Top Banner
Combinatorial Matrix Theory Wayne Barrett August 30, 2013 Brigham Young University Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 1 / 55
252

Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Jun 25, 2020

Download

Documents

dariahiddleston
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Combinatorial Matrix Theory

Wayne Barrett

August 30, 2013

Brigham Young University

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 1 / 55

Page 2: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Combinatorial Matrix Theory

Fusion of Graph Theory and Matrix Theory

Background in Graph Theory.

6 vertices 9 edges

Cycles in a graph

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 2 / 55

Page 3: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Combinatorial Matrix Theory

Fusion of Graph Theory and Matrix Theory

Background in Graph Theory.

6 vertices 9 edges

Cycles in a graph

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 2 / 55

Page 4: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Combinatorial Matrix Theory

Fusion of Graph Theory and Matrix Theory

Background in Graph Theory.

6 vertices 9 edges

Cycles in a graph

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 2 / 55

Page 5: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Combinatorial Matrix Theory

Fusion of Graph Theory and Matrix Theory

Background in Graph Theory.

6 vertices 9 edges

Cycles in a graph

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 2 / 55

Page 6: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Combinatorial Matrix Theory

Fusion of Graph Theory and Matrix Theory

Background in Graph Theory.

6 vertices 9 edges

Cycles in a graph

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 2 / 55

Page 7: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Combinatorial Matrix Theory

Fusion of Graph Theory and Matrix Theory

Background in Graph Theory.

6 vertices 9 edges

Cycles in a graph

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 2 / 55

Page 8: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Combinatorial Matrix Theory

Fusion of Graph Theory and Matrix Theory

Background in Graph Theory.

6 vertices 9 edges

Cycles in a graph

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 2 / 55

Page 9: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian cycles

Definition

A Hamiltonian cycle is one that passes through all the vertices. A graph isHamiltonian if it contains a Hamiltonian cycle.

is Hamiltonian is not

A simple necessary condition for a graph to be Hamiltonian is that everyvertex have degree at least 2.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 3 / 55

Page 10: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian cycles

Definition

A Hamiltonian cycle is one that passes through all the vertices. A graph isHamiltonian if it contains a Hamiltonian cycle.

is Hamiltonian is not

A simple necessary condition for a graph to be Hamiltonian is that everyvertex have degree at least 2.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 3 / 55

Page 11: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian cycles

Definition

A Hamiltonian cycle is one that passes through all the vertices. A graph isHamiltonian if it contains a Hamiltonian cycle.

is Hamiltonian

is not

A simple necessary condition for a graph to be Hamiltonian is that everyvertex have degree at least 2.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 3 / 55

Page 12: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian cycles

Definition

A Hamiltonian cycle is one that passes through all the vertices. A graph isHamiltonian if it contains a Hamiltonian cycle.

is Hamiltonian is not

A simple necessary condition for a graph to be Hamiltonian is that everyvertex have degree at least 2.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 3 / 55

Page 13: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian cycles

Definition

A Hamiltonian cycle is one that passes through all the vertices. A graph isHamiltonian if it contains a Hamiltonian cycle.

is Hamiltonian is not

A simple necessary condition for a graph to be Hamiltonian is that everyvertex have degree at least 2.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 3 / 55

Page 14: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian cycles

Not sufficient K2,3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 4 / 55

Page 15: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian cycles

Not sufficient K2,3

no cycle of odd lengthno 5-cycleno Hamiltonian cycle

No known necessary and sufficient condition(s) for a graph to have aHamiltonian cycle (other than the definition).

NP hard.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 5 / 55

Page 16: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian cycles

Not sufficient K2,3

no cycle of odd length

no 5-cycleno Hamiltonian cycle

No known necessary and sufficient condition(s) for a graph to have aHamiltonian cycle (other than the definition).

NP hard.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 5 / 55

Page 17: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian cycles

Not sufficient K2,3

no cycle of odd lengthno 5-cycle

no Hamiltonian cycle

No known necessary and sufficient condition(s) for a graph to have aHamiltonian cycle (other than the definition).

NP hard.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 5 / 55

Page 18: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian cycles

Not sufficient K2,3

no cycle of odd lengthno 5-cycleno Hamiltonian cycle

No known necessary and sufficient condition(s) for a graph to have aHamiltonian cycle (other than the definition).

NP hard.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 5 / 55

Page 19: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian cycles

Not sufficient K2,3

no cycle of odd lengthno 5-cycleno Hamiltonian cycle

No known necessary and sufficient condition(s) for a graph to have aHamiltonian cycle (other than the definition).

NP hard.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 5 / 55

Page 20: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian cycles

Not sufficient K2,3

no cycle of odd lengthno 5-cycleno Hamiltonian cycle

No known necessary and sufficient condition(s) for a graph to have aHamiltonian cycle (other than the definition).

NP hard.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 5 / 55

Page 21: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Two renowned 3-regular graphs

5-prism Petersen graph

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 6 / 55

Page 22: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Two renowned 3-regular graphs

5-prism Petersen graph

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 6 / 55

Page 23: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Two renowned 3-regular graphs

5-prism Petersen graph

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 7 / 55

Page 24: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Two renowned 3-regular graphs

5-prism Petersen graph

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 8 / 55

Page 25: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Two renowned 3-regular graphs

5-prism Petersen graph

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 9 / 55

Page 26: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Two renowned 3-regular graphs

5-prism Petersen graph

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 10 / 55

Page 27: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory

Adjacency matrix of a graph G

1

3 4

2

A(G ) =

0 1 1 01 0 1 01 1 0 10 0 1 0

.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 11 / 55

Page 28: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory

Adjacency matrix of a graph G

1

3 4

2

A(G ) =

0 1 1 01 0 1 01 1 0 10 0 1 0

.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 11 / 55

Page 29: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory

Adjacency matrix of a graph G

1

3 4

2

A(G ) =

0 1 1 01 0 1 01 1 0 10 0 1 0

.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 11 / 55

Page 30: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory

Adjacency matrix of a graph G

1

3 4

2

A(G ) =

0 1 1 01 0 1 01 1 0 10 0 1 0

.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 12 / 55

Page 31: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory

Adjacency matrix of a graph G

1

3 4

2

A(G ) =

0 1 1 01 0 1 01 1 0 10 0 1 0

.

Definition

If G is a graph on n vertices, its adjacency matrix A(G ) is the n × nsymmetric (0, 1)-matrix defined by

aij =

{1 if {i , j} is an edge of G

0 otherwise.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 13 / 55

Page 32: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory

Adjacency matrix of a graph G

1

3 4

2

A(G ) =

0 1 1 01 0 1 01 1 0 10 0 1 0

.

Definition

If G is a graph on n vertices, its adjacency matrix A(G ) is the n × nsymmetric (0, 1)-matrix defined by

aij =

{1 if {i , j} is an edge of G

0 otherwise.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 13 / 55

Page 33: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory

Adjacency matrix of a graph G

1

3 4

2

A(G ) =

0 1 1 01 0 1 01 1 0 10 0 1 0

.

Definition

If G is a graph on n vertices, its adjacency matrix A(G ) is the n × nsymmetric (0, 1)-matrix defined by

aij =

{1 if {i , j} is an edge of G

0 otherwise.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 13 / 55

Page 34: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory

Adjacency matrix of a graph G

1

3 4

2

A(G ) =

0 1 1 01 0 1 01 1 0 10 0 1 0

.

Definition

If G is a graph on n vertices, its adjacency matrix A(G ) is the n × nsymmetric (0, 1)-matrix defined by

aij =

{1 if {i , j} is an edge of G

0 otherwise.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 13 / 55

Page 35: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory

Goal of spectral graph theory:

Determine properties of G from theeigenvalues of A(G ) or the eigenvalues of other matrices associated withG .

Example:Eigenvalues of Cn.

n

1

6 4

2

5

3 A(Cn) =

0 1 0 0 11 0 1 0 0

0 1 0 1. . .

0 1 0. . .

. . .. . .

. . .. . .

. . . 0

0. . .

. . .. . . 1

1 0 0 1 0

.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 14 / 55

Page 36: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory

Goal of spectral graph theory: Determine properties of G from theeigenvalues of A(G ) or the eigenvalues of other matrices associated withG .

Example:Eigenvalues of Cn.

n

1

6 4

2

5

3 A(Cn) =

0 1 0 0 11 0 1 0 0

0 1 0 1. . .

0 1 0. . .

. . .. . .

. . .. . .

. . . 0

0. . .

. . .. . . 1

1 0 0 1 0

.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 14 / 55

Page 37: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory

Goal of spectral graph theory: Determine properties of G from theeigenvalues of A(G ) or the eigenvalues of other matrices associated withG .

Example:Eigenvalues of Cn.

n

1

6 4

2

5

3 A(Cn) =

0 1 0 0 11 0 1 0 0

0 1 0 1. . .

0 1 0. . .

. . .. . .

. . .. . .

. . . 0

0. . .

. . .. . . 1

1 0 0 1 0

.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 14 / 55

Page 38: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory

Goal of spectral graph theory: Determine properties of G from theeigenvalues of A(G ) or the eigenvalues of other matrices associated withG .

Example:Eigenvalues of Cn.

n

1

6 4

2

5

3

A(Cn) =

0 1 0 0 11 0 1 0 0

0 1 0 1. . .

0 1 0. . .

. . .. . .

. . .. . .

. . . 0

0. . .

. . .. . . 1

1 0 0 1 0

.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 14 / 55

Page 39: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory

Goal of spectral graph theory: Determine properties of G from theeigenvalues of A(G ) or the eigenvalues of other matrices associated withG .

Example:Eigenvalues of Cn.

n

1

6 4

2

5

3 A(Cn) =

0 1 0 0 11 0 1 0 0

0 1 0 1. . .

0 1 0. . .

. . .. . .

. . .. . .

. . . 0

0. . .

. . .. . . 1

1 0 0 1 0

.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 14 / 55

Page 40: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Cn

A(Cn) = Qn + Q−1n Qn =

0 1 0 0

0 0 1. . .

0 0. . . 0

0. . .

. . . 11 0 0 0

.

Qn is orthogonal so all eigenvalues have absolute value 1.

Char poly PQn(t) = tn − 1.

Eigenvalues of Qn are n roots of unity λj = e2πijn , j = 0, 1, 2, . . . , n − 1.

Qn is normal so has an orthonormal basis {u1, . . . , un} of eigenvectors.

Qnuj = λjuj .

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 15 / 55

Page 41: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Cn

A(Cn) = Qn + Q−1n Qn =

0 1 0 0

0 0 1. . .

0 0. . . 0

0. . .

. . . 11 0 0 0

.

Qn is orthogonal so all eigenvalues have absolute value 1.

Char poly PQn(t) = tn − 1.

Eigenvalues of Qn are n roots of unity λj = e2πijn , j = 0, 1, 2, . . . , n − 1.

Qn is normal so has an orthonormal basis {u1, . . . , un} of eigenvectors.

Qnuj = λjuj .

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 15 / 55

Page 42: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Cn

A(Cn) = Qn + Q−1n Qn =

0 1 0 0

0 0 1. . .

0 0. . . 0

0. . .

. . . 11 0 0 0

.

Qn is orthogonal so all eigenvalues have absolute value 1.

Char poly PQn(t) = tn − 1.

Eigenvalues of Qn are n roots of unity λj = e2πijn , j = 0, 1, 2, . . . , n − 1.

Qn is normal so has an orthonormal basis {u1, . . . , un} of eigenvectors.

Qnuj = λjuj .

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 15 / 55

Page 43: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Cn

A(Cn) = Qn + Q−1n Qn =

0 1 0 0

0 0 1. . .

0 0. . . 0

0. . .

. . . 11 0 0 0

.

Qn is orthogonal so all eigenvalues have absolute value 1.

Char poly PQn(t) = tn − 1.

Eigenvalues of Qn are n roots of unity λj = e2πijn , j = 0, 1, 2, . . . , n − 1.

Qn is normal so has an orthonormal basis {u1, . . . , un} of eigenvectors.

Qnuj = λjuj .

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 15 / 55

Page 44: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Cn

A(Cn) = Qn + Q−1n Qn =

0 1 0 0

0 0 1. . .

0 0. . . 0

0. . .

. . . 11 0 0 0

.

Qn is orthogonal so all eigenvalues have absolute value 1.

Char poly PQn(t) = tn − 1.

Eigenvalues of Qn are n roots of unity λj = e2πijn , j = 0, 1, 2, . . . , n − 1.

Qn is normal so has an orthonormal basis {u1, . . . , un} of eigenvectors.

Qnuj = λjuj .

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 15 / 55

Page 45: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Cn

The uj are also eigenvectors of Q−1n

Q−1n uj =

(1

λj

)uj .

So the uj are also eigenvectors of A(Cn).

A(Cn)uj = (Qn + Q−1n )uj =

(λj + 1

λj

)uj

=(

e2πijn + e−

2πijn

)uj = 2 cos 2πj

n uj .

FACT The eigenvalues of A(Cn) are 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 16 / 55

Page 46: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Cn

The uj are also eigenvectors of Q−1n Q−1

n uj =

(1

λj

)uj .

So the uj are also eigenvectors of A(Cn).

A(Cn)uj = (Qn + Q−1n )uj =

(λj + 1

λj

)uj

=(

e2πijn + e−

2πijn

)uj = 2 cos 2πj

n uj .

FACT The eigenvalues of A(Cn) are 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 16 / 55

Page 47: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Cn

The uj are also eigenvectors of Q−1n Q−1

n uj =

(1

λj

)uj .

So the uj are also eigenvectors of A(Cn).

A(Cn)uj = (Qn + Q−1n )uj =

(λj + 1

λj

)uj

=(

e2πijn + e−

2πijn

)uj = 2 cos 2πj

n uj .

