A business repository enrichment process: A case study for manufacturing execution systems

(1)

HAL Id: hal-01533642

https://hal.archives-ouvertes.fr/hal-01533642

Submitted on 6 Jun 2017

HAL is a multi-disciplinary open access archive for the deposit and dissemination of sci- entific research documents, whether they are pub- lished or not. The documents may come from teaching and research institutions in France or abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est destinée au dépôt et à la diffusion de documents scientifiques de niveau recherche, publiés ou non, émanant des établissements d’enseignement et de recherche français ou étrangers, des laboratoires publics ou privés.

A business repository enrichment process: A case study for manufacturing execution systems

Ikbal Arab-Mansour, Pierre-Alain Millet, Valérie Botta-Genoulaz

To cite this version:

Ikbal Arab-Mansour, Pierre-Alain Millet, Valérie Botta-Genoulaz. A business repository enrichment

process: A case study for manufacturing execution systems. Computers in Industry, Elsevier, 2017,

89, pp.13 - 22. �10.1016/j.compind.2017.03.006�. �hal-01533642�

(2)

A business repository enrichment process: A case study for manufacturing execution systems

Ikbal Arab-Mansour, Pierre-Alain Millet*, Valérie Botta-Genoulaz

UnivLyon,INSA-Lyon,DISPEA4570,F-69621,Villeurbanne,France

ARTICLE INFO Articlehistory:

Received5October2016

Receivedinrevisedform7February2017 Accepted30March2017

Availableonlinexxx Keywords:

Manufacturingexecutionsystem(MES) Informationsystemalignment ISO/IEC62264

SCORmodel Interoperability Integration Heterogeneity

ABSTRACT

Akeycharacteristicofthesoftwareapplicationssupportingmanufacturingbusinessprocessesistheir heterogeneity.Thisisduenotonlytodifferencesintheirdevelopmentanddeployment,butalsotothe varietyofprocessesandactorsincomplexorganizations.Heterogeneityatthesemanticlevelisoneofthe major problemsinany processof interoperabilityand/orintegration.Thereis thereforeaneedfor developing new approaches and methods to ensure interoperability between different software solutions.InthecontextofacasestudywithaconsortiumofMES(ManufacturingExecutionSystems) publishers,weproposeasemanticalignmentprocessofrepositoriesusedintheconstructionofaMES solutioncalled“MESOnDemand”,usingmultipleapplicationsanddrivenbybusinessprocesses.Through thestudyofsemanticheterogeneities,weuseanenrichment-basedalignmentforbusinessrepositories appliedtoISO/IEC62264.Finally,weevaluatethecontributionofthisapproachtoenterprisematurityin theapplicationofstandards andreference models,usingNascio’sEnterpriseArchitectureMaturity Model.Thisproposal,whichisusefulforpractitionersandexperts,isacontributiontoacademicstudyon semanticalignmentformasterinteroperability.

1.Introduction

Marketrequirementsareconstantlyevolvingandaredemand- ingthatcompaniesseeknewsolutionsand techniquestobuild informationsystems(IS)inlinewiththeirneedsandstrategies.

EnterpriseISareoftencomposedofseveralbusinessapplications suchasEnterpriseResourcePlanning(ERP)systems,Manufactur- ingExecutionSystems(MES),andProductLifecycleManagement (PLM)systems.Aresultofthistrendistheneedforcompanies’

“integration”.Integrationcanbeseenasa processconsistingin highlevelsofinteractionbetweenpeople,machinesandapplica- tions, which enhances the synergy within a company [1].

Regarding IS, integration essentially involves the application’s components.Insomecases,components arelooselycoupledto preserve their heterogeneity and autonomy. This is called the

“interoperability”of applications,which is onewaytofacilitate integration.

Manyprojectsandstudieshavebeencarriedouttodevelopand design newapproaches and corporate architecturesin orderto ensure the interoperability of systems at the technical and semantic levels. The service-oriented approach provide sound answerstotheproblemsoftechnicalinteroperability.Indeed,itis essentialthatbusinessapplicationsremaincapableofconsistent interpretation of the data exchanged and the functions used.

Semantic conflicts may arise during the exchange between applications. Theseconflictsmay involvenot only data[2] but also the business logic of applications [3]. It becomes vital to identifyresultingconflictsandproceedtotheirresolutionassoon aspossible,preferablyintheearlystagesofanintegrationproject.

The resolution of these semantic heterogeneities requires mechanisms to determine the correspondence areas and/or semanticconﬂictbetweenapplications.

Current research on semantic alignment uses mainly ontologies.Mostoftheseongoingdevelopmentinitiativesafﬁrm thatautomationofalignmentisdifﬁculttoachieve,andhighlight theroleofusersinthevalidationofsemanticalignment[4].The semantic alignment betweenapplicationstherefore remains an unresolved issue in the context of integration and systems interoperability, which requires an alignment process.

Consequently,theroleofexpertsinintegrationprojectsiscritical.

As business applications are based on either proprietary or

* Correspondingauthor.

E-mailaddresses:[email protected](I.Arab-Mansour),

[email protected](P.-A.Millet),[email protected] (V.Botta-Genoulaz).

ComputersinIndustry89(2017)13–22

ContentslistsavailableatScienceDirect

Computers in Industry

j o u r n a lh o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / c o m p i n d

(3)

standardbusinessrepositories,awaytotacklethisresearchissueis toprovidesemanticalignmentsolutionsforbusinessrepositories.

In this paperwe present a case study where a newalignment methodology enriches a given business repository by other businessrepositories.Itprovidesexpertswithanassistedprocess of unidirectional semantic alignment of business repositories, using semantic check rules to help them solve any semantic heterogeneityproblemsthatmayexistbetweenseveralbusiness sources.

Thiscasestudyistheresultofanindustrialinitiativeconducted by a consortium of manufacturing software publishers. They launcheda“MESondemand”platform,usingservicesfromvarious packages,includingManufacturingExecutionSystems(MES)and SupplyChainExecution(SCE).Toallowsuchan“on-demand”MES tomanageheterogeneitybetweenthetrades,softwarepackages, and business reference models involved in the project, we formalize the knowledge of business processes and business objects that will be supported by applications and their interactions.Ofcourse,thisknowledgeisimplementedindifferent waysandoftenpartiallyineachsoftwarepackage.

