Debugging with Timed Automata Mutations

Abstract

Model-based Debugging is an application of Model-based Diagnosis techniques to debugging computer systems. Its basic principle is to compare a model, i.e., a description of the correct behaviour of a system, to the observed behaviour of the system. In this paper we show how this technique can be applied in the context of model-based mutation testing (MBMT) with timed automata. In MBMT we automatically generate a set of test sequences out of a test model. In contrast to general model-based testing, the test cases of MBMT cover a pre-defined set of faults that have been injected into the model (model mutation). Our automatic debugging process is purely black-box. If a test run fails, our tool reports a diagnosis as a set of model mutations. These mutations provide possible explanations why the test case has failed. For reproducing the failure, we also generate a set of minimal test cases leading to the implementation fault. The technique is implemented for Uppaal's timed automata models and is based on a language inclusion check via bounded model checking. It adds debugging capability to our existing test-case generators. A car-alarm system serves as illustrating case study.