NEURAL NETWORK BASED APPROACH FOR IMPROVING COMBINATORIAL COVERAGE IN COMBINATORIAL TESTING APPROACH

摘要

Recent advancements in technology has shown significant impact on social life, where computers have attracted huge attention due to its importance in socio-economic progress. Due to the growth in various computer technologies, software-based application has played pivotal role in the social and economic development. However, poor quality of software module may cause industrial loss; hence software quality improvement remains an attractive research field. Several techniques have been presented for improving the software quality by developing software testing methods. In this field of software testing, combinatorial testing is considered as the most promising scheme for improving software testing and quality improvement by reducing the number of test cases. This combinatorial testing strategy can help to provide a better solution for given software product. In this work, we have focused on software testing using combinatorial testing with the help of IPOG approach which is used for test case generation of the 2-way test scenario. Later, neural network scheme is incorporated for test case generation which provides most suitable test scenario for combinatorial coverage. For given software product, if random testing is performed and its test cases are available, then for this software we can easily identify how much combinatorial coverage is already performed, and how many new test cases are to be added to those available test cases of random testing so that appropriate testing coverage is achieved. A comparative scheme is presented which shows that proposed approach gives the best solution for test case generation for software testing.