The capacity tobuild an “ondemand” multi-publisher MES solutiondepends onthecapacitytoalignbusinessrepositories, companies’ needs, and functional layer components, which containsheterogeneousknowledgeforthefollowingreasons:

a Several software packages are more often involved and are heterogeneous;

b Industrialcompanies’requirementsareexpressedintheirown businesslanguageandhavetobemappedontotherepository language;

cFinally,thecapacitytolinkastandardbusinessrepositoryanda standardwebserviceofapplicationsisaconditionformastering theimplementationoftheindustrialcompany’sbusinessmodel inanorchestrationofexistingwebservices.

More often,aligningtwo structures(models,standards,etc.) meansthatforeachentity(e.g.conceptsandrelationships)inthe ﬁrststructure,wetrytoﬁndacorrespondingentitythathasthe samemeaninginthesecondstructure[5].However,inaproject- orientedapproach,theaimistobuildacommonrepositoryforall consortiummembers,whichfocusesonacertainscopeandaimof theproject.Incontrasttotheapproachesseekingforacomplete andgenericalignment,weproposetomastertheheterogeneity derived from business reference models, which are partially applied. We therefore propose to seek neither to modify the structureofoneoftherepositoriesnortomergethem.Wecreatea newversionVi+1ofa givenbusinessrepositoryViconsideredas ReferenceRepositorybyaddingsomeelementsorsomesemantic relationshipsfromasecondbusinessrepository“A”,dependingon thelevelofgranularityorconsistency.

We applied this methodology in our case study with MES businessrepositoryIEC/ISO62264,whichhasbeenenrichedbythe Supply Chain Operation Reference (SCOR) model. We then implementedittheARISSOAArchitectplatform.

Theremainderofthispaperisorganizedasfollows.Section2 introduces the literature review in terms of techniques and approachesrelatedtoenterprisemodelling,businessrepositories and systems’ interoperability. The alignment methodology, the enrichment process, and its implementation are presented in Section3.In Section4, weevaluate themethodology usingan EnterpriseArchitectureMaturityModelbasedonasetofpractices.

Finally,we draw some conclusions for ﬁrms and outlinesome directionsforfutureresearch,inSection5.

2.Literaturereview

Mostinformationsystems(IS)containseveralapplicationsto meetseveral businessneeds,According toRefs.[3]and[6],the main characteristicsofa company's applicationsareautonomy, distribution and heterogeneity. There are various levels of heterogeneity,includingtechnical,syntacticandsemantic.Inthis paper,we focusontheproblemsofheterogeneityofenterprise applications,particularlyatthesemanticlevel,whichisoneofthe majorproblemsinanyinteroperabilityand/orintegrationprocess.

AccordingtotheEuropeanInteroperabilityFramework[7],there are three aspects of interoperability: (i) organizational, (ii) technical, and (iii) semantic. The different aspects of interoperabilityemphasize thefactthat theheterogeneityofIS indata,applicationsorprocesslevelsmakestheimplementationof interoperabilitybetweenISmoredifﬁcult.Thiscomplexitycreates newstrategies,methodsandarchitecturesforthedesignofIStobe qualiﬁedasinteroperable.

Afteranintroductiononthealignmentissueandabriefsurvey onenterprisemodellingandmodellingframeworks,theremainder ofthissectionpresentsthetwomainrepositoriesthatformthe basisofthisresearchwork,theIEC/ISO62264repositoryandthe SCORmodel.

2.1.Interoperabilityvsalignment

Thisissueofinteroperabilityand heterogeneityiscrucialfor MES when manufacturing resources, processes, and control methods areheterogeneous. The ManufacturingEnterprise Sol- utionsAssociationproposesaformaldeﬁnition[8].MESisoneof thesoftware solutions used tobridgeproduction planningand equipment control systems. To meet the needs of a varietyof manufacturingenvironments,they identiﬁed11mainfunctions, somedirectlylinkedtotheprocess,andothersdescribedascross functions.

Theacademicresearchonintegrationproblemspointsoutthe importanceof the integrationwith otherinformation systems.

Ref. [9] propose a service-oriented framework to address reconﬁgurability required by changing business environments.

Ref. [10] present a literature review of MES systems, showing severalconsiderationsoncommercialsolutions,andmajorissues relatedtotheiruseandimplementation.Theseauthorsreporton in-depthdiscussionsontheresearchareasthatwouldhavetobe explored in order toresolve the increased complexity of MES systems.

Awell-knownbusinessstandardforMESistheISA-95standard publishedby theInstrumentation, Systems&Automation (ISA) Committee.AnacademicversionhasbeenpublishedastheIEC/ISO 62264. Thisstandard dealswiththeformalization ofexchange, around the manufacturing system, with other areas of the company. Its function model puts the manufacturing control functioninacentralposition,intermediarybetweenthevarious departmentsofthecompany.ISA-95iswidelyadoptedbymarket players forthedesign of information ﬂows betweenshopﬂoor applicationsandthoseatahigherlevel.Italsoallowstheindustry tohaveaconsistentterminology.MESareused,interalia,bythe aerospace, automotive, semiconductor, optoelectronic, pharma- ceuticalandpetrochemicalindustries[11–13,41].

Both MESA and ISAorganizations highlight the diversity of softwarecapabilities required,withsoftwarepackages covering different manufacturing domains that often overlap. Semantic alignment is then a major challenge for MES projects for companies.

(4)

2.2.Enterprisemodellingandmodellingframeworks

Enterprise Modelling (EM) can be deﬁned as a means for capturingandcapitalizingontheknowledgeofacompanyandits differentaspects:functional,informational,structural,behaviou- ral,organizationalorother[1].Ref.[14]proposesaclassiﬁcationof thesemodellingmethods:operationalmethodssuchasGIM[15]

or IS and process-oriented methods such as Olympios [16];

referencearchitecture-basedmethodssuchasCIM-OSA(Comput- erIntegratedManufacturing,OpenSystemArchitecture)[17,18]or ARIS (ARchitecturefor integrated InformationSystems)[19],or S-BEAMthatsupportsdecisionsatbothstrategicandoperational levels,usingtheSCORmodel[20].

Anenterprise-modellingframeworkidentifiestheinformation typesandtheirrelationshipstomodelacompanybyorganizing them intological structures. TheENV 12204standard [21] has proposedan initial specification of the main elements for the modelling of “constructs”, and the ISO 15704 [22] defines the requirements for buildingan enterprise reference architecture.

