The spectral gap of random regular graphs

Abstract

We bound the second eigenvalue of random $d$-regular graphs, for a wide range of degrees $d$, using a novel approach based on Fourier analysis. Let $G_{n, d}$ be a uniform random $d$-regular graph on $n$ vertices, and let $\lambda (G_{n, d})$ be its second largest eigenvalue by absolute value. For some constant $c > 0$ and any degree $d$ with $\log^{10} n \ll d \leq c n$, we show that $\lambda (G_{n, d}) = (2 + o(1)) \sqrt{d (n - d) / n}$ asymptotically almost surely. Combined with earlier results that cover the case of sparse random graphs, this fully determines the asymptotic value of $\lambda (G_{n, d})$ for all $d \leq c n$. To achieve this, we introduce new methods that use mechanisms from discrete Fourier analysis, and combine them with existing tools and estimates on $d$-regular random graphs - especially those of Liebenau and Wormald.