Saved in:
Bibliographic Details
Main Authors: Ehlers, Rüdiger, Khalimov, Ayrat
Format: Preprint
Published: 2024
Subjects:
Online Access:https://arxiv.org/abs/2410.01021
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866915463885225984
author Ehlers, Rüdiger
Khalimov, Ayrat
author_facet Ehlers, Rüdiger
Khalimov, Ayrat
contents Chains of co-Buechi automata (COCOA) have recently been introduced as a new canonical representation of omega-regular languages. The co-Buechi automata in a chain assign to each omega-word its natural color, which depends only on the language itself and not on its automaton representation. The automata in such a chain can be minimized in polynomial time and are good-for-games, making the representation attractive for verification and reactive synthesis applications. However, since in such applications, a specification is usually given in linear temporal logic (LTL), to make COCOA useful, the specification first has to be translated into such a chain of automata. Currently, the only known translation procedure involves a detour through deterministic parity automata (LTL to DPW to COCOA), where the first step neglects the natural colors and requires intricate constructions by Safra or Esparza et al. This observation raises the question whether, by leveraging the definition of the natural color of words, these complex constructions can be avoided, leading to a more direct translation from LTL to COCOA. This paper presents a surprisingly simple yet optimal translation from LTL to COCOA. Our procedure relies on standard operations on weak alternating automata, Miyano-Hayashi's breakpoint construction, an augmented subset construction, and simple graph algorithms. With weak alternating automata as a starting point, the procedure can also handle specifications in linear dynamic logic.
format Preprint
id arxiv_https___arxiv_org_abs_2410_01021
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle From LTL to COCOA without Detours
Ehlers, Rüdiger
Khalimov, Ayrat
Formal Languages and Automata Theory
Chains of co-Buechi automata (COCOA) have recently been introduced as a new canonical representation of omega-regular languages. The co-Buechi automata in a chain assign to each omega-word its natural color, which depends only on the language itself and not on its automaton representation. The automata in such a chain can be minimized in polynomial time and are good-for-games, making the representation attractive for verification and reactive synthesis applications. However, since in such applications, a specification is usually given in linear temporal logic (LTL), to make COCOA useful, the specification first has to be translated into such a chain of automata. Currently, the only known translation procedure involves a detour through deterministic parity automata (LTL to DPW to COCOA), where the first step neglects the natural colors and requires intricate constructions by Safra or Esparza et al. This observation raises the question whether, by leveraging the definition of the natural color of words, these complex constructions can be avoided, leading to a more direct translation from LTL to COCOA. This paper presents a surprisingly simple yet optimal translation from LTL to COCOA. Our procedure relies on standard operations on weak alternating automata, Miyano-Hayashi's breakpoint construction, an augmented subset construction, and simple graph algorithms. With weak alternating automata as a starting point, the procedure can also handle specifications in linear dynamic logic.
title From LTL to COCOA without Detours
topic Formal Languages and Automata Theory
url https://arxiv.org/abs/2410.01021