A Hybrid Sine Cosine Algorithm with Cluster Voting for Combinatorial Testing

Combinatorial testing is essential for detecting system faults with minimal test cases. However, traditional methods often struggle to efficiently explore large search spaces, leading to redundant test cases and increased computational costs. This paper introduces a novel hybrid optimization framework, HCV-SCA, which integrates the Sine Cosine Algorithm (SCA), JAYA, and Teaching-Learning-Based Optimization (TLBO) with a dynamic Cluster Voting (CV) mechanism. The CV mechanism dynamically selects and adjusts search operators, penalizing or rewarding them based on their performance, ensuring the production of optimal test suites. Unlike random or exhaustive methods, metaheuristic algorithms like HCV-SCA intelligently navigate complex search spaces, generating test suites with improved coverage and reduced redundancy. Experimental evaluations using benchmark combinatorial testing problems demonstrate that HCV-SCA outperforms existing methods, particularly excelling in smaller test suite configurations (t = 2 to 7) and maintaining competitive performance for larger configurations (t = 8 to 10). The results show a significant reduction in test suite sizes, enhanced fault detection rates, and improved computational efficiency. The adaptive nature of the Cluster Voting mechanism ensures robustness across diverse testing scenarios, making HCV-SCA a powerful and flexible solution for optimizing combinatorial testing. This research sets a new benchmark for effective test case generation in software testing, addressing both efficiency and adaptability in complex systems.