| Summary: | Workflow management systems enable organisations to deal with all aspects of business process management, including analysis, modelling, execution, and administration. Modelling workflow processes involves transformation of the process logic into a formal representation and it always remains a critical success factor for these systems. Workflow modelling languages provide constructs for capturing high-level descriptions of business processes, which are then have to be transformed and encoded into low-level execution semantics with the use of workflow programming languages. However, maintaining these models separately results in a number of issues, particularly when the various interdependencies between them are managed manually. This primarily creates difficulties in adaptation, in terms of identifying changes in high-level descriptions due to modifications of business conditions, and tracing the impact of those changes on the low-level execution semantics. Moreover, certain information included in the high-level descriptions is either partly encoded or omitted from the low-level execution semantics and at the same time, complicated business rules encoded at the execution level are not included in the high-level descriptions, creating major inconsistencies. The above issues result in high maintenance costs, reducing the overall efficiency and performance of workflow management systems. This thesis addresses the aforementioned problems by proposing a framework named SYMEX. SYMEX addresses the issue of integrating high and low-level descriptions in one unified format, from a Systems Theory perspective. SYMEX models have a mathematically defined formalism capable of capturing both high-level descriptions of business processes and low-level workflow execution semantics. Furthermore, SYMEX offers a concise and easy to learn and communicate set of constructs, allowing business analysts, process designers, and programmers to work on the same model, at different levels of abstraction. Apart from the theoretical framework, an XMLbased approach for the application of SYMEX is proposed, along with a constraint- based inference engine. Additionally, SYMEX models are evaluated in terms of their complexity and prove easier to read, understand, and manage than other traditional workflow modelling approaches. However, further research is required to assess the capability of the framework, with respect to modelling workflow processes in a service-oriented environment, where activities of business processes are essentially web-services exposed on the Internet.
|
|---|
