Business Process Model Merging An Approach to Business Process Consolidation

业务流程模型合并:业务流程整合的方法

BusinessProcessModelMerging:AnApproachtoBusinessProcessConsolidation

MARCELLOLAROSA

QueenslandUniversityofTechnology,Australia

MARLONDUMAS

UniversityofTartu,Estonia

REINAUBA

UniversityofTartu,Estonia

REMCODIJKMAN

EindhovenUniversityofTechnology,TheNetherlands

Thispaperaddressestheproblemofconstructingconsolidatedbusinessprocessmodelsoutofcollectionsofprocessmodelsthatsharecommonfragments.Thepaperconsiderstheconstructionofunionsofmultiplemodels(calledmergedmodels)aswellasintersections(calleddigests).Mergedmodelsareintendedforanalystswhowishtocreateamodelthatsubsumesacollectionofprocessmodels–typicallyrepresentingvariantsofthesameunderlyingprocess–withtheaimofreplacingthevariantswiththemergedmodel.Digests,ontheotherhand,areintendedforanalystswhowishtoidentifythemostrecurringfragmentsacrossacollectionofprocessmodels,sothattheycanfocustheire ortsonoptimizingthesefragments.Thepaperpresentsanalgorithmforcomputingmergedmodelsandanalgorithmforextractingdigestsfromamergedmodel.Themerginganddigestextractionalgorithmshavebeenimplementedandtestedagainstcollectionsofprocessmodelstakenfrommultipleapplicationdomains.Thetestsshowthatthemergingalgorithmproducescompactmodelsandscalesuptoprocessmodelscontaininghundredsofnodes.Furthermore,acasestudyconductedinalargeinsurancecompanyhasdemonstratedtheusefulnessofthemerginganddigestextractionoperatorsinapracticalsetting.

CategoriesandSubjectDescriptors:H.4.1[O ceAutomation]:Work owmanagement;D.2.7

[Distribution,Maintenance,andEnhancement]:Versioncontrol

GeneralTerms:Design,Management

AdditionalKeyWordsandPhrases:Businessprocessmodel,variability,modelmerging,graphmatching

1.INTRODUCTION

Inthecontextofcompanymergersandrestructurings,itoftenoccursthatmulti-plealternativebusinessprocesses,previouslybelongingtodi erentcompaniesorunits,needtobeconsolidatedintoasingleoneinordertoeliminateredundanciesandcreatesynergies.Tothisend,teamsofbusinessanalystsneedtocomparesimilarprocessmodelssoastoidentifycommonalitiesanddi erences,andtocre-ateintegratedprocessmodelsthatcanbeusedtodrivetheprocessconsolidatione ort.Thiscomparisonandintegrationtaskistedious,time-consuminganderror-prone.Inoneinstancereportedinthispaper,ittookateamofthreeanalysts130man-hourstomerge25%oftwovariantsofanend-to-endprocessmodel.

Inthissetting,thispaperproposesandevaluatesalgorithmsforsemi-automaticallyconstructingaconsolidatedmodeloutofacollectionofprocessmod-

ACMJournalName,Vol.V,No.N,Month20YY,Pages1–0??.