sQueezeBF: An Effective Preprocessor for QBFs Based on Equivalence Reasoning

Failed Literal Detection for QBF

“…Additionally, unit clauses learnt by PicoSAT are propagated using QBF-specific QBCP rules within QxBF. Table 1 compares the impact of different FL approaches on the performance of QBF solvers based on search (DepQBF [17] and QuBE7.1 [10]) and variable elimination (Quantor [2], squolem [15] and Nenofex [16]) using all benchmarks from QBFEVAL'10 [20]. For QuBE7.1 internal preprocessing was disabled (QuBE7.1-np).…”

“…We used latest publicly available versions of solvers except internal versions of Nenofex and DepQBF, all without proprietary preprocessing 6 . Results using sQueezeBF [11], a state-of-the-art QBF preprocessor, are reported for reference. We cannot expect FL to be competitive with sQueezeBF as the latter applies a larger pool of inference rules (details are given below).…”

“…In order to overcome this drawback, several approaches for preprocessing PCNFs have been suggested. Some were ported from SAT to QBF such as binary clause reasoning, equivalence detection and variable elimination by resolution or expansion [2,5,6,11,21,25,26].…”

Solving QBF with Counterexample Guided Refinement

“…If there is no counterexample to the move winning the abstraction, then this move is also a winning move for the given problem and the function terminates. If there is a counterexample to the move winning the abstraction, the abstraction must be refined (line 14).…”

“…While nonrandom SAT solving has been dominated by the DPLL procedure, it has proven to be far from a silver bullet for QBF solving. Indeed, a number of solving techniques have been proposed for QBF [12,3,4,19,15], complemented by a variety of preprocessing techniques [7,14,21,5].…”