EXTREMAL EIGENVALUES OF CRITICAL ERDOS-RENYI GRAPHS

被引:20
作者
Alt, Johannes [1 ]
Ducatez, Raphael [1 ]
Knowles, Antti [1 ]
机构
[1] Univ Geneva, Sect Math, Geneva, Switzerland
基金
欧洲研究理事会; 瑞士国家科学基金会;
关键词
Erdos-Renyi graphs; sparse random matrix; eigenvalue rigidity; SPECTRAL TECHNIQUES;
D O I
10.1214/20-AOP1483
中图分类号
O21 [概率论与数理统计]; C8 [统计学];
学科分类号
020208 ; 070103 ; 0714 ;
摘要
We complete the analysis of the extremal eigenvalues of the adjacency matrix A of the Erdos-Renyi graph G(N, d/N) in the critical regime d asymptotic to log N of the transition uncovered in (Ann. Inst. Henri Poincare Probab. Stat. 56 (2020) 2141-2161; Ann. Probab. 47 (2019) 1653-1676), where the regimes d >> log N and d << log N were studied. We establish a one-to-one correspondence between vertices of degree at least 2d and nontrivial (excluding the trivial top eigenvalue) eigenvalues of A/root d outside of the asymptotic bulk [-2, 2]. This correspondence implies that the transition characterized by the appearance of the eigenvalues outside of the asymptotic bulk takes place at the critical value d = d(*) = 1/log 4-1 log N. For d < d(*), we obtain rigidity bounds on the locations of all eigenvalues outside the interval [-2, 2], and for d > d(*), we show that no such eigenvalues exist. All of our estimates are quantitative with polynomial error probabilities. Our proof is based on a tridiagonal representation of the adjacency matrix and on a detailed analysis of the geometry of the neighbourhood of the large degree vertices. An important ingredient in our estimates is a matrix inequality obtained via the associated nonbacktracking matrix and an Ihara-Bass formula (Ann. Inst. Henri Poincare Probab. Stat. 56 (2020) 2141-2161). Our argument also applies to sparse Wigner matrices, defined as the Hadamard product of A and a Wigner matrix, in which case the role of the degrees is replaced by the squares of the l(2)-norms of the rows.
引用
收藏
页码:1347 / 1401
页数:55
相关论文
共 21 条
[1]   Spectral techniques in graph algorithms [J].
Alon, N .
LATIN '98: THEORETICAL INFORMATICS, 1998, 1380 :206-215
[2]  
Alt J., 2020, ARXIV200514180
[3]   Spectral radii of sparse random matrices [J].
Benaych-Georges, Florent ;
Bordenave, Charles ;
Knowles, Antti .
ANNALES DE L INSTITUT HENRI POINCARE-PROBABILITES ET STATISTIQUES, 2020, 56 (03) :2141-2161
[4]   LARGEST EIGENVALUES OF SPARSE INHOMOGENEOUS ERDOS-RENYI GRAPHS [J].
Benaych-Georges, Florent ;
Bordenave, Charles ;
Knowles, Antti .
ANNALS OF PROBABILITY, 2019, 47 (03) :1653-1676
[5]  
Bollobas B., 2001, Random Graphs, VSecond
[6]  
Chung F. R. K., 1997, SPECTRAL GRAPH THEOR, V92
[7]   SPECTRAL STATISTICS OF ERDOS-RENYI GRAPHS I: LOCAL SEMICIRCLE LAW [J].
Erdos, Laszlo ;
Knowles, Antti ;
Yau, Horng-Tzer ;
Yin, Jun .
ANNALS OF PROBABILITY, 2013, 41 (3B) :2279-2375
[8]   Spectral Statistics of ErdAs-R,nyi Graphs II: Eigenvalue Spacing and the Extreme Eigenvalues [J].
Erdos, Laszlo ;
Knowles, Antti ;
Yau, Horng-Tzer ;
Yin, Jun .
COMMUNICATIONS IN MATHEMATICAL PHYSICS, 2012, 314 (03) :587-640
[9]   Spectral techniques applied to sparse random graphs [J].
Feige, U ;
Ofek, E .
RANDOM STRUCTURES & ALGORITHMS, 2005, 27 (02) :251-275
[10]   THE EIGENVALUES OF RANDOM SYMMETRIC-MATRICES [J].
FUREDI, Z ;
KOMLOS, J .
COMBINATORICA, 1981, 1 (03) :233-241