Towards a Taxonomy of Process Flexibility

Abstract

E®ective business processes must be able to accommodate changes in the environment in which they operate. Many approaches have been proposed in literature and some of these approaches have been implemented in °exible work°ow management systems. However, a comprehensive classi¯cation of the various approaches has been missing. In this paper, we take a ¯rst step towards a taxonomy of process °exibility by distinguishing four types of process °exibility based on an extensive literature study. An evaluation shows that each of the types can be found in contemporary o®erings.

Introduction

The need for process °exibility has long been recognised [ 8, 15 ] in the work°ow and process technology communities as a critical quality of e®ective business processes to adapt to changing business circumstances, e.g., new business strategies. The notion of °exibility is often viewed in terms of the ability of an organisation's processes and supporting technologies to adapt to these changes [ 7, 16 ]. Others [ 12 ] consider °exibility from the opposite perspective, i.e., they focus on the part of the process which remain unchanged, rather than focusing on which parts have to be changed. Indeed, a process can only be considered to be °exible if it is possible to change it without needing to replace it completely [ 13 ].

There have been a series of proposals for classifying °exibility [ 2, 6, 8, 13 ], both in terms of the factors which motivate it and the ways in which it can be achieved within business processes. The individual °exibility types discussed in this paper are informed by a multitude of research initiatives in the work°ow and BPM ¯elds. Nevertheless, a comprehensive overview of distinct approaches has been missing.

Based on an extensive literature study [ 9 ], in Section 2 we identify four distinct types of process °exibility that improve the ability of business processes to respond to changes in their operating environment without necessitating a complete redesign of the underlying process speci¯cation. Then in Section 3, we explore the support of the °exibility types and show that each of them exists in contemporary o®erings. Finally, the conclusion and future work are presented in Section 4.

Flexibility types

First, we present four distinct types of process °exibility and describe how each of the °exibility types operates [ 9 ].

Flexibility by design: for handling anticipated changes in the operating environment, where supporting strategies can be de¯ned at design-time. Flexibility by deviation: for handling occasional unforeseen behaviour, where di®erences with the expected behaviour are minimal.

Flexibility by underspeci¯cation: for handling anticipated changes in the operating environment, where strategies cannot be de¯ned at design-time, because the ¯nal strategy is not known in advance or is not generally applicable.

Flexibility by change: either for handling occasional unforeseen behaviour, where di®erences require process adaptations, or for handling permanent unforeseen behaviour.

Figure 1 illustrates the distinction between the °exibility types in isolation, in terms of the time at which the speci¯c °exibility options need to be con¯gured (1) at design-time, as part of the process de¯nition, or (2) at run-time via facilities in the process execution environment. It also shows the relative completeness of the process de¯nition for each °exibility type at run-time.

Flexibility by underspeci¯cation works on the basis of an incomplete process de¯nition. Combined with late binding only, it just o®ers design-time con¯guration options, i.e., only the fragments that have been de¯ned during designtime can be selected at run-time. However, when combined with late modelling, also run-time con¯guration options are o®ered by providing means to de¯ne and select fragments at run-time. In the spectrum of options, °exibility by design distinguishes itself by being the °exibility type that is only con¯gurable at design-time, i.e., at design-time the set of possible execution paths is ¯xed. Hence increasing °exibility corresponds to making this set bigger. Both °exibility by deviation and change work with complete process de¯nitions. For both types, the con¯guration options are only available at run-time. Although very similar, only °exibility by change a®ects the process de¯nition both at instance and type level, whereas °exibility by deviation does not a®ect the process de¯nition at all, i.e., the model and the reality no longer match after deviation. 3

Evaluation of Contemporary O®erings

To validate the °exibility types, we investigate the degree of support of them in di®erent Process Aware Information Systems (PAISs) [ 9 ]. We consider ADEPT1 [ 14 ], YAWL1 (version 8.2b) [ 1, 4, 5 ], FLOWer (version 3.0) [ 3 ] and DECLARE (version 1.0) [ 10,11 ]. The selection of these PAISs has been based on the criterion of supporting process °exibility, which excludes classical work°ow systems and 1 The evaluation of YAWL includes the so-called Worklet Service [ 4, 5 ].

Change Deviation Underspecification (Late modelling) Flexibility Configuration

Run−time

E 1 r R

e A PT L WL

ED AWLO EC Flexibility by A Y F D design + + + + deviation { { + + underspeci¯cation { + { { change + { { + most commercial systems. Moreover, the selected systems cover distinct areas of the PAIS technology spectrum, such as adaptive work°ow (ADEPT1), case handling (FLOWer) and declarative work°ow (DECLARE).

The evaluation results are summarized in Figure 2, which shows whether the PAIS supports (+) or does not support ({) the respective °exibility type. A detailed evaluation can be found in [ 9 ]. None of the evaluated systems provides the full range of °exibility alternatives. Flexibility by design is (to some degree) supported by all o®erings. YAWL excels in °exibility by underspeci¯cation, ADEPT1 in °exibility by change, FLOWer in °exibility by deviation and DECLARE excels in two areas, namely deviation and change. 4

Conclusion

In this paper we have identi¯ed four distinct types of process °exibility. These types are based on a wide variety of proposals for process °exibility found in literature and in practice. The distinction is based on the °exibility con¯guration (i.e., design-time vs. run-time), the completeness of the process de¯nition and whether it be done without changing the process de¯nition, or not. Our evaluation shows that each of the types exists in contemporary o®erings. Interestingly, none of these o®erings supports all °exibility types. As future work, we plan to formalise the identi¯ed °exibility types in a taxonomy of process °exibility.

