Workflow Management Systems: Formal Foundation, Conceptual Design, Implementation Aspects

First of all I like to thank Gottfried Vossen for his support of the research reported on in this thesis. In addition, he has been actively involved in the work on WASA from the beginning, and many papers in the early phases of the project were co-authored with him. WASA started as an international cooperation project in 1994, in which the University of Campinas, Brazil, and the University of Münster participated. Valuable insights in the specific properties of scientific applications as well as a generic, workflow-based architecture were developed with our Brazilian partners Prof. Dr. Claudia Bauzer Medeiros and Prof. Dr. João Meidanis. Application studies in the context of workflow management in scientific applications were conducted with our Brazilian partners in molecular biology and geo-processing and with Torsten Reuß in laboratory information management. Their work is gratefully acknowledged. WASA is a team effort. I like to thank Gregor Wit-tkowski and Bernd Focke for their work on implementing the first prototype. Dominik Kuropka, Thomas Serries, Jens Hündling, and Hilmar Schuschel were involved in the development of the second prototype. Many thanks for their interest and involvement in WASA and the valuable work they have put into the system. I like to thank the people to whom we demonstrated the system for their remarks and comments, a considerable number of which have led to improving the system. Among the persons who were patiently enduring WASA demonstrations and asked fruitful questions were Prof. Thanks for the interesting discussions we had and the results we obtained in, let's say, casual environments. Last and foremost I thank Daniela and Jonathan for their patience and support during the months of writing this thesis. Summary This thesis deals with the formal foundation, conceptual design, and prototypical implementation of workflow management systems. Based on properties of information systems applications, research issues in workflow management and requirements of suitable work-flow support are characterized. In particular, the need for flexibility and dynamic adaptations is identified, i.e., the ability of workflow management systems to adapt running workflow instances to changes in the environment of application processes. Re-use of workflow schemata, the integration of application objects as well as high scalability and availability are characterized as additional important properties of workflow management systems. As a formal foundation, a workflow language based on directed graphs is introduced, using a mathematical formalism. In this formalism, workflow schemas and workflow instances are described. Specific properties …