Model-based Testing Sample Clauses
Model-based Testing. The ADVANCE project has developed two mechanisms for model-based test- ing: constraint-based and scope-based, which are described in the deliverable [Leu14]. In the first, the user provides a constraint and the tool finds a test that meets that constraint. In the second, the model checker is used to ex- plore the state space within a scope provided by the user and to generate tests for all paths within that scope. Cyber-physical systems are difficult to test, and by using constraints and scoping, a set of tests can be gener- ated which can be targeted at particular aspects of the system functionality. The coverage obtained by each set of tests is measured by using the ProB simulator to run the generated tests against the model. Chapter 6
Model-based Testing. On top of this, the RT-Tester Model Based Test Case and Test Data Gen- erator (RTT-MBT) [Ver15b] supports model-based testing (MBT), that is, automated generation of test cases, test data, and test procedures from UML /SysML models. A number of common modelling tools can be used as front- ends for this. The most important technical challenge in model-based test automation is the extraction of test cases from test models. RTT-MBT com- ▇▇▇▇▇ an SMT solver with a technique akin to bounded model checking so as to extract finite paths through the test model according to some predefined criterion. This criterion can, for instance, be MC/DC coverage, or it can be requirements coverage (if the requirements are specified as temporal logic formulae within the model). A further aspect is that the environment can be modelled within the test model. For example, the test model may contain a constraint such that a certain input to the system-under-test remains in a predefined range. This aspect becomes important once test automation is lifted from single test models to multi-model cyber-physical systems. The derived test procedures use the RT-Tester Core as a back-end, allowing the system under test to be provided on real hardware, software only, or even just simulation to aid test model development. Further, RTT-MBT includes requirement tracing from test models down to test executions and allows for powerful status reporting in large scale testing projects.
