SlideShare ist ein Scribd-Unternehmen logo
1 von 31
Downloaden Sie, um offline zu lesen
A viewframeworkformodelingandchangevalidation 
of artifact-centricinter-organizationalbusinessprocesses 
SiraYongchareon a, Chengfeiliu b, YuJianc, XiaohuiZhao d 
a Department ofComputing,UnitecInstituteofTechnology,Auckland,NewZealand 
b FacultyofScience,EngineeringandTechnology,SwinburneUniversityofTechnology,Victoria,Australia 
c School ofComputerandMathematicalSciences,AucklandUniversityofTechnology,NewZealand 
d Faculty ofBusiness,GovernmentandLaw,UniversityofCanberra,Canberra,Australia 
a r t i c l e info 
Article history: 
Received27November2012 
Receivedinrevisedform 
13June2014 
Accepted21July2014 
Availableonline1August2014 
Keywords: 
Business processmodeling 
Artifact-centric workflows 
Inter-organizationalbusinessprocesses 
Process views 
Process abstraction 
Model verification 
a b s t r a c t 
Over thepastseveralyears,moreefficientapproacheshavebeenonincreasingdemands 
for designing,modeling,andimplementinginter-organizationalbusinessprocesses.Inthe 
process collaborationacrossorganizationalboundaries,organizationsstillstayautonomic, 
whichmeanseachorganizationcanfreelymodifyitsinternaloperationstomeetits 
privategoalswhilesatisfyingthemutualobjectiveswithitspartners.Recently,artifact- 
centric processmodelinghasbeenevidencedwithhigherflexibilityinprocessmodeling 
and executionthantraditionalactivity-centricmodelingmethods.Althoughsomeefforts 
havebeenputtoexploringhowartifact-centricmodelingfacilitatesthecollaboration 
between organizations,theachievementisstillfarfromsatisfactionlevel,particularlyin 
aspects ofprocessmodelingandvalidating.Tofillinthegaps,weproposea view 
framework for modelingandvalidatingthechangesofinter-organizationalbusiness 
processes. Theframeworkconsistsofanartifact-centricprocessmeta-model,publicview 
constructingmechanism,andprivateviewandchangevalidatingmechanisms,which 
are speciallydesignedtofacilitatetheparticipatingorganizationstocustomizetheir 
internal operationswhileensuringthecorrectnessofthecollaboratingprocesses.Wealso 
implementasoftwaretoolnamed Artifact-M to helporganizationstoautomatically 
constructaminimalandconsistentpublicviewfromtheirprocesses. 
& 2014ElsevierLtd.Allrightsreserved. 
1. Introduction 
Recently,service-orientedarchitecture(SOA)hasbecomea 
predominantITtoolforfacilitating businessestomeetthe 
changingrequirementsofthemarket.SOAparticularlyenables 
thebusinesscollaborationacrossorganizationsbycomposing 
Webservicestoachieveamutualbusinessgoalwithout 
comprisingtheautonomyofparticipatingorganizations.The 
furtherapplicationofSOAinfacilitatingbusinesscollaboration 
calls forefficientapproachesfordesigning,modelingand 
implementinginter-organizationalbusinessprocesses [50]. 
Recently,workbyDesaietal. [20], Ghattas,Montalietal. 
[63] and [83] showsthatthequalityofcoordinatingorganiza- 
tions inaservice-orientedcollaborationreliesonthreemajor 
requirements,viz., compliance, flexibility, and autonomy.Com- 
pliance requiresallpartiesmustprovidetheservicesasthey 
havepromisedinthecollaborationcommitment,suchasa 
servicelevelagreement. Flexibility allowseachpartytoown 
thefreedomofchangingandimplementingitsownprocessin 
thecollaboration.Lastly, autonomy indicateseachparticipating 
organizationactsindependentlyandisnotobligedtorevealits 
Contents listsavailableat ScienceDirect 
journalhomepage: www.elsevier.com/locate/infosys 
InformationSystems 
http://dx.doi.org/10.1016/j.is.2014.07.004 
0306-4379/& 2014ElsevierLtd.Allrightsreserved. 
E-mail addresses: sira@maxsira.com (S. Yongchareon), 
cliu@swin.edu.au (C. liu), jian.yu@aut.ac.nz (J. Yu), 
xiaohui.zhao@canberra.edu.au (X. Zhao). 
Information Systems47(2015)51–81
ownprivateinformation(orprocess)tootherparties. 
Althoughservicechoreography definescommoncollaboration 
behaviorsandkeepstheflexibilityandautonomyofeach 
participant, actualchoreographymodelingapproachesand 
relatedmodelinglanguagesmainlydescribethecollaboration 
fromaproceduralperspective, and focusoncontrol-flow, 
messagesequencing,etc.,insteadoffromadataperspective. 
As such,flexibilityandautonomyisstilllimitedbythenature 
of choreographymodelinglanguages.Theworkondeclarative 
specificationofservicechoreographieshasbeenproposedby 
Montalietal. [63], onthebasisofDecSerFlowlanguage [84], 
toovercomesuchlimitationsandsupportdynamicand 
complexinter-organizationalprocessspecifications.Priorto 
theemergenceofservicechoreography,processviewhasbeen 
adoptedtoaddresstheflexibilityandautonomyissuesin 
businesscollaborations,andtheimprovementhasbeen 
extensivelyevidencedinworkbyVanDerAalstandBasten 
[81], LiuandShen [51,52], SchulzandOrlowska [77], Chiu 
et al. [16,17], Chebbietal. [14], EshuisandGrefen [10], Zhao 
et al. [100,101], Jiangetal. [40] andEshuisetal. [27]. 
Originally,apublic-to-privateviewapproachhasbeenintro- 
duced by [85] andvanderAalstetal., [82] to resolvethe 
privacyandautonomyissuesaswellastosupportchange 
managementindynamiccollaboration.Mostrecently, [83] 
haveproposedamultipartyprocess-orientedcontractto 
supportservicechoreographybasedontheconceptsof 
public/privateviewsandaccordancealongwithoperating 
guidelines [61,57]. However,allofaboveworksfollowthe 
traditionalactivity-centricbusinessprocessmodelingpara- 
digm andtherebyinheritthelimitationsindatamanagement, 
processintegration,andprocessmodification,becausetradi- 
tional modelingapproacheslackadequatesupportsofauto- 
matedtoolsforbusinessprocessinter-operationandprocess 
schemareuse [38,39]. 
In thepastfewyears,anewmodelingapproachhasemer- 
ged,i.e., artifact-centric (operational)businessprocessmodel- 
ing [68]. Insteadofcontrolflowsofabusinessprocess,business 
documentsandtheirevolutionthroughabusinessprocess 
becomethemainmodelingobjects.Thisapproachdepictsa 
businessprocessinfourdimensions,viz.,businessartifacts, 
lifecycleofartifact,services,andassociationsbetweenartifacts 
and services [34]. Thelifecycleofanartifactisdefinedinterms 
of “businessstages” and thepossibleevolutionoftheartifact. 
Theevolutionofanartifactandoperationsofrelatedservices 
arespecifiedintermsoftheirassociations,whichcanbe 
expressedinadeclarativemanner,e.g.,usingECArules. 
As anemergingtoolGuard–Stage–Milestonemeta-model 
is becominganewdeclarativeapproachtomodeling 
artifactlifecyclebasedonavariantofstatemachines 
[35,36,21,22,80]. Withtheeffortsofnumerousacademic 
researchersandindustrialpractitioners,artifact-centric 
modelingapproachhasbeenextensivelyrecognizedto 
be withhigherlevelofrobustnessandflexibilityfordescrib- 
ingprocessspecificationcomparedtotraditionalactivity- 
centricapproaches.Artifact-centricprocessmodeling 
receivescomprehensivecontributionsintermsofbusiness 
transformationpractices [6,7,13], foundations [5,59], design 
methodologies [8,19,56,64,65,73,74,86], modelspecifica- 
tion,construction,andverification [54,30,31,44,25,29,102, 
103,21], workflowrealization/execution [18,53,66,91,67,88], 
and monitoring/conformancesupports [55,28]. Uptopresent, 
artifact-centric approachhasbeenappliedtoseveralindus- 
try domainssuchas healthcare (e.g., PHILharmonicFlows 
framework [45,15]), insurance (e.g., in [44]), and finance (e. 
g., IBMGlobalFinancing [13]). However,comparedwith 
traditionalactivity-centricapproaches,furtherresearchis 
sought afterintheareaofbusinesscollaboration. 
By nowtwomainapproacheshavebeenproposedfor 
artifact-centricinter-organizationalprocesses. Theinitial 
attempt usesartifact-centricinteroperationhubtofacilitate 
and supportinter-organizationalworkflows(inanorchestra- 
tion perspective)amongmultipleautonomousstakeholders 
[37]. Recently,thisworkhasbeenbroughtforwardtoanEU- 
funded projectcalledArtifact-CentricServiceInteroperation 
(ACSI) [3]. Itispromisedtosupportalargenumberofservice 
collaborationbyusingartifact-centricinter-operationsandto 
achievedramaticsavingsoverconventionalapproaches.On 
the otherhand,theartifact-centricchoreographyapproach 
hasdefinedinteractingartifact-centricprocesses [58,79]. 
Althoughflexibilityisnaturallydeemedasoneofthebenefits 
fromartifact-centricmodelingapproaches,acomprehensive 
studyonsupportingorganizationstoachieveallthethree 
collaborationrequirementsisstillmissing.Basedonliterature 
andpractices,wehaveobservedthatview-basedapproaches 
tointer-organizationalbusinessprocessmanagementcan 
provideapromisingandefficientwayofprocessmodeling 
and changemanagementtoaddresssuchrequirements; 
nevertheless,ithasnotbeenyetmuchexploredinthecontext 
of artifact-centricinter-organizational businessprocesses. 
Therefore,inthisarticle,wearetoexploretheideaofprocess 
viewinanartifact-centricperspectiveanddevelopaframe- 
workthatcanhelporganizationstomeettheaforementioned 
requirementsinacollaboration environment.Wesummarize 
our contributionsasfollows: 
 Weproposeaformalartifact-centricviewframework 
based onLTS(LabeledTransitionSystem).Thisframe- 
workconsistsofthreeparts:(1)anartifact-centricMeta 
model forinter-organizationalbusinessprocesses,(2) 
notion of privateview for capturinglocalprocessesof 
participating organizations,and(3)notionof public view 
for servingasanagreedcontractofthecollaboration. 
With public/privateviews,organizationsareableto 
autonomouslyparticipateinthecollaborationwhile 
being freetochangetheirlocalprocesses. 
 Wedesignanalgorithmforautomaticallyconstructinga 
consistent,minimalpublicviewbasedonlocalpro- 
cesses ofanorganization.Tothebestofourknowledge, 
this isthefirstalgorithmforautomaticallyconstructing 
a collaborationcontractthattakesintoaccountinterac- 
tion behaviorsofartifacts. 
 Wedevelopaverificationmechanismforartifact-centric 
processesthatallowsorganizationstochangetheirlocal 
processes,throughtheuseofprivateviews,while 
preservingthecorrectnessandconsistencyoftheover- 
all collaboration.Astheverificationisperformedlocally, 
our mechanismdoesnotsufferfromthestateexplosion 
issue thatmayoccuringlobalverificationapproaches. 
The remainderofthisarticleisorganizedasfollows: 
Section 2 introducesthemotivationofourartifact-centric 
52 S. Yongchareonetal./InformationSystems47(2015)51–81
approachandtheconceptofprocessviews. Section 3 
presentsourviewframeworkforartifact-centricbusiness 
processes. Sections 4 and 5 address howtocreatepublic 
viewsandprivateviewsaswellashowtoensureview 
consistencyandvalidatelocalprocesschanges,respec- 
tively. Section 6 discusses animplementedprototypefor 
the proof-of-conceptpurpose. Section 7 reviewstherelated 
work.Finally,conclusionandfutureworkaregivenin 
Section 8. 
2. Motivatingexample 
In thissection,weuseapurchasingprocessinthesupply 
chaincollaborationdomainasanexampletoillustrateand 
motivatetheartifact-centricapproachtomodelinginter- 
organizationalbusinessprocesscollaboration.In Fig.1, a 
completepurchasingprocessmodelinthecollaborationis 
illustratedbasedonanartifact-centricperspective,which 
involvesthreerolesofparticipatingorganizations: Buyer, 
Supplier, and Logistics. Weinitiatethediscussionofthis 
examplebyidentifyinginvolvedbusinessartifactsand 
describinghowtheyaremodeledinthiscollaboration. 
At thebeginningstage,allpartiesidentifyandmodel 
their requiredbusinessartifactsoftheirlocalprocesses. 
This stepincludesdefiningorganization-ownedartifacts 
(called local artifacts) thatareinternallyused/managedby 
individualpartyaswellastheircommonlyagreedartifacts 
(called sharedartifacts) thatareusedforthecoordination 
betweenpartiesinthecollaboration.Weconsiderthatthe 
lifecycleofasharedartifactshouldrepresentitsagreed 
business stagesandpossiblestepstowardsthecompletion 
of theprocess.Inotherwords,sharedartifactsshouldbe 
used asamutualpointofinterestofallparties;therefore, 
the coordinationoccursatsomepointswheretheyare 
processed.Intheimplementation,thesesharedartifactsact 
as messagesthataresentandreceivedbytheseorganiza- 
tions. Baseduponthecurrentprocessingstateoftheshared 
artifact, aresponsibleorganizationthathasreceivedthis 
artifact willinvokeaspecificserviceaccordingtoacorre- 
sponding businessruledefinedbythatorganization.This 
service willthenreadorupdatethesharedartifactand 
otherlocalartifactsdefinedinthespecification. 
In Fig. 1, wecanseethattheinter-organizationalprocess 
consists of PurchaseOrder (PO), Shipping Order (SO), and 
Invoice (IV) as sharedartifacts. Apartfromthem,Buyer, 
Supplier,andLogisticshave Quote (Q) and Payment (P), 
Picking List (PL) and DeliveryNote (DN), and Shipping List 
(SL) astheir local artifacts, respectively.Inthefigure,we 
also usedashedlinetodepictthe synchronizationdepen- 
dencies betweenartifacts(local-local,shared-shared,or 
local-shared). 
Nextwebrieflydescribetheprocessin Fig. 1 from an 
artifact-centric perspective.Atthebeginningoftheprocess, 
a buyerinitiatesthecreationofthe Quote and the PO 
documents/artifacts. Oncethe quote is approved,the PO is 
confirmed andsenttoaselectedsupplier.Whenthe 
supplier receivesthe PO, itcreatesa PL document forthe 
purpose ofacquiringgoodsforthat PO. Ifthegoodsrunout, 
then thesupplierrejectstosupplythemandthencancels 
the PO. Otherwise,thegoodsarefilled,andthenthe 
supplier generatesaninternal DN document andcreatesa 
SO document forthedesignatelogisticscompany.Oncethe 
SO is received,thelogisticscompanycreatesa SL document 
that isusedforpickingupthegoodsfromthesupplier’s 
shipping pointandalsodeliversthegoodstothebuyer. 
Afterthat,thesuppliercreatesan IV document andsendsit 
to thebuyer.Sometimelater,thebuyerclearsthetotal 
amount owinginthe IV, consequently,thesuppliermarks 
the PO as closed.Thispurchasingprocesscompleteswhen 
Buyer (L1) Supplier (L2) Logistics(L3) 
Purchase Order(PO) 
Picking List(PL) 
ready tofill Filled order 
checking 
Quote (Q) 
created approving 
approved 
In stock 
Shipping Order(SO) 
In transit 
arrived 
Invoice (IV) 
cleared 
Shipping List(SL) 
Queued 
completed picked 
created confirmed 
L1 
closed 
L2 billing L2 
canceled 
accepted filled 
acquiring 
L2 
L2 
delivering 
L2 
ready toship 
L2 
L3 
created 
L2 
L2 
Out ofstock 
L2 
L1 
sent issued 
L2 
L2 
Payment (P) 
approving created 
sent 
Delivery Note(DN) 
prepared 
transferring 
L3 
scheduled 
dispatched 
L3 
unpaid 
L1 
L1 
rejected 
L1 on hold 
L1 
clearing 
L2 
Fig. 1. A completeartifact-centricinter-organizationalpurchasingprocess 
S. Yongchareonetal./InformationSystems47(2015)51–81 53
the PO is inthe closed state, the SO is inthe arrived state, 
and the IV is inthe cleared state. 
In thecontextofinter-organizationalprocesscollabora- 
tion, someconcerns—including changeflexibility,change 
verification,andprivacy—raisedinthetraditionalactivity- 
centric approach(e.g.,in [85,32]) shouldalsobeconsidered 
in anartifact-centricsettingasdiscussedbelow. 
First, organizationsprefertokeepthe freedom of mod- 
ifying theirinternalprocesseswithoutrevealingtheirlocal 
changestootherparticipatingparties.Inotherword,once 
they haveagreedontheoverallprocess,theyshouldhave, 
possibly,thehighestlevelof flexibility tomodifytheirown 
local processeswhileremaining autonomous in thecolla- 
boration [85,82,83]. Forexample,Suppliermaymodify 
some processingstepsofits PL artifact, andsuchchange 
should notbeexposedtoBuyerandLogistics.Apartfrom 
the changeofexistinglocalartifacts,theyshouldbe 
allowedtoaddnewlocalartifactstotheirprocesscaused 
by processexpansion/improvement.Forinstance,Supplier 
may needtoincorporate InventoryList (IL), whichisused 
for inventorymanagement,totheirprocess.The IL artifact 
needs tointeractwithexistinglocalartifact PL. Italso 
implies that IL indirectlycontributestothepartofshared 
artifact PO through PL. Inordertosupportthechangeof 
local processes,organizationsneedtoknowwhatthey 
wanttomodifyandhowsuchmodificationcanbeapplied. 
Second, stemmedfromthefirstreason,thechangeto 
anylocalprocessshouldnotaffectthebehaviorofthe 
overallcollaboration.Thisimpliesthatalocalprocess 
should always comply with thecontractagreedbyall 
parties, andthechangemadetolocalprocessesshould 
not affecttheoverallprocess.Therefore,allpartiesmust 
ensure thattheirlocalchangesdonotviolatethecollabora- 
tion contract.Forexample,ifSupplierchangestheir PL 
artifact byremovingatransition ready_to_f ill-f illed_ 
order, thenthismayaffectthetransition acquiring-f 
illed of PO; consequently,Logisticsisnotabledetermine 
whether thegoodsarereadytobepickedfordelivery.In 
addition, allparticipatingpartiesshouldalsobeawareof 
the localchangesthatwillpropagateandeventuallyaffect 
the overallprocess.Forinstance,iftherearesomechanges 
made on IL artifact, thentheydirectlyaffectthe PL artifact 
and eventuallyaffectthe PO artifact. Thisraisestheissueof 
how wecanguaranteethatlocalchangesmadebyan 
individualpartydonotleadtoincorrectbehaviorofthe 
overallcollaboration.Inotherwords,organizationsmust 
ensure thatchangesintheirlocalprocessesarein compli- 
ance with whattheyhavepromisedtoprovide. 
Last, organizationsconcernabouttheir privacy. Inthe 
collaboration,itisnecessaryfortheparticipatingorganiza- 
tions torevealcertaindetailsoftheirinternalprocesses 
among themselvesatanadequatelevelofvisibilityasto 
establish anoverallpictureofthecollaboration,whichis 
used asaprocessagreementorcontractamongthem.In 
our artifact-centricinter-organizationalprocessmodeling 
approach,thelevelofvisibilitycanbedeterminedbasedon 
the typeofartifacts.Inotherwords,detailsoflocalartifacts 
should bekeptinvisibletoexternalpartiesasmuchas 
possible whiletheyalsosupporttheoverallprocessby 
means ofdependencyassociationswithsomeprocessing 
part (lifecycle)ofthesharedartifact(s).Considerthe DN 
artifact in Fig. 1 for instance.The DN is privatelyusedby 
Supplier;however,wecanseethatithasdependency 
associations withthesub-lifecyclesofthe PO (f illed-ready_ 
to_ship-dispatched) and SO (-created-scheduled- 
in_transit), whicharesharedartifacts.Simi- 
larly,both Quote and PL artifacts areusedtosupportsome 
processingstepsofthe PO artifact. Toachieveprivacy,ifa 
processingstepofasharedartifactisexclusivelycontrolled 
by thelocalartifact(s)ofoneparty,thenthatstepshould 
notbevisibletootherparties.Forexample,the ready_ 
to_ship state anditsrelatedtransitionsof PO should bekept 
invisibletoexternalpartiesbecause DN is alocalartifact 
of Supplier.Apartfromthat,wecanseethatthestep 
created-scheduled of SO can behiddentoexternalparties 
as well.Inordertosupporttheprivacyrequirement,an 
organization needstohaveamechanismtoidentifyan 
artifact and/oritspartsthatcanbeinvisibletotheother 
parties. Thisbringsupinthequestionoftowhatextentofa 
local processcanbekeptprivatewhilenotaffectingthe 
successful establishmentofthecollaboration. 
The threemajorconcernsdiscussedabovecallforan 
approachtoefficientmodelingandchangemanagementof 
artifact-centric inter-organizationalbusinessprocesses.As 
previouslydiscussed,inthisarticle,westudyhoworgani- 
zations canapplytheconceptof view to supportand 
facilitatethemodelingoftheircollaboratingbusiness 
processesinanartifact-centricparadigm.Particularly,we 
borrowtheideaofpublicandprivateviewsapproachto 
inter-organizationalworkflowswhichisoriginallystudied 
in activity-centricbusinessprocessmodelingapproaches 
[85,82,83], andthenexploreitinthecontextofartifact- 
centric processes. 
3. Viewframeworkforartifact-centricinter- 
organizationalprocesses 
In thissection,weintroduceanddiscussourview 
frameworkforartifact-centricinter-organizationalbusiness 
processes.In Section 3.1, weoverviewourviewframework 
that aimsataddressingtheaforementionedrequirements 
for efficientinter-organizationalbusinessprocessescolla- 
boration.Thenin Section 3.2, weformallydefinetheview 
frameworkanddiscusshowtouseittomodelinter- 
organizational businessprocessesfollowedbythediscus- 
sion ofbehaviorpropertiesin Section 3.3. 
3.1.Overview 
Westartthissectionbyintroducingourviewframe- 
workformodelingartifact-centricinter-organizational 
business processes-whichisinfluencedbytheprocess- 
orientedcontractapproachproposedforservicechoreo- 
graphy [83]. Ourartifact-centricviewframeworkconsists 
of thefollowingfourparts:(1)Artifact-Centricbusiness 
Processmodelforinter-organizationalbusinessprocesses 
(ACP-imodel), (2)publicviewanditsconstructionmethod, 
(3) processchangemechanism,and(4)processchange 
validation. Fig. 2 depicts theoverallpictureoftheframe- 
workbytakingourmotivatingpurchasingprocessesasan 
illustrativeexample. 
54 S. Yongchareonetal./InformationSystems47(2015)51–81
Our viewframeworkisdevelopedbasedonthenotion 
of privateview and public view proposedby [85]. A private 
view is usedtocapturethe local processes of eachindividual 
organization, whilea public view of aparticularcollabora- 
tion processisan abstract representationoftheoverall 
processthatisnecessaryforthecoordinationandhidesthe 
details oftheprivateprocessesofeachorganizationas 
much aspossible.Organizationscanachieveanefficient 
collaboration(regardingthethreerequirements)basedon 
the public–privateviewsapproachbythefollowingfour 
stepsasillustratedin Fig. 2. 
(1) Constructacompleteartifact-centricmodelspecifica- 
tion ofinter-organizationalbusinessprocess 
(2) Createa public view that isservedasamutualagree- 
ment, i.e.,contract,ofthecollaboration. 
(3) Eachorganizationcanchangeandvalidatetheirlocal 
processes,viatheuseof privateview, withouttheneed 
for globalverification. 
First, allparticipatingpartiesspecifyacompletespeci- 
fication oftheirinter-organizationalbusinessprocessasto 
achievetheirgoalofthecollaboration.Aspreviouslydis- 
cussed, thecoordinationamongthemcanbespecifiedby 
defining allsharedartifactsandtheirinteractionswithlocal 
artifacts (fromeachparty).WecalltheArtifact-Centric 
Processmodelofthecompleteinter-organizationalprocess 
an ACP-imodel. Onceallparticipatingorganizationshave 
agreedonthecompletemodeloftheartifact-centric 
collaborationprocess,theyconstructanagreedpublicview 
of suchprocess.Thispublicviewrevealsonlythenecessary 
information ofartifactsthatisrequiredtobeusedforthe 
coordinationamongparties.Intheperspectiveofartifact- 
centric processmodeling,apublicviewshouldonlycontain 
sharedartifacts and shouldnotrevealanylocalartifactor 
the partofasharedartifactthatissupportedbylocal 
artifact(s). 
Once thepublicviewisbuiltasacontract,each 
participating partyhastheresponsibilitytoensurethat 
its localprocesses(i.e.,responsiblepartsofsharedartifacts 
and theirlocalartifacts)canprovidewhathasbeen 
specified inthecontracttootherpartiesinvolvedinthe 
collaboration.Themostimportantthingsareasfollows:(1) 
the constructedpublicviewmustbesoundandconsistent 
with itsbaseprocess;and,(2)whenanorganization 
modifies itsownpartsofbothtypesofartifactsusedin 
the collaboration,itmustguaranteethatsuchlocalchanges 
do notcompromisethecorrectnessoftheoverallcollabora- 
tion. Wewilldiscusstheconstructionofapublicviewand 
changemanagementinalocalprocessindetailin Section 4 
and 5. 
3.2. SyntaxforACP-imodel 
Here,weformallydefinetheartifact-centricprocess 
model forinter-organizationalbusinessprocesses,or ACP-i 
model, whichisanextendedversionoftheACPmodel 
presentedinourpreviousworks [97,98,95,96,94]. An ACP-i 
model consistsoffourmaincomponents: roles, artifacts, tasks, 
and businessrules.Roles defineasetofparticipatingorganiza- 
tionrolesinthecollaboration.An artifact isabusinessentity 
or anobjectinvolvedininter-organizationalbusinesspro- 
cesses. A task is ownedbyoneorganizationinthecollabora- 
tionandisusedtoperformread/updateoperationsonartifact 
(s).A business rule is definedbyafamilyofconstraintsin 
a Condition-Action styletodescribewhichserviceisin- 
vokedandwhichstateofanartifactischangedunderwhat 
condition. 
Definition 1. (Artifact class).Anartifactclassabstractsa 
group ofbusinessartifactswiththeirattributesandstates. 
Artifact classC (or artifact if thecontextisclear)isatuple 
(A; S; sinit ; Sf ) where, 
 A ¼ fa1; a2; …; axg; aiAAð1rirxÞ is aname-value 
pairedattributevariable, 
 S ¼ fs1; s2; …; syg; siASð1riryÞ is astate, 
 sinit =2 S denotesthepseudoinitialstate, 
 Sf S is asetoffinalstates. 
ACP-i 
ACP ACP ACP 
ACP-i (modified) 
ACP ACP ACP 
PUBLIC VIEW 
pa(ACP-i) 
private viewchangevalidation 
public viewconstruction 2 
4 
inter-org processconstruction 1 
changes inlocalprocesses 
Purchase Order 
Shipping Order 
Invoice 
Buyer Supplier Logistics 
Picking List Quote 
Shipping List 
Payment 
Deliver Note 3 
′ 
Fig. 2. Our viewframeworkforartifact-centricinter-organizationalbusinessprocesses. 
S. Yongchareonetal./InformationSystems47(2015)51–81 55
Definition 2. (Artifact schema).An artifactschema, 
denotedas Z, containsasetofartifactclasses,i.e., Z ¼ 
fC1; C2; :::; Cxg where CiAZð1rirxÞ is anartifactclass. 
Wealsodefinesomebasicpredicatesoverschema Z to be 
used fordefiningbusinessrulesasfollows: 
 def inedðC; aÞ iff attribute aAC:A of artifactofclass C is 
defined; 
 instateðC; sÞ iff state sAC:S of artifactofclass C is active. 
Next,wedefineabusinessruletoexpressthecontrol 
logic ofabusinessprocess. 
Definition 3. (Business Rule).Abusinessruleregulates 
which taskcanbeinvokedunderwhatpre-condition.The 
conditional effectisalsodefinedtowhatpost-condition 
needs tobesatisfiedafterperformingsuchtask. Business 
rule r is triple ðλ; β; vÞ where 
 λ and β are thepre-conditionandthepost-condition, 
respectively.Bothconditionsmaycontainthefollowing 
twotypesofpropositionsoverschema Z: (1)state 
proposition(the instate predicate)and(2)attribute 
proposition(the def ined predicate).Tomakethetarget 
statedecidable,wedonotallowthepost-conditionto 
include disjunctioninstatepropositions. 
 vAV is ataskoracompositetask(i.e.,morethanone 
atomictask); V ¼ fv1; v2; :::; vxg is asetoftasksof 
which performsoperationsonsomeartifacts C1, C2,…, 
Cy where CjAZð1rjryÞ. 
In ordertomaintaintheexistenceofvalidandexplicit 
state changesofanartifactinbusinessrule r, werequire 
that thereexistsacoupleof instate predicatesofthat 
artifact inboththepre-conditionandthepost-condition 
of r, i.e.,wehavestates sx; syAC:S such that instateðC; sxÞ 
existsin r:λ and instateðC; syÞ existsin r:β. Thestatechange 
referstoeitheratransitionfromonestatetoanotherstate, 
or toitself. Table1 lists somebusinessrulesusedinour 
purchasingprocessin Fig. 1. 
Wealsoclassifybusinessrulesintotwotypesbasedon 
the existenceofthe instate predicateinthepre-andpost- 
conditions ofabusinessrule.Thefirsttypeonlychanges 
the stateofonesingleartifact,whilethesecondtype 
simultaneouslychangesthestatesofmultipleartifacts,i. 
e., morethanonepairof instate predicates(onepairforone 
artifact) mustappearinthepre-andpost-conditionsofa 
single businessrule.Wecallthesecondtype synchroniza- 
tion (sync)rules as theyareusedforexpressingsynchroni- 
zation betweenartifacts. 
Definition 4. (Sync rule).Businessrule r is a sync rule for 
artifact Cx and artifact Cy if thereexists instate ðCx ; siÞ and 
instate ðCy ; smÞ in r. and instateðCx ; sjÞ and instateðCy ; snÞ in 
r:β, where si; sjACx:S and sm; snACy:S. 
As mentionedabove,asingle sync rule can beusedto 
synchronizemorethantwoartifacts. 
Example 1. In Table1, wecanseethatbusinessrules r2 is 
used toexpressonlythestatechangeofthe PO artifact 
(conf irmed-accepted), whilebusiness(sync)rule r1 is used 
tosimultaneouslychangestatesoflocalartifact Quote 
(approving-approvedÞ and sharedartifact PO (on_hold-conf 
irmedÞ. Similarly,business(sync)rule r3 is usedforthe 
synchronizationof DN and SO; andsyncrule r4 is usedfor 
synchronizingfourartifacts PO, DN, SO, and IV (wheretwo 
tasks aredefinedinitsaction). 
Next,wedefine ACP-i model to capturethecomplete 
specification ofaparticularinter-organizationalbusiness 
processinthecollaboration.Asdiscussedin Section 2, 
sharedartifactsshouldbeusedascoordinationmeansin 
inter-organizationalbusinessprocesses.However,ashared 
artifact cannotbecompletelymodeledsolelybyasingle 
Table1 
An exampleofbusinessrulesforourpurchasingprocess. 
r1: Buyerapproves Quote q to confirm PurchaseOrderpo for aselectedsupplier 
Pre-condition instate(q, approving) 4 instate(po, on_hold) 4 defined(po,OrderID) 4 defined(po.SupplierID) 
Task approve(q,po) 
Post-condition instate(q, approved) 4 instate(po, confirmed) 4 defined(po.SubmitDate) 
r2: Supplier accepts PurchaseOrderpo 
Pre-condition instate(po, confirmed) 4 defined(po,OrderID) 4 defined(po.SupplierID) 
Task acceptPO(po) 
Post-condition instate(po, accepted) 
r3: Supplier creates Shipping Orderso from DeliveryNotedn 
Pre-condition instate(dn, prepared) 4 defined(dn.ShipperID) 4 instate(so, init) 
Task createShipping (dn,so) 
Post-condition instate(dn, transferring) 4 instate(so, created) 4 defined(so.ShippingID) 
r4: Supplier dispatchesgoodsfor PurchaseOrderpo that tobeshippedby Shipping Orderso, 
and simultaneouslyissues Invoiceiv to theBuyer 
Pre-condition instate(po, ready_to_ship) 4 instate(dn, transferring) 4 instate(so, scheduled) 4 instate(iv,init) 
Task dispatchGoods(po, dn,so)issueInvoie(po,iv) 
Post-condition instate(po, delivering) 4 instate(dn, dispatched) 4 instate(so, in_transit) 4 instate(iv,issued) 
4 defined(iv.InvoiceID) 4 defined(iv.OrderID) 
56 S. Yongchareonetal./InformationSystems47(2015)51–81
organization duetoitssharingnature.Therefore,allparti- 
cipating organizationsmustagreeonboththestructure 
(data model)andthebehavior(lifecycle)oftheshared 
artifacts inordertodefinetheACP-imodel.Itisworth 
noting thatinordertomaintaintheintegrityofashared 
artifact togetherwithalltheinvolvedtasksandbusiness, 
we assumethatthenameofitsstateandattributemustbe 
uniquelyidentifiedforthesamemeanacrosscollaborating 
organizations. 
Definition 5. (Organization Role).Wedenote L ¼ 
fl1; l2; …; lxg for asetof organization roles, where 
liALð1ixÞ is aroleoforganizationsthatparticipateinan 
inter-organizationalbusinessprocess. 
Definition 6. (ACP-imodel).Givenasetoforganization 
roles Linvolvedinacollaboration,wedefinean ACP-i 
model, denotedas Π^ ¼ ðZ; V; R; L; γÞ, fortheirinter- 
organizational businessprocesswhere 
 Z is anartifactschema, V is asetoftasks,and R is asetof 
business rules, 
 L is asetofparticipatingorganizationroles, 
 γ: Z [ V [ R-2L is arolemappingfunctionfroman 
artifact class,abusinessrule,orataskontoanorganiza- 
tion role(s)asfollows: 
(c) γðCÞ returnsasetofroles fl1; l2; …; lxg, where 
liALð1rirxÞ is arolethatcanaccess(read/write) 
artifact CAZ and theownerof C must beamong 
roles γðCÞ. Notethata sharedartifactC implies that 
jγðCÞj41, andif C is a local artifact then jγðCÞj ¼ 1. 
(c) γðvÞ returnsrole lAL of organizationwhoownstask 
vAV. Notethatataskcanperformread/write 
operationsoneitherlocalartifactorsharedartifact 
or both. 
(c) γðrÞ returnsrole lAL of organizationwhoowns 
business rule rAR. 
In addition,wedefinetwoauxiliaryfunctionsover 
business rules R and artifactschema Z in Π^ 
 function pre_sðr; CÞ returnsasetofstates{s1; s2; :::; sxg 
whereforsomebusinessrule rAR, state 
siAC:Sð1rirxÞ occurs inatleastone instate predicate 
of thepre-conditionof r; and, 
 function post_sðr; CÞ returnsasetofstatesofartifact C 
appearing inthepost-conditionof r. 
Givenan ACP-i model, wecanderivealocalACPmodel 
for anorganization'slocalprocess.Notethatinalocal 
processofanorganization,theattributesandstatesofits 
sharedartifactcanbeobtainedfromthe ACP-i model if they 
arespecifiedinthebusinessrulesoflocalprocess. 
Definition 7. (local ACPmodel).Given Π^ ¼ ðZ; V; R; L; γÞ be an ACP-imodel, alocalACPmodelofrole lAL can be 
derivedfrom Π^ , whichisdefinedas Π^ l 
¼ ðZl; Vl; Rl 
Þ where, 
 Zl 
¼ fCAZ j lAγðCÞg is alocalartifactschema,suchthat 
each sharedartifact in Zl contains astate s if andonlyif 
instateðsÞ appears inthepre-orpost-conditionofa 
business rulein Rl, i.e., 
8CAfCAZl 
j lAγðCÞ 4 γðCÞ  
 
