TELKOM
NIKA
, Vol.11, No
.11, Novemb
er 201
3, pp. 6427
~6
433
e-ISSN: 2087
-278X
6427
Re
cei
v
ed Ma
y 13, 201
3; Revi
sed
Jun
e
20, 2013; Accepted July 1
1
,
2013
A Multi-agent Supply Chain Information Coordination
Mode based on Cloud Computing
Wuxu
e Jian
g*
1,2
, Jing Zhang
1
, Junhu
ai Li
1
1
School of Co
mputer Scie
nc
e and En
gi
neer
ing,
Xi’
an Un
iv
ersit
y
of T
e
chn
o
lo
g
y
,
Xi
’an
71
004
8, Shaa
n
x
i,
Chin
a
2
Departme
n
t of Computer En
g
i
ne
erin
g, Don
g
gua
n
Pol
y
tec
h
nic, Don
g
g
uan
523
80
8, Guan
gdo
ng, Ch
ina
*Corres
p
o
ndi
n
g
author, e-ma
i
l
: jian
g
w
u
xue
_
1
@1
63.com
A
b
st
r
a
ct
In order to i
m
prove th
e hig
h
efficiency a
n
d
secu
rity of su
pply ch
ai
n info
rmati
on co
ordi
natio
n
und
er clo
ud c
o
mputi
ng e
n
vir
o
n
m
e
n
t, this paper pr
op
os
es
a supp
ly cha
i
n infor
m
ati
on
coord
i
nati
on
mod
e
base
d
on cl
oud
comp
utin
g. T
h
is mo
de h
a
s two basic stat
use
s
w
h
ich are onl
ine status
an
d
offline status. A
t
the onl
in
e status, cloud co
mp
utin
g
cente
r
is respons
ib
le for coord
i
n
a
ting the w
h
o
l
e sup
p
ly ch
a
i
n
infor
m
ati
on. At
the offli
ne st
atu
s
, infor
m
ati
on
excha
n
g
e
ca
n
be r
eal
i
z
e
d
a
m
ong
differe
nt n
odes
by
usin
g t
h
e
pre-b
a
ckup s
u
pply ch
ain i
n
f
o
rmatio
n
an
d netw
o
rk
addr
e
ss of other n
odes. T
h
is co
ordi
natio
n mo
de
coord
i
nates
th
e w
hol
e su
pp
l
y
chai
n
infor
m
ation
throu
g
h
the cl
ou
d co
mp
uting
cent
er
at
the
onl
ine
sta
t
us
and
m
a
ximi
z
e
s
the
overall benefit. Me
anwhile, the
offline s
t
atus effect
ively avoids the r
i
sks to the whole
supp
ly cha
i
n b
r
oug
ht by the
over
d
epe
nd
en
ce of clou
d co
mp
utin
g on
ne
tw
ork and the
hig
h
infor
m
atio
n
conce
n
tratio
n. T
he tw
o statuses hav
e
g
ood
customers
’
s
a
ti
sfaction a
nd
hi
gher
order
finis
hed r
a
te i
n
su
p
p
ly
chai
n. T
he si
mu
lati
on ex
per
iment bas
ed
o
n
multi-Ag
ent
verifies the
effect
iven
ess of this
mod
e
in t
h
is
pap
er.
Ke
y
w
ords
:
su
pply ch
ain, i
n
fo
rmati
on co
ordi
nat
io
n, clou
d computi
ng, risk avoi
danc
e
Copy
right
©
2013 Un
ive
r
sita
s Ah
mad
Dah
l
an
. All rig
h
t
s r
ese
rved
.
1. Introduc
tion
With the
dev
elopme
n
t of i
n
formatio
n te
chn
o
logy
and
network te
chnolo
g
y in th
e 21
st
centu
r
y, sup
p
ly chai
n ha
s ente
r
ed i
n
a ne
w er
a
of informatio
nizatio
n
an
d
netwo
rk. In
the
comp
etition
betwe
en e
n
terp
rises, the
quality of
sup
p
ly ch
ain
has
played
an in
creasi
ngly
importa
nt rol
e
in term
s
of determi
ning t
he future an
d
destiny of th
e ente
r
pri
s
e.
The e
n
terp
rises
maste
r
ing
an
d re
spo
ndin
g
to the su
pply
chai
n informa
t
ion rapi
dly are more po
ssi
b
le to wi
n in t
h
e
fierce
b
u
sin
e
s
s comp
etitions. T
he
su
pp
ly chai
n
co
n
s
ists of sup
p
lier,
ma
nufa
c
ture
r,
di
strib
u
tor,
vendor an
d
cu
stome
r
an
d ea
ch
of th
em con
s
is
t
s
of many enti
t
ies. Mo
reov
er, ea
ch
enti
t
y
in
c
l
ud
es
c
o
un
tle
s
s
s
u
pp
ly in
fo
r
m
a
t
ion a
n
d
d
e
man
d
info
rmatio
n. Ho
w t
o
coordi
nate
m
a
ss
informatio
n from different obje
c
ts in su
pply chai
n a
nd how to m
a
ximize the overall be
nefit of
sup
p
ly chai
n have alre
ady become a bu
rning issu
e
in the cu
rrent su
pply chai
n informatio
nization
p
r
oc
es
s
.
