DECLARATION

I, hereby declare that this dissertation entitled, ‘A STUDY ON DIGRAPHS’,is a

bona fide record of research work done by me and this has not previously formed
the basis for the award of any Degree, Diploma, Associate ship, Fellowship or any
other similar title or recognition.

Calicut KRIPA NARAYANAN C.

Date : 05/08/2016 Department of Mathematics

Malabar Christian College, Calicut

Calicut

1
 CERTIFICATE

This is to certify that the dissertation entitled ‘A STUDY ON DIGRAPHS’ is a

bona fide record of research work carried out by Ms.KRIPA NARAYANAN C.
during the academic year 2014-2016 under my guidance and supervision.

Calicut

2
 ACKNOWLEDGEMENT

With immense pleasure, I extend my deepest gratitude to all those who helped and
supported me to mould my thoughts and words into this successful project. At first
I bow before God Almighty who always strengthens me in all my endeaviours. I
express sincere gratitude to, Malabar Christian College, Calicut
who initiated me into the subject. Without his conscientious guidance and patients,
I would not have successfully completed this dissertation. I also thank all other
members of the Department who ignited my curiosity in the world of Mathematics
during the period of my study in this institution. I am grateful to the co-operation
extended to me by Mr. , Librarian, Malabar Christian College, Calicut
and other members of the staff in the library for their timely assistance. I cannot
fail to thank my parents and friends who fill me their optimism.

KRIPA NARAYANAN C.

3
CONTENTS
1. INTRODUTION 5
2. BASIC DEFINITIONS AND
ISOMORPHISM 6
2.1. DEFINITIONS 6
2.2. TYPES OF DIGRAPHS 9
2.3. ISOMORPHISM 12
3. DIGRAPHS AND BINARY
RELATIONS 15
4. CONNECTEDNESS OF DIGRAPHS 17
4.1. CONNECTEDNESS 18
4.2. EULERIAN DIGRAPH 21
5. TREES WITH DIRECTED EDGES 24
5.1. DIRECTED TREES 24
5.2 ACYCLIC GRAPHS 25
6. MATRIX REPRESENTSTION
OF DIGRAPHS 27
6.1. INCIDENCE MATRIX 27
6.2. ADJACENCY MATRIX 29
6.3. CONNECTEDNESS AND
ADJACENCY MATRIX 31
7. TOURMENTS 32
8. CONCLUSION 39

1. INTRODUCTION

4
 Graph theory is a very popular area of discrete mathematics with not only
numerous theoretical developments, but also countless applications to practical
problems. As a research area, graph theory is still relatively young, but it is
maturing rapidly with many deep results having been discovered over the last
couple of decades.

The theory of graphs can be roughly partitioned into two branches: the areas of
undirected graphs and directed graphs (digraphs). Even though both areas have
numerous important applications, for various reasons, undirected graphs have been
studied much more extensively than directed graphs. One of the reasons is that
undirected graphs form in a sense a special class of directed graphs (symmetric
digraphs) and hence problems that can be formulated for both directed and
undirected graphs are often easier for the latter.

The theory of directed graphs has developed enormously within the last three
decades. There is an extensive literature on digraphs (more than 3000 papers). Here
I am discussing some basic definitions related to digraphs and some basic concepts
related to digraphs like isomorphism, connectedness, matrix representation,
tournament etc.

2. BASIC DEFINITIONS

5
2.1 DEFINITIOS : DIRECTED GRAPH
A directed graph D is a pair (V,A), where V is a finite nonempty set and A is a
subset of V × V . The elements of V and A are respectively called vertices
(points) and arcs.

If (u,v) ∈ A then the arc (u,v) is said to have u has it’s initial vertex (tail) and v
as it’s terminal vertex (head). Also the arc (u,v) is said to join u to v.

Example:

Figure 2.1.1

D = ({1,2,3,4} ,{(1,2) , (2,3) , (1,3) , (3,1)} is a directed graph.

UNDERLYING GRAPH

If D is a digraph, the graph obtained from D by ‘removing the arrows’ from the
directed edges is called the underlying graph of D. This graph is also called the
undirected graph corresponding to D.

