A heuristic approach for discovering reference models by mining process model variants

Abstract. Recently, a new generation of adaptive Process-Aware Infor-mation Systems (PAISs) has emerged, which enables structural process changes during runtime while preserving PAIS robustness and consis-tency. Such flexibility, in turn, leads to a large number of process variants derived from the same model, but differing in structure. Generally, such variants are expensive to configure and maintain. This paper provides a heuristic search algorithm which fosters learning from past process changes by mining process variants. The algorithm discovers a reference model based on which the need for future process configuration and adap-tation can be reduced. It additionally provides the flexibility to control the process evolution procedure, i.e., we can control to what degree the discovered reference model differs from the original one. As benefit, we can not only control the effort for updating the reference model, but also gain the flexibility to perform only the most important adaptations of the current reference model. Our mining algorithm is implemented and evaluated by a simulation using more than 7000 process models. Simula-tion results indicate strong performance and scalability of our algorithm even when facing large-sized process models.