Anda di halaman 1dari 8

International Journal of Advanced Engineering Research and Science (IJAERS)

[Vol-1, Issue-3, Aug- !1"#


ISS$% 3"&-'"&(
A review paper on Multi agent base intelligent manufacturing
system
D. B. Upadhyay

Department of Mechanical Engineering, Dr. J.N.Mehta Government polytechnic Amreli, Gujarat
Abstract Today, hot topic for researchers is intelligent
manufacturing system and its new paradigms to face the
problems and challenges of the 21st centurys
manufacturing industry. A new approach with this new
paradigms is problem solving through decomposition and
agent base solution is seen as one of the promising in
responding to the needs of next generation manufacturing
system. Researchers have attempted to apply multiagent
base technology to achieve autonomy, decentrali!ation,
flexibility, reliability, efficiency, learning, and self
regeneration characteristics in manufacturing system.
This paper provides a review on the recent achievements
in these areas, and discusses some "ey issues in
implementing agentbased intelligent manufacturing
systems such as #$% structure, concept of $A% and its
architecture, application domains, advantages,
challenges, agent development tools and migration to
$A%.
Keywords multi-agent system (MAS), intelligent
manufacturing system (IMS), distributed
manufacturing systems (DMS), flexible manufacturing
systems (FMS)
I. INTRODUTION
n the la!t three decade! manufacture concept! have had
!everal redefinition!, in the eightie!, the concept of
fle"i#le manufacturing !y!tem! $%M&' (a! introduced to
develop a ne( family of product! (ith !imilar dimen!ion!
and con!traint! #ut no(aday!, the capacity of
reconfiguration ha! #ecome a major i!!ue for improving
the functioning of indu!trial proce!!e!. ndeed, today a
main o#jective i! to adapt )uic*ly in order to !tart a ne(
production or to react in a failure occurrence +,-.
ntelligent manufacturing !y!tem! ha! #oth .fle"i#ility
and reconfigura#ility, in fact thi! concept #ring! more
than a fe( idea! of !oft(are intelligence meaning!, (hich
contemplate! characteri!tic! !uch a! autonomy,
decentrali/ation, fle"i#ility, relia#ility, efficiency,
learning, and !elf regeneration, all of the!e facilitie! lead
to the concept of agent0#a!ed manufacturing !y!tem!.
1he current challenge i! to develop colla#orative and
reconfigura#le manufacturing control !y!tem! that
!upport efficiently !mall #atche!, product diver!ity, high
)uality and lo( co!t!, #y introducing innovative
characteri!tic! of adaptation, agility and modulari/ation.
nformation and communication technologie!, and
artificial intelligence techni)ue!, have #een u!ed for more
than three decade! to addre!!ing thi! challenge. 1he
Multi0Agent !y!tem $MA&' theory in di!tri#uted artificial
intelligence $DA' provide! fea!i#le technical !upport for
modeling and reali/ation of intelligent manufacturing
!y!tem $M&'. Manufacturing proce!! i! a typical Multi0
Agent )ue!tion! !olution proce!!, and every department
$or !egment' in manufacturing !y!tem i! e)ual to an
Agent in the proce!!. Every !u#0mi!!ion or unit
e)uipment in manufacturing !y!tem could #e acted and
reali/ed #y !ingle Agent or (ell0organi/ed Agent group,
and complete the manufacturing ta!*! together through
their interaction and mutual coordination and cooperation.
1he manufacturing !y!tem !hall #e !timulated a! the
MA&, (hich ma*e! the !y!tem ea!ier to de!ign, reduce!
the comple"ity of the !y!tem, inten!ifie! the
recom#ination, e"panda#ility and relia#ility of the
!y!tem, and improve! the fle"i#ility, adapta#ility and
de"terity of the !y!tem.
n thi! paper, intelligent manufacturing #a!ed on multi
agent !y!tem pre!ented. 1he o#jective i! to give a general
idea of the comple"ity of the future manufacturing
!y!tem!, 1he organi/ation of thi! paper i! a! follo(!,
&ection e"plain the &tructure of intelligent M&.
&ection revie(! the concept of !oft(are agent, MA&
and it! architecture. &ection 2 give! application domain
and control !y!tem of MA& in M&. &ection 2 !ho(! the
advantage! and challenge! of the MA&. &ection 2
e"plain! methodology to migration from traditional to
MA&. &ection 2 deal (ith agent development tool! and
&ection 2 pre!ent! the paper conclu!ion!.