ORIENTATION OF A GRAPH

Given a graph G, if there is a digraph D such that G is the underlying graph of D

then D is called an orientation of G.

In figure 2.1.2, D is the orientation of G.

6
 D G

Figure 2.1.2

PARALLEL EDGES

Two (directed) edges e and e′ of a digraph D are said to be parallel if e and e′ have
the same initial vertex and the same terminal vertex.(figure 2.1.3)

Figure 2.1.3

SELF LOOP

A directed edge e is said to be a self loop if the initial and terminal vertex of e are
same.(figure 2.1.4)

Figure 2.1.4

IN-DEGREE AND OUT-DEGREE

If v is a vertex of a digraph D, the number of edges incident out of v is called the

outdegree of v and the number of edges incident into v is called the indegree of v.
The outdegree of v is denoted by d+(v) and the indegree of v is denoted by d-(v).

From figure 2.1.1, we have,

d+(1) = 2 d+(2) = 1 d+(3) = 1 d+(4) = 0

d-(1) = 1 d-(2) = 1 d-(3) = 2 d-(4) = 0.

7
  (d+(v) , d-(v)) is called the degree pair of v.
 The total degree of v is d+(v) + d-(v).
−¿ ( D )
Minimum and maximum indegree of a graph D denoted by δ
¿ and
−¿ ( D)
∆
¿ .

+¿ ( D ) +¿ ( D ) .
Similarly minimum and maximum outdegree of D by δ
¿ and ∆
¿

ISOLATED VERTEX

If v is a vertex of a digraph D then v is called an isolated vertex of D if

d+(v) = d-(v) = 0.

PENDANT VERTEX

If v is a vertex of a digraph D then v is called a pendant vertex of D if

d+(v) + d-(v) = 1.

SOURCE

If v is a vertex of a digraph D then v is called a source of D if d-(v) = 0.

SINK

If v is a vertex of a digraph D then v is called a sink of D if d+(v) = 0.

DISUBGRAPH

Let D = (V,A) be a given digraph . Then the digraph G =(U,B) is called a

disubgraph if U and B are subsets of V and A respectively.

SPANNING DIGRAPH

Let D = (V,A) be a digraph. A digraph H = (U,B) is said to be spanning if V = U

and B is a subset of A.

CONVERSE DIGRAPH

8
Let D be a given digraph. The digraph D′ obtained by reversing the directions of all
the arcs is called the converse digraph of D.

THEOREM 2.1

HAND - SHAKING DILEMMA

In a digraph D, the sum of the outdegree of all vertices is equal to the sum of the
indegrees of all vertices, each sum being equal to the number of edges in D.
+¿ ( v ) = ∑ d −¿ ( v )=q
v ∈V
d¿
∑¿
v ∈V

Where “q” is the number of arcs.

Proof :

Let “q” denote the number of arcs in D=(V,A).

d +¿ ( v ) d−¿ ( v )
Let B = ∑¿ and C = ∑¿
v ∈V v ∈V

An arc (u,w) contributes one to the outdegree of u and one to the indegree of w.
Hence each arc contributes 1 to the sum of B and 1 to the sum of C .

Hence B = C = q .

2.2 TYPES OF DIGRAPHS

Simple Digraphs

A digraphs that has no self-loop or parallel edges is called a simple digraph.

 The underlying graph of a simple digraph need not be simple.(figure 2.2.1)

9
 Figure 2.2.1

In figure 2.2.1, the digraph is simple but the underlying graph is not simple.

Asymmetric Digraphs

Digraphs that have atmost one directed edge between a pair of vertices, but are
allowed to have self-loops, are called asymmetric or antisymmetric digraph.

Figure 2.2.2

Symmetric Digraph

Digraphs in which for every edge (a, b) (i.e., from vertex a to b) there is also an
edge (b, a).(figure 2.2.3)

Figure 2.2.3

Balanced Digraphs

10
 A digraph D is said to be a balanced digraph or an isograph if d +(v) = d-(v) for
every vertex v of D.(figure 2.2.4)

Figure 2.2.4

Regular Digraph

A balanced digraph is said to be regular if every vertex has the same indegree and
outdegree as every other vertex.(figure 2.2.5)

Figure 2.2.5

 There are two kinds of complete digraphs.

