lc-Ol 5
Copyright © 1996 lFAC 13th TrienniaJ World Congres~ > San Francisr.;o, USA
GIM (GRAI INTEGRATED METHODOLOGY) AND ITS MAPPING ONTO GERAM D. Cllea. G. Do...ela.ts. B. VaDesplr GRAIiLAP, Univa'Sity BoIdcaux I 351, Coon de la iibtraIioo, 33405 TaIcnce cedex, France TeI: (33) 56 84 6S 30, Fax: (33) 56 84 66 44
Atngpq . The papcI" peseDts the Ilaaic coocepIB IIId components of GERAM (Generic EnJ.erprise Reference An:hiII:cIIR and MeIbodoIogy) developed wiIbiD !be Iiamc of IFAC/IFlP Task Fon:e and !be GRAllnlegrated MetbodoIOIIY (GIM) eiJOOralcd by ORAJlLAP. The mapping of OIM onto GERAM is demonstrated and how GIM can INWCf !be set or requiremenIs or OERAM is discussed
Keywgn!s : Enterprise Integmlion, Enterprise ModeIIiDg, Architecture, MeIhodoIogy,
1. BACKGROUND Wi!hin !be frame IFAC/IFlP Task Pcccc on An:bi1ecturcs fir Enterprise InIegraIion, Ihrcc approaches wee selected and a detailed comparative analysis was performed in Irdcr to ~ recommendaIions for futme work in Ihis ..... The Ihree aJlillUKbea arc : CIMOSA developed by Amice Conaortium wilhin !be frame of Esprit Pmgnunmc (Amk:c, 1992), GlM developed by GRAI/LAP of University Bordmux I (Doumcingts, et 1992) and PERA deveIopcd by Purdue Laboratory for Applied Induslrial Cootrol (Williams, 1992). A fust proposal made by Task Fcrce, named as GERAM (Generic Enterprise Reference An:hitectme and Methodology), was anthored by Peter (Bemus, et 1994a). This paper intends to demonstrale how the GIM fits with !be GERAM framewcrk IDd how GIM CIIII meet !be set of requirements of GERAM (Williams 1995).
methods, models and tools which are needed to build !be 1994b). GERAM itself integrated enlerprise" (Bemus, et is a framewcrk of six major elements as follows (Bemus, et al., 1994a; WiUiams, cl 1994) :
m.,
m.,
(1) Qrmris Entmwic Bcfnmre ArclJjtectyre (GFRAl : This is the definition of enlerprise related coocepIS, with the primary focus on the life-cycle of the enterprise. Since the life-cycle can be consideled as a design process the architecture will also have to identify the inrermediary results and componentS of this design process. It can be considered as the model of the life cycle.
m.,
m..
(2) Generic Enterpri:sc; Engipr&ring Methodology CGEEM) : This is the descriplion, on a generic level, of the processes of enterprise integration. In other words, the methodology is a detailed process model, with instrnctions for each step of the integration projecL
2. GERAM: GENERIC ENTERPRISE
(3) Generic Entqpdsc Modelling tools Md LaoRlllges (QEMT4L) : The engineering of the integrakld enterprise is 8 highly sophisticated, mUltidisciplinary management, design and implementstion ellCrCise during which various firmS of descriptions and models of the target enterprise need to be created. To cxPRSS
REFERENCE ARCHITECIURE AND
ME'IHODOLOGY GERAM was built upon the wOlb carried out earlier by CIMOSA, GRAIlGIM and PERA. "GERAM is about those
648
these models more than one modelling language may (2) GlM Mgdc:Iling Fgmcwork : it is composed of I!Iree axis: !be view and SYSIem axis (function, phyaical, decisionaJ , inflll1llalional), the life cycle axis (analysis, design, user and IechnicaJ specifJCalions), abstraction level (concepIWII, otructuraI, realisariooaJ).
be.-bl. (4) Generic Entcmrjse MQdcls
(3) GIM Reference; Ary;hi'C&tnre : an architecture for enterprise inlegralion can be defined as a sII\ICtUrcd set of 'models' wbicb represent the invariant building blocks of the Encerprisc sySlem. The GIM Referm:e arcbiteclUre is defmed wilhin Ihe GlM Modelling Framework. It was devdoped witbin Ihe Esprit Project IMPACS.
