A generalized temporal logic approach for control problems of a class of nondeterministic discrete event systems

A linear-time temporal logic is generalized to an uncertain world in order to apply it to the verification of a class of nondeterministic discrete event systems with probabilities known. Such systems are modeled by bounded stochastic models which specify structures for the language and the proof system. Properties of this class of systems are verified by deducing temporal logic specifications of desired behavior from descriptions of the system dynamics. The formulas do not mention probabilities explicitly (except probability one) so that the analysis of the control problem is completed qualitatively without using probability theory. Applications are illustrated by an example of a flexible manufacturing system.<>