Semantic recovery of traceability links between system artifacts

Research Projects
Organizational Units
Journal Issue
This paper introduces a mechanism to recover traceability links between the requirements and logical models in the context of critical systems development. Currently, lifecycle processes are covered by a good number of tools that are used to generate different types of artifacts. One of the cornerstone capabilities in the development of critical systems lies in the possibility of automatically recovery traceability links between system artifacts generated in different lifecycle stages. To do so, it is necessary to establish to what extent two or more of these work products are similar, dependent or should be explicitly linked together. However, the different types of artifacts and their internal representation depict a major challenge to unify how system artifacts are represented and, then, linked together. That is why, in this work, a concept-based representation is introduced to provide a semantic and unified description of any system artifact. Furthermore, a traceability function is defined and implemented to exploit this new semantic representation and to support the recovery of traceability links between different types of system artifacts. In order to evaluate the traceability function, a case study in the railway domain is conducted to compare the precision and recall of recovery traceability links between text-based requirements and logical model elements. As the main outcome of this work, the use of a concept-based paradigm to represent that system artifacts are demonstrated as a building block to automatically recover traceability links within the development lifecycle of critical systems.
software traceability, software system artifact representation, software reuse
Bibliographic citation
Álvarez-Rodríguez, J.M., Mendieta, R., Moreno, V., Sánchez-Puebla, M., Llorens, J. (2020). Semantic recovery of traceability links between system artifacts. International Journal of Software Engineering and Knowledge Engineering, 30(10), pp. 1415-1442