Matthias Weidlich, Jan Mendling, Mathias Weske jan.mendling@wiwi.hu-berlin.de

1. Matthias Weidlich, Jan Mendling, Mathias Weskejan.mendling@wiwi.hu-berlin.de BehavioralProfilesAn Abstraction for Efficient Calculation of Consistency between Process Models

2. Poster auf Berliner BPM-Offensive http://www.bpmb.de

4. Agenda • WhyConsistencybetween Process Models? • WhyBehaviouralProfiles? • Howtovalidatetheconcept? • Whatarefurtherapplications? • Whattotakehome?

5. The Essence of Modeling is model of is model of correspond

6. Vertical Alignment of Process Models Different purposes for the creation of process models Process automation Staff planning Decision support Business certification Results in significant differences between models describing (parts of) the very same process Slicing of process models Modelling granularity Behavioural differences

7. Horizontal Alignment of Process Models Different variants of a common process, due to Scope of the process Organisational context IT-landscape No big differences in modelling granularity Still, defined behaviour might be different

8. Correspondences Model 1 Correspondences Model 2

9. Agenda • WhyConsistencybetween Process Models? • WhyBehaviouralProfiles? • Howtovalidatetheconcept? • Whatarefurtherapplications? • Whattotakehome?

10. Simply Comparing Activities is not enough

11. Existing Notions like Equivalence of Traces are too strict

12. Behavioural Profiles Need for a behavioural abstraction that is less sensitive to model projections or extensions, respectively Behavioural Profiles capture behavioural characteristics by means of relations between activities Strict order Exclusiveness Interleaving order Based on weak order: weak order between A and B, if there is a trace in which B occurs after A

13. Behavioural Profiles • Strict order between A and D • Exclusiveness between F and G • Interleaving order between C and E

14. Behavioural Profile

15. Properties • Close toTraceEquivalence • Computable in O(n3) for Free Choice nets • Easy tocalculatesimilarity, consistency, etc.

16. What about Trace Equivalence?

17. Agenda • WhyConsistencybetween Process Models? • WhyBehaviouralProfiles? • Howtovalidatetheconcept? • Whatarefurtherapplications? • Whattotakehome?

18. Case Study with SAP Reference Model Computation based on results proved for Petri nets Transformation BPMN to PN EPC to PN UML AD to PN Computation in low polynomial time for certain class of models EPC is sound EPC has unambiguous instantiation semantics

19. VaryingDegree of Profile Consistency

20. Consistent but not traceequivalent

21. Inconsistencies

22. Agenda • WhyConsistencybetween Process Models? • WhyBehaviouralProfiles? • Howtovalidatetheconcept? • Whatarefurtherapplications? • Whattotakehome?

23. Change in Process Model 1 • Assumptions • Change can be localized as a single node • Behavioural profile is consistent for aligned nodes • Find boundary nodes for change • Aligned with target model • Closest nodes in strict order preceding and succeeding change

24. Change Propagation • Derivation of change region supports • Analysis, whether a change should be applied • Application of a change in a consistent manner • Change region might be empty • No flow arc in target model meets requirements for change • Boundary nodes and inter-boundary nodes guide adaptation

25. Action Patterns Derivation of abstract actions from activities Mining of abstract patterns between activities in a repository Co-occurrences and behavioural relations Usage of these patterns for modelling support

26. ICoP Architecture • Architecture for the creation of matchers • Multi-step heuristic approach • Reuse of matching components • Adaptable & extendable • Concrete matching components • Exemplify and evaluate the architecture • Generalise existing approaches

27. Measurement of Compliance Different grouding of behaviouralprofilesforprocessmodelsandforlogs VS Strictness of order relations of Behavioural Profile Subsumption relation Forinstance, interleaving order in processmodelssubsumesstrict order in process log

28. Event Query Optimization MonitoringQueries Process Models Domain Expertfor Processes Analyst Alert if A -> B and … ExtractingBehavioralProfiles Query Translation warnings Process Tailored Execution Plans sub(A)  pull(B) …

29. Process Model Comprehension

30. Publications • M. Weidlich, J. Mendling, M. Weske: Efficient Consistency Measurement based on Behavioural Profiles of Process Models. IEEE Transactions on Software Engineering (TSE). To appear, 2010. • M. Weidlich, A. Polyvyanyy, J. Mendling, M. Weske: Efficient Calculation of Causal Behavioural Profiles using Structural Decomposition. In: 31st International Conference on the Application and Theory of Petri nets 2010, Braga, Portugal, 21-25 June 2010. • M. Weidlich, R. Dijkman, J. Mendling: The ICoP Framework: Identification of Correspondencesbetween Process Models. In: 22nd International Conference on Advanced Information Systems Engineering (CAiSE 2010), Hammamet, Tunesia, 07-11 June 2010. • M. Weidlich, A. Polyvyanyy, N. Desai, J. Mendling: Process Compliance Measurement based on BehaviouralProfiles. In: 22nd International Conference on Advanced Information Systems Engineering (CAiSE 2010), Hammamet, Tunesia, 07-11 June 2010. • S. Smirnov, M. Weidlich, J. Mendling, M. Weske: Action Patterns in Business Process Models. In: 7th International Joint Conference on Service Oriented Computing (ICSOC 2009), Stockholm, Sweden, 24-27 November 2009. • M. Weidlich, M. Weske, J. Mendling: Change Propagation in Process Models usingBehaviouralProfiles. In: IEEE International Conference on Services Computing (SCC 2009), Bangalore, India, 21-25 September 2009.

31. Other Selected Publications • H.A. Reijers, J. Mendling: A Study intotheFactorsthatInfluencetheUnderstandability of Business Process Models. IEEE Transactions on Systems Man & Cybernetics, Part A (SMCA), accepted. • I. Weber, J. Hoffmann, J. Mendling: Beyond Soundness: On theVerification of Semantic Business Process Models. Distributed and Parallel Databases (DPD). Volume 27, Number 3, pages 271-343, 2010, Springer-Verlag. • J. Mendling, H.A. Reijers, W.M.P. van der Aalst: Seven Process Modeling Guidelines (7PMG). Information and Software Technology (IST). Volume 52, Number 2, pages 127-136, 2010. • J. Mendling, H.A. Reijers, J. Recker: ActivityLabeling in Process Modeling: Empirical InsightsandRecommendations. Information Systems (IS). Volume 35, Number 4, pages 467-482. 2010. • G. Decker, J. Mendling: Process Instantiation. Data & Knowledge Engineering (DKE). Volume 68, pages 777-792. 2009. Elsevier B.V. • C. Ouyang, M. Dumas, W. van der Aalst, A. ter Hofstede, and J. Mendling: From Business Process Models to Process-oriented Software Systems. ACM Transactions on Software Engineering andMethodology (TOSEM). Volume 19, Number 1, pages 2:1-2:37. July 2009. ACM. • J. Mendling, B.F. van Dongen, W.M.P. van der Aalst: GettingRid of OR-Joinsand Multiple Start Events in Business Process Models. Enterprise Information Systems (EIS). Special Issue on EDOC 2007 Best Papers. Volume 2, Number 4, pages 403-419. October 2008. Taylor & Francis.

32. Agenda • WhyConsistencybetween Process Models? • WhyBehaviouralProfiles? • Howtovalidatetheconcept? • Whatarefurtherapplications? • Whattotakehome?

33. Whattotakehome • BehaviouralProfilesprovideusefulabstraction • Profilescanbecalculatedefficiently • Profilescanbeused in variousscenarios