(5) Qmeric EntrmjJC McvIpk' : Modules &le products, wbich are SWldard implementations of components Ihat are likely to be used in enrerprise inlegraDon project or by Ibe encerprisc itself. Generic modules C8II be configwed to form more complex modulcs fO' die use of an individual enterprise.
(4) QIM MPdcl!jng fgrmaUams : a fonnaJism is a Ianguaae to represent Ibe CIJOCCIlIS of die Enterprise System. Currently, mER) is used for Ihe functional modelling of the enterprise system; IDEFO and SIR (Stoct/Re90urcC) are used to describe a physical system; GRAI grid &; GRAI nets are used to build a decision system model; entity/relationship i. used to model an information system.
(6) r...."...;c Eptmrj:le Jbmir.1 CGTs) : Theoriea wbicb desc:ribe Ibe most generic aspects of cotcrprise.reIaIcd concepts. GeoeniIly called ontological dtccries. They may also be consideRd "meta models" because lbey consider facts and rules &boot Ihe facts and rules of enwprise models.
"GBRAM is tbus not yct another new proposaI for enterprise reference arcbir.ecture, but a framewOllt whicb is meant to organise Wsting CIWlIprise inlegralion knowledge instead of redefining it" (Bemus. ~I al., 1994a).
(5) GIM StrucJpred AJlPIllIICb : it is a guide to perfo-m the analysis and design of an ente'lJrise. It contains I!Iree main phases : exiating system modelling and analysis (AS IS), User Oriented Design (TO BE), Tecbnical Oriented design (Components Specifications). This .auctured approach allows to swt by die actual system (at die rcalisalionallevcl), to model at Ihe conceptual level, to design the new system at Ibe game level based on reference model : die result of GIM is the components specirlcalions at Ibe realisalionaJ level.
3. GIM : GRAI INTEGRATED
ME1HODOLOGY GIM is developed at Ihe GRAI group of Ihe LAP (Uobomtoire d'ADIOI\Iatique et de Productique) of University Bordeaux I. It is based 011 several PhD rcsean:b works (Doumeingts, 1994; Roboam, 1988; Chen, 1988; VaIlespir, 1987; Fari~, 1994) and implemented witbin the frame of sevaal European projects: Esprit project Open CAM System (Roboam, ., al., 1987), Esprit project IMPACS (Inlegrated Manufacturing And Control System) (lmpacs, 1992), Esprit project FLEXQUAR (Adaplative System for Flexible. High Quality &; Reliable Productioo) Downeingts, ., al., 1993), EUREKA project TIME GUIDE (Tool and method for Integration and Management of EWlIution of industriaIlinns).
(6) GIM CASE JQgI : In order to support Ibe use of GlM, a tool "PROGRAI" was developed by GRA.I/l..AP on Macintosh CDvironment Currently a tool "IMAGIM" is being developed within die flame of an EUREKA project "TIME TOOL". IMAGIM wiu run on PC Windows and is now distributed by INTEGRAl, Ihe research transfer company of GRAI/LAP, and three olhers partners (SBLISA, ISERPA, FORDESI (P) .
GIM is composed of the foUowing components : (I) GRAl CpncepwaI
4. nm MAPPING OF GIM ONTO GERAMFRAMEWORK
Modd : it is Ihe reference model
for the control of an enterprise. It describes all basic concepts. The enterprise model is decomposed into I!Iree sub-systems: Physical system (it deacribes Ibe transformation of tbe flow of products by !be resources), Decision syslem which manages and controls the physical system; and InformaliOll system which provides Ibe decision system wilh information needed fO'decision-making.
