EUVIMEDEUVIMED entdecken
Crossref

AUTOMATIC INFERENCE OF SYNCHRONOUS REGULAR INVARIANTS

A.I. Vasenina, Y.O. Kostyukov, D.A. Mordvinov

Программирование / Programming and Computer Software · 2025 · (4) · 33

DOI: 10.7868/s3034584725040038 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Программирование / Programming and Computer Software, 2025. Verfasst von A.I. Vasenina und 2 weiteren Autor:innen.

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Algebraic data types allow representing complex data structures, thus classical verification methods relying on inductive invariant inference are not applicable to programs with such types. The main problem with classical methods is the weakness of the first-order logic language used to represent inductive invariants. In this paper, synchronous automata over trees are considered as an alternative to classical representations of invariants using first-order logic. A method for automatic inductive invariant inference represented by synchronous automata over trees is proposed. The method uses finite-model finding. The implementation of the proposed method finds 65% more invariants than a first-order logic based tool, which is an annual winner of an international verifier competition, on an existing benchmark.

Zitieren

APA

Vasenina, A., Kostyukov, Y., & Mordvinov, D. (2025). AUTOMATIC INFERENCE OF SYNCHRONOUS REGULAR INVARIANTS. Программирование / Programming and Computer Software, 33. https://doi.org/10.7868/s3034584725040038

Vancouver

Vasenina A, Kostyukov Y, Mordvinov D. AUTOMATIC INFERENCE OF SYNCHRONOUS REGULAR INVARIANTS. Программирование / Programming and Computer Software. 2025:33. doi:10.7868/s3034584725040038

BibTeX
@article{Vasenina2025,
  author = {Vasenina, A.I. and Kostyukov, Y.O. and Mordvinov, D.A.},
  title = {AUTOMATIC INFERENCE OF SYNCHRONOUS REGULAR INVARIANTS},
  journal = {Программирование / Programming and Computer Software},
  year = {2025},
  number = {4},
  pages = {33},
  doi = {10.7868/s3034584725040038}
}
Kostenlos starten →