II. T!" #TRUTUR" O$ IM#
ntelligent Manufacturing i! a man machine integrated
intelligent !y!tem compo!ed #y intelligent machine and
human e"pert!, (hich can carry out intelligent activitie!
!uch a! analy!i!, inference, deci!ion0ma*ing, conception
and judgment during the proce!! of manufacturing. 1he
cooperation #et(een intelligent machine and human (ill
e"pand, e"tend and partially replace the (or* of human
e"pert! during the proce!! of manufacturing +3-. At the
Page | 16
International Journal of Advanced Engineering Research and Science (IJAERS)
[Vol-1, Issue-3, Aug- !1"#
ISS$% 3"&-'"&(
!ame time, it (ill collect, !tore, perfect, !hare, inherit and
develop the manufacturing intelligence of human e"pert!
+4-.
1he main type! of M& are5 M& that ta*e! improving
manufacturing !y!tem intelligence a! the o#ject and
intelligent ro#ot and Agent a! the tool6 M& that integrate!
the modeling, proce!!ing, mea!uring and operating of
corporation! through nternet6 #iological M& that adopt!
!olution procedure for #iological pro#lem!. At pre!ent,
the di!tri#uted net(or* M& model #a!ed on Agent i!
mainly adopted +7-, the architecture of (hich i! illu!trated
in %igure ,. 8n the one hand, each manufacturing unit i!
endo(ed autonomy #y Agent to #ecome the entity (ith
perfect function! and autonomy independence6 on the
other hand, the !y!tem i! endo(ed (ith !elf organi/ation
capacity through the coordination and cooperation
#et(een Agent!.
$igure %. Networ& mo'el bloc& grap( )*+,
9hen M& integrate! the intelligent machine and human
on production !ite, it fle"i#ly centrali/e! all activitie! !uch
a! order, de!ign, produce and !ale! of the company
through *no(ledge #a!e, data#a!e, computer! and
communication net(or* in order to improve the overall
efficiency. 1he !egment! of order, de!ign, produce and
!ale! are functional independent from each other, and
their !olution procedure! are )uite different. 1he
functional !u#!y!tem! of each !egment independently
complete manufacturing !u#0ta!*! a! (ell a! coordinate
(ith each other. 1herefore, in!ide the manufacturing
company, the (hole manufacturing proce!! from de!ign
to !ale i! a typical !olution procedure for Multi0Agent
)ue!tion!.
n order to improve the adapta#ility of manufacturing
!y!tem to the !tate change! from in!ide to out!ide, the
general !tructure de!ign of M& !hall o#ey the open
principle!, (hich are !ho(ed a!5 $,' openne!! of the ta!*5
the ta!* could #e input and proce!!ed at any time6 $:'
openne!! of the !y!tem5 the !y!tem !hall hold the change!
from interior !y!tem, and accept the interference from
out!ide6 $;'openne!! of !olution procedure5 the !olution
procedure !hall accept the change! of information and
*no(ledge. n order to reali/e the openne!! of the!e three
a!pect!, di!tri#uted !tructure i! re!orted to in the M&,
(hich endo(! every component entity and !u#!y!tem of
the !y!tem (ith larger autonomy to form the intelligent
autonomou! agent +:;-. ntelligent autonomou! agent i!
connected (ith computer communication net(or* #y
mean! of intelligent node!, (hich are e)ual in logic
di!per!ive in phy!ic and independent in function. 1he
node! have the loo!ely coupled relation! +<-, (hich
contact each other #y tran!ferring me!!age!. Ba!ed on the
common communication language, they coordinate and
cooperate to complete the manufacturing ta!*!.
III. A-"NT AND MU.TI/A-"NT
#0#T"M#
+.% Agent
An agent i! a real or virtual entity a#le to act on it!elf and
on !urrounding (orld, generally populated #y other
agent!. t! #ehavior i! #a!ed on it! o#!ervation,
*no(ledge and interaction (ith the (orld of other agent!.
An agent ha! capa#ilitie! of perception and partial
repre!entation of the environment, can communicate (ith
other agent!, can reproduce child agent!, and have o(n
o#jective! and autonomou! #ehavior.
