CS2800

Graphs and Relations

c Renee Mirka & Anke van Zuylen

Spring 2023

1
Graphs

2
Graphs

A graph consists of vertices (aka nodes) and edges (aka links, arcs)
between vertices.

A directed graph (digraph) is one where the edges have a direction,

usually denoted with an arrow; and undirected graph is one where the
edges do not have a direction.

1 2 1 2

4 3 4 3

3
Graph notation

A graph G is a pair (V , E ), where V is the set of vertices and E is the

set of edges.

• We write G = (V , E ) so we can use G to refer to the entire graph,

and V and E to refer to just its vertices or just its edges.
• If G is introduced without introducing (V , E ), we can also write
V (G ) and E (G ) to indicate the vertices and edges in G .

4
Edge notation

To denote an edge between vertices u and v (where u and v are two, not
necessarily distinct, vertices), we use di↵erent notation depending on the
type of graph:

• if G is a digraph, then we denote the edge by an ordered pair (u, v ).

• If G is an undirected graph, then we denote the edge by an
unordered pair {u, v }.
• If we don’t want to specify whether an edge is directed or
undirected, we’ll use notation uv .

Although it is possible to have a graph that has both directed and

undirected edges, this is not common.
In addition, it is common to explicitly mention it if a graph is allowed to
have self-loops (edges of the form vv ) or duplicate edges.

5
Graph notation examples

Describe these two graphs using the notation from the previous two
slides.

1 2 1 2

4 3 4 3

6
Some more graph terminology

In an undirected graph:

• Two vertices in a graph are said to be adjacent i↵ they are the

endpoints of the same edge, and an edge is said to be incident to
each of its endpoints.
• The number of edges incident to a vertex v is called the degree of
the vertex, and denoted deg(v ).
• We say u is a neighbor of v if u and v are adjacent.

In a directed graph:

• For a directed edge (u, v ), we call u the tail of the edge, and we call
v the head.
• In a digraph, the in-degree (deg (v )) of a vertex v is the number of
edges of which v is the head, and the out-degree (deg+ (v )) of v is
the number of edges of which v is the tail.

7
Degree examples

Write down the degree of all vertices.

1 2 1 2

4 3 4 3

8
Edges and degrees

(Handshake Lemma) Let G = (V , E ) be an undirected graph. Then

X
2|E | = deg (v ).
v 2V

(Note that this applies even if duplicate edges and loops are present.)

9
Edges and degrees

How about for directed graphs? Describe the relation between the
following quantities:

X X
|E | deg (v ) deg+ (v )
v 2V v 2V

10
Graph applications

Graphs come up everywhere.

• We can view the internet as a graph (in many ways);

• niche graphs in ecology view the species as vertices which are
connected by an edge if they compete (use the same food resources,
etc.);
• contact tracing sars-cov-2 infections can be viewed as finding
“neighborhoods” in a graph;
• finding kidney exchanges has been defined as finding cycles in a
directed graph;
• ...

11
Connectivity

12
More graph terminology: paths and circuits/cycles

From now on, we’ll assume graphs are simple, that is, they do not have
duplicate edges.

A path in a graph G is a sequence (ordered list) of vertices v0 , v1 , . . . , vk

with the property that vi vi+1 2 E for i = 0, . . . , k 1.

• We say the path v0 , v1 , . . . , vk goes from v0 to vk .

• The number of edges on the path v0 , v1 , . . . , vk is k and this is
called the length of the path.
• Note that the number of vertices in a path is 1 more than the
number of edges.
• A path v0 , v1 , . . . , vk is called simple if all vertices except possibly v0
and vk are distinct.
• A path v0 , v1 , . . . , vk is called a circuit or cycle if v0 = vk .
• A simple circuit is also called a cycle.
13
Paths and circuits example

• In the graph on the left, is there a path from vertex 1 to vertex 3?

What is its length?
• In the graph on the right, is there a path of length three
• from vertex 1 to vertex 4?
• from vertex 4 to vertex 1?
• from vertex 1 to vertex 2?
• Give a cyle of length 2 in the graph on the right.
• Question: Is it possible to have a path of length 0?

1 2 1 2

4 3 4 3
14
Warning

Di↵erent authors use di↵erent terminology and notation. For example,

you might see what we call a path be referred to as a walk, and you
might see authors writing down the sequence of edges instead of vertices,
or an alternating sequence of vertices and edges to denote a path.

15
Connectivity in social networks

16
Some special graphs

17
Special graphs

There are many special graphs that have many applications in computer
science, such as:

• complete graphs (simple, undirected graph, where all possible edges

are present);
• hypercube (see book);
• bipartite graphs (discussed next);
• trees (also discussed next);
• and many more...

