Ibuprofen: Improvement and Business Process Refactoring of Embedded Noise

Business process modeling has proven to be beneficial for both enterprise management and software development. As a consequence of this relevance, reverse engineering techniques have appeared for retrieving business process models when these do not exist or are outdated (i.e., from existing information system). Unfortunately, such retrieved business process models may contain quality faults like a lack of completeness, non-relevant elements, ambiguity, among other. These quality faults decrease the understandability and modifiability level of these models. In order to deal with these faults, refactoring has been wide-used in combination with reverse engineering techniques, which modify the internal structure of business process models while semantic is preserved. However, additional knowledge sources are also needed to enrich the semantic of retrieved business process models since the business knowledge is distributed along several sources. Despite the fact there are several studies focused on business process model refactoring, there are no refactoring operators specially designed for those that are retrieved by reverse engineering and undergo the mentioned quality faults. The goal of this Thesis is therefore to provide a refactoring framework to fix these quality faults and thus achieving business process models more accurate, understandable and modifiable. The research proposal, IBUPROFEN, is a framework to improve and refactor business process models retrieved from information systems. IBUPROFEN provides a set of refactoring operators to cope with quality defects of these models. Moreover, IBUPROFEN supplies the best combination of refactoring operators to attain the greatest gain on understandability and modifiability depending of some features of business process models. Additionally, IBUPROFEN proposes a semantic enrichment technique that combines business process models and alternative sources like event logs from which to extract additional knowledge as for example non-retrieved connections between business elements. IBUPROFEN, together with its supporting tool, proved to be suitable to increase the understandability and modifiability of business process models. This result was obtained through the conduction of various case studies. Refactored business process models were proved to be more understandable and modifiable than retrieved ones. Moreover, IBUPROFEN is less time-consuming than a manual refactoring of business process models. As a result, IBUPROFEN provides to engineers with a mechanism for enhancing business process models obtained from existing information systems by reverse engineering