命题逻辑提升到一阶逻辑上的子句消去方法

在基于命题逻辑的可满足性问题（SAT）求解器和基于一阶逻辑的定理证明器上，子句集简化一直是必不可少的步骤，而其中子句消去方法在这些子句集简化方法中是非常重要的组成部分。将命题逻辑中的子句消去方法归结隐藏恒真消去方法（RHTE）和归结隐藏包含消去方法（RHSE）提升到一阶逻辑上，并且利用蕴含模归结原则（IMR）证明了这种提升方式在一阶逻辑上具有可靠性（Soundness），即依据这两种子句消去方法删除一阶逻辑公式集中的子句，并不会改变公式集的可满足性或者不可满足性。此外，将这两个方法与一阶逻辑子句消去方法锁子句消去方法（BCE）和归结包含消去方法（RSE）进行组合推广，发展得到一阶逻辑上新型子句消去方法（BC+RHS）E、（RS+RHT）E和（RHS+RHT）E，并且证明了这3种子句消去方法在一阶逻辑上的可靠性。最后，分析比较了这些子句消去方法的有效性，并且证明了这3种新型子句消去方法比组成它们的原始子句消去方法均具有更高的有效性。     

Incremental preprocessing methods for use in BMC

Traditional incremental SAT solvers have achieved great success in the domain of Bounded Model Checking (BMC). Recently, modern solvers have introduced advanced preprocessing procedures that have allowed them to obtain high levels of performance. Unfortunately, many preprocessing techniques such as variable and (blocked) clause elimination cannot be directly used in an incremental manner. This work focuses on extending these techniques and Craig interpolation so that they can be used effectively together in incremental SAT solving (in the context of BMC). The techniques introduced here doubled the performance of our BMC solver on both SAT and UNSAT problems. For UNSAT problems, preprocessing had the added advantage that Craig interpolation was able to find the fixed point sooner, reducing the number of incremental SAT iterations. Furthermore, our ideas seem to perform better as the benchmarks become larger, and/or deeper, which is exactly when they are needed. Lastly, our methods can be integrated into other SAT based BMC tools to achieve similar speedups.     

Exploiting multivalued knowledge in variable selection heuristics for SAT solvers

We show that we can design and implement extremely efficient variable selection heuristics for SAT solvers by identifying, in Boolean clause databases, sets of Boolean variables that model the same multivalued variable and then exploiting that structural information. In particular, we define novel variable selection heuristics for two of the most competitive existing SAT solvers: Chaff, a solver based on look-back techniques, and Satz, a solver based on look-ahead techniques. Our heuristics give priority to Boolean variables that belong to sets of variables that model multivalued variables with minimum domain size in a given state of the search process. The empirical investigation conducted to evaluate the new heuristics provides experimental evidence that identifying multivalued knowledge in Boolean clause databases and using variable selection heuristics that exploit that knowledge leads to large performance improvements.      

meSAT: multiple encodings of CSP to SAT

One approach for solving Constraint Satisfaction Problems (CSP) (and related Constraint Optimization Problems (COP)) involving integer and Boolean variables is reduction to propositional satisfiability problem (SAT). A number of encodings (e.g., direct, log, support, order) for this purpose exist as well as specific encodings for some constraints that are often encountered (e.g., cardinality constraints, global constraints). However, there is no single encoding that performs well on all classes of problems and there is a need for a system that supports multiple encodings. We present a system that translates specifications of finite linear CSP problems into SAT instances using several well-known encodings, and their combinations. We also present a methodology for selecting a suitable encoding based on simple syntactic features of the input CSP instance. Thorough evaluation has been performed on large publicly available corpora and our encoding selection method improves upon the efficiency of existing encodings and state-of-the-art tools used in comparison.     

Solving SAT problem by heuristic polarity decision-making algorithm

This paper presents a heuristic polarity decision-making algorithm for solving Boolean satisfiability (SAT). The algorithm inherits many features of the current state-of-the-art SAT solvers, such as fast BCP, clause recording, restarts, etc. In addition, a preconditioning step that calculates the polarities of variables according to the cover distribution of Karnaugh map is introduced into DPLL procedure, which greatly reduces the number of conflicts in the search process. The proposed approach is implemented as a SAT solver named DiffSat. Experiments show that DiffSat can solve many "real-life" instances in a reasonable time while the best existing SAT solvers, such as Zchaff and MiniSat, cannot. In particular, DiffSat can solve every instance of Bart benchmark suite in less than 0.03 s while Zchaff and MiniSat fail under a 900 s time limit. Furthermore, DiffSat even outperforms the outstanding incomplete algorithm DLM in some instances.     