FACT The eigenvalues of A(Cn) are 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 16 / 55

Page 48: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Cn

The uj are also eigenvectors of Q−1n Q−1

n uj =

(1

λj

)uj .

So the uj are also eigenvectors of A(Cn).

A(Cn)uj

= (Qn + Q−1n )uj =

(λj + 1

λj

)uj

=(

e2πijn + e−

2πijn

)uj = 2 cos 2πj

n uj .

FACT The eigenvalues of A(Cn) are 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 16 / 55

Page 49: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Cn

The uj are also eigenvectors of Q−1n Q−1

n uj =

(1

λj

)uj .

So the uj are also eigenvectors of A(Cn).

A(Cn)uj = (Qn + Q−1n )uj

=(λj + 1

λj

)uj

=(

e2πijn + e−

2πijn

)uj = 2 cos 2πj

n uj .

FACT The eigenvalues of A(Cn) are 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 16 / 55

Page 50: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Cn

The uj are also eigenvectors of Q−1n Q−1

n uj =

(1

λj

)uj .

So the uj are also eigenvectors of A(Cn).

A(Cn)uj = (Qn + Q−1n )uj =

(λj + 1

λj

)uj

=(

e2πijn + e−

2πijn

)uj = 2 cos 2πj

n uj .

FACT The eigenvalues of A(Cn) are 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 16 / 55

Page 51: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Cn

The uj are also eigenvectors of Q−1n Q−1

n uj =

(1

λj

)uj .

So the uj are also eigenvectors of A(Cn).

A(Cn)uj = (Qn + Q−1n )uj =

(λj + 1

λj

)uj

=(

e2πijn + e−

2πijn

)uj

= 2 cos 2πjn uj .

FACT The eigenvalues of A(Cn) are 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 16 / 55

Page 52: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Cn

The uj are also eigenvectors of Q−1n Q−1

n uj =

(1

λj

)uj .

So the uj are also eigenvectors of A(Cn).

A(Cn)uj = (Qn + Q−1n )uj =

(λj + 1

λj

)uj

=(

e2πijn + e−

2πijn

)uj = 2 cos 2πj

n uj .

FACT The eigenvalues of A(Cn) are 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 16 / 55

Page 53: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Cn

The uj are also eigenvectors of Q−1n Q−1

n uj =

(1

λj

)uj .

So the uj are also eigenvectors of A(Cn).

A(Cn)uj = (Qn + Q−1n )uj =

(λj + 1

λj

)uj

=(

e2πijn + e−

2πijn

)uj = 2 cos 2πj

n uj .

FACT The eigenvalues of A(Cn) are 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 16 / 55

Page 54: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Laplacian matrix of a graph

1

2

3 4 L(G ) =

2 −1 −1 0−1 2 −1 0−1 −1 3 −1

0 0 −1 1

.

Definition

If G is a graph on n vertices, L(G ) is the n × n symmetric integer matrixdefined by

`ij =

deg(i) if i = j

−1 if i 6= j and {i , j} is an edge of G .

0 otherwise.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 17 / 55

Page 55: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Laplacian matrix of a graph

1

2

3 4

L(G ) =

2 −1 −1 0−1 2 −1 0−1 −1 3 −1

0 0 −1 1

.

Definition

If G is a graph on n vertices, L(G ) is the n × n symmetric integer matrixdefined by

`ij =

deg(i) if i = j

−1 if i 6= j and {i , j} is an edge of G .

0 otherwise.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 17 / 55

Page 56: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Laplacian matrix of a graph

1

2

3 4 L(G ) =

2 −1 −1 0−1 2 −1 0−1 −1 3 −1

0 0 −1 1

.

Definition

If G is a graph on n vertices, L(G ) is the n × n symmetric integer matrixdefined by

`ij =

deg(i) if i = j

−1 if i 6= j and {i , j} is an edge of G .

0 otherwise.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 17 / 55

Page 57: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Laplacian matrix of a graph

1

2

3 4 L(G ) =

2 −1 −1 0−1 2 −1 0−1 −1 3 −1

0 0 −1 1

.

Definition

If G is a graph on n vertices, L(G ) is the n × n symmetric integer matrixdefined by

`ij =

deg(i) if i = j

−1 if i 6= j and {i , j} is an edge of G .

0 otherwise.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 17 / 55

Page 58: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Laplacian matrix of a graph

1

2

3 4 L(G ) =

2 −1 −1 0−1 2 −1 0−1 −1 3 −1

0 0 −1 1

.

Definition

If G is a graph on n vertices, L(G ) is the n × n symmetric integer matrixdefined by

`ij =

deg(i) if i = j

−1 if i 6= j and {i , j} is an edge of G .

0 otherwise.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 17 / 55

Page 59: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

L(Cn)

L(Cn) =

2 −1 0 0 −1−1 2 −1 0 0

0 −1 2 −1. . .

0 −1 2. . .

. . .. . .

. . .. . .

. . . 0

0. . .

. . .. . . −1

−1 0 0 −1 2

= 2In − A(Cn).

Eigenvalues: L(Cn)uj = (2In − A(Cn))uj =

(2− 2 cos

2πj

n

)uj .

FACT: Eigenvalues of L(Cn) are 2− 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 18 / 55

Page 60: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

L(Cn)

L(Cn) =

2 −1 0 0 −1−1 2 −1 0 0

0 −1 2 −1. . .

0 −1 2. . .

. . .. . .

. . .. . .

. . . 0

0. . .

. . .. . . −1

−1 0 0 −1 2

= 2In − A(Cn).

Eigenvalues: L(Cn)uj = (2In − A(Cn))uj =

(2− 2 cos

2πj

n

)uj .

FACT: Eigenvalues of L(Cn) are 2− 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 18 / 55

Page 61: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

L(Cn)

L(Cn) =

2 −1 0 0 −1−1 2 −1 0 0

0 −1 2 −1. . .

0 −1 2. . .

. . .. . .

. . .. . .

. . . 0

0. . .

. . .. . . −1

−1 0 0 −1 2

= 2In − A(Cn).

Eigenvalues: L(Cn)uj = (2In − A(Cn))uj =

(2− 2 cos

2πj

n

)uj .

FACT: Eigenvalues of L(Cn) are 2− 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 18 / 55

Page 62: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

L(Cn)

L(Cn) =

2 −1 0 0 −1−1 2 −1 0 0

0 −1 2 −1. . .

0 −1 2. . .

. . .. . .

. . .. . .

. . . 0

0. . .

. . .. . . −1

−1 0 0 −1 2

= 2In − A(Cn).

Eigenvalues:

L(Cn)uj = (2In − A(Cn))uj =

(2− 2 cos

2πj

n

)uj .

FACT: Eigenvalues of L(Cn) are 2− 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 18 / 55

Page 63: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

L(Cn)

L(Cn) =

2 −1 0 0 −1−1 2 −1 0 0

0 −1 2 −1. . .

0 −1 2. . .

. . .. . .

. . .. . .

. . . 0

0. . .

. . .. . . −1

−1 0 0 −1 2

= 2In − A(Cn).

Eigenvalues: L(Cn)uj = (2In − A(Cn))uj

=

(2− 2 cos

2πj

n

)uj .

FACT: Eigenvalues of L(Cn) are 2− 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 18 / 55

Page 64: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

L(Cn)

L(Cn) =

2 −1 0 0 −1−1 2 −1 0 0

0 −1 2 −1. . .

0 −1 2. . .

. . .. . .

. . .. . .

. . . 0

0. . .

. . .. . . −1

−1 0 0 −1 2

= 2In − A(Cn).

Eigenvalues: L(Cn)uj = (2In − A(Cn))uj =

(2− 2 cos

2πj

n

)uj .

FACT: Eigenvalues of L(Cn) are 2− 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 18 / 55

Page 65: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

L(Cn)

L(Cn) =

2 −1 0 0 −1−1 2 −1 0 0

0 −1 2 −1. . .

0 −1 2. . .

. . .. . .

. . .. . .

. . . 0

0. . .

. . .. . . −1

−1 0 0 −1 2

= 2In − A(Cn).

Eigenvalues: L(Cn)uj = (2In − A(Cn))uj =

(2− 2 cos

2πj

n

)uj .

FACT: Eigenvalues of L(Cn) are 2− 2 cos2πj

n, j = 0, . . . n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 18 / 55

Page 66: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Weyl’s inequalities

If A, B are real symmetric n × n matrices, what is the relationship of theeigenvalues of A + B to those of A and B?

Theorem

[Weyl] Let

λ1(A) ≥ λ2(A) ≥ · · · ≥ λn(A)

λ1(B) ≥ λ2(B) ≥ · · · ≥ λn(B)

be the eigenvalues of the real symmetric n × n matrices A, B. Then

λj(A) + λ1(B) ≥ λj(A + B) ≥ λj(A) + λn(B).

Corollary

If B is positive semidefinite, λj(A + B) ≥ λj(A).

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 19 / 55

Page 67: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Weyl’s inequalities

If A, B are real symmetric n × n matrices, what is the relationship of theeigenvalues of A + B to those of A and B?

Theorem

[Weyl] Let

λ1(A) ≥ λ2(A) ≥ · · · ≥ λn(A)

λ1(B) ≥ λ2(B) ≥ · · · ≥ λn(B)

be the eigenvalues of the real symmetric n × n matrices A, B. Then

λj(A) + λ1(B) ≥ λj(A + B) ≥ λj(A) + λn(B).

Corollary

If B is positive semidefinite, λj(A + B) ≥ λj(A).

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 19 / 55

Page 68: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Weyl’s inequalities

If A, B are real symmetric n × n matrices, what is the relationship of theeigenvalues of A + B to those of A and B?

Theorem

[Weyl] Let

λ1(A) ≥ λ2(A) ≥ · · · ≥ λn(A)

λ1(B) ≥ λ2(B) ≥ · · · ≥ λn(B)

be the eigenvalues of the real symmetric n × n matrices A, B.

Then

λj(A) + λ1(B) ≥ λj(A + B) ≥ λj(A) + λn(B).

Corollary

If B is positive semidefinite, λj(A + B) ≥ λj(A).

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 19 / 55

Page 69: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Weyl’s inequalities

If A, B are real symmetric n × n matrices, what is the relationship of theeigenvalues of A + B to those of A and B?

Theorem

[Weyl] Let

λ1(A) ≥ λ2(A) ≥ · · · ≥ λn(A)

λ1(B) ≥ λ2(B) ≥ · · · ≥ λn(B)

be the eigenvalues of the real symmetric n × n matrices A, B. Then

λj(A) + λ1(B) ≥ λj(A + B) ≥ λj(A) + λn(B).

Corollary

If B is positive semidefinite, λj(A + B) ≥ λj(A).

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 19 / 55

Page 70: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Weyl’s inequalities

If A, B are real symmetric n × n matrices, what is the relationship of theeigenvalues of A + B to those of A and B?

Theorem

[Weyl] Let

λ1(A) ≥ λ2(A) ≥ · · · ≥ λn(A)

λ1(B) ≥ λ2(B) ≥ · · · ≥ λn(B)

be the eigenvalues of the real symmetric n × n matrices A, B. Then

λj(A) + λ1(B) ≥ λj(A + B) ≥ λj(A) + λn(B).

Corollary

If B is positive semidefinite,

λj(A + B) ≥ λj(A).

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 19 / 55

Page 71: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Weyl’s inequalities

If A, B are real symmetric n × n matrices, what is the relationship of theeigenvalues of A + B to those of A and B?

Theorem

[Weyl] Let

λ1(A) ≥ λ2(A) ≥ · · · ≥ λn(A)

λ1(B) ≥ λ2(B) ≥ · · · ≥ λn(B)

be the eigenvalues of the real symmetric n × n matrices A, B. Then

λj(A) + λ1(B) ≥ λj(A + B) ≥ λj(A) + λn(B).

Corollary

If B is positive semidefinite, λj(A + B) ≥ λj(A).

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 19 / 55

Page 72: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Second Way to Define L(G)

Example

1

2

3 4

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 20 / 55

Page 73: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Second Way to Define L(G)

Example

1

2

3 4

E{1, 2} =

1 −1 0 0−1 1 0 0

0 0 0 00 0 0 0

,

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 21 / 55

Page 74: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Second Way to Define L(G)

Example

1

2

3 4

E{1, 2} =

1 −1 0 0−1 1 0 0

0 0 0 00 0 0 0

, E{1, 3} =

1 0 −1 00 0 0 0−1 0 1 0

0 0 0 0

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 22 / 55

Page 75: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Second Way to Define L(G)

Example

1

2

3 4

E{1, 2} =

1 −1 0 0−1 1 0 0

0 0 0 00 0 0 0

, E{1, 3} =

1 0 −1 00 0 0 0−1 0 1 0

0 0 0 0

E{2, 3} =

0 0 0 00 1 −1 00 −1 1 00 0 0 0

,Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 23 / 55

Page 76: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Second Way to Define L(G)

Example

1

2

3 4

E{1, 2} =

1 −1 0 0−1 1 0 0

0 0 0 00 0 0 0

, E{1, 3} =

1 0 −1 00 0 0 0−1 0 1 0

0 0 0 0

E{2, 3} =

0 0 0 00 1 −1 00 −1 1 00 0 0 0

, E{3, 4} =

0 0 0 00 0 0 00 0 1 −10 0 −1 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 24 / 55

Page 77: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Second Way to Define L(G)

E{1, 2}+ E{1, 3}+ E{2, 3}+ E{3, 4}

=

26641 −1 0 0−1 1 0 0

0 0 0 00 0 0 0

3775 +

26641 0 −1 00 0 0 0−1 0 1 0

0 0 0 0

3775 +

26640 0 0 00 1 −1 00 −1 1 00 0 0 0

3775 +

26640 0 0 00 0 0 00 0 1 −10 0 −1 1

3775

=

26642 −1 −1 0−1 2 −1 0−1 −1 3 −1

0 0 −1 1

3775 = L(G).

