Using Defect Prediction to Improve the Bug Detection Capability of Search-Based Software Testing

Software systems have a direct and indirect impact on the lives of humans, animals and other living things. They need to be tested thoroughly to minimise software failures. Automated test generators, like search-based software testing (SBST) techniques, replace the tedious and expensive task of manually writing tests. Despite achieving high code coverage, current SBST techniques perform rather poorly in terms of detecting bugs. This thesis proposes novel SBST approaches guided by defect prediction and demonstrates that to effectively and efficiently detect bugs SBST needs to focus test generation more on likely buggy areas in programs guided by defect prediction.