According to Jenning! and 9ooldridge=! +,:- >>an agent i!
a computer !y!tem !ituated in !ome environment, and that
i! capa#le of autonomou! action in thi! environment in
order to meet it! de!ign o#jective!.== Under the conte"t
M&, (e can define an agent a! a !oft(are !y!tem that
communicate! and cooperate! (ith other !oft(are
!y!tem! to !olve a comple" pro#lem that i! #eyond of the
capa#ility of each individual !oft(are !y!tem. An
autonomou! agent !hould #e a#le to act (ithout the direct
intervention of human #eing! or other agent!, and !hould
have control over it! o(n action! and internal !tate!.
Page | 17
International Journal of Advanced Engineering Research and Science (IJAERS)
[Vol-1, Issue-3, Aug- !1"#
ISS$% 3"&-'"&(
An agent operate! in an environment from (hich it i!
clearly !eparated $%igure :'. ?ence, an agent $,' ma*e!
o#!ervation! a#out it! environment, $:' ha! it! o(n
*no(ledge and #elief! a#out it! environment, $;' ha!
preference! regarding the !tate! of the environment, and
finally, $3' initiate! and e"ecute! action! to change the
environment. Agent! operate typically in environment!
that are only partly *no(n, o#!erva#le and predicta#le.
Autonomou! agent! have the opportunity and a#ility to
ma*e deci!ion! of their o(n. @ational agent! act in the
manner mo!t appropriate for the !ituation at hand and do
the #e!t they can do for them!elve!. ?ence, they
ma"imi/e their e"pected utility given their o(n local
goal! and *no(ledge. 1he mo!t important common
propertie! of computational agent! are a! follo(!5
Agent! act on #ehalf of their de!igner or the u!er they
repre!ent in order to meet a particular purpo!e.
Agent! are autonomou! in the !en!e that they control
#oth their internal !tate and #ehavior in the
environment.
Agent! e"hi#it !ome *ind of intelligence, from
applying fi"ed rule! to rea!oning, planning and
learning capa#ilitie!.
Agent! interact (ith their environment, and in a
community, (ith other agent!.
Agent! are ideally adaptive, i.e., capa#le of tailoring
their #ehavior to the change! of the environment
(ithout the intervention of their de!igner.
$igure *. T(e agent an' its environment )%+,.
+.* Multi/agent systems
A multi0agent !y!tem $MA&' mean! a !y!tem in (hich
the *ey a#!traction u!ed i! that of an agent. t i! a loo!ely
coupled net(or* of pro#lem !olver! that (or* together to
!olve pro#lem! that are #eyond their individual
capa#ilitie!. 1he agent! may have only a partial model of
their environment and may po!!e!! a limited !et of mean!
for the ac)ui!ition and integration of ne( *no(ledge into
their model! and for pu!hing the !y!temA! !tate to(ard!
their o(n goal!. 1he *no(ledge of t(o agent!, referring
to the !ame thing!, i! not nece!!arily commen!urate and
may have different repre!entation!. No clo!ed0!y!tem
a!!umption ha! to #e maintained5 the MA& i! !u#merged
into and interact! (ith it! environment, (hich i! not
de!cri#ed completely #y formal mean!. 9henever novel
*ind! of interaction (ith the environment may occur, the
MA& !hould #e open and a#le to evolve.
n a community an agent ha! to coordinate it! action! (ith
tho!e of the other agent!6 i.e., to ta*e the effect! of other
agent!A action! into account (hen deciding (hat to do.
Boordination model! provide #oth media $!uch a!
channel!, #lac*#oard!, pheromone!, mar*et, etc.' and
rule! for managing the interaction! and dependencie! of
agent!. Boordination re)uire! !ome regulated flo( of
information #et(een the agent and it! !urrounding
environment, in other (ord!, communication. Note that in
a MA& coordination i! po!!i#le #oth #y indirect
communication via the environment, or #y direct
information e"change #et(een !pecific agent!. n any
ca!e, communication need! !ome language! (ith !ynta"
and !emantic!, at lea!t partially *no(n for each
communicating agent.
$igure +. -eneric sc(eme of MA# )%1,.
