posted on 2021-07-21, 15:28authored byMurphy Berzish, Joel DayJoel Day, Vijay Ganesh, Mitja Kulczynski, Florin Manea, Federico Mora, Dirk Nowotka
Widespread use of string solvers in formal analysis of stringheavy programs has led to a growing demand for more efficient and reliable techniques which can be applied in this context, especially for real-world cases. Designing an algorithm for the (generally undecidable) satisfiability problem for systems of string constraints requires a thorough understanding of the structure of constraints present in the targeted cases. In this paper, we investigate benchmarks presented in the literature containing regular expression membership predicates, extract different first order logic theories, and prove their decidability, resp. undecidability. Notably, the most common theories in real-world benchmarks are PSPACEcomplete and directly lead to the implementation of a more efficient algorithm to solving string constraints.
History
School
Science
Department
Computer Science
Published in
Combinatorics on Words. 13th International Conference, WORDS 2021 Rouen, France, September 13–17, 2021. Proceedings. (WORDS 2021)
Pages
50-64
Source
13th International Conference on Words (WORDS 2021)