A novel approach to automatic model-based test case generation

This paper proposes a new method for automatic generation of test cases using model-based testing. as test model, class and state diagrams are used and constraints are expressed using object constraint language (ocl). first, the state machine is converted into a mathematical representation in ampl (a mathematical programming language). then, using a search algorithm and based upon coverage criteria, the abstract paths are selected from state machine. second, using symbolic execution, the generated abstract path along with the constraints on this path is converted into the data of generated mathematical model. third, the generated mathematical problem is solved with solvers that have interface with ampl, and the test data are produced for each abstract test case. finally, the generated test data and abstract paths are transformed into executable test cases. all-transitions and all-states coverage criteria are used for conducting the search algorithm as well as the criteria to evaluate the quality of the generated test cases. to validate the work, by utilizing various solvers, the test cases are generated for various problems. the proposed technique is implemented as a tool, named mobateg. the tool shows good result in terms of test case generation execution time, test goals satisfaction rate, source code instructions coverage, and boundary values generation.