Bolla#oration mean! carrying out concerted activitie! !o
a! to achieve !ome !hared goal!. %or in!tance, in a
!cheduling domain machine agent! may agree on
e"ecuting each ta!* of a jo# (ith the aim of completing
an order #y the given due date. 1he !hared goal
$completing an order' can #e achieved only if all agent!
commit them!elve! to carrying out the action! they have
agreed upon. n general, meeting high0level o#jective!
and !ati!fying !y!tem0(ide con!traint! need cooperation
in a multi agent !y!tem (here agent! are !elf0intere!ted
and autonomou!.
Page | 18
International Journal of Advanced Engineering Research and Science (IJAERS)
[Vol-1, Issue-3, Aug- !1"#
ISS$% 3"&-'"&(
1he overall operation of MA& i! affected #y an
organi/ation that i! impo!ed on the individual agent!.
Even though there may #e no glo#al control or centrali/ed
data and the computation! are a!ynchronou!, !ome
organi/ational rule! al(ay! e"i!t. 1he organi/ation
determine! the C!phereD of the activity of agent!, a! (ell
a! their potential interaction! $!ee %igure ;'.
+.+ MA# arc(itectures
1he architecture! propo!ed in the literature for agent0
#a!ed manufacturing !y!tem! fall into three approache!5
the ?ierarchical approach, the %ederation one, and the
Autonomou! Agent one +::-. Any modern manufacturing
enterpri!e i! compo!ed of many, mo!t often di!tri#uted
phy!ically, !emi0autonomou! unit!, all having a certain
degree of control over local re!ource! or having varying
information re)uirement!. n !uch real !ituation!, a certain
num#er of agent0#a!ed indu!trial application! !till u!e the
hierarchical architecture.
A! regard! federation architecture!, the follo(ing
approache! have #een u!ed5 %acilitator!, Bro*er! and
Mediator!. %acilitator! are !everal related agent! (hich
are com#ined into a group. A facilitator i! a
communication interface #et(een agent!. Every
facilitator i! re!pon!i#le for en!uring communication
#et(een a local collection of agent! and remote agent!,
#y5 routing outgoing me!!age! to their de!tination!,
tran!lating incoming me!!age! for it! agent!.
Bro*er! re!em#le the facilitator! having t(o additional
function! !uch a! monitoring and notification. 1he
difference #et(een a facilitator and a #ro*er i! that a
facilitator i! re!pon!i#le only for a given group of agent!,
(herea! any agent may contact any #ro*er in the !ame
!y!tem for finding !ervice agent! to complete a !pecial
ta!*.
n addition to the function! of a facilitator and a #ro*er, a
mediator a!!ume! the role of !y!tem coordinator #y
promoting cooperation among intelligent agent! and
learning from the agent! #ehavior. 1he %ederation multi0
agent architecture! can to coordinate multi agent activity
via facilitation a! a mean! of reducing overhead!,
en!uring !ta#ility, and providing !cala#ility. 1he
Autonomou! Agent approach i! different.
1he autonomou! agent !hould have the follo(ing
characteri!tic! at lea!t5 it i! not controlled or managed #y
any other !oft(are agent! or human #eing!6 it can
communicateEinteract directly (ith any other agent! in the
!y!tem and al!o (ith other e"ternal !y!tem!6 it ha!
*no(ledge a#out other agent! and it! environment6 it ha!
it! o(n goal! and an a!!ociated !et of motivation!. 1he
Autonomou! Agent approach i! (ell !uited for
developing di!tri#uted intelligent de!ign !y!tem! (here
the !y!tem con!i!t! of a !mall num#er of agent! and for
developing autonomou! multiple ro#otic !y!tem!.
I2. A33.IATION DOMAIN# AND
ONTRO. #0#T"M O$ MA#
1echni)ue! from Artificial ntelligence have already #een
u!ed in ntelligent Manufacturing for more than t(o
decade!. ?o(ever, the recent development! in multi0
agent !y!tem! have #rought ne( and intere!ting
po!!i#ilitie!. 1herefore, re!earcher! have #een trying to
apply agent technology different domain of M& li*e,
manufacturing enterpri!e integration, enterpri!e
colla#oration, manufacturing proce!! planning,
!cheduling and !hop floor control, material! handling and
inventory management. %igure 3 illu!trate compari!on
#et(een Bonventional v!. Multi0agent manufacturing
control !y!tem. &ome important application domain! of
agent! in MA& are de!cri#ed #elo( +:F-, +:,-5
$igure 1. onventional vs. Multi/agent manufacturing
control system )%4,
8rder agent, repre!ent! an order to #e accompli!hed
#y the production !y!tem.
