Contributions to Constraint-Based Testing

These last years have seen the development of several constraint solving techniques dedicated to the testing of software systems, in an area called ''Constraint-Based Testing''. Our approach in this research domain consists to explore how Constraint Programming techniques can help the automatic generation of tests for imperative programs. We have addressed hard combinatorial problems resulting from automatic test cases generation by developing constraint propagation and filtering techniques well-tuned for handling control and data structures of imperative programs. Our habilitation tries to establish a first synthesis of the domain through five contributions, namely hybrid constraint solving for automatic test case generation, probabilistic test generation through probabilistic constraint combinator, constraints over abstract memory models, constraint solving over floating-point expressions and testing of constraint programs. We also illustrate these contributions through their application to critical software verification, and draw several research perspectives to our work.