Improved Mixing of Critical Hardcore Model
Abstract
The hardcore model is one of the most classic and widely studied examples of undirected graphical models. Given a graph , the hardcore model describes a Gibbs distribution of -weighted independent sets of . In the last two decades, a beautiful computational phase transition has been established at a precise threshold where denotes the maximum degree, where the task of sampling independent sets transitions from polynomial-time solvable to computationally intractable. We study the critical hardcore model where and show that the Glauber dynamics, a simple yet popular Markov chain algorithm, mixes in time on any -vertex graph of maximum degree , significantly improving the previous upper bound by the recent work arXiv:2411.03413. Our improvement comes from an optimal bound on the -spectral independence for the hardcore model at all subcritical fugacity .
Keywords
Cite
@article{arxiv.2505.07515,
title = {Improved Mixing of Critical Hardcore Model},
author = {Zongchen Chen and Tianhui Jiang},
journal= {arXiv preprint arXiv:2505.07515},
year = {2026}
}
Comments
16 pages, addressed an error in the previous version involving $\ell_\infty$-spectral independence; see Remark 1.2 for details