Complete Symmetric Digraph

A complete symmetric digraph is a simple digraph in which there is exactly one

edge directed from every vertex to every other vertex.(figure 2.2.6)

Figure 2.2.6

11
Complete Asymmetric Digraph

A complete asymmetric digraph is an asymmetric digraph in which there is exactly

one edge between every pair of vertices.(figure 2.2.7)

Figure 2.2.7

2.3 ISOMORPHISM
Two digraphs D1 = (V1,A1) and D2 =(V2,A2) are said to be isomorphic ( D1 ≅ D2 ¿
if there exists a bijection f : V1 → V2 such that (u,w) ∈ A1 if and only if ,

(f(u) , f(w)) ∈ A2 . f is called an isomorphism from D1 to D2 .

Figure 2.3.1

Let us consider the following one-to-one correspondence between the directed

edges in the two digraphs :

(u2, u1) ↔ (v2, v1) (u5, u2) ↔ (v3, v2)

(u2, u4) ↔ (v2, v4) (u3, u2) ↔ (v5, v2)

12
(u3, u4) ↔ (v5, v4) (u5, u4) ↔ (v3, v4)

(u1, u3) ↔ (v1, v5).

These yield the following one-to-one correspondence between the vertices in the
two digraphs : u1 ↔ v1, u2 ↔ v2, u3 ↔ v5, u4 ↔ v4, u5 ↔ v3.

Hence we have in figure 2.3.1, (a) and (b) are isomorphic.

The above mentioned one-to-one correspondences between the vertices and the
directed edges establish the isomorphism between the given digraphs.

 Figure 2.3.2, (a) and (b) are two non isomorphic digraphs.

(a) (b)
Figure 2.3.2

THEOREM 2.2

If two digraphs are isomorphic .Then corresponding points have the same degree
pair.

Proof :

Let D1 =(V1,A1) and D2 =(V2,A2) be isomorphic under isomorphism f .

Let v ∈ V1

Let N(v) = {w : w ∈ V1 and (v,w) ∈ A1 }

N(f(v)) = {f(w) : f(w) ∈ V2 and (f(v),f(w)) ∈ A2 }

13
Now w ∈ N(v) ↔ (v, w) ∈ A1
↔ (f(v),f(w)) ∈ A2 (since f is isomorphism )
↔ f(v) ∈ N(f(v))

i.e. |N(v)| = |N(f(v))|

hence L.H.S and R.H.S are the outdegree of v and f(v) respectively. Hence v, f(v)
have the same outdegree. Similarly same indegree. Hence same degree pair.

14
3. DIGRAPHS AND BINARY RELATIONS

Let A and B be nonempty sets. A (binary) relation R from A to B is a subset

of A×B. If R ⊆ A×B and (a, b) ∈ R, where a ∈ A, b ∈ B, we say a “is related to”
b by R, and we write aRb. A relation R defined on a set X is a subset of X × X.

For example, less than, greater than and equality are the relations in the set of real
numbers. The property “is congruent to” defines a relation in the set of all triangles
in a plane. Also, parallelism defines a relation in the set of all lines in a plane.

A binary relation R on a set X can always be represented by a digraph. In such a

representation, each xi ∈ X is represented by a vertex xi and whenever there is a
relation R from xi to xj, an arc is drawn from xi to xj, for every pair (xi, xj).

For example, the digraph in figure 3.1.1 represents the relation “is greater than” on
a set consisting of five numbers {3,4,7,5,8}.

Figure 3.1.1

Let R define a relation on a nonempty set X. If R relates every element of X to

itself, that is, ∀ xi ∈ X, xiRxi. then the relation R is said to be reflexive. A
relation R is said to be symmetric if for all xi , xj ∈ X , xiRxj implies xjRxi. A relation
R is said to be transitive if for any three elements xi , xj and xk in X, xiRxj and xjRxk
imply xiRxk.

15
 Figure 3.1.2 Figure 3.1.3

Figure 3.1.2 represents a symmetric digraph and figure 3.1.3 represents a reflexive
digraph.

A binary relation is called an equivalence relation if it is reflexive, symmetric and

transitive.