Groce!! planning agent, plan of the !everal
proce!!ing pha!e! to produce a (or* piece of an
order.
Groce!! !cheduling agent, minimi/e the production
time and co!t! from proce!! planning.
Boordinator and &upervi!or agent coordinate and
!upervi!e the action! #et(een different agent!
impo!ing the correct e"ecution of the rule!
e!ta#li!hed in the !y!tem.
Page | 19
International Journal of Advanced Engineering Research and Science (IJAERS)
[Vol-1, Issue-3, Aug- !1"#
ISS$% 3"&-'"&(
@e!ource agent!, have the re!pon!i#ility of manage
different re!ource!. %or e"ample (or* piece agent
manage! the proce!!ing !tate of the (or* piece, the
tran!port agent decide! autonomou!ly in (hich
direction a (or* piece i! for(arded in!ide the
production !y!tem, and the machine agent, control!
the machine.
2. AD2ANTA-"# AND !A.."N-"# O$ MU.TI
A-"NT #0#T"M#
5.% A'vantages6
1here are many advantage! +,H- +,I-, provided #y the
characteri!tic! of MA& related to5
,. 1echnological and application need!5 Multi agent
!y!tem offer a promi!ing and innovative (ay to
under!tand, manage, and u!e di!tri#uted, large0!cale,
dynamic, open, and heterogeneou! compounding
!y!tem.
:. Natural vie( of intelligent !y!tem!5 ntelligent and
interaction are deeply and inevita#ly coupled, and
multi agent !y!tem! reflect thi! in!ight. Natural
intelligent !y!tem!, li*e human, do not function in
i!olation6 they interact in variou! (ay! and at variou!
level!. MA& provide in!ight and under!tanding a#out
poorly under!tood interaction among natural,
intelligent #eing!, a! they organi/e them!elve! into
variou! group!, committee!, !ocietie!, and economie!
in order to achieve improvement.
;. Bomple"ity management5 1here are 3 major
techni)ue! for dealing (ith !i/e and comple"ity of
enterpri!e information !y!tem!6 modularity,
di!tri#ution, a#!traction, intelligence. 1he u!e of
intelligent, di!tri#uted agent! com#ine all four
techni)ue!
3. &peed0up and efficiency5 Agent! can operate
a!ynchronou!ly and in parallel, and thi! can re!ult in
an increa!ed overall.
4. @o#u!tne!! and relia#ility5 1he failure of one or
!everal agent! doe! not nece!!arily ma*e the overall
!y!tem u!ele!!, #ecau!e other agent! already
availa#le in the !y!tem may ta*e over their part.
7. &cala#ility and fle"i#ility5 1he !y!tem can #e adopted
to an increa!ed pro#lem !i/e #y adding ne( agent!,
and thi! doe! not nece!!arily affect the operationality
of other agent!.
<. Bo!t!5 t may #e much more co!t0effective than a
centrali/ed !y!tem, !ince it could #e compo!ed of
!imple !u#!y!tem! of lo( unit co!t.
H. Development and reu!a#ility5 ndividual agent! can
#e developed !eparately #y !peciali!t!, the overall
!y!tem can #e te!ted and maintained more ea!ily, and
it may #e po!!i#le to reconfigure and reu!e agent! in
different application !cenario.
I. Grivacy5 A centrali/ed approach i! not po!!i#le
!ometime! #ecau!e !y!tem and data may #elong to
companie! that for competitive rea!on! (ant to *eep
them private.
