Characterizations of Strongly Regular Graphs: Part II: Bose-Mesner algebras of graphs. Sung Y. Song Iowa State University

Similar documents
Introduction to Association Schemes

A Questionable Distance-Regular Graph

Vertex subsets with minimal width and dual width

Tridiagonal pairs in algebraic graph theory

Chapter 1 Vector Spaces

Advanced Engineering Mathematics Prof. Pratima Panigrahi Department of Mathematics Indian Institute of Technology, Kharagpur

Math 4377/6308 Advanced Linear Algebra

Math 1060 Linear Algebra Homework Exercises 1 1. Find the complete solutions (if any!) to each of the following systems of simultaneous equations:

Review of Linear Algebra

Vector Spaces. distributive law u,v. Associative Law. 1 v v. Let 1 be the unit element in F, then

DS-GA 1002 Lecture notes 0 Fall Linear Algebra. These notes provide a review of basic concepts in linear algebra.

ELEMENTARY LINEAR ALGEBRA

Math Camp Lecture 4: Linear Algebra. Xiao Yu Wang. Aug 2010 MIT. Xiao Yu Wang (MIT) Math Camp /10 1 / 88

1. What is the determinant of the following matrix? a 1 a 2 4a 3 2a 2 b 1 b 2 4b 3 2b c 1. = 4, then det

The Terwilliger algebra of a Q-polynomial distance-regular graph with respect to a set of vertices

Math 4A Notes. Written by Victoria Kala Last updated June 11, 2017

Linear Algebra. Paul Yiu. Department of Mathematics Florida Atlantic University. Fall 2011

The Hamming Codes and Delsarte s Linear Programming Bound

Some Notes on Distance-Transitive and Distance-Regular Graphs

MATH 240 Spring, Chapter 1: Linear Equations and Matrices

Intriguing sets of vertices of regular graphs

3. Vector spaces 3.1 Linear dependence and independence 3.2 Basis and dimension. 5. Extreme points and basic feasible solutions

1 Matrices and Systems of Linear Equations. a 1n a 2n

INTRODUCTION TO LIE ALGEBRAS. LECTURE 2.

Upper triangular matrices and Billiard Arrays

Chapter 3 Transformations

1 Last time: least-squares problems

MATH 423 Linear Algebra II Lecture 33: Diagonalization of normal operators.

The Multiplicities of a Dual-thin Q-polynomial Association Scheme

Theorems of Erdős-Ko-Rado type in polar spaces

Complex Hadamard matrices and 3-class association schemes

MA 265 FINAL EXAM Fall 2012

MAT 242 CHAPTER 4: SUBSPACES OF R n

Chapter 2. Error Correcting Codes. 2.1 Basic Notions

Jim Lambers MAT 610 Summer Session Lecture 1 Notes

2. Every linear system with the same number of equations as unknowns has a unique solution.

Linear Algebra. Session 12

1 Linear Algebra Problems

MIT Algebraic techniques and semidefinite optimization February 16, Lecture 4

6. The scalar multiple of u by c, denoted by c u is (also) in V. (closure under scalar multiplication)

Chapter 6: Orthogonality

ELE/MCE 503 Linear Algebra Facts Fall 2018

Elementary maths for GMT

Lecture Summaries for Linear Algebra M51A

Knowledge Discovery and Data Mining 1 (VO) ( )

Math 309 Notes and Homework for Days 4-6

Quantum Computing Lecture 2. Review of Linear Algebra

Linear Algebra Review. Vectors

Ma/CS 6b Class 25: Error Correcting Codes 2

JOHNSON SCHEMES AND CERTAIN MATRICES WITH INTEGRAL EIGENVALUES

Detailed Proof of The PerronFrobenius Theorem

Triply Regular Graphs

D-bounded Distance-Regular Graphs

Orthogonality and Least Squares

Assignment 1 Math 5341 Linear Algebra Review. Give complete answers to each of the following questions. Show all of your work.

Linear Algebra March 16, 2019

EIGENVALUES OF THE ADJACENCY MATRIX OF CUBIC LATTICE GRAPHS

Graph fundamentals. Matrices associated with a graph

4 Vector Spaces. 4.1 Basic Definition and Examples. Lecture 10

Chapter 7 Network Flow Problems, I

β α : α β We say that is symmetric if. One special symmetric subset is the diagonal

Ma/CS 6b Class 23: Eigenvalues in Regular Graphs

The following definition is fundamental.

CLASS 12 ALGEBRA OF MATRICES

Lecture 8 : Eigenvalues and Eigenvectors

Mathematical Methods wk 1: Vectors

Mathematical Methods wk 1: Vectors

Applications of semidefinite programming in Algebraic Combinatorics

Properties of Matrices and Operations on Matrices

Introduction to Matrix Algebra

There are two things that are particularly nice about the first basis

CSL361 Problem set 4: Basic linear algebra

Nonnegative Matrices I

Elementary linear algebra

1. Let m 1 and n 1 be two natural numbers such that m > n. Which of the following is/are true?

Orthogonality. 6.1 Orthogonal Vectors and Subspaces. Chapter 6

Delsarte s linear programming bound

Tues Feb Vector spaces and subspaces. Announcements: Warm-up Exercise:

Basic Concepts in Linear Algebra

Linear Algebra M1 - FIB. Contents: 5. Matrices, systems of linear equations and determinants 6. Vector space 7. Linear maps 8.

