Expressiveness and suitability of languages for control flow modelling in workflows

Work ow management has been a research area that attracted signi cant interest in the last decade In spite of this little consensus has been reached as to what the essential ingredients of work ow speci cation languages should be Consequently many work ow management systems have been developed that are based on di erent paradigms and include di erent feature sets These di erences result in various levels of suitability and expressive power The challenge which we undertake in this thesis is to provide a comprehensive analy sis of control ow aspects of work ow speci cations We start by analysing modelling languages of eight commercially available work ow systems Next we identify the requirements for work ow control ow through the use of work ow patterns Ranging from very simple to more complex we identify the business requirements that need to be addressed by any work ow management system that is to be used to support a wide range of business processes Through matching these patterns with the modelling capabilities of the work ow systems that we have under review we provide a thorough analysis of these systems and identify shortcomings in their modelling language approaches Finally we establish a theoretical foundation that is subsequently used to provide a formal analysis of the expressive power and suitability of di erent approaches to modelling control ow in work ows We identify four major evaluation strategies and assess their theoretical expressive power Additionally we analyse some of the more advanced concepts related to control ow modelling for work ows such as termination deadlock decomposition and use of data ow to augment control ow speci cation The results obtained as part of this work should not only aid those developing work ow speci cations in practice but also those developing new work ow engines