References

1. W.M.P. van der Aalst and A.H.M. ter Hofstede . YAWL: Yet Another Work°ow Language. Information Systems , 30 ( 4 ): 245 { 275 , 2005 .

2. W.M.P. van der Aalst and S. Jablonski . Dealing with Work°ow Change: Identi¯cation of Issues and Solutions . International Journal of Computer Systems , Science, and Engineering, 15 ( 5 ): 267 { 276 , 2000 .

3. W.M.P. van der Aalst , M. Weske, and D. GruÄnbauer. Case Handling: A New Paradigm for Business Process Support . Data and Knowledge Engineering , 53 ( 2 ): 129 { 162 , 2005 .

4. M. Adams , A.H.M. ter Hofstede , W.M.P. van der Aalst , and D. Edmond . Dynamic, Extensible and Context-Aware Exception Handling for Work°ows . In F. Curbera, F. Leymann , and M. Weske, editors, Proceedings of the OTM Conference on Cooperative information Systems (CoopIS 2007 ), volume 4803 of Lecture Notes in Computer Science, pages 95 { 112 . Springer-Verlag, Berlin, 2007 .

5. M. Adams , A.H.M. ter Hofstede , D. Edmond , and W.M.P. van der Aalst. Worklets: A Service-Oriented Implementation of Dynamic Flexibility in Work°ows . In R. Meersman and Z. Tari et al., editors, Proceeding of the OTM Conference on Cooperative Information Systems (CoopIS 2006 ), volume 4275 of Lecture Notes in Computer Science, pages 291 { 308 . Springer-Verlag, Berlin, 2006 .

6. S. Carlsen , J. Krogstie , A. S¿lvberg , and O.I. Lindland . Evaluating Flexible Work°ow Systems . In Proceedings of the Thirtieth Hawaii International Conference on System Sciences (HICSS-30) , Maui, Hawaii, 1997 . IEEE Computer Society Press.

7. F. Daoudi and S. Nurcan . A Benchmarking Framework for Methods to Design Flexible Business Processes . Software Process Improvement and Practice , 12 : 51 { 63 , 2007 .

8. P. Heinl , S. Horn , S. Jablonski , J. Neeb , K. Stein , and M. Teschke . A Comprehensive Approach to Flexibility in Work°ow Management Systems . In G. Georgakopoulos, W. Prinz , and A .L. Wolf, editors, Work Activities Coordination and Collaboration (WACC'99) , pages 79 { 88 , San Francisco, February 1999 . ACM press.

9. M.H. Schonenberg , R.S. Mans , N.C. Russell , N.A. Mulyar , and W.M.P. van der Aalst . Towards a Taxonomy of Process Flexibility (Extended Version) . BPM Center Report BPM-07-11 , BPMcenter.org, 2007 .

10. M Pesic and W.M.P. van der Aalst . A Declarative Approach for Flexible Business Processes Management . In Business Process Management Workshops , pages 169 { 180 , 2006 .

11. M. Pesic , M. H. Schonenberg , N. Sidorova , and W.M.P. van der Aalst. ConstraintBased Work°ow Models: Change Made Easy . In F. Curbera, F. Leymann , and M. Weske, editors, Proceedings of the OTM Conference on Cooperative information Systems (CoopIS 2007 ), volume 4803 of Lecture Notes in Computer Science, pages 77 { 94 . Springer-Verlag, Berlin, 2007 .

12. G. Regev, I. Bider, and A. Wegmann. De ¯ning Business Process Flexibility with the Help of Invariants . Software Process Improvement and Practice , 12 : 65 { 79 , 2007 .

13. G. Regev, P. So ®er, and R. Schmidt . Taxonomy of Flexibility in Business Processes . In Proceedings of the 7th Workshop on Business Process Modelling, Development and Support(BPMDS'06) , 2006 . http://lamswww.epfl.ch/conference/bpmds06/ taxbpflex.

14. M. Reichert , S. Rinderle , and P. Dadam . Adept work°ow management system . In W.M.P. van der Aalst , A.H.M. ter Hofstede , and M. Weske, editors, Business Process Management, International Conference, BPM 2003, Eindhoven, The Netherlands , June 26-27, 2003 , Proceedings, volume 2678 of Lecture Notes in Computer Science, pages 370 { 379 . Springer, 2003 .

15. H.A. Reijers . Work°ow Flexibility: The Forlorn Promise . In 15th IEEE International Workshops on Enabling Technologies: Infrastructures for Collaborative Enterprises (WETICE 2006 ), 26 - 28 June 2006, Manchester, United Kingdom, pages 271 { 272 . IEEE Computer Society, 2006 .

16. R.A. Snowdon , B.C. Warboys , R.M. Greenwood , C.P. Holland , P.J. Kawalek , and D.R. Shaw . On the Architecture and Form of Flexible Process Support . Software Process Improvement and Practice , 12 : 21 { 34 , 2007 .