41g; 8sAC:S; (rARl; 
sApre_sðr; CÞ [ post_sðr; CÞ; 
 Vl 
¼ fvAV j γðvÞ ¼ lg is asetoflocaltasks, 
 Rl 
¼ frAZ j γðrÞ ¼ lg is asetoflocalbusinessrules. 
Example 2. Fig. 3 showstheSupplier'slocalACPmodel 
which isderivedfromthe ACP-i model of thepurchasing 
processillustratedin Fig. 1. Wecanseethattheshared 
artifacts PO and SO representonlythepartsthatbelongto 
the Supplier'slocalprocess,e.g.,someprocessingstepsof 
PO (beforethe confirm state)and SO (after the in_transit 
state) thatbelongtoBuyerandLogistics,respectively,are 
not capturedintheSupplier’s localACPmodel. 
Next,wediscussthebehaviorpropertiesofartifact- 
centric inter-organizationalbusinessprocesses.Ingeneral, 
it isimportantthatthebehaviorofinter-organizational 
business processesmustbesoundinordertoguaranteethe 
reachabilityofdesiredgoalsofthecollaborationand 
participating localprocesses [83,58,42]. 
3.3. BehaviorpropertiesofACPmodelanditsartifacts 
Weclassifybehavioralpropertiesofartifactsin ACP-i 
model into intra-behavior and inter-behavior. Theintra- 
behaviorofanartifactdescribeshowanartifactchanges 
its statethroughoutitslifecycle.Here,weadoptLabel 
TransitionSystem(LTS)tocapturethelifecycleofan 
individualartifact.Second,theinter-behaviordescribes 
how thelifecycleofoneartifactdependsonthecounter- 
part ofanotherartifact,anditcanberepresentedas 
synchronizationdependencybetweenartifacts,i.e.,a sync 
rule. 
Here, wegeneralize ACP-i model Π^ to ACPmodel, 
denotedas Π¼ ðZ; V; RÞ, bydisregardingtherolesoforga- 
nizations androlemappingof Π^ . 
Definition 8. (Lifecycleofartifact, )n). Let 
Ci ¼ ðAi; sinit 
i ; Si; Sf 
i Þbe anartifactclassin ACPmodel Π. 
Picking List(PL) 
Purchase Order(PO) 
confirmed 
acquiring 
accepted filled 
canceled 
ready tofill Filled order 
checking In stock 
Out ofstock 
delivering 
ready toship 
Delivery Note(DN) 
prepared 
transferring 
dispatched 
closed billing 
Shipping Order(SO) 
In transit created scheduled 
Invoice (IV) 
cleared 
sent issued 
unpaid 
clearing 
Fig. 3. A localACPmodelforSupplier. 
S. Yongchareonetal./InformationSystems47(2015)51–81 57
A lifecycle of Ci, denotedas ℒCi , canbedefinedasatuple 
ðS; sinit ; )Þ where, 
 set ofstates S ¼ Si, initialstate sinit ¼ sinit 
i , 
 statetransitionrelation ) DS Ri Gi S where 
 RiD Π:R is asetbusinessrulesthatareusedtoinduce 
statetransitionsofartifact Ci such that, 
8rAΠ:R; (sx; syACi:S; sxApre_sðr; CiÞ 4sy 
Apost_sðr; CiÞ-rARi; 
 Gi (guards)isaunionsetofstatepreconditionsofeach 
business rulein Ri such thateachpreconditionrefer- 
ences toastateofotherartifactin Π, i.e., 
Gi¼ [jΠ:Zj j ¼ 1 fCj:sj(rARi; (CjAΠ:Z; sApre_sðr; CjÞ4CjaCig ; 
Wealsodenote )n for areflexivetransitiveclosureof 
). Wewrite si)nsj if state sj can bereachedfromstate 
si by somesequenceofbusinessrulesin Π:R. 
Wewritetransition ss) 
r½gst tomeanthatthestateofthe 
artifact willchangefromsourcestate ss totargetstate st if 
business rule r is firedandguard g (state pre-conditionof r) 
holds. Notethatinaclearcontext,wemayuseshorthand 
ss ) st without itssuperscription,andmayuseterm 
artifact for themeanof lifecycleofartifact. 
Based on Definition 8, given ACPmodelΠ, wecanderive 
a lifecyclecorrespondingtoanartifactin Π from asetof 
correspondingbusinessrulesthatareusedtotriggerthe 
state transitionsoftheartifact.Wecanobtainthelifecycle 
of anentireprocessbycomposingallartifactsinthemodel. 
Here, wedefine ACPlifecycle for describingthebehavioral 
aspect ofanACPmodelconsistingofsynchronizedlife- 
cyclesofartifacts.Weadaptastatemachinecomposition 
techniquepresentedin [49] for generatingthelifecycleof 
ACP.Similartechniqueforprocessmodelgenerationbased 
on (object)lifecyclecompositionisalsopresentin [46]. 
Definition 9. (Lifecyclecomposition,composedlifecycle, 
). Let ℒi ¼ ðSi; sinit 
i ; )iÞ, and ℒj ¼ ðSj; sinit 
j ;)jÞ be two 
artifact lifecyclesinACPmodel Π. Lifecyclecomposition 
(i.e., synchronizedproduct)of ℒi and ℒj is denotedas 
ℒc ¼ℒi  ℒj ¼ ðSc; sinit 
c ;)cÞ where, 
 ScDℒi:S ℒj:S is asetofcomposedstates, 
 sinit 
c ¼ ðℒi:sinit ;ℒj:sinit Þ is theinitialstate, 
 —)cD Sc Π:R Gc Sc is atransitionrelationwhere Gc is a 
set ofguards(statepropositions). 
Now,let g½sx 
ℒi=stateðℒi; sxÞ denotethatstate sx 
ℒi in 
guard g is substituted (denotedbysymbol/)by true or 
f alse (of statepredicate)dependingonwhetherthelocal 
stateof ℒi is sx. Wecanformulatetransitionrelation )c of 
composedlifecycle ℒc , byusingthefollowingthreeinfer- 
ence rules. 
ðsx 
ℒi ; r; g1; sy 
ℒi ÞA)i 
ððsx 
ℒi ; sx 
ℒj Þ; r; gc; ðsy 
ℒi ; sx 
ℒj ÞÞA)c; gc ¼ g1½sx 
ℒj=stateðℒj; sxÞ 
ð3:1Þ 
ðsx 
ℒj ; r; g2; sy 
ℒj ÞA)j 
ððsx 
ℒi ; sx 
ℒj Þ; r; gc; ðsx 
ℒi ; sy 
ℒj ÞÞA)c; gc ¼ g2½sx 
ℒi=stateðℒi; sxÞ 
ð3:2Þ 
ðsx 
ℒi ; r; g1; sy 
ℒi ÞA)i4ðsx 
ℒj ; r; g2; sy 
ℒj ÞA)j 
ððsx 
ℒi ; sx 
ℒj Þ; r; gc; ðsy 
ℒi ; sy 
ℒj ÞÞA)c; gc ¼ g1½sx 
ℒj=stateðℒj; sxÞ 4g2ðsx 
ℒi=stateðℒi; sxÞÞ 
ð3:3Þ 
Rule (3.1) and Rule (3.2) are appliedwhenbusinessrule r is 
firedononlyindividuallifecycle ℒi and ℒj, respectively. 
Rule (3.3) is appliedwhen sync ruler is firedonboth 
lifecycles ℒi and ℒj. Asthethreeinferencerulesapply 
the substitutionofstateconditionsoftwolifecyclesinthe 
composition, referencestoexternallifecyclearenot 
replaced. 
Example 3. Fig. 4 showsthecompositionbetweenthe 
lifecycleofartifact C1 and thelifecycleofartifact C2. The 
label ri½g attached toatransitionmeansthatthetransition 
is firedwhenboththeattributepropositioninthepre- 
condition ofbusinessrule ri holds andallstatepropositions 
(of externallifecycles)in g hold. Wedenotethecounter 
stateconditionof C:sx by symbol  C:sx in theguard.We 
can alsoseethatstateconditionsreferencingtoartifacts C3 
and C4 remaininthecomposedlifecyclebutindifferent 
forms, whichdependonthetransitiontheybelong. 
Now,wecandefinethelifecycleofACPbyusing lifecycle 
composition. 
Definition 10. (ACPlifecycle).GivenACPmodel Π, a (ACP) 
lifecycle of Π, denotedas ℒΠ, canbegeneratedbyitera- 
tivelyperforming lifecyclecomposition of everyartifactin 
Π. 
Notethat lifecyclecomposition is associativeandcom- 
mutative,i.e., ℒi  ℒj  ℒk ¼ ℒi  ðℒj  ℒkÞ ¼ ðℒi  
ℒjÞ  ℒk and ℒi  ℒj ¼ℒj  ℒi. Therefore,thefinal 
resultofthecompositionofasetoflifecyclesisnot 
impactedbytheircompositionorder. 
Next,wedefine soundness propertytodescribeadesired 
and correctbehaviorofartifactlifecycleandtheprocess. 
Definition 11. (Safe, goal-reachable,andsoundlifecycle). 
GivenACPmodel Π and lifecycle ℒ¼ ðS; sinit ; )Þ, we 
Fig. 4. An exampleoflifecyclecomposition(takenfrom [96]). 
58 S. Yongchareonetal./InformationSystems47(2015)51–81
define asetoflifecyclestates S ¼ ℒ:S [ fsinitg and asetof 
final states Sf S. Lifecycle ℒ is saidtobe: 
 safe iff thereexistsbusinessrule rAΠ:R such that r 
induces oneandonlyonetransitionin ℒ, i.e., 
Π:R; ðsx; r; g; syÞA 
) ðsm; r; g; snÞ=2 ) 
 goal-reachable iff, foreverynon-finalstate s of ℒ, s can 
be reachedfromtheinitialstateand s can reachoneof 
the finalstatesof ℒ, i.e., 
8sAS Sf ; (sf ASf ; sinit) n 
s s) n 
sf 
 sound iff ℒ is safe and goal-reachable. 
Notethatthe goal-reachability propertyimplies 
deadlock-freeandconnectedlifecycleofanartifact(ora 
process). 
Nowconsiderthecaseofasharedartifactdefinedina 
local ACPmodel.ItisalwaystruethatthelocalACPmodel 
is not goal-reachable as thelifecycleofsharedartifactis 
partially modeledandcanbenon-terminated.However, 
when integratingallthedifferentpartsofasharedartifact 
from eachorganization,thecompletelifecyclemustbe 
goal-reachable. 
In thisarticle,wedonotfocusonthetask-level 
information, i.e.,thespecificationoftaskisomitted.How- 
ever,thespecificationoftaskinartifact-centricprocess 
modeling approachcanbedefinedinthespiritofsematic 
web-servicesspecifiedinOWL-Sproposal[2003]—that isin 
a formofInput,Output,Pre-condition,andEffect(IOPE). 
The pre-conditionandconditionaleffectofataskshould 
conform tothepre-conditionandpost-condition,respec- 
tively,ofthebusinessrulethatspecifiesthetaskinits 
action. Adetaileddiscussionofsemanticweb-services 
togetherwiththeuseofbusinessrules(e.g.,SBVR [69]) 
for modelingartifact-centricbusinessprocesscanbefound 
in [5,19]. Wealsorestrictour soundness discussion onlyon 
the lifecyclebehaviorofartifactswhilediscussionsand 
formal approachestodataverificationofartifact-centric 
business processes(somecallartifactsystems)canbe 
found inseveralexistingliteratures,e.g., [54,31,29,25,21]. 
4. Behavior-consistentpublicviewconstruction 
In thissection,wefirstintroducethedefinitionofpublic 
view anddiscussthebehaviorconsistencybetweena 
createdviewanditsbase ACP-i model. Followingthat,we 
presentanabstractionmethodtoconstructpublicviews 
preservingthebehaviorconsistency,anddiscusshow 
synchronizationdependencybetweenartifactsisaffected 
by suchabstraction.Finally,weproposeatechniquefor 
constructing theminimalpublicviewforagivencollabora- 
tion andformalizeitintoanalgorithmofprotectingthe 
information (i.e.,localprocesses)ofeveryorganizationat 
the highestlevelofprivacyandautonomyduringview 
construction. 
4.1.Publicviewconstructionandbehaviorconsistency 
Generallyspeaking,constructingapublicviewofa 
particular collaborationshouldtakeintoaccountallinter- 
action betweenparticipatingorganizations,particularly 
the processingandexchangesofsharedartifacts.Aslife- 
cycleisthemainmechanismofanartifactforspecifying 
its dynamicbehaviors,andtheassociation/coordination 
betweenartifactsisalsobasedonthestatesinsidelife- 
cycles,wetakeitnecessarytoconstructpublicviewsbased 
on lifecyclesby blinding off the privatepartwhilepreser- 
ving theglobalcoordination.Weproposeourabstraction 
techniquetohelporganizationsrevealonlytheirnecessary 
steps thatarerequiredtocompletethecollaboration. 
Shared artifactsarethemainconcernastheyareusedby 
more thanoneorganization.Thisrequirementraisesthe 
questionofhowtodecidewhichprocessingpartofa 
shared artifactshouldbeabstractedsuchthatthe 
abstractedpartdoesnotaffectthecoordination.Basedon 
this requirement,weobservethatthepartofashared 
artifact interactingwithlocalartifact(s)andaprocessing 
step ofthesharedartifactownedbyasingleorganization 
should beabstracted.Asdiscussedin Section 3.1, thisis 
because suchpartisdeemedaslocalprocessingofthe 
shared artifactanditshouldnotberevealedtoother 
parties (outofinterestandprivacyconcern).Assuch,a 
constructedpublicviewshouldhavealllocalartifactsof 
every organizationhidden,andhaveallpartsofshared 
artifacts thatinteractwithlocalartifactsabstracted. 
Next,wediscussthe ACPabstractionmethod for con- 
structing publicviewsthatderivefromanunderlying ACP-i 
model. Themethodisdiscussedbasedonthefollowingtwo 
points: 
 Abstractionofnon-synchronizedpartoflifecycle. Itiseasy 
to understandthatapartofthelifecycle(calledlifecycle 
fragmentorfragment)isconvertedintoeitherasingle 
stateorasingletransitionduringthelifecycleabstrac- 
tion. In Section 4.2, wewillintroduceageneraltechni- 
queforstate/transitionabstractionthatcanbeapplied 
directlytoalifecyclefragmentofindividualartifact. 
 Abstractionofsynchronizedpartoflifecycle. Apartfrom 
the isolatedabstraction,wealsoconsiderhowmultiple 
synchronizedfragments(via sync rules) ofdifferent 
artifact lifecyclescanbeabstracted.Weobservethat 
the abstractionofonesynchronizedendcallsforthe 
abstractionofthecorrespondingendduetoconsistency 
preservation.Bothabstractedfragmentsoftwolife- 
cyclesmuststillsomehowbeeithercorrectlysynchro- 
nized ornone.Ifanentirelifecyclesynchronizingwitha 
fragmentofanotherlifecycleistobeabstracted,the 
former istotallyunsynchronizedandshouldbeconsid- 
ered asanembeddedfragmentoftheabstractedlife- 
cycleofthelatter. Section 4.3 discusses thisissuein 
more detail. 
Based onan ACP-i model of inter-organizationalpro- 
cesses, organizationscanconstructtheirpublicviewby 
abstractingtheirsharedartifactsandtheirlocalartifacts. 
S. Yongchareonetal./InformationSystems47(2015)51–81 59
Wedefineapublicviewbyapplyingthe abstractionfunc- 
tion to mapacomplete ACP-i model toitspublicview. 
Definition 12. (Public viewandACPabstraction).Given 
ACP-imodel Π^ ¼ ðZ; V; R; L; γÞ, paΠ^ ¼ ðZp; Vp; Rp; Lp; γpÞ denotesapublicviewof Π^ , where Zp; Vp; Rp; Lp; γp are 
sets ofabstractartifacts,abstractservices,abstractbusiness 
rules, organizationroles,androlemappingfunctionofthe 
public view,respectively. ACPabstractionfunction is defined 
as pa: Zp 
[ Vp 
[ Rp 
[ Lp 
[ γp-Z [ V [ R [ L [ γ 
which maps paΠ^ to Π^ . pa is atotalmappingfunctionwith 
the followingconditions: 
 paΠ^ and Π^ haveidenticalsetsoforganizationroles, 
 paΠ^ contains everyabstractsharedartifact,i.e., 
8CAZ; (CiAZp; γðCÞ  
 