These requirements were included in the CEN/ISO 19439 (Enterprise Integration Framework for enterprise modelling) standard [23] thatproposesa modelling frameworkwiththree dimensions:lifecycleofmodels,viewofmodels,andgenericity.

AnotherstandardistheCEN/ISO19440[24]whichhasadoptedthe modelling constructs deﬁned in the ENV 12204 standard, and which complies with the three dimensions deﬁned in the ISO 19439standard.

2.3.Standardsandbusinessrepositoriesfortheenterprise

Repositoriesexist inmedical emergencies,entrepreneurship, professional skills, and so on. There are many standards and businessrepositoriesinmanufacturing,particularlythoseusedby ManufacturingExecutionSystems.

Thesubcommittee65AestablishedtheIEC/ISO62264interna- tional standard [25–27], based on the American standard ANSI/ISA-95.Itdeﬁnesfourdomains:production,quality,maintenance and inventory. The fundamental concept of the IEC/ISO 62264standardisbasedonthebreakdownofbusinessandcontrol functions into sub-processes and activities. The activities use inputs and produce information objects. Information ﬂows between functions are represented by the content of their interfaces.Table1showstheattributesoftheEquipmentProperty class.

TheactivitiesdeﬁnedintheIEC/ISO62264standardareusedto provide activity models that can be adapted and customized accordingtobusinessneeds.

TheSupplyChainOperationReference(SCOR)modelprovided bytheSupplyChainCouncil[28]offersavisionandanapproach that focuses on business processes withinthe supply chain.In Version 9.0 [28], the SCOR model contains 27 key processes detailedin171processelements,274piecesofinformationdefined asinputs/outputsforprocesselements,489bestpractices,and498 metrics. The SCOR model is based on four levels: 1-strategic, 2-tactical,3-operational,andlevel4,whichisnotformalizedinthe model.Ref.[29]hasproposedaclassificationoftheinformation exchanged in theSCORmodel. Hehas identifiedten classes of information (actor, rules, plan, object, resource, condition, decision, flow, journal, and metric), as well as several studies basedontheSCORmodel,usedtodevelopmethods,knowledge bases[30]andtoolsfortheintegration[31],evaluation[32] and alignment [33] of supply chains. MES arepresent in theSCOR model through various features: production, maintenance and inventory.

ThereareotherstandardsrelatedtoMESapplications,suchas:

STEP(StandardsfortheExchangeofProductdata),orISO10303 [34], which is a product-oriented standard for product data exchangebetweenapplications;

ISO9001,whichisaquality-orientedstandard.Theversion2000 promotestheadoptionofaprocessapproachwhiledeveloping, implementing and improving the effectiveness of a quality management system, to enhance customers’ satisfaction by complyingwiththeirrequirements[35];

AFNOR, which deﬁned a maintenance-oriented standard (13306X60-319)in2001.Maintenancemanagementrepresents oneofthefunctionaldomainsoftheIEC/ISO62264standard.

This diversityand heterogeneityof business repositories for MES highlight theneed for a newalignment approach to help practitionerstousesomeoftheserepositorieswhenassembling heterogeneouscomponentstobuildtheirIS.

3.TheMESbusinessrepositorycasestudy

Inthefollowing,wepresentthemethodologyusedtoenrich theIEC/ISO62264MESrepository(Vi)bytheSCORmodel(A).This enrichmentiscomposedofthefollowingsteps:

(i)thecomparisonbetweenthetwobusinessrepositoriesViand A,

(ii)theextractionandthemodellingofthecontentstobealigned, (iii) thesemanticalignmentprocess,

(iv) andthevalidationcycle.

Table1

TheAttributesoftheEquipmentPropertyClass(IEC62264-2).

AttributeName Description Examples

ID Anidentiﬁcationofthespeciﬁcproperty Equipmentname

Runrate Templatesize

Description Additionalinformationabouttheequipmentproperty “Localnameforthewidgetmachine”

“Widgetmakingaveragerunrate”

“Widgetjigtemplatesize”

Value Thevalue,setofvalues,orrangeoftheproperty.

Thevalue(s)isassumedtobewithintherangeofsetofdeﬁnedvaluesfortherelatedequipmentproperty.

BigBertha 59 300

Valueunitofmeasure Theunitofmeasureoftheassociatedpropertyvalue,ifapplicable. [notapplicable]

Widgets/h cm

I.Arab-Mansouretal./ComputersinIndustry89(2017)13–22 15

(5)

WedetaileachofthesestepsthroughillustrationswithARIS models. The ARIS platform [19], which is independent of the consortiummembers,allowsthemodellingofbusinessprocesses andobjectsandtheuseofsemanticrulesandscriptstoanalyse thesemodels.

3.1.Comparisonbetweenrepositories

This step helps to assess the degree of similarity and heterogeneitybetweenthedifferentbusinessmodelsofthetwo repositoriesViandA.Therearefourheterogeneitylevels[36]:(i) syntactic;(ii)terminological;(iii)conceptual;(iv)pragmatic.

Thesyntacticlevelrelatestoknowledgedescriptionlanguages usedtorepresentthedifferentmodelsofthetworepositories.In ourcase,theISO/IEC62264standardexistsinatextualformthat describesthevariousconcepts,objectmodels,etc.,andadigital formrepresentedbyXMLﬁlesB2MML“BusinesstoManufacturing MarkupLanguage”[37]todescribetheinteractionbetweenMES andERPsystems.TheSCORmodelexistsinatextualformproposed bytheSupply ChainCouncilthat describesthecontents ofthe variousprocesses,andadigitalformrepresentedbymodelsusing theARISlanguage. Comparingthe twoversionsof bothtextual repositories,weseea cleardifference inthe lexiconused.This differenceisduetothediversityofthecommitteesthatcreateboth standards and their differing scope. When comparing the two digitalformats,itisclearthattheconceptsarequitedifferent,since the ISO/IEC 62264 describes the exchanged information ﬂows whiletheSCORmodeldescribestheprocessesofthesupplychain.

The terminological level concerns theidentiﬁcation of functions,businessobjectsandprocesses.Itispossibletoidentifythe followingtypicalconﬂictslikesynonymy,polysemyandsyntactic variationsofthesameword(abbreviations,etc.).

