Goal-Aware Neural SAT Solver

Ozoliņš, Emīls; Freivalds, Kārlis; Draguns, Andis; Eliza, Gaile,; Zakovskis, Ronalds; Kozlovičs, Sergejs

doi:10.1109/ijcnn55064.2022.9892733

Cited by 7 publications

(2 citation statements)

References 8 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…(2) Reinforcement learning-based approach: this approach treats the SAT solving process as a reinforcement learning problem, where the SAT solver acts as an intelligent body that chooses an action (i.e., assignment of a variable) based on the current state (i.e., the state of the CNF formula) and receives a reward or penalty based on the outcome of the action (i.e., whether it leads to a conflict). Reinforcement learning algorithms, such as Q-learning or deep reinforcement learning, are then used to learn a policy to maximize future rewards [24,25].…”

Section: Learning-based Solution Methodsmentioning

confidence: 99%

A Review of Research on Algorithms for Solving SAT Problems

Guo

2024

MMAA

View full text Add to dashboard Cite

The SAT problem solving algorithms are an important research area in computer science aimed at solving the Boolean satisfiability problem. With the wide application of SAT problems in the fields of hardware design, automated planning, and artificial intelligence, researchers have proposed a variety of efficient solution algorithms. The classical algorithms include exhaustive search and backtracking search, but they are less efficient when dealing with large-scale problems. Heuristic search methods accelerate the search process by introducing randomization or evolutionary algorithms to improve the solution efficiency. Learning-based methods utilize learning and memorizing the explored search space or learning heuristic information to guide the search process. The CDCL algorithm, on the other hand, is an efficient conflict-driven learning algorithm that achieves impressive results by learning and applying conflict information to prune the search space. Parallel and distributed solving as well as quantum computing methods are hot research topics in recent years, which provide new ideas and technical support for solving large-scale problems. The research on algorithms for solving SAT problems continues to innovate, providing powerful tools and technical support for real-world problem solving.

show abstract

Section: Learning-based Solution Methodsmentioning

confidence: 99%

A Review of Research on Algorithms for Solving SAT Problems

Guo

2024

MMAA

View full text Add to dashboard Cite

show abstract

“…While some solvers are optimized to run on a single-core CPU (e.g., Glusoce and GSAT), others take advantage specific of hardware, e.g., HordeSAT utilizes multiple CPU cores [3,11,4]. At IMCS 2 , we have developed our own solver QuerySAT, which uses graph neural networks internally and requires a GPU 3 for better performance [21].…”

Section: Introductionmentioning

confidence: 99%

Shared SAT Solvers and SAT Memory in Distributed Business Applications

Kozlovičs

2022

Communications in Computer and Information Science

View full text Add to dashboard Cite

We propose a software architecture where SAT solvers act as a shared network resource for distributed business applications. There can be multiple parallel SAT solvers running either on dedicated hardware (a multi-processor system or a system with a specific GPU) or in the cloud. In order to avoid complex message passing between network nodes, we introduce a novel concept of the shared SAT memory, which can be accessed (in the read/write mode) from multiple different SAT solvers and modules implementing the business logic. As a result, our architecture allows for the easy generation, diversification, and solving of SAT instances from existing high-level programming languages without the need to think about the network. We demonstrate our architecture on the use case of transforming the integer factorization problem to SAT.Keywords: SAT • distributed applications • software architecture • integer factorization 1 Though" finding the most direct low-degree polynomial reduction to SAT can be a challenge.

show abstract