Local and global symmetry breaking in itemset mining

The concept of symmetry has been extensively studied in the field of constraint programming and in the propositional satisfiability. Several methods for detection and removal of these symmetries have been developed, and their use in known solvers of these domains improved dramatically their effectiveness on a big variety of problems considered difficult to solve. The concept of symmetry may be exported to other areas where some structures can be exploited effectively. Particularly, in the area of data mining where some tasks can be expressed as constraints or logical formulas. We are interested here, by the detection and elimination of local and global symmetries in the item-set mining problem. Recent works have provided effective encodings as Boolean constraints for these data mining tasks and some idea on symmetry elimination in this area begin to appear, but still few and the techniques presented are often on global symmetry that is detected and eliminated statically in a preprocessing phase. In this work we study the notion of local symmetry and compare it to global symmetry for the itemset mining problem. We show how local symmetries of the boolean encoding can be detected dynamically and give some properties that allow to eliminate theses symmetries in SAT-based itemset mining solvers in order to enhance their efficiency.     

UnitWalk: A new SAT solver that uses local search guided by unit clause elimination

In this paper we present a new randomized algorithm for SAT, i.e., the satisfiability problem for Boolean formulas in conjunctive normal form. Despite its simplicity, this algorithm performs well on many common benchmarks ranging from graph coloring problems to microprocessor verification. Our algorithm is inspired by two randomized algorithms having the best current worst-case upper bounds ([27,28] and [30,31]). We combine the main ideas of these algorithms in one algorithm. The two approaches we use are local search (which is used in many SAT algorithms, e.g., in GSAT [34] and WalkSAT [33]) and unit clause elimination (which is rarely used in local search algorithms). In this paper we do not prove any theoretical bounds. However, we present encouraging results of computational experiments comparing several implementations of our algorithm with other SAT solvers. We also prove that our algorithm is probabilistically approximately complete (PAC).     

UnitWalk: A new SAT solver that uses local search guided by unit clause elimination

In this paper we present a new randomized algorithm for SAT, i.e., the satisfiability problem for Boolean formulas in conjunctive normal form. Despite its simplicity, this algorithm performs well on many common benchmarks ranging from graph coloring problems to microprocessor verification. Our algorithm is inspired by two randomized algorithms having the best current worst-case upper bounds ([27,28] and [30,31]). We combine the main ideas of these algorithms in one algorithm. The two approaches we use are local search (which is used in many SAT algorithms, e.g., in GSAT [34] and WalkSAT [33]) and unit clause elimination (which is rarely used in local search algorithms). In this paper we do not prove any theoretical bounds. However, we present encouraging results of computational experiments comparing several implementations of our algorithm with other SAT solvers. We also prove that our algorithm is probabilistically approximately complete (PAC).     

Efficiently solving quantified bit-vector formulas

In recent years, bit-precise reasoning has gained importance in hardware and software verification. Of renewed interest is the use of symbolic reasoning for synthesising loop invariants, ranking functions, or whole program fragments and hardware circuits. Solvers for the quantifier-free fragment of bit-vector logic exist and often rely on SAT solvers for efficiency. However, many techniques require quantifiers in bit-vector formulas to avoid an exponential blow-up during construction. Solvers for quantified formulas usually flatten the input to obtain a quantified Boolean formula, losing much of the word-level information in the formula. We present a new approach based on a set of effective word-level simplifications that are traditionally employed in automated theorem proving, heuristic quantifier instantiation methods used in SMT solvers, and model finding techniques based on skeletons/templates. Experimental results on two different types of benchmarks indicate that our method outperforms the traditional flattening approach by multiple orders of magnitude of runtime.     

Easy Cases of Probabilistic Satisfiability

The Probabilistic Satisfiability problem (PSAT) can be considered as a probabilistic counterpart of the classical SAT problem. In a PSAT instance, each clause in a CNF formula is assigned a probability of being true; the problem consists in checking the consistency of the assigned probabilities. Actually, PSAT turns out to be computationally much harder than SAT, e.g., it remains difficult for some classes of formulas where SAT can be solved in polynomial time. A column generation approach has been proposed in the literature, where the pricing sub-problem reduces to a Weighted Max-SAT problem on the original formula. Here we consider some easy cases of PSAT, where it is possible to give a compact representation of the set of consistent probability assignments. We follow two different approaches, based on two different representations of CNF formulas. First we consider a representation based on directed hypergraphs. By extending a well-known integer programming formulation of SAT and Max-SAT, we solve the case in which the hypergraph does not contain cycles; a linear time algorithm is provided for this case. Then we consider the co-occurrence graph associated with a formula. We provide a solution method for the case in which the co-occurrence graph is a partial 2-tree, and we show how to extend this result to partial k-trees with k>2.     