FACT. Let G be a graph on n vertices.

For each edge {r , s} ∈ G , let E{r , s} be the n × n matrix defined by

E{r , s}ij =

8><>:1 if i = j = r or i = j = s

−1 if i = r , j = s or i = s, j = r

0 otherwise.

Then L(G) =X

{r,s}∈E(G)

E{r , s}.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 25 / 55

Page 78: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Second Way to Define L(G)

E{1, 2}+ E{1, 3}+ E{2, 3}+ E{3, 4}

=

26641 −1 0 0−1 1 0 0

0 0 0 00 0 0 0

3775 +

26641 0 −1 00 0 0 0−1 0 1 0

0 0 0 0

3775 +

26640 0 0 00 1 −1 00 −1 1 00 0 0 0

3775 +

26640 0 0 00 0 0 00 0 1 −10 0 −1 1

3775

=

26642 −1 −1 0−1 2 −1 0−1 −1 3 −1

0 0 −1 1

3775 = L(G).

FACT. Let G be a graph on n vertices.

For each edge {r , s} ∈ G , let E{r , s} be the n × n matrix defined by

E{r , s}ij =

8><>:1 if i = j = r or i = j = s

−1 if i = r , j = s or i = s, j = r

0 otherwise.

Then L(G) =X

{r,s}∈E(G)

E{r , s}.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 25 / 55

Page 79: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Second Way to Define L(G)

E{1, 2}+ E{1, 3}+ E{2, 3}+ E{3, 4}

=

26641 −1 0 0−1 1 0 0

0 0 0 00 0 0 0

3775 +

26641 0 −1 00 0 0 0−1 0 1 0

0 0 0 0

3775 +

26640 0 0 00 1 −1 00 −1 1 00 0 0 0

3775 +

26640 0 0 00 0 0 00 0 1 −10 0 −1 1

3775

=

26642 −1 −1 0−1 2 −1 0−1 −1 3 −1

0 0 −1 1

3775

= L(G).

FACT. Let G be a graph on n vertices.

For each edge {r , s} ∈ G , let E{r , s} be the n × n matrix defined by

E{r , s}ij =

8><>:1 if i = j = r or i = j = s

−1 if i = r , j = s or i = s, j = r

0 otherwise.

Then L(G) =X

{r,s}∈E(G)

E{r , s}.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 25 / 55

Page 80: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Second Way to Define L(G)

E{1, 2}+ E{1, 3}+ E{2, 3}+ E{3, 4}

=

26641 −1 0 0−1 1 0 0

0 0 0 00 0 0 0

3775 +

26641 0 −1 00 0 0 0−1 0 1 0

0 0 0 0

3775 +

26640 0 0 00 1 −1 00 −1 1 00 0 0 0

3775 +

26640 0 0 00 0 0 00 0 1 −10 0 −1 1

3775

=

26642 −1 −1 0−1 2 −1 0−1 −1 3 −1

0 0 −1 1

3775 = L(G).

FACT. Let G be a graph on n vertices.

For each edge {r , s} ∈ G , let E{r , s} be the n × n matrix defined by

E{r , s}ij =

8><>:1 if i = j = r or i = j = s

−1 if i = r , j = s or i = s, j = r

0 otherwise.

Then L(G) =X

{r,s}∈E(G)

E{r , s}.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 25 / 55

Page 81: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Second Way to Define L(G)

E{1, 2}+ E{1, 3}+ E{2, 3}+ E{3, 4}

=

26641 −1 0 0−1 1 0 0

0 0 0 00 0 0 0

3775 +

26641 0 −1 00 0 0 0−1 0 1 0

0 0 0 0

3775 +

26640 0 0 00 1 −1 00 −1 1 00 0 0 0

3775 +

26640 0 0 00 0 0 00 0 1 −10 0 −1 1

3775

=

26642 −1 −1 0−1 2 −1 0−1 −1 3 −1

0 0 −1 1

3775 = L(G).

FACT. Let G be a graph on n vertices.

For each edge {r , s} ∈ G , let E{r , s} be the n × n matrix defined by

E{r , s}ij =

8><>:1 if i = j = r or i = j = s

−1 if i = r , j = s or i = s, j = r

0 otherwise.

Then L(G) =X

{r,s}∈E(G)

E{r , s}.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 25 / 55

Page 82: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Second Way to Define L(G)

E{1, 2}+ E{1, 3}+ E{2, 3}+ E{3, 4}

=

26641 −1 0 0−1 1 0 0

0 0 0 00 0 0 0

3775 +

26641 0 −1 00 0 0 0−1 0 1 0

0 0 0 0

3775 +

26640 0 0 00 1 −1 00 −1 1 00 0 0 0

3775 +

26640 0 0 00 0 0 00 0 1 −10 0 −1 1

3775

=

26642 −1 −1 0−1 2 −1 0−1 −1 3 −1

0 0 −1 1

3775 = L(G).

FACT. Let G be a graph on n vertices.

For each edge {r , s} ∈ G , let E{r , s} be the n × n matrix defined by

E{r , s}ij =

8><>:1 if i = j = r or i = j = s

−1 if i = r , j = s or i = s, j = r

0 otherwise.

Then L(G) =X

{r,s}∈E(G)

E{r , s}.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 25 / 55

Page 83: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Second Way to Define L(G)

E{1, 2}+ E{1, 3}+ E{2, 3}+ E{3, 4}

=

26641 −1 0 0−1 1 0 0

0 0 0 00 0 0 0

3775 +

26641 0 −1 00 0 0 0−1 0 1 0

0 0 0 0

3775 +

26640 0 0 00 1 −1 00 −1 1 00 0 0 0

3775 +

26640 0 0 00 0 0 00 0 1 −10 0 −1 1

3775

=

26642 −1 −1 0−1 2 −1 0−1 −1 3 −1

0 0 −1 1

3775 = L(G).

FACT. Let G be a graph on n vertices.

For each edge {r , s} ∈ G , let E{r , s} be the n × n matrix defined by

E{r , s}ij =

8><>:1 if i = j = r or i = j = s

−1 if i = r , j = s or i = s, j = r

0 otherwise.

Then L(G) =X

{r,s}∈E(G)

E{r , s}.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 25 / 55

Page 84: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Edge Deletion

Definition

Given a graph G and an edge e = {i , j} of G , G − e is the graph obtainedfrom G by deleting e.

Observation

1. L(G ) = L(G − e) + E{i , j} 2. E{i , j} is positive semidefinite.

Corollary

Let G be a graph, let e be an edge of G, and let

λ1(G ) ≥ λ2(G ) ≥ · · · ≥ λn(G )

λ1(G − e) ≥ λ2(G − e) ≥ · · · ≥ λn(G − e)

be the eigenvalues of L(G ) and L(G − e).

Then λj(G ) ≥ λj(G − e), j = 1, . . . , n.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 26 / 55

Page 85: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Edge Deletion

Definition

Given a graph G and an edge e = {i , j} of G , G − e is the graph obtainedfrom G by deleting e.

Observation

1. L(G ) = L(G − e) + E{i , j}

2. E{i , j} is positive semidefinite.

Corollary

Let G be a graph, let e be an edge of G, and let

λ1(G ) ≥ λ2(G ) ≥ · · · ≥ λn(G )

λ1(G − e) ≥ λ2(G − e) ≥ · · · ≥ λn(G − e)

be the eigenvalues of L(G ) and L(G − e).

Then λj(G ) ≥ λj(G − e), j = 1, . . . , n.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 26 / 55

Page 86: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Edge Deletion

Definition

Given a graph G and an edge e = {i , j} of G , G − e is the graph obtainedfrom G by deleting e.

Observation

1. L(G ) = L(G − e) + E{i , j} 2. E{i , j} is positive semidefinite.

Corollary

Let G be a graph, let e be an edge of G, and let

λ1(G ) ≥ λ2(G ) ≥ · · · ≥ λn(G )

λ1(G − e) ≥ λ2(G − e) ≥ · · · ≥ λn(G − e)

be the eigenvalues of L(G ) and L(G − e).

Then λj(G ) ≥ λj(G − e), j = 1, . . . , n.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 26 / 55

Page 87: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Edge Deletion

Definition

Given a graph G and an edge e = {i , j} of G , G − e is the graph obtainedfrom G by deleting e.

Observation

1. L(G ) = L(G − e) + E{i , j} 2. E{i , j} is positive semidefinite.

Corollary

Let G be a graph, let e be an edge of G, and let

λ1(G ) ≥ λ2(G ) ≥ · · · ≥ λn(G )

λ1(G − e) ≥ λ2(G − e) ≥ · · · ≥ λn(G − e)

be the eigenvalues of L(G ) and L(G − e).

Then λj(G ) ≥ λj(G − e), j = 1, . . . , n.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 26 / 55

Page 88: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Edge Deletion

Definition

Given a graph G and an edge e = {i , j} of G , G − e is the graph obtainedfrom G by deleting e.

Observation

1. L(G ) = L(G − e) + E{i , j} 2. E{i , j} is positive semidefinite.

Corollary

Let G be a graph, let e be an edge of G, and let

λ1(G ) ≥ λ2(G ) ≥ · · · ≥ λn(G )

λ1(G − e) ≥ λ2(G − e) ≥ · · · ≥ λn(G − e)

be the eigenvalues of L(G ) and L(G − e).

Then λj(G ) ≥ λj(G − e), j = 1, . . . , n.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 26 / 55

Page 89: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian Graphs

Observation

If G is a Hamiltonian graph on n vertices, there exist edges e1, e2, . . . , ek

of G such that G − e1 − e2 − · · · − ek = Cn.

Corollary

Let G be a Hamiltonian graph on n vertices, let

λ1(G ) ≥ λ2(G ) ≥ · · · ≥ λn(G ) be the eigenvalues of L(G )

and λ1(Cn) ≥ λ2(Cn) ≥ · · · ≥ λn(Cn) be the eigenvalues of L(Cn).

Then λj(G ) ≥ λj(Cn), j = 1, . . . , n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 27 / 55

Page 90: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian Graphs

Observation

If G is a Hamiltonian graph on n vertices, there exist edges e1, e2, . . . , ek

of G such that G − e1 − e2 − · · · − ek = Cn.

Corollary

Let G be a Hamiltonian graph on n vertices, let

λ1(G ) ≥ λ2(G ) ≥ · · · ≥ λn(G ) be the eigenvalues of L(G )

and λ1(Cn) ≥ λ2(Cn) ≥ · · · ≥ λn(Cn) be the eigenvalues of L(Cn).

Then λj(G ) ≥ λj(Cn), j = 1, . . . , n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 27 / 55

Page 91: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian Graphs

Observation

If G is a Hamiltonian graph on n vertices, there exist edges e1, e2, . . . , ek

of G such that G − e1 − e2 − · · · − ek = Cn.

Corollary

Let G be a Hamiltonian graph on n vertices,

let

λ1(G ) ≥ λ2(G ) ≥ · · · ≥ λn(G ) be the eigenvalues of L(G )

and λ1(Cn) ≥ λ2(Cn) ≥ · · · ≥ λn(Cn) be the eigenvalues of L(Cn).

Then λj(G ) ≥ λj(Cn), j = 1, . . . , n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 27 / 55

Page 92: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian Graphs

Observation

If G is a Hamiltonian graph on n vertices, there exist edges e1, e2, . . . , ek

of G such that G − e1 − e2 − · · · − ek = Cn.

Corollary

Let G be a Hamiltonian graph on n vertices, let

λ1(G ) ≥ λ2(G ) ≥ · · · ≥ λn(G ) be the eigenvalues of L(G )

and λ1(Cn) ≥ λ2(Cn) ≥ · · · ≥ λn(Cn) be the eigenvalues of L(Cn).

Then λj(G ) ≥ λj(Cn), j = 1, . . . , n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 27 / 55

Page 93: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian Graphs

Observation

If G is a Hamiltonian graph on n vertices, there exist edges e1, e2, . . . , ek

of G such that G − e1 − e2 − · · · − ek = Cn.

Corollary

Let G be a Hamiltonian graph on n vertices, let

λ1(G ) ≥ λ2(G ) ≥ · · · ≥ λn(G ) be the eigenvalues of L(G )

and λ1(Cn) ≥ λ2(Cn) ≥ · · · ≥ λn(Cn) be the eigenvalues of L(Cn).

Then λj(G ) ≥ λj(Cn), j = 1, . . . , n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 27 / 55

Page 94: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Hamiltonian Graphs

Observation

If G is a Hamiltonian graph on n vertices, there exist edges e1, e2, . . . , ek

of G such that G − e1 − e2 − · · · − ek = Cn.

Corollary

Let G be a Hamiltonian graph on n vertices, let

λ1(G ) ≥ λ2(G ) ≥ · · · ≥ λn(G ) be the eigenvalues of L(G )

and λ1(Cn) ≥ λ2(Cn) ≥ · · · ≥ λn(Cn) be the eigenvalues of L(Cn).

