Integration and verification of semantic constraints in adaptive process management systems

  • Authors:

  • Linh Thao Ly;Stefanie Rinderle;Peter Dadam

  • Affiliations:

  • Ulm University, Faculty of Engineering and Computer Science, Institute of Databases and Information Systems, James-Franck-Ring, 89069 Ulm, Germany;Ulm University, Faculty of Engineering and Computer Science, Institute of Databases and Information Systems, James-Franck-Ring, 89069 Ulm, Germany;Ulm University, Faculty of Engineering and Computer Science, Institute of Databases and Information Systems, James-Franck-Ring, 89069 Ulm, Germany

  • Venue:
  • Data & Knowledge Engineering
  • Year:
  • 2008

Quantified Score

Hi-index 0.00

Visualization

Abstract

Adaptivity in process management systems is key to their successful applicability in practice. Approaches have been already developed to ensure system correctness after arbitrary process changes at the syntactical level (e.g., avoiding inconsistencies such as deadlocks or missing input parameters after a process change). However, errors may be still caused at the semantical level (e.g., violation of business rules). Therefore, the integration and verification of domain knowledge will flag a milestone in the development of adaptive process management technology. In this paper, we introduce a framework for defining semantic constraints over processes in such a way that they can express real-world domain knowledge on the one hand and are still manageable concerning the effort for maintenance and semantic process verification on the other hand. This can be used to detect semantic conflicts (e.g., drug incompatibilities) when modeling process templates, applying ad hoc changes at process instance level, and propagating process template modifications to already running process instances, even if they have been already individually modified themselves; i.e., we present techniques to ensure semantic correctness for single and concurrent changes which are, in addition, minimal regarding the set of semantic constraints to be checked. Together with further optimizations of the semantic checks based on certain process meta model properties this allows for efficiently verifying processes. Altogether, the framework presented in this paper provides the basis for process management systems which are adaptive and semantic-aware at the same time.