The ma
rket of traditional
supply
chai
n is re
l
a
tively stable; ho
wever, its te
chni
cal
prog
re
ss is sl
ow an
d the produ
ct life cycle is lo
n
g
. Therefo
r
e, u
s
u
a
lly the previous
sup
p
ly chai
n
informatio
n coordi
nation
a
dopted t
he
chaine
d info
rmation flo
w
mode. In
ord
e
r to
solve t
he
probl
em
s in the ch
aine
d in
formation flo
w
su
ch
as
inf
o
rmatio
n dist
ortion, lon
g
transfe
r cy
cle, low
utilization rat
e
and hard-t
o
-sha
re,
the schola
r
s pro
posed
a se
ri
es
of info
rma
t
ion co
ordi
na
tion
mode
s. M
r
. McCull
en
and
Mr. T
o
will [1]
pro
p
o
se
d a
sup
p
ly ch
ain i
n
formation
m
ode
stren
g
theni
ng
dema
n
d
informatio
n. Thi
s
m
ode
add
ed the
chan
nels feedi
ng
ba
ck custo
m
e
r
informatio
n to supplie
r, ma
nufactu
re
r, distri
buto
r
and
vendor. Altho
ugh su
ch a m
ode ha
s solv
ed
the pro
b
lem
of cu
stome
r
deman
d information di
stortion in the supply ch
ain,
it still does
n
o
t
reali
z
e information sh
ari
ng of the other fou
r
nod
es in
cludin
g
sup
p
lier.
R.
Garcia-Flores [2]
prop
osed a n
e
twork info
rm
ation flow mo
de. At this
mode, ea
ch no
de tran
sfers i
n
formatio
n to the
other
nod
es
and m
a
kes
d
e
ci
sion
s to th
e be
st of
thei
r o
w
n a
d
vant
age
s. Ho
wev
e
r, this
mod
e
can
not maximize
the overall benefit of su
pply c
hai
n. Mr. Hau L. L
ee and M
r
. Seungjin
Wh
ang
prop
osed an
integrate
d
informatio
n flow operatio
n mode [3] which
concentrates all informatio
n
in an i
n
form
ation
cente
r
for ce
ntrali
zed
p
r
ocessi
n
g
. Su
ch m
ode
ha
d
the a
d
vantag
es
of a
c
curat
e
Evaluation Warning : The document was created with Spire.PDF for Python.
e-ISSN: 2
087-278X
TELKOM
NIKA
Vol. 11, No
. 11, Novemb
er 201
3: 642
7 – 6433
6428
informatio
n, fast tra
n
sfe
r
speed
and
suff
icient info
rma
t
ion sh
arin
g, however, a
s
all inform
atio
n is
saved i
n
one
informatio
n center, it is e
a
s
y to form inf
o
rmatio
n man
ageme
n
t mon
opoly which
can
cau
s
e th
e informatio
n asy
mmetry at e
a
ch
node i
n
sup
p
ly chai
n. The do
cum
ents [4] and
[5]
prop
ose a virtual inform
ation cente
r
sch
e
me ba
se
d o
n
clo
ud
com
p
uting to solve
the problem
of
sup
p
ly chain
coo
r
din
a
tion.
It takes the
cl
oud
com
put
in
g platform a
s
the informatio
n center at th
e
integrate
d
inf
o
rmatio
n flo
w
mo
de. It has
so
lve
d
the p
r
oble
m
s of inform
ation ma
nage
m
ent
monop
oly an
d se
cu
rity; howeve
r
, a
s
it adopt
s
the centralized
in
formation ma
nagem
ent,
th
is
sup
p
ly chai
n informatio
n flow mod
e
will
suffer
a dev
astating bl
ow if the cloud computing
cen
t
er
is atta
cked.
Some
sch
o
lars impl
em
ent the
re
la
ted stu
d
ies
from a
nothe
r persp
ective
o
f
informatio
n coordi
nation.
The do
cu
me
nt [6] analyze
s
the
so
cial ri
sks a
nd te
ch
nical
risks in
the
pro
c
e
s
s
of cloud com
puti
ng sup
p
ly
chain
i
n
form
ation
coo
r
din
a
tion in
de
pth. Aiming
at the
traditional ag
ricultu
r
al
p
r
o
duct sup
p
ly
mode
s
which
are hi
ghly di
screte, the d
o
cum
ents [7]
and
[8] implement
studie
s
to th
e clo
s
e
d
ag
ri
cultural p
r
od
u
c
t su
pply cha
i
n inform
ation
i
zation
platform
based on co
mputing. The
docum
ent [9, 10] propose
s
a para
-
virtu
a
l enterp
r
ise logisti
cs mo
d
e
and an
alyze
s
the main logi
stics relatio
n
ship
of this logi
stics mode by
actual cases.
Clou
d co
mpu
t
ing is a n
e
w IT technol
og
y following g
r
id com
puting.