18
Trees

A simple undirected graph is called a tree if

• it is connected (this means that there must be a path between every

pair of vertices), and
• it has no cycles.

19
Not a tree

20
Bipartite graphs

A simple graph G is called bipartite if its vertex set V can be partitioned

into two disjoint sets V1 and V2 such that every edge in the graph
connects a vertex in V1 and a vertex in V2 (so that no edge in G
connects either two vertices in V1 or two vertices in V2 ).

When this condition holds, we call the pair (V1 , V2 ) a bipartition of the
vertex set V of G .

21
Bipartite graph example

Are the following graphs bipartite?

1 2

4 3

22
Some useful facts

Theorem 1: A tree on n vertices has n 1 edges.

23
Some useful facts

Theorem 2: A graph is bipartite if and only if it doesn’t have odd cycles.

24
Binary relations

25
Binary relations

Recall the cartesian product of sets S and T :

S ⇥ T = {(s, t) : s 2 S, t 2 T }.

A (binary) relation on S and T is a subset R of S ⇥ T .

• Notation: (s, t) 2 R or s R t
• Example: Let S = {a, b, c}, T = {1, 2} and
R = {(a, 1), (a, 2), (b, 1), (c, 2)}.
• Representation as a bipartite graph and as a matrix:

26
Binary relations and functions

Recall that a function from S to T was formally defined as a subset f of

S ⇥ T such that for each s 2 S, there is a unique t 2 T such that
(s, t) 2 f .
In other words, if f is a function from S to T then f is is a binary
relation on S and T (but the converse is not true)!

27
Binary relations on a single set

A binary relation on a single set A is a subset of A ⇥ A.

From now on, we will restrict attention to such binary relations.

Examples:

• < relation on R
• “is a family member of” relation on people
•  relation on the cardinalities of infinite sets
• “bought the same product as” relation on Amazon customers
• ✓ relation on sets
• ...

28
Properties of relations

A relation R on A is

• reflexive if 8x (xRx)
• symmetric if 8x8y (xRy ) yRx).
• transitive if 8x8y 8z (xRy ^ yRz ) xRz).

Examples:
Set A reals Amazon customers
Relation R <  “bought same product” “lives on same address’
reflexive
symmetric
transitive

29
Properties of the ⇢ relation

Question: The “strict-subset” relation ⇢ is defined as follows:

X ⇢ Y is true i↵ X ✓ Y and X 6= Y . Select which of the following is
true (choose one):

(a) ⇢ is reflexive.
(b) ⇢ is symmetric.
(c) ⇢ is transitive.

30
One more property

A relation R on A is anti-symmetric if 8x8y (xRy ^ yRx ) x = y ).

Examples:
Set A integers sets
Relation R <  “has same parity” ✓
antisymmetric

31
Properties of binary relations

Question: Choose the statement that is true:

(a) If R is anti-symmetric, then R must be reflexive.

(b) If R is not symmetric, then R is anti-symmetric.
(c) If R is transitive and not anti-symmetric, then R must be reflexive.
(d) None of the above.

32
Inverting relations

For functions, we defined the inverse f 1 (which exists only if f is

bijective).
A binary relation always has an inverse.
If R is a relation on A, then R 1 is the relation on A satisfying:
1
8x, y xRy , yR x .

Reminder: We can also write this as

1
8x, y (x, y ) 2 R , (y , x) 2 R .

33
Example: inverting a relation

Example: Let A = {1, 2, 3, 4} and R = {(1, 1), (1, 2), (2, 1), (3, 4)}.
1
• Find R .
• Is R anti-symmetric?

34
Exercise 1: revisiting anti-symmetry

The identity (or diagonal) relation on a set A is the set

{(a, a) : a 2 A} (in other words, a b holds if and only if b = a)
Express the definition for R being anti-symmetric using only the following
symbols:
1
• R and R
•
• [, \, \, ⇢, ✓, =

and prove your new definition is equivalent to the original definition.

35
51
Composition and powers of relations

For functions, we defined composition f g . The same definition applies

for composing binary relations.
Using the definition of composition, we see that R R is the relation
satisfying:

8x8y ((x, y ) 2 R R, ).

36
Composition and powers of relations

The powers R n of R, for n = 1, 2, 3, . . ., are defined recursively as

• R 1 = R,
• R n+1 = R n R for n 1.

37
Example: composition and powers of relations

Example: Let A = {1, 2, 3, 4} and R = {(1, 1), (1, 2), (2, 3), (3, 4)}.

• Find R n for n = 1, 2, 3, . . .
• Is R transitive?

38
Exercise 2: revisiting transitivity

Express the definition for R being transitive using only the following
symbols:

• R, R R, R n (for n = 1, 2, 3, . . .)
• [, \, \, ⇢, ✓, =

and prove your new definition is equivalent to the original definition.