41-ðpaðCiÞ 4 γpðCiÞ  
 
41Þ 
 each abstractsharedartifactin paΠ^ has anidenticalrole 
set toitscorrespondingconcreteartifactin Π^ , i.e., 
8CiAZp; (CAZ; paðCiÞ ¼ C - 8lAγpðCiÞ; lAγðCÞ 
 every ruleandtaskdefinedin Π^ must beprojectedto 
some abstractruleandabstracttask,respectively,defined 
in paΠ^ , i.e., 
8riAR; (rARp; paðrÞ ¼ ri 48viAV; (vAVp; paðvÞ ¼ vi 
 if anentirelocalartifactin Π^ is hiddenin paΠ^ then the 
followingholds: 
instate 
 rolemapping γp for abstractartifacts,tasks,andbusiness 
rules holds 
8CiAZp; γpðCiÞDL48viAVp; γpðviÞAL48riARp; γpðriÞAL 
Example 4. Fig. 5 illustratesanexampleofanagreed 
public viewthatcanbeconstructedfromthepurchasing 
processintroducedin Fig. 1. Wecanseethatthepublic 
view isachievedbyabstractingthreesharedartifacts PO, 
SO, and IV and removingalllocalartifactsofeachorganiza- 
tion. Itispossiblethat,duetoaimingtoachievehigher 
level ofabstraction,abstractstatesmaybeintroducedin 
the publicview,e.g., approving, supplying, unpaid are 
abstractstatesin Fig. 5. 
Notethelabelingofabstractstateisdonemanuallyafter 
lifecycleabstraction.Participatingorganizationsshould 
agree onchoosingareasonablenameofanabstractstate 
in theirpublicview.Onceapublicviewisconstructed 
based onitsunderlying ACP-i model, itisveryimportantto 
ensure thevalidityofthepublicview.Asalreadymen- 
tioned, inthisarticleweonlyfocusonthebehavior 
perspectiveofACPmodels,andthuswedefinethevalidity 
of publicviewsintermofthebehaviorconsistency 
betweenaviewanditsbasemodel.Inourrecentwork 
[96], wehaveproposedabehaviorconsistencychecking 
approachcalled B-consistency tocheckwhetheraspecia- 
lized processderivedfromthebaseprocessisconsistently 
observable.Here,weusethe B-consistency notion tocheck 
whether thebehaviorofapublicviewisconsistentwithits 
underlyingmodel. 
Let ℒy be thelifecycleofapublicviewconstructedby 
abstractingitsbase ACP-i lifecycle ℒx. Wearetocheck 
whether abstractlifecycle ℒy has consistentbehaviorswith 
its baselifecycle ℒx. Intuitively,ifeachfiringsequenceof 
transitionsin ℒx, disregardingstatesandtransitionsinan 
abstractedL-fragment,isobservableasthesamesequence 
as of ℒy, ℒy is saidtobe behavior consistent with ℒx. Our B-consistency 
relationbetweentwolifecyclesisdefinedwith 
the helpofbi-simulationequivalenceinprocessalgebra [9]. 
By abstractingalifecyclefragmentintoa silent (τ) action, 
we canapplyweakbi-simulationbetweenalifecycleand 
its abstractedone. 
Definition 13. (B-consistent, C). Let ℒx ¼ ðSx; sinit 
x ;)xÞ 
and r:be twolifecyclesand Sxy ¼ Sx  Sy be asetofstates 
that commonlyexistin ℒx and ℒy. ℒy and ℒx aresaidto 
be B-consistent (denotedas ℒyCℒx) iff, 
 8si; sjASxy; (ðsi; r; g; sjÞA)x; 8skASy Sxy; si)n 
ysk-sk)n 
ysj 
 8si; sjASxy; ∄ðsi; r; g; sjÞA)x; 8skASy Sxy; :ðsi)n 
ysk -sk)n 
ysjÞ 
Notethatasdefinitionof B-consistency is generalized,it 
can beusedforcheckinganytypeofalifecycle(i.e.,ACP 
lifecycles,artifactlifecycles,orcompositelifecycles). 
Example 5. The lifecyclein Fig. 6(a) isnot B-consistent 
with the lifecycleinFigure(b),i.e., . Thisisbecause, 
in somefiringsequencesoftransitionsinthelifecyclein 
Fig. 6(b), state a can reachstate c(throughstate x2) without 
passing state b; and,state a can reachitselfviastate x4 
without passingstate b. Incontrast,wecanseethat 
ℒaCℒc and ℒaCℒd in Fig. 6(c) and Fig. 6(d), respectively. 
4.2. Abstractionofnon-synchronizedlifecycle 
Next,weintroducenotionoflifecyclefragment(called 
L-fragment) forcapturingthepartofanartifactlifecycleto 
be abstractedinthepublicview.OnceanL-fragmentis 
identified inalifecycle,weapplytheabstractionfunction 
tomap L-fragment to aspecifiedabstractstateorabstract 
transition(s)inanabstractACPmodel(i.e.,apublicview). 
Definition 14. (LifecyclefragmentorL-fragment, Π^ ¼ 
ðZ; V; R; L; γÞ, find_Lf ). GivenACPmodel Π, L-fragment 
γðCÞ of lifecycle ℒx is anonemptyconnectedsub-lifecycleof 
ℒx. Itcanbedefinedas ℓℒx ¼ ðS;); )in; )out Þ where, 
Purchase Order(PO) 
closed delivering 
confirmed 
Shipping Order(SO) 
In transit arrived 
Invoice (IV) 
cleared unpaid 
billing 
approving 
canceled 
supplying filled 
issued 
Fig. 5. An exampleofanagreedpublicviewbasedontheprocessin 
Fig. 1. 
S. Yongchareonetal./InformationSystems47(2015)5160 –81
SDℒx:S fsinitg [ Sf is anon-emptysetofstatesof ℓℒx , 
where Sf is asetof final statesof ℒx, 
 ) DS RℒxGℒx SDℒx: ) is asetoftransitionsof ℓℒx , 
where Rℒx and Gℒx aresubsetsofbusinessrulesand 
guards,respectively,definedin ℒx, 
 )in ¼ℒx:) ððℒx:S SÞRℒx GℒxSÞÞ is asetof entry tran- 
sitions into ℓℒx , 
 )out ¼ℒx:) ðS RℒxGℒx ðℒx:S SÞÞ is asetof exit transi- 
tions from ℓℒx . 
Wedenote ℓ!ℒc (or ℒcgℓ) if ℓ is anL-fragmentof 
lifecycle ℒc of artifact C. Inaddition,wealsodefine 
function f ind_Lf ðC; SÞ to returnL-fragment ℓ if ℓ consists 
of asetofstates S of artifact C such that ℓ:S ¼ S and ℓ!ℒc; 
otherwise,return null if suchanL-fragmentcannotbe 
found. 
Noticethat,givenavalidinputsetofstatesofanartifact, 
thereisonlyonecasethat f ind_Lf returns null – that is 
when thesetcontainsonlyan init stateand final state(s)of 
that artifact. 
ToensureL-fragment ℓℒx is correctlyformedbya 
connectedsub-lifecycleofitsentirelifecycle ℒx, weconfine 
that foreachstate s in ℓℒx :S, thereexistsasequenceof 
transitionsfromanentrytransitionin )in to s and from s 
to anexittransitionin )out. soundness propertycanalso 
be appliedto L-fragment providingthatanyL-fragmentisa 
sub-lifecycleofitsentirelifecycle. 
Notethatbasedontheconditionofentryandexit 
transitionsofL-fragment,anentry/exittransitionmustbe 
fired from/toastateinsidetheL-fragment.However,there 
is norestrictiononthenumberofentrystatesandexit 
statesofL-fragment. 
Next,weidentifyaspecifictypeofL-fragmentbasedon 
its atomicitypropertywhichisrestrictedbymeansof 
single-entry-single-exit(SESE)fragmentoflifecycle.How- 
ever,weadaptittoourL-fragmentdefinitionbyallowing 
the structureofmultiple-entrytransitionsandmultiple- 
exittransitionsinsteadofsingleentryandsingleexitstates. 
Here, wedefinesuchL-fragmentas AtomicL-fragment. 
Definition 15. (AL-fragment,NAL-fragment).Given L-fragment 
ℓℒx ¼ ðS;); )in; )out Þ of lifecycle ℒx, ℓℒx is 
called an AL-fragment iff, allentrytransitionsin )in have 
the samesourcestateandallexittransitionsin )out have 
the sametargetstate.Otherwise, ℓℒx is classifiedas NAL- 
fragment (non-atomicL-fragment). 
Example 6. Fig. 7 showsexamplesofdifferenttypesofL- 
fragments.In Fig. 7(a), L-fragments ℓ1 and ℓ2 havesingle 
entry state s1 and singleexitstate s4; therefore,both ℓ1 and 
ℓ2 are AL-fragments.WecanseethatL-fragments ℓ3, ℓ4, 
and ℓ5 in Fig. 7(b) areNAL-fragmentsas ℓ3 has multiple 
exitstates,andboth ℓ4 and ℓ5 havemultipleexitstatesand 
multiple entrystates. 
Based onabovenotionofL-fragment,givenacomplete 
ACP-i model wecanusethesetwoabstractionmechanisms 
togetherwithL-fragmentstoconstructanabstract ACP-i 
model whichcanbeusedtorepresentthepublicviewofits 
base model.Next,wediscussthetwooutputtypesofan 
abstractionoperation: abstracttransition and abstractstate. 
In mostcases,anabstractionoperationshouldyieldan 
abstracttransitionasthetransitionrepresentsanatomic 
and uninterruptablestepfromonestatetoanotherstatein 
the lifecycle.However,insomecases,wemayseethatan 
abstractstateisyieldedinstead.Forinstance,considerthe 
PO artifact inthepurchasingprocessshownin Fig. 1 and 
compare itwithitspublicviewin Fig. 5, wecanseethatthe 
approving state inthepublicviewisanabstractstateofa 
lifecyclefragmentconsistingofstates created and on_hold. 
There aretwopossibleunderlyingreasonssupportingthis 
case. First,anabstractstateisspecifiedinthedesignofan 
organization orinthemutualagreementofthecollabora- 
tion. Second,afragmentcannotbeabstractedintoasingle 
abstracttransition—this isbecausethefragmentisnot 
atomic(i.e., NAL-fragment). Theresultofmultipleabstract 
transitionsforL-fragmentmakesitdifficulttodecideon 
drawingaprojectionfromapartofthefragmenttoan 
abstracttransition.Therefore,apossiblesolutionistouse 
an abstractstatetoremovetheambiguityofwhatis 
abstractedinsuchtransitions.Here,weshowdifferent 
cases ofL-fragmentabstractionin Fig. 8. 
Example 7. Fig. 8(a) and Fig. 8 (b) showtheabstractions 
for abstracttransitionofL-fragmentsinartifact A1 and 
artifact A2, respectively,while Fig. 8(c) showsacaseof 
abstractionforabstractstateofartifact A2. Wecanseethat 
both L-fragmentsin A1 and A2 are AL-fragments. However, 
consider theL-fragmentofartifact A3 in Fig. 8(d) whichis 
an NAL-fragment (due tomultipleexitstates s4 and s5), its 
abstractiondoesnotresultinasingleabstracttransition.As 
alreadydiscussed,multipleabstracttransitionsareambig- 
uous. Therefore,totacklethisissue,theabstractionneeds 
to resultinanabstractstateinstead,whichisshownin 
Fig. 8(e). Notethatwediscussthisabstractstateindetailin 
Definition 17 with anexamplein Fig. 9. 
Next,wedefine lifecycleabstractionmapping function to 
map anabstractelementinabstractlifecycleontoastateor 
transitioninthebaselifecycle,whichisshownin Definition 
a b 
c 
d a b 
c 
x d 
x x 
a b 
c 
d 
x x 
x 
x x 
x 
a 
c 
x x d 
x 
l1 
l2 
l3 
l4 
Fig. 6. Examples ofabstractlifecycle(a),itsbaselifecycles(b),(c)and(d)(takenfrom [96]). 
S. Yongchareonetal./InformationSystems47(2015)51–81 61
16. Thenin Definition 17, wedefinetwo abstractionfunc- 
tions that areusedtoconstructanabstractlifecycle.These 
functions arebasedonthepreferredoutputoftheabstrac- 
tion—that isabstractstateorabstracttransition. 
Definition 16. (Lifecycleabstraction(la)).Letartifactlife- 
cyclele ℒ0Cx ¼ ðS0; sinit ; )0Þ in anabstractACPmodel Π0be 
an abstractlifecycleofartifactlifecycle ℒCx ¼ ðS0; sinit ; )Þ 
in abaseACPmodel Π. Wedefine lifecycleabstraction 
mapping function laΠ0-Π 
ℒ0Cx 
-ℒCx 
: S0[ )0 -S[ ), where 
laΠ0-Π 
ℒ0Cx 
-ℒCx 
is atotalfunctionthatmapsanabstracttransition 
in )0 and anabstractstatein S0 onto astateanda 
transitionin ℒCx . Wewrite la without itssuperscription 
or itssubscriptionifacontextisclear. 
Definition 17. (Abstractionfunctionsforabstracttransi- 
tion (la_tran)andforabstractstate(la_state)).LetL- 
fragment ℓCx of artifactlifecycle ℒCx ¼ ðS; sinit ; )Þ to be 
abstracted.Wecanabstract ℓCx into asetof abstract 
transitions and an abstractstate (if applicable)in abstract 
lifecycle ℒ0Cx ¼ ðS0; sinit ; )0Þ via thetwofollowingabstrac- 
tion constructionfunctions: 
 function la_tranðℒCx ; ℓCx Þ returns ℒ0Cx by abstracting ℓCx 
into asingle abstracttransition, 
 function la_stateðℒCx ; ℓCx ; s0Þ returns ℒ0Cx by abstracting 
ℓCx intosingle abstractstates0 and corresponding 
abstracttransitionsrelatedto s0. 
Functions la_tran and la_state can beexpressedby 
lifecycleabstractionmappinglaℒ0Cx 
-ℒCx as follows. 
(a) Let ℓCx be an AL-fragment with entrystate si and exit 
state sj. ℓCx can beabstractedintoabstracttransition 
C1 
s1 
s4 
s2 s3 
C2 
s1 
s4 
s2 s3 
C4 
s1 
s4 
s2 
s6 
s5 
l1 l2 l4 
C3 
s1 
s4 s5 
s3 
l3 
C5 
s1 
s4 
s2 s3 
s6 
s5 
l5 
s2 s3 
Fig. 7. Examples ofL-fragmentsandNAL-fragments. 
A’2 
s1 
s4 
A2 
s1 
s4 
s2 s3 
A’’2 
s1 
s4 
A2 
s1 
s4 
s2 s3 sx 
A’’3 
s1 
s4 
A3 
s1 
s4 
s2 s3 sx 
s5 s5 
A’1 
s1 
s4 
A1 
s1 
s4 
s2 s3 
A’3 
s1 
s4 
A3 
s5 s5 
s3 
s1 
s4 
s2 
Fig. 8. Abstractionsforabstracttransitionsandabstractstates. 
A1 A’1 
s1 
s4 
sx 
s6 
s5 
A1 
s0 
s7 
A1 
s0 
s7 
OR sy 
s1 
s2 s3 
s5 
s0 
s7 
s4 s6 
A 
B 
Fig. 9. An abstractiononanexpandedNAL-fragment. 
S. Yongchareonetal./InformationSystems47(2015)5162 –81
si)0Sj by applyingfunction la_tran. Wehave 
ℒ0Cx ¼ la_tranðℒCx ; ℓCx Þ such that, 
 laðsiÞ¼s 
i 4 la s j ¼ sj, 
 8sAℓCx :S; laðsi )0 sjÞ ¼ s, 
 8)lAℓCx : ); laðsi )0 sjÞ¼)l. 
(b) 
Let ℓCx be an AL-fragment with entrystate si and exit 
state sj. ℓCx can beabstractedinto abstractstates0AS0 
with asetof abstracttransitions ) 0z by applying 
function la_state. Wehave ℒ0Cx ¼ la_stateðℒCx ; ℓCx ; s0Þ 
such that, 
 8sAℓCx :S; laðs0Þ ¼ s, 
 8)lAℓCx :) ðℓCx :)in [ ℓCx :)out Þ; laðs0Þ¼)l, 
 8)nAℓCx :)in; laðsi )0 s0Þ¼)n, 
 8)oAℓCx :)out ; laðs0 )0 sjÞ ¼)o. 
(c) 
Let ℓCx be an NAL-fragment with asetofentrystate Sen 
and exitstate Sex. ℓCx can beabstractedinto abstract 
state s0AS0 with asetof abstracttransitions ) 0z by 
applyingfunction la_state. Wehave ℒ0Cx ¼ la_stateðℒCx ; ℓCx ; s0Þ such that, 
 8sAℓCx :S; laðs0Þ ¼ s, 
 8)lAℓCx :) ðℓCx :)in [ ℓCx :)out Þ; laðs0Þ¼)l, 
 8smAℓCx :Sen; (snAℓCx :S; sm ) s0 ) 0z-laðsm ) s0Þ ¼ sm ) sn, 
 8soAℓCx :Sex; (spAℓCx :S; s0 ) so ) 0z-laðs0 ) soÞ ¼ sp ) so. 
It isworthmentioningthatanabstractbusinessrule, 
which firesanabstracttransition,containsno defined 
predicates.Ifabusinessruleinabaselifecycleisabstracted 
in anabstractlifecycle,thenany defined predicateinthe 
rule shouldberemovedfromtheabstractedruletoavoid 
overrestrictiononartifactattributesthatmayprohibit 
firing theabstracttransition. 
Example 8. Fig. 8(a) and Fig. 8(b) showtheabstract 
transitionsresultedfromapplyingfunction la_tran to AL- 
fragments,while Fig. 8(c) showstheabstractstategener- 
atedfromfunction la_tran on anAL-fragment.Ontheother 
hand, Fig. 8 (e) showstheabstractstateandrelatedabstract 
transitionsasanoutputofabstractionbyapplyingfunction 
la_state on aNAL-fragment. 
Next,wedefine B-consistent abstraction based onthe 
lifecycleabstractionmapping,andthenweshow,in 
Theorem 1, thatapplyingeitherfunction la_tran or func- 
tion la_state togenerateanabstractlifecyclefromitsbase 
lifecyclealwayspreserves B-consistency. 
Definition 18. (B-consistentabstraction, Cla). Letartifact 
lifecycle ℒ0Cx ¼ ðS0; sinit ; )0Þ in abstractACPmodelΠ0 be an 
abstractlifecycleofartifactlifecycle ℒCx ¼ ðS; sinit ; )Þ in 
ACPmodel Π based on lifecycleabstractionmapping func-tion 
laΠ0-Π 
ℒ0Cx 
-ℒCx 
. If ℒ0CxCℒCx , thenwesaythat ℒ0Cx is a B-consistent 
abstraction of ℒCx , denotedas ℒ0CxClaℒCx . 
Theorem1. (Non-synchronizedlifecycleb-consistent 
abstraction). Let lifecycle ℒ0Cx 
be anabstractlifecycleof ℒCx 
by abstractingL-fragment ℓCx where ℓCx!ℒCx . Given 
ℒ0Cx ¼ la_tranðℒCx ; ℓCx Þ, we have ℒ0CxClaℒCx . Correspond- 
ingly, ℒ0CxClaℒCx holds for ℒ0Cx ¼ la_stateðℒCx ; ℓCx ; s0Þ where 
s0Aℒ0Cx :S. 
Wecanprove Theorem1 by applying B-consistency 
checkingtothecomparisonbetweeninputlifecycleand 
output (abstract)lifecyclebasedonthethreeusecasesof 
those twoabstractionfunctions.Considerthefirsttwo 
cases (a)and(b)in Definition 17. Weknowthatabstracting 
an AL-fragment alwaysproducesasingleabstracttransition 
or abstractstate(withsingleentrytransitionandsingleexit 
transition)thatstillpreservestheatomicityofthefragment 
to beabstracted.Thus,theoutputlifecycleisconsistent 
with itsbase.Forthethirdcase(c)in Definition 17, we 
generateasingleabstractstatewiththerestrictiononits 
abstractentrytransitionsandexittransitions.Similarto 
case (b),theabstractstaterepresentstheinternalbehavior 
of theabstractedfragmentanditisatomic. 
Please notethat,in Definition 17(c), weuseabstract 
states toallowtheabstractionof NAL-fragments; however, 
this abstractionhastomeetcertainconditionsforthe 
abstractentryandexittransitionsoftheabstractstate. 
For instance,considertheabstractionin Fig. 8(e). Entry 
transition s1)sx abstractstwooriginalentrytransitionsof 
its fragment(s1)s2 and s1)s3) andexittransition sx)s4 
abstractstwooriginalexittransitions(s2)s4 and s3)s4). 
Nevertheless,wecanseethatexittransition sx)s5 should 
abstractonlyforoneexittransition s3)s5. Alternatively,if 
we donotwanttohaveanabstractstateastheresultof 
abstractionofNAL-fragmenttodecidetheappropriate 
trigger conditionsfortransitionsduetotheaboveaddition 
mechanism, thenwemayexpandtheNAL-fragmenttillit 
satisfies theconditionofAL-fragment(ifpossible).This 
requiresthemodelertofindapossibleAL-fragmentfrom 
the expanded-boundaryofNAL-fragment.Ifsuchan 
expandedfragmentcannotbefound,themodelermay 
decide toimplementanabstractstatefortheabstraction. 
Taketwoabstractions(A)and(B)in Fig. 9, forexamples.On 
one hand,abstraction(A)on NAL-fragment consisting of 
states s2 and s3 can resultinabstractstate sx and four 
abstracttransitions.Ontheotherhand,withalternative 
abstraction(B),theboundaryofthefragmentkeeps 
expandingtillitsatisfiestheconditionofAL-fragment. 
Wehavetheexpandedfragmentthatcoversallstatesand 
transitionsbetweenstates s0 and s7. Then,wecanabstract 
the fragmentintoeitherabstractiontransition s0)s7 or 
abstractstate sy. 
Next,wepresentanalgorithmtofindtheminimalAL- 
fragment ofaninputL-fragment.Thealgorithmgetsan 
inputfragmentandexpandsitsboundaryuntilthe 
expandedfragmentsatisfiestheconditionofAL-fragment 
(in Definition 15). TheoutputAL-fragmentisminimalasin 
an iterationofthealgorithm,itfindsthenearestsourceand 
target statesandaddsthemintothefragment(Lines8and 
10).IftheinputisalreadyaqualifiedAL-fragmentthenthe 
fragment itselfisreturned(Line11).Otherwise,ifthe 
function cannotfindavalidAL-fragment,thenreturns null. 
Algorithm1.(function find_minAL). Finding aminimal 
AL-fragmentfromanL-fragment 
Input: L-fragment ℓCx of artifact Cx in ACP-i model Π^ . 
Output: AL-fragment ℓ0Cx if found;otherwise null is returned. 
1. ℒyℓ0Cx ¼ ℓCx ; 
2. repeat 
3. Sen 
¼a setof entry states of ℓ0Cx ; 
4. Sex 
¼a setof exit states of ℓ0Cx ; 
5. if (jSen 
j 41 or jSex 
j 41) 
6. then 
7. if jSen 
j 41 then 
S. Yongchareonetal./InformationSystems47(2015)51–81 63
8. ℓ0Cx includes allstatesin Sen and theirrelated exit transitions; 
9. if jSex 
j 41 then 
10. ℓ0Cx includes allstatesin Sex and theirrelated entry 
transitions; 
11. else if (jSen 
j ¼ 1 and jSex 
j ¼ 1) then return ℓ0Cx ; 
12. until Sen 
¼ ∅ and Sex 
¼ ∅ 
13.return null; 
Based on Theorem 1 along withthehelpof f ind_minAL 
function forNAL-fragmentabstraction,wecanconstructa 
consistentpublicviewofanindividualartifactlifecycleby 
abstractingitsnon-synchronizedpartofthelifecycle.Next, 
we discusshowafragmentofalifecyclethatsynchronizes 
with afragmentofanotherlifecyclecanbeconsistently 
abstracted. 
4.3. AbstractionofsynchronizedL-fragments 
In thissection,wediscusshowthesynchronization(via 
sync rules) betweentwolifecyclescanbeabstractedin 
detail. Technically,weneedtoanswerthefollowing 
questions. 
 What istheresultofabstractinganL-fragment(ora 
whole) ofonelifecyclesynchronizedwithanL-fragment 
of anotherlifecycle? 
 What istheconditionthatmakesanabstractsynchro- 
nization consistentwithitsoriginalsynchronization? 
Here, weusethesameL-fragmentabstractionmethod 
to abstractthesynchronizedpartswiththeconsideration 
of synchronizationstructureandbehavior.However,in 
order tocapturesynchronizationdependenciesbetween 
lifecycles,weneedtoextendthedefinitionofL-fragment 
for asynchronizedregion(called S-region) whichrepre- 
sents synchronizedL-fragmentsbetweenlifecycles(called 
SL-fragments). Withatargetfragmentofanartifactandthe 
sync rulesusedwithinthefragment,wecanidentifya 
counter-synchronizedpart(s)oftheotherartifact(s)thatit 
interactswith. 
Definition 19. (Sync ruleforsynchronizedL-fragments, φ). 
GivenACPmodel Π, asetofsyncrulesthatisusedwithin 
two synchronizedL-fragments ℓx and ℓy can bedefinedas 
follow: 
φðℓx; ℓyÞ ¼fr 
AΠ:Rj (ðsi; r; g; sjÞA ℓx: 
) 4(ðsm; r; g; snÞAℓy: )g 
It isnotedthatasyncruleistransitive,i.e., 
(rAφðℓx; ℓyÞ  φðℓy; ℓzÞrAφðℓx; ℓzÞ. 
Definition 20. (SL-fragmentandS-region).GivenACP 
model Π, wedenote ω¼ ðΓ; Rsync) fora synchronization 
region (S-region) where, 
 a setofsynchronizedL-fragments Γ¼ fℓC1 ; ℓC2 ; …; 
ℓCx g, ℓCi AΓð1rirxÞ is asynchronizedL-fragment, 
called as SL-fragment, of artifactlifecycle ℒCi ðCiAΠ:ZÞ, 
 !RsyncDΠ:R is asetof sync rules that isusedto 
synchronizetransitionsbetweenL-fragmentsin Γ such 
that, 
8rAΠ:R; (ℓCi ; ℓCj AΓ; 8rAφðℓCi ; ℓCj Þ; rARsync 
Example 9. In Fig. 10(a), wehaveS-region ωa with SL- 
fragments l1 of artifact A1 synchronizedwith SL-fragmentl2 
of artifact A2 via syncrules r1 and r2. In Fig. 10 (b), S-region 
ωb has two SL-fragmentsl3 and l4 with syncrules 
Rsync 
¼ fr1; r2; r3g. Noticethatsyncrule r4 is excludedfrom 
ωb as itisnotusedforthesynchronizationbetween l3 and 
l4. 
Next,westudytheatomicitypropertyofS-regionby 
determiningthecomposabilityofcontainedSL-fragments 
and theboundnessoftheirsynchronizationbehavior. 
4.3.1.AtomicityofS-region 
Here, weproposeafragmentalcompositiontechnique 
tochecktheatomicityofS-region.Firstweseethatthe 
composition betweentwosynchronizedL-fragments 
resultsinacompositeL-fragment.Thenwecanapply 
atomicitycheckingtothecompositeL-fragment.Assuch, 
we needtoobservetheconditionsforSL-fragmentsthat 
make thecompositeL-fragmentatomic.Now,wedefine 
compositeS-region based on lifecyclecomposition (in 
Definition 9). 
Definition 21. (CompositeS-region).GivenACPmodel Π, 
let S-region ω¼ ðfℓCx ; ℓCy g; Rsync) ofL-fragment ℓCx of artifact 
CxAΠ:Z and L-fragment ℓCy of artifact CyAΠ:Z where ℓCx 
and ℓCy aresynchronizedviabusinessrules RsyncRsync. The 
compositeS-region of ω, ω¼ ℓCx  ℓCy , istuple ðS;); 
)in;)out Þ where eachsetelementin ω has thesame 
definition correspondingtotheelementofL-fragment,i.e., 
inω can beconsideredasanL-fragment(orSL-fragment 
if thecompositefragmentstillhassomesynchronizationto 
otherfragmentofdifferentartifact). 
It isnotedthata compositeS-region is consideredasa 
sub-lifecycleofthecompositionbetweentwoentirelife- 
cycles.Tohavea(minimalandsufficient)completeviewof 
the compositionwedrawadashedarrowforatransition 
betweenacompositestateexcludedfromtheS-regionand 
a compositestatethatisanentryorexitstateoftheS- 
region,asexemplifiedin Fig. 11. Fig. 11 (a) and(b)showthe 
resultsofSL-fragmentcomposition, compositeS-regions 
ωa and ωb, forS-regions ωa and ωb in Fig. 10 (a) and 
in Fig. 10 (b), respectively. 
Example 10. In Fig. 11, ωa has compositestate(s2; s5), 
and (s4; s7) asitsentrystateandexitstate,respectively. 
Likewise, ωb has twoentrystates(s2; s5) and(s9; s6), and 
twoexitstates(s4; s7) and(s4; s8). Notethatcomposite 
state(s1; s5) isoutofscopeof ωa and ωb, soitisexcluded 
from ωa and ωb, respectively. 
Next,wevalidateatomicitypropertyof S-region by 
checkingwhetherSL-fragmentsoftheS-regioncanbe 
composed intoanatomiccompositeS-region.Weconsider 
the propertyof AL-fragment to define atomicS-region (AS-region), 
i.e., AS-region must haveallentrytransitionfired 
fromthesame(composite)sourcestateandallexittransi- 
tions firedtothesame(composite)targetstate. 
64 S. Yongchareonetal./InformationSystems47(2015)51–81
Definition 22. (AS-regionandASL-fragment).GivenACP 
model Π, let S-region ω¼ ðΓ; Rsync) and ZΓDΠ:Z be asetof 
artifacts whereofwhichhasitsL-fragmentdefinedin Γ. 
The compositionofallSL-fragmentsin Γ satisfies the 
propertyof AL-fragment iff, forevery ℓCi AΓ, 
 ℓCi is an AL-fragment, 
 8ℓCx ; ℓCy AΓ; 8rAφðℒCx ;ℒCy Þ; (ssAℓCi :S; ss)rsAℒCi : 
)-ss) r 
sAℓCi : ) 4rARsync 
 8ℓCx ; ℓCy AΓ; 8rAφðℒCx ;ℒCy Þ; (stAℓCi :S; s)rstAℒCi : 
)-s) r 
stAℓCi : ) 4rARsync 
 8ℒCj ðCjAΠ:Z ZΓ 
Þ;φðℓCi ;ℒCjÞ ¼ ∅. 
By holdingaboveconditions, ω can beconsideredasan 
AS-region. WealsocalleachL-fragmentin Γ as ASL- 
fragment of ω. 
Notethattheconditionsin Definition 22 are usedto 
restricttwoSL-fragments(tobecomposedforS-region)to 
include everytransitionandcorrespondingsyncrulethat 
areusedforonlythesynchronizationbetweenL-fragments 
in Γ. IfanS-regioncontainsmorethantwoSL-fragments, 
then, likelifecyclecompositionforACPmodel,wecan 
check theAS-regionbyperformingiterativecomposition 
for eachSL-fragmentinthatS-region.Itisalsoworth 
mentioning thatthecompositionofSL-fragmentsforAS- 
region holdsassamecharacteristicsasforthelifecycle 
composition – that iscommutativeandassociative,i.e., 
ℓCx  ℓCy ¼ ℓCy  ℓCx and ℓCx  ℓCy  ℓCz ¼ ℓCx  ðℓCy  
ℓCzÞ ¼ðℓCx  ℓCyÞ  ℓCz . 
Example 11. In Fig. 10 (a), ωa is anAS-regionasboth 
L-fragments l1 and l2 areAL-fragmentswithallrelatedsync 
rules (r1 and r2) residingwithinthem.Assuch,the 
resultingfragmentfromthecompositionbetween l1 and 
l2 is thenatomic,asshownin Fig. 11 (a). Incontrast,in 
Fig. 10 (b), wecanseethatL-fragment l4 cannotsatisfythe 
propertyofAL-fragment,andL-fragment l3 does not 
include transition s5)r4 s6 wheresyncrule r4 exitsinentry 
transition s5)r4 s6 of l4. Therefore, ωb cannotbeconsidered 
as AS-region. 
Example 12. Now,weillustratethecasethatanS-region 
contains morethantwoSL-fragmentswhereoneofwhich 
synchronizedonthe(nested)sub-fragmentofSL-fragment, 
as shownin Fig. 12. AssumeS-region ω1 for SL-fragments 
{l1; l2; l3}, ω1 cannotbeconsideredasanAS-regionas l1 has 
some syncrulesthatareusedforthesynchronization 
betweenitssubL-fragment l5 and L-fragment l4 of the 
lifecycleofartifact A4. Therefore,weneedtoinclude l1 into 
the S-regioninordertosatisfythepropertyofAS-region. 
In thenextsection,wediscussabouthow AS-region can 
be usedfortheabstractionofsynchronization. 
A1 A2 
s1 
s2 s4 
s3 s5 s7 
r1 
s6 
r2 
l1 A1 A2 
s1 
s2 s4 
s3 s5 s7 
r1 
s6 
r2 
s9 
s8 
l2 
l3 
l4 
a b 
r4 
r3 
ω ω 
Fig. 10. An exampleofS-regionsandSL-fragments. 
Fig. 11. Composite S-regionsofSL-fragmentsbasedon Fig. 10. 
A1 
s1 
A3 A2 
A4 
s2 
s1 
s3 
s2 
s1 s2 s4 s3 
l4 
s3 s5 
s4 
s6 
l1 
s2 
s3 
s1 
l2 
l3 l5 
Fig. 12. An S-regionwithsubSL-fragments. 
S. Yongchareonetal./InformationSystems47(2015)51–81 65
4.3.2. SL-fragmentabstraction 
Here, wedefinetheabstractionrelationofthesynchro- 
nization betweentwoabstractartifactsintheabstractACP 
model andthesynchronizationbetweentwoartifactsinits 
base ACPmodel. 
Definition 23. (Synchronization(Sync)abstraction).Let 
artifact lifecycles ℒ0Cx and ℒ0Cy in ACPmodel Π0 abstract 
artifact lifecycles ℒ0Cx and ℒ0Cy in baseACPmodel Π with 
lifecycleabstractionmappings laℒ0Cx 
-ℒCx and laℒ0Cy-ℒCy , 
respectively.Wecandefine sync abstractionmapping func-tion 
saΠ0- Π 
ðℒ0Cx ;ℒ0Cy ÞðℒCx ;ℒCy Þ 
:φðℒ0Cx ;ℒ0Cy ÞφðℒCx ;ℒCy Þ that isused 
to projectthe abstractsyncrule for ℒ0Cx and ℒ0Cy onto its 
base syncrulefor ℒ0Cx and ℒ0Cy . 
Notethat saΠ0 - Π 
ðℒ0Cx ;ℒ0Cy Þ ðℒCx ;ℒCy Þ 
is atotalfunctionandwe 
may write sa without itssuperscriptionandsubscriptionin 
a clearcontext. 
Now,wewanttoperformanabstractiononanL- 
fragment thatsynchronizeswithotherL-fragment(s).Simi- 
lar toabstractionfunctionfornon-synchronizedlifecycle, 
we define sync abstractionfunction for abstracting AS-region 
that containssynchronizedL-fragments. 
Definition 24. (Sync abstractionfunction(sa_f)).Given 
ACPmodel Π, letAS-region ω¼ ðΓ; Rsync 
Þ to beabstractedin 
abstractACPmodel Π0, where Γ ¼ fℓC1 ; ℓC2 ; …; ℓCx g and 
ℓCi ð1rirxÞAΓ is anL-fragmentoflifecycle ℒCi in Π. We 
define syncabstraction function sa_f ðΠ;ωÞ to return Π0 
with asetof abstractartifactlifecycles L ¼ fℒ0C1 ;ℒ0C2 ; ::;ℒ0Cx g 
and an abstractsyncruler0, where ℒ0Ci ð1rirxÞAL is a 
lifecycleofartifact CiAΠ0:Z such that, 
 for every ℓCi AΓ, ℓCi is abstractedinto abstracttransi- 
tions in ℒ0Ci , i.e., 
ℒ0Ci ¼ la_tranðℒCi ; ℓCi Þ; 
 for everysyncrule rAΠ:R that isusedtosynchronize 
betweenanytwoL-fragmentsin Γ, r is abstractedinto 
r0AΠ0:R, i.e., 
8ℓCx ; ℓCy AΓ; 8rAφðℓCx ; ℓCy Þ; (!r0Aφðℒ0Cx ;ℒ0Cy Þ; 
sa 
ðℒ0Cx ;ℒ0Cy Þ ðℒCx ;ℒCy Þðr0Þ ¼ r 
Abstractsyncrule r0 that isusedtosynchronizeall 
abstracttransitionstogethercanbedefinedasfollow.For 
everyL-fragment ℓCi ¼ ðS; );)in;)out ÞAΓ, wehave, 
(ss; stAℒCi :S S;ðss)in) n 
)outst Þ 
 ssApre_sðr0; CiÞ4stApost_sðr0; CiÞ 
Example 13. In Fig. 13 (a), wecanseethatAS-region ωa 
contains twofragments l1 and l2 with syncrules r1 and r2. 
When applying sa_f ðΠ;ωa), wehaveabstractlifecycleof 
ℒA1 and abstractlifecycleof ℒA2 with abstractsyncrule r0. 
In addition, Fig. 13 (b) showsacaseofsyncabstractionfor 
AS-region ωb which containsmultipleentrytransitions 
and multipleexittransitionsASL-fragments(bothASL- 
fragments l3 and l4). 
In theaboveexample,wedemonstratethesyncabstrac- 
tion oftwosynchronizedfragments.However,itispossible 
that anAS-regioncontainsmorethantwoASL-fragments. 
Forwiderunderstandingofsyncabstraction,weillustratea 
sync abstractionofmorethantwoSL-fragmentsin Fig. 14 
(with artifact A3 extendedtotheexamplein Fig. 13 (b)). 
Example 14. In Fig. 14, wecanseethatAS-region ωb 
contains synchronizedfragments l3 and l4 of artifact A1 and 
A2, respectively,and l5 of artifact l4. Asallthreefragments 
can beconsideredasASL-fragmentsin ωb, wecanvalidly 
apply function sa_f ðΠ;ωb) andtheabstractlifecyclesof 
artifacts fA01; A02; A03g with abstractsyncrules fr0; r″g are 
returned. 
A1 A2 
s1 
s2 s4 
s3 s5 s7 
r1 
s6 
r2 
l1 
l2 
a A’1 A’2 
s1 
s2 s4 
s5 
s7 
r' 
A1 A2 
s2 s4 
s3 s5 s7 
r1 
s6 
r2 
s9 
s8 
l3 
l4 
b r4 
r3 
s1 
A’1 A’2 
s2 s4 
s5 
s8 
r' 
s1 
ω 
ω 
Fig. 13. Examples ofsyncabstraction. 
S. Yongchareonetal./InformationSystems47(2015)5166 –81
Next,wediscussthecaseofanAS-regioncontainingSL- 
fragmentswithanested(sub)SL-fragmentthatsynchro- 
nizes withotherlifecycle.Intuitively,thesubSL-fragment 
and itssynchronizedlifecycleshouldbealsotakeninto 
account whenitssuperfragmenthastobeabstracted. 
Therefore,weneedtoinducetheabstractiontoitssub 
fragmenttogetherwithitscounterpartiftheybothcan 
satisfy thepropertyofAS-region(whichisconsideredasa 
sub AS-regionofthewhole).Inotherwords,wecansay 
that theentireAS-regionshouldcontainsuchcounterpart 
in ordertohaveavalidabstraction.Weshowanexampleof 
AS-regionconsistingofsubSL-fragmentin Fig. 15. 
Example 15. In Fig. 14, wecanseethatL-fragment l4 is a 
synchronizedfragmentofL-fragment l5 which isnested 
under L-fragment l1. Theabstractionyieldsabstracttransi- 
tions withabstractsyncrule r0z- thatis s1)r0z s3 in the 
lifecycleof A04 and s1)r0z s6 in thelifecycleof A01. 
Next,weconsiderthecasethatthesyncabstraction 
performs onanL-fragmentofoneartifactthatsynchronizes 
with theentirelifecycleofoneanotherartifact.Suchwhole 
lifecyclecanbeconsideredasfully-embeddedexternal 
lifecycleintheL-fragment.Wesaythatlifecycle ℒCj can 
be fully-embeddedinlifecycle ℒCi if thereexists 
L-fragment ℓCi!ℒCi that completelysynchronizesthe 
entire lifecycle ℒCj , i.e.,theentrystateandtheexitstate 
(s) of ℓCj are the init state andallthe final state(s)of ℒCj , 
respectively.Itcanbeunderstoodthattheabstractlifecycle 
of artifact Cj containing onlyabstracttransition(s)firing 
from its init statetoits final state(s) impliesasinglestep 
lifecycleandtheredoesnotexistsanyL-fragmentwithinit, 
i.e., f ind_Lf ðCj; Cj:SÞ ¼ null. Thebenefitofintroducingthis 
type oflifecycleabstractionisdiscussedlaterwhentaking 
into accounttheconstructionofpublicview.Itisnotedthat 
a lifecycleofoneartifactcanbeconsideredasfully- 
embedded externallifecycleofothermultiplelifecycles. 
Example 16. In Fig. 16 (a), L-fragment l1 synchronizeswith 
L-fragment l2 which representsthewholelifecycleof 
artifact A2. Therefore,wehave ℒA2 as anfully-embedded 
lifecycleof ℒA1 . Similarly,in Fig. 16 (b), wehave ℒA3 as a 
fully-embeddedlifecycleofboth ℒA1and ℒA2 . 
Next,weshowthat sync abstractionfunctionsa_f pre-serves 
B-consistency of betweentwosynchronizedlife- 
cyclesoftheinputandtwooutputtedabstractlifecycles. 
Theorem2. (synchronizedfragmentsb-consistentabstrac- 
tion). Let ω¼ ðΓ; Rsync 
Þ be anAS-regioninACPmodel Π that 
is tobeabstracted in ACPmodel Π0 and let L be asetof 
A3 A1 A2 
s10 
s11 s2 s4 
s3 s5 s7 
r1 
s6 
r2 
s9 
s8 
l3 l4 
b 
r4 
s1 
r8 
r r3 7 
s13 
s14 
r6 
A’1 A’2 
s2 s4 
s5 
s8 
r' 
s1 
s12 
r5 l5 
A’3 
s10 
s11 s14 
r'’ 
r8 
ω 
Fig. 14. An exampleofanabstractionoftwoormoreASL-fragments. 
A1 
s1 
A3 A2 
A4 
A’1 
s1 s6 
A’3 
s1 
s3 s4 
s2 
s1 
s3 
s2 
s1 s2 s4 s3 
l4 
A’4 
s1 s3 
r'z 
s3 s5 
s4 
s6 
l1 
s2 
s3 
s1 
l2 
l3 
A’2 
s1 
s3 
r'x r'y 
l5 
Fig. 15. An exampleofanabstractiononnestedsub-SL-fragments(cf. Fig. 12). 
S. Yongchareonetal./InformationSystems47(2015)51–81 67
abstractartifactlifecyclesresultedfromapplyingsync 
abstractionfunctionsa_f ðΠ;ωÞ. Then thefollowingstatement 
holds. 
8ℒCi ;ℒCj Afℒjℓ!ℒ4ℓAΓg; (ℒ0Ci ;ℒ0Cj AL; 
ℒ0CiClaℒCi 
4ℒ0CjClaℒj-ℒ0Ci  ℒ0CjCℒCi  ℒCi 
Wecanprove Theorem 2 by using Definition 21 and 
Definition 22for the B-consistency checkingbetweentwo 
abstractlifecyclesandtheirbaselifecycles,andbetween 
the compositionbetweentheformerandthecomposition 
betweenthelatter.From Definition 24, syncabstraction 
sa_f alwaysreturnsasingleabstracttransitionineach 
abstractlifecycleandasinglesyncrulebetweentwo 
abstracttransitions,sothecompositionofsuchabstract 
transitionsalwaysyieldsanatomiccompositefragment. 
The compositionofASL-fragmentsintheirbaselifecycleis 
atomic andcanentirelyberepresentedbythecomposition 
of abstracttransitions. 
4.3.3. FindingminimalAS-region 
Next,wediscusssimilarrequirementsfortheneedof 
NAL-fragmentexpansion(toAL-fragment).GivenanL- 
fragment ofoneartifactlifecyclethatsynchronizeswith 
otherartifactlifecycle(s),wecanfindtheminimalAS- 
region (whichcontainsthatL-fragment)anditsminimal 
counterpart(s)inwhichtheycanbeusedasinputsfor sync 
abstractionfunctionsa_f . Here,weproposetwoalgorithms 
to findtheminimalAS-regionofanL-fragment.Notethat 
the synchronizedcounterpartscanbemanyastheL- 
fragment cansynchronizewithmultiplelifecycles.Two 
proposed algorithms f ind_minASR and f ind_SR areusedto 
expandtheboundaryofanL-fragmentalongwithits 
synchronizedcounterpart(s)untilallsatisfythecondition 
of AS-regionandASL-fragment(in Definition 22). IfnoAS- 
region canbeconstructedoranL-fragmenthasnosyn- 
chronizedcounterpartofanyotherlifecycle,thenthe null 
valueisreturned. 
Algorithm2.(function find_minASR). Finding aminimal 
AS-regionfromanL-fragment 
Input: L-fragment ℓCx in ACP model Π. 
Output: AS-region ω¼ ðΓ; Rsync 
Þ if found;otherwise null is 
returned. 
1. ω’ðfℓCx g; ∅); 
2. ω¼ find_SRðωÞ; 
3. return ω; 
Algorithm3.(function find_SR). Finding anexpanded 
S-regionfromaninputtedS-region 
Input: S-region ω¼ ðΓ; Rsync 
Þ in ACPmodel Π. 
Output: ExpandedS-regionof ω or null if anySL-fragmentin ω 
cannot satisfythepropertyofAL-fragment. 
1. for each ℓiAω:Γ do 
2. L-fragment ℓ0i ¼ find_minALðℓi); 
3. if (ℓ0i ¼ nullÞ then return null; 
4. else 
5. ω:Γ ¼ ω:Γ [ fℓ0ig fℓig; 
6. end if 
7. for each ℒjAfℒCj j CjAΠ^ :Zg do 
8. states Sj ¼ ∅; 
9. for each rkAφðℓ0i ;ℒjÞ do 
10. Sj’pre_sðrk; CjÞ [ post_sðrk; CjÞ; 
11. if (∄rkAω:Rsync 
Þ then rkAω:Rsync’rk; 
12.endfor 
13. L-fragment ℓj ¼ f ind_Lf ðCj ; SjÞ; 
14. L-fragment ℓ0j ¼ find_minALðℓj); 
15. if(ℓ0janullÞ 
16. then 
17. if ð∄ℓ0jAω:ΓÞ 
18. then 
19. ω:Γ ¼ ω:Γ [ fℓ0jg; 
20. ω¼find_SRðω Þ; 
21. return ω; 
22. end if 
23. else return null; 
24. end if 
25. end for 
26. end for 
Here, weusearunningexampleillustratedin Fig. 17 to 
explainthealgorithmsoffunctions f ind_minASR and 
f ind_SR. 
Webeginwith Fig. 17 (a) byapplying f ind_minASRðl1Þ in 
this example.FirstittakesL-fragment l1 of artifact A2 as an 
inputandinitializesanS-regionfrom l1 and anemptysetof 
sync rule.ThenitfindsthepossiblecorrespondingS-region 
for l1 (bycalling f ind_SR function). Function f ind_SR starts 
checkingwhether l1 can satisfythepropertyofAL- 
fragment(Line3).Ifso,thenitcontinuessearchingforall 
sync rulesandtheirrelatedsynchronizedstatesthatare 
used tosynchronize l1 with anyotherlifecycletransitions; 
otherwise,itimmediatelyreturns null. Intheiterationof 
finding synchronizedpartofotherartifact,ifasetof 
synchronizedstatesisfound(Lines9–12),thenafragment 
consisting ofsuchsetisconstructedbycalling find_Lf 
function (definedin Definition 14) – that isL-fragment l2 
of artifact A1 shown in Fig. 17 (b). Next, l2 is tobechecked 
whether itisqualifiedasAL-fragment.Ifitsatisfies,thenit 
is addedintotheS-region(Lines13-21);otherwise,ithas 
tobeexpandeduntilitcansatisfyAL-fragment.Once 
qualified,itisaddedintotheS-region(Line19).After l2 is 
added totheS-region,werecursivelycallfunction f ind_SR 
again withtheexpandedS-regionasaninput(Line20). 
Fig. 17 (c) showsAL-fragment l02 that isexpandedfrom l2. 
Wecanseethat l02 introducesnewsyncrules fr5; r6g that 
arenotincludedintheS-region.Therecursionwillcon- 
tinue untilalltheSL-fragmentsintheS-regionareAL- 
fragmentsandincludeallthepossiblesyncrulesthatare 
used intheS-region.Consequently, l1 is requiredtoexpand 
itself andbecomesnewAL-fragment l01, asshownin Fig. 17 
(d). Afterthat,wecanseenewsyncrules fr7; r8g appear in 
l01. TheS-regionneedsanotherexpansionagaintocover 
those syncrules – that isL-fragment l3 of artifact A3. The 
exitconditionoftherecursioniswhenthefragmentisnull 
(Line 15withtheexitonLine23).Thefinaloutputofthe 
f ind_SR function isanS-regionthatallthesyncrulesare 
discoveredandthatcontainsonlyAL-fragments.Finally,by 
completing theiterationandrecursionwithallthecondi- 
tions ofAS-regionsatisfied,thefunctionreturnsthemini- 
mal AS-regionconsistingofASL-fragments fl01; l02; l3g as the 
output fromtheprovidedinputL-fragment l1. IftheAS- 
regioncannotbeconstructedthenthefunctionreturns 
S. Yongchareonetal./InformationSystems47(2015)5168 –81
A1 A2 
sa 
s1 s3 
sa 
s1 
r1 
s2 
s6 
s5 
s4 
s6 
r4 s5 
s2 
r3 
r2 
l1 
s3 r5 
A1 A2 
sa 
s1 s3 
s2 
s6 
s5 
s4 
sa 
s1 
s6 
s5 
s2 
r1 
r4 
r3 
r2 
s4 
r6 
s3 r5 
s4 
r6 
l'2 l1 
A1 A2 
sa 
s1 s3 
s2 
s6 
s5 
s4 
sa 
s1 
s6 
s5 
s2 
r1 
r4 
r3 
r2 
s3 r5 
s4 
r6 
l'2 
A1 A2 
sa 
s1 s3 
sa 
s1 
l2 r1 
s2 
s6 
s5 
s4 
s6 
r4 s5 
s2 
r3 
r2 
l1 
s3 r5 
s4 
r6 
A3 
s1 
s2 
s3 
r7 
r8 
A3 
s1 
s2 
s3 
r7 
r8 
A3 
s1 
s2 
s3 
r7 
r8 
A3 
s1 
s2 
s3 
r7 
r8 
l'1 
A1 A2 
sa 
s1 s3 
s2 
s6 
s5 
s4 
sa 
s1 
s6 
s5 
s2 
r1 
r4 
r3 
r2 
s3 r5 
s4 
r6 
l'2 
A3 l'1 
s1 
s2 
s3 
r7 
r8 
l3 
Fig. 17. A runningexampleforthe f ind_minASR function. 
A1 A3 A2 
s1 
sf s4 
A’2 
s1 
s3 
s4 
A’1 
s1 
s4 
A2 
A’1 
s1 
s3 
A1 
s1 
s3 
s2 
s1 
s4 
s2 s3 
s1 
s3 
s2 
l1 
l1 l2 
sf 
A2 
sf 
r' 
s1 
s2 
r2 
r1 
l2 
A’3 
sf 
Fig. 16. An exampleoffully-embeddedexternallifecycles. 
S. Yongchareonetal./InformationSystems47(2015)51–81 69
false. WecansaytheresultedAS-regionisminimalasthe 
function usethe f ind_minAL function toguaranteethe 
minimal AL-fragmentexpansionintheAS-region. 
Based on Theorem 2 and theuseofthe f ind_minASR and 
f ind_SR functions, wecanconstructaconsistentpublic 
view ofsynchronizedlifecyclesbyabstractingtheirsyn- 
chronizedpartsofthelifecycles. 
4.4. Consistentpublicviewconstruction 
Now,weuse Theorems1and2 to formulatethe B-consistency 
for theentireartifact-centricinter-organiza- 
tional businessprocess. 
Theorem3. (b-consistent publicview). Let paΠ^ be apublic 
view of ACP-imodel Π^ constructed byapplyingsyncabstrac- 
tion functionsa_f and lifecycleabstractionfunctions la_tran 
and sa_state. Then paΠ^ is aB-consistentabstraction of Π^ , i.e., 
ℒpa ^ ΠClaℒΠ^ . 
Theorem 3isderivedfrom Theorems1and2. 
Recallthatinthepublicview,alllocalartifactsshouldbe 
invisible.Onlyabstractsharedartifactsarerevealedforthe 
collaboration.Inordertohidethoselocalartifactsofeach 
party,thepartymustensurethatifalocalartifactsynchro- 
nizes withafragmentofasharedartifact,thenthis 
fragment mustbehiddenaswell.Thishidingproperty 
refers totheabstractionofsuchfragment.However,ifthe 
entire lifecycleofalocalartifactisabstracted,thenitcanbe 
validlyhiddeninthepublicview.In Section 4.3.2, we 
discuss thefully-embeddingpropertytocapturethiskind 
of lifecycle.Inotherwords,ifthelifecycleofalocalartifact 
is afully-embeddedexternallifecycleofthesharedartifact, 
then thelocalartifactcanbehidden.Assuch,thecorre- 
sponding fragment(s)inthesharedartifact(s)isalso 
abstracted(byusingsyncabstractionfunction sa_f ). How- 
ever,thereisacaseifthatfragmentisnotanAL-fragment, 
then the sa_f function cannotbeappliedduetothe 
requirementoftheinputthatmustbeAS-region(consist- 
ing ofASL-fragmentsofsharedartifact(s)andlocalartifact). 
Tocopewiththisissue,thenweproposetouselifecycle 
abstractionfunction la_state to abstractitintoanabstract 
state. Thenthelocalartifactthatsynchronizeswiththat 
fragment canbeabstracted.Althoughwecanabstractan 
NAL-fragmentintoanabstractstate,itishardtodecide 
whether thatabstractionisvalidandconsistentintermsof 
synchronizationbehavior.Thelocalartifactthatsynchro- 
nizes withanypartofsuchNAL-fragmentmustbeexclu- 
sivelyencapsulatedintheabstractstate.Inaddition,the 
abstractstateitselfisnotconsideredasatomicifithas 
multiple entryandmultipleexittransitionsfrom/todiffer- 
ent states.Therefore,weallowhavingsyncabstractionfor 
the abstractstateforthecasethatthewholelifecycleofthe 
local artifactissynchronizedwithintheNAL-fragment.As 
previouslydiscussed,representinganNAL-fragmentinan 
abstracttransitionisdeemedinconsistent. 
Example 17. Revisitourpurchasingprocessexamplein 
Fig. 1. Wecanconstructafragmentforthe PO artifact that 
consists ofstates fcreated; on_holdg. Thefragmentiscon- 
sideredasNAL-fragment(duetotwoexitstates).Wecan 
see thatitsynchronizeswiththeentirelifecycleofthe 
Quote artifact, therefore,theabstractionofthisfragment 
must yieldanabstractstatewhichrequiresthewhole 
lifecycleof Quote tobeabstracted,i.e.,fully-embedded – 
that isthe approving stateofabstract PO in thepublicview 
shownin Fig. 5. Similarly,theentirelifecycleof PL artifact 
can beabstractedandfully-embeddedinthe supplying 
stateoftheabstract PO in thepublicview. 
Next,weproposeanalgorithm(forafunctionnamed 
f ind_minPV) tohelporganizationstoautomaticallyfindthe 
minimal, consistentpublicviewfromtheir ACP-i model. 
Due totheinconsistencyissueonNAL-fragmentabstrac- 
tion, wedonottakeNAL-fragmentsintoaccountinthis 
algorithm.Afterpresentingthealgorithm,wethenshow 
howitcanguaranteethe B-consistency. 
Algorithm4.(function find_minPV). Finding themini- 
mal, consistentpublicviewofan ACP-imodel 
Input: ACP-i model Π^ ¼ ðZ; V; R; L; γÞ. 
Output: theminimalpublicviewof Π^ . 
1. public ACP-imodel Π^ 0 ¼ Π^ ; 
2. for each CiAfCAΠ^ 0Z jγðCÞj ¼ 1g do 
3. L-fragment ℓi ¼ f ind_Lf ðCi ; Ci :SÞ; 
4. AS-region ω¼ find_minASRðℓiÞ; 
5. i f(ω anull) 
6. then 
7. Π^ 0:¼ sa_f ð Π^ 0; ωÞ; 
8. artifacts Zl’Ci; 
9. for each CjAfCAΠ^ 0:Z jγðCÞj ¼ 1g do 
10. L-fragment ℓj ¼ f ind_Lf ðCj ; Cj :SÞ; 
11. if ðℓj ¼ nullÞ then Zl’Cj ; 
12. end for 
13. remove all artifactsin Zl and theirrelatedabstractsyncrules 
from Π^ 0; 
14. Zl 
¼ ∅; 
15.endif 
16.endfor 
17. return Π^ 0; 
Now,weexplainthealgorithmforthe f ind_minPV 
function. First,itinitializesapublicviewasidenticalto 
an input ACP-imodel. Then,itsearchesforalllocalartifacts 
in thepublicview(Line2).Foreachlocalartifactfound,it 
attemptstoconstructanAS-regionthatconsistsofsuch 
artifact (Lines3–4). IfitisabletoconstructtheAS-region 
(Line 5),thenabstractionfunction sa_f is appliedonthe 
regionandtheartifactisaddedtotheset Zl and allartifacts 
in Zl will belaterremovedfromthepublicview(Lines7–8). 
As sa_f finds allcorrespondingASL-fragments(ifconstruc- 
tible) thatsynchronizedwiththelocalartifactintoaccount 
for theabstraction,thealgorithmalsosearchesforother 
local artifactthatqualifiesasfully-embeddedexternallife- 
cycletobeincludedin Zl (Lines 9–12).Thiswilleliminate 
an unnecessarylocalartifactinthepublicview;therefore, 
the numberofiterationsforfindinglocalartifactsis 
reducedinthemainloop. 
Fromthealgorithmwewillgettheminimal, B-consistent 
public viewofthe ACP-i model. However,itdoesnot 
guaranteethatalllocalartifactsareabstracted.This 
depends onwhetherthelifecycleofsuchartifactscanbe 
S. Yongchareonetal./InformationSystems47(2015)5170 –81
fully-embeddedaswellasthesynchronizedpartofthe 
sharedartifactisanASL-fragment.Inotherwords,ifthe 
fragmentisanNAL-fragment,then,aspreviouslydiscussed, 
the abstractionofthoselocalartifactsmustbeachieved 
manually – by abstractingeachofsuchfragmentsintoan 
abstractstateinstead.Thisprocessrequiresapredefined 
abstractstateforanabstractsharedartifact. 
Next,weshowin Theorem 4 that our f ind_minPV 
function producesaminimal, B-consistent public view. 
Theorem4. f ind_minPVðΠ^ Þ returns aminimal, B-consistent 
public viewof Π^ . 
Wecanprove Theorem4 as follows.First,weprovethat 
f ind_minPVð Π^ Þ returns acorrectpublicviewof Π^ . Thiscan 
be donebyinductionoverallmappingconditionsin 
Definition 12. Notethatremovinglocalartifactfromthe 
public viewwhereitsentirelifecyclecanbeabstractedalso 
conforms tothedefinitionofpublicview.Second,weprove 
that given ACP-i model Π^ , f ind_minPVð Π^ Þ is guaranteedto 
returnminimalpublicviewof Π^ . Asweusefunction 
find_minASR to searchfortheminimalAS-regionthatcan 
be usedasaninputoftheabstractionfunction sa_f , then 
this statementisnaturallysatisfied.Last,weprovethe B-consistency 
of thegeneratedpublicview.Basedon Theorem 
3 and Definition 24, thepublicviewpreserves B-consistency 
as function sa_f alwaysyieldsa B-consistent abstractACP 
model oftheinputtedACPmodel. 
Wenowconclude Section4. Insummary,thissectionhas 
discussedthepublicviewconstructionmethodologyand 
behaviorconsistencyandformulatesseveralfunctionsand 
theoremsthatcanbeusedtoconstructthe B-consistent public 
viewofthecollaboration.Italsopresentsouralgorithmsthat 
helporganizationsautomaticallygeneratetheminimal 
abstractpublicviewwiththeassuranceofthe B-consistency. 
5. Consistentprocesschangesandprivateviews 
In thissection,weproposeamechanismthatallowsa 
party inthecollaborationtochangetheirlocal(private) 
processwithoutaffectingthecorrectnessandconsistency 
of theoverallprocess. 
5.1.Localprocesschanges 
Organizations mayneedtochangetheirlocalprocesses 
due totheirnew(orupdated)setofbusinessrequirements 
or regulationsthattheyhavetofollow.Weobservethatin 
artifact-centric processes,changestolocalprocessescanbe 
classified intothreetypesbasedonthethreecomponents 
of the ACPmodel. 
 Changes toartifacts. Anorganizationmaymodify/delete/ 
add anattributeorastateofitslocalartifact,whichcan 
be seenasstructuralchangestothedatamodelofan 
artifact. Inaddition,exceptforchangestoexisting 
artifacts, itispossiblethatanorganizationmayincor- 
porateasetofnewartifactsintothelocalprocess. 
 Changes totasks. Stemmedfromservice-orientedarchi- 
tecture,anorganizationmayseektoaggregateexisting 
local tasksintoacompositetask(i.e.,aservice).Onthe 
otherhand,acompositetaskcanbedecomposedinto 
smaller tasks.Duetothesepossiblechangestotasks,the 
specification ofatask,includingtheinput,output,and 
pre/post-conditionsofthetask,issubjecttoreflecting 
the changes –. 
 Changes tobusinessrules. Businessrulescanbechanged 
due tothestructuralchangesofartifacts,thespecifica- 
tion changesoftasks,andthechangesoftheprocess 
logic. Itisworthnotingthatachangemayaffectan 
existinginteractionbehaviorbetweenartifacts. 
Next,wedefinethree changeoperators that canbeusedto 
expresswhatandhowtheabove-mentionedthreetypesof 
changecanbeachieved.Asaforementioned,inthisarticlewe 
onlyfocusonthebehavioraspectofartifact-centricbusiness 
processes.Werestrictourdiscussioninbehavioralchangesof 
localprocesses.Therefore,forsimplicity,weassumethat 
capturingthechangeofabusinessrulecanreflectthechange 
of artifact(s)andtask(s)involvedinthisrule.Thismakes 
sense astheassociationsbetweenartifactsandtasksare 
defined inthepre/post-conditionsandactions,respectively, 
in businessrules.Inthecaseofaddinganewartifacttoalocal 
process,wecanderivethelifecycleofthenewartifactandits 
interactionfromtheaddedbusinessrulesthatareusedto 
inducethestatetransitionoftheartifactandthesyncrules 
thatareusedtosynchronizeother existingartifact(s)withit, 
respectively. 
Definition 25. (Change operators).Let Π^ l 
¼ ðZl; Vl; Rl 
Þ be 
a localACPmodeloforganizationrolelinthecollaboration. 
An organizationcanchangeitsownlocalprocessby 
applyingthefollowingchangeoperatorsoverbusiness 
rules Rl. 
 Add operator ðRþ Þ. Wewrite Rþ ðrÞ to meanthatnew 
business rulerisaddedintoR,i.e., Rþ ðrÞ ¼ Rl 
[ frg. 
 Delete operator ðR Þ. Wewrite R ðrÞ to meanthat 
existingbusinessrulerisdeletedfromR,i.e., 
R ðrÞ ¼ Rl 
frg. 
 Replaceoperator (R%). Wewrite R%ðrx; ryÞ tomeanthat 
existingbusinessrulerx is replacedbynewbusiness 
rule ry in R,i.e., R%ðrx; ryÞ ¼ Rl 
[ fryg frxg. 
It isnotedthatthereplaceoperatoridenticallyperforms 
as thecombinationoftheaddoperatorandthedelete 
operator;however,itprovidesbettertraceabilityofchanges 
by maintainingtherelationforthereplacementofanold 
business rulewithanewbusinessrule. 
Definition 26. (Modified localACPmodel,processchange 
function (pc)).Let Π^ l 
¼ ðZl; Vl; Rl 
Þ be a local ACPmodel of 
role l.Wecanobtain modified localACPmodel Π^ l0 
from Π^ l 
by applying processchangefunctionpc Π^ l :Π^ l 
 X-Π^ l0 
where 
X is aunionsetof changeoperationsRþ, R, andR% that 
perform onsomebusinessrulesinRl. 
Next,weexplaintheconceptofprivateviewandhowit 
can beusedtovalidatethechangesoflocalprocess. 
S. Yongchareonetal./InformationSystems47(2015)51–81 71
5.2. Privateviewsandchangevalidation 
In theoverallpicture,wevalidatethechangesoflocal 
processbycheckingwhetherthe B-consistency of the 
modified inter-organizationalbusinessprocess(afterlocal 
processchangesapplied)andtheagreedpublicviewcanbe 
preserved.A privateview of anorganizationisusedto 
capturethemodifiedlocalprocessforthepurposeoflocally 
checkingwhetherthelocalprocessisstillabletoprovide 
what promisedintheagreedcontract,i.e.,publicview.We 
illustrateanoverviewofprocesschangevalidationin 
Fig. 18. 
Next,wedefinethe privateview of aparticularorgani- 
zation roleinthecollaborationbasedontheagreedpublic 
view.Basedoneachversionoflocalprocesschanges,each 
role hasacorrespondingprivateviewthatcapturesits 
(modified) local ACPmodel plus abstractsharedartifacts 
defined inthepublicview. 
Definition 27. (Privateview).Let paΠ^ ¼ ðZp; Vp; Rp; Lp; γpÞ 
be apublicviewofACP-imodel Π^ and Π^ l 
¼ ðZl; Vl; Rl 
Þ be a 
local ACPmodelofrole lALp. The privateview of role l can 
be definedby privateviewmapping function pv: Zx 
[ Vx 
[ Rx 
[ γx-ðZp 
[ Zl 
Þ [ Vl 
[ ðRp 
[ Rl 
Þ [ Lp 
[ γp such thatthe 
followingshold. 
 each abstractsharedartifactin Zp existsin Zx, i.e., 
8CiAfCAZp 
j γpðCÞ  
 