Math 443 Differential Geometry Spring Handout 3: Bilinear and Quadratic Forms This handout should be read just before Chapter 4 of the textbook.

Linear algebra and applications to graphs Part 1

Stat 159/259: Linear Algebra Notes

and let s calculate the image of some vectors under the transformation T.

MATH 167: APPLIED LINEAR ALGEBRA Least-Squares

MODEL ANSWERS TO THE FIRST QUIZ. 1. (18pts) (i) Give the definition of a m n matrix. A m n matrix with entries in a field F is a function

Linear Algebra review Powers of a diagonalizable matrix Spectral decomposition

Econ Slides from Lecture 7

Complex Hadamard matrices and 3-class association schemes

MATH 235. Final ANSWERS May 5, 2015

Review of Basic Concepts in Linear Algebra

Algebraic Methods in Combinatorics

ECS130 Scientific Computing. Lecture 1: Introduction. Monday, January 7, 10:00 10:50 am

Strongly Regular Decompositions of the Complete Graph

Kernels of Directed Graph Laplacians. J. S. Caughman and J.J.P. Veerman

Supplementary Material for MTH 299 Online Edition

Real symmetric matrices/1. 1 Eigenvalues and eigenvectors

Vector spaces. EE 387, Notes 8, Handout #12

1. Vectors.

Meaning of the Hessian of a function in a critical point

Transcription:

Characterizations of Strongly Regular Graphs: Part II: Bose-Mesner algebras of graphs Sung Y. Song Iowa State University sysong@iastate.edu

Notation K: one of the fields R or C X: a nonempty finite set Mat X (K): the K-algebra consisting of the matrices over K, and whose rows and columns are indexed by X K X : the K-vector space of column vectors over K, and whose coordinates are indexed by X Observe Mat X (K) acts on K X by left multiplication.

We endow R X with the dot product u, v = u t v (u, v R X ), and we endow C X with the Hermitean dot product u, v = u t v (u, v C X ). For each x in X, let ˆx denote the vector in K X with a 1 in coordinate x, and zeros in all other coordinates. Observe that {ˆx : x X} is an orthonormal basis for K X.

Let Γ = (X, R) denote a graph. The binary relation on X of being connected by a path is an equivalence relation; the equivalence classes are known as the connected components of Γ. Γ is said to be connected whenever X is a connected component. Assume Γ is connected, and pick x, y X. By the distance from x to y, we mean the scalar (x, y) = min{l : path of length l from x to y}. By the diameter of Γ, we mean the scalar D := max{ (x, y) : x, y X}.

Let Γ be a finite simple (undirected) connected graph. The adjacency matrix for Γ is the matrix A Mat X (C) with xy entry 1 if {x, y} R A xy = ( x, y X). 0 if {x, y} R The Bose-Mesner algebra of Γ over K, is the subalgebra M of Mat X (K) generated by the adjacency matrix A of Γ. Observe that: the dimension of M as a K-vector space = the degree of the minimal polynomial of A = the number of distinct eigenvalues of Γ.

Lemma. Let Γ be a connected graph with diameter D. Let M denote the Bose-Mesner algebra of Γ over K. (i) The dimension of M as a K-vector space is at least D + 1. (ii) Γ has at least D + 1 distinct eigenvalues.

Let Γ = (X, R) denote a connected graph with diameter D. For each integer i (0 i D), let A i denote the matrix in Mat X (C) with x, y entry 1 if (x, y) = i (A i ) xy = ( x, y X). 0 if (x, y) i We call A i the ith distance matrix for Γ.

Let Γ = (X, R) denote a connected graph with diameter D, and distance matrices A 0, A 1,...,A D. (i) A 0, A 1,...,A D are linearly independent. (ii) For all x X, and for all integers i (0 i D), A iˆx = y X (x,y)=i ŷ.

Lemma. Let Γ = (X, R) be a connected graph with diameter D, and distance matrices A 0, A 1,...,A D. (i) A 0 = I. (ii) A 1 = A. (iii) A 0 + A 1 + + A D = J (the all 1s matrix). (iv) A t i = A i (0 i D).

Let Γ = (X, R) denote a connected graph with diameter D. Γ is said to be distance-regular whenever for all integers i, 0 i D and for all x, y X at distance (x, y) = i, the scalars c i := {z X : (x, z) = i 1, (y, z) = 1}, a i := {z X : (x, z) = i, (y, z) = 1}, b i := {z X : (x, z) = i + 1, (y, z) = 1} are constants that are independent of x and y. We refer to the c i, a i, b i as the intersection numbers of Γ.