Theconceptualheterogeneitiesinvolvedifferencesin:

Coverage: theISO/IEC62264 coversproduction,maintenance, quality and inventory domains, while the SCOR model is designedandmaintainedtosupportsupplychainswithvarious complexitiesandacrossmultipleindustries;

Granularity: the entities of the two business repositories describerealobjectsinvariousdegreesofdetail;

Perspective:thetwobusinessrepositoriescoverdifferentpoints ofview.

Thepragmaticlevelconcernstheheterogeneousinterpretation of therepositories. Theseproblems canoccur whenexperts or differentcommunitiesinterpretdifferentrepositoriesindifferent contexts.Indeed,thislevelisquitecomplex,sincetheresultofthe alignmentcanchange,dependingontheinterpretationsofthetwo repositories.Table 2 shows a comparative studyof the ISO/IEC 62264 and the SCOR model, applied to the MES domain and classiﬁedaccordingtofourlevelsofheterogeneity.

3.2.Extractionandmodellingofcontenttobealigned

Thefirststepistochoosethebusinesscontentfromrepository AtoalignitwiththeMESrepositoryVi,dependingoftheobjectives ofpartnersinthecontextofourcasestudy,whichinfluencesthe scopeofexpectedalignment.Themainpurposeistoreducethe coverageissuesidentifiedinthecomparisonphase,andtochoose onlythecontentclosetotheMESdomainfromabusinesspointof view.PartsofthecontentoftherepositoryAcanberemovedfrom thealignment process.Thischoiceis usuallymadeby business expertsinordertomeetspecificbusinessneeds.Inourcasestudy, werestrictthealignmenttotheMESdomainasdefinedbyISO/IEC 62264,excludingthepartsofthesecondreferencethatisoutofthe domain.Moreprecisely,wekeep onlytheMAKEprocess ofthe SCORmodel,whichincludestheMESfunctionsandobjects.We haveidentified28processelementsand107piecesofinformation thatcanbeused.ThechoiceofMAKEprocess’contentisbasedon theresultsofRef.[29].

Themodellingoftheextractedcontentensuresasinglereading ofthemodelsofthetwobusinessrepositories.Thisisdoneusinga simpliﬁedmodellingframeworkbasedontheISO19439standard [23]focusedonmodellingconstructsrequiredbytheISO62264 repository.Inthissimpliﬁedframework,therepositorylevelisan aggregationofthe‘generic’and“partial”levelsoftheISO19439

Table2

ResultsofcomparisonstepbetweenISO/IEC62264andtheSCORmodel.

ISO/IEC62264 SCORmodel

Syntactic level

Description language

Textual Digital(B2MML)

Textual

Digital(ARISmodels)

Concepts Information,Activities Processesandprocesselements,Inputs/Outputs,

Metrics,Bestpractices Terminological

level

Presenceof Synonymy

Yes Yes

Presenceof Polysemy

Yes Yes

Presenceof Syntactic variations

Yes Yes

Conceptual level

Coverage MES

Production Maintenance Quality Inventory

Supplychainoperations Severalapplications,ofwhichMES

Granularity Weak Variesaccordingtothelevels

Perspective Informational,Functional Processes,performancemanagementofthesupply

chain Pragmatic

level

Variesaccordingtotheinterpretationofthesemanticcontentofeachrepository

(6)

standard,whiletheprojectlevelusedtomodelclientrequirements andsystemsistheimplementationofthe“particular”levelofthe standard.Likewise,weuseonlytheinformationalandfunctional viewsofthestandard,excludingtheorganizationalandresource views, because the ISO 62264 standard deﬁnes only data and functionmodels.Regardingthelifecyclephase dimensionofthe ISO 19439 standard, we focus onthe “concept deﬁnition” and

“requirement”phases.Thebusinessandfunctionalmodelsinthe genericitydimensionattherepositorylevelarethenconsideredin the“conceptdeﬁnition”phase,whereasattheprojectlevel,they are considered in the “requirement” phase. This ensures the conformityofmodelscreatedinthisframeworkwiththegeneric level, and then for the industrial companies, with the MES repository.

Thepurposeofthisframeworkistomodeltheseheterogeneous repositories using the same constructs and a certain syntactic interoperabilitybetweenthem.Wethusproposesomesolutions forproblemsofheterogeneity,regardingsyntax.

GiventhelargenumberofSCORelementstobemodelled,we usethe“matrix models” of the ARISplatform torepresent the relationshipsbetweenthefunctions (28process elements) and theirinputs/outputs(107inputs/outputs).Fig.1showsanexcerpt fromanoverviewof thematrix, whichis representativeof the MAKE process, with the input/output (considered as business objectsmodelledusingtheARISobjecttype“clusters”)inlinesand thefunctionsincolumns.

3.3.Semanticalignmentprocess

Thisprocessconcernsthedefinitionofsemanticrelationships betweenbusinessrepositories.Thecontentextractionstepreduces coverage problems. However, it is difficult to find the right granularitytoalignthetworepositories.Ontheonehand,a“high- level”alignment(top-down)maynotprovideenoughinformation onthedifferencesandsimilarities;ontheotherhand,a“verylow” (bottom-up)alignmentisnotaneasytask,giventhelargenumber ofpotentialconnectionsbetweenthetworepositories.Wepropose to analyse the level of granularity in both repositories, by

examining separatelytheelementsofeach viewdeﬁned inthe modelling framework. This implies the creation of semantic relationships between the elements of each view. Reasoning independently in each view is a simpliﬁcation factor, which reduces the number of elements and therefore thenumber of connections. However, this independence can lead tosemantic inconsistenciesintherelationsbetweentheviews.Wetacklethis point in Section 3.4, in which we propose solutions based on semanticcontrolrules.

InFig.2,weshowanexampleofarelationship(specialization relation) between the two constructs “Repository Business Activity” and “Repository BusinessOperation”.Thisrelationship is reﬂected in the alignment step by one or more semantic relationships between entities in the two repositories. The example therefore contains two types of semantic relation:

“Equivalentto”(corresponds-to)and“Belongingto”(asub-state) betweenafunctionofISO/IEC62264“TrackingOperations” and three functionsofthe SCORmodel. Regardingthedifference in perspective(functionalvsprocesses),itcontributestoenriching theMESrepositorybycreatingnewsemanticrelationships. We have added a new relationship (Is_predecessor) between the functions of the SCOR model belonging to the same process element. In the following, we propose a “hybrid” semantic alignmentapproach,whichisbothtop-downandbottom-up.