Then λj(G ) ≥ λj(Cn), j = 1, . . . , n − 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 27 / 55

Page 95: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

The Petersen Graph

1

2

3

9

5

4

106

7

8

L(P) =

3 −1 0 0 −1 −1 0 0 0 0−1 3 −1 0 0 0 −1 0 0 0

0 −1 3 −1 0 0 0 −1 0 00 0 −1 3 −1 0 0 0 −1 0−1 0 0 −1 3 0 0 0 0 −1−1 0 0 0 0 3 0 −1 −1 0

0 −1 0 0 0 0 3 0 −1 −10 0 −1 0 0 −1 0 3 0 −10 0 0 −1 0 −1 −1 0 3 00 0 0 0 −1 0 −1 −1 0 3

Eigenvalues: 5, 5, 5, 5, 2, 2, 2, 2, 2, 0.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 28 / 55

Page 96: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

The Petersen Graph

1

2

3

9

5

4

106

7

8

L(P) =

3 −1 0 0 −1 −1 0 0 0 0−1 3 −1 0 0 0 −1 0 0 0

0 −1 3 −1 0 0 0 −1 0 00 0 −1 3 −1 0 0 0 −1 0−1 0 0 −1 3 0 0 0 0 −1−1 0 0 0 0 3 0 −1 −1 0

0 −1 0 0 0 0 3 0 −1 −10 0 −1 0 0 −1 0 3 0 −10 0 0 −1 0 −1 −1 0 3 00 0 0 0 −1 0 −1 −1 0 3

Eigenvalues: 5, 5, 5, 5, 2, 2, 2, 2, 2, 0.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 28 / 55

Page 97: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

The Petersen Graph

1

2

3

9

5

4

106

7

8

L(P) =

3 −1 0 0 −1 −1 0 0 0 0−1 3 −1 0 0 0 −1 0 0 0

0 −1 3 −1 0 0 0 −1 0 00 0 −1 3 −1 0 0 0 −1 0−1 0 0 −1 3 0 0 0 0 −1−1 0 0 0 0 3 0 −1 −1 0

0 −1 0 0 0 0 3 0 −1 −10 0 −1 0 0 −1 0 3 0 −10 0 0 −1 0 −1 −1 0 3 00 0 0 0 −1 0 −1 −1 0 3

Eigenvalues: 5, 5, 5, 5, 2, 2, 2, 2, 2, 0.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 28 / 55

Page 98: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

L(C10)

Eigenvalues: 2− 2 cos2πj

10, j = 0, . . . , 9.

Ordered:

4, 2− 2 cos4π

5(2), 2− 2 cos

5(2), 2− 2 cos

5(2), 2− 2 cos

π

5(2), 0.

If the Petersen graph is Hamiltonianλ5(P) ≥ λ5(C10)

⇔2 ≥ 2− 2 cos 3π

5⇔

cos 3π5 ≥ 0

3π5

cos3π

5< 0 The Petersen graph is not Hamiltonian.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 29 / 55

Page 99: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

L(C10)

Eigenvalues: 2− 2 cos2πj

10, j = 0, . . . , 9.

Ordered:

4, 2− 2 cos4π

5(2), 2− 2 cos

5(2), 2− 2 cos

5(2), 2− 2 cos

π

5(2), 0.

If the Petersen graph is Hamiltonianλ5(P) ≥ λ5(C10)

⇔2 ≥ 2− 2 cos 3π

5⇔

cos 3π5 ≥ 0

3π5

cos3π

5< 0 The Petersen graph is not Hamiltonian.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 29 / 55

Page 100: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

L(C10)

Eigenvalues: 2− 2 cos2πj

10, j = 0, . . . , 9.

Ordered:

4, 2− 2 cos4π

5(2), 2− 2 cos

5(2), 2− 2 cos

5(2), 2− 2 cos

π

5(2), 0.

If the Petersen graph is Hamiltonianλ5(P) ≥ λ5(C10)

⇔2 ≥ 2− 2 cos 3π

5⇔

cos 3π5 ≥ 0

3π5

cos3π

5< 0 The Petersen graph is not Hamiltonian.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 29 / 55

Page 101: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

L(C10)

Eigenvalues: 2− 2 cos2πj

10, j = 0, . . . , 9.

Ordered:

4, 2− 2 cos4π

5(2), 2− 2 cos

5(2), 2− 2 cos

5(2), 2− 2 cos

π

5(2), 0.

If the Petersen graph is Hamiltonianλ5(P) ≥ λ5(C10)

⇔2 ≥ 2− 2 cos 3π

5

⇔cos 3π

5 ≥ 0

3π5

cos3π

5< 0 The Petersen graph is not Hamiltonian.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 29 / 55

Page 102: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

L(C10)

Eigenvalues: 2− 2 cos2πj

10, j = 0, . . . , 9.

Ordered:

4, 2− 2 cos4π

5(2), 2− 2 cos

5(2), 2− 2 cos

5(2), 2− 2 cos

π

5(2), 0.

If the Petersen graph is Hamiltonianλ5(P) ≥ λ5(C10)

⇔2 ≥ 2− 2 cos 3π

5⇔

cos 3π5 ≥ 0

3π5

cos3π

5< 0 The Petersen graph is not Hamiltonian.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 29 / 55

Page 103: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

L(C10)

Eigenvalues: 2− 2 cos2πj

10, j = 0, . . . , 9.

Ordered:

4, 2− 2 cos4π

5(2), 2− 2 cos

5(2), 2− 2 cos

5(2), 2− 2 cos

π

5(2), 0.

If the Petersen graph is Hamiltonianλ5(P) ≥ λ5(C10)

⇔2 ≥ 2− 2 cos 3π

5⇔

cos 3π5 ≥ 0

3π5

cos3π

5< 0 The Petersen graph is not Hamiltonian.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 29 / 55

Page 104: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

L(C10)

Eigenvalues: 2− 2 cos2πj

10, j = 0, . . . , 9.

Ordered:

4, 2− 2 cos4π

5(2), 2− 2 cos

5(2), 2− 2 cos

5(2), 2− 2 cos

π

5(2), 0.

If the Petersen graph is Hamiltonianλ5(P) ≥ λ5(C10)

⇔2 ≥ 2− 2 cos 3π

5⇔

cos 3π5 ≥ 0

3π5

cos3π

5< 0

The Petersen graph is not Hamiltonian.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 29 / 55

Page 105: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

L(C10)

Eigenvalues: 2− 2 cos2πj

10, j = 0, . . . , 9.

Ordered:

4, 2− 2 cos4π

5(2), 2− 2 cos

5(2), 2− 2 cos

5(2), 2− 2 cos

π

5(2), 0.

If the Petersen graph is Hamiltonianλ5(P) ≥ λ5(C10)

⇔2 ≥ 2− 2 cos 3π

5⇔

cos 3π5 ≥ 0

3π5

cos3π

5< 0 The Petersen graph is not Hamiltonian.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 29 / 55

Page 106: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

√5 > 1

2 cos3π

5=

1−√

5

2

λ5(P) ≥ λ5(C10)⇔

2 ≥ 2− 2 cos 3π5

⇔2 ≥ 2− 1−

√5

2⇔1−√

5

2≥ 0

⇔1 ≥√

5

But√

5 > 1. The Petersen graph is not Hamiltonian

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 30 / 55

Page 107: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

√5 > 1

2 cos3π

5=

1−√

5

2

λ5(P) ≥ λ5(C10)⇔

2 ≥ 2− 2 cos 3π5

⇔2 ≥ 2− 1−

√5

2⇔1−√

5

2≥ 0

⇔1 ≥√

5

But√

5 > 1. The Petersen graph is not Hamiltonian

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 30 / 55

Page 108: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

√5 > 1

2 cos3π

5=

1−√

5

2

λ5(P) ≥ λ5(C10)

⇔2 ≥ 2− 2 cos 3π

5⇔

2 ≥ 2− 1−√

5

2⇔1−√

5

2≥ 0

⇔1 ≥√

5

But√

5 > 1. The Petersen graph is not Hamiltonian

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 30 / 55

Page 109: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

√5 > 1

2 cos3π

5=

1−√

5

2

λ5(P) ≥ λ5(C10)⇔

2 ≥ 2− 2 cos 3π5

⇔2 ≥ 2− 1−

√5

2⇔1−√

5

2≥ 0

⇔1 ≥√

5

But√

5 > 1. The Petersen graph is not Hamiltonian

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 30 / 55

Page 110: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

√5 > 1

2 cos3π

5=

1−√

5

2

λ5(P) ≥ λ5(C10)⇔

2 ≥ 2− 2 cos 3π5

⇔2 ≥ 2− 1−

√5

2

⇔1−√

5

2≥ 0

⇔1 ≥√

5

But√

5 > 1. The Petersen graph is not Hamiltonian

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 30 / 55

Page 111: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

√5 > 1

2 cos3π

5=

1−√

5

2

λ5(P) ≥ λ5(C10)⇔

2 ≥ 2− 2 cos 3π5

⇔2 ≥ 2− 1−

√5

2⇔1−√

5

2≥ 0

⇔1 ≥√

5

But√

5 > 1. The Petersen graph is not Hamiltonian

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 30 / 55

Page 112: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

√5 > 1

2 cos3π

5=

1−√

5

2

λ5(P) ≥ λ5(C10)⇔

2 ≥ 2− 2 cos 3π5

⇔2 ≥ 2− 1−

√5

2⇔1−√

5

2≥ 0

⇔1 ≥√

5

But√

5 > 1. The Petersen graph is not Hamiltonian

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 30 / 55

Page 113: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

√5 > 1

2 cos3π

5=

1−√

5

2

λ5(P) ≥ λ5(C10)⇔

2 ≥ 2− 2 cos 3π5

⇔2 ≥ 2− 1−

√5

2⇔1−√

5

2≥ 0

⇔1 ≥√

5

But√

5 > 1.

The Petersen graph is not Hamiltonian

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 30 / 55

Page 114: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

√5 > 1

2 cos3π

5=

1−√

5

2

λ5(P) ≥ λ5(C10)⇔

2 ≥ 2− 2 cos 3π5

⇔2 ≥ 2− 1−

√5

2⇔1−√

5

2≥ 0

⇔1 ≥√

5

But√

5 > 1. The Petersen graph is not Hamiltonian

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 30 / 55

Page 115: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory and Combinatorial Matrix Theory

Spectral Graph Theory

Matrix information (eigenvalues) −→ Graph information.

Combinatorial Matrix Theory is the Reverse direction

Graph information −→ Matrix information

Symmetric matrices that arise in applications often have a specific pattern.

A graph is the best way to indicate the pattern of zeros and nonzeros insuch a matrix.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 31 / 55

Page 116: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory and Combinatorial Matrix Theory

Spectral Graph Theory

Matrix information (eigenvalues) −→ Graph information.

Combinatorial Matrix Theory is the Reverse direction

Graph information −→ Matrix information

Symmetric matrices that arise in applications often have a specific pattern.

A graph is the best way to indicate the pattern of zeros and nonzeros insuch a matrix.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 31 / 55

Page 117: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory and Combinatorial Matrix Theory

Spectral Graph Theory

Matrix information (eigenvalues) −→ Graph information.

Combinatorial Matrix Theory is the Reverse direction

Graph information −→ Matrix information

Symmetric matrices that arise in applications often have a specific pattern.

A graph is the best way to indicate the pattern of zeros and nonzeros insuch a matrix.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 31 / 55

Page 118: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory and Combinatorial Matrix Theory

Spectral Graph Theory

Matrix information (eigenvalues) −→ Graph information.

Combinatorial Matrix Theory is the Reverse direction

Graph information −→ Matrix information

Symmetric matrices that arise in applications often have a specific pattern.

A graph is the best way to indicate the pattern of zeros and nonzeros insuch a matrix.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 31 / 55

Page 119: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory and Combinatorial Matrix Theory

Spectral Graph Theory

Matrix information (eigenvalues) −→ Graph information.

Combinatorial Matrix Theory is the Reverse direction

Graph information −→ Matrix information

Symmetric matrices that arise in applications often have a specific pattern.

A graph is the best way to indicate the pattern of zeros and nonzeros insuch a matrix.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 31 / 55

Page 120: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Spectral Graph Theory and Combinatorial Matrix Theory

Spectral Graph Theory

Matrix information (eigenvalues) −→ Graph information.

Combinatorial Matrix Theory is the Reverse direction

Graph information −→ Matrix information

Symmetric matrices that arise in applications often have a specific pattern.

A graph is the best way to indicate the pattern of zeros and nonzeros insuch a matrix.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 31 / 55

Page 121: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Symmetric Matrix associated with a Graph

Sn - set of all n × n real symmetric matrices

Given A ∈ Sn, let G (A) be the graph with

vertex set V = {1, 2, . . . , n} and

edge set E = {{i , j}|aij 6= 0}

For any graph G , let S(G ) = {A ∈ Sn |G (A) = G}

1

2

3 4b

a

c

dG A =

d1 a b 0a d2 c 0b c d3 d0 0 d d4

∈ S(G )

paw

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 32 / 55

Page 122: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Symmetric Matrix associated with a Graph

Sn - set of all n × n real symmetric matrices

Given A ∈ Sn, let G (A) be the graph with

vertex set V = {1, 2, . . . , n}

and

edge set E = {{i , j}|aij 6= 0}

For any graph G , let S(G ) = {A ∈ Sn |G (A) = G}

1

2

3 4b

a

c

dG A =

d1 a b 0a d2 c 0b c d3 d0 0 d d4

∈ S(G )

paw

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 32 / 55

Page 123: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Symmetric Matrix associated with a Graph

Sn - set of all n × n real symmetric matrices

Given A ∈ Sn, let G (A) be the graph with

vertex set V = {1, 2, . . . , n} and

edge set E = {{i , j}|aij 6= 0}

For any graph G , let S(G ) = {A ∈ Sn |G (A) = G}

1

2

3 4b

a

c

dG A =

d1 a b 0a d2 c 0b c d3 d0 0 d d4

∈ S(G )

paw

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 32 / 55

Page 124: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Symmetric Matrix associated with a Graph

Sn - set of all n × n real symmetric matrices

Given A ∈ Sn, let G (A) be the graph with

vertex set V = {1, 2, . . . , n} and

edge set E = {{i , j}|aij 6= 0}

For any graph G , let S(G ) = {A ∈ Sn |G (A) = G}

1

2

3 4b

a

c

dG A =

d1 a b 0a d2 c 0b c d3 d0 0 d d4

∈ S(G )

paw

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 32 / 55

Page 125: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Symmetric Matrix associated with a Graph

Sn - set of all n × n real symmetric matrices

Given A ∈ Sn, let G (A) be the graph with

vertex set V = {1, 2, . . . , n} and

edge set E = {{i , j}|aij 6= 0}

For any graph G , let S(G ) = {A ∈ Sn |G (A) = G}

1

2

3 4b

a

c

dG

A =

d1 a b 0a d2 c 0b c d3 d0 0 d d4

∈ S(G )

paw

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 32 / 55

Page 126: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Symmetric Matrix associated with a Graph

Sn - set of all n × n real symmetric matrices

Given A ∈ Sn, let G (A) be the graph with

vertex set V = {1, 2, . . . , n} and

edge set E = {{i , j}|aij 6= 0}

For any graph G , let S(G ) = {A ∈ Sn |G (A) = G}

1

2

3 4b

a

c

dG A =

d1 a b 0a d2 c 0b c d3 d0 0 d d4

∈ S(G )

paw

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 32 / 55

Page 127: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Symmetric Matrix associated with a Graph

Sn - set of all n × n real symmetric matrices

Given A ∈ Sn, let G (A) be the graph with

vertex set V = {1, 2, . . . , n} and

edge set E = {{i , j}|aij 6= 0}

For any graph G , let S(G ) = {A ∈ Sn |G (A) = G}

1

2

3 4b

a

c

dG A =

d1 a b 0a d2 c 0b c d3 d0 0 d d4

∈ S(G )

pawWayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 32 / 55

Page 128: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Questions

What possible eigenvalues can a matrix in S(G ) have?

