Test Data Generation

The quality of the specifications used for test generation plays an important role in the quality of the generated tests. One approach to improve the quality of the UML specification is the use of executable models specified using the Foundational Subset for Executable UML Models (fUML) and the Action language for fUML (Alf). Due to their precise semantics, fUML and Alf models can be simulated or executed using an fUML execution engine. However, in order to execute the models exhaustively, one must provide input data required to reach and cover all essential elements not only in the graphical fUML models, but also in the textual Alf code associated with the graphical models. In this paper, we present an approach for exhaustive simulation and test generation from fUML activity diagrams containing Alf code. The proposed approach translates fUML activity diagrams and associated Alf code into equivalent Java code and then automatically generates: (1) input data needed to cover or execute all paths in the executable fUML and Alf models and (2) test cases and test oracle (expected output) for testing the actual implementation of the system under development. We also present a tool chain and demonstrate our proposed approach with the help of an example.

Event-B is a formal modeling language having set theory as its mathematical foundation and abstract state machines as its behavioral specifications. The language has very good tool support based on theorem proving and model checking technologies, but very little support for test generation. Motivated by industrial interest in the latter domain, this paper presents an approach based on genetic algorithms that generates test data for Event-B test paths. For that, new fitness functions adapted to the set-theoretic nature of Event-B are devised. The approach was implemented and its efficiency was proven on a carefully designed benchmark using statistically sound evaluations.