­Automated Generation of Covering Array Using Model Checking and Modified Gravitational Search Algorithm

Abstract

The widespread usage of combinatorial testing methodologies in software development has necessitated efficient strategies for generating high-quality test suites. Covering Array (CA) has emerged as a significant component of combinatorial testing, offering various types that cater to different testing requirements. Several strategies have been proposed to generate CAs, each with its unique trade-offs in terms of performance and efficiency. However, there remains a lack of a strategy that simultaneously addresses both aspects effectively. Furthermore, manually acquiring software information introduces a higher chance of errors and also poses challenges in terms of the complexity involved in extracting relevant data. To overcome these challenges, this research incorporates the powerful GROOVE model checker to automate the extraction of variables and their interactions within the software. Subsequently, by modifying the Gravitational Search Algorithm (GSA), an optimal test suite is produced with improved efficiency. This paper aims to construct a software model using the GROOVE model checker and leverage its features to extract pertinent software information. The proposed methodology then employs a modified Gravitational Search Algorithm (MGSA) to generate CAs with both uniform and variable strength. Additionally, a mechanism is introduced to expedite search operations within data structures. To evaluate the effectiveness of the proposed solution, it is implemented within the GROOVE environment, alongside several other meta-heuristic algorithms. Furthermore, the proposed algorithm is also implemented externally for comparison with existing strategies. The evaluation results demonstrate that the proposed solution outperforms other strategies in terms of efficiency and performance.