What possible ranks can a matrix in S(G ) have?

In order to determine all possible ranks, it suffices to determine theminimum possible rank.

Every rank larger than the minimum rank is attainable.

Determining the minimum possible rank is called the minimum rankproblem.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 33 / 55

Page 129: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Questions

What possible eigenvalues can a matrix in S(G ) have?

What possible ranks can a matrix in S(G ) have?

In order to determine all possible ranks, it suffices to determine theminimum possible rank.

Every rank larger than the minimum rank is attainable.

Determining the minimum possible rank is called the minimum rankproblem.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 33 / 55

Page 130: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Questions

What possible eigenvalues can a matrix in S(G ) have?

What possible ranks can a matrix in S(G ) have?

In order to determine all possible ranks, it suffices to determine theminimum possible rank.

Every rank larger than the minimum rank is attainable.

Determining the minimum possible rank is called the minimum rankproblem.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 33 / 55

Page 131: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Questions

What possible eigenvalues can a matrix in S(G ) have?

What possible ranks can a matrix in S(G ) have?

In order to determine all possible ranks, it suffices to determine theminimum possible rank.

Every rank larger than the minimum rank is attainable.

Determining the minimum possible rank is called the minimum rankproblem.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 33 / 55

Page 132: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Questions

What possible eigenvalues can a matrix in S(G ) have?

What possible ranks can a matrix in S(G ) have?

In order to determine all possible ranks, it suffices to determine theminimum possible rank.

Every rank larger than the minimum rank is attainable.

Determining the minimum possible rank is called the minimum rankproblem.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 33 / 55

Page 133: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank Problem

Let mr(G ) be the minimum rank over all matrices in S(G ).

Let M(G ) to be the maximum nullity over all matrices in S(G ).

Since mr(G ) + M(G ) = n, computing the minimum rank and themaximum nullity are equivalent problems.

Computing M(G ) or mr(G ) for a general graph is hard.

Easy for n < 6.

mr(paw) = 2

d1 a b 0a d2 c 0b c d3 d0 0 d d4

1 1 1 01 1 1 01 1 2 10 0 1 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 34 / 55

Page 134: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank Problem

Let mr(G ) be the minimum rank over all matrices in S(G ).

Let M(G ) to be the maximum nullity over all matrices in S(G ).

Since mr(G ) + M(G ) = n, computing the minimum rank and themaximum nullity are equivalent problems.

Computing M(G ) or mr(G ) for a general graph is hard.

Easy for n < 6.

mr(paw) = 2

d1 a b 0a d2 c 0b c d3 d0 0 d d4

1 1 1 01 1 1 01 1 2 10 0 1 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 34 / 55

Page 135: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank Problem

Let mr(G ) be the minimum rank over all matrices in S(G ).

Let M(G ) to be the maximum nullity over all matrices in S(G ).

Since mr(G ) + M(G ) = n, computing the minimum rank and themaximum nullity are equivalent problems.

Computing M(G ) or mr(G ) for a general graph is hard.

Easy for n < 6.

mr(paw) = 2

d1 a b 0a d2 c 0b c d3 d0 0 d d4

1 1 1 01 1 1 01 1 2 10 0 1 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 34 / 55

Page 136: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank Problem

Let mr(G ) be the minimum rank over all matrices in S(G ).

Let M(G ) to be the maximum nullity over all matrices in S(G ).

Since mr(G ) + M(G ) = n, computing the minimum rank and themaximum nullity are equivalent problems.

Computing M(G ) or mr(G ) for a general graph is hard.

Easy for n < 6.

mr(paw) = 2

d1 a b 0a d2 c 0b c d3 d0 0 d d4

1 1 1 01 1 1 01 1 2 10 0 1 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 34 / 55

Page 137: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank Problem

Let mr(G ) be the minimum rank over all matrices in S(G ).

Let M(G ) to be the maximum nullity over all matrices in S(G ).

Since mr(G ) + M(G ) = n, computing the minimum rank and themaximum nullity are equivalent problems.

Computing M(G ) or mr(G ) for a general graph is hard.

Easy for n < 6.

mr(paw) = 2

d1 a b 0a d2 c 0b c d3 d0 0 d d4

1 1 1 01 1 1 01 1 2 10 0 1 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 34 / 55

Page 138: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank Problem

Let mr(G ) be the minimum rank over all matrices in S(G ).

Let M(G ) to be the maximum nullity over all matrices in S(G ).

Since mr(G ) + M(G ) = n, computing the minimum rank and themaximum nullity are equivalent problems.

Computing M(G ) or mr(G ) for a general graph is hard.

Easy for n < 6.

mr(paw) = 2

d1 a b 0a d2 c 0b c d3 d0 0 d d4

1 1 1 01 1 1 01 1 2 10 0 1 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 34 / 55

Page 139: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank Problem

Let mr(G ) be the minimum rank over all matrices in S(G ).

Let M(G ) to be the maximum nullity over all matrices in S(G ).

Since mr(G ) + M(G ) = n, computing the minimum rank and themaximum nullity are equivalent problems.

Computing M(G ) or mr(G ) for a general graph is hard.

Easy for n < 6.

mr(paw) = 2

d1 a b 0a d2 c 0b c d3 d0 0 d d4

1 1 1 01 1 1 01 1 2 10 0 1 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 34 / 55

Page 140: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank Problem

Let mr(G ) be the minimum rank over all matrices in S(G ).

Let M(G ) to be the maximum nullity over all matrices in S(G ).

Since mr(G ) + M(G ) = n, computing the minimum rank and themaximum nullity are equivalent problems.

Computing M(G ) or mr(G ) for a general graph is hard.

Easy for n < 6.

mr(paw) = 2

d1 a b 0a d2 c 0b c d3 d0 0 d d4

1 1 1 01 1 1 01 1 2 10 0 1 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 34 / 55

Page 141: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Extreme examples

1. Complete graph Kn

K K K K1 2 3 4

Jn =

1 1 . . . 11 1 . . . 1...

.... . .

...1 1 . . . 1

∈ S(Kn) =⇒ mr(G ) = 1, n ≥ 2

Kn is the only connected graph with minimum rank 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 35 / 55

Page 142: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Extreme examples

1. Complete graph Kn

K K K K1 2 3 4

Jn =

1 1 . . . 11 1 . . . 1...

.... . .

...1 1 . . . 1

∈ S(Kn) =⇒ mr(G ) = 1, n ≥ 2

Kn is the only connected graph with minimum rank 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 35 / 55

Page 143: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Extreme examples

1. Complete graph Kn

K K K K1 2 3 4

Jn =

1 1 . . . 11 1 . . . 1...

.... . .

...1 1 . . . 1

∈ S(Kn)

=⇒ mr(G ) = 1, n ≥ 2

Kn is the only connected graph with minimum rank 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 35 / 55

Page 144: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Extreme examples

1. Complete graph Kn

K K K K1 2 3 4

Jn =

1 1 . . . 11 1 . . . 1...

.... . .

...1 1 . . . 1

∈ S(Kn) =⇒ mr(G ) = 1, n ≥ 2

Kn is the only connected graph with minimum rank 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 35 / 55

Page 145: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Extreme examples

1. Complete graph Kn

K K K K1 2 3 4

Jn =

1 1 . . . 11 1 . . . 1...

.... . .

...1 1 . . . 1

∈ S(Kn) =⇒ mr(G ) = 1, n ≥ 2

Kn is the only connected graph with minimum rank 1.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 35 / 55

Page 146: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Extreme examples

2. Pn . . .

Any A ∈ S(Pn) has the form A =

a1 b1

b1 a2 b2

b2 a3. . .

. . .. . . bn−1

bn−1 an

, bi 6= 0

Deleting the first column and last row gives an invertible lower triangularmatrix =⇒ rank A ≥ n − 1 =⇒ mr(Pn) ≥ n − 1

L(Pn) =

1 −1−1 2 −1

−1 2. . .

. . .. . . −1−1 1

∈ S(Pn)

Rows sum to 0, so L is singular =⇒ rank L = n − 1 =⇒ mr(Pn) = n − 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 36 / 55

Page 147: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Extreme examples

2. Pn . . .

Any A ∈ S(Pn) has the form A =

a1 b1

b1 a2 b2

b2 a3. . .

. . .. . . bn−1

bn−1 an

, bi 6= 0

Deleting the first column and last row gives an invertible lower triangularmatrix =⇒ rank A ≥ n − 1 =⇒ mr(Pn) ≥ n − 1

L(Pn) =

1 −1−1 2 −1

−1 2. . .

. . .. . . −1−1 1

∈ S(Pn)

Rows sum to 0, so L is singular =⇒ rank L = n − 1 =⇒ mr(Pn) = n − 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 36 / 55

Page 148: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Extreme examples

2. Pn . . .

Any A ∈ S(Pn) has the form A =

a1 b1

b1 a2 b2

b2 a3. . .

. . .. . . bn−1

bn−1 an

, bi 6= 0

Deleting the first column and last row gives an invertible lower triangularmatrix

=⇒ rank A ≥ n − 1 =⇒ mr(Pn) ≥ n − 1

L(Pn) =

1 −1−1 2 −1

−1 2. . .

. . .. . . −1−1 1

∈ S(Pn)

Rows sum to 0, so L is singular =⇒ rank L = n − 1 =⇒ mr(Pn) = n − 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 36 / 55

Page 149: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Extreme examples

2. Pn . . .

Any A ∈ S(Pn) has the form A =

a1 b1

b1 a2 b2

b2 a3. . .

. . .. . . bn−1

bn−1 an

, bi 6= 0

Deleting the first column and last row gives an invertible lower triangularmatrix =⇒ rank A ≥ n − 1

=⇒ mr(Pn) ≥ n − 1

L(Pn) =

1 −1−1 2 −1

−1 2. . .

. . .. . . −1−1 1

∈ S(Pn)

Rows sum to 0, so L is singular =⇒ rank L = n − 1 =⇒ mr(Pn) = n − 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 36 / 55

Page 150: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Extreme examples

2. Pn . . .

Any A ∈ S(Pn) has the form A =

a1 b1

b1 a2 b2

b2 a3. . .

. . .. . . bn−1

bn−1 an

, bi 6= 0

Deleting the first column and last row gives an invertible lower triangularmatrix =⇒ rank A ≥ n − 1 =⇒ mr(Pn) ≥ n − 1

L(Pn) =

1 −1−1 2 −1

−1 2. . .

. . .. . . −1−1 1

∈ S(Pn)

Rows sum to 0, so L is singular =⇒ rank L = n − 1 =⇒ mr(Pn) = n − 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 36 / 55

Page 151: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Extreme examples

2. Pn . . .

Any A ∈ S(Pn) has the form A =

a1 b1

b1 a2 b2

b2 a3. . .

. . .. . . bn−1

bn−1 an

, bi 6= 0

Deleting the first column and last row gives an invertible lower triangularmatrix =⇒ rank A ≥ n − 1 =⇒ mr(Pn) ≥ n − 1

L(Pn) =

1 −1−1 2 −1

−1 2. . .

. . .. . . −1−1 1

∈ S(Pn)

Rows sum to 0, so L is singular =⇒ rank L = n − 1 =⇒ mr(Pn) = n − 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 36 / 55

Page 152: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Extreme examples

2. Pn . . .

Any A ∈ S(Pn) has the form A =

a1 b1

b1 a2 b2

b2 a3. . .

. . .. . . bn−1

bn−1 an

, bi 6= 0

Deleting the first column and last row gives an invertible lower triangularmatrix =⇒ rank A ≥ n − 1 =⇒ mr(Pn) ≥ n − 1

L(Pn) =

1 −1−1 2 −1

−1 2. . .

. . .. . . −1−1 1

∈ S(Pn)

Rows sum to 0, so L is singular

=⇒ rank L = n − 1 =⇒ mr(Pn) = n − 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 36 / 55

Page 153: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Extreme examples

2. Pn . . .

Any A ∈ S(Pn) has the form A =

a1 b1

b1 a2 b2

b2 a3. . .

