| The interdisciplinary nature of Systems Engineering (SE), having stakeholders from diverse domains with orthogonal facets, and need to consider all stages of lifecycle of system during conception, can benefit tremendously by employing Knowledge Engineering (KE) to achieve semantic agreement among all stakeholders for all stages of life cycle. Present practices in SE, such as using Vocabularies, Processes, Standards, and Modeling Languages are oriented towards manual reasoning of system concepts and relations to achieve semantic agreement. Semantic Web (SW) is being developed to provide KE and automated reasoning to the unstructured information, in huge amounts, present on the World Wide Web, using Knowledge Representation (KR) methods. Borrowing these KE and KR methods from SW and improvising them to solve the problems in SE is the main theme of my research. Employing KE methods from SW in SE requires ontologies that are capable to represent semantic information in a system model. In my research, I introduced Ontology for System Engineering, OSE, for this purpose. Using OSE, a Knowledge Base (KB) is constructed with extracted semantic information from system model. The information is translated, aligned, reasoned and queried, during different stages of proposed framework, to gain semantic agreement. The conception of a system model with agreed semantics, using UML or SysML, is based on requirements model conceived by Requirements Engineering (RE) process. As system design proceeds, the requirements model needs to be integrated into the system model through translation of concepts. Present research in RE involves methodologies such as use of an ontology or a Requirements Modeling Language, which have been developed independently from the system modeling environment. As a consequence, manual integration of requirements and systems models introduces inconsistencies. To overcome this major limitation, I introduced an Ontology for Requirements Engineering (ORE). ORE is designed to be completely compatible with UML and SysML, thus enabling a seamless integration. Complete mapping between ORE (and OSE) and UML (or SysML) is demonstrated. Reasoning is used to infer implicit information in the model and to detect hidden inconsistencies. Two comprehensive case studies have been presented to demonstrate the efficiency of the proposed ontologies and framework. |