5.* (allenges6
Although the advantage! di!cu!! a#ove, there are !till
challenging )ue!tion! to #e an!(ered !uch a! +,I-,
,. ?o( to ena#le agent! to decompo!e their goal! and
ta!*!, to allocate !u# goal! and !u# ta!*! to other
agent!, and to !ynthe!i/e partial re!ult! and
!olution!J
:. ?o( to ena#le agent! to communicateJ 9hat type of
communication language! and protocol! to u!eJ
;. ?o( to ena#le agent! to repre!ent and rea!on a#out
the action!, plan!, and *no(ledge of other agent! in
order to appropriately interact (ith themJ
3. ?o( to ena#le agent! to recogni/e and reconcile
di!parate vie(point! and conflict!J ?o( to
!ynthe!i/e vie(! and re!ult!J
4. ?o( to engineer and con!traint practical multi agent
!y!temJ ?o( to de!ign technology platform! and
development methodologie! for manufacturing
!y!temJ
7. ?o( to effectively #alance local computation and
communicationJ
<. ?o( to ena#le agent! to negotiate and contractJ
9hat negotiation and contract protocol! !hould they
u!edJ
H. ?o( to ena#le agent! to form and di!!olve
organi/ational !tructure!J
I. ?o( to formally de!cri#e multi agent !y!tem! and
the interaction among agent!J ?o( to ma*e !ure that
they correctly !pecifiedJ
,F. ?o( to reali/e >intelligent proce!!e!= !uch a!
pro#lem !olving, planning, deci!ion ma*ing, and
learning in multi agent conte"t!J ?o( to ena#le
agent! to collectively carry out !uch proce!!e! in a
coherent (ayJ
2I. T!" M"T!ODO.O-0 O$
MI-RATION TO MA#
&ince multi0agent technology ha! #een recogni/ed a! a
*ey concept in #uilding a ne( generation of highly
di!tri#uted, intelligent, !elf0organi/ing and ro#u!t
manufacturing !y!tem, the traditional concept of
manufacturing !y!tem! ha! #ecome vulnera#le to change!
+I-. Environmental change!, failure detection,
reconfigura#ility, and e"panda#ility6 are a !et of
capa#ilitie! that ma*e an attractive option migration from
Page | 20
International Journal of Advanced Engineering Research and Science (IJAERS)
[Vol-1, Issue-3, Aug- !1"#
ISS$% 3"&-'"&(
traditional to MA&. 1he migration !hould #e !uch that it
i! ea!ier to adopt ne( production infra!tructure! (ithout
dramatic hard(are change! and long !etup time!.
Gre!ently, there are !everal topologie! of manufacturing
cell!, !uch a! centrali/ed, hierarchical, and heterarchical
!tructure! +,F-. Each of thi! topology could #e con!idered
a! optimal and a#le to accept migration, ta*ing into
account that each element !hould #e related (ithout
complete dependency and (ell functioning !hall not #e
compromi!ed (ith any other element from the cell, after
migration i! implemented. A generic platform (a!
de!igned in order to apply MA&. 1he platform de!ign (a!
implemented in !uch a (ay that any .fle"i#le
manufacturing cell could #e evolved into agent0#a!ed
!tructure. 1he clue i! to adopt the platform !tructure, and
!hape each element $ro#ot!, numeric control machinery,
conveyor!' of a cell to ac)uire agent per!onality. 8nce the
pro#lem or pro#lem! are identified the MA& de!ign
pha!e, !tart!, (hich i! more oriented to(ard the
implementation of the generic platform6 ho(ever a
methodology !hould #e committed.
$igure 5. !ierarc(y mo'el of communication an'
interaction)%%,
%igure 4 !ho(! ho( agent! can #e integrated and ho(
data (ould #e interchanged, (herein the three layer!, that
computer !y!tem in manufacturing management u!e, are
illu!trated. 1he generic plat0form i! to(ard from general
to particular application, !o #efore !tart (or*ing on
developing intelligence, i! crucial to ma*e independent
each element, (hich i! !uppo!ed to emerge from a
centrali/ed and !e)uential architecture that actually !hall
#e !u#!tituted #y the ne( platform. 1he !uperior part of
thi! pyramid i! performed #y management layer, (hich
are !ati!fied (ith a manufacturing planning level, and a
manufacturing e"ecution level. Both could #e
programma#le holon!, purely !oft(are #a!ed. n addition
pyramid #ottom i! formed #y e"ecuta#le holon!, (hich
ha! direct contact (ith machinery, and hard(are !y!tem!,
al!o thi! part of the pyramid fre)uently i! the one (ith
more con!traint! in manufacturing environment!. 1he
effort! on thi! !ection (ill #e driven to get the pyramid
#a!e prepared to #e adapted (ithout neither hard(are
change! nor de!ign, on the other hand ready to #ecome
reconfigura#le, and holonic0ready +I-, the methodology a!
!ho(n in figure 7 !hall #e de!cri#e in four !tep!+,,-.
