Improving Test Data Generation for Critical Paths in Software Programs Through Automation Using an Enhanced Coati Optimization Algorithm

Software testing constitutes approximately ۴۰% of development expenditure, with test data generation being a particularly resource-intensive component. Manual creation of test data frequently introduces inefficiencies, elevated costs, and increased error rates. Automated methods have thus been developed, framing test data generation as an optimization problem to maximize defect detection while minimizing time and cost. Traditional approaches such as Genetic Algorithms (GA) and Particle Swarm Optimization (PSO) are widely used but exhibit limitations. This study utilizes the Coati Optimization Algorithm (COA), which offers reduced complexity, balanced exploration–exploitation phases, and lower sensitivity to initial parameters. However, COA occasionally fails to converge to optimal solutions. To address this, an enhanced corrective search mechanism and a novel fitness function have been introduced to improve convergence speed and solution quality. Additionally, due to the infeasibility of covering all program paths, a new path prioritization strategy based on control flow graph analysis was developed to focus efforts on critical paths. The enhanced COA was evaluated on several benchmark programs and compared with GA, PSO, Pelican Optimization Algorithm (POA), Lyrebird Optimization Algorithm (LOA), Teaching–Learning‑Based Optimization (TLBO), and the original COA. Experimental results indicate that the enhanced COA consistently generates superior test data and enhances testing efficiency. This work presents a robust, optimization-driven framework for automated test data generation and contributes to the advancement of software testing methodologies.

کلیدواژه ها:

Software Testing ، Test Data Generation ، Coati Optimization Algorithm ، Exploration and Exploitation ، Control Flow Graph ، Critical Paths