An overview of parallel SAT solving

Boolean satisfiability (SAT) solvers are currently very effective in practice. However, there are still many challenging problems for SAT solvers. Nowadays, extra computational power is no longer coming from higher processor frequencies. At the same time, multicore architectures are becoming predominant. Exploiting this new architecture is essential for the evolution of SAT solvers. Due to the increasing interest in parallel SAT solving, it is important to give an overview of what has been done so far. This paper presents an overview of parallel SAT solving and it is expected to be a valuable document for researchers in this field. This overview covers the main topics of parallel SAT solving, namely, different approaches and a variety of clause sharing strategies. Additionally, an evaluation of multicore SAT solvers is presented, showing the evolution of multicore SAT solvers over the last years.     

Beyond the structure of SAT formulas

Nowadays, many real-world problems are encoded into SAT instances and efficiently solved by modern SAT solvers. These solvers, usually known as Conflict-Driven Clause Learning (CDCL) SAT solvers, include a variety of sophisticated techniques, such as clause learning, lazy data structures, conflict-based adaptive branching heuristics, or random restarts, among others. However, the reasons of their efficiency in solving real-world, or industrial, SAT instances are still unknown. The common wisdom in the SAT community is that these technique exploit some hidden structure of real-world problems.In this thesis, we characterize some important features of the underlying structure of industrial SAT instances. Namely, they are the community structure and the self-similar structure. We observe that most industrial SAT formulas, viewed as graphs, have these two properties. This means that (i) in a graph with a clear community structure, i.e. having high modularity, we can find a partition of its nodes into communities such that most edges connect nodes of the same community; and (ii) in a graph with a self-similar pattern, i.e. being fractal, its shape is kept after re-scalings, i.e., grouping sets of nodes into a single node. We also analyze how these structures are affected by the effects of CDCL techniques during the search.Using the previous structural studies, we propose three applications. First, we face the problem of generating pseudo-industrial random SAT instances using the notion of modularity. Our model generates instances similar to (classical) random SAT formulas when the modularity is low, but when this value is high, our model is also adequate to model realistic pseudo-industrial problems. Second, we propose a method based on the community structure of the instance to detect relevant learnt clauses. Our technique augments the original instance with this set of relevant clauses, and this results into an overall improvement of the efficiency of several state-of-the-art CDCL SAT solvers. Finally, we analyze the classification of industrial SAT instances into families using the previously analyzed structure features, and we compare them to other classifiers commonly used in portfolio SAT approaches.In summary, this dissertation extends the understandings of the structure of SAT instances, with the aim of better explaining the success of CDCL techniques and possibly improve them, and propose a number of applications based on this analysis of the underlying structure of SAT formulas.     

Limitations of restricted branching in clause learning

The techniques for making decisions, that is, branching, play a central role in complete methods for solving structured instances of constraint satisfaction problems (CSPs). In this work we consider branching heuristics in the context of propositional satisfiability (SAT), where CSPs are expressed as propositional formulas. In practice, there are cases when SAT solvers based on the Davis-Putnam-Logemann-Loveland procedure (DPLL) benefit from limiting the set of variables the solver is allowed to branch on to so called input variables which provide a strong unit propagation backdoor set to any SAT instance. Theoretically, however, restricting branching to input variables implies a super-polynomial increase in the length of the optimal proofs for DPLL (without clause learning), and thus input-restricted DPLL cannot polynomially simulate DPLL. In this paper we settle the case of DPLL with clause learning. Surprisingly, even with unlimited restarts, input-restricted clause learning DPLL cannot simulate DPLL (even without clause learning). The opposite also holds, and hence DPLL and input-restricted clause learning DPLL are polynomially incomparable. Additionally, we analyze the effect of input-restricted branching on clause learning solvers in practice with various structured real-world benchmarks. This is an extended version of a paper [27] presented at the 13th International Conference on Principles and Practice of Constraint Programming (CP 2007) in Providence, RI, USA. The first author gratefully acknowledges financial support from Helsinki Graduate School in Computer Science and Engineering, Academy of Finland (grants #211025 and #122399), Emil Aaltonen Foundation, Jenny and Antti Wihuri Foundation, Finnish Foundation for Technology Promotion TES, and Nokia Foundation. The second author gratefully acknowledges the financial support from Academy of Finland (grant #112016).     

Compressing BMC Encodings with QBF

Symbolic model checking is PSPACE complete. Since QBF is the standard PSPACE complete problem, it is most natural to encode symbolic model checking problems as QBF formulas and then use QBF decision procedures to solve them. We discuss alternative encodings for unbounded and bounded safety checking into SAT and QBF. One contribution is a linear encoding of simple path constraints, which usually are necessary to make k-induction complete. Our experimental results show that indeed a large reduction in the size of the generated formulas can be obtained. However, current QBF solvers seem not to be able to take advantage of these compact formulations. Despite these mostly negative results the availability of these benchmarks will help improve the state of the art of QBF solvers and make QBF based symbolic model checking a viable alternative.     

