Balanced FSM Generation for Empirical Studies

The experiments conducted to compare and analyse different FSM-based testing methods do not always provide the details of FSM generation. This paper focuses on automated generation of data, which can be used to run FSM- and X-machine test generation experiments. The novelty of the work is generation of what we call "balanced", state machines where the number of incoming and outgoing transitions from each state is a parameter specified by a user. It is also possible to build machines where a user-defined proportion of states have a few times higher number of incoming/outgoing transitions.