|
Zusammenfassung
|
Todays workflow management systems are inflexible and have problems dealing with changes.
As a result, the workflow management system is not used to support dynamically changing workflow
processes or the workflow process is supported in a rigid manner, i.e., changes are not allowed
or handled outside of the workflow management system. Typically, there are two types of changes:
- ad-hoc changes and
- evolutionary changes.
Ad-hoc changes are handled on a case-by-case basis. In order to provide customer specific
solutions or to handle rare events, the process is adapted for a single case or a limited
group of cases. Evolutionary change is often the result of reengineering efforts. This paper
addresses two notorious problems related to adaptive workflow:
- providing management information at the right aggregation level, and
- supporting dynamic change, i.e., migrating cases from an old to a new workflow.
Most of the problems related to change stem from the fact that flexibility (the ability to
handle changes) on the one hand and process support (enactment, control, and management
information) on the other hand impose (partially) conflicting constraints. It is our belief
that the inheritance-preserving transformation rules developed at EUT are a good starting
point for solving the problems identified. The rules correspond to design constructs often
used in practice, namely choice, parallel composition, sequential composition, and iteration.
In fact, the applicability of the inheritance notions and inheritance-preserving transformation
rules is not limited to adaptive workflow. Other application domains are ERP-systems
(reference models), electronic commerce (trade procedures), and O.O. methods.
|