. . .. . . bn−1

bn−1 an

, bi 6= 0

Deleting the first column and last row gives an invertible lower triangularmatrix =⇒ rank A ≥ n − 1 =⇒ mr(Pn) ≥ n − 1

L(Pn) =

1 −1−1 2 −1

−1 2. . .

. . .. . . −1−1 1

∈ S(Pn)

Rows sum to 0, so L is singular =⇒ rank L = n − 1

=⇒ mr(Pn) = n − 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 36 / 55

Page 154: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Extreme examples

2. Pn . . .

Any A ∈ S(Pn) has the form A =

a1 b1

b1 a2 b2

b2 a3. . .

. . .. . . bn−1

bn−1 an

, bi 6= 0

Deleting the first column and last row gives an invertible lower triangularmatrix =⇒ rank A ≥ n − 1 =⇒ mr(Pn) ≥ n − 1

L(Pn) =

1 −1−1 2 −1

−1 2. . .

. . .. . . −1−1 1

∈ S(Pn)

Rows sum to 0, so L is singular =⇒ rank L = n − 1 =⇒ mr(Pn) = n − 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 36 / 55

Page 155: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced subgraphs

G v G − v P4

Definition

An induced subgraph H of a graph G is a graph that can be obtained fromG by successively deleting vertices.

Example

−→ −→

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 37 / 55

Page 156: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced subgraphs

G v

G − v P4

Definition

An induced subgraph H of a graph G is a graph that can be obtained fromG by successively deleting vertices.

Example

−→ −→

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 37 / 55

Page 157: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced subgraphs

G v G − v

P4

Definition

An induced subgraph H of a graph G is a graph that can be obtained fromG by successively deleting vertices.

Example

−→ −→

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 37 / 55

Page 158: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced subgraphs

G v G − v P4

Definition

An induced subgraph H of a graph G is a graph that can be obtained fromG by successively deleting vertices.

Example

−→ −→

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 37 / 55

Page 159: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced subgraphs

G v G − v P4

Definition

An induced subgraph H of a graph G is a graph that can be obtained fromG by successively deleting vertices.

Example

−→ −→

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 37 / 55

Page 160: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced subgraphs

G v G − v P4

Definition

An induced subgraph H of a graph G is a graph that can be obtained fromG by successively deleting vertices.

Example

−→ −→

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 37 / 55

Page 161: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced subgraphs

G v G − v P4

Definition

An induced subgraph H of a graph G is a graph that can be obtained fromG by successively deleting vertices.

Example

−→

−→

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 37 / 55

Page 162: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced subgraphs

G v G − v P4

Definition

An induced subgraph H of a graph G is a graph that can be obtained fromG by successively deleting vertices.

Example

−→ −→

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 37 / 55

Page 163: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Trees

Definition

A graph T is a tree if

T is connected

T contains no cycle

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 38 / 55

Page 164: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Trees

Definition

A graph T is a tree if

T is connected

T contains no cycle

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 38 / 55

Page 165: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Path Cover Number

Definition

A path cover of a graph G is a set of disjoint, induced paths in G thatcover all the vertices.

The path cover number P(G ) is the minimum number of paths in a pathcover of G .

Theorem (Duarte-Johnson)

If T is a tree, then M(T ) = P(T ).

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 39 / 55

Page 166: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Path Cover Number

Definition

A path cover of a graph G is a set of disjoint, induced paths in G thatcover all the vertices.

The path cover number P(G ) is the minimum number of paths in a pathcover of G .

Theorem (Duarte-Johnson)

If T is a tree, then M(T ) = P(T ).

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 39 / 55

Page 167: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Path Cover Number

Definition

A path cover of a graph G is a set of disjoint, induced paths in G thatcover all the vertices.

The path cover number P(G ) is the minimum number of paths in a pathcover of G .

Theorem (Duarte-Johnson)

If T is a tree, then M(T ) = P(T ).

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 39 / 55

Page 168: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Path Cover Number

Definition

A path cover of a graph G is a set of disjoint, induced paths in G thatcover all the vertices.

The path cover number P(G ) is the minimum number of paths in a pathcover of G .

Theorem (Duarte-Johnson)

If T is a tree, then M(T ) = P(T ).

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 39 / 55

Page 169: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Unicyclic Graphs

Definition

A connected graph is unicyclic if it contains exactly one cycle.

1

3 4

2

Theorem (Barioli,Fallat,Hogben)

If G is unicyclic, then either M(G ) = P(G ) or else M(G ) = P(G )− 1.

Furthermore, the circumstances in which each occur are characterized.

Best reference: p. 49 of John Sinkovic’s Ph.D. dissertation

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 40 / 55

Page 170: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Unicyclic Graphs

Definition

A connected graph is unicyclic if it contains exactly one cycle.

1

3 4

2

Theorem (Barioli,Fallat,Hogben)

If G is unicyclic, then either M(G ) = P(G ) or else M(G ) = P(G )− 1.

Furthermore, the circumstances in which each occur are characterized.

Best reference: p. 49 of John Sinkovic’s Ph.D. dissertation

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 40 / 55

Page 171: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Unicyclic Graphs

Definition

A connected graph is unicyclic if it contains exactly one cycle.

1

3 4

2

Theorem (Barioli,Fallat,Hogben)

If G is unicyclic, then either M(G ) = P(G ) or else M(G ) = P(G )− 1.

Furthermore, the circumstances in which each occur are characterized.

Best reference: p. 49 of John Sinkovic’s Ph.D. dissertation

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 40 / 55

Page 172: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Unicyclic Graphs

Definition

A connected graph is unicyclic if it contains exactly one cycle.

1

3 4

2

Theorem (Barioli,Fallat,Hogben)

If G is unicyclic, then either M(G ) = P(G ) or else M(G ) = P(G )− 1.

Furthermore, the circumstances in which each occur are characterized.

Best reference: p. 49 of John Sinkovic’s Ph.D. dissertation

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 40 / 55

Page 173: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Unicyclic Graphs

Definition

A connected graph is unicyclic if it contains exactly one cycle.

1

3 4

2

Theorem (Barioli,Fallat,Hogben)

If G is unicyclic, then either M(G ) = P(G ) or else M(G ) = P(G )− 1.

Furthermore, the circumstances in which each occur are characterized.

Best reference: p. 49 of John Sinkovic’s Ph.D. dissertation

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 40 / 55

Page 174: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Outerplanar Graphs

Definition

A graph is outerplanar if it has an embedding in the plane with everyvertex adjacent to the unbounded face.

Example

Theorem (John Sinkovic)

If G is outerplanar, then M(G ) ≤ P(G ).

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 41 / 55

Page 175: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Outerplanar Graphs

Definition

A graph is outerplanar if it has an embedding in the plane with everyvertex adjacent to the unbounded face.

Example

Theorem (John Sinkovic)

If G is outerplanar, then M(G ) ≤ P(G ).

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 41 / 55

Page 176: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Outerplanar Graphs

Definition

A graph is outerplanar if it has an embedding in the plane with everyvertex adjacent to the unbounded face.

Example

Theorem (John Sinkovic)

If G is outerplanar, then M(G ) ≤ P(G ).

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 41 / 55

Page 177: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced Subgraphs and Minimum Rank

1

2 3

4 5

1

2

4 5

house house delete vertex 3 (P4)

A =

266664d1 u v 0 0u d2 w x 0v w d3 0 y0 x 0 d4 z0 0 y z d5

377775 ∈ S(house) A(3) =

266664d1 u v 0 0u d2 w x 0v w d3 0 y0 x 0 d4 z0 0 y z d5

377775

=

2664d1 u 0 0u d2 x 00 x d4 z0 0 z d5

3775 ∈ S(P4)

rank A ≥ rank A(3)⇒ mr(house) ≥ mr(P4) = 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 42 / 55

Page 178: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced Subgraphs and Minimum Rank

1

2 3

4 5

1

2

4 5

house house delete vertex 3 (P4)

A =

266664d1 u v 0 0u d2 w x 0v w d3 0 y0 x 0 d4 z0 0 y z d5

377775 ∈ S(house)

A(3) =

266664d1 u v 0 0u d2 w x 0v w d3 0 y0 x 0 d4 z0 0 y z d5

377775

=

2664d1 u 0 0u d2 x 00 x d4 z0 0 z d5

3775 ∈ S(P4)

rank A ≥ rank A(3)⇒ mr(house) ≥ mr(P4) = 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 42 / 55

Page 179: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced Subgraphs and Minimum Rank

1

2 3

4 5

1

2

4 5

house house delete vertex 3 (P4)

A =

266664d1 u v 0 0u d2 w x 0v w d3 0 y0 x 0 d4 z0 0 y z d5

377775 ∈ S(house) A(3) =

266664d1 u v 0 0u d2 w x 0v w d3 0 y0 x 0 d4 z0 0 y z d5

377775

=

2664d1 u 0 0u d2 x 00 x d4 z0 0 z d5

3775 ∈ S(P4)

rank A ≥ rank A(3)⇒ mr(house) ≥ mr(P4) = 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 42 / 55

Page 180: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced Subgraphs and Minimum Rank

1

2 3

4 5

1

2

4 5

house house delete vertex 3 (P4)

A =

266664d1 u v 0 0u d2 w x 0v w d3 0 y0 x 0 d4 z0 0 y z d5

377775 ∈ S(house) A(3) =

266664d1 u v 0 0u d2 w x 0v w d3 0 y0 x 0 d4 z0 0 y z d5

377775

=

2664d1 u 0 0u d2 x 00 x d4 z0 0 z d5

3775 ∈ S(P4)

rank A ≥ rank A(3)⇒ mr(house) ≥ mr(P4) = 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 42 / 55

Page 181: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced Subgraphs and Minimum Rank

1

2 3

4 5

1

2

4 5

house house delete vertex 3 (P4)

A =

266664d1 u v 0 0u d2 w x 0v w d3 0 y0 x 0 d4 z0 0 y z d5

377775 ∈ S(house) A(3) =

266664d1 u v 0 0u d2 w x 0v w d3 0 y0 x 0 d4 z0 0 y z d5

377775

=

2664d1 u 0 0u d2 x 00 x d4 z0 0 z d5

3775 ∈ S(P4)

rank A ≥ rank A(3)

⇒ mr(house) ≥ mr(P4) = 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 42 / 55

Page 182: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced Subgraphs and Minimum Rank

1

2 3

4 5

1

2

4 5

house house delete vertex 3 (P4)

A =

266664d1 u v 0 0u d2 w x 0v w d3 0 y0 x 0 d4 z0 0 y z d5

377775 ∈ S(house) A(3) =

266664d1 u v 0 0u d2 w x 0v w d3 0 y0 x 0 d4 z0 0 y z d5

377775

=

2664d1 u 0 0u d2 x 00 x d4 z0 0 z d5

3775 ∈ S(P4)

rank A ≥ rank A(3)⇒ mr(house) ≥ mr(P4)

= 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 42 / 55

Page 183: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced Subgraphs and Minimum Rank

1

2 3

4 5

1

2

4 5

house house delete vertex 3 (P4)

A =

266664d1 u v 0 0u d2 w x 0v w d3 0 y0 x 0 d4 z0 0 y z d5

377775 ∈ S(house) A(3) =

266664d1 u v 0 0u d2 w x 0v w d3 0 y0 x 0 d4 z0 0 y z d5

377775

=

2664d1 u 0 0u d2 x 00 x d4 z0 0 z d5

3775 ∈ S(P4)

rank A ≥ rank A(3)⇒ mr(house) ≥ mr(P4) = 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 42 / 55

Page 184: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced Subgraphs and Minimum Rank

Observation

Let G be a graph and v a vertex of G . Then mr(G ) ≥ mr(G − v).

Observation

If H is an induced subgraph of G , then mr(G ) ≥ mr(H).

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 43 / 55

Page 185: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Induced Subgraphs and Minimum Rank

Observation

Let G be a graph and v a vertex of G . Then mr(G ) ≥ mr(G − v).

Observation

If H is an induced subgraph of G , then mr(G ) ≥ mr(H).

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 43 / 55

Page 186: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum rank 2 graphs

Problem

Which graphs G have minimum rank ≤ 2?

Work with Hein van der Holst and Raphael Loewy.

Necessary condition. If mr(H) ≥ 3, then H cannot be induced in G .

Definition

If mr(H) ≥ 3, but mr(H − v) ≤ 2 for any vertex v of H, then H is aminimal forbidden subgraph for the minimum rank 2 problem.

Example: P4 mr(P4) = 3

P4 − v 2 possibilities up to isomorphism

P3 P2 ∪ K1

mr(P3) = 2 mr(P2 ∪ K1) = 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 44 / 55

Page 187: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum rank 2 graphs

Problem

Which graphs G have minimum rank ≤ 2?

Work with Hein van der Holst and Raphael Loewy.

Necessary condition. If mr(H) ≥ 3, then H cannot be induced in G .

Definition

If mr(H) ≥ 3, but mr(H − v) ≤ 2 for any vertex v of H, then H is aminimal forbidden subgraph for the minimum rank 2 problem.

Example: P4 mr(P4) = 3

P4 − v 2 possibilities up to isomorphism

P3 P2 ∪ K1

mr(P3) = 2 mr(P2 ∪ K1) = 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 44 / 55

Page 188: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum rank 2 graphs

Problem

Which graphs G have minimum rank ≤ 2?

Work with Hein van der Holst and Raphael Loewy.

Necessary condition. If mr(H) ≥ 3, then H cannot be induced in G .

Definition

If mr(H) ≥ 3, but mr(H − v) ≤ 2 for any vertex v of H, then H is aminimal forbidden subgraph for the minimum rank 2 problem.

Example: P4 mr(P4) = 3

P4 − v 2 possibilities up to isomorphism

P3 P2 ∪ K1

mr(P3) = 2 mr(P2 ∪ K1) = 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 44 / 55

Page 189: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum rank 2 graphs

Problem

Which graphs G have minimum rank ≤ 2?