,. Define Bommunication &tructure
:. !olate from glo#al !y!tem
;. Bonvert from general to particular
3. Breate relation!hip! #ut not dependence!
$igure 7. T(e met(o'ology mo'el implemente' to
ac(ieve migration feature )%%,
1he re!ult after thi! methodology (ould #e (hat (e call a
Kholonic0ready agentK, (hich meaning contemplate! an
entity (ith characteri!tic! and attri#ute! nece!!arie! to
adopt intelligence #loc*! $to #ecome a ?olon' in order to
achieve deferent function! or ta!*!. An over0all vie( of
the re!ulting platform i! !ho(n in a more oriented (ay
the methodology applied on the commercial !oft(are
u!ed to develop the generic platform. 1he methodology
ma*e! reconfigura#ility po!!i#le into the manufacturing
cell.
Page | 21
International Journal of Advanced Engineering Research and Science (IJAERS)
[Vol-1, Issue-3, Aug- !1"#
ISS$% 3"&-'"&(
2II. A-"NT# D"2".O3M"NT# TOO.#
Development of the agent0#a!ed manufacturing !y!tem!
(a! u!ually carried out !o far u!ing !uch programming
language! li*e BLL, Java, Mi!p, Grolog, 8#jective B and
&mall1al*. ?o(ever, efficient agent development tool!
are urgently needed $1a#le :'.
Effort! have #een made to !et forth !tandard! for agent0
#a!ed !y!tem!6 ho(ever there are no accepted !tandard!
for developing agent #a!ed manufacturing !y!tem!. 8nly
a fe( project! can #e named, li*e NOMM intended to #e a
common communication language for agent!, or N%
#eing a common content format. 1he %oundation for
ntelligent Ghy!ical Agent! $%GA' promote! the
development of !pecification! of generic agent
technologie! that ma"imi/e interopera#ility (ithin and
acro!! agent0#a!ed application! +::-.
Table %. "8emplary agent 'evelopment tools )**,
2III. ON.U#ION
MA& ma*e the manufacturing !y!tem ea!ier to de!ign,
reduce the comple"ity, inten!ify the recom#ination,
e"panda#ility and relia#ility of the !y!tem, and improve
the fle"i#ility, adapta#ility and de"terity of the !y!tem. t
can offer di!tri#uted intelligent control action! to create
evolva#le !y!tem! re)uired on the fle"i#le and di!tri#uted
manufacturing !y!tem! already needed in our! day! and
e!!ential for the future. 1hi! paper pre!ent! the revie( on
multi agent #a!e intelligent manife!ting !y!tem and dealt
(ith !tructure, architecture, application domain!,
advantage!, challenge!, migration to MA& and agent
development tool!.1he main characteri!tic! to implement
M& #a!ed on autonomou! agent! are pre!ented and
concept and u!e of MA& i! e"plained.
R"$"R"N"#
+,- E.J. Mee, A. 1oguyeni, N. Dangoumau, CA Getri Net #a!ed
approach for the &ynthe!i! of Gart!A Bontroller! for
@econfigura#le Manufacturing &y!tem!,K &BE0BA&E
nternational Joint Bonference :FF7, !ice, pp. 447<044<:,
:FF7.
+:- Armando Bolom#o, @onald &choop and @alf Neu#ert, CAn
Agent0 Ba!ed ntelligent Bontrol Glatform for ndu!trial
?olonic Manufacturing &y!tem!D, EEE 1ran!. nd. Elect.,
vol. 4;, no. ,, pp. ;::0;;<, %e#. :FF7.
+;- &. Bu!!mann and N. &child, CAn Agent0#a!ed Approach to
the Bontrol of %le"i#le Groduction &y!tem!D, in Groc. Hth
EEE nt. Bonf. on Emerging 1echnologie! and %actory
Automation, vol. :, pp. 3H,03HH, :FF,.