It has a cert
ain ri
sk
while
it bri
n
g
s
the
ne
w o
p
portunitie
s
fo
r su
pply
chain
inform
ation
coordi
nation
m
ode. Ba
sed
o
n
the clou
d co
mputing tech
nology, this
pape
r ta
ke
s
the advantag
es an
d disad
v
antage
s of its
appli
c
ation
s
in sup
p
ly chai
n informatio
n
coordi
natio
n
into a comp
rehe
nsive
co
nsid
eratio
n a
n
d
prop
oses
a supply chain
coordi
nation
mode whi
c
h
i
s
cap
able of resi
sting
the clou
d comp
uting
ris
k
.
2. Supply
Chain Coordina
tion
2.1. Definitio
n
of Supply
Chain
Some Chine
s
e schol
ars
su
ch a
s
M
r
. MA
Shihua
belie
ve that su
ppl
y chain
is
a f
unctio
nal
netwo
rk
chai
n st
ru
ctural
mode
integ
r
a
t
ing suppli
e
r,
man
u
factu
r
e
r
, di
stributo
r
,
retaile
r a
nd fi
nal
cu
stome
r
. It surroun
ds th
e
core ent
erp
r
i
s
e
s
an
d c
ontrols the i
n
form
ation flow, lo
gistics flo
w
a
nd
capital flo
w
to produ
ce inte
rmediate p
r
od
ucts
a
nd finished produ
cts
star
ting from purcha
s
in
g ra
w
material
s. Fin
a
lly the produ
cts a
r
e delive
r
ed
to consu
m
ers throu
g
h
the selling n
e
twork.
2.2. Supply
Chain Infor
m
ation Coo
r
dination
Information fl
ow reflect
s
the pro
c
e
ss
of informatio
n transfe
rri
ng a
m
ong differe
nt node
s
in su
pply ch
ai
n. As ea
ch n
ode bel
ong
s t
o
diffe
rent e
n
t
ities and e
a
ch node
only knows thei
r o
w
n
situation, the informatio
n of
other node
s
can only be
maste
r
ed by informatio
n co
mmuni
cation.
In
orde
r to e
n
h
ance the u
n
derstandi
ng
and ma
ste
r
in
g
of inform
ation of on
e n
ode to the
o
t
her
node
s
and to
the
whole
su
pply chain, e
a
ch
no
de n
e
eds
to tra
n
sf
er thei
r
own
i
n
formatio
n to
the
other.
The fun
c
tion
and p
u
rp
ose of su
pply
chai
n in
form
ation coordin
a
tion are rea
lizing th
e
operation and
market data sha
r
ing
an
d
e
x
chan
ge of
di
fferent no
de
s
and im
proving the
re
spo
n
se
spe
ed
and
q
uality of ea
ch no
de. Su
p
p
ly ch
ai
n i
n
formatio
n
coo
r
dinatio
n m
a
i
n
ly incl
ude
s
the
following items:
1) Interfa
c
e coordi
nation
The goal of i
n
terface co
ordination i
s
re
aliz
in
g the co
ordin
a
tion of each nod
e in
terface
and
ope
ratin
g
mod
e
a
nd i
m
provin
g the
effectivene
ss of informatio
n exchan
ge b
e
twee
n the
u
s
er
and the wh
ole sup
p
ly chain. Inte
rfa
c
e co
ordinati
on inclu
d
e
s
t
he similarity
of information
manag
eme
n
t interfa
c
e
of ea
ch
nod
e i
n
supply
c
h
a
i
n an
d the
si
milarity of
system
ope
rati
ng
mode.
2
)
D
a
ta
c
o
or
din
a
t
io
n
The goal
of data coo
r
din
a
tion
i
s
kee
p
i
ng
the
con
s
istency
of
information definit
ion
a
n
d
format of e
a
ch nod
e in
su
pply ch
ain a
n
d
re
alizin
g
se
amless i
n
formation utili
za
tion of one
n
ode
to the other
node
s. Data
coo
r
din
a
tion
is very
imp
o
rtant to the
whole
syste
m
coo
r
din
a
tion.
Without
data coo
r
din
a
tion, different
no
d
e
s can
n
o
t
communi
cate
with ea
ch oth
e
r, like that t
hey
are commu
ni
cating with dif
f
erent lang
ua
ges. Data
co
ordin
a
tion ca
n be measured and evalu
a
ted
from three
aspect
s
: firstly,
the data
avai
lability
deg
re
e of different
syst
em
s; secondly, the d
a
t
a
repetition
deg
ree of differe
nt system
s; thirdly,
the dat
a com
patibilit
y degre
e
of d
i
fferent syste
m
s
in terms of se
mantic limit a
nd format.
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
e-ISSN:
2087
-278X
A Multi-agent
Supply Chai
n Inform
ation Coo
r
di
n
a
tion
Mode ba
se
d on Clo
ud… (Wu
x
ue
Jian
g)
6429
3) Co
ntrol
co
ordin
a
tion
The goal of control coordin
a
tion is com
b
ining
the different functio
n
s of the whole sup
p
ly
chai
n syste
m
into an organi
c whole
and achi
evi
ng the sam
e
goal toget
her. The
co
ntrol
coo
r
din
a
tion
degree in
clud
es the availa
bility degr
ee
of services p
r
ovided by th
is nod
e and the
availability degree of
servi
c
es
provided b
y
the other no
des.