3.3.1.Informationalview

TheISO/IEC62264standard deﬁnesonlynineobjectmodels.

We can classify these object models in two main categories:

commonobject models(“Common Object Models”), and object models for the management of manufacturing operations (“Operations Management Information”). The purpose of this classificationis,firstly,toreducetherisk,thecostandtheerrors associated with the implementation of these interfaces, and secondly,toprovideasemanticlevelallowingotherstandardsor repositories to adapt. Regarding the informational view of the SCOR model, we identify in the “MAKE” process 140 objects defined as inputs/outputs of process elements. We propose a classificationoftheseelementsintofiveclasses(Resource,Object,

Fig.1.ExampleofthealignmentmodelledbymeansofanARISmatrix.

(7)

Plan,Flow, and Journal) based onthe resultsof Ref. [29].This classiﬁcationenablesaﬁrstleveloftop-downalignment.

Wethereforeseektoprovideanumberofclassesthatisclose enoughandreasonableincomparisontotheamountofinforma- tioninISO/IEC62264.Table3showsthecorrespondencebetween the ﬁve information classes of the SCOR model and the nine informationclassesfromtheISO/IEC62264repository.

ToimplementthealignmentprocessontheARISplatform,we initiallycreate ﬁve IEdata models includingSCOR information (top-downapproach).ThisalignmentisreﬂectedonARISusingan associationrelationshipbetweenacluster–representingoneof theninemodelsofobjects–andone(ormore)model(s)oftheIE data– representingoneor moreSCORinformation classes. For example,the“Plan”and“Flow”classesaremodelledusingtwo“IE data models” having an association relationship with the

“OperationsSchedule”objectmodelrepresentedbyacluster.

We complete the semantic alignment with a bottom-up approach,startingfromthelowestlevel,i.e.directlyfromI/Oof the SCOR model “MAKE” process. For example, we align the elements of the “Resources” class with “Common Object” and

“OperationCapability”information (Table 3).Theresultsof this unidirectionalalignmentare:

Someinformationremainsoutoftheareadespitetheextraction step (i.e.information onsupply chainperformance, historical payroll,etc.).

Weidentifytwotypesofrelationshipbetweentheelementsof ISO/IEC62264andthoseoftheSCORmodel:

A link of the 1.1 type: information from the SCOR model corresponds to a single object from the ISO/IEC 62264

repository.Inthiscase,thetwoobjectspotentiallyhavethe samelevelofgranularityandare“equivalent”fromasemantic pointofview.

A link of the 1.N type: information from the SCOR model correspondstooneormoreobjectsfromtheISO/IEC62264re- pository.Wenolongerhavethesamelevelofgranularity.Inthis case,theinformation fromtheSCORmodelis composite.For example, the “Production Orders Planned & Actual Reports” information is composed of (Is_Composed_Of relathionship)

“OperationsCapability” and“OperationsPerformance” objects Functionalview.

TheISO/IEC62264standarddeﬁnesmanagement modelsfor manufacturingoperationsthatallowtheintegrationofenterprise controlsystems. Eachmodelis a setofeight activitiesthat act betweentheplanningfunctionsandtheprocesscontrolfunctions.

ConcerningthefunctionsdefinedintheMAKEprocessoftheSCOR model, we identify28 functions.We keep thesame alignment processastheonefollowedintheinformationalview,i.e.westart withahighlevelofabstractiontofacilitatethealignment.Wethus usetheclassificationfunctionsofferedbytheSCORmodel,which consistsin defining fourvariants for theMAKE process, corre- spondingtodifferentbusinesstypologies:Make-to-Stock(MTS), Make-to-Order(MTO)Engineer-to-Order(ETO)andEnable-Make (EM).Table4 showsafirstleveloftheunidirectionalalignment betweenthefunctionsoftheSCORmodel(lines)andtheactivities oftherepositoryVi(columns).

Inthisalignmentstep,whichaddressesthedimensionofthe granularity,wehighlightthestrongappearanceoftheperspective dimensionbetweenthetworepositories.Infact,wetrytoalign Fig.2.AlignmentatFunctionalLevelUsingAlignmentRelationship.

Table3

High-levelinformationalalignmentbetweenSCORandIOS62264.

SCORInformationClasses ISO/IEC62264InformationClasses

Resources CommonObject,OperationCapability

Object OperationDeﬁnition

Plan OperationSchedule,OperationCapability

Flow OperationSchedule,OperationCapability,OperationPerformance

Journal OperationSchedule,OperationCapability,OperationPerformance

(8)

four processes with eight activities. In Table 4 the boxes are completedbasedonthetextualdescriptionofeachsub-process.

This step requires the expert intervention to interpret the semantics and imagine the potential activities of the standard thatmayhavearelationshipwitheachprocess.Herewestartfrom thelowest level of theMAKE process, i.e. thefunctions of the variantprocesses,andthusalignthe28functionsofSCORonthe eightoftheISO/IEC62264activities.Wehavethusidentiﬁedtwo functionsoutsidethearea:riskmanagement,andmanagementof theregulatoryenvironment.Wealsoidentiﬁedtwokindsoflinks:

a1.1typelink“isequivalentto”whichcorrespondstoelements withsamelevelofgranularityandperspective,anda1.Ntypelink

“isthemasterof”.Inthelattercase,thefunctionsofSCORmodel haveahighergranularitylevelthantheactivitiesoftherepository Vi,andsecondly,theyallpresentprocesselements.

Table5showsthedifferentpossiblealignmentsituationsthat expertsmayencounterinthealignmentprocess,andthenamesof correspondingrelationships.

3.4.Validationcyclebybusinessexperts

Humanexpertiseisrequiredinouralignmentprocess.Thegoal istoalignthetworepositoriesatalevelofgranularityclosetoeach other.The top-downapproachof informationalsemanticalign- ment signiﬁcantly reduces the number of potential mapping operations between elements. However, in the bottom-up alignment, the expert hasto solve semantic interpretations of differencesbetweenrepositories.Forexample,he/shemusttake intoaccounttheprocessapproachof theSCORmodel,whichis absentintheISO/IEC62264,ortoexcludetheriskmanagement usedintheSCORmodel,whichisnotrequiredintherepository Vi+1.