41g; (CjAZx; pvðCjÞ ¼ Ci  γxðCjÞ ¼ γp ðCiÞ 
 each localartifactin Zl (not in Zp) existsin Zx, i.e., 
8CiAfCAZl 
jjγpðCÞj ¼ 1g; (CjAZx; pvðCjÞ ¼ Ci 
 γxðCjÞ ¼ l 
each taskin Vl existsin Vx, i.e., 
8viAVl; (vjAVx; pvðvjÞ ¼ vi  γxðvjÞ ¼ l 
 each businessrulein Rl existsin Rx, i.e., 
8riARl; (rjARx; pvðrjÞ ¼ ri; 
 each abstractbusinessrulein Rp that isnotusedforthe 
synchronizationbetweenanabstractsharedartifactin 
Zp and alocalartifactin Zl existsin Rx, i.e., 
8riAfrARp 
jl=2γpðrÞg; (rjARx; pvðrjÞ ¼ ri 
Notethatwemayusethetermprivateviewforthe 
lifecycleofthisprivateviewinanunambiguouscontext. 
Example 18. Fig. 19 depicts thelifecyclesofartifactsinthe 
privateviewoftheoriginallocalprocessofSupplier 
(Π^ Supplier 
1 ) whichisextractedfromthecompletepurchasing 
processshownin Fig. 1. Suchprivateviewcanbecon- 
structedbasedonthepublicviewshownin Fig. 5. Apart 
fromthelocalartifacts PL and DN, comparedwiththe 
public view,wecanseelocalprocessdetails(ingray- 
shaded areas)ofsharedartifacts PO, SO, and IV. 
Example 19. Now,considerthecasethatifSupplierwants 
tochangeitslocalprocessbasedontheexistingone(cf. 
Fig. 19). Theresultofchangesisillustratedinprivateview 
Π^ Supplier 
2 in Fig. 20. Sincethe IL artifact isaddedintothelocal 
process,wecanseesomesynchronizationbetweenexisting 
artifact PL and newartifact IL, aswellasthechangeof 
state’s namesof PL (from checking to scheduled, from out of 
stock to unavailable, andfrom in stock to picking). Existing 
business rulesthatcorrespondtothesechangesare 
affectedandneededtobeadjustedaccordingly.Obviously, 
a newsetofbusinessrulesisneededtoexpressthe 
lifecycleoftheaddedartifact.Thissetincludessyncrules 
that areusedforthesynchronizationbetween PL and IL. 
Next,wedefine processconformance and itsconditions 
that canbeusedtocheckwhetherchangesinlocalprocess 
can beimplementedwhilepreservingthecorrectnessand 
consistencyoftheoverallprocess.Basically,wereusethe 
definition of B-consistency to define processconformance for 
Fig. 18. An overviewofprivateviewsandviewconformance. 
S. Yongchareonetal./InformationSystems47(2015)5172 –81
consistentlocalprocesschangesastosatisfythefollowing 
statements. 
 Changesshouldnotleadtoanunsoundglobalprocess, 
i.e., themodifiedglobalprocessshouldbeabletoreach 
its goalstatesasitsoriginalglobalprocessdoes,and, 
 Changesshouldbeguaranteedthatthe B-consistency of 
the modifiedglobalprocessanditsoriginalglobal 
processispreserved. 
As anagreedpublicviewisconstructedforthecolla- 
borationwhichactslikeacontract,sowecanexpress 
consistent processchanges by meansofnotbreakingthe 
original publicview.Inotherwords,themodified ACP-i 
model must beconsistentwiththeagreedpublicview. 
Regardingthebehavioralequivalencenotioninprocess 
algebras [9], wecansaythatourapproachforconsistent 
processchangespreservesthecongruencepropertyofthe 
modified localprocessanditsbaselocalprocess,i.e.,they 
can behaveinterchangeablywithoutaffectingtheoverall 
process. 
Definition 28. (Processconformance, F). Let Π^ ¼ ðZ; 
V; R; L; γÞ be anACP-imodeland paΠ^ be its public view. 
Let pv Π^ l be aprivateviewof local ACPmodel Π^ l 
for role lAL. 
Wesaythat Π^ l 
conformspa Π^ , writtenas Π^ l 
FpaΠ^ , iffthe 
Picking List(PL) 
Purchase Order(PO) 
confirmed 
canceled 
Out ofstock 
closed 
Shipping Order(SO) 
In transit 
Invoice (IV) 
approving 
acquiring 
accepted filled 
ready tofill Filled order 
checking Instock 
Delivery Note(DN) 
prepared 
transferring 
dispatched 
billing 
issued 
cleared 
sent 
unpaid 
clearing 
arrived 
delivering 
ready toship 
created scheduled 
Fig. 19. Supplier’s privateview Π^ Supplier 
1 . 
Picking List(PL) 
Purchase Order(PO) 
confirmed 
canceled 
unavailable 
closed 
Shipping Order(SO) 
Invoice (IV) 
approving 
acquiring 
accepted filled 
ready tofill Filled order 
scheduled picking 
Delivery Note(DN) 
prepared 
transferring 
dispatched 
billing 
issued 
cleared 
sent 
unpaid 
clearing 
Inventory List(IL) 
checking 
ready topick 
sourcing 
canceled 
delivering 
ready toship 
In transit created scheduled arrived 
Fig. 20. Supplier’s modifiedprivateview Π^ Supplier 
2 . 
S. Yongchareonetal./InformationSystems47(2015)51–81 73
A view framework for modeling and change validation of artifact centric inter-organizational business processes
A view framework for modeling and change validation of artifact centric inter-organizational business processes
A view framework for modeling and change validation of artifact centric inter-organizational business processes
A view framework for modeling and change validation of artifact centric inter-organizational business processes
A view framework for modeling and change validation of artifact centric inter-organizational business processes
A view framework for modeling and change validation of artifact centric inter-organizational business processes
A view framework for modeling and change validation of artifact centric inter-organizational business processes
A view framework for modeling and change validation of artifact centric inter-organizational business processes

Weitere ähnliche Inhalte

Ähnlich wie A view framework for modeling and change validation of artifact centric inter-organizational business processes

A framework for realizing artifact centric business processes in SOA
A framework for realizing artifact centric business processes in SOAA framework for realizing artifact centric business processes in SOA
A framework for realizing artifact centric business processes in SOADr. Sira Yongchareon
 
Process aware information system at amikom oct 2015
Process aware information system at amikom oct 2015Process aware information system at amikom oct 2015
Process aware information system at amikom oct 2015Rumah Studio
 
Rapid Knowledge Deployment in an Organizational-Memory-Based Workflow Environ...
Rapid Knowledge Deployment in an Organizational-Memory-Based Workflow Environ...Rapid Knowledge Deployment in an Organizational-Memory-Based Workflow Environ...
Rapid Knowledge Deployment in an Organizational-Memory-Based Workflow Environ...Ralf Klamma
 
ISS Service Innovation Leadership Seminar, 28 March - Jochen Wirtz
ISS Service Innovation Leadership Seminar, 28 March - Jochen WirtzISS Service Innovation Leadership Seminar, 28 March - Jochen Wirtz
ISS Service Innovation Leadership Seminar, 28 March - Jochen WirtzNUS-ISS
 
Kanchan Ghangrekar_SrTestingAnalyst
Kanchan Ghangrekar_SrTestingAnalystKanchan Ghangrekar_SrTestingAnalyst
Kanchan Ghangrekar_SrTestingAnalystKanchan Ghangrekar
 
Towards a Software Framework for Automatic Business Process Redesign
Towards a Software Framework for Automatic Business Process RedesignTowards a Software Framework for Automatic Business Process Redesign
Towards a Software Framework for Automatic Business Process RedesignIDES Editor
 
LinearProgrammingModelsOverviewandRationaleT.docx
LinearProgrammingModelsOverviewandRationaleT.docxLinearProgrammingModelsOverviewandRationaleT.docx
LinearProgrammingModelsOverviewandRationaleT.docxwashingtonrosy
 
LinearProgrammingModelsOverviewandRationaleT.docx
LinearProgrammingModelsOverviewandRationaleT.docxLinearProgrammingModelsOverviewandRationaleT.docx
LinearProgrammingModelsOverviewandRationaleT.docxcroysierkathey
 
Process Innovation Redesigning an Enterprise Backbone System
Process Innovation Redesigning an Enterprise Backbone SystemProcess Innovation Redesigning an Enterprise Backbone System
Process Innovation Redesigning an Enterprise Backbone SystemCONFENIS 2012
 
9 madderi sivalingamsaravanan_106-110
9 madderi sivalingamsaravanan_106-1109 madderi sivalingamsaravanan_106-110
9 madderi sivalingamsaravanan_106-110Alexander Decker
 
Presentation of aviation
Presentation of aviationPresentation of aviation
Presentation of aviationcohtech
 
Workshop on Knowledge-based Service Innovation
Workshop on Knowledge-based Service InnovationWorkshop on Knowledge-based Service Innovation
Workshop on Knowledge-based Service Innovation2016
 
Ceviche framework for dynamic adaptation business process using complex event...
Ceviche framework for dynamic adaptation business process using complex event...Ceviche framework for dynamic adaptation business process using complex event...
Ceviche framework for dynamic adaptation business process using complex event...eSAT Journals
 

Ähnlich wie A view framework for modeling and change validation of artifact centric inter-organizational business processes (20)

A framework for realizing artifact centric business processes in SOA
A framework for realizing artifact centric business processes in SOAA framework for realizing artifact centric business processes in SOA
A framework for realizing artifact centric business processes in SOA
 
Process aware information system at amikom oct 2015
Process aware information system at amikom oct 2015Process aware information system at amikom oct 2015
Process aware information system at amikom oct 2015
 
5 5-norbert jastroch
5 5-norbert jastroch5 5-norbert jastroch
5 5-norbert jastroch
 
PAVAN KUMAR DANGETI
PAVAN KUMAR DANGETIPAVAN KUMAR DANGETI
PAVAN KUMAR DANGETI
 
PAVAN KUMAR DANGETI
PAVAN KUMAR DANGETIPAVAN KUMAR DANGETI
PAVAN KUMAR DANGETI
 
Rapid Knowledge Deployment in an Organizational-Memory-Based Workflow Environ...
Rapid Knowledge Deployment in an Organizational-Memory-Based Workflow Environ...Rapid Knowledge Deployment in an Organizational-Memory-Based Workflow Environ...
Rapid Knowledge Deployment in an Organizational-Memory-Based Workflow Environ...
 
