Intelligent Concepts for the Management of Information in Workflow Systems

Workflow systems are commonly used in industry, commerce and government. They provide computerized support for owners of repetitive, highly standardized business processes, with a means of controlling the execution of instances of those processes according to predefined process templates. However, many real-life business processes are characterized by various forms of unpredictability and uncertainty. For workflow systems to be applicable in these environments therefore, they must incorporate methods of addressing uncertainty, vagueness, variability, exceptional cases and missing information. Methods that have been previously been applied include dynamic instance adaptation, partial completion and case handling - not to mention manual over-riding in the case of exceptions. Intelligent approaches have included stochastic and fuzzy Petri Nets. In this paper, we discuss the further potential of intelligent concepts, in particular rough set theory, for the support of the management of information in workflow systems. Since its introduction in the beginning of the nineteen eighties, rough set theory has gained increasing attention and has established itself as a useful intelligent concept and an important method within soft computing. We show how rough sets can be utilized to set up an early warning system in cases where information is missing in the workflow system. We also show the potential of rough sets to detect excessive or redundant information in a workflow management system's design.