So,therepositoryVi+1issubmittedforvalidationtomultiple businessexperts,includingexpertsonISO/IEC62264andonthe SCORmodel.Wehaveproposedareﬁnementcycleattendedby experts,toresolve semanticconﬂicts duetotheinterpretation ofthesemanticmodels, objectsand businessprocesses,which varyfrom one expert toanother.As mentionedin Section 3.1, we faced different situations at the pragmatic heterogeneity levelwherevariousexperts interpretthesemanticsof thetwo repositories.The choice to perform a unidirectionalalignment (to theISO/IEC62264) guarantees theconvergenceof experts’ alignments.

Tofacilitatetheexperts’decision-making,wehavedeveloped some semantic control rules to check and detect semantic inconsistenciesofnewadditionalelements.InFig.3,weassume

thatwehaveoneelementofinformationXandtwoactivitiesBand CfromtheISO/IEC62264,suchthat“XistheinputofB,notofC”. SupposethattheinformationX1andfunctionC1aretwovariants oftheSCORmodel,respectivelyofinformationXandactivityC, suchthat“X1istheentryofC1”.Thesemanticcheckdetectsan inconsistency,eveniftherelationshipbetweenX1andC1iswell deﬁned in SCOR. The role of the business expertis tomake a decisionaboutthisrelationship:

ThealignmentofthetwoentitiesX1andC1oftheSCORmodel on,respectively,XandCisfalse.Inthiscase,itisnecessaryto correctortodeleteoneorbothofthesevariantrelationships.

Thealignmentiscorrect.TheexpertvalidatestherepositoryVi+1

despitethisapparentinconsistency.

Thedeﬁnedsemanticcontrolruledoesnotapplyinsomecases.

Table 6 shows an excerptof a report presenting results for semanticcontrolappliedtoabusinessscenario:“Schedulingand launchingofaProductionOrderinjobshop”.Thesamesemantic controlrulemaybeapplieddirectlytothemodels.

4.Evaluationandkeylessonslearned 4.1.Evaluation

The beneﬁt assessment of this work is evaluated using the EnterpriseArchitectureMaturityModelofRef.[38],basedonaset ofbestpractices.ItcontainsMaturityLevelsrangingfrom1to5, depending on the best practices integration levels. We have consideredNascio’sbestpractices,which areclearly inrelation withourproposal,excludingissueslikeadministration,planning or communication. Wefocused on thefollowing best practices regarding standards, processes and models to evaluate the contributionofourmethodology:

Framework to organize processes and templates used for EnterpriseArchitecture(EA)

Blueprintascollectionoftheactualstandardsandspeciﬁcations Compliance,toevaluatetheadherencetopublishedstandards, processesandotherEAelements,andtheprocessestodocument andtrackvariancesfromthosestandards

Integrationtoevaluate consistencyof management processes withEA

Weevaluatedretrospectivelytherequiredlevelofmaturityand thepotentialcontributionofouralignmentmethodology,toreach Table5

ImplementationinARISmodellinglanguageofalignmentrelationships.

Views Situation Semanticrelationcreatedtomodelthealignment

Granularity InformationalView ISO/IEC62264=SCOR Variant

ISO/IEC62264<SCOR “iscomposedof”

FunctionalView ISO/IEC62264=SCOR “isequivalentto”

ISO/IEC62264>SCOR “belongsto”

ISO/IEC62264<SCOR “ismasterof”

Perspective FunctionalView “ispredecessorof”

Table4

High-levelfunctionalalignmentbetweentheMAKEprocessandISO/IEC62264activities.

ResourceManagement DeﬁnitionManagement Dispatching Tracking DataCollection Analysis DetailScheduling ExecutionManagement

MTS

MTO

ETO

EM

(9)

new levels of maturity regarding each of these best practices (Table7).

Obviously, the objective ofmastering heterogeneitythrough alignmentofdifferentbusinessmodelsrequirestheprocessestobe plannedand tracked, and the organization tobe able toreuse methodsforcapturingcriticalEinformation(Level2/Framework).

Our methodology reaches three of the four characteristics expected for Level 3. The architecture based on the resulting repository Vi+1 helps the industrial company to customize the processesinaspeciﬁc/local context,toprepareprocessmodels, andtousetemplates.Thelastcharacteristicregardingprocesses’ lifecyclesisnotconsidered.

Regarding the blueprint, a prerequisite is that technology standardsaredocumented(Level1).Usingourproposaltobuilda MES repositoryable to integrateother heterogeneousbusiness

models is an opportunity to build an enterprise architecture repositorytoensureconsistencyofdocumentation(Level3).

Compliance ofprocessesand procedureswithstandardsisa corecharacteristicofourmethodology.Aclearrequirementisthe needforcompliancewithstandardstobeidentiﬁed,evenifthe complianceisinformalandunstructured,andcannotbemeasured effectively,becauseprocessesandproceduresarenotconsistent acrossareasand/orprojects(Level1).TheresultofaMESproject usingourproposalisthat thecompanyhasbeguntodevelopa compliance process to ensure consistency with Enterprise Architecture standards (Level 2).Furthermore, it is a driver to organizea formalcomplianceprocessasanintegralpartofthe enterprisearchitecturelifecycleprocesses,consistentlythrough- outtheenterprise(Level3).ItcanhelptoensurethataBusiness Caseisrequiredforvariancefromthestandards.

Finally,regarding integration,theuseofa repositoryimplies thattheneedtodocumentcommonfunctionshasbeenidentified, even if MES projects are carried out in isolation (Level 1). In addition,itimpliesthattheneedforintegrationofprocessesinthe EAFrameworkhasbeenidentified,evenifthevarioustouch-points between the management processes and the EA Program Framework have been mapped, without details on how the integrationworks.Thisisanintermediatesituationbetweenlevel 1andlevel2.Ourmethodologyhelpstodefineconsistentlythe Fig.3.Exampleofsemanticinconsistency.

Table6

SemanticcontrolreportwithcommentsonISO62264compliance.

Model:BP2SchedulingandlaunchingofaProductionOrderinjobshop Bringthetools

Cluster1:WorkshopChief Theobject“WorkshopChief”hasamasterintherepositoryassociatedtothemasterofthe function.

TheISO62264standardiscomplied with.

Cluster2:HandTools Theobject“HandTools” TheISO62264standardiscomplied

with.

Printtheinventoriesforaproductionorder

Cluster1:WorkshopChief Theobject“WorkshopChief”hasamasterintherepositoryassociatedtothemasterofthe function.