Work with Hein van der Holst and Raphael Loewy.

Necessary condition. If mr(H) ≥ 3, then H cannot be induced in G .

Definition

If mr(H) ≥ 3, but mr(H − v) ≤ 2 for any vertex v of H, then H is aminimal forbidden subgraph for the minimum rank 2 problem.

Example: P4 mr(P4) = 3

P4 − v 2 possibilities up to isomorphism

P3 P2 ∪ K1

mr(P3) = 2 mr(P2 ∪ K1) = 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 44 / 55

Page 190: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum rank 2 graphs

Problem

Which graphs G have minimum rank ≤ 2?

Work with Hein van der Holst and Raphael Loewy.

Necessary condition. If mr(H) ≥ 3, then H cannot be induced in G .

Definition

If mr(H) ≥ 3, but mr(H − v) ≤ 2 for any vertex v of H, then H is aminimal forbidden subgraph for the minimum rank 2 problem.

Example: P4

mr(P4) = 3

P4 − v 2 possibilities up to isomorphism

P3 P2 ∪ K1

mr(P3) = 2 mr(P2 ∪ K1) = 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 44 / 55

Page 191: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum rank 2 graphs

Problem

Which graphs G have minimum rank ≤ 2?

Work with Hein van der Holst and Raphael Loewy.

Necessary condition. If mr(H) ≥ 3, then H cannot be induced in G .

Definition

If mr(H) ≥ 3, but mr(H − v) ≤ 2 for any vertex v of H, then H is aminimal forbidden subgraph for the minimum rank 2 problem.

Example: P4 mr(P4) = 3

P4 − v 2 possibilities up to isomorphism

P3 P2 ∪ K1

mr(P3) = 2 mr(P2 ∪ K1) = 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 44 / 55

Page 192: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum rank 2 graphs

Problem

Which graphs G have minimum rank ≤ 2?

Work with Hein van der Holst and Raphael Loewy.

Necessary condition. If mr(H) ≥ 3, then H cannot be induced in G .

Definition

If mr(H) ≥ 3, but mr(H − v) ≤ 2 for any vertex v of H, then H is aminimal forbidden subgraph for the minimum rank 2 problem.

Example: P4 mr(P4) = 3

P4 − v 2 possibilities up to isomorphism

P3 P2 ∪ K1

mr(P3) = 2 mr(P2 ∪ K1) = 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 44 / 55

Page 193: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum rank 2 graphs

Problem

Which graphs G have minimum rank ≤ 2?

Work with Hein van der Holst and Raphael Loewy.

Necessary condition. If mr(H) ≥ 3, then H cannot be induced in G .

Definition

If mr(H) ≥ 3, but mr(H − v) ≤ 2 for any vertex v of H, then H is aminimal forbidden subgraph for the minimum rank 2 problem.

Example: P4 mr(P4) = 3

P4 − v 2 possibilities up to isomorphism

P3 P2 ∪ K1

mr(P3) = 2 mr(P2 ∪ K1) = 1

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 44 / 55

Page 194: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum rank 2 graphs

Problem

Which graphs G have minimum rank ≤ 2?

Work with Hein van der Holst and Raphael Loewy.

Necessary condition. If mr(H) ≥ 3, then H cannot be induced in G .

Definition

If mr(H) ≥ 3, but mr(H − v) ≤ 2 for any vertex v of H, then H is aminimal forbidden subgraph for the minimum rank 2 problem.

Example: P4 mr(P4) = 3

P4 − v 2 possibilities up to isomorphism

P3 P2 ∪ K1

mr(P3) = 2 mr(P2 ∪ K1) = 1Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 44 / 55

Page 195: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimal Forbidden Subgraphs

1 2

3

4 5

A =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

folding stool

B = A({2, 5}, {1, 4}) =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

=

a b 0c d3 y0 x 0

which is always invertible (det = −axy)

=⇒ rank B = 3 =⇒ rank A ≥ 3 =⇒ mr(folding stool) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 45 / 55

Page 196: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimal Forbidden Subgraphs

1 2

3

4 5

A =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

folding stool

B = A({2, 5}, {1, 4}) =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

=

a b 0c d3 y0 x 0

which is always invertible (det = −axy)

=⇒ rank B = 3 =⇒ rank A ≥ 3 =⇒ mr(folding stool) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 45 / 55

Page 197: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimal Forbidden Subgraphs

1 2

3

4 5

A =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

folding stool

B = A({2, 5}, {1, 4}) =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

=

a b 0c d3 y0 x 0

which is always invertible (det = −axy)

=⇒ rank B = 3 =⇒ rank A ≥ 3 =⇒ mr(folding stool) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 45 / 55

Page 198: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimal Forbidden Subgraphs

1 2

3

4 5

A =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

folding stool

B = A({2, 5}, {1, 4}) =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

=

a b 0c d3 y0 x 0

which is always invertible (det = −axy)

=⇒ rank B = 3 =⇒ rank A ≥ 3 =⇒ mr(folding stool) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 45 / 55

Page 199: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimal Forbidden Subgraphs

1 2

3

4 5

A =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

folding stool

B = A({2, 5}, {1, 4}) =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

=

a b 0c d3 y0 x 0

which is always invertible (det = −axy)

=⇒ rank B = 3 =⇒ rank A ≥ 3 =⇒ mr(folding stool) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 45 / 55

Page 200: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimal Forbidden Subgraphs

1 2

3

4 5

A =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

folding stool

B = A({2, 5}, {1, 4}) =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

=

a b 0c d3 y0 x 0

which is always invertible (det = −axy)

=⇒ rank B = 3 =⇒ rank A ≥ 3 =⇒ mr(folding stool) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 45 / 55

Page 201: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimal Forbidden Subgraphs

1 2

3

4 5

A =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

folding stool

B = A({2, 5}, {1, 4}) =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

=

a b 0c d3 y0 x 0

which is always invertible (det = −axy)

=⇒ rank B = 3 =⇒ rank A ≥ 3 =⇒ mr(folding stool) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 45 / 55

Page 202: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimal Forbidden Subgraphs

1 2

3

4 5

A =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

folding stool

B = A({2, 5}, {1, 4}) =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

=

a b 0c d3 y0 x 0

which is always invertible (det = −axy)

=⇒ rank B = 3

=⇒ rank A ≥ 3 =⇒ mr(folding stool) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 45 / 55

Page 203: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimal Forbidden Subgraphs

1 2

3

4 5

A =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

folding stool

B = A({2, 5}, {1, 4}) =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

=

a b 0c d3 y0 x 0

which is always invertible (det = −axy)

=⇒ rank B = 3 =⇒ rank A ≥ 3

=⇒ mr(folding stool) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 45 / 55

Page 204: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimal Forbidden Subgraphs

1 2

3

4 5

A =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

folding stool

B = A({2, 5}, {1, 4}) =

d1 a b 0 0a d2 c 0 0b c d3 x y0 0 x d4 00 0 y 0 d5

=

a b 0c d3 y0 x 0

which is always invertible (det = −axy)

=⇒ rank B = 3 =⇒ rank A ≥ 3 =⇒ mr(folding stool) ≥ 3Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 45 / 55

Page 205: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimal Forbidden Subgraphs

1 2

3

4 5

mr(folding stool) ≥ 3

Also mr(folding stool − v) ≤ 2 for each vertex v .

folding stool is a minimal forbidden subgraph.

A similar argument shows that

is a minimal forbidden subgraph.

dart

For a month or more in 2001 Raphi Loewy and I thought these were all.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 46 / 55

Page 206: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimal Forbidden Subgraphs

1 2

3

4 5

mr(folding stool) ≥ 3

Also mr(folding stool − v) ≤ 2 for each vertex v .

folding stool is a minimal forbidden subgraph.

A similar argument shows that

is a minimal forbidden subgraph.

dart

For a month or more in 2001 Raphi Loewy and I thought these were all.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 46 / 55

Page 207: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimal Forbidden Subgraphs

1 2

3

4 5

mr(folding stool) ≥ 3

Also mr(folding stool − v) ≤ 2 for each vertex v .

folding stool is a minimal forbidden subgraph.

A similar argument shows that

is a minimal forbidden subgraph.

dart

For a month or more in 2001 Raphi Loewy and I thought these were all.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 46 / 55

Page 208: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimal Forbidden Subgraphs

1 2

3

4 5

mr(folding stool) ≥ 3

Also mr(folding stool − v) ≤ 2 for each vertex v .

folding stool is a minimal forbidden subgraph.

A similar argument shows that

is a minimal forbidden subgraph.

dart

For a month or more in 2001 Raphi Loewy and I thought these were all.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 46 / 55

Page 209: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimal Forbidden Subgraphs

1 2

3

4 5

mr(folding stool) ≥ 3

Also mr(folding stool − v) ≤ 2 for each vertex v .

folding stool is a minimal forbidden subgraph.

A similar argument shows that

is a minimal forbidden subgraph.

dart

For a month or more in 2001 Raphi Loewy and I thought these were all.Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 46 / 55

Page 210: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 d3 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 d6 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 d8 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

What is the smallest possible rank of A?

Case 1a d1 d2 d3 6= 0 rank A ≥ 31b d4 d5 d6 6= 0 rank A ≥ 31c d7 d8 d9 6= 0 rank A ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 47 / 55

Page 211: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 d3 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 d6 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 d8 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

What is the smallest possible rank of A?

Case 1a d1 d2 d3 6= 0 rank A ≥ 31b d4 d5 d6 6= 0 rank A ≥ 31c d7 d8 d9 6= 0 rank A ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 47 / 55

Page 212: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 d3 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 d6 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 d8 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

What is the smallest possible rank of A?

Case 1a d1 d2 d3 6= 0 rank A ≥ 31b d4 d5 d6 6= 0 rank A ≥ 31c d7 d8 d9 6= 0 rank A ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 47 / 55

Page 213: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 d3 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 d6 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 d8 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

What is the smallest possible rank of A?

Case 1a d1 d2 d3 6= 0

rank A ≥ 31b d4 d5 d6 6= 0 rank A ≥ 31c d7 d8 d9 6= 0 rank A ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 47 / 55

Page 214: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 d3 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 d6 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 d8 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

What is the smallest possible rank of A?

Case 1a d1 d2 d3 6= 0 rank A ≥ 3

1b d4 d5 d6 6= 0 rank A ≥ 31c d7 d8 d9 6= 0 rank A ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 47 / 55

Page 215: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 d3 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 d6 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 d8 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

What is the smallest possible rank of A?

Case 1a d1 d2 d3 6= 0 rank A ≥ 31b d4 d5 d6 6= 0 rank A ≥ 3

1c d7 d8 d9 6= 0 rank A ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 47 / 55

Page 216: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 d3 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 d6 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 d8 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

What is the smallest possible rank of A?

Case 1a d1 d2 d3 6= 0 rank A ≥ 31b d4 d5 d6 6= 0 rank A ≥ 31c d7 d8 d9 6= 0 rank A ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 47 / 55

Page 217: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 d3 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 d6 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 d8 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

Case 2

For some integers i , j , k ,

1 ≤ i ≤ 3, 4 ≤ j ≤ 6, 7 ≤ k ≤ 9,

di = dj = dk = 0.

All cases are alike. So say d3 = d6 = d8 = 0

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 48 / 55

Page 218: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 d3 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 d6 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 d8 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

Case 2 For some integers i , j , k ,

1 ≤ i ≤ 3, 4 ≤ j ≤ 6, 7 ≤ k ≤ 9,

di = dj = dk = 0.

All cases are alike. So say d3 = d6 = d8 = 0

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 48 / 55

Page 219: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 d3 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 d6 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 d8 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

Case 2 For some integers i , j , k ,

1 ≤ i ≤ 3, 4 ≤ j ≤ 6, 7 ≤ k ≤ 9,

di = dj = dk = 0.

All cases are alike.

So say d3 = d6 = d8 = 0

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 48 / 55

Page 220: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 d3 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 d6 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 d8 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

Case 2 For some integers i , j , k ,

1 ≤ i ≤ 3, 4 ≤ j ≤ 6, 7 ≤ k ≤ 9,

di = dj = dk = 0.

All cases are alike. So say d3 = d6 = d8 = 0Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 48 / 55

Page 221: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 0 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 0 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 0 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

Case 2 For some integers i , j , k ,

1 ≤ i ≤ 3, 4 ≤ j ≤ 6, 7 ≤ k ≤ 9,

di = dj = dk = 0.

