Artifact-driven process monitoring: dynamically binding real-world objects to running processes

Monitoring inter-organizational business processes requires explicit knowledge about when activities start and complete. This is a challenge because no single system controls the process, activities might not be directly recorded, and the overall course of execution might only be determined at runtime. In this paper, we address these problems by integrating process monitoring with sensor data from real-world objects. We formalize our approach using the E-GSM artifact-centric language. Since the association between real-world objects and process instances is often only determined at runtime, our approach also caters for dynamic binding and unbinding at runtime.