4) Prog
ram
coordi
nation
The g
oal of
prog
ram
co
ordination i
s
re
alizin
g sub
s
ystem
s
interaction to su
ppo
rt som
e
spe
c
ific
prog
rams.
T
he progra
m
coo
r
di
nation deg
re
e in
clud
es th
e follo
wing
two l
e
vels: firstly,
wheth
e
r the
related
su
bsystem com
b
i
nation of
the
same p
r
o
g
ram is smoot
h; secondly,
the
perfo
rman
ce
and efficie
n
cy
of related su
bsyste
m colla
boratio
n.
3. Cloud Co
mputing
3.1. Definitio
n
of Cloud
Computing
Clou
d com
p
u
t
ing is a new comme
rci
a
l comp
ut
ing m
odel ba
sed o
n
the develo
p
ment of
distrib
u
ted co
mputing and
virtualizatio
n techn
o
l
ogy.
No
wad
a
ys, t
here
is no
a
b
sol
u
tely unif
o
rm
definition to
cloud
com
puti
ng yet. Mr. LI
U Pen
g
, a
clo
ud
comp
uting
expert i
n
Chi
na beli
e
ves that
clou
d
comp
u
t
ing di
stribut
es th
e
comp
uting ta
sks
i
n
the re
sou
r
ce pool co
nsisting of
a
lo
t
of
comp
uters a
nd ma
ke
s dif
f
erent a
ppli
c
a
t
ion sy
ste
m
s
gain the
co
m
puting p
o
wer,
storage
sp
a
c
e
and differe
nt softwa
r
e
serv
ice
s
on dem
a
nd.
3.2. Cloud Computing Ri
sk
It does not
m
ean th
at the
r
e is no
ri
sk a
t
all
by a
dopti
ng
clou
d
com
puting. Fi
rstly
,
clou
d
comp
uting is suppo
rted b
y
high-spee
d
network.
If
the use
r
’s n
e
twork conn
ection is b
r
o
k
en
because
of the low
bandwi
dth or any emergency,
the user
will not gain t
he required servi
c
es
by
netwo
rk.
Se
c
o
nd
ly, th
e u
s
er
mus
t
gain
a
n
y
se
r
v
ic
e th
rou
gh the
clou
d comput
ing center.
O
n
ce th
e
clou
d comp
uting cente
r
fai
l
s to
work
by atta
ck, the
supply chain
will b
r
ea
k d
o
w
n. If the u
s
er
saves data i
n
cloud
without loca
l
backup, once the cloud platfo
rm breaks down,
the user
will
s
u
ffer ines
timable los
s
.
4. Supply
Chain, Information Coordin
a
tion, Cloud
Compu
t
ing, Risk Av
oidance
4.1. Coordin
a
tion of O
n
line and Offlin
e Statu
ses
After taking
an overall consi
deration
of
advantag
es a
nd ri
sks of applying
clou
d
comp
uting, th
is pa
per
pro
p
o
se
s a
su
ppl
y chain info
rmation coo
r
di
nation mo
de
whi
c
h i
s
ca
pa
ble
of re
si
sting
clou
d
comp
u
t
ing ri
sk a
s
sh
own
in
Figure 1
an
d Fig
u
re
2.
Thi
s
info
rm
ation
coo
r
din
a
tion
mode ha
s two basi
c
statu
s
es wh
ich
are online status and
offline
st
atus.
Figure 1. Online Status Co
ordin
a
tion Mo
de
Figure 2. Offline Status Co
ordin
a
tion Mo
de
Evaluation Warning : The document was created with Spire.PDF for Python.
e-ISSN: 2
087-278X
TELKOM
NIKA
Vol. 11, No
. 11, Novemb
er 201
3: 642
7 – 6433
6430
At the onli
n
e
statu
s
, the
netwo
rk i
s
smoot
h and
the clou
d co
mputing
cent
er run
s
norm
a
lly. At su
ch
statu
s
,
each n
ode
in
su
pply
c
hai
n
only ex
cha
n
ges info
rmati
on
with the
cl
oud
comp
uting ce
nter and g
a
in
s the re
quire
d se
rvice
s
fro
m
the cloud
comp
uting ce
nter. The cl
o
u
d
comp
uting
ce
nter ta
ke
s
ch
arge
of
co
ord
i
nating th
e
in
formation
of different nod
es and
reali
z
i
ng
the optimal in
formation int
egratio
n and
intere
st ma
ximization of th
e whol
e su
pp
ly chain. At the
online
statu
s
, each n
ode
can ba
ckup
th
e su
pply
chai
n inform
ation
and th
e net
work
add
re
ss of
the other no
d
e
s in re
al time within its o
w
n aut
h
o
ri
zat
i
on ran
ge for
usin
g at the off-line statu
s
.
Once the
net
work fail
s to
work
or th
e cloud
com
puting cente
r
fail
s to
work
normally by
attack, it will enter into the offline status.
At the offline status, each n
o
d
e
can excha
nge
informatio
n with the other
node
s by usi
ng the
ba
cku
p
sup
p
ly chai
n informatio
n
and add
re
ss of
the othe
r n
o
d
e
s
com
p
leted
at the o
n
line
statu
s
fo
r its
own
ben
efit maximizatio
n
. This mod
e
can
guarantee
th
e no
rmal
op
eration
of
su
pply chain
in
the
ca
se
of network p
a
ralysis or cl
o
ud
comp
uting ce
nter brea
kdo
w
n.
