Maintaining Requirements and Test Cases Traceability in an Agile Environment
Lucas Sales Vieira, Regina Moraes · 2025
The increasing adoption of agile methodologies in software development has highlighted challenges related to the dynamic management of requirements and the continuous updating of test cases. In an environment where changes are frequent and requirements evolve rapidly, ensuring the quality and consistency of testing is essential for project success. This problem is common among various companies that adopt agile methodologies, facing difficulties in maintaining software quality while managing frequent changes in requirements. In this context, this research proposes a semi-automated approach to managing requirements and test cases by automatically identifying the impacts of requirement changes on test cases. For this purpose, the natural language process is applied with the support of the BERT language model. The methodology proposed in this study aims to enable the addition, editing, and removal of requirements and test cases in an integrated manner, ensuring that all changes are reflected efficiently and accurately. The suggested solution is based on tokenization and embeddings provided by the BERT model, combined with cosine similarity analysis, to identify the requirements and test cases that may be affected by changes in the requirement set. It is expected that the approach brings significant improvements in the efficiency of the requirement and test case management process and, consequently, in the quality of the software developed. By automatically identifying and updating test cases, it should contribute to reducing failures and rework, resulting in a more agile development process aligned with stakeholder expectations.