We note that every binary relation on a finite set can be represented by a digraph
without parallel edges and vice versa.Clearly, the digraph of a reflexive relation
contains a loop at every vertex. A digraph representing a reflexive binary relation is
called a reflexive digraph.The digraph of a symmetric relation is a symmetric
digraph because for every arc from xi to xj, there is an arc from xj to xi.A digraph
representing a transitive relation on its vertex set is called a transitive digraph.

A binary relation R on a set M can also be represented by a matrix, called a relation

matrix. This is a (0, 1), n×n matrix MR = [mij], where n is the number of elements
in M, and is defined by,

mij = {1 if0 otx Rh erwise

i jx is true

16
4. CONNECTEDNESS OF DIGRAPHS

Definitions 4.1:

DIRECTED WALK, DIRECTED TRAIL, DIRECTED PATH AND

DIRECTED CYCLES

A directed walk in D is a finite sequence whose terms are alternately vertices and
edges in D such that each edge is incident out of the vertex preceeding it in the
sequence and incident into the vertex following it.

A directed trail : A directed walk in which all arcs are distinct. A vertex can
appear more than once in a directed walk but not an edge. The vertex with which a
directed walk begins is called its initial vertex and the vertex with which it ends is
called its final or terminal vertex.

Directed path: An open directed walk in which no vertex is repeated is called a

directed path.

Directed circuit : A closed directed walk in which no vertices, except the initial
and final vertices are repeated is called a directed circuit or a directed cycle.

 The number of edges present in a directed walk, directed path, directed

circuit is called its length.

Definitions 4.2 :

SEMI-WALK, SEMI-PATH, SEMI-CIRCUIT

Semi-Walk : A semi-walk in a digraph D is a walk in the underlying graph of D,

but is not a directed walk in D. A walk in D can mean either a directed walk or a
semi-walk in D.

Semi-path : A semi-path in a digraph D is a path in the underlying graph of D, but

is not a directed path in D. A path in D can mean either a directed path or a semi-
path in D.

17
Semi-circuit : A semi-circuit in a digraph D is a circuit in the underlying graph of
D, but is not a directed circuit in D. A circuit in D can mean either a directed circuit
or a semi-circuit in D.

REACHABILITY

Given two vertices u and v of a digraph D, we say that v is reachable (or

accessible) from u if there exist atleast one directed path in D from u to v.

4.1. CONNECTEDNESS
Weakly connected : If underlying simple graph is connected.

Unilaterally connetcted : If given any two vertices u and v atleast one reachable
from the other.

Strongly connected : If every vertex is reachable from every other vertex.

COMPONENTS

Let W be the maximal subset of V. Then the subgraph D(W) induced by W is said
to be,

Weak component of D if the corresponding subgraph of the underlying graph is

connected.

Unilateral component of D if given any two vertices u and v, atleast one is

reachable from the other.

Strong component of D if every vertex of W is reachable from every other vertex

of W.

 A digraph is said to be disconnected if it is not even weak.

ORIENTABLE GRAPH

A graph D is said to be orientable if there exists a strongly connected digraph G for

which D is the underlying graph.(figure 4.1.1)
18
 Figure 4.1.1
Here D is orientable.

Definition 4.3 : An arc sequence in a digraph D is an alternating sequence of

vertices and arcs of D.

The following results characterise various types of connectivity in digraphs.

THEOREM 4.1

A digraph is strong if and only if it has a spanning closed arc sequence.

Proof :

Necessity : Let D = (V, A) be a strong digraph with V = {v1, v2, ..., vn}. Then there
is an arc sequence from each vertex in V to every other vertex in V.

Therefore, there exists in D, arc sequences Q1, Q2, ..., Qn-1 such that the first vertex
of Qi is vi and the last vertex of Qi is vi+1, for i = 1, 2, ..., n−1.

Also, there exists an arc sequence, say Qn , with first vertex vn and the last vertex
v1. Then the arc sequence obtained by traversing the arc sequences Q1, Q2, ...,Qn in
succession is a spanning closed arc sequence of D.

