
r
o
c
e
s
s, some ente
r
prises d
e
velo
p rapidly,
and their b
u
siness conta
c
ts are fa
st increa
sed.
But most of the enterp
r
i
s
e
s
n
ode
s have little
busi
n
e
ss
con
t
acts
with the
rest
rictio
n of
busi
ness capa
city,
geog
raphi
cal and other rea
s
o
n
s
,
then enterpri
s
e
s
with larg
e-scal
e bu
sin
e
ss ca
n attra
c
t more ne
w
node
s to est
ablish busi
n
e
ss,
and the
ne
wl
y establi
s
he
d
busi
n
e
s
s co
ntacts will l
e
ad to that th
e ente
r
pri
s
e
s
node
with l
a
rge
amount b
u
si
ness
conta
c
t
s
grow m
o
re
quickly,
and
these n
ode
s will develo
p
to be the core
enterp
r
i
s
e
s
in
the indu
strial
clu
s
ter, thu
s
point
strength
of enterp
r
ise
s
sh
o
w
"the rich get
richer"
power-la
w
be
havior in the
simulatio
n
m
odel.
3.2. Degre
e
Distribu
tion
of Weigh
t
e
d
Industrial Cl
uster
Compl
ex Ne
t
w
o
r
k
The deg
ree
d
i
stributio
n of weig
hted
i
n
d
u
strial
c
l
us
ter c
o
mplex
netw
o
rk is ac
cord w
i
th the
power la
w distribution. It is sho
w
n in Fig
u
re 3.
10
0
10
1
10
2
10
3
10
-4
10
-3
10
-2
10
-1
10
0
P
o
i
n
t
s
t
r
e
ngt
h s
P(
s
)
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Model Con
s
truction a
nd Sim
u
lation of Weighted Ind
u
strial Clu
s
ter
Com
p
lex… (Lij
un Wa
ng)
7333
Figure 3.
Deg
r
ee Di
stri
buti
o
n
In the wei
g
h
t
ed indu
strial
clu
s
ter
com
p
lex network model, a
n
e
w e
n
terp
ri
se nod
e
cho
o
se to establish b
u
sin
e
ss co
ntact
s
accordi
ng to point stre
n
g
th of the original ente
r
prise
node, the
r
efo
r
e, the hig
h
e
r
point st
rengt
h enter
pri
s
e node
s can g
e
t
more bu
si
ness conta
c
t
s
of
the ne
wly ad
ded e
n
terp
ri
se nod
es, a
n
d
these poi
n
t
stren
g
th is increa
sed
with addin
g
n
e
w
busi
n
e
ss
co
n
t
acts,
while
a
larg
e num
be
r of lo
we
r poi
nt stre
ngth e
n
terp
rise no
d
e
s
can
only
get
less ne
w bu
si
ness conta
c
ts, so the powe
r
law
ph
eno
m
enon of "the rich get ri
che
r
"
is appe
are
d
.
3.3. Point Strength
-
poin
t
Streng
th Co
r
r
elation
Point strengt
h-poi
nt st
ren
g
th correlatio
n refle
c
ts an
enterp
r
i
s
e n
o
de
choi
ce to
the oth
e
rs
according to the scale of b
u
sin
e
ss co
ntact. Larg
e
-scale busi
n
e
ss
conta
c
ts me
a
n
the large p
o
int
stren
g
th
of en
terpri
se
no
de.
If the e
n
terprise
nod
e
with
large
p
o
int
strength i
s
incli
n
ed to
conne
ct
to the enterp
r
ise
nod
e wit
h
small p
o
int
stren
g
th
, the co
rrel
a
tion
is neg
ative. If the enterp
r
i
s
e
node
with l
a
rge poi
nt st
re
ngth is in
clin
ed to
con
n
e
c
t to the ente
r
pri
s
e
nod
e
with la
rge
po
int
stren
g
th, the
co
rrel
a
tion
is po
sitive.
