English

Invertibility of Sparse non-Hermitian matrices

Probability 2017-02-06 v3

Abstract

We consider a class of sparse random matrices of the form An=(ξi,jδi,j)i,j=1nA_n =(\xi_{i,j}\delta_{i,j})_{i,j=1}^n, where {ξi,j}\{\xi_{i,j}\} are i.i.d.~centered random variables, and {δi,j}\{\delta_{i,j}\} are i.i.d.~Bernoulli random variables taking value 11 with probability pnp_n, and prove a quantitative estimate on the smallest singular value for pn=Ω(lognn)p_n = \Omega(\frac{\log n}{n}), under a suitable assumption on the spectral norm of the matrices. This establishes the invertibility of a large class of sparse matrices. For pn=Ω(nα)p_n =\Omega( n^{-\alpha}) with some α(0,1)\alpha \in (0,1), we deduce that the condition number of AnA_n is of order nn with probability tending to one under the optimal moment assumption on {ξi,j}\{\xi_{i,j}\}. This in particular, extends a conjecture of von Neumann about the condition number to sparse random matrices with heavy-tailed entries. In the case that the random variables {ξi,j}\{\xi_{i,j}\} are i.i.d.~sub-Gaussian, we further show that a sparse random matrix is singular with probability at most exp(cnpn)\exp(-c n p_n) whenever pnp_n is above the critical threshold pn=Ω(lognn)p_n = \Omega(\frac{ \log n}{n}). The results also extend to the case when {ξi,j}\{\xi_{i,j}\} have a non-zero mean. We further find quantitative estimates on the smallest singular value of the adjacency matrix of a directed Erd\H{o}s-R\'{e}yni graph whenever its edge connectivity probability is above the critical threshold Ω(lognn)\Omega(\frac{\log n}{n}).

Keywords

Cite

@article{arxiv.1507.03525,
  title  = {Invertibility of Sparse non-Hermitian matrices},
  author = {Anirban Basak and Mark Rudelson},
  journal= {arXiv preprint arXiv:1507.03525},
  year   = {2017}
}

Comments

46 pages, minor changes in V3. To appear in Advances in Mathematics

R2 v1 2026-06-22T10:10:54.832Z