Toward Multilevel Textual Requirements Traceability Using Model-Driven Engineering and Information Retrieval

Abstract — In complex industrial projects, textual information remains the main vector of information at the project level. Consequently, requirements are scattered throughout multiple documents expressing different levels of requirements and different kinds of requirements. Formalizing this information and tracing different relationships among documents and organizing this environment present a challenging question. Domain-specific modeling and traceability modeling are Model-Driven Engineering (MDE) techniques that could address various aspects of requirements formalization. Text-based high level requirements can be formalized as document concepts can be gathered and represented. Still, relationships cannot always be determined using sole MDE approaches and, as a consequence, relationships and traceability issue remains. Information retrieval (IR) approaches have already proved to work in an efficient way on large text corpora for requirements traceability analysis but do only consider similarity aspects of flatten documents, losing their organization and hierarchy. This paper aims to introduce how a combined use of both MDE and IR can lead to improved requirements organization and traceability while handling textual ambiguous requirements documents