Structure-Based Local Search Heuristics for Circuit-Level Boolean Satisfiability

TL;DR

This paper enhances stochastic local search methods for Boolean satisfiability by integrating circuit-level structural heuristics, leading to improved performance on industrial SAT instances.

Contribution

It introduces novel circuit-structure-based heuristics for CRSat, advancing the state-of-the-art in local search for SAT solving.

Findings

New heuristics improve SAT solving efficiency

Circuit-level knowledge enhances local search performance

Most heuristics are novel in the context of SLS for SAT

Abstract

This work focuses on improving state-of-the-art in stochastic local search (SLS) for solving Boolean satisfiability (SAT) instances arising from real-world industrial SAT application domains. The recently introduced SLS method CRSat has been shown to noticeably improve on previously suggested SLS techniques in solving such real-world instances by combining justification-based local search with limited Boolean constraint propagation on the non-clausal formula representation form of Boolean circuits. In this work, we study possibilities of further improving the performance of CRSat by exploiting circuit-level structural knowledge for developing new search heuristics for CRSat. To this end, we introduce and experimentally evaluate a variety of search heuristics, many of which are motivated by circuit-level heuristics originally developed in completely different contexts, e.g., for electronic design automation applications. To the best of our knowledge, most of the heuristics are novel in the context of SLS for SAT and, more generally, SLS for constraint satisfaction problems.