Model-based test suite generation for graph transformation system using model simulation and search-based techniques

Abstract

Abstract Context Test generation by model checking is a useful technique in model-based testing that allows automatic generation of test cases from models by utilizing the counter-examples/witnesses produced through a model checker. However, generating redundant test cases and state space explosion problem are two major obstacles to transfer this technique into industrial practice. Objective An idea to cope with these challenges consists in an intelligent model checking for exploring only a portion of the state space according to the test objectives. Motivated by this idea, we propose an approach that exploits meta-heuristic algorithms to adapt a model checker when used for integration testing of systems formally specified by graph transformations. Method This method is not based on model checking algorithms, but rather uses the modeling and simulation features of the underlying model checker. In the proposed approach, a population of test suites that each of which is a set of paths on the state space, is evolved towards satisfying the all def-use test objectives. Consequently, a test suite with high coverage is generated. Results To assess the efficiency of our approach, it is implemented in GROOVE, an open source toolset for designing and model checking graph transformation systems. Empirical results based on some case studies, confirm a significant improvement in terms of coverage, speed and memory usage, in comparison with the state of the art techniques. Conclusion Our analysis reveals that intelligent model checking can appropriately address the challenges of traditional model-checking-assisted testing. We further conclude that graph transformation specification is an efficient modeling solution to behavioral testing and graph transformation tools have a great potential for developing a model-based testing tool.