Sufficiency : Let u and v be two distinct vertices of V. If v follows u in any

spanning closed arc sequence, say Q of D, then there exist a sequence of the arcs of
Q forming an arc sequence from u to v. If u follows v in Q, then there is an arc
sequence from u to the last vertex of Q and an arc sequence from that vertex to v.
An arc sequence from u to v is then obtained by traversing these two arc sequences
in succession.

Hence the proof.

19
THEOREM 4.2

A digraph D is unilateral if and only if it has a spanning arc sequence.

Proof :

Necessity : Assume D is unilateral. Let Q be an arc sequence in D which contains

maximum number of vertices and let Q begin at the vertex v1 of D and end at the
vertex v2 of D. If Q is a spanning arc sequence, there is nothing to prove. Assume
that Q is not a spanning arc sequence.

Then there exists a vertex, say u of D that is not in Q. Also in D, there is neither an
arc sequence from u to v1, nor from v2 to u. Since D is unilateral and does not
contain an arc sequence from u to v1, D contains an arc sequence from v1 to u.

Let w( ≠ v 2 ) be the last vertex of Q from which an arc sequence from w to u exists
in D. Let Q1 be an arc sequence from w to u in D. Let z be the vertex in D which is
the immediate successor of the last appearance of w in Q.

Clearly, D does not contain an arc sequence from z to u. Since D is unilateral, there
is an arc sequence, say Q2 from u to z in D. Traversing Q from v1 to the last
appearance of w, then traversing u, then traversing Q2 to vertex z and finally
traversing Q to v2, we obtain an arc sequence from v1 to v2 which has more distinct
vertices than Q. This is a contradiction and thus Q is a spanning arc sequence in D.
(figure 4.1.2)

The sufficiency follows from the definition.

Figure 4.1.2

20
THEOREM 4.3

A digraph is weak if and only if it has a spanning semi arc sequence.

Proof :

Let D = (V, A) be a weak digraph with V = {v1, v2, ..., vn}. Since D is weak, there is
a semi arc sequence, say Qi from vi to vi+1 in D for i = 1, 2, ..., n− 1.

The semi arc sequence obtained by traversing the semi arc sequences Q1, Q2, ..., Qn
in succession is a spanning semi arc sequence of D.

Conversely, let D be a digraph containing a spanning semi arc sequence, say Q.

Let v1 and v2 be two distinct vertices of D. Clearly, v1 and v2 are in Q, since Q is
spanning. The part of Q which begins at any appearance of v1 (v2) and ends at any
appearance of v2 (v1) represents a semi arc sequence from v1 to v2 (from v2 to v1) in
D. Thus there is either a semi arc sequence from v1 to v2, or from v2 to v1 in D.
Hence D is weak.

4.2 EULERIAN GRAPH

In a digraph D, a closed spanning walk in which each arc of D occurs exactly
once is called an Eulerian trail. A digraph is called Eulerian if it has an Eulerian
trail.(figure 4.2.1)

Figure 4.2.1

1d3e2f4a2b1c1 is an Eulerian trail in the above Eulerian graph.

21
THEOREM 4.4

A connected digraph D is Eulerian if and only if for each vertex ‘v’ of D

outdeg(v)=indeg(v).

Proof :

Let D be Eulerian and T be an Eulerian trail in D. Each occurance of a given point

v in T contributes one to d-(v) and one to d+(v).Since each arc of D occurs exactly
once in T, the contribution of each arc of D to d -(v) and d+(v) can be accounted in
this way .

Hence d-(v)=d+(v) ∀ v ∈ D(V,E).

Conversely let d-(v)=d+(v) ∀ v ∈ D(V,E).

Since the trivial digraph is vacuously Eulerian, let D have at least two point.
Hence every points of D has +ve indegree and outdegree. Hence D contains a
cycle Z (Since if you reach a point for the first time you can always move out).The
removal of the lines of Z results in a spanning subdigraph D1 in which again d-
(v)=d+(v) for every point v.

If D1 has no arcs, then Z is an Eulerian trail in D. Otherwise D 1 has a cycle Z,

containing the above process, when a digraph D n with no arc is obtained, we have
a partition of the arcs of D into n cycles, n ≥ 2.