Most of tbe existing elements of GRAI Integrated Methodology find lheir place in Ibe GERAM framework in Ihe following way as summarised by Ihe table I (Cheu, ~I al., 1995) :
649
GB&I£iilld
6EB6M •
• GIM Modelling PmmewoII<
• GERA : Generic EnIr:2prise
• GIM Reftaeoce An:hi1ecmre
Refcftnr.e Arcbi_
• GEEM : Generic Enterprise EngineeriDg
• GIM StrucIIJnId ApJxoaclt
Met·· >doIogy
• OEMT&L : Generic Enterprise
• GIM Modelling Forma1isms &: tools
Modelling Tools &Languages
• ORAl grid and nets ·IDEFO • Enlity/Rela!ionship
••• PROGRAI, IMAGIM • GEMs : Generic EnIr:2prise Models
• OIM Rcfemx:e ModeJs • GRAI reference models
• GMs : Gencric Enterprise Modules
• GIM Tecbnological CompooenlS Jilnry (10 be developed)
• GTs : Generic Enterprise Thecries
• ORAl ConcepIUaJ Model It is based on :
• General sysIeI1llhea'y • Theory of hiemrt:hica1 systems • Theory of discreet activilies • Production mt
meo..v
Table I : Mappina of ORAI/GIM 01110 OERAM
- The OIM Modelling Fn!mcwork is onc of the elcmcnlS of OIM approach. One of the orijpnalilies of OIM Modelling Fnnework is ilS gencricity I!Id openness because i. offers !he possibility 10 define various reference architcctures widlin !be same framework. Foe... is made OD !be fact Ibat
At the Conceptual level, !be analysis and user oriented
design are performed; III the SllUCturallevel, the four views models (FUIIClion, Ioformaliou, Decision and Physical) are IrIInsfonned into three domains models (Infonnation Technology components, Technology components and
the Life cycle axis and Abstraction level axis are independent At a given phase of the Life cycle, one can build • model III the SIrUClInIIcvel, andIor concepIII8Ilcvel and/or IaIisationaIlevel, accorcIiDa 10 !be type of the project 8Dd/or die type oi die mlr:2prise concerned. At die suuctural level, !he four models (Function, Physical, Decision and Information) are IrIInsformed into !brec domains models (Informaticn Technology, TccbnoIogy and Organisation). This is because Ibat modeJs are user orimted and conceptual subject wbilc system COIIIJIOIIeIIIS are technical oriented and impbnclllation concerned.
According to the agreed undetstandiog within !be group of Task Force, the !enn "archilecture" here should be understood as the architecture for designing and implementing the enterprise model, instead of the architecture itself. This leads 10 !he classification of two types of architectures : type I (those architectures representing an enterprise) or Type 11 (archi!eCtures of the program or project for developing an integrated enterprise or part of thmot) (Wi11inms. 1994).
- The OIM Refeieoce Arcbi_ ddincs die various models that are necessary 10 build the global model along the eotapriae life cycle at the two levels: - !be abslraclion levels - the views
- The OIM structured approach corresponds TO the GEEM of GERAM. It consists of !brec main phases: • Analysis, • User Oriented Design, • Technical Oriented Design. The originality of GIM is 10 start by the .wysis of the existing system. Even if !he system is NEW, there is always an existing situation : knowledge about techniques, organization, example of realization of the prodUCL We call
Organisation).
650
IbiJ pbue "ANALYSIS" or "M Is" analysis. It c:onsists in eI8bonIing a model of "cxislin& sySlelll". 1be interest of
IbiJ "exiIlini model" or "poc:udo exiIlini Model" (in cue of aeaIioa) is to all.". an ..- to bave • clear and precUe doIcripIioD of the problem. TIIia phase bcIpa the user to refiac the Uaen Requi"liaallld to defiDc in a JRcise _y. his objecIiYCS. It is well m-. Iba! die initial -am of the U.... Requiremcoll and Objectives of • Future lie detWld in vtI'/ VIIPJC lenDS.
S,....
- GIM Modelling Fonna1isma oorrespood to die GEMTdtT of GERAM. 1be modeIa cIabInted uaing GIM fonnaJiams must be cobcrent ~ tbcm. All the iofOllD8lion used and 1Icnera&ed in lbe functiooal. lbe physical 8IId the decisioaBI models lR iqJiesenk:d by 811 unified data model in die ioformatim view.
