Department of Mathematics, Faculty of Science, Beni-Suef University, Egypt; Department of Computer Science, College of Computers and Information Systems, Taif University, Saudi Arabia

---

Abstract—The widespread use of the object-oriented programs (OOPs) makes the requirement for testsgeneration strategies for testing the OOPs increases from day to day. In this paper, we present a multi-stage genetic algorithm (MSGA) to generate a suite of tests for testing the OOPs. MSGA includes two optimization stages. The first stage concentrates on finding test cases (sequences of called methods), which satisfy a given test criterion. The second stage focuses on generating test data (values of the arguments of the called methods). In addition, we introduce a new chromosome representation, which consists of two concatenated one-dimensional arrays. Each array contains set of homogeneous genes. In addition, we introduce set of strategies for encoding and decoding the tests. Furthermore, we present set of new genetic operators and the required pre- and post-conditions for applying these operators. In order to determine the applicability and practicability of MSGA, we introduce a new testing tool by implementing MSGA. Also, we conduct a case study by the new tool to assess the efficiency of MSGA in data-flow testing of OOPs.

Index Terms—Object-Oriented Testing, Test Generation, Genetic Algorithms, Class Control-Flow Graph, Dominance.

[PDF]

Cite: Ahmed S. Ghiduk, " Automatic Generation of Object-Oriented Tests With a Multistage-Based Genetic Algorithm," Journal of Computers vol. 5, no. 10, pp. 1560-1569, 2010.