Inductive Reference Modelling Based on Simulated Social Collaboration

Abstract

Organizations nowadays possess huge repositories of process models. Inductive reference modelling can save costs and time by reusing process parts of process models belonging to a common domain. The inductive development of a reference model for a large corpus of process models is a difficult problem. Quite a few, primarily heuristic approaches have been proposed to the research community that require an approximate matching between the single processes. With our approach, we introduce a new concept that brings in for the first time an abstract efficiency simulation of the social collaboration around knowledge-based process models. A reference model is assembled featuring at least the topological minimum requirements to be significantly more efficient than the input process models. Our evaluation indicates that the approach is able to generate reference process models that are more efficient than the input process models and at least as a reference model designed by an expert.