Model Checking Linear Duration Invariants of Networks of Automata

Linear duration invariants (LDIs) are important safety properties of real-time systems. In this paper, we reduce the problem of verification of a network of timed automata against an LDI to an equivalent problem of model checking whether a failure state is never reached. Our approach is first to transform each component automaton Ai of the network A to an automaton Gi. The transformation helps us to record entry and exit to critical locations that appear in the LDI. We then introduce an auxiliary checker automaton S and define a failure state to verify the LDI on a given interval. Since a model checker checks exhaustively, a failure of the checker automaton to find the failure state will prove that the LDI holds.