Verifyavailabilityofmaterialforaproductionorder Cluster2:Documentofaproduction

order

Theobject“WorkshopChief”hasamasterintherepositoryassociatedtothemasterofthe function.

Table7

Acontributionoflevelsofmaturity.

BestPractices RequiredLevel Contributiontoupperlevel

Framework Level2 ³/4ofLevel3

Blueprint Level1 Level3

Compliance Level1 ²/3ofLevel3

Integration Level2 ¹/2ofLevel3

(10)

relationofmanagementprocessestotheenterprisearchitecture, even without any strategic planning and budgeting processes (level3).

Themaincontributionsofourproposaltoenterprisematurity usingmodelsinenterprisearchitectureistheroleoftherepository obtained in the enhancement of business capabilities and processes documentation (i.e. Blueprint), and consequently in thecapacitytoevaluatecompliancewithstandards.

4.2.Keylessonslearnedfromthecasestudy

TheBusinessRepositorybasedonISO62624andenrichedwith differentbusinessknowledgeisbuilttobeusedbyanindustrial companytoensurethattheirbusinessobjectsandprocessesare

“standardcompliant”.Theobjectivesaretoensuretheconsistency of the MES information system and to reduce the cost of its implementationinon-demandsolutions.Thishasbeendoneby integratingourrepositoryintotheplatformoftheconsortium[39].

ThisplatformoffersorchestrationofMESservicesbasedonthe partners’existingapplications[40].

Theevaluationofourrepositoryusedvariousbusinesscases proposedbythesoftwarepublishersinvolvedintheplatform.Each ofthesecaseswasmodelledusingstandardbusinessobjectsand functions from the repository, irrespective of their source functions, or from SCOR model, ISO/IEC 62264, or software publishers.ThisdemonstratesthecapacityoftheMESrepository obtainedtoallowthemodellingofactualbusinesscasesinauser perspective,usingheterogeneousrepositories.

Basedonthiscasestudy,weproposeanalignmentmethodolo- gytoallowtheuseofexperts’knowledge,usingtherepositoryasa decisionsupporttoolforsemanticalignment.

To take into account the variety of alignment situations, dependingongranularity,perspectiveorcoverageofrepositories, weformalizedthealignmentresultsbylistingsemanticrelation- shipsbetweenthedifferentelementsofbothbusinessrepositories.

Fig.4showsthemeta-modelforsemanticalignmentoftheSCOR modelonISO/IEC62264.

Thesealignmentrelationsareaﬁrstapproachofameta-model ofalignmenttoextendamodellingplatformasanalignmenttool.

5.Conclusions

In this paper,we developa unidirectionalapproach toalign businessrepositoriesforenrichingacore-businessrepository.The approachconsistsoffoursteps:comparison,modelling,alignment and validation. This comparison is based on four levels of heterogeneity:syntactic,terminological,conceptualandpragmat- ic. On each of these levels, different situations of alignment, regarding granularity, views, or consistency, serve to classify differentalignmentrelationships, andthen todeﬁneanenrich- ment-based alignment approach for business repositories. This approachusessemanticcontrolrulesthatallowexpertstocheck theconsistencyofalignmentdecisions.

We assess the contribution of this approach to enterprise maturity in the use of standards and reference models when expertsaredealingwithheterogeneityinthecontextofmodel- driven projects and have to evaluate the compliance of their enrichedmodelswithstandards.

This new approach helps experts in the management of semantic heterogeneities between business repositories, as experimentedwiththroughtheintegrationofpartsoftheSCOR modelintheISO/IEC62246repository.

Firms can use this approach to comply with a standard repositorywhen expressingtheirrequirementsand toensurea reliableimplementationusingstandardcomponents.Forsoftware publishers,this“on-demand”approachisakeysuccessfactortobe selectedincompanies’projects.

Furtherresearchcouldbedonetobuildindicatorsofalignment levelsbetweenrepositories,basedonthemeta-modelandrules.

Such indicators could give the actors of a MES project (ﬁrms,

Fig.4.Meta-modelforsemanticalignmentbetweenbusinessrepositories.

(11)

publishers, and experts) an overall view of the alignment, in relation to the project’s progress and at different levels of granularity, so as to facilitate decisions on a greater or lesser degreeofintegration.

Acknowledgments

ThisresearchwaspartofaresearchprojectfundedbyFEDER andtheRégionRhône-Alpes(France),involvinga consortiumof eighteditorsofMESsolutions.

References

[1]F.B. Vernadat, Enterprise Modeling and Integration: Principles and Applications,Chapman&Hall,London,UK,1996496p.ISBN:978- 0412605505.

[2]C.H. Goh, Representing and Reasoning about Semantic Conf licts in HeterogeneousInformationSources".Thèsededoctorat,Massachusetts InstituteOfTechnology,Massachusetts,1997113p..

[3]C.Bussler,SemanticWebservices:reﬂectionsonWebservicemediationand composition,ProceedingsoftheFourthInternationalConferenceonWeb InformationSystemsEngineering,Galway,Ireland,2003,pp.253–260.

[4]S.M.Falconer, N.F.Noy,M.A.Storey,Towards understandingtheneedsof cognitivesupportforontologymapping,ProcOntologyMatchingWorkshop (2006)25–36.

[5]M.,Klein,Combiningandrelatingontologies:ananalysisofproblemsand solutions.,inProceedingsofWorkshoponOntologiesandInformationSharing atthe17thInternationalJointConferenceonArtiﬁcialIntelligence(IJCAI- 2001),G.A.Perez,H.H.Gruninger,H.Stuckenschmidt,M.Uschold,Editors.

2001:Seattle,WA,USA.

[6]W.Hasselbring,Informationsystemintegration,Commun.ACM43(6)(2000) 33–38.

[7]European Interoperability Framework for pan – European eGovernment Services,InteroperableDeliveryofEuropeaneGovernmentServicestopublic Administrations,BusinessesandCitizens(IDABC),.Rapportderecherchen 1.0:Luxembourg:2004,26p.

[8]MESA, 1997. https://services.mesa.org/ResourceLibrary/ShowResource/

2cedfe75-daed-4b9c-b187-f421cf90fdd2.

