African Buffalo Optimization Algorithm Based T-Way Test Suite Generation Strategy for Electronic-Payment Transactions

• Main Authors: Odili, Julius Beneoluchi, Noraziah, Ahmad, Nasser, Abdullah B., Abd. Wahab, Mohd Helmy, Ahmed, Mashuk
• Format: Conference or Workshop Item
• Language: English; English
• Published: Springer, Cham 2021
• Subjects: QA75 Electronic computers. Computer science
• Online Access: http://umpir.ump.edu.my/id/eprint/33101/7/African%20Buffalo%20Optimization%20Algorithm%20Based%20T-Way.pdf; http://umpir.ump.edu.my/id/eprint/33101/8/African%20Buffalo%20Optimization%20Algorithm%20Based%20T-Way1.pdf; http://umpir.ump.edu.my/id/eprint/33101/; https://link.springer.com/chapter/10.1007/978-3-030-82616-1_15; https://doi.org/10.1007/978-3-030-82616-1_15

The use of meta-heuristics in Combinatorial Interaction Testing (CIT) is becoming more and more popular due to their effectiveness and efficiency over the traditional methods espe-cially in authenticating electronic payment (e-payment) transactions. Concomitantly, over the past two decades, there has been a rise both in the development of metaheuristics and their application to diverse theoretical and practical areas including CIT in e-payments. In the implementation of t-way strategies (the t is used to represent the interaction strength), mixed results have been reported; some very exciting but, in other cases, the performance of metaheuristics has been, to say the least, below par. This mixed trend has led many re-searchers to explore alternate ways of improving the effectiveness and efficiency of me-taheuristics in CIT, hence this study. It must be emphasized, however, that available litera-ture indicates that no particular metaheuristic testing strategy has had consistent superior performance over the others in diverse testing environments and configurations. The need for effectiveness, therefore, necessitates the need for algorithm hybridization to deploy only the component parts of algorithms that have been proven to enhance overall search capa-bilities while at the same time eliminating the demerits of particular algorithms in the hybrid-ization procedure. In this paper, therefore, a hybrid variant of the African Buffalo Optimi-zation (ABO) algorithm is proposed for CIT. Four hybrid variants of the ABO are proposed through a deliberate improvement of the ABO with four algorithmic components. Experi-mental procedures indicate that the hybridization of the ABO with these algorithmic com-ponents led to faster convergence and greater effectiveness superior to the outcomes of existing techniques, thereby placing the algorithm among the best when compared with other methods/techniques.