- In GIM, RcfcIaIcc models can be used 8IlWO SleJlS : onc is to help the deleCtion of inconsiSleocies of an existinll system by comparing die mode1 of an existing sysiCllt to die reference model; anochcr is It lbe step of ..- oriented desigu. Cmrenlly -wrk is beinl! performed to popular the GRAI pjd Reference models. We ccnsidtz paiSible to define reference models according to die type and size of enterprises. For example. within the Esprit project IMPACS. a decisiona1 reference model is elaborated for the IMPACS type systems.
(WiIliams, 1995). In the following we will GIM can meet tbcsc requin:ments.
discUS8
how
R"IIPilJ!!!l!!nl I. "GERAM md be a Complete Architecture and McthodoIogy. i.e., it must cover all of the deIaIIs of the life history of any coterpriac, entity or sysiCllt from its initial concepl Jbrooghout all aspects of its life history nntil and including its operation and final disposal". The Architecture part of GIM is composed of 'GIM Modelling framework' and 'GIM reference architecture'; \be MeIhodology part is composed of 'ORAl conceptual model', 'modelling formalisms' and 'struclUred approach'. Complied with die definition of die GERAM Requirement I, GIM today does nOl cover "all aspects of its life history". But in the frame of \be EUREKA project TIME (Tool and method for Integation and Management of Evolution of industrial firms). the TIME GUIDE Methodology whicb is in developmcllt (\be version 2.0 will be ready in September 95) and which used GIM will cover all aspects of the life cycle includinl! the development pbasc, the implemenlalion pbue and the opezation phase. Req1ri!'!l!!lCnt 11. "GERAM should specific class or type oC systems information systems, CIM, elC.), bandlinl! the description and co-enlerprise, entity or system".
nOl be confmed to any (discrete manufacture, but should be able of development of any
• Concerning the GMs (Generic Enterprise Modules). work
wiII be performed 10 cIaborate a technolol!icaI c:ompmcnlll hllrary. These components will be classified as : • IT (Information Tccbnology) components including applications. dal8bases. ... • T (TcchnoloiY) components such as machines. human opmIIOi'S. computers ... • Generic Enterprise Theories (GTs) describe die most I!CDCric aspects of enterprise related concepts. In GIM. the GRAI Model is developed to represent lbe set of generic cOllCcplS used to modeIlIId design a manuflCluring system. Work is concentrated at die level of die Decisiona1 System. Example of the set of generic c:oocepIS are : Decision Level which is characterised by Horizon and Period; Decision CenW. Decision Flame. etc. 1be GRAI Model is based on scveraIlheoretical approaches such as : the Genend SysiCllt Theory (Le Moigne. 1984). the Theory of Discrete Activities (Pun, 1977; Pun, 1984), contributions having an criI!in in Conb'ol Theory (MtJese, 1979) and the Hierarcbical Coob'ol Theory (Mesarovic, et al., 1970) as well as contributions from Production Mansgement theories.
5. HOW GIM CAN ANSWER GERAM REQUIREMENTS Within lbe frame of IFACJIFIP Task Force, a set of GERAM requirements are developed in seven points
GIM was developed 81 the bel!inning for the purpose of the analysis and design of manufacturing systems. Howevet, because of the 1!Cftericity of the GRAI model, GIM can be used as well to analyse and to design other !rind of enlelprise system as one-of-a \rind production, projects, services. 1be latest applications of GIM concern BANK (in Bordeaux), Educational depaJtment (Institute University of TecbnoIolY in Agen near Bordeaux) and Msembly line (SCANIA, Sweden). Reqnirement DI "GERAM should be able 10 incorporate, present and utilize all of the usefu1 capabilities of any and all pertinent type 2 architectures. Thus, it should conlain the IOtaI capability of all type 2 architectures proposed by die Task ~ and others". 1be architecture part of GIM is generic in a SCDse Ihat GIM modelling framework is open to other architectures depending on die concepts used. If we come back 10 die SlIt.systems and view dimension, GIM modelling framework is able 10 accept new dimension on the condition that the integration with existing model is coherent For abstraction axis and Life cycle axis, the concepIs are generic enough to be coherent with the concepts used by other archileClUreS.