The me
as
ure index of
point st
ren
g
th-poi
nt stren
g
th
correl
ation i
s
the average
point st
ren
g
th of all n
e
igh
bor
enterpri
s
e nod
es for e
n
terp
rise no
d
e
i,
whi
c
h is d
e
fin
ed by:
1
,
()
nn
j
i
j
jv
j
ss
s
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 10, Octobe
r 2014: 733
0
– 7336
7334
Figure 4.
Point Strength-p
o
int Strength
Correl
ation
Figure 4 sho
w
s p
o
int stre
ngth-p
o
int strength
co
rrela
t
ion in weigh
t
ed indust
r
ial
cluste
r
compl
e
x
net
work. Wh
en
t
he
point stre
ngth
of a nod
e is small, th
e ave
r
age
po
int strength
o
f
its
neigh
bor no
d
e
s is large, with the increa
sing of
point stren
g
th, the averag
e poin
t
strength of its
neigh
bor
no
d
e
s de
crea
se rapidly,
but th
e continui
ng i
n
crea
sing
of t
he p
o
int
stre
n
g
th can
lead
t
o
the neigh
bor
averag
e point
strengt
h stab
ilize at a fixed values.
3.4. Relation
ship of Point Strength an
d Degr
ee
Figure 5
sho
w
s the
rel
a
tionship
of poi
nt st
rength
a
nd d
e
g
r
ee
in
wei
ghted
in
dustri
a
l
clu
s
ter
co
mpl
e
x network,
whi
c
h
(a
) i
s
i
n
the li
nea
r a
nd (b) is in th
e do
uble
loga
rithmic coordi
nate
sy
st
em.
(a)
(b)
Figure 5. Rel
a
tionship of
Point Strength
and Degree
There is a st
rong lin
ear
re
lationship bet
wee
n
node p
o
int stren
g
th and its deg
re
e, and
greate
r
deg
ree can
l
ead
to
gre
a
ter po
int
stre
ngth.
This
mea
n
s t
hat the e
n
terprise n
ode
with
large b
u
si
ne
ss co
ntact
s
generally have
greate
r
logistics amo
unt, for example,
the upstream
enterp
r
i
s
e wit
h
more bu
sin
e
ss conta
c
ts
gene
rally
hav
e stro
ng a
b
ility to supply, and d
o
wn
stre
am
enterp
r
i
s
e wit
h
more
con
n
e
c
tion ge
neral
l
y
have strong
purcha
s
ing
capa
city.
In doubl
e log
a
rithmi
c coordinate
s
, in th
e ca
se
of sm
aller n
ode
de
gree,
som
e
e
n
terp
rise
node
s have
same
deg
ree
but different
node p
o
int
stren
g
th. Thi
s
mea
n
s th
at in the wei
g
h
t
ed
indu
strial
clu
s
ter n
e
two
r
k, the node
p
o
int stre
ngth
with simil
a
r busin
ess
contact
s
may
be
different. In the evolution
of indust
r
ial cluster
netwo
rk, each nod
e whi
c
h join
ed i
n
to the indu
strial
clu
s
ter net
wo
rk ha
s differe
nt point stre
n
g
th
and ea
ch
edge ha
s different logi
stics amou
nt. When
0
200
400
600
800
0
50
10
0
15
0
20
0
poi
nt
s
t
r
e
ngth
s
av
er
ag
e
s
nn
of
ne
ig
hbor
s
0
10
0
20
0
30
0
40
0
0
10
0
20
0
30
0
40
0
50
0
60
0
70
0
d
e
g
r
ee k
point s
t
r
e
ngth
10
0
10
1
10
2
10
3
10
0
10
1
10
2
10
3
d
e
g
r
ee k
point s
t
r
e
n
g
th
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
ISSN:
2302-4
046
Model Con
s
truction a
nd Sim
u
lation of Weighted Ind
u
strial Clu
s
ter
Com
p
lex… (Lij
un Wa
ng)
7335
the poi
nt st
re
ngth of
ente
r
prise n
ode
s is
smalle
r,
the
logi
stics a
m
ount of
ea
ch
busi
n
e
s
s con
t
act
is the majo
r factor in affecting the
whole
no
de p
o
int stren
g
th.
