Synthesis of Distinguishing Test Cases for Timed Finite State Machines

This paper is devoted to the generation of distinguishing experiments with completely specified timed finite state machines. It is shown, in particular, that two completely specified nondeterministic finite state machines can be distinguished by a multiple preset experiment if and only if these finite state machines are not equivalent. Two finite state machines can be distinguished by a simple adaptive experiment if and only if they are r-distinguishable, i.e., have no common completely specified reduction. The corresponding adaptive experiment is described by a special timed finite state machine. The procedure for constructing such an r-distinguishing timed finite state machine is proposed.