Semantic annotation of business process models

In the last decades, business process models have increasingly been used by companies with different purposes, such as documenting enacted processes or enabling and improving the communication among stakeholders (e.g., designers and implementers). Aside from the differences, all the roles played by process models involve human actors (e.g., business designers, business analysts, re-engineers) and hence demand for readability and ease of use, beyond correctness and reasonable completeness. It often happens, however, that process models are large and intricate, thus resulting potentially difficult to understand and to manage. In this thesis we propose some techniques aimed at supporting business designers and analysts in the management of business process models. The core of the proposal is the enrichment of process models with semantic annotations from domain ontologies and the formalization of both structural and domain information in a shared knowledge base, thus opening to the possibility of exploiting reasoning for supporting business experts in their work. In detail, this thesis investigates some of the services that can be provided on top of the process semantic annotation, as for example, the automatic verification of process constraints, the automated querying of process models or the semi-automatic mining, documentation and modularization of crosscutting concerns. Moreover, special care is devoted to support designers and analysts when process models are not available or they have to be semantically annotated. Specifically, an approach for recovering process models from (Web) applications and some metrics for evaluating the understandability of the recovered models are investigated. Techniques for suggesting candidate semantic annotations are also proposed. The results obtained by applying the presented techniques have been validated by means of case studies, performance evaluations and empirical investigations.