Dipartimento di Ingegneria informatica, automatica e gestionale

So far, business process monitoring approaches have mainly focused on monitoring executions with respect to a single process model. This setting aptly captures monolithic scenarios from domains in which all possible behaviors can be folded into a single model. However, this strategy cannot be applied to domains where multiple interacting (procedural) sub-processes work under additional (declarative) constraints. For example, in healthcare, co-morbid patients may be subject to multiple clinical pathways at once, in the presence of additional, general constraints capturing basic medical knowledge. To support monitoring of thus emerging hybrid specifications, we propose a Multi-Model Monitoring Framework. On the one hand, the framework allows for a hybrid representation of a process, using both procedural and declarative models. This admits more flexible process model design as domain experts can focus on specific procedures and domain constraints without needing to merge them into one single specification. On the other hand, the framework includes an automata-based monitoring technique to simultaneously account for multiple models within one execution while resolving conflicts caused by the interplay of such models. We describe the overall framework, report on a prototypical implementation of the monitoring technique, and demonstrate its feasibility with a healthcare scenario.