Relations and Graphs

Composition and
Graph Powers
Sections 6.3-6.5
Prof. Sandy Irani
 Directed Graphs
A directed graph G = (V, E) consists of
V: set of verties
E: set of edges. E ⊆ V x V

Example: V = {a, b, c, d}
E = { (a,b), (b,b), (b,c), (c,b), (c,d) }

a b
A directed graph G=(V,E)
is the same as a relation
d c E on set V
 Directed Graph Terminology

The tail of edge (b,d) is b

a The head of edge (b,d) is d
b

The in-degree of v = |{u: (u,v) ∈ E}|

d c The out-degree of v = |{u: (v,u) ∈ E}|

In-degree of c =

Out-degree of b =
 A Walk in A Directed Graph
The walk in a directed graph is an alternating
sequence of vertices and edges, starting and ending
with a vertex.
Each edge comes after its tail and before its head.
< b, (b,a), a, (a,b), b, (b,d), d, (d,e), e, (e,c), c, (c,c), c, (c,d), d >

b Including the edges is redundant:

c <b, a, b, d, e, c, c, d>
a
<b, a, e> is not a walk because
d (a, e) is not an edge.
e
 A Path in A Directed Graph
The walk in a directed graph is also a path if it has no
repeated vertices.
b
<a, b, d, e, c>
c
a <b, a, d, e>

<c, d, e>
d e
<d, e>
 A Circuit in A Directed Graph
The walk in a directed graph is also a circuit if it
starts and ends at the same vertex.
b
<e, c, d, e>
c
a <a, b, a>

<c, c>
d e
<e, c, c, d, e, c, d, e>
 A Circuit in A Directed Graph
The circuit in a directed graph is also a cycle if there are
no repeated vertices except for the first and the last.

b
<e, d, c, e>
c
a <a, b, a>

<c, c>
d e
< a, d, e, c, b, a >
 Walks

Paths

Circuits

Cycles
 Directed Graph Terminology *
Pick the correct description for the sequence

<c, d, a, b, d, e, c>

in the graph below.

b A) A walk but not a path, circuit or cycle

B) A walk and a circuit, but not a path or cycle

c
a
C) A walk, circuit, and cycle but not a path.

D) A walk and a path but not a circuit or cycle

d e
E) Not a walk
 The length of a walk
The length of a walk (path, circuit, cycle) is the number
of edges.

b <e, d, c, e>

c <a, b, a>
a
<c, c>
d e <b, a, b, e, d, c, c, e>

<d>
 Transitive Closure of a Directed Graph
Is it possible to travel from b to c
b by a walk?

c From c to b?
a
We can build a new graph with
The same vertex set.
d e
The new graph has an edge (x,y)
Whenever Y can be reached from x by a walk.
What is the smallest set of edges we need to add to make edges
in the graph transitive?
….the Transitive Closure.
Composition of Binary Relations
a b a b

d c d c
R S
a b
Add an edge from x to y
If there is a red-blue walk
d c x z y
S◦R Can have x = z or x = y or z = y
Composing a Binary Relation
with Itself
a a b
b

d c
d c
R
R◦R
Add an edge from x to y
if there is a walk of length 2
From x to y in R.

x z y
Can have x = z or x = y or z = y
 Composition of Binary Relations *

A)3(R◦R)9
R is a binary relation on ℤ

For any x,y ∈ ℤ, xRy if x2 = y B) 3(R◦R)81

Select the statement that is true: C) 9(R◦R)3

D) 81(R◦R)3
 Powers of a Relation
We can compose a relation multiple times with
itself:
R1 = A
R2 = R ◦ R
R3 = R ◦ R◦ R = R2 ◦ R
kth power of R
…
Rk = R ◦ R◦ R ◦… ◦ R = Rk-1 ◦ R
k times
 Graph Powers
Graph G = (V, E)

Graph Gk has the same vertex set as G.

(x,y) is an edge in Gk if and only if there is a walk
of length k in G from x to y.

Theorem: The edges in Gk are the relation Ek

 Graph Power Example
G G2
2 3 2 3

1 4 1 4

6 5 6 5
G3
2 3

1 4

6 5
 Graph Powers *
What is the out-degree of vertex 2 in G4?

2 3 A) 2

1 4 B) 3

C) 4
6 5
G D) 5
 Graph Powers
Graph Gk has the same vertex set as G.
(x,y) is an edge in Gk if and only if there is a walk
of length k in G from x to y.

Theorem: The edges in Gk are the relation Ek

Why?.... If there is a walk of length k from x to y in G:

x z y
walk of length k-1 edge from z to y in G
from x to z in G

(x,z) is an edge in Gk-1

Transitive Closure of a Directed Graph
Transitive closure of G: G+ = G1 ∪ G2 ∪ G3∪ G4…..
(Same vertex set. Take the union of the edges.)

(x,y) is an edge in G+ if and only if there is a walk in G

of any length.

Don’t actually need an infinite union:

If there is a walk from x to y in G, then there is a walk of
length at most n from x to y. (n is the number of vertices in G.)

G+ = G1 ∪ G2 ∪ G3∪ G4 ∪ ….. ∪ Gn
If there is a walk from x to y in G, then there is a walk of length at
most n from x to y. (n is the number of vertices in G.)

Worst Case:
Transitive Closure of a Directed Graph
Transitive closure of G: G+ = G1 ∪ G2 ∪ G3∪ G4…..
(Same vertex set. Take the union of the edges.)

(x,y) is an edge in G+ if and only if there is a walk in G

of any length.

The edge set of G+ is transitive:

Walk from x to y + Walk from y to z  Walk from x to z

y
x z
 Transitive Closure of a Directed Graph *
Which pair is not an edge in the transitive closure of G?

2 3 A) (1,1)

1 4 B) (1,4)

C) (5,5)
6 5
G D) (3,6)
Transitive Closure of a Binary Relation
Transitive closure of Binary Relation R:
R+ = R1 ∪ R2 ∪ R3∪ R4…..
(Same domain. Take the union of the pairs.)

R+ is the smallest relation that includes all of R and is

transitive.
Alternative way to find the transitive closure of R:

Repeat until no new pairs are added:

If (x,y) ∈ R and (y,z) ∈ R and (x,z) ∉ R
Add (x,z) to R

2 3

1 4

6 5