Equivalence Class Partitions for Exhaustive Model-Based Testing

For safety-critical systems testing the justification of test case selections is mandatory. In the case of systems under test (SUT) with large data types (floats, doubles etc.) test cases are identified using the equivalence class partitioning principle. While this heuristic is useful for reducing the otherwise infeasible number of test cases to be performed, justification of the class partitions selected is currently mainly performed on an intuitive level which is inadequate for safety or mission critical systems. In this paper, we therefore introduce a rigorous justification principle for equivalence class partitions which is based on the fact that the classes selected lead to an exhaustive test suite of the SUT, provided that it fulfills certain fault hypotheses.