Reqnirement IY. "The methodology associated with GERAM should be able 10 provide all die necessary IlUideIincs and management techniques for the initiation .wI pursuit of a project or program of development and operation of any enterprise or entity .....
651
6. CONCLUSION The SInII:IIIICd approacb of OIM provides IIlOSI of guidelincs and. manaaemeat tcc:bniques necessary 10 conduct IIlI enlerpri8e dcaigD or re-eop-nnll projecL The GIM structured approacb defines the oumber IIlId the II8IIIre of groupI tbal are "CC"""'Y, die reIaIions betwcm the poups and. work proc:edures, the ~ wilb ~ Idnds of meetinp 10 be beId, the lOt of c:becIr: points, intermediate tqIOI1s ••• One of die fundtmmlwl concern of OIM lIInICIuIed approach is 10 invol~ as much as pOlSiblc people of the cntapriIe. 'This the deaipc:d sylllCm 10 bcuer meet the user requimnents, and. c:onsequendy increase the user's accepcabiIily of the design syIIaD. CertsioIy the lIInICIuIed approecb is a IIrong advaoll&C of GIM due 10 tbc nmnber of applicalions in indusay and in services.
.no.,.
RcqujmnmlS V. "The framework ... overaJI graphical fotm of GERAM should be able 10 assist users io tbe interpretation and use of the associaled methodologies of GERAM .... •
The GIM modelliog framework and the GIM structured approech allow users 10 understand and 10 apply GIM methodology. Based 00 the more than ooe bundred experiments, a lot of improvements were introduced since the last decade. A precise documentation is giveo today 10 the user and the software support tool (lMAGIM or PROGRAI) can provide with a lot of helps. ReoujmneplS VI. ·GERAM must be able 10 show the place of buman in all aspects of their involvemenl in the
EnWprise ... •. The hwnan aspect must be tateu inlO account as much as possible. The current versioo of GIM takea tbc human aspect inlO accounl by two considerations. The first one is 10 analyse buman decision-making structure of a manufacturing system or of any organisation. The second one is 10 use a structured approecb 10 inyol~ as much as possible the users in the design of a manufacturing system or of an organisation. Moreover GIM allows 10 make choice in the design phase between the activities that wiD be aUlomated and tbc ooes that will remain manual. New research work undertaken with specialists of cognitive ergonomics will also improve tbc place of human in GIM. Rcqujremem
vrr. ·GERAM should presenl a standardized
glossary f... use in enterprise integration engineering and other systems eogineering effOlU and provide a semantics and syntax 10 promoIC ~ understanding in projects and other cooperative efforts in this &res •• There exists a basic set of concepts and ICmI definitions in GRAI/GIM approach. A standardized Illoasary is in developmenl and tbc software support tool IMAGIM wiD allow 10 lint the vocabulary of the enterprise 10 tbc generic
glossary.
In Ibis paper, we disc:uased tbc mapping of GIM 0010 GERAM. Globally speaking, GIM fits within the GERAM framework. Most of tbc OIM elements can be mapped ODIO tbc GERAM and. the basic aJIICepIS of tbcBe two IAJI08Chea COD1IeI'Je. ID geoeral, IIlOSI of the requirements of GERAM can be met by GIM. We consider that one of the importanl contributions of GERAM is lbat it in fact allows 10 standardiz.e the terms by federating concepts of ~ approacbea and thus enrich the theory on enterprise
architecture and methodology. REFERENCES Amice (lm). Project D~liw,abl~s, Mi1estone M-2, Project 5288, VOLUME 3, Formal Reference Base, Esprit COIISIB'Iimn AMICE.
