Degree Name

Doctor of Philosophy


School of Electrical, Computer and Telecommunications Engineering - Faculty of Engineering and Information Sciences


Business process management has emerged as a focus of considerable industry and research interest in recent times. Organizations routinely find themselves in situations where they need to manage a very large number distinct processes. A single organization might need to manage hundreds or even thousands of business process models that are stored in a process repository. In addition, the complexity of individual process models (designs) has also grown. Business process designs within an enterprise process repository are often closely interrelated. Some designs represent reference process models, while others are context-specific generalizations or specializations of other process models. Processes are also interrelated via choreographies. Further, an organization may practice a continuous improvement approach such that it needs to frequently change its interrelated business processes in order to maintain its performance.

In such changes, an initial modification made to a process model may impact a number of other process models in the repository in order to preserve their interdependencies. However, it is not trivial to manage such changes in a process repository consisting of a large number of process models. Such change management imposes us to clearly define and to precisely preserve any existing inter-process relationship. Currently, there are not enough approaches dealing with how to define inter-process relationships which may occur in a process repository and how to maintain such relationships which may be perturbed due to changes made on a process model. We, however, recognize a little work on preserving any inter-process relationship due to process changes in process model collections through change propagation. Further, defining relationships among process models still requires more works.

Therefore, in this dissertation, we introduce process ecosystems to view large and complex business process repositories, which emphasize the identification and maintenance of normative inter-process relationships. We argue this framework can deliver significant value to fill the existing gap in dealing with process change management in a process repository. We leverage the Software Engineering Research Methodology (SERM) including conceptualization, formalization and development processes in order to construct the proposed framework.

During conceptualization, we identify problems in managing changes in business process repositories and ground them with the theoretical constructs. In the formalization process, we have identified and formalized three inter-process relationship types, i.e. part-whole, inter-operation and generalization-specialization, leveraging a machinery for semantic effect annotation of process models. In the development process, we further develop process ecosystem framework as the solution for problems identified in dealing with process changes management in a process repository. We leverage techniques from constraint networks to define procedures to manage and propagate change in process ecosystems. We then evaluate such proposed framework to study its performance and accuracy. Our experimental results suggest that such procedures are efficient to propagating changes within medium-sized process repositories. Finally, we expect this new framework to significantly assist process analysts in dealing with process changes management in process repositories.

FoR codes (2008)




Unless otherwise indicated, the views expressed in this thesis are those of the author and do not necessarily represent the views of the University of Wollongong.