ISS Service Innovation Leadership Seminar, 28 March - Jochen Wirtz
ISS Service Innovation Leadership Seminar, 28 March - Jochen WirtzISS Service Innovation Leadership Seminar, 28 March - Jochen Wirtz
ISS Service Innovation Leadership Seminar, 28 March - Jochen Wirtz
 
Agile It 20091020
Agile It 20091020Agile It 20091020
Agile It 20091020
 
Kanchan Ghangrekar_SrTestingAnalyst
Kanchan Ghangrekar_SrTestingAnalystKanchan Ghangrekar_SrTestingAnalyst
Kanchan Ghangrekar_SrTestingAnalyst
 
Towards a Software Framework for Automatic Business Process Redesign
Towards a Software Framework for Automatic Business Process RedesignTowards a Software Framework for Automatic Business Process Redesign
Towards a Software Framework for Automatic Business Process Redesign
 
LinearProgrammingModelsOverviewandRationaleT.docx
LinearProgrammingModelsOverviewandRationaleT.docxLinearProgrammingModelsOverviewandRationaleT.docx
LinearProgrammingModelsOverviewandRationaleT.docx
 
LinearProgrammingModelsOverviewandRationaleT.docx
LinearProgrammingModelsOverviewandRationaleT.docxLinearProgrammingModelsOverviewandRationaleT.docx
LinearProgrammingModelsOverviewandRationaleT.docx
 
Process Innovation Redesigning an Enterprise Backbone System
Process Innovation Redesigning an Enterprise Backbone SystemProcess Innovation Redesigning an Enterprise Backbone System
Process Innovation Redesigning an Enterprise Backbone System
 
klh-case-study
klh-case-studyklh-case-study
klh-case-study
 
9 madderi sivalingamsaravanan_106-110
9 madderi sivalingamsaravanan_106-1109 madderi sivalingamsaravanan_106-110
9 madderi sivalingamsaravanan_106-110
 
PROACTVE
PROACTVEPROACTVE
PROACTVE
 
Presentation of aviation
Presentation of aviationPresentation of aviation
Presentation of aviation
 
Workshop on Knowledge-based Service Innovation
Workshop on Knowledge-based Service InnovationWorkshop on Knowledge-based Service Innovation
Workshop on Knowledge-based Service Innovation
 
ebizQ publication
ebizQ publicationebizQ publication
ebizQ publication
 
Ceviche framework for dynamic adaptation business process using complex event...
Ceviche framework for dynamic adaptation business process using complex event...Ceviche framework for dynamic adaptation business process using complex event...
Ceviche framework for dynamic adaptation business process using complex event...
 

Mehr von Dr. Sira Yongchareon

Efficient Process Model Discovery Using Maximal Pattern Mining
Efficient Process Model Discovery Using Maximal Pattern MiningEfficient Process Model Discovery Using Maximal Pattern Mining
Efficient Process Model Discovery Using Maximal Pattern MiningDr. Sira Yongchareon
 
A workflow execution platform for collaborative artifact centric business pro...
A workflow execution platform for collaborative artifact centric business pro...A workflow execution platform for collaborative artifact centric business pro...
A workflow execution platform for collaborative artifact centric business pro...Dr. Sira Yongchareon
 
An artifact centric view-based approach to modeling inter-organizational busi...
An artifact centric view-based approach to modeling inter-organizational busi...An artifact centric view-based approach to modeling inter-organizational busi...
An artifact centric view-based approach to modeling inter-organizational busi...Dr. Sira Yongchareon
 
An Artifact-centric View-based Approach to Modeling Inter-organizational Busi...
An Artifact-centric View-based Approach to Modeling Inter-organizational Busi...An Artifact-centric View-based Approach to Modeling Inter-organizational Busi...
An Artifact-centric View-based Approach to Modeling Inter-organizational Busi...Dr. Sira Yongchareon
 
A framework for behavior consistent specialization of artifact-centric busine...
A framework for behavior consistent specialization of artifact-centric busine...A framework for behavior consistent specialization of artifact-centric busine...
A framework for behavior consistent specialization of artifact-centric busine...Dr. Sira Yongchareon
 
A framework for realizing artifact centric business processes in soa
A framework for realizing artifact centric business processes in soaA framework for realizing artifact centric business processes in soa
A framework for realizing artifact centric business processes in soaDr. Sira Yongchareon
 
An artifact centric approach to generating web-based business process driven ...
An artifact centric approach to generating web-based business process driven ...An artifact centric approach to generating web-based business process driven ...
An artifact centric approach to generating web-based business process driven ...Dr. Sira Yongchareon
 
An artifact centric approach to generating web-based business process driven ...
An artifact centric approach to generating web-based business process driven ...An artifact centric approach to generating web-based business process driven ...
An artifact centric approach to generating web-based business process driven ...Dr. Sira Yongchareon
 
A process view framework for artifact centric business processes
A process view framework  for artifact centric business processesA process view framework  for artifact centric business processes
A process view framework for artifact centric business processesDr. Sira Yongchareon
 
Process view framework for artifact centric business processes
Process view framework for artifact centric business processesProcess view framework for artifact centric business processes
Process view framework for artifact centric business processesDr. Sira Yongchareon
 

Mehr von Dr. Sira Yongchareon (11)

Efficient Process Model Discovery Using Maximal Pattern Mining
Efficient Process Model Discovery Using Maximal Pattern MiningEfficient Process Model Discovery Using Maximal Pattern Mining
Efficient Process Model Discovery Using Maximal Pattern Mining
 
A workflow execution platform for collaborative artifact centric business pro...
A workflow execution platform for collaborative artifact centric business pro...A workflow execution platform for collaborative artifact centric business pro...
A workflow execution platform for collaborative artifact centric business pro...
 
An artifact centric view-based approach to modeling inter-organizational busi...
An artifact centric view-based approach to modeling inter-organizational busi...An artifact centric view-based approach to modeling inter-organizational busi...
An artifact centric view-based approach to modeling inter-organizational busi...
 
An Artifact-centric View-based Approach to Modeling Inter-organizational Busi...
An Artifact-centric View-based Approach to Modeling Inter-organizational Busi...An Artifact-centric View-based Approach to Modeling Inter-organizational Busi...
An Artifact-centric View-based Approach to Modeling Inter-organizational Busi...
 
A framework for behavior consistent specialization of artifact-centric busine...
A framework for behavior consistent specialization of artifact-centric busine...A framework for behavior consistent specialization of artifact-centric busine...
A framework for behavior consistent specialization of artifact-centric busine...
 
A framework for realizing artifact centric business processes in soa
A framework for realizing artifact centric business processes in soaA framework for realizing artifact centric business processes in soa
A framework for realizing artifact centric business processes in soa
 
An artifact centric approach to generating web-based business process driven ...
An artifact centric approach to generating web-based business process driven ...An artifact centric approach to generating web-based business process driven ...
An artifact centric approach to generating web-based business process driven ...
 
An artifact centric approach to generating web-based business process driven ...
An artifact centric approach to generating web-based business process driven ...An artifact centric approach to generating web-based business process driven ...
An artifact centric approach to generating web-based business process driven ...
 
BPMN process views construction
BPMN process views constructionBPMN process views construction
BPMN process views construction
 
A process view framework for artifact centric business processes
A process view framework  for artifact centric business processesA process view framework  for artifact centric business processes
A process view framework for artifact centric business processes
 
Process view framework for artifact centric business processes
Process view framework for artifact centric business processesProcess view framework for artifact centric business processes
Process view framework for artifact centric business processes
 

Kürzlich hochgeladen

The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxMalak Abu Hammad
 
FULL ENJOY 🔝 8264348440 🔝 Call Girls in Diplomatic Enclave | Delhi
FULL ENJOY 🔝 8264348440 🔝 Call Girls in Diplomatic Enclave | DelhiFULL ENJOY 🔝 8264348440 🔝 Call Girls in Diplomatic Enclave | Delhi
FULL ENJOY 🔝 8264348440 🔝 Call Girls in Diplomatic Enclave | Delhisoniya singh
 
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking MenDelhi Call girls
 
Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...
Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...
Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...Neo4j
 
Install Stable Diffusion in windows machine
Install Stable Diffusion in windows machineInstall Stable Diffusion in windows machine
Install Stable Diffusion in windows machinePadma Pradeep
 
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationBeyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationSafe Software
 
Factors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptxFactors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptxKatpro Technologies
 
AI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsAI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsMemoori
 
Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...Alan Dix
 
How to Remove Document Management Hurdles with X-Docs?
How to Remove Document Management Hurdles with X-Docs?How to Remove Document Management Hurdles with X-Docs?
How to Remove Document Management Hurdles with X-Docs?XfilesPro
 
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationFrom Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationSafe Software
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Patryk Bandurski
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024Scott Keck-Warren
 
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024BookNet Canada
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerThousandEyes
 
Azure Monitor & Application Insight to monitor Infrastructure & Application
Azure Monitor & Application Insight to monitor Infrastructure & ApplicationAzure Monitor & Application Insight to monitor Infrastructure & Application
Azure Monitor & Application Insight to monitor Infrastructure & ApplicationAndikSusilo4
 
Presentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreterPresentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreternaman860154
 
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024Rafal Los
 
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsHandwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsMaria Levchenko
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Allon Mureinik
 

Kürzlich hochgeladen (20)

The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptx
 
FULL ENJOY 🔝 8264348440 🔝 Call Girls in Diplomatic Enclave | Delhi
FULL ENJOY 🔝 8264348440 🔝 Call Girls in Diplomatic Enclave | DelhiFULL ENJOY 🔝 8264348440 🔝 Call Girls in Diplomatic Enclave | Delhi
FULL ENJOY 🔝 8264348440 🔝 Call Girls in Diplomatic Enclave | Delhi
 
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
 
Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...
Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...
Neo4j - How KGs are shaping the future of Generative AI at AWS Summit London ...
 
Install Stable Diffusion in windows machine
Install Stable Diffusion in windows machineInstall Stable Diffusion in windows machine
Install Stable Diffusion in windows machine
 
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry InnovationBeyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
Beyond Boundaries: Leveraging No-Code Solutions for Industry Innovation
 
Factors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptxFactors to Consider When Choosing Accounts Payable Services Providers.pptx
Factors to Consider When Choosing Accounts Payable Services Providers.pptx
 
AI as an Interface for Commercial Buildings
AI as an Interface for Commercial BuildingsAI as an Interface for Commercial Buildings
AI as an Interface for Commercial Buildings
 
Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...Swan(sea) Song – personal research during my six years at Swansea ... and bey...
Swan(sea) Song – personal research during my six years at Swansea ... and bey...
 
How to Remove Document Management Hurdles with X-Docs?
How to Remove Document Management Hurdles with X-Docs?How to Remove Document Management Hurdles with X-Docs?
How to Remove Document Management Hurdles with X-Docs?
 
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationFrom Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
 
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
Integration and Automation in Practice: CI/CD in Mule Integration and Automat...
 
SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024SQL Database Design For Developers at php[tek] 2024
SQL Database Design For Developers at php[tek] 2024
 
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
#StandardsGoals for 2024: What’s new for BISAC - Tech Forum 2024
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
 
Azure Monitor & Application Insight to monitor Infrastructure & Application
Azure Monitor & Application Insight to monitor Infrastructure & ApplicationAzure Monitor & Application Insight to monitor Infrastructure & Application
Azure Monitor & Application Insight to monitor Infrastructure & Application
 
Presentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreterPresentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreter
 
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024
 
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsHandwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed texts
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)
 

A view framework for modeling and change validation of artifact centric inter-organizational business processes