Bemus, P. and. Nemes, L (1994a). A. F'QJIIeW(),k 10 defiM a GeM,ic Ent~,p'ise R"/~,ellce A.,chit~ctu,~ alld M~thodology (GERAM), Divisional report n· : MTM 366. CSIRO Division of Manufacwring Technology. Bemus, P. and Nemes, L (l994b). A. F'OJIItWo,k 10 dejiM a G~M'ic ElIt~,p'is~ R~/"'''"c~ A.,chitectu,~ and Methodology, ICARCV'94, Singapore. Chcn, D. (l988). DlagllOstic des IfIColtirmces MS sysUtMS M gurio1I M prodMCtioll assiste par o,dillOle/IT, ThNe de Doctorat. AUIOIDatique. Universire Bordeaux L Chcn, D., Va11eapir, B. and. Doumeingts, G. (1995). GIM (G,al Ift/egr/lllld Methodology) aNi Us mapping 01110 GERAM, 'CIM at Work' Conference, Kaatsheuvel, The Netherlaoclol. Doumeingts, G., Valle.pir, B., Zaneuin, M. and Cben, D. (1992). GRM Illt~g,ated M~thodoloBJ : A. ""'thodology 10' Msignlng CIM syst.ms, version 1.0, GRAI/LAP, University Bordeaux 1. DoumeinglS, G., R~gnier, P. and Feoie. P. (1993).
FLEXQUAR Consortium Esp,it project III CIME Fl~"quar project : GRA! I nt~grat~d
6408
M~thodo/ogy·, Deliverable MI2T11. Doumeingts, G. (1994). M.tltode GRM: m.tMM M
conceptioll MS syS/l"",s ~II prodMCtique, Th~ d'~tal en AutomaIiquc, Universire de BmIeaux I. Rni~, P. (1994). GRMCO : Method. d'alUJlys~ ., M cOllc~ptioll du syst~m" d'apioitatioll, ~ de Doclmlt, Auwmatique, Universire Bonleaux I. Impacs, (1992). IlItegr/ll~d MalWj'actJVing Planning aNi COlltrol SySl~m, Milestone 6 Deliverables, Esprit Project 2338.
Le Moigne, J.L. (1984). La thiorie du system~ glnirGJ", Th/orie de la ",odllisatioll, Paris, Presses Universitafta de France. J. (1979). Approclt~s
M~lese,
systlmiqu~s des o,g
652
Pw!, L. (1977). ApprocM mitllodow8ique de motUlhIuiolt ell ...., de la IfItIItri.re GUUt« ,. ID protIMcdoll. ~
de I'AFCBT, VerIIiIlea, Faoce. Edilions Hommes el Techniques, pp. 466 - 478. Pun. L. (1984). Systillles ilUhutrids d'iltuIli8e"ct ardfk:ielk. outiIa de productique. EditcSlS. Paris. Roboam, M. (1988). MotUlu,. rIP~na et iNisrQlio" des mitltodu d'fINlIyu pow ID CO/ICepdo/t des SJaUwtu de productioll. Tb~se de Doctoral, Automatique.
Univcnit61lo1doMx L. Roboam, M. et al. (1981). Ope,. CAM System I Esprit project 418, Pllblic domain rqJOII. Vallespir. B. (1987). Exploitatiolt des .,,,emu de prodlu:tioll dUcrets-coll/i,.., : cOlllributiolt .t UIte mlthode de cOltceptiolt, Tb~se de doctoral, AulOOlalique. Univn1l! de Bordeaux I .• WiIliams, T J . (1992). TIle PurM ElIlerpriu Reference Arcllilectun,lnatnuneat Society of America, Resean:b Triangle Part, North Carolina, USA. WiIliams, T J. (1994). COlltributiolLS of tM I'urdue EllUrprise Reference ArclUtectwe alld Melllodology (PERIl) to tM devcwpmelll of a 8eneral Elllerprise Reference ArclUleclllre d: MethotJq/oBY (GERAM),
ICARCV'94. Sinppcrc. Williams, T J. &. HOIIi U (1995). A specificatiolt alld _lIlofr~lIISlorGERAM,~n·159.
Purdue Labonlory for Applied Industrial Conttol. version 1.0. Williams. P. et al., (1994). ArclUtectwes for illle8ralilt8 llUWl/acturi1t8 actillities aM eIIlerprises. Cornputets In Industry, vo!. 24, n"2-3.
653