Using semantics and reasoning to detect inconsistency in UML class diagrams

Software designers typically work in different design teams and use different sets of terminologies with possibly different intended meanings. This makes inconsistencies in UML class diagrams become inconspicuous. Based on the existing literature, yet, there is no approach found to supporting designers clarify conceptual ideas used in their models while detecting inconsistencies. Furthermore, while there are a number of works in inconsistency detection for UML class diagrams, none of them provides a comprehensive classification of inconsistencies. In this thesis, we provide an inconsistency classification for UML class diagrams. The classification is based on the results of an analysis of inconsistencies in literature. We propose a formal approach to detecting inconsistencies using Description Logic (DL) and Semantic Web Rule Language (SWRL). Based on our approach, designers can clarify their conceptual ideas through an ontology constructed from the (DL + SWRL) knowledge base. We validate our ideas through a number of scenarios. These scenarios are based on class diagrams extracted from literature which are prototypical.