4.2. Multi-ag
ent Inform
ati
on Coordin
a
tion Meth
od
The la
st se
cti
on de
scrib
e
s
the informatio
n net
wo
rk at the onlin
e coo
r
dinatio
n stat
us an
d
at the offline
coo
r
din
a
tion
status. A
s
these
two inf
o
rmatio
n net
works a
r
e dif
f
erent, su
ppo
rting
two differe
nt informatio
n n
e
tworks by u
s
ing o
ne coordination meth
od is the
key of the probl
e
m
.
This
pap
er
so
lves the
probl
em by u
s
in
g the mult
i-A
g
e
n
t inform
ation
co
ordi
nation
method [1
0]. At
the onlin
e
status, the
r
e a
r
e total five-la
y
er onl
i
ne Ag
ent entities in
the cl
oud
co
mputing
ce
nter
inclu
d
ing
sup
p
lier, m
anufa
c
ture
r, di
stri
b
u
tor,
ven
dor
and
cu
stom
e
r
. Each laye
r Agent i
n
cl
ud
es
online A
gent
for o
r
de
r m
anag
ement,
online A
gent
for mate
rial
planni
ng, o
n
line Ag
ent
for
prod
uctio
n
pl
annin
g
, onlin
e Agent fo
r capa
city pl
ann
ing, onlin
e A
gent for
man
u
facturi
ng, o
n
line
Agent for inventory, online
Agent for su
pplier
m
anag
ement and o
n
line Agent for marketing, etc.
The fun
c
tion
s of each on
line Agent are diffe
rent. F
o
r exampl
e, the online A
gent for
orde
r man
a
g
e
ment is respon
sible for
cu
stome
r
’s
o
r
de
r pro
c
e
s
si
ng su
ch a
s
receivin
g ord
e
r,
che
c
king inve
ntory and deli
v
ering a
c
cord
ing to the
informatio
n tran
sferred by dif
f
erent nod
es
in
real time; the
online Ag
ent for mate
rial pl
annin
g
is
re
spon
sible fo
r p
u
rcha
sing
an
d mana
ging
raw
material
s, formulating p
u
rchase plan
s a
nd providi
ng
the material
available q
u
a
n
tity information
according
to
each n
ode
in
formation.
Th
e cl
oud
com
puting
ce
nter provid
es services for
ea
ch
node
by e
s
ta
blishi
ng th
e fi
ve-layer Age
n
t entities
fro
m
supplie
r to
cu
stom
er an
d by th
e mut
ual
coo
r
din
a
tion
and commu
ni
cation am
ong
different Age
n
ts.
There is
an
offline Age
n
t co
rre
sp
on
ding to e
a
ch
online A
g
e
n
t in ea
ch
n
ode. Th
e
function
s of the offline Age
n
t are almo
st exactly
same
as tho
s
e of the online Age
n
t. When it is at
the online
status, the offline
Agent wi
ll synch
r
oni
ze with the o
n
line Agent. The differe
nce
betwe
en the
m
is that the offline Agent only wo
rks in
ca
se
of netwo
rk
p
a
ralysi
s o
r
cloud
comp
uting ce
nter breakdo
wn; in that ca
se, the
offline Agent will co
mmuni
cate with each Age
n
t
distrib
u
ted in
the network and gai
n
the info
rmation tran
sfe
rre
d from its up
strea
m
and
downstream
busi
n
e
ss Age
n
ts; based on
such informa
t
ion, the offlin
e Agent will provide service
s
for the re
spe
c
tive node.
4.3. Critical Steps Es
tabl
ishing the Coordinatio
n
Mode
The info
rmati
on security p
e
rform
a
n
c
e
o
f
cloud
comp
uting cente
r
i
s
mu
ch
co
ncerne
d by
th
e
us
er
to
o. T
h
e
c
l
ou
d c
o
mp
u
t
in
g
c
e
n
t
er
mu
st
have
goo
d
inform
ation
cla
s
sificatio
n
an
d
confid
entiality and e
s
tablish the autho
rity managem
e
n
t
to informati
on. Accordi
n
g to the different
confid
entiality levels
of ent
erp
r
ise info
rmation,
the i
n
formatio
n
can b
e
divide
d into th
ree
basi
c
levels. Th
e in
formation
at the hig
h
e
s
t le
vel ca
n
only
be viewed
by the ente
r
p
r
ise itself a
nd t
h
e
core
coo
perative enterpri
s
e
s
; in ad
dition,
su
ch i
n
formatio
n can
only be
sa
ved insi
de t
h
e
enterp
r
i
s
e. T
he informatio
n at the intermediate leve
l
is op
en to th
e wh
ole
sup
p
ly chain; h
o
we
ver,
su
ch inform
ation can o
n
ly be edited by the enterp
r
ise itself. The informatio
n at the lowest le
vel
can b
e
browsed thro
ugh
ou
t the whole n
e
twork. A
go
od inform
atio
n cla
ssifi
catio
n
system i
s
g
ood
for the inform
ation se
cu
rity of
the whole sup
p
ly chain
and strength
e
n
s the tru
s
t of the user to th
e
se
curity of the clou
d com
p
uting ce
nter.