  • 1. A viewframeworkformodelingandchangevalidation of artifact-centricinter-organizationalbusinessprocesses SiraYongchareon a, Chengfeiliu b, YuJianc, XiaohuiZhao d a Department ofComputing,UnitecInstituteofTechnology,Auckland,NewZealand b FacultyofScience,EngineeringandTechnology,SwinburneUniversityofTechnology,Victoria,Australia c School ofComputerandMathematicalSciences,AucklandUniversityofTechnology,NewZealand d Faculty ofBusiness,GovernmentandLaw,UniversityofCanberra,Canberra,Australia a r t i c l e info Article history: Received27November2012 Receivedinrevisedform 13June2014 Accepted21July2014 Availableonline1August2014 Keywords: Business processmodeling Artifact-centric workflows Inter-organizationalbusinessprocesses Process views Process abstraction Model verification a b s t r a c t Over thepastseveralyears,moreefficientapproacheshavebeenonincreasingdemands for designing,modeling,andimplementinginter-organizationalbusinessprocesses.Inthe process collaborationacrossorganizationalboundaries,organizationsstillstayautonomic, whichmeanseachorganizationcanfreelymodifyitsinternaloperationstomeetits privategoalswhilesatisfyingthemutualobjectiveswithitspartners.Recently,artifact- centric processmodelinghasbeenevidencedwithhigherflexibilityinprocessmodeling and executionthantraditionalactivity-centricmodelingmethods.Althoughsomeefforts havebeenputtoexploringhowartifact-centricmodelingfacilitatesthecollaboration between organizations,theachievementisstillfarfromsatisfactionlevel,particularlyin aspects ofprocessmodelingandvalidating.Tofillinthegaps,weproposea view framework for modelingandvalidatingthechangesofinter-organizationalbusiness processes. Theframeworkconsistsofanartifact-centricprocessmeta-model,publicview constructingmechanism,andprivateviewandchangevalidatingmechanisms,which are speciallydesignedtofacilitatetheparticipatingorganizationstocustomizetheir internal operationswhileensuringthecorrectnessofthecollaboratingprocesses.Wealso implementasoftwaretoolnamed Artifact-M to helporganizationstoautomatically constructaminimalandconsistentpublicviewfromtheirprocesses. & 2014ElsevierLtd.Allrightsreserved. 1. Introduction Recently,service-orientedarchitecture(SOA)hasbecomea predominantITtoolforfacilitating businessestomeetthe changingrequirementsofthemarket.SOAparticularlyenables thebusinesscollaborationacrossorganizationsbycomposing Webservicestoachieveamutualbusinessgoalwithout comprisingtheautonomyofparticipatingorganizations.The furtherapplicationofSOAinfacilitatingbusinesscollaboration calls forefficientapproachesfordesigning,modelingand implementinginter-organizationalbusinessprocesses [50]. Recently,workbyDesaietal. [20], Ghattas,Montalietal. [63] and [83] showsthatthequalityofcoordinatingorganiza- tions inaservice-orientedcollaborationreliesonthreemajor requirements,viz., compliance, flexibility, and autonomy.Com- pliance requiresallpartiesmustprovidetheservicesasthey havepromisedinthecollaborationcommitment,suchasa servicelevelagreement. Flexibility allowseachpartytoown thefreedomofchangingandimplementingitsownprocessin thecollaboration.Lastly, autonomy indicateseachparticipating organizationactsindependentlyandisnotobligedtorevealits Contents listsavailableat ScienceDirect journalhomepage: www.elsevier.com/locate/infosys InformationSystems http://dx.doi.org/10.1016/j.is.2014.07.004 0306-4379/& 2014ElsevierLtd.Allrightsreserved. E-mail addresses: sira@maxsira.com (S. Yongchareon), cliu@swin.edu.au (C. liu), jian.yu@aut.ac.nz (J. Yu), xiaohui.zhao@canberra.edu.au (X. Zhao). Information Systems47(2015)51–81
  • 2. ownprivateinformation(orprocess)tootherparties. Althoughservicechoreography definescommoncollaboration behaviorsandkeepstheflexibilityandautonomyofeach participant, actualchoreographymodelingapproachesand relatedmodelinglanguagesmainlydescribethecollaboration fromaproceduralperspective, and focusoncontrol-flow, messagesequencing,etc.,insteadoffromadataperspective. As such,flexibilityandautonomyisstilllimitedbythenature of choreographymodelinglanguages.Theworkondeclarative specificationofservicechoreographieshasbeenproposedby Montalietal. [63], onthebasisofDecSerFlowlanguage [84], toovercomesuchlimitationsandsupportdynamicand complexinter-organizationalprocessspecifications.Priorto theemergenceofservicechoreography,processviewhasbeen adoptedtoaddresstheflexibilityandautonomyissuesin businesscollaborations,andtheimprovementhasbeen extensivelyevidencedinworkbyVanDerAalstandBasten [81], LiuandShen [51,52], SchulzandOrlowska [77], Chiu et al. [16,17], Chebbietal. [14], EshuisandGrefen [10], Zhao et al. [100,101], Jiangetal. [40] andEshuisetal. [27]. Originally,apublic-to-privateviewapproachhasbeenintro- duced by [85] andvanderAalstetal., [82] to resolvethe privacyandautonomyissuesaswellastosupportchange managementindynamiccollaboration.Mostrecently, [83] haveproposedamultipartyprocess-orientedcontractto supportservicechoreographybasedontheconceptsof public/privateviewsandaccordancealongwithoperating guidelines [61,57]. However,allofaboveworksfollowthe traditionalactivity-centricbusinessprocessmodelingpara- digm andtherebyinheritthelimitationsindatamanagement, processintegration,andprocessmodification,becausetradi- tional modelingapproacheslackadequatesupportsofauto- matedtoolsforbusinessprocessinter-operationandprocess schemareuse [38,39]. In thepastfewyears,anewmodelingapproachhasemer- ged,i.e., artifact-centric (operational)businessprocessmodel- ing [68]. Insteadofcontrolflowsofabusinessprocess,business documentsandtheirevolutionthroughabusinessprocess becomethemainmodelingobjects.Thisapproachdepictsa businessprocessinfourdimensions,viz.,businessartifacts, lifecycleofartifact,services,andassociationsbetweenartifacts and services [34]. Thelifecycleofanartifactisdefinedinterms of “businessstages” and thepossibleevolutionoftheartifact. Theevolutionofanartifactandoperationsofrelatedservices arespecifiedintermsoftheirassociations,whichcanbe expressedinadeclarativemanner,e.g.,usingECArules. As anemergingtoolGuard–Stage–Milestonemeta-model is becominganewdeclarativeapproachtomodeling artifactlifecyclebasedonavariantofstatemachines [35,36,21,22,80]. Withtheeffortsofnumerousacademic researchersandindustrialpractitioners,artifact-centric modelingapproachhasbeenextensivelyrecognizedto be withhigherlevelofrobustnessandflexibilityfordescrib- ingprocessspecificationcomparedtotraditionalactivity- centricapproaches.Artifact-centricprocessmodeling receivescomprehensivecontributionsintermsofbusiness transformationpractices [6,7,13], foundations [5,59], design methodologies [8,19,56,64,65,73,74,86], modelspecifica- tion,construction,andverification [54,30,31,44,25,29,102, 103,21], workflowrealization/execution [18,53,66,91,67,88], and monitoring/conformancesupports [55,28]. Uptopresent, artifact-centric approachhasbeenappliedtoseveralindus- try domainssuchas healthcare (e.g., PHILharmonicFlows framework [45,15]), insurance (e.g., in [44]), and finance (e. g., IBMGlobalFinancing [13]). However,comparedwith traditionalactivity-centricapproaches,furtherresearchis sought afterintheareaofbusinesscollaboration. By nowtwomainapproacheshavebeenproposedfor artifact-centricinter-organizationalprocesses. Theinitial attempt usesartifact-centricinteroperationhubtofacilitate and supportinter-organizationalworkflows(inanorchestra- tion perspective)amongmultipleautonomousstakeholders [37]. Recently,thisworkhasbeenbroughtforwardtoanEU- funded projectcalledArtifact-CentricServiceInteroperation (ACSI) [3]. Itispromisedtosupportalargenumberofservice collaborationbyusingartifact-centricinter-operationsandto achievedramaticsavingsoverconventionalapproaches.On the otherhand,theartifact-centricchoreographyapproach hasdefinedinteractingartifact-centricprocesses [58,79]. Althoughflexibilityisnaturallydeemedasoneofthebenefits fromartifact-centricmodelingapproaches,acomprehensive studyonsupportingorganizationstoachieveallthethree collaborationrequirementsisstillmissing.Basedonliterature andpractices,wehaveobservedthatview-basedapproaches tointer-organizationalbusinessprocessmanagementcan provideapromisingandefficientwayofprocessmodeling and changemanagementtoaddresssuchrequirements; nevertheless,ithasnotbeenyetmuchexploredinthecontext of artifact-centricinter-organizational businessprocesses. Therefore,inthisarticle,wearetoexploretheideaofprocess viewinanartifact-centricperspectiveanddevelopaframe- workthatcanhelporganizationstomeettheaforementioned requirementsinacollaboration environment.Wesummarize our contributionsasfollows: Weproposeaformalartifact-centricviewframework based onLTS(LabeledTransitionSystem).Thisframe- workconsistsofthreeparts:(1)anartifact-centricMeta model forinter-organizationalbusinessprocesses,(2) notion of privateview for capturinglocalprocessesof participating organizations,and(3)notionof public view for servingasanagreedcontractofthecollaboration. With public/privateviews,organizationsareableto autonomouslyparticipateinthecollaborationwhile being freetochangetheirlocalprocesses. Wedesignanalgorithmforautomaticallyconstructinga consistent,minimalpublicviewbasedonlocalpro- cesses ofanorganization.Tothebestofourknowledge, this isthefirstalgorithmforautomaticallyconstructing a collaborationcontractthattakesintoaccountinterac- tion behaviorsofartifacts. Wedevelopaverificationmechanismforartifact-centric processesthatallowsorganizationstochangetheirlocal processes,throughtheuseofprivateviews,while preservingthecorrectnessandconsistencyoftheover- all collaboration.Astheverificationisperformedlocally, our mechanismdoesnotsufferfromthestateexplosion issue thatmayoccuringlobalverificationapproaches. The remainderofthisarticleisorganizedasfollows: Section 2 introducesthemotivationofourartifact-centric 52 S. Yongchareonetal./InformationSystems47(2015)51–81
  • 3. approachandtheconceptofprocessviews. Section 3 presentsourviewframeworkforartifact-centricbusiness processes. Sections 4 and 5 address howtocreatepublic viewsandprivateviewsaswellashowtoensureview consistencyandvalidatelocalprocesschanges,respec- tively. Section 6 discusses animplementedprototypefor the proof-of-conceptpurpose. Section 7 reviewstherelated work.Finally,conclusionandfutureworkaregivenin Section 8. 2. Motivatingexample In thissection,weuseapurchasingprocessinthesupply chaincollaborationdomainasanexampletoillustrateand motivatetheartifact-centricapproachtomodelinginter- organizationalbusinessprocesscollaboration.In Fig.1, a completepurchasingprocessmodelinthecollaborationis illustratedbasedonanartifact-centricperspective,which involvesthreerolesofparticipatingorganizations: Buyer, Supplier, and Logistics. Weinitiatethediscussionofthis examplebyidentifyinginvolvedbusinessartifactsand describinghowtheyaremodeledinthiscollaboration. At thebeginningstage,allpartiesidentifyandmodel their requiredbusinessartifactsoftheirlocalprocesses. This stepincludesdefiningorganization-ownedartifacts (called local artifacts) thatareinternallyused/managedby individualpartyaswellastheircommonlyagreedartifacts (called sharedartifacts) thatareusedforthecoordination betweenpartiesinthecollaboration.Weconsiderthatthe lifecycleofasharedartifactshouldrepresentitsagreed business stagesandpossiblestepstowardsthecompletion of theprocess.Inotherwords,sharedartifactsshouldbe used asamutualpointofinterestofallparties;therefore, the coordinationoccursatsomepointswheretheyare processed.Intheimplementation,thesesharedartifactsact as messagesthataresentandreceivedbytheseorganiza- tions. Baseduponthecurrentprocessingstateoftheshared artifact, aresponsibleorganizationthathasreceivedthis artifact willinvokeaspecificserviceaccordingtoacorre- sponding businessruledefinedbythatorganization.This service willthenreadorupdatethesharedartifactand otherlocalartifactsdefinedinthespecification. In Fig. 1, wecanseethattheinter-organizationalprocess consists of PurchaseOrder (PO), Shipping Order (SO), and Invoice (IV) as sharedartifacts. Apartfromthem,Buyer, Supplier,andLogisticshave Quote (Q) and Payment (P), Picking List (PL) and DeliveryNote (DN), and Shipping List (SL) astheir local artifacts, respectively.Inthefigure,we also usedashedlinetodepictthe synchronizationdepen- dencies betweenartifacts(local-local,shared-shared,or local-shared). Nextwebrieflydescribetheprocessin Fig. 1 from an artifact-centric perspective.Atthebeginningoftheprocess, a buyerinitiatesthecreationofthe Quote and the PO documents/artifacts. Oncethe quote is approved,the PO is confirmed andsenttoaselectedsupplier.Whenthe supplier receivesthe PO, itcreatesa PL document forthe purpose ofacquiringgoodsforthat PO. Ifthegoodsrunout, then thesupplierrejectstosupplythemandthencancels the PO. Otherwise,thegoodsarefilled,andthenthe supplier generatesaninternal DN document andcreatesa SO document forthedesignatelogisticscompany.Oncethe SO is received,thelogisticscompanycreatesa SL document that isusedforpickingupthegoodsfromthesupplier’s shipping pointandalsodeliversthegoodstothebuyer. Afterthat,thesuppliercreatesan IV document andsendsit to thebuyer.Sometimelater,thebuyerclearsthetotal amount owinginthe IV, consequently,thesuppliermarks the PO as closed.Thispurchasingprocesscompleteswhen Buyer (L1) Supplier (L2) Logistics(L3) Purchase Order(PO) Picking List(PL) ready tofill Filled order checking Quote (Q) created approving approved In stock Shipping Order(SO) In transit arrived Invoice (IV) cleared Shipping List(SL) Queued completed picked created confirmed L1 closed L2 billing L2 canceled accepted filled acquiring L2 L2 delivering L2 ready toship L2 L3 created L2 L2 Out ofstock L2 L1 sent issued L2 L2 Payment (P) approving created sent Delivery Note(DN) prepared transferring L3 scheduled dispatched L3 unpaid L1 L1 rejected L1 on hold L1 clearing L2 Fig. 1. A completeartifact-centricinter-organizationalpurchasingprocess S. Yongchareonetal./InformationSystems47(2015)51–81 53
  • 4. the PO is inthe closed state, the SO is inthe arrived state, and the IV is inthe cleared state. In thecontextofinter-organizationalprocesscollabora- tion, someconcerns—including changeflexibility,change verification,andprivacy—raisedinthetraditionalactivity- centric approach(e.g.,in [85,32]) shouldalsobeconsidered in anartifact-centricsettingasdiscussedbelow. First, organizationsprefertokeepthe freedom of mod- ifying theirinternalprocesseswithoutrevealingtheirlocal changestootherparticipatingparties.Inotherword,once they haveagreedontheoverallprocess,theyshouldhave, possibly,thehighestlevelof flexibility tomodifytheirown local processeswhileremaining autonomous in thecolla- boration [85,82,83]. Forexample,Suppliermaymodify some processingstepsofits PL artifact, andsuchchange should notbeexposedtoBuyerandLogistics.Apartfrom the changeofexistinglocalartifacts,theyshouldbe allowedtoaddnewlocalartifactstotheirprocesscaused by processexpansion/improvement.Forinstance,Supplier may needtoincorporate InventoryList (IL), whichisused for inventorymanagement,totheirprocess.The IL artifact needs tointeractwithexistinglocalartifact PL. Italso implies that IL indirectlycontributestothepartofshared artifact PO through PL. Inordertosupportthechangeof local processes,organizationsneedtoknowwhatthey wanttomodifyandhowsuchmodificationcanbeapplied. Second, stemmedfromthefirstreason,thechangeto anylocalprocessshouldnotaffectthebehaviorofthe overallcollaboration.Thisimpliesthatalocalprocess should always comply with thecontractagreedbyall parties, andthechangemadetolocalprocessesshould not affecttheoverallprocess.Therefore,allpartiesmust ensure thattheirlocalchangesdonotviolatethecollabora- tion contract.Forexample,ifSupplierchangestheir PL artifact byremovingatransition ready_to_f ill-f illed_ order, thenthismayaffectthetransition acquiring-f illed of PO; consequently,Logisticsisnotabledetermine whether thegoodsarereadytobepickedfordelivery.In addition, allparticipatingpartiesshouldalsobeawareof the localchangesthatwillpropagateandeventuallyaffect the overallprocess.Forinstance,iftherearesomechanges made on IL artifact, thentheydirectlyaffectthe PL artifact and eventuallyaffectthe PO artifact. Thisraisestheissueof how wecanguaranteethatlocalchangesmadebyan individualpartydonotleadtoincorrectbehaviorofthe overallcollaboration.Inotherwords,organizationsmust ensure thatchangesintheirlocalprocessesarein compli- ance with whattheyhavepromisedtoprovide. Last, organizationsconcernabouttheir privacy. Inthe collaboration,itisnecessaryfortheparticipatingorganiza- tions torevealcertaindetailsoftheirinternalprocesses among themselvesatanadequatelevelofvisibilityasto establish anoverallpictureofthecollaboration,whichis used asaprocessagreementorcontractamongthem.In our artifact-centricinter-organizationalprocessmodeling approach,thelevelofvisibilitycanbedeterminedbasedon the typeofartifacts.Inotherwords,detailsoflocalartifacts should bekeptinvisibletoexternalpartiesasmuchas possible whiletheyalsosupporttheoverallprocessby means ofdependencyassociationswithsomeprocessing part (lifecycle)ofthesharedartifact(s).Considerthe DN artifact in Fig. 1 for instance.The DN is privatelyusedby Supplier;however,wecanseethatithasdependency associations withthesub-lifecyclesofthe PO (f illed-ready_ to_ship-dispatched) and SO (-created-scheduled- in_transit), whicharesharedartifacts.Simi- larly,both Quote and PL artifacts areusedtosupportsome processingstepsofthe PO artifact. Toachieveprivacy,ifa processingstepofasharedartifactisexclusivelycontrolled by thelocalartifact(s)ofoneparty,thenthatstepshould notbevisibletootherparties.Forexample,the ready_ to_ship state anditsrelatedtransitionsof PO should bekept invisibletoexternalpartiesbecause DN is alocalartifact of Supplier.Apartfromthat,wecanseethatthestep created-scheduled of SO can behiddentoexternalparties as well.Inordertosupporttheprivacyrequirement,an organization needstohaveamechanismtoidentifyan artifact and/oritspartsthatcanbeinvisibletotheother parties. Thisbringsupinthequestionoftowhatextentofa local processcanbekeptprivatewhilenotaffectingthe successful establishmentofthecollaboration. The threemajorconcernsdiscussedabovecallforan approachtoefficientmodelingandchangemanagementof artifact-centric inter-organizationalbusinessprocesses.As previouslydiscussed,inthisarticle,westudyhoworgani- zations canapplytheconceptof view to supportand facilitatethemodelingoftheircollaboratingbusiness processesinanartifact-centricparadigm.Particularly,we borrowtheideaofpublicandprivateviewsapproachto inter-organizationalworkflowswhichisoriginallystudied in activity-centricbusinessprocessmodelingapproaches [85,82,83], andthenexploreitinthecontextofartifact- centric processes. 3. Viewframeworkforartifact-centricinter- organizationalprocesses In thissection,weintroduceanddiscussourview frameworkforartifact-centricinter-organizationalbusiness processes.In Section 3.1, weoverviewourviewframework that aimsataddressingtheaforementionedrequirements for efficientinter-organizationalbusinessprocessescolla- boration.Thenin Section 3.2, weformallydefinetheview frameworkanddiscusshowtouseittomodelinter- organizational businessprocessesfollowedbythediscus- sion ofbehaviorpropertiesin Section 3.3. 3.1.Overview Westartthissectionbyintroducingourviewframe- workformodelingartifact-centricinter-organizational business processes-whichisinfluencedbytheprocess- orientedcontractapproachproposedforservicechoreo- graphy [83]. Ourartifact-centricviewframeworkconsists of thefollowingfourparts:(1)Artifact-Centricbusiness Processmodelforinter-organizationalbusinessprocesses (ACP-imodel), (2)publicviewanditsconstructionmethod, (3) processchangemechanism,and(4)processchange validation. Fig. 2 depicts theoverallpictureoftheframe- workbytakingourmotivatingpurchasingprocessesasan illustrativeexample. 54 S. Yongchareonetal./InformationSystems47(2015)51–81
  • 5. Our viewframeworkisdevelopedbasedonthenotion of privateview and public view proposedby [85]. A private view is usedtocapturethe local processes of eachindividual organization, whilea public view of aparticularcollabora- tion processisan abstract representationoftheoverall processthatisnecessaryforthecoordinationandhidesthe details oftheprivateprocessesofeachorganizationas much aspossible.Organizationscanachieveanefficient collaboration(regardingthethreerequirements)basedon the public–privateviewsapproachbythefollowingfour stepsasillustratedin Fig. 2. (1) Constructacompleteartifact-centricmodelspecifica- tion ofinter-organizationalbusinessprocess (2) Createa public view that isservedasamutualagree- ment, i.e.,contract,ofthecollaboration. (3) Eachorganizationcanchangeandvalidatetheirlocal processes,viatheuseof privateview, withouttheneed for globalverification. First, allparticipatingpartiesspecifyacompletespeci- fication oftheirinter-organizationalbusinessprocessasto achievetheirgoalofthecollaboration.Aspreviouslydis- cussed, thecoordinationamongthemcanbespecifiedby defining allsharedartifactsandtheirinteractionswithlocal artifacts (fromeachparty).WecalltheArtifact-Centric Processmodelofthecompleteinter-organizationalprocess an ACP-imodel. Onceallparticipatingorganizationshave agreedonthecompletemodeloftheartifact-centric collaborationprocess,theyconstructanagreedpublicview of suchprocess.Thispublicviewrevealsonlythenecessary information ofartifactsthatisrequiredtobeusedforthe coordinationamongparties.Intheperspectiveofartifact- centric processmodeling,apublicviewshouldonlycontain sharedartifacts and shouldnotrevealanylocalartifactor the partofasharedartifactthatissupportedbylocal artifact(s). Once thepublicviewisbuiltasacontract,each participating partyhastheresponsibilitytoensurethat its localprocesses(i.e.,responsiblepartsofsharedartifacts and theirlocalartifacts)canprovidewhathasbeen specified inthecontracttootherpartiesinvolvedinthe collaboration.Themostimportantthingsareasfollows:(1) the constructedpublicviewmustbesoundandconsistent with itsbaseprocess;and,(2)whenanorganization modifies itsownpartsofbothtypesofartifactsusedin the collaboration,itmustguaranteethatsuchlocalchanges do notcompromisethecorrectnessoftheoverallcollabora- tion. Wewilldiscusstheconstructionofapublicviewand changemanagementinalocalprocessindetailin Section 4 and 5. 3.2. SyntaxforACP-imodel Here,weformallydefinetheartifact-centricprocess model forinter-organizationalbusinessprocesses,or ACP-i model, whichisanextendedversionoftheACPmodel presentedinourpreviousworks [97,98,95,96,94]. An ACP-i model consistsoffourmaincomponents: roles, artifacts, tasks, and businessrules.Roles defineasetofparticipatingorganiza- tionrolesinthecollaboration.An artifact isabusinessentity or anobjectinvolvedininter-organizationalbusinesspro- cesses. A task is ownedbyoneorganizationinthecollabora- tionandisusedtoperformread/updateoperationsonartifact (s).A business rule is definedbyafamilyofconstraintsin a Condition-Action styletodescribewhichserviceisin- vokedandwhichstateofanartifactischangedunderwhat condition. Definition 1. (Artifact class).Anartifactclassabstractsa group ofbusinessartifactswiththeirattributesandstates. Artifact classC (or artifact if thecontextisclear)isatuple (A; S; sinit ; Sf ) where, A ¼ fa1; a2; …; axg; aiAAð1rirxÞ is aname-value pairedattributevariable, S ¼ fs1; s2; …; syg; siASð1riryÞ is astate, sinit =2 S denotesthepseudoinitialstate, Sf S is asetoffinalstates. ACP-i ACP ACP ACP ACP-i (modified) ACP ACP ACP PUBLIC VIEW pa(ACP-i) private viewchangevalidation public viewconstruction 2 4 inter-org processconstruction 1 changes inlocalprocesses Purchase Order Shipping Order Invoice Buyer Supplier Logistics Picking List Quote Shipping List Payment Deliver Note 3 ′ Fig. 2. Our viewframeworkforartifact-centricinter-organizationalbusinessprocesses. S. Yongchareonetal./InformationSystems47(2015)51–81 55
  • 6. Definition 2. (Artifact schema).An artifactschema, denotedas Z, containsasetofartifactclasses,i.e., Z ¼ fC1; C2; :::; Cxg where CiAZð1rirxÞ is anartifactclass. Wealsodefinesomebasicpredicatesoverschema Z to be used fordefiningbusinessrulesasfollows: def inedðC; aÞ iff attribute aAC:A of artifactofclass C is defined; instateðC; sÞ iff state sAC:S of artifactofclass C is active. Next,wedefineabusinessruletoexpressthecontrol logic ofabusinessprocess. Definition 3. (Business Rule).Abusinessruleregulates which taskcanbeinvokedunderwhatpre-condition.The conditional effectisalsodefinedtowhatpost-condition needs tobesatisfiedafterperformingsuchtask. Business rule r is triple ðλ; β; vÞ where λ and β are thepre-conditionandthepost-condition, respectively.Bothconditionsmaycontainthefollowing twotypesofpropositionsoverschema Z: (1)state proposition(the instate predicate)and(2)attribute proposition(the def ined predicate).Tomakethetarget statedecidable,wedonotallowthepost-conditionto include disjunctioninstatepropositions. vAV is ataskoracompositetask(i.e.,morethanone atomictask); V ¼ fv1; v2; :::; vxg is asetoftasksof which performsoperationsonsomeartifacts C1, C2,…, Cy where CjAZð1rjryÞ. In ordertomaintaintheexistenceofvalidandexplicit state changesofanartifactinbusinessrule r, werequire that thereexistsacoupleof instate predicatesofthat artifact inboththepre-conditionandthepost-condition of r, i.e.,wehavestates sx; syAC:S such that instateðC; sxÞ existsin r:λ and instateðC; syÞ existsin r:β. Thestatechange referstoeitheratransitionfromonestatetoanotherstate, or toitself. Table1 lists somebusinessrulesusedinour purchasingprocessin Fig. 1. Wealsoclassifybusinessrulesintotwotypesbasedon the existenceofthe instate predicateinthepre-andpost- conditions ofabusinessrule.Thefirsttypeonlychanges the stateofonesingleartifact,whilethesecondtype simultaneouslychangesthestatesofmultipleartifacts,i. e., morethanonepairof instate predicates(onepairforone artifact) mustappearinthepre-andpost-conditionsofa single businessrule.Wecallthesecondtype synchroniza- tion (sync)rules as theyareusedforexpressingsynchroni- zation betweenartifacts. Definition 4. (Sync rule).Businessrule r is a sync rule for artifact Cx and artifact Cy if thereexists instate ðCx ; siÞ and instate ðCy ; smÞ in r. and instateðCx ; sjÞ and instateðCy ; snÞ in r:β, where si; sjACx:S and sm; snACy:S. As mentionedabove,asingle sync rule can beusedto synchronizemorethantwoartifacts. Example 1. In Table1, wecanseethatbusinessrules r2 is used toexpressonlythestatechangeofthe PO artifact (conf irmed-accepted), whilebusiness(sync)rule r1 is used tosimultaneouslychangestatesoflocalartifact Quote (approving-approvedÞ and sharedartifact PO (on_hold-conf irmedÞ. Similarly,business(sync)rule r3 is usedforthe synchronizationof DN and SO; andsyncrule r4 is usedfor synchronizingfourartifacts PO, DN, SO, and IV (wheretwo tasks aredefinedinitsaction). Next,wedefine ACP-i model to capturethecomplete specification ofaparticularinter-organizationalbusiness processinthecollaboration.Asdiscussedin Section 2, sharedartifactsshouldbeusedascoordinationmeansin inter-organizationalbusinessprocesses.However,ashared artifact cannotbecompletelymodeledsolelybyasingle Table1 An exampleofbusinessrulesforourpurchasingprocess. r1: Buyerapproves Quote q to confirm PurchaseOrderpo for aselectedsupplier Pre-condition instate(q, approving) 4 instate(po, on_hold) 4 defined(po,OrderID) 4 defined(po.SupplierID) Task approve(q,po) Post-condition instate(q, approved) 4 instate(po, confirmed) 4 defined(po.SubmitDate) r2: Supplier accepts PurchaseOrderpo Pre-condition instate(po, confirmed) 4 defined(po,OrderID) 4 defined(po.SupplierID) Task acceptPO(po) Post-condition instate(po, accepted) r3: Supplier creates Shipping Orderso from DeliveryNotedn Pre-condition instate(dn, prepared) 4 defined(dn.ShipperID) 4 instate(so, init) Task createShipping (dn,so) Post-condition instate(dn, transferring) 4 instate(so, created) 4 defined(so.ShippingID) r4: Supplier dispatchesgoodsfor PurchaseOrderpo that tobeshippedby Shipping Orderso, and simultaneouslyissues Invoiceiv to theBuyer Pre-condition instate(po, ready_to_ship) 4 instate(dn, transferring) 4 instate(so, scheduled) 4 instate(iv,init) Task dispatchGoods(po, dn,so)issueInvoie(po,iv) Post-condition instate(po, delivering) 4 instate(dn, dispatched) 4 instate(so, in_transit) 4 instate(iv,issued) 4 defined(iv.InvoiceID) 4 defined(iv.OrderID) 56 S. Yongchareonetal./InformationSystems47(2015)51–81
  • 7. organization duetoitssharingnature.Therefore,allparti- cipating organizationsmustagreeonboththestructure (data model)andthebehavior(lifecycle)oftheshared artifacts inordertodefinetheACP-imodel.Itisworth noting thatinordertomaintaintheintegrityofashared artifact togetherwithalltheinvolvedtasksandbusiness, we assumethatthenameofitsstateandattributemustbe uniquelyidentifiedforthesamemeanacrosscollaborating organizations. Definition 5. (Organization Role).Wedenote L ¼ fl1; l2; …; lxg for asetof organization roles, where liALð1ixÞ is aroleoforganizationsthatparticipateinan inter-organizationalbusinessprocess. Definition 6. (ACP-imodel).Givenasetoforganization roles Linvolvedinacollaboration,wedefinean ACP-i model, denotedas Π^ ¼ ðZ; V; R; L; γÞ, fortheirinter- organizational businessprocesswhere Z is anartifactschema, V is asetoftasks,and R is asetof business rules, L is asetofparticipatingorganizationroles, γ: Z [ V [ R-2L is arolemappingfunctionfroman artifact class,abusinessrule,orataskontoanorganiza- tion role(s)asfollows: (c) γðCÞ returnsasetofroles fl1; l2; …; lxg, where liALð1rirxÞ is arolethatcanaccess(read/write) artifact CAZ and theownerof C must beamong roles γðCÞ. Notethata sharedartifactC implies that jγðCÞj41, andif C is a local artifact then jγðCÞj ¼ 1. (c) γðvÞ returnsrole lAL of organizationwhoownstask vAV. Notethatataskcanperformread/write operationsoneitherlocalartifactorsharedartifact or both. (c) γðrÞ returnsrole lAL of organizationwhoowns business rule rAR. In addition,wedefinetwoauxiliaryfunctionsover business rules R and artifactschema Z in Π^ function pre_sðr; CÞ returnsasetofstates{s1; s2; :::; sxg whereforsomebusinessrule rAR, state siAC:Sð1rirxÞ occurs inatleastone instate predicate of thepre-conditionof r; and, function post_sðr; CÞ returnsasetofstatesofartifact C appearing inthepost-conditionof r. Givenan ACP-i model, wecanderivealocalACPmodel for anorganization'slocalprocess.Notethatinalocal processofanorganization,theattributesandstatesofits sharedartifactcanbeobtainedfromthe ACP-i model if they arespecifiedinthebusinessrulesoflocalprocess. Definition 7. (local ACPmodel).Given Π^ ¼ ðZ; V; R; L; γÞ be an ACP-imodel, alocalACPmodelofrole lAL can be derivedfrom Π^ , whichisdefinedas Π^ l ¼ ðZl; Vl; Rl Þ where, Zl ¼ fCAZ j lAγðCÞg is alocalartifactschema,suchthat each sharedartifact in Zl contains astate s if andonlyif instateðsÞ appears inthepre-orpost-conditionofa business rulein Rl, i.e., 8CAfCAZl j lAγðCÞ 4 γðCÞ 41g; 8sAC:S; (rARl; sApre_sðr; CÞ [ post_sðr; CÞ; Vl ¼ fvAV j γðvÞ ¼ lg is asetoflocaltasks, Rl ¼ frAZ j γðrÞ ¼ lg is asetoflocalbusinessrules. Example 2. Fig. 3 showstheSupplier'slocalACPmodel which isderivedfromthe ACP-i model of thepurchasing processillustratedin Fig. 1. Wecanseethattheshared artifacts PO and SO representonlythepartsthatbelongto the Supplier'slocalprocess,e.g.,someprocessingstepsof PO (beforethe confirm state)and SO (after the in_transit state) thatbelongtoBuyerandLogistics,respectively,are not capturedintheSupplier’s localACPmodel. Next,wediscussthebehaviorpropertiesofartifact- centric inter-organizationalbusinessprocesses.Ingeneral, it isimportantthatthebehaviorofinter-organizational business processesmustbesoundinordertoguaranteethe reachabilityofdesiredgoalsofthecollaborationand participating localprocesses [83,58,42]. 3.3. BehaviorpropertiesofACPmodelanditsartifacts Weclassifybehavioralpropertiesofartifactsin ACP-i model into intra-behavior and inter-behavior. Theintra- behaviorofanartifactdescribeshowanartifactchanges its statethroughoutitslifecycle.Here,weadoptLabel TransitionSystem(LTS)tocapturethelifecycleofan individualartifact.Second,theinter-behaviordescribes how thelifecycleofoneartifactdependsonthecounter- part ofanotherartifact,anditcanberepresentedas synchronizationdependencybetweenartifacts,i.e.,a sync rule. Here, wegeneralize ACP-i model Π^ to ACPmodel, denotedas Π¼ ðZ; V; RÞ, bydisregardingtherolesoforga- nizations androlemappingof Π^ . Definition 8. (Lifecycleofartifact, )n). Let Ci ¼ ðAi; sinit i ; Si; Sf i Þbe anartifactclassin ACPmodel Π. Picking List(PL) Purchase Order(PO) confirmed acquiring accepted filled canceled ready tofill Filled order checking In stock Out ofstock delivering ready toship Delivery Note(DN) prepared transferring dispatched closed billing Shipping Order(SO) In transit created scheduled Invoice (IV) cleared sent issued unpaid clearing Fig. 3. A localACPmodelforSupplier. S. Yongchareonetal./InformationSystems47(2015)51–81 57
