A review on meta-heuristic search techniques for automated test data generation: Applicability towards improving automatic programming assessment

Abstract

Meta-Heuristic Search Techniques (MHST) recently have become among of the popular techniques applied in software testing research particularly in supporting Automated Test Data Generation (ATDG). MHST are generally defined as high-level frameworks to carry out the most optimal test data searching for efficient software testing so as to achieving effective cost reduction. ATDG is commonly known as a method used in software testing to automatically furnish a set of inputs or test data that do satisfy certain selected criteria. Having to integrate ATDG as part of testing in the area of computer science education, hence there should be a promising potential by embracing MHST with the current practice of Automatic Programming Assessment (APA). APA is identified as a technique applied in computer science education to execute automated marking and grading on students’ programming exercises. Thus, it motivates us to study on the gaps occurred between software testing and computer science education research areas in term of the significance of adapting MHST in deriving the desired test data to perform a dynamic testing in APA. In this paper, we provide an analysis of a review on the related topic of interest as part of the conducted Systematic Literature Review (SLR). This SLR is expected to highlight the current state and significance of MHST implementation in supporting ATDG to cover both of the dynamic functional and structural testing as well as to provide insight on MHST applicability towards improving the test data generation process in APA. As the result of this study, it depicts that MHST have been burgeoning most of the interest by researchers in ATDG and yet have been successfully carried out in achieving test data adequacy for efficient testing. With this current state, it positively offers a promising insight towards improving the state of the art of APA.