Improved Bounds for Sampling Solutions of Random CNF Formulas
Abstract
Let be a random -CNF formula on variables and clauses, where each clause is a disjunction of literals chosen independently and uniformly. Our goal is to sample an approximately uniform solution of (or equivalently, approximate the partition function of ). Let be the density. The previous best algorithm runs in time for any [Galanis, Goldberg, Guo, and Yang, SIAM J. Comput.'21]. Our result significantly improves both bounds by providing an almost-linear time sampler for any . The density captures the \emph{average degree} in the random formula. In the worst-case model with bounded \emph{maximum degree}, current best efficient sampler works up to degree bound [He, Wang, and Yin, FOCS'22 and SODA'23], which is, for the first time, superseded by its average-case counterpart due to our bound. Our result is the first progress towards establishing the intuition that the solvability of the average-case model (random -CNF formula with bounded average degree) is better than the worst-case model (standard -CNF formula with bounded maximal degree) in terms of sampling solutions.
Cite
@article{arxiv.2207.11892,
title = {Improved Bounds for Sampling Solutions of Random CNF Formulas},
author = {Kun He and Kewen Wu and Kuan Yang},
journal= {arXiv preprint arXiv:2207.11892},
year = {2023}
}
Comments
51 pages, all proofs added, and bounds slightly improved