  • 8. A lifecycle of Ci, denotedas ℒCi , canbedefinedasatuple ðS; sinit ; )Þ where, set ofstates S ¼ Si, initialstate sinit ¼ sinit i , statetransitionrelation ) DS Ri Gi S where RiD Π:R is asetbusinessrulesthatareusedtoinduce statetransitionsofartifact Ci such that, 8rAΠ:R; (sx; syACi:S; sxApre_sðr; CiÞ 4sy Apost_sðr; CiÞ-rARi; Gi (guards)isaunionsetofstatepreconditionsofeach business rulein Ri such thateachpreconditionrefer- ences toastateofotherartifactin Π, i.e., Gi¼ [jΠ:Zj j ¼ 1 fCj:sj(rARi; (CjAΠ:Z; sApre_sðr; CjÞ4CjaCig ; Wealsodenote )n for areflexivetransitiveclosureof ). Wewrite si)nsj if state sj can bereachedfromstate si by somesequenceofbusinessrulesin Π:R. Wewritetransition ss) r½gst tomeanthatthestateofthe artifact willchangefromsourcestate ss totargetstate st if business rule r is firedandguard g (state pre-conditionof r) holds. Notethatinaclearcontext,wemayuseshorthand ss ) st without itssuperscription,andmayuseterm artifact for themeanof lifecycleofartifact. Based on Definition 8, given ACPmodelΠ, wecanderive a lifecyclecorrespondingtoanartifactin Π from asetof correspondingbusinessrulesthatareusedtotriggerthe state transitionsoftheartifact.Wecanobtainthelifecycle of anentireprocessbycomposingallartifactsinthemodel. Here, wedefine ACPlifecycle for describingthebehavioral aspect ofanACPmodelconsistingofsynchronizedlife- cyclesofartifacts.Weadaptastatemachinecomposition techniquepresentedin [49] for generatingthelifecycleof ACP.Similartechniqueforprocessmodelgenerationbased on (object)lifecyclecompositionisalsopresentin [46]. Definition 9. (Lifecyclecomposition,composedlifecycle, ). Let ℒi ¼ ðSi; sinit i ; )iÞ, and ℒj ¼ ðSj; sinit j ;)jÞ be two artifact lifecyclesinACPmodel Π. Lifecyclecomposition (i.e., synchronizedproduct)of ℒi and ℒj is denotedas ℒc ¼ℒi ℒj ¼ ðSc; sinit c ;)cÞ where, ScDℒi:S ℒj:S is asetofcomposedstates, sinit c ¼ ðℒi:sinit ;ℒj:sinit Þ is theinitialstate, —)cD Sc Π:R Gc Sc is atransitionrelationwhere Gc is a set ofguards(statepropositions). Now,let g½sx ℒi=stateðℒi; sxÞ denotethatstate sx ℒi in guard g is substituted (denotedbysymbol/)by true or f alse (of statepredicate)dependingonwhetherthelocal stateof ℒi is sx. Wecanformulatetransitionrelation )c of composedlifecycle ℒc , byusingthefollowingthreeinfer- ence rules. ðsx ℒi ; r; g1; sy ℒi ÞA)i ððsx ℒi ; sx ℒj Þ; r; gc; ðsy ℒi ; sx ℒj ÞÞA)c; gc ¼ g1½sx ℒj=stateðℒj; sxÞ ð3:1Þ ðsx ℒj ; r; g2; sy ℒj ÞA)j ððsx ℒi ; sx ℒj Þ; r; gc; ðsx ℒi ; sy ℒj ÞÞA)c; gc ¼ g2½sx ℒi=stateðℒi; sxÞ ð3:2Þ ðsx ℒi ; r; g1; sy ℒi ÞA)i4ðsx ℒj ; r; g2; sy ℒj ÞA)j ððsx ℒi ; sx ℒj Þ; r; gc; ðsy ℒi ; sy ℒj ÞÞA)c; gc ¼ g1½sx ℒj=stateðℒj; sxÞ 4g2ðsx ℒi=stateðℒi; sxÞÞ ð3:3Þ Rule (3.1) and Rule (3.2) are appliedwhenbusinessrule r is firedononlyindividuallifecycle ℒi and ℒj, respectively. Rule (3.3) is appliedwhen sync ruler is firedonboth lifecycles ℒi and ℒj. Asthethreeinferencerulesapply the substitutionofstateconditionsoftwolifecyclesinthe composition, referencestoexternallifecyclearenot replaced. Example 3. Fig. 4 showsthecompositionbetweenthe lifecycleofartifact C1 and thelifecycleofartifact C2. The label ri½g attached toatransitionmeansthatthetransition is firedwhenboththeattributepropositioninthepre- condition ofbusinessrule ri holds andallstatepropositions (of externallifecycles)in g hold. Wedenotethecounter stateconditionof C:sx by symbol C:sx in theguard.We can alsoseethatstateconditionsreferencingtoartifacts C3 and C4 remaininthecomposedlifecyclebutindifferent forms, whichdependonthetransitiontheybelong. Now,wecandefinethelifecycleofACPbyusing lifecycle composition. Definition 10. (ACPlifecycle).GivenACPmodel Π, a (ACP) lifecycle of Π, denotedas ℒΠ, canbegeneratedbyitera- tivelyperforming lifecyclecomposition of everyartifactin Π. Notethat lifecyclecomposition is associativeandcom- mutative,i.e., ℒi ℒj ℒk ¼ ℒi ðℒj ℒkÞ ¼ ðℒi ℒjÞ ℒk and ℒi ℒj ¼ℒj ℒi. Therefore,thefinal resultofthecompositionofasetoflifecyclesisnot impactedbytheircompositionorder. Next,wedefine soundness propertytodescribeadesired and correctbehaviorofartifactlifecycleandtheprocess. Definition 11. (Safe, goal-reachable,andsoundlifecycle). GivenACPmodel Π and lifecycle ℒ¼ ðS; sinit ; )Þ, we Fig. 4. An exampleoflifecyclecomposition(takenfrom [96]). 58 S. Yongchareonetal./InformationSystems47(2015)51–81
  • 9. define asetoflifecyclestates S ¼ ℒ:S [ fsinitg and asetof final states Sf S. Lifecycle ℒ is saidtobe: safe iff thereexistsbusinessrule rAΠ:R such that r induces oneandonlyonetransitionin ℒ, i.e., Π:R; ðsx; r; g; syÞA ) ðsm; r; g; snÞ=2 ) goal-reachable iff, foreverynon-finalstate s of ℒ, s can be reachedfromtheinitialstateand s can reachoneof the finalstatesof ℒ, i.e., 8sAS Sf ; (sf ASf ; sinit) n s s) n sf sound iff ℒ is safe and goal-reachable. Notethatthe goal-reachability propertyimplies deadlock-freeandconnectedlifecycleofanartifact(ora process). Nowconsiderthecaseofasharedartifactdefinedina local ACPmodel.ItisalwaystruethatthelocalACPmodel is not goal-reachable as thelifecycleofsharedartifactis partially modeledandcanbenon-terminated.However, when integratingallthedifferentpartsofasharedartifact from eachorganization,thecompletelifecyclemustbe goal-reachable. In thisarticle,wedonotfocusonthetask-level information, i.e.,thespecificationoftaskisomitted.How- ever,thespecificationoftaskinartifact-centricprocess modeling approachcanbedefinedinthespiritofsematic web-servicesspecifiedinOWL-Sproposal[2003]—that isin a formofInput,Output,Pre-condition,andEffect(IOPE). The pre-conditionandconditionaleffectofataskshould conform tothepre-conditionandpost-condition,respec- tively,ofthebusinessrulethatspecifiesthetaskinits action. Adetaileddiscussionofsemanticweb-services togetherwiththeuseofbusinessrules(e.g.,SBVR [69]) for modelingartifact-centricbusinessprocesscanbefound in [5,19]. Wealsorestrictour soundness discussion onlyon the lifecyclebehaviorofartifactswhilediscussionsand formal approachestodataverificationofartifact-centric business processes(somecallartifactsystems)canbe found inseveralexistingliteratures,e.g., [54,31,29,25,21]. 4. Behavior-consistentpublicviewconstruction In thissection,wefirstintroducethedefinitionofpublic view anddiscussthebehaviorconsistencybetweena createdviewanditsbase ACP-i model. Followingthat,we presentanabstractionmethodtoconstructpublicviews preservingthebehaviorconsistency,anddiscusshow synchronizationdependencybetweenartifactsisaffected by suchabstraction.Finally,weproposeatechniquefor constructing theminimalpublicviewforagivencollabora- tion andformalizeitintoanalgorithmofprotectingthe information (i.e.,localprocesses)ofeveryorganizationat the highestlevelofprivacyandautonomyduringview construction. 4.1.Publicviewconstructionandbehaviorconsistency Generallyspeaking,constructingapublicviewofa particular collaborationshouldtakeintoaccountallinter- action betweenparticipatingorganizations,particularly the processingandexchangesofsharedartifacts.Aslife- cycleisthemainmechanismofanartifactforspecifying its dynamicbehaviors,andtheassociation/coordination betweenartifactsisalsobasedonthestatesinsidelife- cycles,wetakeitnecessarytoconstructpublicviewsbased on lifecyclesby blinding off the privatepartwhilepreser- ving theglobalcoordination.Weproposeourabstraction techniquetohelporganizationsrevealonlytheirnecessary steps thatarerequiredtocompletethecollaboration. Shared artifactsarethemainconcernastheyareusedby more thanoneorganization.Thisrequirementraisesthe questionofhowtodecidewhichprocessingpartofa shared artifactshouldbeabstractedsuchthatthe abstractedpartdoesnotaffectthecoordination.Basedon this requirement,weobservethatthepartofashared artifact interactingwithlocalartifact(s)andaprocessing step ofthesharedartifactownedbyasingleorganization should beabstracted.Asdiscussedin Section 3.1, thisis because suchpartisdeemedaslocalprocessingofthe shared artifactanditshouldnotberevealedtoother parties (outofinterestandprivacyconcern).Assuch,a constructedpublicviewshouldhavealllocalartifactsof every organizationhidden,andhaveallpartsofshared artifacts thatinteractwithlocalartifactsabstracted. Next,wediscussthe ACPabstractionmethod for con- structing publicviewsthatderivefromanunderlying ACP-i model. Themethodisdiscussedbasedonthefollowingtwo points: Abstractionofnon-synchronizedpartoflifecycle. Itiseasy to understandthatapartofthelifecycle(calledlifecycle fragmentorfragment)isconvertedintoeitherasingle stateorasingletransitionduringthelifecycleabstrac- tion. In Section 4.2, wewillintroduceageneraltechni- queforstate/transitionabstractionthatcanbeapplied directlytoalifecyclefragmentofindividualartifact. Abstractionofsynchronizedpartoflifecycle. Apartfrom the isolatedabstraction,wealsoconsiderhowmultiple synchronizedfragments(via sync rules) ofdifferent artifact lifecyclescanbeabstracted.Weobservethat the abstractionofonesynchronizedendcallsforthe abstractionofthecorrespondingendduetoconsistency preservation.Bothabstractedfragmentsoftwolife- cyclesmuststillsomehowbeeithercorrectlysynchro- nized ornone.Ifanentirelifecyclesynchronizingwitha fragmentofanotherlifecycleistobeabstracted,the former istotallyunsynchronizedandshouldbeconsid- ered asanembeddedfragmentoftheabstractedlife- cycleofthelatter. Section 4.3 discusses thisissuein more detail. Based onan ACP-i model of inter-organizationalpro- cesses, organizationscanconstructtheirpublicviewby abstractingtheirsharedartifactsandtheirlocalartifacts. S. Yongchareonetal./InformationSystems47(2015)51–81 59
  • 10. Wedefineapublicviewbyapplyingthe abstractionfunc- tion to mapacomplete ACP-i model toitspublicview. Definition 12. (Public viewandACPabstraction).Given ACP-imodel Π^ ¼ ðZ; V; R; L; γÞ, paΠ^ ¼ ðZp; Vp; Rp; Lp; γpÞ denotesapublicviewof Π^ , where Zp; Vp; Rp; Lp; γp are sets ofabstractartifacts,abstractservices,abstractbusiness rules, organizationroles,androlemappingfunctionofthe public view,respectively. ACPabstractionfunction is defined as pa: Zp [ Vp [ Rp [ Lp [ γp-Z [ V [ R [ L [ γ which maps paΠ^ to Π^ . pa is atotalmappingfunctionwith the followingconditions: paΠ^ and Π^ haveidenticalsetsoforganizationroles, paΠ^ contains everyabstractsharedartifact,i.e., 8CAZ; (CiAZp; γðCÞ 41-ðpaðCiÞ 4 γpðCiÞ 41Þ each abstractsharedartifactin paΠ^ has anidenticalrole set toitscorrespondingconcreteartifactin Π^ , i.e., 8CiAZp; (CAZ; paðCiÞ ¼ C - 8lAγpðCiÞ; lAγðCÞ every ruleandtaskdefinedin Π^ must beprojectedto some abstractruleandabstracttask,respectively,defined in paΠ^ , i.e., 8riAR; (rARp; paðrÞ ¼ ri 48viAV; (vAVp; paðvÞ ¼ vi if anentirelocalartifactin Π^ is hiddenin paΠ^ then the followingholds: instate rolemapping γp for abstractartifacts,tasks,andbusiness rules holds 8CiAZp; γpðCiÞDL48viAVp; γpðviÞAL48riARp; γpðriÞAL Example 4. Fig. 5 illustratesanexampleofanagreed public viewthatcanbeconstructedfromthepurchasing processintroducedin Fig. 1. Wecanseethatthepublic view isachievedbyabstractingthreesharedartifacts PO, SO, and IV and removingalllocalartifactsofeachorganiza- tion. Itispossiblethat,duetoaimingtoachievehigher level ofabstraction,abstractstatesmaybeintroducedin the publicview,e.g., approving, supplying, unpaid are abstractstatesin Fig. 5. Notethelabelingofabstractstateisdonemanuallyafter lifecycleabstraction.Participatingorganizationsshould agree onchoosingareasonablenameofanabstractstate in theirpublicview.Onceapublicviewisconstructed based onitsunderlying ACP-i model, itisveryimportantto ensure thevalidityofthepublicview.Asalreadymen- tioned, inthisarticleweonlyfocusonthebehavior perspectiveofACPmodels,andthuswedefinethevalidity of publicviewsintermofthebehaviorconsistency betweenaviewanditsbasemodel.Inourrecentwork [96], wehaveproposedabehaviorconsistencychecking approachcalled B-consistency tocheckwhetheraspecia- lized processderivedfromthebaseprocessisconsistently observable.Here,weusethe B-consistency notion tocheck whether thebehaviorofapublicviewisconsistentwithits underlyingmodel. Let ℒy be thelifecycleofapublicviewconstructedby abstractingitsbase ACP-i lifecycle ℒx. Wearetocheck whether abstractlifecycle ℒy has consistentbehaviorswith its baselifecycle ℒx. Intuitively,ifeachfiringsequenceof transitionsin ℒx, disregardingstatesandtransitionsinan abstractedL-fragment,isobservableasthesamesequence as of ℒy, ℒy is saidtobe behavior consistent with ℒx. Our B-consistency relationbetweentwolifecyclesisdefinedwith the helpofbi-simulationequivalenceinprocessalgebra [9]. By abstractingalifecyclefragmentintoa silent (τ) action, we canapplyweakbi-simulationbetweenalifecycleand its abstractedone. Definition 13. (B-consistent, C). Let ℒx ¼ ðSx; sinit x ;)xÞ and r:be twolifecyclesand Sxy ¼ Sx Sy be asetofstates that commonlyexistin ℒx and ℒy. ℒy and ℒx aresaidto be B-consistent (denotedas ℒyCℒx) iff, 8si; sjASxy; (ðsi; r; g; sjÞA)x; 8skASy Sxy; si)n ysk-sk)n ysj 8si; sjASxy; ∄ðsi; r; g; sjÞA)x; 8skASy Sxy; :ðsi)n ysk -sk)n ysjÞ Notethatasdefinitionof B-consistency is generalized,it can beusedforcheckinganytypeofalifecycle(i.e.,ACP lifecycles,artifactlifecycles,orcompositelifecycles). Example 5. The lifecyclein Fig. 6(a) isnot B-consistent with the lifecycleinFigure(b),i.e., . Thisisbecause, in somefiringsequencesoftransitionsinthelifecyclein Fig. 6(b), state a can reachstate c(throughstate x2) without passing state b; and,state a can reachitselfviastate x4 without passingstate b. Incontrast,wecanseethat ℒaCℒc and ℒaCℒd in Fig. 6(c) and Fig. 6(d), respectively. 4.2. Abstractionofnon-synchronizedlifecycle Next,weintroducenotionoflifecyclefragment(called L-fragment) forcapturingthepartofanartifactlifecycleto be abstractedinthepublicview.OnceanL-fragmentis identified inalifecycle,weapplytheabstractionfunction tomap L-fragment to aspecifiedabstractstateorabstract transition(s)inanabstractACPmodel(i.e.,apublicview). Definition 14. (LifecyclefragmentorL-fragment, Π^ ¼ ðZ; V; R; L; γÞ, find_Lf ). GivenACPmodel Π, L-fragment γðCÞ of lifecycle ℒx is anonemptyconnectedsub-lifecycleof ℒx. Itcanbedefinedas ℓℒx ¼ ðS;); )in; )out Þ where, Purchase Order(PO) closed delivering confirmed Shipping Order(SO) In transit arrived Invoice (IV) cleared unpaid billing approving canceled supplying filled issued Fig. 5. An exampleofanagreedpublicviewbasedontheprocessin Fig. 1. S. Yongchareonetal./InformationSystems47(2015)5160 –81
  • 11. SDℒx:S fsinitg [ Sf is anon-emptysetofstatesof ℓℒx , where Sf is asetof final statesof ℒx, ) DS RℒxGℒx SDℒx: ) is asetoftransitionsof ℓℒx , where Rℒx and Gℒx aresubsetsofbusinessrulesand guards,respectively,definedin ℒx, )in ¼ℒx:) ððℒx:S SÞRℒx GℒxSÞÞ is asetof entry tran- sitions into ℓℒx , )out ¼ℒx:) ðS RℒxGℒx ðℒx:S SÞÞ is asetof exit transi- tions from ℓℒx . Wedenote ℓ!ℒc (or ℒcgℓ) if ℓ is anL-fragmentof lifecycle ℒc of artifact C. Inaddition,wealsodefine function f ind_Lf ðC; SÞ to returnL-fragment ℓ if ℓ consists of asetofstates S of artifact C such that ℓ:S ¼ S and ℓ!ℒc; otherwise,return null if suchanL-fragmentcannotbe found. Noticethat,givenavalidinputsetofstatesofanartifact, thereisonlyonecasethat f ind_Lf returns null – that is when thesetcontainsonlyan init stateand final state(s)of that artifact. ToensureL-fragment ℓℒx is correctlyformedbya connectedsub-lifecycleofitsentirelifecycle ℒx, weconfine that foreachstate s in ℓℒx :S, thereexistsasequenceof transitionsfromanentrytransitionin )in to s and from s to anexittransitionin )out. soundness propertycanalso be appliedto L-fragment providingthatanyL-fragmentisa sub-lifecycleofitsentirelifecycle. Notethatbasedontheconditionofentryandexit transitionsofL-fragment,anentry/exittransitionmustbe fired from/toastateinsidetheL-fragment.However,there is norestrictiononthenumberofentrystatesandexit statesofL-fragment. Next,weidentifyaspecifictypeofL-fragmentbasedon its atomicitypropertywhichisrestrictedbymeansof single-entry-single-exit(SESE)fragmentoflifecycle.How- ever,weadaptittoourL-fragmentdefinitionbyallowing the structureofmultiple-entrytransitionsandmultiple- exittransitionsinsteadofsingleentryandsingleexitstates. Here, wedefinesuchL-fragmentas AtomicL-fragment. Definition 15. (AL-fragment,NAL-fragment).Given L-fragment ℓℒx ¼ ðS;); )in; )out Þ of lifecycle ℒx, ℓℒx is called an AL-fragment iff, allentrytransitionsin )in have the samesourcestateandallexittransitionsin )out have the sametargetstate.Otherwise, ℓℒx is classifiedas NAL- fragment (non-atomicL-fragment). Example 6. Fig. 7 showsexamplesofdifferenttypesofL- fragments.In Fig. 7(a), L-fragments ℓ1 and ℓ2 havesingle entry state s1 and singleexitstate s4; therefore,both ℓ1 and ℓ2 are AL-fragments.WecanseethatL-fragments ℓ3, ℓ4, and ℓ5 in Fig. 7(b) areNAL-fragmentsas ℓ3 has multiple exitstates,andboth ℓ4 and ℓ5 havemultipleexitstatesand multiple entrystates. Based onabovenotionofL-fragment,givenacomplete ACP-i model wecanusethesetwoabstractionmechanisms togetherwithL-fragmentstoconstructanabstract ACP-i model whichcanbeusedtorepresentthepublicviewofits base model.Next,wediscussthetwooutputtypesofan abstractionoperation: abstracttransition and abstractstate. In mostcases,anabstractionoperationshouldyieldan abstracttransitionasthetransitionrepresentsanatomic and uninterruptablestepfromonestatetoanotherstatein the lifecycle.However,insomecases,wemayseethatan abstractstateisyieldedinstead.Forinstance,considerthe PO artifact inthepurchasingprocessshownin Fig. 1 and compare itwithitspublicviewin Fig. 5, wecanseethatthe approving state inthepublicviewisanabstractstateofa lifecyclefragmentconsistingofstates created and on_hold. There aretwopossibleunderlyingreasonssupportingthis case. First,anabstractstateisspecifiedinthedesignofan organization orinthemutualagreementofthecollabora- tion. Second,afragmentcannotbeabstractedintoasingle abstracttransition—this isbecausethefragmentisnot atomic(i.e., NAL-fragment). Theresultofmultipleabstract transitionsforL-fragmentmakesitdifficulttodecideon drawingaprojectionfromapartofthefragmenttoan abstracttransition.Therefore,apossiblesolutionistouse an abstractstatetoremovetheambiguityofwhatis abstractedinsuchtransitions.Here,weshowdifferent cases ofL-fragmentabstractionin Fig. 8. Example 7. Fig. 8(a) and Fig. 8 (b) showtheabstractions for abstracttransitionofL-fragmentsinartifact A1 and artifact A2, respectively,while Fig. 8(c) showsacaseof abstractionforabstractstateofartifact A2. Wecanseethat both L-fragmentsin A1 and A2 are AL-fragments. However, consider theL-fragmentofartifact A3 in Fig. 8(d) whichis an NAL-fragment (due tomultipleexitstates s4 and s5), its abstractiondoesnotresultinasingleabstracttransition.As alreadydiscussed,multipleabstracttransitionsareambig- uous. Therefore,totacklethisissue,theabstractionneeds to resultinanabstractstateinstead,whichisshownin Fig. 8(e). Notethatwediscussthisabstractstateindetailin Definition 17 with anexamplein Fig. 9. Next,wedefine lifecycleabstractionmapping function to map anabstractelementinabstractlifecycleontoastateor transitioninthebaselifecycle,whichisshownin Definition a b c d a b c x d x x a b c d x x x x x x a c x x d x l1 l2 l3 l4 Fig. 6. Examples ofabstractlifecycle(a),itsbaselifecycles(b),(c)and(d)(takenfrom [96]). S. Yongchareonetal./InformationSystems47(2015)51–81 61
  • 12. 16. Thenin Definition 17, wedefinetwo abstractionfunc- tions that areusedtoconstructanabstractlifecycle.These functions arebasedonthepreferredoutputoftheabstrac- tion—that isabstractstateorabstracttransition. Definition 16. (Lifecycleabstraction(la)).Letartifactlife- cyclele ℒ0Cx ¼ ðS0; sinit ; )0Þ in anabstractACPmodel Π0be an abstractlifecycleofartifactlifecycle ℒCx ¼ ðS0; sinit ; )Þ in abaseACPmodel Π. Wedefine lifecycleabstraction mapping function laΠ0-Π ℒ0Cx -ℒCx : S0[ )0 -S[ ), where laΠ0-Π ℒ0Cx -ℒCx is atotalfunctionthatmapsanabstracttransition in )0 and anabstractstatein S0 onto astateanda transitionin ℒCx . Wewrite la without itssuperscription or itssubscriptionifacontextisclear. Definition 17. (Abstractionfunctionsforabstracttransi- tion (la_tran)andforabstractstate(la_state)).LetL- fragment ℓCx of artifactlifecycle ℒCx ¼ ðS; sinit ; )Þ to be abstracted.Wecanabstract ℓCx into asetof abstract transitions and an abstractstate (if applicable)in abstract lifecycle ℒ0Cx ¼ ðS0; sinit ; )0Þ via thetwofollowingabstrac- tion constructionfunctions: function la_tranðℒCx ; ℓCx Þ returns ℒ0Cx by abstracting ℓCx into asingle abstracttransition, function la_stateðℒCx ; ℓCx ; s0Þ returns ℒ0Cx by abstracting ℓCx intosingle abstractstates0 and corresponding abstracttransitionsrelatedto s0. Functions la_tran and la_state can beexpressedby lifecycleabstractionmappinglaℒ0Cx -ℒCx as follows. (a) Let ℓCx be an AL-fragment with entrystate si and exit state sj. ℓCx can beabstractedintoabstracttransition C1 s1 s4 s2 s3 C2 s1 s4 s2 s3 C4 s1 s4 s2 s6 s5 l1 l2 l4 C3 s1 s4 s5 s3 l3 C5 s1 s4 s2 s3 s6 s5 l5 s2 s3 Fig. 7. Examples ofL-fragmentsandNAL-fragments. A’2 s1 s4 A2 s1 s4 s2 s3 A’’2 s1 s4 A2 s1 s4 s2 s3 sx A’’3 s1 s4 A3 s1 s4 s2 s3 sx s5 s5 A’1 s1 s4 A1 s1 s4 s2 s3 A’3 s1 s4 A3 s5 s5 s3 s1 s4 s2 Fig. 8. Abstractionsforabstracttransitionsandabstractstates. A1 A’1 s1 s4 sx s6 s5 A1 s0 s7 A1 s0 s7 OR sy s1 s2 s3 s5 s0 s7 s4 s6 A B Fig. 9. An abstractiononanexpandedNAL-fragment. S. Yongchareonetal./InformationSystems47(2015)5162 –81
  • 13. si)0Sj by applyingfunction la_tran. Wehave ℒ0Cx ¼ la_tranðℒCx ; ℓCx Þ such that, laðsiÞ¼s i 4 la s j ¼ sj, 8sAℓCx :S; laðsi )0 sjÞ ¼ s, 8)lAℓCx : ); laðsi )0 sjÞ¼)l. (b) Let ℓCx be an AL-fragment with entrystate si and exit state sj. ℓCx can beabstractedinto abstractstates0AS0 with asetof abstracttransitions ) 0z by applying function la_state. Wehave ℒ0Cx ¼ la_stateðℒCx ; ℓCx ; s0Þ such that, 8sAℓCx :S; laðs0Þ ¼ s, 8)lAℓCx :) ðℓCx :)in [ ℓCx :)out Þ; laðs0Þ¼)l, 8)nAℓCx :)in; laðsi )0 s0Þ¼)n, 8)oAℓCx :)out ; laðs0 )0 sjÞ ¼)o. (c) Let ℓCx be an NAL-fragment with asetofentrystate Sen and exitstate Sex. ℓCx can beabstractedinto abstract state s0AS0 with asetof abstracttransitions ) 0z by applyingfunction la_state. Wehave ℒ0Cx ¼ la_stateðℒCx ; ℓCx ; s0Þ such that, 8sAℓCx :S; laðs0Þ ¼ s, 8)lAℓCx :) ðℓCx :)in [ ℓCx :)out Þ; laðs0Þ¼)l, 8smAℓCx :Sen; (snAℓCx :S; sm ) s0 ) 0z-laðsm ) s0Þ ¼ sm ) sn, 8soAℓCx :Sex; (spAℓCx :S; s0 ) so ) 0z-laðs0 ) soÞ ¼ sp ) so. It isworthmentioningthatanabstractbusinessrule, which firesanabstracttransition,containsno defined predicates.Ifabusinessruleinabaselifecycleisabstracted in anabstractlifecycle,thenany defined predicateinthe rule shouldberemovedfromtheabstractedruletoavoid overrestrictiononartifactattributesthatmayprohibit firing theabstracttransition. Example 8. Fig. 8(a) and Fig. 8(b) showtheabstract transitionsresultedfromapplyingfunction la_tran to AL- fragments,while Fig. 8(c) showstheabstractstategener- atedfromfunction la_tran on anAL-fragment.Ontheother hand, Fig. 8 (e) showstheabstractstateandrelatedabstract transitionsasanoutputofabstractionbyapplyingfunction la_state on aNAL-fragment. Next,wedefine B-consistent abstraction based onthe lifecycleabstractionmapping,andthenweshow,in Theorem 1, thatapplyingeitherfunction la_tran or func- tion la_state togenerateanabstractlifecyclefromitsbase lifecyclealwayspreserves B-consistency. Definition 18. (B-consistentabstraction, Cla). Letartifact lifecycle ℒ0Cx ¼ ðS0; sinit ; )0Þ in abstractACPmodelΠ0 be an abstractlifecycleofartifactlifecycle ℒCx ¼ ðS; sinit ; )Þ in ACPmodel Π based on lifecycleabstractionmapping func-tion laΠ0-Π ℒ0Cx -ℒCx . If ℒ0CxCℒCx , thenwesaythat ℒ0Cx is a B-consistent abstraction of ℒCx , denotedas ℒ0CxClaℒCx . Theorem1. (Non-synchronizedlifecycleb-consistent abstraction). Let lifecycle ℒ0Cx be anabstractlifecycleof ℒCx by abstractingL-fragment ℓCx where ℓCx!ℒCx . Given ℒ0Cx ¼ la_tranðℒCx ; ℓCx Þ, we have ℒ0CxClaℒCx . Correspond- ingly, ℒ0CxClaℒCx holds for ℒ0Cx ¼ la_stateðℒCx ; ℓCx ; s0Þ where s0Aℒ0Cx :S. Wecanprove Theorem1 by applying B-consistency checkingtothecomparisonbetweeninputlifecycleand output (abstract)lifecyclebasedonthethreeusecasesof those twoabstractionfunctions.Considerthefirsttwo cases (a)and(b)in Definition 17. Weknowthatabstracting an AL-fragment alwaysproducesasingleabstracttransition or abstractstate(withsingleentrytransitionandsingleexit transition)thatstillpreservestheatomicityofthefragment to beabstracted.Thus,theoutputlifecycleisconsistent with itsbase.Forthethirdcase(c)in Definition 17, we generateasingleabstractstatewiththerestrictiononits abstractentrytransitionsandexittransitions.Similarto case (b),theabstractstaterepresentstheinternalbehavior of theabstractedfragmentanditisatomic. Please notethat,in Definition 17(c), weuseabstract states toallowtheabstractionof NAL-fragments; however, this abstractionhastomeetcertainconditionsforthe abstractentryandexittransitionsoftheabstractstate. For instance,considertheabstractionin Fig. 8(e). Entry transition s1)sx abstractstwooriginalentrytransitionsof its fragment(s1)s2 and s1)s3) andexittransition sx)s4 abstractstwooriginalexittransitions(s2)s4 and s3)s4). Nevertheless,wecanseethatexittransition sx)s5 should abstractonlyforoneexittransition s3)s5. Alternatively,if we donotwanttohaveanabstractstateastheresultof abstractionofNAL-fragmenttodecidetheappropriate trigger conditionsfortransitionsduetotheaboveaddition mechanism, thenwemayexpandtheNAL-fragmenttillit satisfies theconditionofAL-fragment(ifpossible).This requiresthemodelertofindapossibleAL-fragmentfrom the expanded-boundaryofNAL-fragment.Ifsuchan expandedfragmentcannotbefound,themodelermay decide toimplementanabstractstatefortheabstraction. Taketwoabstractions(A)and(B)in Fig. 9, forexamples.On one hand,abstraction(A)on NAL-fragment consisting of states s2 and s3 can resultinabstractstate sx and four abstracttransitions.Ontheotherhand,withalternative abstraction(B),theboundaryofthefragmentkeeps expandingtillitsatisfiestheconditionofAL-fragment. Wehavetheexpandedfragmentthatcoversallstatesand transitionsbetweenstates s0 and s7. Then,wecanabstract the fragmentintoeitherabstractiontransition s0)s7 or abstractstate sy. Next,wepresentanalgorithmtofindtheminimalAL- fragment ofaninputL-fragment.Thealgorithmgetsan inputfragmentandexpandsitsboundaryuntilthe expandedfragmentsatisfiestheconditionofAL-fragment (in Definition 15). TheoutputAL-fragmentisminimalasin an iterationofthealgorithm,itfindsthenearestsourceand target statesandaddsthemintothefragment(Lines8and 10).IftheinputisalreadyaqualifiedAL-fragmentthenthe fragment itselfisreturned(Line11).Otherwise,ifthe function cannotfindavalidAL-fragment,thenreturns null. Algorithm1.(function find_minAL). Finding aminimal AL-fragmentfromanL-fragment Input: L-fragment ℓCx of artifact Cx in ACP-i model Π^ . Output: AL-fragment ℓ0Cx if found;otherwise null is returned. 1. ℒyℓ0Cx ¼ ℓCx ; 2. repeat 3. Sen ¼a setof entry states of ℓ0Cx ; 4. Sex ¼a setof exit states of ℓ0Cx ; 5. if (jSen j 41 or jSex j 41) 6. then 7. if jSen j 41 then S. Yongchareonetal./InformationSystems47(2015)51–81 63
  • 14. 8. ℓ0Cx includes allstatesin Sen and theirrelated exit transitions; 9. if jSex j 41 then 10. ℓ0Cx includes allstatesin Sex and theirrelated entry transitions; 11. else if (jSen j ¼ 1 and jSex j ¼ 1) then return ℓ0Cx ; 12. until Sen ¼ ∅ and Sex ¼ ∅ 13.return null; Based on Theorem 1 along withthehelpof f ind_minAL function forNAL-fragmentabstraction,wecanconstructa consistentpublicviewofanindividualartifactlifecycleby abstractingitsnon-synchronizedpartofthelifecycle.Next, we discusshowafragmentofalifecyclethatsynchronizes with afragmentofanotherlifecyclecanbeconsistently abstracted. 4.3. AbstractionofsynchronizedL-fragments In thissection,wediscusshowthesynchronization(via sync rules) betweentwolifecyclescanbeabstractedin detail. Technically,weneedtoanswerthefollowing questions. What istheresultofabstractinganL-fragment(ora whole) ofonelifecyclesynchronizedwithanL-fragment of anotherlifecycle? What istheconditionthatmakesanabstractsynchro- nization consistentwithitsoriginalsynchronization? Here, weusethesameL-fragmentabstractionmethod to abstractthesynchronizedpartswiththeconsideration of synchronizationstructureandbehavior.However,in order tocapturesynchronizationdependenciesbetween lifecycles,weneedtoextendthedefinitionofL-fragment for asynchronizedregion(called S-region) whichrepre- sents synchronizedL-fragmentsbetweenlifecycles(called SL-fragments). Withatargetfragmentofanartifactandthe sync rulesusedwithinthefragment,wecanidentifya counter-synchronizedpart(s)oftheotherartifact(s)thatit interactswith. Definition 19. (Sync ruleforsynchronizedL-fragments, φ). GivenACPmodel Π, asetofsyncrulesthatisusedwithin two synchronizedL-fragments ℓx and ℓy can bedefinedas follow: φðℓx; ℓyÞ ¼fr AΠ:Rj (ðsi; r; g; sjÞA ℓx: ) 4(ðsm; r; g; snÞAℓy: )g It isnotedthatasyncruleistransitive,i.e., (rAφðℓx; ℓyÞ φðℓy; ℓzÞrAφðℓx; ℓzÞ. Definition 20. (SL-fragmentandS-region).GivenACP model Π, wedenote ω¼ ðΓ; Rsync) fora synchronization region (S-region) where, a setofsynchronizedL-fragments Γ¼ fℓC1 ; ℓC2 ; …; ℓCx g, ℓCi AΓð1rirxÞ is asynchronizedL-fragment, called as SL-fragment, of artifactlifecycle ℒCi ðCiAΠ:ZÞ, !RsyncDΠ:R is asetof sync rules that isusedto synchronizetransitionsbetweenL-fragmentsin Γ such that, 8rAΠ:R; (ℓCi ; ℓCj AΓ; 8rAφðℓCi ; ℓCj Þ; rARsync Example 9. In Fig. 10(a), wehaveS-region ωa with SL- fragments l1 of artifact A1 synchronizedwith SL-fragmentl2 of artifact A2 via syncrules r1 and r2. In Fig. 10 (b), S-region ωb has two SL-fragmentsl3 and l4 with syncrules Rsync ¼ fr1; r2; r3g. Noticethatsyncrule r4 is excludedfrom ωb as itisnotusedforthesynchronizationbetween l3 and l4. Next,westudytheatomicitypropertyofS-regionby determiningthecomposabilityofcontainedSL-fragments and theboundnessoftheirsynchronizationbehavior. 4.3.1.AtomicityofS-region Here, weproposeafragmentalcompositiontechnique tochecktheatomicityofS-region.Firstweseethatthe composition betweentwosynchronizedL-fragments resultsinacompositeL-fragment.Thenwecanapply atomicitycheckingtothecompositeL-fragment.Assuch, we needtoobservetheconditionsforSL-fragmentsthat make thecompositeL-fragmentatomic.Now,wedefine compositeS-region based on lifecyclecomposition (in Definition 9). Definition 21. (CompositeS-region).GivenACPmodel Π, let S-region ω¼ ðfℓCx ; ℓCy g; Rsync) ofL-fragment ℓCx of artifact CxAΠ:Z and L-fragment ℓCy of artifact CyAΠ:Z where ℓCx and ℓCy aresynchronizedviabusinessrules RsyncRsync. The compositeS-region of ω, ω¼ ℓCx ℓCy , istuple ðS;); )in;)out Þ where eachsetelementin ω has thesame definition correspondingtotheelementofL-fragment,i.e., inω can beconsideredasanL-fragment(orSL-fragment if thecompositefragmentstillhassomesynchronizationto otherfragmentofdifferentartifact). It isnotedthata compositeS-region is consideredasa sub-lifecycleofthecompositionbetweentwoentirelife- cycles.Tohavea(minimalandsufficient)completeviewof the compositionwedrawadashedarrowforatransition betweenacompositestateexcludedfromtheS-regionand a compositestatethatisanentryorexitstateoftheS- region,asexemplifiedin Fig. 11. Fig. 11 (a) and(b)showthe resultsofSL-fragmentcomposition, compositeS-regions ωa and ωb, forS-regions ωa and ωb in Fig. 10 (a) and in Fig. 10 (b), respectively. Example 10. In Fig. 11, ωa has compositestate(s2; s5), and (s4; s7) asitsentrystateandexitstate,respectively. Likewise, ωb has twoentrystates(s2; s5) and(s9; s6), and twoexitstates(s4; s7) and(s4; s8). Notethatcomposite state(s1; s5) isoutofscopeof ωa and ωb, soitisexcluded from ωa and ωb, respectively. Next,wevalidateatomicitypropertyof S-region by checkingwhetherSL-fragmentsoftheS-regioncanbe composed intoanatomiccompositeS-region.Weconsider the propertyof AL-fragment to define atomicS-region (AS-region), i.e., AS-region must haveallentrytransitionfired fromthesame(composite)sourcestateandallexittransi- tions firedtothesame(composite)targetstate. 64 S. Yongchareonetal./InformationSystems47(2015)51–81
