Cost-effective metamorphic testing techniques for failure detection in software with oracle problem

Many software applications are developed to find answers to unsolved problems in the first place. Detecting failures or incorrect outputs in these applications is fundamentally challenging because the correct expected outputs (oracles) are unknown. This thesis investigates failure detections in these software applications through a property-based testing approach known as metamorphic testing. Six novel cost saving strategies have been proposed and adopted in designing cost effective metamorphic testing techniques to detect failures in the absence of oracles. The proposed testing techniques enable cost-effective failure detections in software applications without knowing the correct expected outputs and contribute towards software quality improvement.