Whe
n
this co
ordin
a
tion m
ode i
s
at the
off
line statu
s
, the cl
oud
co
mputing
ce
nter d
o
e
s
not play
any
role
any m
o
re. In thi
s
ca
se, the
offline
Agent of
ea
ch no
de
whi
c
h
wa
s depl
oye
d
in
advan
ce
sh
al
l play the
role
. The foll
owi
n
g item
s n
eed
to b
e
n
o
tice
d in th
e
pro
c
ess of
depl
oying
the offline Agent: firstly, whe
n
the offline Agent
i
s
at the onlin
e statu
s
, it shall ba
ckup t
he
informatio
n whi
c
h can b
e
browsed
within
the
a
u
thority limit of the no
d
e
from th
e
cloud
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
e-ISSN:
2087
-278X
A Multi-agent
Supply Chai
n Inform
ation Coo
r
di
n
a
tion
Mode ba
se
d on Clo
ud… (Wu
x
ue
Jian
g)
6431
comp
uting ce
nter. On
ce th
e netwo
rk fail
s to wo
rk a
n
d
the node ent
erp
r
ise fails to comm
uni
ca
te
with the clou
d comp
uting cente
r
, the node ca
n
co
m
p
lete the sub
s
eq
uent wo
rks by using th
e
backu
p info
rmation
whi
c
h
wa
s
co
mple
ted in
re
al ti
me p
r
eviou
s
l
y
until the
n
e
twork re
cov
e
rs.
Secon
d
ly, if the net
work
works n
o
rm
a
lly and
th
e cloud com
puti
ng cente
r
b
r
eaks do
wn,
t
h
e
offline appli
c
ation shall
co
mmuni
cate
with the ot
he
r
node
s a
nd g
a
in the
req
u
ired info
rmatio
n by
usin
g the ba
ckup net
work addre
s
s of the other n
o
d
e
s. In this case, the coo
r
dinatio
n mo
de
pre
s
ent
s the
netwo
rk flo
w
mode a
nd ev
ery nod
e ca
n
work no
rmall
y
. The only difference is th
a
t
the goal of informatio
n co
o
r
dinatio
n is transfo
rme
d
to benefit maximization of p
a
rtial nod
es from
the benefit maximization of
the whole
su
pply chai
n.
4.4. Coordin
a
tion Model
This pap
er p
r
esented
a
fair
coo
r
di
nati
on m
odel
ba
sed
on
tra
n
saction
hi
story, it is
defined a
s
fol
l
ows.
,,
,
d
M
NS
V
W
(1)
1,
2
3
:
{
,
,
..
.,
}
n
NN
a
a
a
a
is Agent se
t involved in supply ch
ai
n coo
r
din
a
tio
n
, n is th
e
numbe
r of the Agent.
1,
1
2
2
:
{
,
,
,
...
,
,
,
}
mm
SS
s
z
s
z
s
z
is pa
rtici
p
a
n
ts’ p
r
opo
siti
on, incl
udin
g
tasks,
r
e
sour
ces
,
et
c
.
:
{
,
,
,
..
.,
}
3
12
SS
SS
mm
mm
V
V
VVV
V
aa
n
aa
is issue’
s vector-valu
ed set which is rai
s
ed by Agent.
:
{
,
,
.
..,
}
12
SS
S
mm
m
WW
w
w
w
k
is wei
ght value of ea
ch issue of p
r
op
osition assi
gn
ed by
an individual
agent an
d
1
1
k
k
k
w
.
In this pape
r, a satisfa
c
tio
n
degree eva
l
uation fun
c
tion ba
sed o
n
coo
r
din
a
tion fairne
ss
wa
s de
sign
ed
and it is described a
s
follo
ws.
1
(,
)
(
,
)
K
ii
i
VT
v
t
(2)
In which,
(,
)
ii
vt
is satisfactio
n
de
gree of
i
x
and it’s definition is as follo
ws.
0
(1
)
(,)
(
(
)
)
(
)
i
t
ii
i
i
i
i
ii
th
vt
t
h
v
v
w
vv
(3)
Among them,
th
is minim
u
m
threshold fo
r satisfa
c
ti
on
degree, al
so
is the retenti
on of
sat
i
sf
a
c
t
i
on.
0
i
v
is initial
val
ue of i
s
sue
vector
i
x
, while
0
ii
vv
, the sati
sfaction
deg
re
e of
individual ag
e
n
t to the issu
e of
i
a
is
1
.
5. Experimental Simulation
Figure 3. The Simulated
Supply Chai
n
Evaluation Warning : The document was created with Spire.PDF for Python.
e-ISSN: 2
087-278X
TELKOM
NIKA
Vol. 11, No
. 11, Novemb
er 201
3: 642
7 – 6433
6432
JACK si
mulat
i
on platform f
a
cin
g
to multi-A
gent imple
m
ents
simula
tion verificati
on to the
sup
p
ly chain i
n
formatio
n
co
ordin
a
tion m
o
de p
r
op
osed i
n
this pap
er.
