Certified DQBF Solving by Definition Extraction

TL;DR

This paper introduces a novel decision procedure for DQBF that leverages interpolation-based definition extraction within a CEGIS loop, enabling certifiable solutions and improved performance on benchmarks.

Contribution

It presents a new certification-friendly DQBF solving method using arbiter variables and definition extraction, enhancing model generation and proof certification capabilities.

Findings

Successfully solves standard benchmarks with competitive performance.

Generates and validates models for all true instances solved.

Provides a certifying procedure with proof derivation for false formulas.

Abstract

We propose a new decision procedure for dependency quantified Boolean formulas (DQBF) that uses interpolation-based definition extraction to compute Skolem functions in a counter-example guided inductive synthesis (CEGIS) loop. In each iteration, a family of candidate Skolem functions is tested for correctness using a SAT solver, which either determines that a model has been found, or returns an assignment of the universal variables as a counterexample. Fixing a counterexample generally involves changing candidates of multiple existential variables with incomparable dependency sets. Our procedure introduces auxiliary variables -- which we call arbiter variables -- that each represent the value of an existential variable for a particular assignment of its dependency set. Possible repairs are expressed as clauses on these variables, and a SAT solver is invoked to find an assignment that deals with all previously seen counterexamples. Adding arbiter variables defines the values of Skolem functions for assignments where they were previously undefined, and may lead to the detection of Skolem functions by definition extraction in subsequent iterations. A key feature of the proposed procedure is that it is certifying by design: for true DQBF, models can be returned at minimal overhead. Towards certification of false formulas, we prove that clauses can be derived in an expansion-based proof system for DQBF. In an experimental evaluation on standard benchmark sets, an implementation was able to match (and in some cases, surpass) the performance of state-of-the-art DQBF solvers. Moreover, models could be generated and validated for all true instances that were solved.