Lemma Let Γ = (X, R) denote a distance-regular graph with diameter D. (a) c 0 = 0, a 0 = 0, b D = 0, c 1 = 1, and that c i > 0 (1 i D), b i > 0 (0 i D 1). (b) Γ is regular with valency k := b 0. (c) k = c i + a i + b i (0 i D).

Distance-regular graphs with diameter D The Hamming graph H(D, N), (N 2). X = the set of D-tuples of elements from the set {1, 2,...,N}, R = {xy : x, y X, x and y differ in exactly one coordinate}. Here c i = (N 1)i, b i = (D i)(n 1) (0 i D). H(D, 2) is called the D-cube.

The Johnson graph J(D, N), (N 2D). X = the set of D-element subsets of {1, 2,...,N}, R = {xy : x, y X, x y = D 1}. Here c i = i 2, b i = (D i)(n D i) (0 i D).

The q-johnson graph J q (D, N), (N 2D). Let V denote an N-dimensional vector space over a finite field GF(q). X = the set of D-dimensional subspaces of V. R = {xy : x, y X, dim(x y) = D 1}. Here c i = ( q i 1 ) 2 (0 i D), q 1 b i = q2i+1 (q D i 1)(q N D i 1) (q 1) 2 (0 i D).

Lemma. Let Γ = (X, R) be a distance-regular graph with diameter D. Then the distance matrices satisfy AA i = c i+1 A i+1 + a i A i + b i 1 A i 1 (1 i D), where A D+1 = 0, and for convenience c D+1 = 1.

Lemma. Let Γ = (X, R) denote a distance-regular graph with diameter D. Let M denote the Bose-Mesner algebra of Γ over K. (i) The distance matrices A 0, A 1,...,A D form a basis for M. (ii) The dimension of M as a K-vector space equals D + 1. (iii) Γ has exactly D + 1 distinct eigenvalues.

How to compute the eigenvalues of a distance-regular graph from the intersection numbers? Def. Let Γ = (X, R) be a distance-regular graph with diameter D. We define the polynomials f 0, f 1,...,f D+1 C[λ] by f 0 = 1, f 1 = λ, λf i = c i+1 f i+1 + a i f i + b i 1 f i 1 (1 i D), where for convenience we set c D+1 = 1.

Lemma. Let the polynomials f 0, f 1,...,f D+1 be as in the above definition. (i) deg f i = i (0 i D + 1). (ii) The coefficient of λ i in f i is (c 1 c 2 c i ) 1 (0 i D + 1). (iii) f i (A) = A i (0 i D + 1). In particular, f D+1 (A) = 0. (iv) The distinct eigenvalues of Γ are precisely the zeros of f D+1.

We now wish to expand our notion of an intersection number. Let Γ = (X, R) denote a distance-regular graph with diameter D. Since the distance matrices A 0, A 1,...,A D are a basis for the Bose-Mesner algebra, there exist scalars p h ij C (0 h, i, j D) such that D A i A j = (0 i, j D). h=0 Since A i and A j commute, p h ija h p h ij = p h ji (0 h, i, j D).

We find that for all integers h, i, j (0 h, i, j D), and for all x, y X such that (x, y) = h, p h ij = {z X (x, z) = i, (y, z) = j}. In particular, each p h ij definitions, we see that is a nonnegative integer. Comparing the c i = p i i 1,1 a i = p i i1 (1 i D), (0 i D), b i = p i i+1,1 (0 i D 1). In view of this, we often refer to any p h ij number of Γ. as an intersection

It turns out that all of the p h ij b 0, b 1,...,b D 1, c 1, c 2,...,c D. can be computed from Set k i := p 0 ii (0 i D). With h = 0, i = j, x = y, we find that for all integers i (0 i D), and for all x X, k i = {z X : (x, z) = i}. In particular, and k 0 = 1, k 1 = k, k i 0 (0 i D). We refer to k i as the ith valency of Γ.

Lemma. Let Γ = (X, R) be a distance-regular graph with diameter D. Then (i) k h p h ij = k ip i hj = k jp j ih (0 h, i, j D). (ii) k i 1 b i 1 = k i c i (iii) (1 i D). k i = b 0b 1 b i 1 c 1 c 2 c i (0 i D).

Lemma. Let Γ = (X, R) denote a distance-regular graph with diameter D. Then for all integers h, i, j (0 h, i, j D), (i) p h ij (ii) p h ij = 0 if one of h, i, j is greater than the sum of the other two; 0 if one of h, i, j is equal to the sum of the other two.

In the sequel, we hope to continue to study on distance-regular graphs. However, (considering the schedule having been booked up for the semester,) next time it is illuminating to consider a slightly more general object called a commutative association scheme. For these schemes one still has intersection numbers p h ij ; the distance regular graphs correspond to the case where the intersection numbers satisfy (i) and (ii) in the last lemma.