Among these n cycles,take two cycles Zi and Zj having a point v in common. The
walk beginning at v and consisting of the cycles Zi and Zj in succession is closed
trail containing the lines of these two cycles. Continuing this process, we can
construct a closed trail containing all the arcs of D. Hence D is Eulerian

Principle of duality for digraphs

While changing a digraph D to its converse D′, we observe that the properties
about D get changed to the corresponding properties about D′. When D′ is changed
to D′′ = D, the original properties of D are obtained. Such type of a pair of
properties are called dual properties, (transmitter, receiver), (source, sink),
(indegree, outdegree), (isolate, isolate) and (carrier, carrier). The dual of a
statement P about a digraph is the statement P′ obtained from P by changing every

22
concept in P to its dual. For any statements P and Q for digraphs, P ⇒ Q in D is
true if and only if P′ ⇒ Q′ in D′ and for any digraph D there is a converse D′.
Therefore for every result in digraphs we get a dual result by changing every
property to its dual. This is called the principle of duality for digraphs.

23
5. TREES WITH DIRECTED EDGES

We know that a tree in undirected graphs is a connected graph without

cycles. But in case of digraphs, a structure similar to that of a tree needs absence of
cycles as well as absence of semi cycles. We have the following definition.

5.1 DIRECTED TREES

A directed tree is a connected digraph without cycles, neither directed cycles nor
semi cycles. We observe that a tree with n vertices has n−1 directed edges, and has
properties analogous to those of trees with undirected edges. A digraph whose
weak components are trees is called a forest. The given figure 5.1.1 is a tree.

Figure 5.1.1

ARBORESCENCE

A directed tree is said to be an arborescence if it contains exactly one vertex,

called the root, with no arcs directed towards it and if all the arcs on any semipath
are directed away from the root. That is, every vertex other than the root has
indegree exactly one. Arborescence is also called an out-tree.(figure 5.1.2). If the
direction of every arc in an arborescence is reversed, we get a tree called an in-tree.

Figure 5.1.2

24
THEOREM 5.1

In an arborescence, there is a directed path from the root v to every other vertex.
Conversely, a digraph D without cycles is an arborescence if there is a vertex v in
D such that every other vertex is reachable from v and v is not reachable from any
other vertex.

Proof :

In an arborescence, consider a directed path P starting from the root v and

continuing as far as possible. Clearly, P can end only at a pendant vertex, since
otherwise, we get a vertex whose indegree is two or more, which is a contradiction.
Since an arborescence is connected, every vertex lies on some directed path from
the root v to each of the pendant vertices.

Conversely, since every vertex in D is reachable from v and D has no cycle, D is a

tree. Further, since v is not reachable from any other vertex, d-(v) = 0. Every other
vertex is reachable from v and therefore indegree of each of these vertices is at
least one. The indegree is not greater than one, because there are only n−1 arcs in
D, n being the number of vertices of D.

5.2 ACYCLIC DIRECTED GRAPH

A directed graph with at least one directed circuit is said to be cyclic. A directed
graph is acyclic otherwise. Obviously, directed trees are acyclic but the reverse
implication is not true.

Figure 5.2.1

The digraph is acyclic but it is not a directed tree.

25
THEOREM 5.2

In an acyclic digraph, there exist at least one source (a vertex whose indegree is
zero) and at least one sink (a vertex whose outdegree is zero).

Proof :

Let G be an acyclic digraph. If G has no arcs, then it is obvious. Otherwise, let us

consider the directed path

vi0,ej1,vi1,ej2,…...,ejk,vik,

which has the maximum path length k. Since G is acyclic, vi0 ≠ vik. If (v,vi0) is an
arc, then one of the following is true :

• v ≠ vit for every value of t = 0,…..,k. Then,

v,(v,vi0),vi0,ej1,vi1,ej2,…..,ejk,vik

is a directed path with length k + 1.

• v = vit for some value of t. We choose the smallest such t. Then, t > 0
because there are no loops in G and

vi0,ej1,vi1,ej2,...,ejt,vit,(v,vi0),vi0

is a directed circuit.

Hence, d-(vi0) = 0. Using a similar technique, we can show that d+(vik) = 0 as well.

26
6. MATRIX REPRESENTATION OF DIGRAPHS

The matrices associated with a digraph are almost similar to those discussed
for an undirected graph, with the difference that in matrices of digraphs consist of
1, 0, −1 instead of only 0 and 1 for undirected graphs.

