A Comparison of Four Metaheuristic Algorithms for the Problem of Test Redundancy Reduction

Abstract. Finding the optimal solution out of all reasonable solutions is the goal of an optimization problem. Numerous metaheuristic algorithms have been created in the literature during the past 30 years. It is essential to assess each algorithm's performance using broad case studies in order to assist engineers in selecting the optimal metaheuristic algorithm for the given problem. In this research, we give a comparative analysis of four metaheuristic algorithms used to solve the test redundancy reduction problem: the teaching-learning-based optimization (TLBO), the jaya algorithm (JA), the sine-cosine algorithm (SCA), and the sparrow-search algorithm (SSA). To achieve statistical significance, the evaluation of these algorithms' performance is carried out by running every algorithm thirty times. Finding a minimal subset of the test suite that satisfies the specified test criteria is the aim of test redundancy reduction. It was discovered that, in terms of average reduction rate and runtime effectiveness, the SSA is the more efficient metaheuristic algorithm for the test redundancy reduction problem among all the competing metaheuristic algorithms.