When the
point stre
ngt
h o
f
enterp
r
i
s
e n
o
des is l
a
rger,
the l
ogi
stics
amount fa
cto
r
of e
a
ch
bu
siness
conta
c
t
be
come
min
o
r,
and the
num
ber
of bu
sine
ss
co
ntact
s
b
e
com
e
the
m
a
jor fa
ctor in
affecting the
whol
e no
de p
o
int
stren
g
th. So, in the indu
strial cl
uste
r netwo
rk the
enterp
r
ise node
s with
smalle
r bu
si
ness
conta
c
ts m
a
y have simila
r busi
n
e
ss
cont
acts n
u
mbe
r
while h
a
ve different point st
rength.
3.5. Rich-clu
b Property
The rich-club
pro
p
e
r
ty
is chara
c
te
rized by
the
rich
-cl
ub
con
n
e
c
tivity and the
no
de-n
ode
edge di
stri
but
ion, whi
c
h m
easure th
e interconn
ecti
o
n
betwe
en ri
ch
node
s. No
de
s in the net
work
are
so
rted
by deg
ree
s
of
node
ente
r
pri
s
e
s
. The
ri
ch
-clu
b p
r
op
ert
y
)
/
(
N
r
is d
e
fined
as th
e
ratio of the total actual nu
mber of edg
e
s
to
the maximum po
ssibl
e numbe
r of edge
s amo
n
g
the
membe
r
s
with maximal de
gree
s. The m
a
ximum po
ssible numb
e
r o
f
links betwee
n
node
s is
n(n
-
1)/2
. In the weighted in
du
strial clu
s
ter
complex net
work, the p
o
int
stren
g
th is u
s
ed tco in
ste
a
d
the degree o
f
node ente
r
prises. Th
e ri
ch-clu
b prop
er
ty of weigh
t
ed indu
strial
cluste
r com
p
lex
netwo
rk is de
fined a
s
the
ratio of the th
e total
a
c
tual
numbe
r of
ed
ges to the m
a
ximum po
ssible
numbe
r of ed
ges am
ong th
e membe
r
s
with maximal point stren
g
th.
Figure 6. Rich-cl
ub Pro
p
e
r
ty with Point
Strength
Figure 6
sh
o
w
s the
rich-cl
ub p
r
op
erty
with poi
nt st
rength
of wei
ghted in
du
stri
al clu
s
te
r
compl
e
x net
work. At the
start of th
e ri
ch-clu
b
curv
e
,
the de
cline
spe
ed i
s
slo
w
er tha
n
the
e
nd of
the cu
rve. Th
is phe
nom
en
on sh
ows th
at rich
-cl
ub
i
s
existed i
n
the wei
ghted
indu
strial
clu
s
ter
compl
e
x network.
4. Conclusio
n
In this pape
r we co
nst
r
u
c
t a weighted
indust
r
ial cl
uster
com
p
le
x network m
odel to
resea
r
ch the
evolution of t
he ind
u
st
rial
clu
s
ter
netwo
rk. T
he
weig
h
t
is set a
s
lo
g
i
stics am
ount
of
the ente
r
p
r
ise conn
ectio
n
. With
gro
w
th
of the
n
e
two
r
k, every
edg
e weight
rem
a
ins un
ch
ang
ed,
and the
ente
r
pri
s
e
nod
e p
o
int strength
increa
se
s.
Simulation
sh
o
w
that the
d
egre
e
a
nd p
o
int
stren
g
th di
stribution a
r
e
a
c
cord
with
p
o
we
r la
w,
p
o
int strength
-
point st
ren
g
th co
rrelation
is
negative,
a
n
d
rel
a
tion
ship
with
d
e
g
r
ee
and point st
rength i
s
stro
ngly line
a
r. In
re
al
worl
d, e
dge
weig
ht in ind
u
strial
clu
s
ter may chan
ge
as t
he net
work
gro
w
s, so, how to e
s
tablish the m
o
re
reali
s
tic mo
d
e
l of industri
a
l cluste
r net
work
acco
rdin
g
to different network evolut
ion mechani
sm,
is the future rese
arch di
re
ction of industrial clu
s
ter net
work.
