Generalised computation of behavioural profiles based on petri-net unfoldings

Abstract. Behavioural profiles have been proposed as a concept to judge on the behavioural consistency of process models that depict different perspectives of a process. These profiles describe the observable relations between the activities of a process model. Consistency criteria based on behavioural profiles are less sensitive to model projections than common equivalence criteria, such as trace equivalence. Existing algorithms derive those profiles for unlabelled sound free-choice workflow nets efficiently. In this paper, we generalise the computation of behavioural profiles by relaxing the aforementioned assumptions. First, we introduce an algorithm that derives behavioural profiles from the complete prefix unfolding of a bounded Petri net. Hence, it is applicable in a more general case. Second, we lift the concept to the level of labelled Petri nets. We also elaborate on findings of applying our approach to a collection of industry models.