The
simul
a
te
d supply
chai
n
can b
e
se
en in Figure 3, which
con
s
i
s
ts
of one
su
ppli
e
r, one ma
nu
facture
r
, one
distrib
u
tor, on
e
vendor a
nd o
ne cu
stome
r
.
See the detail
ed simul
a
tion
param
eters in Table 1.
Table 1. Simulation Para
meters
Parameter Value
customer’s order
request
200
supply
lea
d
time
4
production lead time
3
mobile time span
1
finished goods inventor
y
g
oal
600
ra
w
mate
rial inventor
y
goal
600
initial raw
mate
rial inventor
y
200
initial number of finished goods
100
initial number of stock-out
0
initial unfulfilled o
r
der
0
initial number of r
a
w
material
100
In the process
of s
i
mulation, five times
tes
t
s ha
d do
ne, the total simulatio
n
du
ration is
200, Onli
ne a
nd offline takes 1
00 ea
ch.
Acco
rdi
ng to
formula
(3),
some
re
sult o
f
the satisfa
c
ti
on
of online an
d offline is sh
o
w
n a
s
Figu
re
4 and Figu
re
5.
Figure 4. Satisfactio
n
De
gree between
Online
and Offline
Figure 5. Differen
c
e of Sat
i
sfactio
n
De
gree
betwe
en Onli
ne and Offlin
e
From Figu
re
4, in the prop
ose
d
coo
r
di
n
a
tion model, offline satisfa
c
tion is lo
wer than the
satisfa
c
tion
d
egre
e
of the
online
statu
s
, on the
wh
ol
e, they are very simil
a
r. I
n
Figu
re
5, the
differen
c
e
of satisfa
c
tion
b
e
twee
n onli
n
e and
offline t
end
s to b
e
st
able d
u
rin
g
th
e five tests,
a
n
d
the differen
c
e
value is between 0.19 to 0.
21.
Switch the
su
pply chai
n to the offline
sta
t
us
from
the online status whe
n
100
t
. See
the averag
e o
r
de
r finishi
ng
rate befo
r
e a
nd after statu
s
switchi
ng in
Table 2.
Table 2. Average O
r
d
e
r Fi
nishi
ng Ra
te at the Online/
Offline Status
Average finished rate
online status
99.7%
offline status
96.8%
50
60
70
80
90
100
110
1234
5
satisfaciton
times
o
n
lin
e
0
0.
5
1
1.
5
2
2.
5
1234
5
diff of
satisfacation
time
s
dif
of
…
Evaluation Warning : The document was created with Spire.PDF for Python.
TELKOM
NIKA
e-ISSN:
2087
-278X
A Multi-agent
Supply Chai
n Inform
ation Coo
r
di
n
a
tion
Mode ba
se
d on Clo
ud… (Wu
x
ue
Jian
g)
6433
It can
be
see
n
from
the
ta
ble that
the
supply
chai
n o
perate
s
well
whe
n
it i
s
at the o
n
line
status an
d th
e ave
r
age
o
r
der finishi
ng
rate i
s
nea
rly 100%. It i
s
becau
se th
e
clou
d
comp
uting
cente
r
gath
e
r
s the i
n
formation of all
node
s a
n
d
coo
r
din
a
tes the su
pply
chai
n inform
ation
intensively to
improve th
e p
r
odu
ction
and
transpo
rtatio
n ca
pability of the
su
pply chain. When t
he
sup
p
ly chai
n is switch
ed t
o
the offline status
, e
a
ch
node o
n
ly op
timizes th
eir
own
status
a
nd
only pursue
s
for the benefi
t
maximization of thei
r own
instead of pu
rsui
ng for the
overall benef
it
maximizatio
n
. As a
co
nse
q
uen
ce, the
a
v
erage
fini
shi
ng rate falls;
however, the
falling rang
e
is
small an
d the
average fini
shing rate
still rea
c
he
s 96.8
%
. Viewed from the simul
a
tion re
sult, the
sup
p
ly
chain informatio
n coordi
nation
m
ode pro
p
o
s
e
d
in thi
s
pap
er avoi
ds ri
sks effe
ctively and
improve
s
the
prod
uctio
n
a
nd tran
sp
orta
tion cap
abilit
y of supply chain by u
s
ing
the advanta
ges
of cloud
com
puting center.
6. Conclusio
n
This pa
pe
r a
nalyze
s
the
advantag
es
and ri
sks of
clou
d com
p
u
t
ing and pro
poses a
sup
p
ly ch
ain
coo
r
din
a
tion
mode
whi
c
h
can
avoid the
clou
d compu
t
ing risks. Thi
s
mo
de h
a
s t
w
o
basi
c
statu
s
e
s
whi
c
h a
r
e o
n
line statu
s
a
nd off
line stat
us. At the online statu
s
, cl
oud computin
g
cente
r
is re
spon
sible fo
r
coo
r
din
a
ting
the wh
ole
su
pply ch
ain in
formation;
while at the of
fline
status, info
rm
ation exch
an
ge ca
n be
re
alize
d
am
o
n
g
different nod
es by u
s
ing t
he pre-b
a
cku
p
sup
p
ly chain
information
and netwo
rk add
re
ss o
f
other nod
es. This
co
ordin
a
tion m
ode
coo
r
din
a
tes t
he wh
ole su
pply chai
n informatio
n
thro
ugh the cl
ou
d com
puting
cente
r
at onli
ne
status
and m
a
ximize
s the
overall be
ne
fit. Meanwhil
e
, the offline status effe
ctively avoids the
risks to the
whole
sup
p
ly chain b
r
ou
ght
by t
he overd
epen
den
ce
of clou
d comp
u
t
ing on n
e
two
r
k
and
th
e high
informatio
n concentratio
n
.
