Simplifying Process Model Abstraction: Techniques for Generating Model Names

The increased adoption of business process management approaches, tools, and practices has led organizations
to accumulate large collections of business process models. These collections can easily include from a
hundred to a thousand models, especially in the context of multinational corporations or as a result of
organizational mergers and acquisitions. A concrete problem is thus how to maintain these large repositories
in such a way that their complexity does not hamper their practical usefulness as a means to describe and
communicate business operations. This paper proposes a technique to automatically infer suitable names
for business process models and fragments thereof. This technique is useful for model abstraction scenarios,
as for instance when user-specific views of a repository are required, or as part of a refactoring initiative
aimed to simplify the repository's complexity. The technique is grounded in an adaptation of the theory
of meaning to the realm of business process models. We implemented the technique in a prototype tool
and conducted an extensive evaluation using three process model collections from practice and a case study
involving process modelers with different experience.