39
51
Reviewing binary relations using
their graph representation

40
Graph representation of a binary relation

A binary relation R on a set A can be represented as a digraph G with

vertex set A and directed edge set R.
For example, if A = {1, 2, 3, 4} and
R = {(1, 2), (2, 2), (2, 3), (3, 2), (1, 3), (4, 3)}, then the following graph
represents R:

Note: We can use a graph to represent a binary relation R on A only if

|A| is finite but binary relations also exist on infinite sets.

41
Binary relations and their graph representation

Question:
Let R be represented by the graph above. Which of the following
properties does R have?

(a) reflexive (8x xRx)

(b) symmetric (8x8y (xRy ) yRx))
(c) antisymmetric (8x8y (xRy ^ yRx ) x = y ))
(d) transitive (8x8y 8z (xRy ^ yRz ) xRz))

42
Composing binary relations

Question: Let R be represented by the following graph.

Which of the following is not in R R?

(a) (1,3)
(b) (1,2)
(c) (4,3)

43
Composition and powers of relations

Finish the following theorem. How would you prove it?

Theorem Let digraph G represent a binary relation R on A. Then

relation R n contains (a, b) if and only if G has the property that ...
.

44
Transitivity

Remember that in Exercise 2 we gave the following alternative

(equivalent) definition for a binary relation R to be transitive:

R R ✓ R.

Can you rephrase this requirement as a requirement on the graph

representing R?

45
The transitive closure

46
Transitive closure of a binary relation

Definition: The transitive closure of a relation R on A is the relation

R trans on A such that

(i) R ✓ R trans ,
(ii) R trans is transitive,
(iii) For any transitive relation R 0 on A that contains R, it holds that
R trans ✓ R 0 .

Note that that A ⇥ A always satisfies the two properties! But if (iii)
doesn’t hold, then there is a smaller relation that satisfies them.
In other words, we want the smallest relation that contains all of R, and
that is transitive.

47
Example of transitive closure

Question: Suppose A = {1, 2, 3, 4} and R = {(1, 2), (2, 3), (3, 4)}.
What is R trans ?

(a) R trans = {(1, 3), (2, 4)}

(b) R trans = {1, 2, 3, 4} ⇥ {1, 2, 3, 4}
(c) R trans = R [ {(1, 3), (2, 4), (1, 4)}
(d) R trans = R [ {(1, 3), (2, 4)}

48
The connectivity relation

Given graph representing a relation R, the connectivity relation R ⇤

consists of the pairs (a, b) such that there is a path of length at least one
from a to b.
Express R ⇤ in terms of the powers of R.

49
Transitive closure and the connectivity relation

Theorem: The transitive closure equals the connectivity relation of R.

Proof:

50
51
Equivalence Relations

52
Equivalence relations

A relation R is an equivalence relation if it is reflexive, symmetric, and

transitive.
Examples: Verify that the following relations are equivalence relations:

• = on N
• ⇠P on N, where x ⇠P y i↵ x y is even.
• =1 on N ⇥ N, where (x1 , y1 ) =1 (x2 , y2 ) i↵ x1 = x2 .

53
Equivalence class

If R is an equivalence relation, the equivalence class of a is denoted [a],

and is defined as [a] = {b : aRb}
Examples: What are the equivalence classes of the two equivalence
relations given on the previous slide?

54
Partition

Definition: Let A be a set. A partition of A is a collection A1 , . . . , Ak of

subsets of A such that every element of A is in exactly one of
A1 , . . . , Ak .
In other words, A1 , . . . , Ak ✓ A form a partition of A if

• the subsets A1 , . . . , Ak are disjoint,

Sk
• i=1 Ai = A.

Examples:

• The equivalence classes of the parity relation ⇠P give a partition of

the positive integers into {odd positive integers} and {even positive
integers}.
• The equivalence classes of the = relation give a partition of the
positive integers into .......

55
Equivalence classes

Theorem 1: The equivalence classes of an equivalence relation on A

form a partition of A.
In other words, for any two elements a, b of A, either their equivalence
class are the same (so [a] = [b]) or their equivalence classes are disjoint
(so [a] \ [b] = ;).
Proof:

56
57
Partial orders

58
Partial orders

A relation is (weak or non-strict) partial order if it is reflexive, transitive,

and antisymmetric.
Examples:

•  and are (weak) partial orders on numbers.

• ✓ is a (weak) partial order on sets.

59
The graph for a partial order

Question: What can we say about the graph representing a partial

order?

(a) Every edge is bidirected.

(b) There is an edge (in one or both possible directions) between every
pair of vertices.
(c) It has no self-loops.
(d) It has no cycles except for self-loops.

60