Seminator 2 Can Complement Generalized Büchi Automata via Improved Semi-determinization

Abstract

We present the second generation of the tool Seminator that transforms transition-based generalized Büchi automata (TGBAs) into equivalent semi-deterministic automata. The tool has been extended with numerous optimizations and produces considerably smaller automata than its first version. In connection with the state-of-the-art LTL to TGBAs translator Spot, Seminator 2 produces smaller (on average) semi-deterministic automata than the direct LTL to semi-deterministic automata translator ltl2ldgba of the Owl library. Further, Seminator 2 has been extended with an improved NCSB complementation procedure for semi-deterministic automata, providing a new way to complement automata that is competitive with state-of-the-art complementation tools.

Sky Is Not the Limit: Tighter Rank Bounds for Elevator Automata in Büchi Automata Complementation (Technical Report)

“…Intuitively, an elevator automaton is a BA whose strongly connected components (SCCs) are all either inherently weak or deterministic. Clearly, the class substantially generalizes the popular inherently weak [14] and semi-deterministic BAs [15,16,17]). The structure of elevator automata allows us to provide tighter estimates of the TRUBs, not only for elevator automata per se, but also for BAs where elevator automata occur as a sub-structure (which is even more common).…”

“…We tested the correctness of our implementation using S 's autcross on all BAs in our benchmark. We compared modified R with other state-of-the-art tools, namely, G [22] (implementing P [23], S [8], S [24], and F [25]), S 2.9.3 [26] (implementing Redziejowski's algorithm [27]), S 2 [17], LTL2 0.5.4 [28], and R [29]. All tools were set to the mode where they output an automaton with the standard state-based Büchi acceptance condition.…”

Reducing (to) the Ranks: Efficient Rank-based Büchi Automata Complementation (Technical Report)

“…We implemented the optimizations described in the previous sections in a tool called Ranker 1 in C++. We compared our complementation approach with other state-of-the-art tools, namely, GOAL [20] (including the Fribourg plugin [21]), Spot 2.9.3 [22], ROLL [23], and Seminator 2 [24]. All tools were set to the mode where they output an automaton with the standard state-based Büchi acceptance condition.…”

“…In this setting, we focused the evaluation on the size of the output BA after postprocessing (we, again, used autfilt with simplification level --high; we denote this using "+PP"). We evaluated the following algorithms: Safra [27] implemented in GOAL (parameter -m safra), Piterman [28] implemented in GOAL (parameter -m piterman), Fribourg [21] implemented as a plugin of GOAL, the complementation implemented in Spot (we could not find which particular algorithm(s) is/are used, though it seems from the source code that a particular determinization-based algorithm is selected based on the features of the input BA), a learning-based algorithm [23] implemented in ROLL, and a semideterminization-based algorithm [24] implemented in Seminator 2.…”

“…Except for complementation of standard nondeterministic Büchi automata, there are specific approaches for complementation of special types of BAs, such as deterministic [30], semi-deterministic [31], or unambiguous [32]. Semi-determinization based complementation then uses a conversion of a standard BA to its semi-deterministic version [33] followed by its complementation [24].…”