EvoMBT: Evolutionary model based testing

Writing tests for software systems is an important but expensive activity that plays a critical role in the success of the software. This is particularly true in systems where the interaction space is fine grained and in continuous change, as in the case of computer games or cyber-physical systems. In such situations model-based testing offers reasonable testing solutions as it allows to abstract away from details and focus on the relevant aspects from the point of view of testing. In this paper we present our tool EvoMBT that combines model-based testing with search algorithms for the generation of test cases for systems with complex and fine grained interactions. We illustrate the basic principles behind EvoMBT and provide examples along with empirical data from experiments on a self-driving car simulator where EvoMBT is used to produce road configurations that challenge the model responsible for driving the car.