Interactive and Minimal Repair of Declarative Process Models

We present an approach for resolving inconsistencies in declarative process models while guaranteeing a minimal information loss (w.r.t. the number of deleted elements). To this aim, we show how smallest correction sets, i.e., the smallest sets of constraints that need to be deleted in order to resolve inconsistencies, can be computed via an application of Reiter’s hitting set theorem. In this context, as deleting certain constraints might be highly sensitive or not plausible in a real-life sense, we extend our approach with functionalities for enabling a close human-in-the-loop interaction, such as prioritizing constraints, as well as metrics that offer modelers insights into the impact of deleting constraints. Furthermore, we implement our approach and show that our inconsistency resolution approach outperforms existing approaches in terms of runtime and information loss in experiments with real-life data sets.