Descriptive business process models at run-time

Redlich, David (2018) Descriptive business process models at run-time. PhD thesis, UNSPECIFIED.

[img]
Preview
PDF (2018redlichphd)
thesis.pdf - Published Version
Available under License Creative Commons Attribution-NoDerivs.

Download (11MB)

Abstract

Today's competitive markets require organisations to react proactively to changes in their environment if financial and legal consequences are to be avoided. Since business processes are elementary parts of modern organisations they are also required to efficiently adapt to these changes in quick and flexible ways. This requirement demands a more dynamic handling of business processes, i.e. treating business processes as run-time artefacts rather than design-time artefacts. One general approach to address this problem is provided by the community of models@run.time, which promotes methodologies concerned with self-adaptive systems where models reflect the system's current state at any point in time and allow immediate reasoning and adaptation mechanisms. However, in contrast to common self-adaptive systems the domain of business processes features two additional challenges: (i) a bigger than usual abstraction gap between the business process models and the actual run-time information of the enterprise system and (ii) the possibility of run-time deviations from the planned models. Developing an understanding of such processes is a crucial necessity in order to optimise business processes and dynamically adapt to changing demands. This thesis explores the potential of adopting and enhancing principles and mechanisms from the models@run.time domain to the business process domain for the purpose of run-time reasoning, i.e. investigating the potential role of Descriptive Business Process Models at Run-time (DBPMRTs) in the business process management domain. The DBPMRT is a model describing the enterprise system at run-time and thus enabling higher-level reasoning on the as-is state. Along with the specification of the DBPMRT, algorithms and an overall framework are proposed to establish and maintain a causal link from the enterprise system to the DBPMRT at run-time. Furthermore, it is shown that proactive higher-level reasoning on a DBPMRT in the form of performance prediction allows for more accurate results. By taking these steps the thesis addresses general challenges of business process management, e.g. dealing with frequently changing processes and shortening the business process life cycle. At the same time this thesis contributes to research in models@run.time by providing a complex real-world use case as well as a reference approach for dealing with volatile models@run.time of a higher abstraction level.

Item Type:
Thesis (PhD)
ID Code:
124261
Deposited By:
Deposited On:
28 Mar 2018 08:28
Refereed?:
No
Published?:
Published
Last Modified:
27 Sep 2020 07:22