Minimization of Incompletely Specified Finite State Machines Based on Distinction Graphs

Many generation methods are available for deriving a test suite from reduced finite state machines. When the machine in question is not reduced, what happens on several situations, those methods cannot be directly applied, requiring first the minimization of the machine. However, the minimization of incompletely specified finite state machines is a known NP-complete problem Due to the usefulness of such models in systems development and test, the search for heuristics to obtain feasible solutions to the problem has received great attention. Improvements in this area bring clear benefits to model based testing approaches, since it allows the usage of many generation methods that can only be applied to reduced machines. In this paper, we propose an approach to minimize partial finite state machines based on the selection of compatible states using maximum cliques on the distinction graph. We empirically measured the performance of our method against two previous methods and the data show that our new approach generates good results in less time.