TOTeM: Temporal Object Type Model for Object-Centric Process Mining

System behavior emerges from multiple subprocesses operating on interacting objects of different types. The relations between types of subprocesses are essential to understand the system’s behavior. This includes temporal relations as well as cardinalities of relationships. Whether a product is produced before or after a customer order, which is essential with respect to lean management, would be an example of a temporal relationship. Conversely, the number of products for each customer order would be an example of a cardinality constraint. Current object-centric process modeling approaches focus on precedence constraints between activities, which is not sufficient to capture temporal and cardinality relationships between types. This paper introduces the temporal object type model (TOTeM) to model and discover process-specific type-level relations. We propose three type-level relations: a temporal relation, an overall log cardinality, and an event cardinality. The contributions of this paper include a definition of the temporal object type model, an algorithm to compute them, a publicly available implementation, and an evaluation.