Timed Automata with Integer Resets: Language Inclusion and Expressiveness

Are Timed Automata Bad for a Specification Language? Language Inclusion Checking for Timed Automata

“…Naturally, this conclusion leads to work on identifying determinizable subclasses of timed automata, with reduced expressiveness. Several subclasses of timed automata have been identified, i.e., event-clock timed automata [3,17], timed automata with integer resets [18] or with one clock [16] and strongly non-Zeno timed automata [4].…”

“…In order to avoid undecidability, a number of subclasses of timed automata which are determinizable (and perhaps serve as a good specification language) have been identified, e.g., event-clock timed automata [3,17], timed automata restricted to at most one clock [16] and integer resets timed automata [18]. Recently, Baier et al [4] described a method for determinizing arbitrary timed automaton, which under a boundedness condition, yields an equivalent deterministic timed automaton in finite time.…”

A Game Approach to Determinize Timed Automata

“…Under a clockboundedness assumption, the infinite tree with infinitely many clocks can be folded up into a timed automaton (with finitely many locations and clocks). The clock-boundedness assumption is satisfied for several classes of timed automata, such as event-clock TAs [2], TAs with integer resets [13] and strongly non-Zeno TAs [3], which can thus be determinized by this procedure. The resulting deterministic timed automaton is doubly exponential in the size of A.…”

“…Timed automata with integer resets, introduced in [13], form a determinizable subclass of timed automata, where every edge ( , g, a, X , ) satisfies X = ∅ if and only if g contains an atomic constraint of the form x = c for some clock x.…”

“…deterministic TAs accepting more timed words. For the first approach, several classes of determinizable TAs have been identified, such as strongly non-Zeno TAs [3], event-clock TAs [2], or TAs with integer resets [13]. In a recent paper, Baier, Bertrand, Bouyer and Brihaye [4] propose a procedure which does not terminate in general, but allows one to determinize TAs in a class covering all the aforementioned determinizable classes.…”