StructureSAT
收藏资源简介:
StructureSAT是一个大规模的饱和问题(SAT)数据集,由新南威尔士大学计算机科学与工程学院创建。该数据集包含来自不同问题领域的多样化SAT问题,旨在研究图神经网络(GNN)在SAT问题上的泛化能力。数据集涵盖了随机、设计、伪工业和工业等11个SAT领域,并研究了9种对传统SAT求解器有影响的图形结构属性。数据集通过结构属性对问题领域进行细分,以研究不同结构属性对GNN求解器泛化能力的影响。
StructureSAT is a large-scale Satisfiability Problem (SAT) dataset developed by the School of Computer Science and Engineering at the University of New South Wales. This dataset includes diverse SAT problems from various problem domains, with the aim of researching the generalization capability of Graph Neural Networks (GNNs) on SAT problems. The dataset covers 11 SAT domains such as random, crafted, pseudo-industrial and industrial ones, and involves 9 graph structural attributes that affect traditional SAT solvers. The dataset subdivides problem domains based on structural attributes to study the impact of different structural attributes on the generalization ability of GNN-based solvers.




