Abstract

Over the last decade, workflow management has become a significant tool in the effort of organizations to improve the efficiency of their processes. However, the scope for its adoption has been constrained by the variability, uncertainty and impreciseness that is inherent in business process execution. A number of attempts have been made by the workflow community to address this weakness, but no clear winner has emerged. In this paper, we consider the potential for applying Rough Set theory. One widely used modelling language for workflows is that offered by Petri Nets. In any process, the focal points of uncertainty are when decisions have to be taken. In the Petri Net model, this is represented by resolving, for each transition, whether it should fire or not. In particular at each OR-split (conditional branch) in a process instance, one has to decide the route to be taken by the tokens in the Petri Net. In our paper, therefore, we point out the potential of rough sets to resolve cases where contradictory information is available - or where information is missing. We introduce the concepts of rough places, rough tokens and rough transitions, and show how they can be utilized for workflow management. © 2008 Springer-Verlag Berlin Heidelberg.