Clustering of Solutions in the Random Satisfiability Problem 论文
2005Physical Review Letters引用 220
Constraint Satisfaction and OptimizationData Management and AlgorithmsRough Sets and Fuzzy Logic
摘要
Using elementary rigorous methods we prove the existence of a clustered phase in the random K-SAT problem, for K > or = 8. In this phase the solutions are grouped into clusters which are far away from each other. The results are in agreement with previous predictions of the cavity method and give a rigorous confirmation to one of its main building blocks. It can be generalized to other systems of both physical and computational interest.