6.1 INCIDENCE MATRIX

The incidence matrix of a digraph with n vertices, m arcs and no self-loops is an
n×m matrix A = [aij], whose rows correspond to vertices and columns correspond
to arcs, such that

{
1 ,if v i is t h e initial vertex of e j
aij = −1,if v i is t h e terminal vertex of e j
0 , ot h erwise

Figure 6.1.1

A directed graph (figure 6.1.1) and it’s incidence matrix.

27
THEOREM 6.1

The determinant of every square submatrix of A, the incident matrix of a digraph is

1, – 1 or 0.

Proof :

The theorem can be proved directly by expanding the determinant of a square

submatrix of A.

Consider a k by k submatrix M of A. If M has any column or row consisting of all

zeros det M is clearly zero. Also det M = 0 if every column of M contains the two
non zero entries, a1 and a-1 .

Now if det M ≠ 0 (i.e., M is non singular), then the sum of entries in each column
of M cannot be zero. Therefore, M must have a column in which there is a single
non zero element that either + 1 or – 1.

Let this single element be in the (i, j)th position in M. Thus,

det M = ± 1 det Mij , where Mij is the submatrix of M

with its ith row and jth column deleted.

The (k – 1) by (k – 1) submatrix Mij is also nonsingular. Therefore it too must have

atleast one column with a single non zero entry, say, in the (p,q)th position.
Expanding det Mij about this element in the (p, q)th position.

det Mij = ±1 [determinant of a non sigular (k – 2) by (k – 2) submatrix of

M]

Repeated application of this procedure yields

det M = ± 1.

Hence the theorem.

6.2 ADJACENCY MATRIX

28
Let D be a directed graph with n vertices. The adjacency matrix A(D) of D is a n
× n matrix [aij] with

aij = {
1 if ( v i , v j ) is an arc of D
0 ot h erwise

Figure 6.1.2

A digraph (figure 6.1.2) with it’s adjacency matrix.

We have the following observations about the adjacency matrix X of a digraph D.

 X is a symmetric matrix if and only if D is a symmetric digraph.

 Every non-zero element on the main diagonal element represents a loop at
the corresponding vertex.
 The parallel arcs cannot be represented by X and therefore X is defined only
for a digraph without parallel arcs.
 The sum of each row equals the outdegree of the corresponding vertex and
the sum of each column equals the indegree of the corresponding vertex. The
number of non-zero entries of X equals the number of arcs in D

THEOREM 6.2

29
The (i,j)th entry An is the number of walks of length n from vi to vj, where A is the
adjacency matrix.

Proof :

Proof by induction on ‘n’.

From definition of A(D) the theorem holds for n=1.

Now assume that theorem holds for n-1.

Let An-1=(bij).

Hence bij= number of vi-vj walk of length n-1. ---(a)

Now An=An-1A.

Therefore (i,j)th entry of An = bi1a1j+bi2a2j+….….+bipapj. ---(b)

By definition of A(D), Akj=number of vk-vj walks of length 1.

Hence by (a) ∀ k, 1 ≤ k ≤ p,

bikakj = number of vi-vj walks of length n whose last arc is (vkvj) .

Since any vi - vj walk has one among v1vjv2vj,….vpvj as the last arc.

The R.H.S of (b) gives the number of vi-vj walks of length n.

Hence the (i,j)th entry of An is the number of vi - vj walks of length n.

30
6.3 CONNECTEDNESS AND ADJACENCY MATRIX
A digraph is disconnected if and only if its vertices can be ordered in such a way
that its adjacency matrix X can be expressed as the direct sum of two square
submatrices X1 and X2 as

This partitioning is possible if and only if the vertices in the submatrix X 1 have no
arc going to or coming from the vertex of X2.

Similarly, a digraph is weakly connected if and only if its vertices can be ordered in
such a way that its adjacency matrix can be expressed as

where X1 and X2 are square submatrices.

Form (b) represents the case when there is no arc going from the subdigraph
corresponding to X1 to the one corresponding to X2. Form (c) represents the case
when there is no arc going from the subdigraph corresponding to X 2 to the
subdigraph corresponding to X1.

