Business process collaborations among multiple partners require particular considerations regarding flexibility and change management. Indeed, each change or process redesign originated by a partner may cause ripple effects on other partners participating in the choreography. Consequently, a change request could spread over partners in an unexpected way with relevant costs due to its transitivity (e.g., in supply chains). In order to avoid costly negotiations or propagation failures, understanding this behavior becomes critical. This paper focuses on analyzing the behavior of change requests in process choreographies, i.e., the change propagation behavior. The input data might be available in two different formats, i.e., as change logs or change propagation logs. In order to understand the data and to explore potential analysis models and techniques, we employ exploratory data analysis as well analysis techniques from process mining and change management to simulation data. The results yield the requirements for designing a mining algorithm that derives the propagation behavior behind change logs. This algorithm is a memetic algorithm that is based on different heuristics. Its feasibility is shown based on a comparison with the other mining techniques.

Top