English

Spectrum of Random $d$-regular Graphs Up to the Edge

Probability 2025-02-04 v3 Mathematical Physics Combinatorics math.MP

Abstract

Consider the normalized adjacency matrices of random dd-regular graphs on NN vertices with fixed degree d3d\geq3. We prove that, with probability 1N1+ε1-N^{-1+{\varepsilon}} for any ε>0{\varepsilon} >0, the following two properties hold as NN \to \infty provided that d3d\geq3: (i) The eigenvalues are close to the classical eigenvalue locations given by the Kesten-McKay distribution. In particular, the extremal eigenvalues are concentrated with polynomial error bound in NN, i.e. λ2,λN2+Nc\lambda_2, |\lambda_N|\leq 2+N^{-c}. (ii) All eigenvectors of random dd-regular graphs are completely delocalized.

Keywords

Cite

@article{arxiv.2102.00963,
  title  = {Spectrum of Random $d$-regular Graphs Up to the Edge},
  author = {Jiaoyang Huang and Horng-Tzer Yau},
  journal= {arXiv preprint arXiv:2102.00963},
  year   = {2025}
}

Comments

Updated version with corrected typos