Ogrenme-Ogretme Temelli ve Armoni Arama Algoritmalari ile Test Verisi Uretimi

The increase in the size of software made hard to manage the projects and caused software crisis such as late delivering, delivering with missing functions, not meeting the requirements. The main reason of software crisis is that testing is not carried out at each stage of software development cycle. When the test is left to post-delivery stages, both time and money budget increases exponentially. To conduct an efficient testing, all possibilities and senarios should be considered and test data supplied should produce maximum-coverage on the control-flow graph of the associated program. Meta-heuristic algorithms instead of applying an exhaustive search can produce high-quality test data that provide a maximum coverage. In this study, two of recent meta-heuristics, teachinglearning based optimization algorithm and harmony search algorithm has been applied to test data generations and their search ability has been analyzed on seven code fragments.