Changing Work ows : On the Interplay Between Flexibility

Foreword The management of business processes comprises their planning, speciication, modeling, analysis and computer-supported execution. Current trends in information technology such as workkow management systems or packaged software systems indicate the increasing importance of a systematic approach to business process management. Many graphical notations are in everyday use for business modeling purposes, usually resembling some version of data ow diagrams. Often, the lack of formal semantics and methodological support limits their computer-based analysis and execution. In most cases, the design and the improvement of business processes in industrial applications does with intuition and experience of experts only. Petri nets have been in use as a graphical modeling language for more than 30 years. Mathematical analysis techniques allow for analytical veriication of many relevant properties of systems' behaviour. Petri nets nevertheless are seldom used in business applications. Even the few commercially available Petri net based software development tools exploit Petri net theory to a very limited extend only. There is an apparent gap between practical needs of business process management in industry and theoretical investigations of Petri nets in the academic sector. Recently, diierent variants of Petri nets have been suggested for various purposes in the context of business processes: As a reference language for the semantics of other, formal or semi-formal modeling languages, or as a means to compare and relate diierent modeling languages, or as a tool to execute process models, or in order to check reachability of distinguished states. Central problems of this area include: 1 Which speciic requirements for modeling languages arise in the area of business process management and workkow management (related to ooce documents, organizational dependencies of tasks, cost, time and quality aspects)? How can the analytical and methodological potential of Petri nets be exploited for business process management? Which requirements are not met by classical Petri net models? Which extensions of the model are sensible or necessary? Will Petri nets eventually play a role for business process modeling which compares to today's role of the relational model for data modeling? Will there develop a normalizing theory for business processes similar to the normalization theory of relational data bases? May Petri nets formally link application oriented modeling languages to particular workkow programming languages? This report contains abstracts from contributions to an IBFI Seminar held in Schloss Dagstuhl from July 6 to July 10, 1998. Moreover, it contains statements of three panel discussions. The seminar had …