Source-linked AI summary
Ramanujan Graphs
Alexander Lubotzky
TL;DR
Ramanujan graphs are regular graphs whose nontrivial adjacency eigenvalues match the spectrum of the universal covering tree, making them spectrally optimal expanders. The paper explains their connection to Ramanujan’s conjecture on Hecke eigenvalues and describes arithmetic constructions of such graphs.
Problem
Constructing k-regular graphs with optimally bounded nontrivial eigenvalues is difficult: random regular graphs are known to be expanders, but it is not known whether they are Ramanujan.
Method
The constructions use quotients of Bruhat–Tits trees by arithmetic subgroups whose associated representation-theoretic temperedness conditions follow from the Ramanujan–Peterson conjecture and the Jacquet–Langlands correspondence.
Results
Ramanujan graphs have all nontrivial eigenvalues within the spectrum of the infinite k-regular tree and are therefore optimal expanders.
Takeaways & Limitations
Their eigenvalue bounds yield the fastest possible convergence of random walks to the uniform distribution and support applications in networks, algorithms, combinatorics, and pure mathematics.
Abstract
from arXiv · showhide
This is an item on Ramanujan Graphs for a planned encyclopedia on Ramanujan. The notion of Ramanujan graphs is explained, as well as the reason to name these graphs after Ramanujan.