基于布尔可满足性的组合电路ATPG算法

布尔可满足性被深入研究并广泛应用于电子设计自动化等领域。该文提出了一种基于布尔可满足性的组合电路ATPG改进算法。在采用当前最新布尔可满足性求解程序加速策略的基础上,比如冲突驱动训练、冲突导向回跳和重启动技术等,引入电路结构信息来实现基于结构的分支决策。通过新增的电路结构信息层,布尔可满足性求解程序只需稍加修改,就能利用和及时更新此信息。最后给出的实验结果表明了算法的可行性和有效性。     

Solving parity games by a reduction to SAT

This paper presents a reduction from the problem of solving parity games to the satisfiability problem in propositional logic (SAT). The reduction is done in two stages, first into difference logic, i.e. SAT combined with the theory of integer differences, an instance of the SAT modulo theories (SMT) framework. In the second stage the integer variables and constraints of the difference logic encoding are replaced with a set of Boolean variables and constraints on them, giving rise to a pure SAT encoding of the problem. The reduction uses Jurdziński?s characterisation of winning strategies via progress measures. The reduction is motivated by the success of SAT solvers in symbolic verification, bounded model checking in particular. The paper reports on prototype implementations of the reductions and presents some experimental results.     

基于不可满足原因的最小纠正集求解

布尔公式的最小纠正集MCS是子句的集合。对于一个不可满足公式,移除MCS后,所得到的新公式可满足。任一MCS中的子句保留在公式中,所得到的新公式不可满足。通过求解MCS 并调整约束集合,能够求解最小不可满足核心、MaxSAT 问题和最大（小）可满足解问题;还能够应用于故障定位、模型检查配置优化等实际问题中。 提出了一种基于不可满足原因的MCS求解算法,实现了相应的CUC工具。通过与目前最好的MCS求解工具LBX进行比较,得到了CUC性能优于LBX的结论。CUC比LBX平均多解出5%（65个）的公式。对于CUC和LBX均可解出的公式,CUC的平均求解时间比LBX快2.5倍。     

On Modern Clause-Learning Satisfiability Solvers

In this paper, we present a perspective on modern clause-learning SAT solvers that highlights the roles of, and the interactions between, decision making and clause learning in these solvers. We discuss two limitations of these solvers from this perspective and discuss techniques for dealing with them. We show empirically that the proposed techniques significantly improve state-of-the-art solvers.     

Propagation via lazy clause generation

Finite domain propagation solvers effectively represent the possible values of variables by a set of choices which can be naturally modelled as Boolean variables. In this paper we describe how to mimic a finite domain propagation engine, by mapping propagators into clauses in a SAT solver. This immediately results in strong nogoods for finite domain propagation. But a naive static translation is impractical except in limited cases. We show how to convert propagators to lazy clause generators for a SAT solver. The resulting system introduces flexibility in modelling since variables are modelled dually in the propagation engine and the SAT solver, and we explore various approaches to the dual modelling. We show that the resulting system solves many finite domain problems significantly faster than other techniques. This paper is an extension of results first published in [29, 30].     

Solution Validation and Extraction for QBF Preprocessing

In the context of reasoning on quantified Boolean formulas (QBFs), the extension of propositional logic with existential and universal quantifiers, it is beneficial to use preprocessing for solving QBF encodings of real-world problems. Preprocessing applies rewriting techniques that preserve (satisfiability) equivalence and that do not destroy the formula’s CNF structure. In many cases, preprocessing either solves a formula directly or modifies it such that information helpful to solvers becomes better visible and irrelevant information is hidden. The application of a preprocessor, however, prevented the extraction of proofs for the original formula in the past. Such proofs are required to independently validate the correctness of the preprocessor’s rewritings and the solver’s result as well as for the extraction of solutions in terms of Skolem functions. In particular for the important technique of universal expansion efficient proof checking and solution generation was not possible so far. This article presents a unified proof system with three simple rules based on quantified resolution asymmetric tautology (\(\mathsf {QRAT}\)). In combination with an extended version of universal reduction, we use this proof system to efficiently express current preprocessing techniques including universal expansion. Further, we develop an approach for extracting Skolem functions. We equip the preprocessor bloqqer with \(\mathsf {QRAT}\) proof logging and provide a proof checker for \(\mathsf {QRAT}\) proofs which is also able to extract Skolem functions.