Weighted Rank Ant Colony Metaheuristics Optimization Based Test Suite Reduction in Combinatorial Testing for Improving Software Quality

Software Product Lines (SPLs) contains a large number of feature combinations therefore pose challenges for testing software application program. In combinatorial testing, pair wise coverage maximization and test suite reduction plays a major role to reduce the testing cost of software program application. Few research works have been developed for combinatorial testing using different test suite reduction techniques. But, the conventional genetic algorithm does not present an optimal solution for test suite optimization problem because of process of randomly mutating bits of chromosome in mutation operation. In addition to that, genetic algorithm requires more computational time. In order to solve this limitation, Weighted Rank Ant Colony Metaheuristic Test Suite Optimization (WRACMTSO) Technique is proposed. WRACMTSO Technique designed a Weighted Rank Ant Colony Metaheuristic Optimization algorithm for test suite reduction in combinatorial testing. The WRACMO algorithm begins with initialization of ACO parameters with help of test cases in test suite. Then, ant chooses the one of two vertices to reach food source using trail's probability with weighted rank value and consequently updates pheromone trails. This process is repetitive until the ant attains food source. In WRACMO algorithm, vertices (test cases) elected by an ant to arrive at the food source are considered as optimal for performing combinatorial testing. By using this process, WRACMTSO Technique efficiently optimizes the test cases in test suite with minimum time. Therefore, WRACMTSO Technique enhances the test suite reduction rate and reducing the computational time of test cases for efficient combinatorial testing. The WRACMTSO Technique conducts the experimental works on metrics such as test suite reduction rate, computational time, testing cost and coverage rate with respect to different number of test suite and size of software product lines.