[9]H.Guangke,H.Weiping,H.Yanli,Y.Hui,Researchonreconﬁgurabilityof service-orientedmanufacturingexecutionsystem,IEEE10thInternational ConferenceonComputer-AidedIndustrialDesign&ConceptualDesign,26–

29Nov(2009).

[10]D.U.Saenz,A.Artiba,R.Pellerin,Manufacturingexecutionsystem–aliterature review,Prod.Plann.ControlJ.20(6)(2009)525–539.

[11]J.A. Vinhasis,Manufacturingexecutionsystems:theone-stopinformation source,Qual.Digest18(9)(1998)39–40.

[12]B.Hakanson,MESstandsformanufacturingimprovements,IIESolut.28(2) (1996)26–29.

[13]D.J.Adler,J.Herkamp,D.Henricks,R.Moss,Doesamanufacturingexecution systemreducethecostofproductionforbulkpharmaceuticals?ISATrans.34 (4)(1995)343–347.

[14]J.-S.Ulmer,Apivotal-basedapproachforenterprisebusinessprocessandIS integration,Enterp.Inf.Syst.(EIS)7(1)(2013)61–78.

[15]G. Doumeingts, B. Vallespir,D. Chen,GRAI griddecisional modelling,in handbookonarchitectureofinformationsystems,in:P.Bernus,K.Mertins, G.Schmith(Eds.),HandbookonInformationSystems,SpringerVerlag,Berlin, 1998,pp.313–337.

[16]C.Braesch,F.Théroude,A.Haurat,OLYMPIOS:amodeltocontrolbyprocess reengineering,InternationalConferenceSystemsEngineeringandInformation

&CommunicationTechnology(TIC'2000),Nîmes,France,2000.

[17]M.Zelm,F.B.Vernadat,K.Kosanke,TheCIMOSAbusinessmodellingprocess, Comput.Ind.27(2)(1995)123–142.

[18]K. Kosanke, J.G. Nell, Enterprise Engineering and Integration: Building InternationalConsensus,Springer-Verlag,Torino,Italy,1997pp.634.

[19]A.-W. Scheer, K. Schneider, ARIS—architecture of integrated information systems,in:P.Bernus,K.Mertins,G.Schmidt(Eds.),Handbookon ArchitecturesofInformationSystems,Springer,BerlinHeidelberg,2006,pp.

605–623ISBN:978-3-540-26661-7.

[20]K.Medini,J.P.Bourey,SCOR-basedenterprisearchitecturemethodology,Int.J.

Comput.Integr.Manuf.(IJCIM)25(7)(2012)594–607.

[21]CEN,ENV12204:AdvancedManufacturingTechnology–SystemsArchitecture –ConstructsforEnterpriseModelling,(1996) 44p..

[22]ISO,ISO15704:IndustrialAutomationSystems–RequirementsforEnterprise- referenceArchitecturesandMethodologies,(2000) 43p..

[23]ISO/TC and CEN/TC, "Enterprise integration – Framework for enterprise modelling(ISO19439)".2006,47p.

[24]ISO/TC, "DRAFT INTERNATIONAL STANDARD ISO/DIS 19440". Enterprise integration–Constructsforenterprisemodelling,2004,63p.

[25]ISO/IEC-1, "IEC 62264-1: Enterprise-control system integration". Part 1:

Modelsandterminology,2003,128p.

[26]ISO/IEC-2,"IEC62264-2:Enterprise-controlsystemintegration".Part2:Model objectattributes,2004,96p.

[27]ISO/IEC-3, "IEC 62264-3: Enterprise-control system integration". Part3:

Activitymodelsofmanufacturingoperationsmanagement,2007,212p.

[28]SupplyChainCouncil,SupplyChainOperations ReferenceModel (SCOR), Version10.0[online],(2010)Availablefrom:http://supplychain.org/scor/10.0 (Accessed02June2015).

[29]P.-A.Millet, AnAnalysis ofOrganizational andInformationalIntegration.

ApplicationtoERPBasedInformationSystems.ThèseDeDoctorat,INSAde Lyon,Lyon,2008308p..

[30]M.Zdravkovic,M.Trajanovic,H.Panetto,M.Lezoche,A.Aubry,Ontology-based supplychainprocessconﬁguration,in:M.Trajanovic(Ed.),InProceedingsof the34thInternationalConferenceonProductionEngineering,Niš,Serbia, 2011,pp.399–402.

[31]G.Stewart,Supply-chainoperationsreferencemodel(SCOR):theﬁrstcross- industryframeworkforintegratedsupply-chainmanagement,Logist.Inf.

Manage.10(2)(1997)62–67.

[32]Y.-D.Hwang,Y.-C.Lin,L.J.Jung,TheperformanceevaluationofSCORsourcing process–thecasestudyofTaiwan'sTFT–LCDindustry,Int.J.Prod.Econ.115 (2)(2008)411–423.

[33]P.-A.Millet,P.Schmitt,V.Botta-Genoulaz,TheSCORmodelforthealignment ofbusinessprocessesandinformationsystems,Enterp.Inf.Syst.3(4)(2009) 393–407.

[34]ISO,ISO10303:IndustrialautomationsystemsandintegrationProductdata representationandexchange–OverviewandFundamentalPrinciples.ISO TC184/SC4,1994,17p.

[35]R.,Tricker,"ISO9001:2000:TheQualityManagementProcess":VanHaren Pub,2006,186p.(B.Practice).

[36]KnowledgeWeb,2004.DeliverablesofKWEBProjectEU-IST-2004-507482.

Researchreport.

[37]WBF,"B2MML".Availableat:<http://http://www.mesa.org/en/B2MML.asp>

(Accessed07September2016).

[38]NationalAssociationofStateShiefInformationOfﬁcers,NASCIOEnterprise ArchitectureMaturityModel[online],(2003) Availablefrom:http://www.

nascio.org/publications/documents/NASCIO-EAMM.pdf[AccessedJuly2015].

[39]AGILIUM,"Le trophéedu LABELMESTRIA",http://www.agilium.com/spip.

php?page=article&id_article=184&connect=awensys.Accessed07September 2016.

[40]H.Verjus,MESTRIA:LePilotaged'AtelieràLaCarte,(2011) Availablefrom:

http://www.jiteconline.com/mestria___le_pilotage_d_atelier_a_la_carte- fr1827.html.Accessed07September2016.

[41]M.Larson,Gainshop-ﬂoorcontrolwithMESsoftware,Quality36(6)(1997) 68–70.