A SAT Solver for Circuits Based on the Tableau Method

Abstract

We present an extension of the BC tableau, a calculus for determining satisfiability of constrained Boolean circuits. We argue that a satisfiability decision procedure based on the BC tableau can be implemented as a nonclausal DPLL procedure and that therefore, advances to the DPLL framework can be integrated into such a tableau procedure. We present a prototypical implementation of these ideas and evaluate it using a set of benchmark instances. We show that the extensions increase the efficiency of the basic BC tableau considerably and compare the performance of our solver with that of the non-clausal solver NoClause and the CNF-based SAT solver MiniSat.

Current Trends in Automated Deduction

“…Most SAT solvers are based on sophisticated refinements of the classical DPLL algorithm developed in [8], but there are also SAT solvers based on stochastic local search algorithms. While most SAT solvers operate on formulas in conjunctive normal form, there exist solvers operating on non-clausal formulas as well, see e.g., [9]. SAT solvers are used in numerous application areas, in particular in hardware and software verification [11].…”