  • 15. Definition 22. (AS-regionandASL-fragment).GivenACP model Π, let S-region ω¼ ðΓ; Rsync) and ZΓDΠ:Z be asetof artifacts whereofwhichhasitsL-fragmentdefinedin Γ. The compositionofallSL-fragmentsin Γ satisfies the propertyof AL-fragment iff, forevery ℓCi AΓ, ℓCi is an AL-fragment, 8ℓCx ; ℓCy AΓ; 8rAφðℒCx ;ℒCy Þ; (ssAℓCi :S; ss)rsAℒCi : )-ss) r sAℓCi : ) 4rARsync 8ℓCx ; ℓCy AΓ; 8rAφðℒCx ;ℒCy Þ; (stAℓCi :S; s)rstAℒCi : )-s) r stAℓCi : ) 4rARsync 8ℒCj ðCjAΠ:Z ZΓ Þ;φðℓCi ;ℒCjÞ ¼ ∅. By holdingaboveconditions, ω can beconsideredasan AS-region. WealsocalleachL-fragmentin Γ as ASL- fragment of ω. Notethattheconditionsin Definition 22 are usedto restricttwoSL-fragments(tobecomposedforS-region)to include everytransitionandcorrespondingsyncrulethat areusedforonlythesynchronizationbetweenL-fragments in Γ. IfanS-regioncontainsmorethantwoSL-fragments, then, likelifecyclecompositionforACPmodel,wecan check theAS-regionbyperformingiterativecomposition for eachSL-fragmentinthatS-region.Itisalsoworth mentioning thatthecompositionofSL-fragmentsforAS- region holdsassamecharacteristicsasforthelifecycle composition – that iscommutativeandassociative,i.e., ℓCx ℓCy ¼ ℓCy ℓCx and ℓCx ℓCy ℓCz ¼ ℓCx ðℓCy ℓCzÞ ¼ðℓCx ℓCyÞ ℓCz . Example 11. In Fig. 10 (a), ωa is anAS-regionasboth L-fragments l1 and l2 areAL-fragmentswithallrelatedsync rules (r1 and r2) residingwithinthem.Assuch,the resultingfragmentfromthecompositionbetween l1 and l2 is thenatomic,asshownin Fig. 11 (a). Incontrast,in Fig. 10 (b), wecanseethatL-fragment l4 cannotsatisfythe propertyofAL-fragment,andL-fragment l3 does not include transition s5)r4 s6 wheresyncrule r4 exitsinentry transition s5)r4 s6 of l4. Therefore, ωb cannotbeconsidered as AS-region. Example 12. Now,weillustratethecasethatanS-region contains morethantwoSL-fragmentswhereoneofwhich synchronizedonthe(nested)sub-fragmentofSL-fragment, as shownin Fig. 12. AssumeS-region ω1 for SL-fragments {l1; l2; l3}, ω1 cannotbeconsideredasanAS-regionas l1 has some syncrulesthatareusedforthesynchronization betweenitssubL-fragment l5 and L-fragment l4 of the lifecycleofartifact A4. Therefore,weneedtoinclude l1 into the S-regioninordertosatisfythepropertyofAS-region. In thenextsection,wediscussabouthow AS-region can be usedfortheabstractionofsynchronization. A1 A2 s1 s2 s4 s3 s5 s7 r1 s6 r2 l1 A1 A2 s1 s2 s4 s3 s5 s7 r1 s6 r2 s9 s8 l2 l3 l4 a b r4 r3 ω ω Fig. 10. An exampleofS-regionsandSL-fragments. Fig. 11. Composite S-regionsofSL-fragmentsbasedon Fig. 10. A1 s1 A3 A2 A4 s2 s1 s3 s2 s1 s2 s4 s3 l4 s3 s5 s4 s6 l1 s2 s3 s1 l2 l3 l5 Fig. 12. An S-regionwithsubSL-fragments. S. Yongchareonetal./InformationSystems47(2015)51–81 65
  • 16. 4.3.2. SL-fragmentabstraction Here, wedefinetheabstractionrelationofthesynchro- nization betweentwoabstractartifactsintheabstractACP model andthesynchronizationbetweentwoartifactsinits base ACPmodel. Definition 23. (Synchronization(Sync)abstraction).Let artifact lifecycles ℒ0Cx and ℒ0Cy in ACPmodel Π0 abstract artifact lifecycles ℒ0Cx and ℒ0Cy in baseACPmodel Π with lifecycleabstractionmappings laℒ0Cx -ℒCx and laℒ0Cy-ℒCy , respectively.Wecandefine sync abstractionmapping func-tion saΠ0- Π ðℒ0Cx ;ℒ0Cy ÞðℒCx ;ℒCy Þ :φðℒ0Cx ;ℒ0Cy ÞφðℒCx ;ℒCy Þ that isused to projectthe abstractsyncrule for ℒ0Cx and ℒ0Cy onto its base syncrulefor ℒ0Cx and ℒ0Cy . Notethat saΠ0 - Π ðℒ0Cx ;ℒ0Cy Þ ðℒCx ;ℒCy Þ is atotalfunctionandwe may write sa without itssuperscriptionandsubscriptionin a clearcontext. Now,wewanttoperformanabstractiononanL- fragment thatsynchronizeswithotherL-fragment(s).Simi- lar toabstractionfunctionfornon-synchronizedlifecycle, we define sync abstractionfunction for abstracting AS-region that containssynchronizedL-fragments. Definition 24. (Sync abstractionfunction(sa_f)).Given ACPmodel Π, letAS-region ω¼ ðΓ; Rsync Þ to beabstractedin abstractACPmodel Π0, where Γ ¼ fℓC1 ; ℓC2 ; …; ℓCx g and ℓCi ð1rirxÞAΓ is anL-fragmentoflifecycle ℒCi in Π. We define syncabstraction function sa_f ðΠ;ωÞ to return Π0 with asetof abstractartifactlifecycles L ¼ fℒ0C1 ;ℒ0C2 ; ::;ℒ0Cx g and an abstractsyncruler0, where ℒ0Ci ð1rirxÞAL is a lifecycleofartifact CiAΠ0:Z such that, for every ℓCi AΓ, ℓCi is abstractedinto abstracttransi- tions in ℒ0Ci , i.e., ℒ0Ci ¼ la_tranðℒCi ; ℓCi Þ; for everysyncrule rAΠ:R that isusedtosynchronize betweenanytwoL-fragmentsin Γ, r is abstractedinto r0AΠ0:R, i.e., 8ℓCx ; ℓCy AΓ; 8rAφðℓCx ; ℓCy Þ; (!r0Aφðℒ0Cx ;ℒ0Cy Þ; sa ðℒ0Cx ;ℒ0Cy Þ ðℒCx ;ℒCy Þðr0Þ ¼ r Abstractsyncrule r0 that isusedtosynchronizeall abstracttransitionstogethercanbedefinedasfollow.For everyL-fragment ℓCi ¼ ðS; );)in;)out ÞAΓ, wehave, (ss; stAℒCi :S S;ðss)in) n )outst Þ ssApre_sðr0; CiÞ4stApost_sðr0; CiÞ Example 13. In Fig. 13 (a), wecanseethatAS-region ωa contains twofragments l1 and l2 with syncrules r1 and r2. When applying sa_f ðΠ;ωa), wehaveabstractlifecycleof ℒA1 and abstractlifecycleof ℒA2 with abstractsyncrule r0. In addition, Fig. 13 (b) showsacaseofsyncabstractionfor AS-region ωb which containsmultipleentrytransitions and multipleexittransitionsASL-fragments(bothASL- fragments l3 and l4). In theaboveexample,wedemonstratethesyncabstrac- tion oftwosynchronizedfragments.However,itispossible that anAS-regioncontainsmorethantwoASL-fragments. Forwiderunderstandingofsyncabstraction,weillustratea sync abstractionofmorethantwoSL-fragmentsin Fig. 14 (with artifact A3 extendedtotheexamplein Fig. 13 (b)). Example 14. In Fig. 14, wecanseethatAS-region ωb contains synchronizedfragments l3 and l4 of artifact A1 and A2, respectively,and l5 of artifact l4. Asallthreefragments can beconsideredasASL-fragmentsin ωb, wecanvalidly apply function sa_f ðΠ;ωb) andtheabstractlifecyclesof artifacts fA01; A02; A03g with abstractsyncrules fr0; r″g are returned. A1 A2 s1 s2 s4 s3 s5 s7 r1 s6 r2 l1 l2 a A’1 A’2 s1 s2 s4 s5 s7 r' A1 A2 s2 s4 s3 s5 s7 r1 s6 r2 s9 s8 l3 l4 b r4 r3 s1 A’1 A’2 s2 s4 s5 s8 r' s1 ω ω Fig. 13. Examples ofsyncabstraction. S. Yongchareonetal./InformationSystems47(2015)5166 –81
  • 17. Next,wediscussthecaseofanAS-regioncontainingSL- fragmentswithanested(sub)SL-fragmentthatsynchro- nizes withotherlifecycle.Intuitively,thesubSL-fragment and itssynchronizedlifecycleshouldbealsotakeninto account whenitssuperfragmenthastobeabstracted. Therefore,weneedtoinducetheabstractiontoitssub fragmenttogetherwithitscounterpartiftheybothcan satisfy thepropertyofAS-region(whichisconsideredasa sub AS-regionofthewhole).Inotherwords,wecansay that theentireAS-regionshouldcontainsuchcounterpart in ordertohaveavalidabstraction.Weshowanexampleof AS-regionconsistingofsubSL-fragmentin Fig. 15. Example 15. In Fig. 14, wecanseethatL-fragment l4 is a synchronizedfragmentofL-fragment l5 which isnested under L-fragment l1. Theabstractionyieldsabstracttransi- tions withabstractsyncrule r0z- thatis s1)r0z s3 in the lifecycleof A04 and s1)r0z s6 in thelifecycleof A01. Next,weconsiderthecasethatthesyncabstraction performs onanL-fragmentofoneartifactthatsynchronizes with theentirelifecycleofoneanotherartifact.Suchwhole lifecyclecanbeconsideredasfully-embeddedexternal lifecycleintheL-fragment.Wesaythatlifecycle ℒCj can be fully-embeddedinlifecycle ℒCi if thereexists L-fragment ℓCi!ℒCi that completelysynchronizesthe entire lifecycle ℒCj , i.e.,theentrystateandtheexitstate (s) of ℓCj are the init state andallthe final state(s)of ℒCj , respectively.Itcanbeunderstoodthattheabstractlifecycle of artifact Cj containing onlyabstracttransition(s)firing from its init statetoits final state(s) impliesasinglestep lifecycleandtheredoesnotexistsanyL-fragmentwithinit, i.e., f ind_Lf ðCj; Cj:SÞ ¼ null. Thebenefitofintroducingthis type oflifecycleabstractionisdiscussedlaterwhentaking into accounttheconstructionofpublicview.Itisnotedthat a lifecycleofoneartifactcanbeconsideredasfully- embedded externallifecycleofothermultiplelifecycles. Example 16. In Fig. 16 (a), L-fragment l1 synchronizeswith L-fragment l2 which representsthewholelifecycleof artifact A2. Therefore,wehave ℒA2 as anfully-embedded lifecycleof ℒA1 . Similarly,in Fig. 16 (b), wehave ℒA3 as a fully-embeddedlifecycleofboth ℒA1and ℒA2 . Next,weshowthat sync abstractionfunctionsa_f pre-serves B-consistency of betweentwosynchronizedlife- cyclesoftheinputandtwooutputtedabstractlifecycles. Theorem2. (synchronizedfragmentsb-consistentabstrac- tion). Let ω¼ ðΓ; Rsync Þ be anAS-regioninACPmodel Π that is tobeabstracted in ACPmodel Π0 and let L be asetof A3 A1 A2 s10 s11 s2 s4 s3 s5 s7 r1 s6 r2 s9 s8 l3 l4 b r4 s1 r8 r r3 7 s13 s14 r6 A’1 A’2 s2 s4 s5 s8 r' s1 s12 r5 l5 A’3 s10 s11 s14 r'’ r8 ω Fig. 14. An exampleofanabstractionoftwoormoreASL-fragments. A1 s1 A3 A2 A4 A’1 s1 s6 A’3 s1 s3 s4 s2 s1 s3 s2 s1 s2 s4 s3 l4 A’4 s1 s3 r'z s3 s5 s4 s6 l1 s2 s3 s1 l2 l3 A’2 s1 s3 r'x r'y l5 Fig. 15. An exampleofanabstractiononnestedsub-SL-fragments(cf. Fig. 12). S. Yongchareonetal./InformationSystems47(2015)51–81 67
  • 18. abstractartifactlifecyclesresultedfromapplyingsync abstractionfunctionsa_f ðΠ;ωÞ. Then thefollowingstatement holds. 8ℒCi ;ℒCj Afℒjℓ!ℒ4ℓAΓg; (ℒ0Ci ;ℒ0Cj AL; ℒ0CiClaℒCi 4ℒ0CjClaℒj-ℒ0Ci ℒ0CjCℒCi ℒCi Wecanprove Theorem 2 by using Definition 21 and Definition 22for the B-consistency checkingbetweentwo abstractlifecyclesandtheirbaselifecycles,andbetween the compositionbetweentheformerandthecomposition betweenthelatter.From Definition 24, syncabstraction sa_f alwaysreturnsasingleabstracttransitionineach abstractlifecycleandasinglesyncrulebetweentwo abstracttransitions,sothecompositionofsuchabstract transitionsalwaysyieldsanatomiccompositefragment. The compositionofASL-fragmentsintheirbaselifecycleis atomic andcanentirelyberepresentedbythecomposition of abstracttransitions. 4.3.3. FindingminimalAS-region Next,wediscusssimilarrequirementsfortheneedof NAL-fragmentexpansion(toAL-fragment).GivenanL- fragment ofoneartifactlifecyclethatsynchronizeswith otherartifactlifecycle(s),wecanfindtheminimalAS- region (whichcontainsthatL-fragment)anditsminimal counterpart(s)inwhichtheycanbeusedasinputsfor sync abstractionfunctionsa_f . Here,weproposetwoalgorithms to findtheminimalAS-regionofanL-fragment.Notethat the synchronizedcounterpartscanbemanyastheL- fragment cansynchronizewithmultiplelifecycles.Two proposed algorithms f ind_minASR and f ind_SR areusedto expandtheboundaryofanL-fragmentalongwithits synchronizedcounterpart(s)untilallsatisfythecondition of AS-regionandASL-fragment(in Definition 22). IfnoAS- region canbeconstructedoranL-fragmenthasnosyn- chronizedcounterpartofanyotherlifecycle,thenthe null valueisreturned. Algorithm2.(function find_minASR). Finding aminimal AS-regionfromanL-fragment Input: L-fragment ℓCx in ACP model Π. Output: AS-region ω¼ ðΓ; Rsync Þ if found;otherwise null is returned. 1. ω’ðfℓCx g; ∅); 2. ω¼ find_SRðωÞ; 3. return ω; Algorithm3.(function find_SR). Finding anexpanded S-regionfromaninputtedS-region Input: S-region ω¼ ðΓ; Rsync Þ in ACPmodel Π. Output: ExpandedS-regionof ω or null if anySL-fragmentin ω cannot satisfythepropertyofAL-fragment. 1. for each ℓiAω:Γ do 2. L-fragment ℓ0i ¼ find_minALðℓi); 3. if (ℓ0i ¼ nullÞ then return null; 4. else 5. ω:Γ ¼ ω:Γ [ fℓ0ig fℓig; 6. end if 7. for each ℒjAfℒCj j CjAΠ^ :Zg do 8. states Sj ¼ ∅; 9. for each rkAφðℓ0i ;ℒjÞ do 10. Sj’pre_sðrk; CjÞ [ post_sðrk; CjÞ; 11. if (∄rkAω:Rsync Þ then rkAω:Rsync’rk; 12.endfor 13. L-fragment ℓj ¼ f ind_Lf ðCj ; SjÞ; 14. L-fragment ℓ0j ¼ find_minALðℓj); 15. if(ℓ0janullÞ 16. then 17. if ð∄ℓ0jAω:ΓÞ 18. then 19. ω:Γ ¼ ω:Γ [ fℓ0jg; 20. ω¼find_SRðω Þ; 21. return ω; 22. end if 23. else return null; 24. end if 25. end for 26. end for Here, weusearunningexampleillustratedin Fig. 17 to explainthealgorithmsoffunctions f ind_minASR and f ind_SR. Webeginwith Fig. 17 (a) byapplying f ind_minASRðl1Þ in this example.FirstittakesL-fragment l1 of artifact A2 as an inputandinitializesanS-regionfrom l1 and anemptysetof sync rule.ThenitfindsthepossiblecorrespondingS-region for l1 (bycalling f ind_SR function). Function f ind_SR starts checkingwhether l1 can satisfythepropertyofAL- fragment(Line3).Ifso,thenitcontinuessearchingforall sync rulesandtheirrelatedsynchronizedstatesthatare used tosynchronize l1 with anyotherlifecycletransitions; otherwise,itimmediatelyreturns null. Intheiterationof finding synchronizedpartofotherartifact,ifasetof synchronizedstatesisfound(Lines9–12),thenafragment consisting ofsuchsetisconstructedbycalling find_Lf function (definedin Definition 14) – that isL-fragment l2 of artifact A1 shown in Fig. 17 (b). Next, l2 is tobechecked whether itisqualifiedasAL-fragment.Ifitsatisfies,thenit is addedintotheS-region(Lines13-21);otherwise,ithas tobeexpandeduntilitcansatisfyAL-fragment.Once qualified,itisaddedintotheS-region(Line19).After l2 is added totheS-region,werecursivelycallfunction f ind_SR again withtheexpandedS-regionasaninput(Line20). Fig. 17 (c) showsAL-fragment l02 that isexpandedfrom l2. Wecanseethat l02 introducesnewsyncrules fr5; r6g that arenotincludedintheS-region.Therecursionwillcon- tinue untilalltheSL-fragmentsintheS-regionareAL- fragmentsandincludeallthepossiblesyncrulesthatare used intheS-region.Consequently, l1 is requiredtoexpand itself andbecomesnewAL-fragment l01, asshownin Fig. 17 (d). Afterthat,wecanseenewsyncrules fr7; r8g appear in l01. TheS-regionneedsanotherexpansionagaintocover those syncrules – that isL-fragment l3 of artifact A3. The exitconditionoftherecursioniswhenthefragmentisnull (Line 15withtheexitonLine23).Thefinaloutputofthe f ind_SR function isanS-regionthatallthesyncrulesare discoveredandthatcontainsonlyAL-fragments.Finally,by completing theiterationandrecursionwithallthecondi- tions ofAS-regionsatisfied,thefunctionreturnsthemini- mal AS-regionconsistingofASL-fragments fl01; l02; l3g as the output fromtheprovidedinputL-fragment l1. IftheAS- regioncannotbeconstructedthenthefunctionreturns S. Yongchareonetal./InformationSystems47(2015)5168 –81
  • 19. A1 A2 sa s1 s3 sa s1 r1 s2 s6 s5 s4 s6 r4 s5 s2 r3 r2 l1 s3 r5 A1 A2 sa s1 s3 s2 s6 s5 s4 sa s1 s6 s5 s2 r1 r4 r3 r2 s4 r6 s3 r5 s4 r6 l'2 l1 A1 A2 sa s1 s3 s2 s6 s5 s4 sa s1 s6 s5 s2 r1 r4 r3 r2 s3 r5 s4 r6 l'2 A1 A2 sa s1 s3 sa s1 l2 r1 s2 s6 s5 s4 s6 r4 s5 s2 r3 r2 l1 s3 r5 s4 r6 A3 s1 s2 s3 r7 r8 A3 s1 s2 s3 r7 r8 A3 s1 s2 s3 r7 r8 A3 s1 s2 s3 r7 r8 l'1 A1 A2 sa s1 s3 s2 s6 s5 s4 sa s1 s6 s5 s2 r1 r4 r3 r2 s3 r5 s4 r6 l'2 A3 l'1 s1 s2 s3 r7 r8 l3 Fig. 17. A runningexampleforthe f ind_minASR function. A1 A3 A2 s1 sf s4 A’2 s1 s3 s4 A’1 s1 s4 A2 A’1 s1 s3 A1 s1 s3 s2 s1 s4 s2 s3 s1 s3 s2 l1 l1 l2 sf A2 sf r' s1 s2 r2 r1 l2 A’3 sf Fig. 16. An exampleoffully-embeddedexternallifecycles. S. Yongchareonetal./InformationSystems47(2015)51–81 69
  • 20. false. WecansaytheresultedAS-regionisminimalasthe function usethe f ind_minAL function toguaranteethe minimal AL-fragmentexpansionintheAS-region. Based on Theorem 2 and theuseofthe f ind_minASR and f ind_SR functions, wecanconstructaconsistentpublic view ofsynchronizedlifecyclesbyabstractingtheirsyn- chronizedpartsofthelifecycles. 4.4. Consistentpublicviewconstruction Now,weuse Theorems1and2 to formulatethe B-consistency for theentireartifact-centricinter-organiza- tional businessprocess. Theorem3. (b-consistent publicview). Let paΠ^ be apublic view of ACP-imodel Π^ constructed byapplyingsyncabstrac- tion functionsa_f and lifecycleabstractionfunctions la_tran and sa_state. Then paΠ^ is aB-consistentabstraction of Π^ , i.e., ℒpa ^ ΠClaℒΠ^ . Theorem 3isderivedfrom Theorems1and2. Recallthatinthepublicview,alllocalartifactsshouldbe invisible.Onlyabstractsharedartifactsarerevealedforthe collaboration.Inordertohidethoselocalartifactsofeach party,thepartymustensurethatifalocalartifactsynchro- nizes withafragmentofasharedartifact,thenthis fragment mustbehiddenaswell.Thishidingproperty refers totheabstractionofsuchfragment.However,ifthe entire lifecycleofalocalartifactisabstracted,thenitcanbe validlyhiddeninthepublicview.In Section 4.3.2, we discuss thefully-embeddingpropertytocapturethiskind of lifecycle.Inotherwords,ifthelifecycleofalocalartifact is afully-embeddedexternallifecycleofthesharedartifact, then thelocalartifactcanbehidden.Assuch,thecorre- sponding fragment(s)inthesharedartifact(s)isalso abstracted(byusingsyncabstractionfunction sa_f ). How- ever,thereisacaseifthatfragmentisnotanAL-fragment, then the sa_f function cannotbeappliedduetothe requirementoftheinputthatmustbeAS-region(consist- ing ofASL-fragmentsofsharedartifact(s)andlocalartifact). Tocopewiththisissue,thenweproposetouselifecycle abstractionfunction la_state to abstractitintoanabstract state. Thenthelocalartifactthatsynchronizeswiththat fragment canbeabstracted.Althoughwecanabstractan NAL-fragmentintoanabstractstate,itishardtodecide whether thatabstractionisvalidandconsistentintermsof synchronizationbehavior.Thelocalartifactthatsynchro- nizes withanypartofsuchNAL-fragmentmustbeexclu- sivelyencapsulatedintheabstractstate.Inaddition,the abstractstateitselfisnotconsideredasatomicifithas multiple entryandmultipleexittransitionsfrom/todiffer- ent states.Therefore,weallowhavingsyncabstractionfor the abstractstateforthecasethatthewholelifecycleofthe local artifactissynchronizedwithintheNAL-fragment.As previouslydiscussed,representinganNAL-fragmentinan abstracttransitionisdeemedinconsistent. Example 17. Revisitourpurchasingprocessexamplein Fig. 1. Wecanconstructafragmentforthe PO artifact that consists ofstates fcreated; on_holdg. Thefragmentiscon- sideredasNAL-fragment(duetotwoexitstates).Wecan see thatitsynchronizeswiththeentirelifecycleofthe Quote artifact, therefore,theabstractionofthisfragment must yieldanabstractstatewhichrequiresthewhole lifecycleof Quote tobeabstracted,i.e.,fully-embedded – that isthe approving stateofabstract PO in thepublicview shownin Fig. 5. Similarly,theentirelifecycleof PL artifact can beabstractedandfully-embeddedinthe supplying stateoftheabstract PO in thepublicview. Next,weproposeanalgorithm(forafunctionnamed f ind_minPV) tohelporganizationstoautomaticallyfindthe minimal, consistentpublicviewfromtheir ACP-i model. Due totheinconsistencyissueonNAL-fragmentabstrac- tion, wedonottakeNAL-fragmentsintoaccountinthis algorithm.Afterpresentingthealgorithm,wethenshow howitcanguaranteethe B-consistency. Algorithm4.(function find_minPV). Finding themini- mal, consistentpublicviewofan ACP-imodel Input: ACP-i model Π^ ¼ ðZ; V; R; L; γÞ. Output: theminimalpublicviewof Π^ . 1. public ACP-imodel Π^ 0 ¼ Π^ ; 2. for each CiAfCAΠ^ 0Z jγðCÞj ¼ 1g do 3. L-fragment ℓi ¼ f ind_Lf ðCi ; Ci :SÞ; 4. AS-region ω¼ find_minASRðℓiÞ; 5. i f(ω anull) 6. then 7. Π^ 0:¼ sa_f ð Π^ 0; ωÞ; 8. artifacts Zl’Ci; 9. for each CjAfCAΠ^ 0:Z jγðCÞj ¼ 1g do 10. L-fragment ℓj ¼ f ind_Lf ðCj ; Cj :SÞ; 11. if ðℓj ¼ nullÞ then Zl’Cj ; 12. end for 13. remove all artifactsin Zl and theirrelatedabstractsyncrules from Π^ 0; 14. Zl ¼ ∅; 15.endif 16.endfor 17. return Π^ 0; Now,weexplainthealgorithmforthe f ind_minPV function. First,itinitializesapublicviewasidenticalto an input ACP-imodel. Then,itsearchesforalllocalartifacts in thepublicview(Line2).Foreachlocalartifactfound,it attemptstoconstructanAS-regionthatconsistsofsuch artifact (Lines3–4). IfitisabletoconstructtheAS-region (Line 5),thenabstractionfunction sa_f is appliedonthe regionandtheartifactisaddedtotheset Zl and allartifacts in Zl will belaterremovedfromthepublicview(Lines7–8). As sa_f finds allcorrespondingASL-fragments(ifconstruc- tible) thatsynchronizedwiththelocalartifactintoaccount for theabstraction,thealgorithmalsosearchesforother local artifactthatqualifiesasfully-embeddedexternallife- cycletobeincludedin Zl (Lines 9–12).Thiswilleliminate an unnecessarylocalartifactinthepublicview;therefore, the numberofiterationsforfindinglocalartifactsis reducedinthemainloop. Fromthealgorithmwewillgettheminimal, B-consistent public viewofthe ACP-i model. However,itdoesnot guaranteethatalllocalartifactsareabstracted.This depends onwhetherthelifecycleofsuchartifactscanbe S. Yongchareonetal./InformationSystems47(2015)5170 –81
  • 21. fully-embeddedaswellasthesynchronizedpartofthe sharedartifactisanASL-fragment.Inotherwords,ifthe fragmentisanNAL-fragment,then,aspreviouslydiscussed, the abstractionofthoselocalartifactsmustbeachieved manually – by abstractingeachofsuchfragmentsintoan abstractstateinstead.Thisprocessrequiresapredefined abstractstateforanabstractsharedartifact. Next,weshowin Theorem 4 that our f ind_minPV function producesaminimal, B-consistent public view. Theorem4. f ind_minPVðΠ^ Þ returns aminimal, B-consistent public viewof Π^ . Wecanprove Theorem4 as follows.First,weprovethat f ind_minPVð Π^ Þ returns acorrectpublicviewof Π^ . Thiscan be donebyinductionoverallmappingconditionsin Definition 12. Notethatremovinglocalartifactfromthe public viewwhereitsentirelifecyclecanbeabstractedalso conforms tothedefinitionofpublicview.Second,weprove that given ACP-i model Π^ , f ind_minPVð Π^ Þ is guaranteedto returnminimalpublicviewof Π^ . Asweusefunction find_minASR to searchfortheminimalAS-regionthatcan be usedasaninputoftheabstractionfunction sa_f , then this statementisnaturallysatisfied.Last,weprovethe B-consistency of thegeneratedpublicview.Basedon Theorem 3 and Definition 24, thepublicviewpreserves B-consistency as function sa_f alwaysyieldsa B-consistent abstractACP model oftheinputtedACPmodel. Wenowconclude Section4. Insummary,thissectionhas discussedthepublicviewconstructionmethodologyand behaviorconsistencyandformulatesseveralfunctionsand theoremsthatcanbeusedtoconstructthe B-consistent public viewofthecollaboration.Italsopresentsouralgorithmsthat helporganizationsautomaticallygeneratetheminimal abstractpublicviewwiththeassuranceofthe B-consistency. 5. Consistentprocesschangesandprivateviews In thissection,weproposeamechanismthatallowsa party inthecollaborationtochangetheirlocal(private) processwithoutaffectingthecorrectnessandconsistency of theoverallprocess. 5.1.Localprocesschanges Organizations mayneedtochangetheirlocalprocesses due totheirnew(orupdated)setofbusinessrequirements or regulationsthattheyhavetofollow.Weobservethatin artifact-centric processes,changestolocalprocessescanbe classified intothreetypesbasedonthethreecomponents of the ACPmodel. Changes toartifacts. Anorganizationmaymodify/delete/ add anattributeorastateofitslocalartifact,whichcan be seenasstructuralchangestothedatamodelofan artifact. Inaddition,exceptforchangestoexisting artifacts, itispossiblethatanorganizationmayincor- porateasetofnewartifactsintothelocalprocess. Changes totasks. Stemmedfromservice-orientedarchi- tecture,anorganizationmayseektoaggregateexisting local tasksintoacompositetask(i.e.,aservice).Onthe otherhand,acompositetaskcanbedecomposedinto smaller tasks.Duetothesepossiblechangestotasks,the specification ofatask,includingtheinput,output,and pre/post-conditionsofthetask,issubjecttoreflecting the changes –. Changes tobusinessrules. Businessrulescanbechanged due tothestructuralchangesofartifacts,thespecifica- tion changesoftasks,andthechangesoftheprocess logic. Itisworthnotingthatachangemayaffectan existinginteractionbehaviorbetweenartifacts. Next,wedefinethree changeoperators that canbeusedto expresswhatandhowtheabove-mentionedthreetypesof changecanbeachieved.Asaforementioned,inthisarticlewe onlyfocusonthebehavioraspectofartifact-centricbusiness processes.Werestrictourdiscussioninbehavioralchangesof localprocesses.Therefore,forsimplicity,weassumethat capturingthechangeofabusinessrulecanreflectthechange of artifact(s)andtask(s)involvedinthisrule.Thismakes sense astheassociationsbetweenartifactsandtasksare defined inthepre/post-conditionsandactions,respectively, in businessrules.Inthecaseofaddinganewartifacttoalocal process,wecanderivethelifecycleofthenewartifactandits interactionfromtheaddedbusinessrulesthatareusedto inducethestatetransitionoftheartifactandthesyncrules thatareusedtosynchronizeother existingartifact(s)withit, respectively. Definition 25. (Change operators).Let Π^ l ¼ ðZl; Vl; Rl Þ be a localACPmodeloforganizationrolelinthecollaboration. An organizationcanchangeitsownlocalprocessby applyingthefollowingchangeoperatorsoverbusiness rules Rl. Add operator ðRþ Þ. Wewrite Rþ ðrÞ to meanthatnew business rulerisaddedintoR,i.e., Rþ ðrÞ ¼ Rl [ frg. Delete operator ðR Þ. Wewrite R ðrÞ to meanthat existingbusinessrulerisdeletedfromR,i.e., R ðrÞ ¼ Rl frg. Replaceoperator (R%). Wewrite R%ðrx; ryÞ tomeanthat existingbusinessrulerx is replacedbynewbusiness rule ry in R,i.e., R%ðrx; ryÞ ¼ Rl [ fryg frxg. It isnotedthatthereplaceoperatoridenticallyperforms as thecombinationoftheaddoperatorandthedelete operator;however,itprovidesbettertraceabilityofchanges by maintainingtherelationforthereplacementofanold business rulewithanewbusinessrule. Definition 26. (Modified localACPmodel,processchange function (pc)).Let Π^ l ¼ ðZl; Vl; Rl Þ be a local ACPmodel of role l.Wecanobtain modified localACPmodel Π^ l0 from Π^ l by applying processchangefunctionpc Π^ l :Π^ l X-Π^ l0 where X is aunionsetof changeoperationsRþ, R, andR% that perform onsomebusinessrulesinRl. Next,weexplaintheconceptofprivateviewandhowit can beusedtovalidatethechangesoflocalprocess. S. Yongchareonetal./InformationSystems47(2015)51–81 71
  • 22. 5.2. Privateviewsandchangevalidation In theoverallpicture,wevalidatethechangesoflocal processbycheckingwhetherthe B-consistency of the modified inter-organizationalbusinessprocess(afterlocal processchangesapplied)andtheagreedpublicviewcanbe preserved.A privateview of anorganizationisusedto capturethemodifiedlocalprocessforthepurposeoflocally checkingwhetherthelocalprocessisstillabletoprovide what promisedintheagreedcontract,i.e.,publicview.We illustrateanoverviewofprocesschangevalidationin Fig. 18. Next,wedefinethe privateview of aparticularorgani- zation roleinthecollaborationbasedontheagreedpublic view.Basedoneachversionoflocalprocesschanges,each role hasacorrespondingprivateviewthatcapturesits (modified) local ACPmodel plus abstractsharedartifacts defined inthepublicview. Definition 27. (Privateview).Let paΠ^ ¼ ðZp; Vp; Rp; Lp; γpÞ be apublicviewofACP-imodel Π^ and Π^ l ¼ ðZl; Vl; Rl Þ be a local ACPmodelofrole lALp. The privateview of role l can be definedby privateviewmapping function pv: Zx [ Vx [ Rx [ γx-ðZp [ Zl Þ [ Vl [ ðRp [ Rl Þ [ Lp [ γp such thatthe followingshold. each abstractsharedartifactin Zp existsin Zx, i.e., 8CiAfCAZp j γpðCÞ 41g; (CjAZx; pvðCjÞ ¼ Ci γxðCjÞ ¼ γp ðCiÞ each localartifactin Zl (not in Zp) existsin Zx, i.e., 8CiAfCAZl jjγpðCÞj ¼ 1g; (CjAZx; pvðCjÞ ¼ Ci γxðCjÞ ¼ l each taskin Vl existsin Vx, i.e., 8viAVl; (vjAVx; pvðvjÞ ¼ vi γxðvjÞ ¼ l each businessrulein Rl existsin Rx, i.e., 8riARl; (rjARx; pvðrjÞ ¼ ri; each abstractbusinessrulein Rp that isnotusedforthe synchronizationbetweenanabstractsharedartifactin Zp and alocalartifactin Zl existsin Rx, i.e., 8riAfrARp jl=2γpðrÞg; (rjARx; pvðrjÞ ¼ ri Notethatwemayusethetermprivateviewforthe lifecycleofthisprivateviewinanunambiguouscontext. Example 18. Fig. 19 depicts thelifecyclesofartifactsinthe privateviewoftheoriginallocalprocessofSupplier (Π^ Supplier 1 ) whichisextractedfromthecompletepurchasing processshownin Fig. 1. Suchprivateviewcanbecon- structedbasedonthepublicviewshownin Fig. 5. Apart fromthelocalartifacts PL and DN, comparedwiththe public view,wecanseelocalprocessdetails(ingray- shaded areas)ofsharedartifacts PO, SO, and IV. Example 19. Now,considerthecasethatifSupplierwants tochangeitslocalprocessbasedontheexistingone(cf. Fig. 19). Theresultofchangesisillustratedinprivateview Π^ Supplier 2 in Fig. 20. Sincethe IL artifact isaddedintothelocal process,wecanseesomesynchronizationbetweenexisting artifact PL and newartifact IL, aswellasthechangeof state’s namesof PL (from checking to scheduled, from out of stock to unavailable, andfrom in stock to picking). Existing business rulesthatcorrespondtothesechangesare affectedandneededtobeadjustedaccordingly.Obviously, a newsetofbusinessrulesisneededtoexpressthe lifecycleoftheaddedartifact.Thissetincludessyncrules that areusedforthesynchronizationbetween PL and IL. Next,wedefine processconformance and itsconditions that canbeusedtocheckwhetherchangesinlocalprocess can beimplementedwhilepreservingthecorrectnessand consistencyoftheoverallprocess.Basically,wereusethe definition of B-consistency to define processconformance for Fig. 18. An overviewofprivateviewsandviewconformance. S. Yongchareonetal./InformationSystems47(2015)5172 –81
  • 23. consistentlocalprocesschangesastosatisfythefollowing statements. Changesshouldnotleadtoanunsoundglobalprocess, i.e., themodifiedglobalprocessshouldbeabletoreach its goalstatesasitsoriginalglobalprocessdoes,and, Changesshouldbeguaranteedthatthe B-consistency of the modifiedglobalprocessanditsoriginalglobal processispreserved. As anagreedpublicviewisconstructedforthecolla- borationwhichactslikeacontract,sowecanexpress consistent processchanges by meansofnotbreakingthe original publicview.Inotherwords,themodified ACP-i model must beconsistentwiththeagreedpublicview. Regardingthebehavioralequivalencenotioninprocess algebras [9], wecansaythatourapproachforconsistent processchangespreservesthecongruencepropertyofthe modified localprocessanditsbaselocalprocess,i.e.,they can behaveinterchangeablywithoutaffectingtheoverall process. Definition 28. (Processconformance, F). Let Π^ ¼ ðZ; V; R; L; γÞ be anACP-imodeland paΠ^ be its public view. Let pv Π^ l be aprivateviewof local ACPmodel Π^ l for role lAL. Wesaythat Π^ l conformspa Π^ , writtenas Π^ l FpaΠ^ , iffthe Picking List(PL) Purchase Order(PO) confirmed canceled Out ofstock closed Shipping Order(SO) In transit Invoice (IV) approving acquiring accepted filled ready tofill Filled order checking Instock Delivery Note(DN) prepared transferring dispatched billing issued cleared sent unpaid clearing arrived delivering ready toship created scheduled Fig. 19. Supplier’s privateview Π^ Supplier 1 . Picking List(PL) Purchase Order(PO) confirmed canceled unavailable closed Shipping Order(SO) Invoice (IV) approving acquiring accepted filled ready tofill Filled order scheduled picking Delivery Note(DN) prepared transferring dispatched billing issued cleared sent unpaid clearing Inventory List(IL) checking ready topick sourcing canceled delivering ready toship In transit created scheduled arrived Fig. 20. Supplier’s modifiedprivateview Π^ Supplier 2 . S. Yongchareonetal./InformationSystems47(2015)51–81 73