A Novel Software Test Data Generation Framework based on Multi-level Fuzzy Clustering Algorithm

Software testing plays a crucial role in ensuring the quality and reliability of software products. This study presents a novel software test data generation framework based on the multi-level fuzzy clustering algorithm. The framework aims to optimize the testing process by efficiently generating test cases and prioritizing them for execution. The proposed methodology integrates hierarchical reinforcement learning and hierarchical clustering techniques to improve the effectiveness and comprehensiveness of software testing. A detailed review of recent advances in software testing methodologies is provided, focusing on differential regression testing for REST APIs, test case prioritization, program repair using neural translation models, and other key areas. The proposed framework is evaluated through experiments on real-world software datasets, demonstrating its superiority in terms of non-redundancy rate and vulnerability count detection compared to existing methods. The results highlight the effectiveness and relevance of the proposed framework in improving the efficiency and reliability of software testing.