All cases are alike. So say d3 = d6 = d8 = 0

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 49 / 55

Page 222: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 0 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 0 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 0 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

Let B =

0 c3 c5

c3 0 z2

c5 z2 0

det B = 2c3c5z2 6= 0 =⇒ rank B = 3 =⇒ rank A ≥ 3

In either case, rank A ≥ 3 =⇒ mr(K3,3,3) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 50 / 55

Page 223: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 0 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 0 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 0 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

Let B =

0 c3 c5

c3 0 z2

c5 z2 0

det B = 2c3c5z2 6= 0 =⇒ rank B = 3 =⇒ rank A ≥ 3

In either case, rank A ≥ 3 =⇒ mr(K3,3,3) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 50 / 55

Page 224: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 0 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 0 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 0 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

Let B =

0 c3 c5

c3 0 z2

c5 z2 0

det B = 2c3c5z2 6= 0

=⇒ rank B = 3 =⇒ rank A ≥ 3

In either case, rank A ≥ 3 =⇒ mr(K3,3,3) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 50 / 55

Page 225: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 0 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 0 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 0 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

Let B =

0 c3 c5

c3 0 z2

c5 z2 0

det B = 2c3c5z2 6= 0 =⇒ rank B = 3

=⇒ rank A ≥ 3

In either case, rank A ≥ 3 =⇒ mr(K3,3,3) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 50 / 55

Page 226: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 0 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 0 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 0 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

Let B =

0 c3 c5

c3 0 z2

c5 z2 0

det B = 2c3c5z2 6= 0 =⇒ rank B = 3 =⇒ rank A ≥ 3

In either case, rank A ≥ 3 =⇒ mr(K3,3,3) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 50 / 55

Page 227: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 0 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 0 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 0 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

Let B =

0 c3 c5

c3 0 z2

c5 z2 0

det B = 2c3c5z2 6= 0 =⇒ rank B = 3 =⇒ rank A ≥ 3

In either case, rank A ≥ 3

=⇒ mr(K3,3,3) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 50 / 55

Page 228: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3

A =

d1 0 0 a1 a2 a3 a4 a5 a6

0 d2 0 b1 b2 b3 b4 b5 b6

0 0 0 c1 c2 c3 c4 c5 c6

a1 b1 c1 d4 0 0 x1 x2 x3

a2 b2 c2 0 d5 0 y1 y2 y3

a3 b3 c3 0 0 0 z1 z2 z3

a4 b4 c4 x1 y1 z1 d7 0 0a5 b5 c5 x2 y2 z2 0 0 0a6 b6 c6 x3 y3 z3 0 0 d9

∈ S(K3,3,3)

Let B =

0 c3 c5

c3 0 z2

c5 z2 0

det B = 2c3c5z2 6= 0 =⇒ rank B = 3 =⇒ rank A ≥ 3

In either case, rank A ≥ 3 =⇒ mr(K3,3,3) ≥ 3

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 50 / 55

Page 229: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3 − v

K3,3,3 − v ∼= K3,3,2 for every v

A =

0 0 0 1 1 1 1 10 0 0 1 1 1 1 10 0 0 1 1 1 1 11 1 1 0 0 0 1 −11 1 1 0 0 0 1 −11 1 1 0 0 0 1 −11 1 1 1 1 1 2 0−1 −1 −1 1 1 1 0 2

∈ S(K3,3,2)

rank A = 2 =⇒ mr(K3,3,2) ≤ 2

So K3,3,3 is also a minimal forbidden subgraph for the minimum rank 2problem.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 51 / 55

Page 230: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3 − v

K3,3,3 − v ∼= K3,3,2 for every v

A =

0 0 0 1 1 1 1 10 0 0 1 1 1 1 10 0 0 1 1 1 1 11 1 1 0 0 0 1 −11 1 1 0 0 0 1 −11 1 1 0 0 0 1 −11 1 1 1 1 1 2 0−1 −1 −1 1 1 1 0 2

∈ S(K3,3,2)

rank A = 2 =⇒ mr(K3,3,2) ≤ 2

So K3,3,3 is also a minimal forbidden subgraph for the minimum rank 2problem.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 51 / 55

Page 231: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3 − v

K3,3,3 − v ∼= K3,3,2 for every v

A =

0 0 0 1 1 1 1 10 0 0 1 1 1 1 10 0 0 1 1 1 1 11 1 1 0 0 0 1 −11 1 1 0 0 0 1 −11 1 1 0 0 0 1 −11 1 1 1 1 1 2 0−1 −1 −1 1 1 1 0 2

∈ S(K3,3,2)

rank A = 2

=⇒ mr(K3,3,2) ≤ 2

So K3,3,3 is also a minimal forbidden subgraph for the minimum rank 2problem.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 51 / 55

Page 232: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3 − v

K3,3,3 − v ∼= K3,3,2 for every v

A =

0 0 0 1 1 1 1 10 0 0 1 1 1 1 10 0 0 1 1 1 1 11 1 1 0 0 0 1 −11 1 1 0 0 0 1 −11 1 1 0 0 0 1 −11 1 1 1 1 1 2 0−1 −1 −1 1 1 1 0 2

∈ S(K3,3,2)

rank A = 2 =⇒ mr(K3,3,2) ≤ 2

So K3,3,3 is also a minimal forbidden subgraph for the minimum rank 2problem.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 51 / 55

Page 233: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

K3,3,3 − v

K3,3,3 − v ∼= K3,3,2 for every v

A =

0 0 0 1 1 1 1 10 0 0 1 1 1 1 10 0 0 1 1 1 1 11 1 1 0 0 0 1 −11 1 1 0 0 0 1 −11 1 1 0 0 0 1 −11 1 1 1 1 1 2 0−1 −1 −1 1 1 1 0 2

∈ S(K3,3,2)

rank A = 2 =⇒ mr(K3,3,2) ≤ 2

So K3,3,3 is also a minimal forbidden subgraph for the minimum rank 2problem.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 51 / 55

Page 234: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank 2 Theorem

These are all the minimal forbidden subgraphs.

Theorem

Let G be a connected graph. Then mr(G ) ≤ 2 if and only if none of the 4graphs

P4 dart folding stool K3,3,3

is induced in G .

Reverse implication is much longer and depends on a characterization ofthe complements of the class of graphs that are {P4, dart, folding stool,K3,3,3}−free

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 52 / 55

Page 235: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank 2 Theorem

These are all the minimal forbidden subgraphs.

Theorem

Let G be a connected graph. Then mr(G ) ≤ 2 if and only if none of the 4graphs

P4 dart folding stool K3,3,3

is induced in G .

Reverse implication is much longer and depends on a characterization ofthe complements of the class of graphs that are {P4, dart, folding stool,K3,3,3}−free

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 52 / 55

Page 236: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank 2 Theorem

These are all the minimal forbidden subgraphs.

Theorem

Let G be a connected graph. Then mr(G ) ≤ 2 if and only if none of the 4graphs

P4 dart folding stool K3,3,3

is induced in G .

Reverse implication is much longer and depends on a characterization ofthe complements of the class of graphs that are {P4, dart, folding stool,K3,3,3}−free

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 52 / 55

Page 237: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Other formulations

Let G be a connected graph.

Then M(G ) ≥ n − 2 if and only if none of the 4 graphs P4, dart, foldingstool, K3,3,3 is induced in G .

M(G ) ≤ n − 3 if and only if one of the 4 graphs P4, dart, folding stool,K3,3,3 is induced in G .

Eigenvalue formulation: A matrix in S(G ) can have an eigenvalue ofmaximum multiplicity at most n − 3 if and only if one of the graphs P4,dart, folding stool, K3,3,3 is induced in G .

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 53 / 55

Page 238: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Other formulations

Let G be a connected graph.

Then M(G ) ≥ n − 2 if and only if none of the 4 graphs P4, dart, foldingstool, K3,3,3 is induced in G .

M(G ) ≤ n − 3 if and only if one of the 4 graphs P4, dart, folding stool,K3,3,3 is induced in G .

Eigenvalue formulation: A matrix in S(G ) can have an eigenvalue ofmaximum multiplicity at most n − 3 if and only if one of the graphs P4,dart, folding stool, K3,3,3 is induced in G .

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 53 / 55

Page 239: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Other formulations

Let G be a connected graph.

Then M(G ) ≥ n − 2 if and only if none of the 4 graphs P4, dart, foldingstool, K3,3,3 is induced in G .

M(G ) ≤ n − 3 if and only if one of the 4 graphs P4, dart, folding stool,K3,3,3 is induced in G .

Eigenvalue formulation: A matrix in S(G ) can have an eigenvalue ofmaximum multiplicity at most n − 3 if and only if one of the graphs P4,dart, folding stool, K3,3,3 is induced in G .

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 53 / 55

Page 240: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Other formulations

Let G be a connected graph.

Then M(G ) ≥ n − 2 if and only if none of the 4 graphs P4, dart, foldingstool, K3,3,3 is induced in G .

M(G ) ≤ n − 3 if and only if one of the 4 graphs P4, dart, folding stool,K3,3,3 is induced in G .

Eigenvalue formulation: A matrix in S(G ) can have an eigenvalue ofmaximum multiplicity at most n − 3 if and only if one of the graphs P4,dart, folding stool, K3,3,3 is induced in G .

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 53 / 55

Page 241: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank 2 Theorem over any field

For any field F , one can define S(F ,G ) to be the class of all symmetricmatrices corresponding to G with entries in F .

For each such field there is a characterization of the graphs such thatmr(F ,G ) ≤ 2 in terms of forbidden subgraphs.

For any infinite field with char F 6= 2, the result is the same.

For Hermitian matrices, the result is the same except K3,3,3 drops off thelist.

For disconnected graphs, the result is the same except that one has toinclude two disconnected graphs

P3 ∪ K2 3K2

in the list of forbidden subgraphs.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 54 / 55

Page 242: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank 2 Theorem over any field

For any field F , one can define S(F ,G ) to be the class of all symmetricmatrices corresponding to G with entries in F .

For each such field there is a characterization of the graphs such thatmr(F ,G ) ≤ 2 in terms of forbidden subgraphs.

For any infinite field with char F 6= 2, the result is the same.

For Hermitian matrices, the result is the same except K3,3,3 drops off thelist.

For disconnected graphs, the result is the same except that one has toinclude two disconnected graphs

P3 ∪ K2 3K2

in the list of forbidden subgraphs.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 54 / 55

Page 243: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank 2 Theorem over any field

For any field F , one can define S(F ,G ) to be the class of all symmetricmatrices corresponding to G with entries in F .

For each such field there is a characterization of the graphs such thatmr(F ,G ) ≤ 2 in terms of forbidden subgraphs.

For any infinite field with char F 6= 2, the result is the same.

For Hermitian matrices, the result is the same except K3,3,3 drops off thelist.

For disconnected graphs, the result is the same except that one has toinclude two disconnected graphs

P3 ∪ K2 3K2

in the list of forbidden subgraphs.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 54 / 55

Page 244: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank 2 Theorem over any field

For any field F , one can define S(F ,G ) to be the class of all symmetricmatrices corresponding to G with entries in F .

For each such field there is a characterization of the graphs such thatmr(F ,G ) ≤ 2 in terms of forbidden subgraphs.

For any infinite field with char F 6= 2, the result is the same.

For Hermitian matrices, the result is the same except K3,3,3 drops off thelist.

For disconnected graphs, the result is the same except that one has toinclude two disconnected graphs

P3 ∪ K2 3K2

in the list of forbidden subgraphs.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 54 / 55

Page 245: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank 2 Theorem over any field

For any field F , one can define S(F ,G ) to be the class of all symmetricmatrices corresponding to G with entries in F .

For each such field there is a characterization of the graphs such thatmr(F ,G ) ≤ 2 in terms of forbidden subgraphs.

For any infinite field with char F 6= 2, the result is the same.

For Hermitian matrices, the result is the same except K3,3,3 drops off thelist.

For disconnected graphs, the result is the same except that one has toinclude two disconnected graphs

P3 ∪ K2 3K2

in the list of forbidden subgraphs.

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 54 / 55

Page 246: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

Minimum Rank 2 Theorem over any field

For any field F , one can define S(F ,G ) to be the class of all symmetricmatrices corresponding to G with entries in F .

For each such field there is a characterization of the graphs such thatmr(F ,G ) ≤ 2 in terms of forbidden subgraphs.

For any infinite field with char F 6= 2, the result is the same.

For Hermitian matrices, the result is the same except K3,3,3 drops off thelist.

For disconnected graphs, the result is the same except that one has toinclude two disconnected graphs

P3 ∪ K2 3K2

in the list of forbidden subgraphs.Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 54 / 55

Page 247: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

The minimum rank 3 Problem

What graphs G have mr(G ) ≤ 3?

What are the forbidden subgraphs?

Major question of my Fulbright proposal for leave in Israel in 2007.

Tracy Hall demolished any hope of solving this problem a week or twobefore I left by showing that the list of forbidden subgraphs is infinite.

Picture description

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 55 / 55

Page 248: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

The minimum rank 3 Problem

What graphs G have mr(G ) ≤ 3? What are the forbidden subgraphs?

Major question of my Fulbright proposal for leave in Israel in 2007.

Tracy Hall demolished any hope of solving this problem a week or twobefore I left by showing that the list of forbidden subgraphs is infinite.

Picture description

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 55 / 55

Page 249: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

The minimum rank 3 Problem

What graphs G have mr(G ) ≤ 3? What are the forbidden subgraphs?

Major question of my Fulbright proposal for leave in Israel in 2007.

Tracy Hall demolished any hope of solving this problem a week or twobefore I left by showing that the list of forbidden subgraphs is infinite.

Picture description

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 55 / 55

Page 250: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

The minimum rank 3 Problem

What graphs G have mr(G ) ≤ 3? What are the forbidden subgraphs?

Major question of my Fulbright proposal for leave in Israel in 2007.

Tracy Hall demolished any hope of solving this problem a week or twobefore I left

by showing that the list of forbidden subgraphs is infinite.

Picture description

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 55 / 55

Page 251: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

The minimum rank 3 Problem

What graphs G have mr(G ) ≤ 3? What are the forbidden subgraphs?

Major question of my Fulbright proposal for leave in Israel in 2007.

Tracy Hall demolished any hope of solving this problem a week or twobefore I left by showing that the list of forbidden subgraphs is infinite.

Picture description

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 55 / 55

Page 252: Combinatorial Matrix Theory...Combinatorial Matrix Theory Fusion of Graph Theory and Matrix Theory Background in Graph Theory. 6 vertices 9 edges Cycles in a graph Wayne Barrett (BYU

The minimum rank 3 Problem

What graphs G have mr(G ) ≤ 3? What are the forbidden subgraphs?

Major question of my Fulbright proposal for leave in Israel in 2007.

Tracy Hall demolished any hope of solving this problem a week or twobefore I left by showing that the list of forbidden subgraphs is infinite.

Picture description

Wayne Barrett (BYU ) Combinatorial Matrix Theory August 30, 2013 55 / 55