The
si
mula
ti
on
re
sult
sho
w
s that
wh
en
the m
ode
i
s
at
the online
status, it has v
e
ry good
pro
c
e
ssi
ng c
apa
city, includin
g
cu
stome
r
s’
satisfa
c
tion
and
orde
r fini
sh
ed
rate.
Wh
en it
is at the
offline
stat
us,
its pro
c
e
ssi
ng capa
city
falls
slightly; even
so,
its processi
ng capaci
ty is st
ill excellent.
Ackn
o
w
l
e
dg
ement
This wo
rk
wa
s sup
porte
d by
the
Nation
al
Natural
Science Fo
und
ation of
Chin
a (G
ra
nt
No. 611
720
18), an
d by
the Nation
al Natu
ral S
c
ien
c
e F
oun
dation of China (Gra
nt
No.
6117
2124
).
Referen
ces
[1]
McCull
enP, T
o
w
i
ll
D. Dia
gnos
i
s
and re
ductio
n
of bull
w
h
i
p in
suppl
y ch
ai
ns.
Supply C
h
a
i
n
Mana
ge
me
nt:
an Internati
o
n
a
l
Journ
a
l.
200
2; 7(3): 164-1
7
9
.
[2]
R Garcia-F
lor
e
s, XZ
W
ang,
GE Goltz. Agent-base
d
infor
m
ation fl
o
w
fo
r process i
n
d
u
s
tries’ sup
p
l
y
chai
n mode
li
ng
.
Comp
uters & Che
m
ic
al En
gi
neer
ing.
2
000;
24(1): 11
35-
11
41.
[3]
Sean Marston, Zhi Li, Subhajy
o
ti B
andy
opadhy
ay
, et al. Cloud computing
the business
perspectiv
e
.
Decisi
on Su
pp
ort Systems.
2
011; 51(
1): 176
-189.
[4]
Susan
a
Azeve
do, Paul
a Prat
a, Paulo F
a
z
e
nde
ir
o, V Crua
-Macha
do. Assessment of S
upp
l
y
Cha
i
n
Agilit
y i
n
a
Clo
ud C
o
mputi
ng-
base
d
Frame
w
ork.
Scala
b
le
Co
mp
uting: Pr
actice a
nd Ex
p
e
rie
n
ce.
2
012;
13(4): 29
5-3
0
1
.
[5]
I Mohamme
d,
R Sha
n
kar, D
Ban
w
et. Cr
eati
ng fle
x
-
l
ea
n-a
g
ile v
a
lu
e ch
ain
b
y
outs
ourci
n
g
: An ISM-
base
d
interv
en
tiona
l roadm
ap
.
Business Pro
c
ess Mana
ge
ment Journ
a
l.
20
08; 14(3): 3
38–
389.
[6]
J Yan, K Ye, H W
ang, Z
Hua. Ontolog
y
of colla
bor
ative m
anufactur
i
n
g
: Ali
gnm
ent of ser
v
ice-ori
ente
d
frame
w
ork
w
i
t
h
service-d
o
min
ant log
i
c.
Expe
rt Systems w
i
th Applic
ations.
2
010; 37(
3): 222
2–2
23
1.
[7]
Olson G. Creat
ing an Enterp
r
i
se-lev
el Gr
e
en Stra
te
gy
.
Jo
ur
nal of Bus
i
nes
s Strategy.
200
8; 29(2): 22-
30.
[8]
S Hua
ng, S Z
eng, Y F
a
n, G Hua
ng. Optim
a
l serv
ice s
e
le
ction a
nd c
o
m
positi
on for s
e
r
v
ice-ori
ente
d
manufactur
i
ng
net
w
o
rk.
Inter
n
ation
a
l J
our
nal
of Co
mput
er In
tegrated
Ma
nuf
acturin
g
.
20
11;
24(
5): 41
6-
430.
[9
]
An
th
o
n
y
Bi
song
, Sy
ed
M Ra
h
m
an
. An
Ove
r
vi
e
w
o
f
th
e Se
cu
ri
ty
C
once
r
n
s
i
n
En
te
rp
ri
se
Cl
ou
d
Comp
uting.
Internati
ona
l Jour
nal of Netw
ork Security & Its
Appl
icatio
ns.
2
011; 1(1): 3
0
-4
5.
[10]
Haluk D
e
mirk
a
n
, Hsing Ke
ne
th Chen
g, Sub
hha
j
y
oti Ba
nd
yopa
dh
ya
y
.
C
o
ordi
natio
n Stra
tegies i
n
an
Saas Sup
p
l
y
C
hai
n.
Journ
a
l o
f
Manage
ment
Information Sy
stems.
20
10; 2
6
(4): 119-
14
3.
Evaluation Warning : The document was created with Spire.PDF for Python.