+3- @affaello, M.D Advanced humanPmachine !y!tem for
intelligent manufacturingD. Journal of ntelligent
Manufacturing ,<$7', 74;P777 $:FF7'
+4- &ha(, B.DGreliminary de!ign and manufacturing planning
integration u!ing (e#0#a!ed intelligent agent!D, Journal of
ntelligent Manufacturing ,7$3', 3:;P3;< $:FF4'
+7- Qang, G., Miao, N.B.DA mi"ed i!omorphi!m approach for
*inematic !tructure enumeration graph! #a!ed on intelligent
de!ign and manufacturingD, 1he nternational Journal of
Advanced Manufacturing 1echnology ;,$I', H3,PH34
$:FF<'
+<- ndira, M.D%u//y0#a!ed methodology for multi0o#jective
!cheduling in a ro#ot0centered fle"i#le manufacturing cellD,
Journal of ntelligent Manufacturing ,I$:', I;PI< $:FFH'
+H- Jo!e Ma!tra and van Delamer, C&emantic 9e# &ervice! in
%actory Automation5 %undamental n!ight! and @e!earch
@oadMapD, EEE 1ran!. ndu!trial nformatic!, vol. :, no.
,, %e#. :FF7.
+I- 2an Meeu(en, E.?., Norrie, D.Dntelligent manufacturing5
holon! and holorachie!D, Manufacturing Engineer,
<7$:',H70HH, $,II<'.
+,F- &aad, N. Na(amura, and G. Bi!(a!, Gerformance
evaluating of contract Net0Ba!ed hierarchical !cheduling
for %le"i#le manufacturing !y!tem!.
+,,- Jorge Gam#oa0@evilla and Miguel @amRSre/0Badena.
Cntelligent Manufacturing &y!tem!5 a methodology for
technological migrationD, Groceeding! of the 9orld
Bongre!! on Engineering, :FFH 2ol. 9BE :FFH, Mondon,
U.N.
+,:- N.@. Jenning!, M.J. 9ooldridge, Application! of intelligent
agent!, in5 N.@. Jenning!, M.J. 9ooldridge $Ed!.', Agent
1echnology5 %oundation!, Application!, and Mar*et!,
&pringer, ,IIH, pp. ;P:H.
+,;- @u!!el, &., Norvig, G., ,II4, Artificial ntelligence5 A
Modern Approach, Grentice ?all.
+,3- Jenning!, N.@., :FF,, An Agent0Ba!ed Approach for
Building Bomple" &oft(are &y!tem!, Bommunication! of
the ABM, 33E35 ;403,.
Page | 22
International Journal of Advanced Engineering Research and Science (IJAERS)
[Vol-1, Issue-3, Aug- !1"#
ISS$% 3"&-'"&(
+,4- 9arnec*e, ?. J.5 1he %ractal Bompany. &pringer02erlag,
Berlin, Germany $,II:'
+,7- Mandel#rot, B.B.5 1he %ractal Geometry of Nature. &an
%ranci!co. 9. ?. %reeman $,IH:'
+,<- n!titute for Manufacturing $%M', Univer!ity of
Bam#ridge, http5EE(((.ifm.eng.cam.ac.u*E
automationEre!earchEdi!tri#utedTintelligent.html
+,H- ?uhn! M.N., &tephen! M.M., ,III Multi agent !y!tem! and
!ocietie! of agent!, in Multi agent !y!tem, ed. 9ei!!, G.,
&BN F0:7:0:;:F:0F.
+,I- Multi agent !y!tem!, A modern approach to di!tri#uted
artificial intelligence, ,III, ed. 9ei!!, G., &BN F0:7:0
:;:F:0F.
+:F- Armando Bolom#o, @onald &choop and @alf Neu#ert, CAn
Agent0 Ba!ed ntelligent Bontrol Glatform for ndu!trial
?olonic Manufacturing &y!tem!D, EEE 1ran!. nd. Elect.,
vol. 4;, no. ,, pp. ;::0;;<, %e#. :FF7.
+:,- &. Bu!!mann and N. &child, CAn Agent0#a!ed Approach to
the Bontrol of %le"i#le Groduction &y!tem!D, in Groc. Hth
EEE nt. Bonf. on Emerging 1echnologie! and %actory
Automation, vol. :, pp. 3H,03HH, :FF,.
+::- J. Madej!*i, C&urvey of the agent0#a!ed approach to
intelligent manufacturingD, Journal of achievement! in
material and manufacturing engineering, volume :,, i!!ue
,, March :FF<
+:;- Oinglin Guo and Ming Uhang, C@e!earch on ntelligent
Manufacturing &y!tem Ba!ed on Multi0AgentD B@A
:FFH, Gart , pp. H:IPH;H
Page | 23

Anda mungkin juga menyukai