Ackn
o
w
l
e
dg
ements
This
pap
er i
s
sup
p
o
r
ted b
y
the Youth
Found
ation f
o
r So
cial S
c
i
ence Develo
pment a
nd
Re
sea
r
ch of Hebei Provi
n
ce of Chi
n
a unde
r Gra
n
t No. 2012
0412
4 and the Fund
ame
n
tal
Re
sea
r
ch Fu
nds for th
e Central Universities u
nde
r Grant
No. 13
MS113.
10
-4
10
-3
10
-2
10
-1
10
0
10
-3
10
-2
10
-1
10
0
r/
N
(r/
N
)
Evaluation Warning : The document was created with Spire.PDF for Python.
ISSN: 23
02-4
046
TELKOM
NI
KA
Vol. 12, No. 10, Octobe
r 2014: 733
0
– 7336
7336
Referen
ces
[1] Z
hong
lu
Z
e
n
g
.
Industrial cl
u
s
ter and re
gio
nal ec
on
o
m
ic
deve
l
op
ment
. Nank
ai Eco
n
o
m
ic Studies
.
199
7; 1: 69-73.
[2] Chiles.
Man
a
g
i
ng th
e e
m
erg
ence
of cluste
r: an incre
a
sin
g
returns a
ppr
oach to strate
gic cha
n
g
e
.
Emerge
nce. 20
01;
3
: 58-89.
[3] Rull
an
i
E.
Co
mplexity
and In
d
u
strial C
l
usters
: Dyna
mics a
n
d
Mod
e
ls i
n
T
h
eory a
nd Practi
ce
. Sprin
ger
.
Heid
el
berg. 20
02; 35-6
1
.
[4]
Coo
per H, Jaime R and Jerem
y
Hal
l
, 2004,
T
he Complex
i
ty of Canad
ia
n Clusters Opera
t
ing in Globa
l
Scienc
e
. Unive
r
sit
y
of Cal
gar
y
.
W
o
rking Pap
e
r
.
[5] Xi
ng
hua
Li.
R
e
searc
h
on S
e
lf-org
ani
z
a
t
i
o
n
Mecha
n
is
m and Co
nd
itio
n of High-t
e
c
h
Enterpris
e
s
Cluster
. Bei
jin
g
:
Economic Sci
ence Press. 2
0
04; 88: 92-
102,
218-2
1
9
[6] Jixiang
Chen.
Industrial clu
s
ters and co
mp
lexity
. Sha
ngh
ai: Shan
gh
ai Univ
ersit
y
of F
i
nance an
d
Econom
ics Pre
ss. 2005.
[7] Gang
L
i
.
Preli
m
i
nary D
i
scuss
ion o
n
F
o
r
m
ati
on a
nd Evo
l
ve
me
nt of Industr
ial C
l
usters V
i
e
w
point bas
e
d
on the self-or
g
ani
z
a
t
i
on th
eor
y
.
Academic Exch
an
ge. 20
05
; 78-82.
[8] Caimei
Lu.
T
he complex
i
ty of the analysis of
enterprise cl
u
s
ters
. Science &
T
e
chnolo
g
y
Progress a
n
d
Polic
y
.
20
07; 6: 76-79.
[9]
Albert R, BarabasiA
L.
Statistical
mec
han
ics
of comp
lex net
w
o
rks
. Phy
s
. R
e
v
.
Mod. 200
2; 74(1): 47.
[10]
Xi
an
g Li, Guan
rong C
hen.
A
l
o
cal-w
o
rl
d evol
ving n
e
tw
ork mode
l
. Ph
y
s
ica
A. 2003; 32
8, 274 – 28
6.
Evaluation Warning : The document was created with Spire.PDF for Python.