A methodology for the semantic and structural restructuring of BPMN models

Abstract

Purpose Restructuring a business process (BP) model may enhance the BP performance and improve its understandability. So-far proposed restructuring methods use either refactoring which focuses on structural aspects, social network discovery which uses semantic information to guide the affiliation process during its analysis, or social network rediscovery which uses structural information to identify clusters of actors according to their relationships. The purpose of this paper is to propose a hybrid method that exploits both the semantic and structural aspects of a BP model. Design/methodology/approach The proposed method first generates a social network from the BP model. Second, it applies hierarchical clustering to determine the performers’ partitions; this step uses the social context which specifies features related to performers, and two new distances that account for semantic and structural information. Finally, it applies a set of behavioral and organizational restructuring rules adapted from the graph optimization domain; each rule uses the identified performers’ partitions and the business context to reduce particular quality metrics. Findings The efficiency of the proposed method is illustrated through well-established complexity metrics. The illustration is made through the development of a tool that fully supports the proposed method and proposes a strategy for the application of the restructuring rules. Originality/value The proposed method has the merit of combining the semantic and structural aspects of a Business Process Modeling Notation model to identify restructuring operations whose ordered application reduces the complexity of the initial model.

Cite this work

No public copy is stored for this record, so only the metadata is cited.

Stored only when both papers are in the corpus — 16 citations exist upstream.

Papers that cite the same references — the relational fallback while embeddings are still backfilling.