The expander mixing lemma intuitively states that the edges of certain -regular graphs are evenly distributed throughout the graph. In particular, the number of edges between two vertex subsets and is always close to the expected number of edges between them in a random -regular graph, namely .

d-Regular Expander Graphs edit

Define an  -graph to be a  -regular graph   on   vertices such that all of the eigenvalues of its adjacency matrix   except one have absolute value at most   The  -regularity of the graph guarantees that its largest absolute value of an eigenvalue is   In fact, the all-1's vector   is an eigenvector of   with eigenvalue  , and the eigenvalues of the adjacency matrix will never exceed the maximum degree of   in absolute value.

If we fix   and   then  -graphs form a family of expander graphs with a constant spectral gap.

Statement edit

Let   be an  -graph. For any two subsets  , let   be the number of edges between S and T (counting edges contained in the intersection of S and T twice). Then

 

Tighter Bound edit

We can in fact show that

 

using similar techniques.[1]

Biregular Graphs edit

For biregular graphs, we have the following variation, where we take   to be the second largest eigenvalue.[2]

Let   be a bipartite graph such that every vertex in   is adjacent to   vertices of   and every vertex in   is adjacent to   vertices of  . Let   with   and  . Let  . Then

 

Note that   is the largest eigenvalue of  .

Proofs edit

Proof of First Statement edit

Let   be the adjacency matrix of   and let   be the eigenvalues of   (these eigenvalues are real because   is symmetric). We know that   with corresponding eigenvector  , the normalization of the all-1's vector. Define   and note that  . Because   is symmetric, we can pick eigenvectors   of   corresponding to eigenvalues   so that   forms an orthonormal basis of  .

Let   be the   matrix of all 1's. Note that   is an eigenvector of   with eigenvalue   and each other  , being perpendicular to  , is an eigenvector of   with eigenvalue 0. For a vertex subset  , let   be the column vector with   coordinate equal to 1 if   and 0 otherwise. Then,

 .

Let  . Because   and   share eigenvectors, the eigenvalues of   are  . By the Cauchy-Schwarz inequality, we have that  . Furthermore, because   is self-adjoint, we can write

 .

This implies that   and  .

Proof Sketch of Tighter Bound edit

To show the tighter bound above, we instead consider the vectors   and  , which are both perpendicular to  . We can expand

 

because the other two terms of the expansion are zero. The first term is equal to  , so we find that

 

We can bound the right hand side by   using the same methods as in the earlier proof.

Applications edit

The expander mixing lemma can be used to upper bound the size of an independent set within a graph. In particular, the size of an independent set in an  -graph is at most   This is proved by letting   in the statement above and using the fact that  

An additional consequence is that, if   is an  -graph, then its chromatic number   is at least   This is because, in a valid graph coloring, the set of vertices of a given color is an independent set. By the above fact, each independent set has size at most   so at least   such sets are needed to cover all of the vertices.

A second application of the expander mixing lemma is to provide an upper bound on the maximum possible size of an independent set within a polarity graph. Given a finite projective plane   with a polarity   the polarity graph is a graph where the vertices are the points a of  , and vertices   and   are connected if and only if   In particular, if   has order   then the expander mixing lemma can show that an independent set in the polarity graph can have size at most   a bound proved by Hobart and Williford.

Converse edit

Bilu and Linial showed[3] that a converse holds as well: if a  -regular graph   satisfies that for any two subsets   with   we have

 

then its second-largest (in absolute value) eigenvalue is bounded by  .

Generalization to hypergraphs edit

Friedman and Widgerson proved the following generalization of the mixing lemma to hypergraphs.

Let   be a  -uniform hypergraph, i.e. a hypergraph in which every "edge" is a tuple of   vertices. For any choice of subsets   of vertices,

 

Notes edit

  1. ^ Vadhan, Salil (Spring 2009). "Expander Graphs" (PDF). Harvard University. Retrieved December 1, 2019.
  2. ^ See Theorem 5.1 in "Interlacing Eigenvalues and Graphs" by Haemers
  3. ^ Expander mixing lemma converse

References edit

  • Alon, N.; Chung, F. R. K. (1988), "Explicit construction of linear sized tolerant networks", Discrete Mathematics, 72 (1–3): 15–19, CiteSeerX 10.1.1.300.7495, doi:10.1016/0012-365X(88)90189-6.
  • F.C. Bussemaker, D.M. Cvetković, J.J. Seidel. Graphs related to exceptional root systems, Combinatorics (Proc. Fifth Hungarian Colloq., Keszthely, 1976), volume 18 of Colloq. Math. Soc. János Bolyai (1978), 185-191.