Since a strongly connected digraph is neither disconnected nor weakly connected,

a digraph is strongly connected if and only if the vertices of D cannot be ordered
such that its adjacency matrix X is expressible in the form (a), or (b), or (c).

31
7. TOURNAMENTS

A tournament is an orientation of a complete graph. Therefore in a tournament each

pair of distinct vertices vi and vj is joined by one and only one of the oriented arcs
(vi, vj) or (vj, vi). If the arc (vi, vj) is in T, then we say vi dominates vj and is denoted
by vi → vj. The relation of dominance thus defined is a complete, irreflexive,
antisymmetric binary relation.

Figure 7.1.1 displays all tournaments on two, three and four vertices.

Figure 7.1.1

Definition : A triple in a tournament T is the subdigraph induced by any three

vertices. A triple (u, v, w) in T is said to be transitive if whenever (u, v) ∈ A(T) and
(v, w) ∈ A(T), then (u, w) ∈ A(T). That is, whenever u → v and v → w,

then u → w.

Definition : A bipartite tournament is an orientation of a complete bipartite graph.

A k-partite tournament is an orientation of a complete k-partite graph.

Figure 7.1.2 displays a bipartite and a tripartite tournament.

32
 Figure 7.1.2

THEOREM 7.1

If v is a vertex having maximum outdegree in the tournament T, then for every

vertex w of T there is a directed path from v to w of length at most 2.

Proof :

Let T be a tournament with n vertices and let v be a vertex of maximum outdegree

in T. Let d+(v) = m and let v1, v2, ..., vm be the vertices in T such that there are arcs
from v to vi , 1 ≤ i≤ m. Since T is a tournament, there are arcs from the remaining
n−m−1 vertices, say u1, u2,..., un-m-1 to v.

That is, there are arcs from uj to v, 1 ≤ j ≤ n−m−1 (Figure 7.1.3)

Figure 7.1.3

33
Then for each i, 1 ≤ i ≤ m, the arc from v to vi gives a directed path of length 1 from
v to vi. We now show that there is a directed path of length 2 from v to uj for each j,
1 ≤ j ≤ n−m−1.

Given such a vertex uj, if there is an arc from vi to uj for some i, then vviuj is a
directed path of length 2 from v to uj. Now, let there be a vertex uk, 1 ≤ k ≤ n−m−1,
such that no vertex vi,1 ≤ i ≤ m, has an arc from vi to uk. Since T is tournament,
there is an arc from uk to each of the m vertices vi. Also, there is an arc from uk to v
and therefore d+(uk) ≥ m+1. This contradicts the fact that v has maximum outdegree
with d+(v) = m. Thus each uj must have an arc joining it from some vi and the proof
is complete by using the directed path vviuj.

Remark :

 Let T be a tournament with n vertices and let v be any vertex of T. Then

T − v is the digraph obtained from T by removing v and all arcs incident with
v. Clearly, any two vertices of T −v are joined by exactly one arc, since these
two vertices are joined by exactly one arc in T. Thus T −v is again a
tournament.

Definition : A directed Hamiltonian path of a digraph D is the directed path in D

that includes every vertex of D exactly once.

The following result, due to Redei, shows that a tournament contains a direct
Hamiltonian path.

34
THEOREM 7.2 (Redei)

Every tournament T has a directed Hamiltonian path.

Proof :

Let T be a tournament with n vertices. We induct on n. When n = 1, 2, or 3, the

result is trivially true.(figure 7.1.4)

Figure 7.1.4

Let n ≥ 4. Assume that the result is true for all tournaments with fewer than n
vertices. Let v be any vertex of T. Then T −v is a tournament with n−1 vertices and
by induction hypothesis has a directed Hamiltonian path, say P = v1v2 ...vn-1.

In case there is an arc from v to v1, then P1 = vv1 v2 .....vn-1 is a directed Hamiltonian
path in T. Similarly if there is an arc from vn-1 to v, then P2 = v1 v2 ...vn-1v is a
directed Hamiltonian path in T. If there is an arc from v1 to v.

Let vi+1 be the first vertex on P for which the arc (v,vi+1) is in T, so that the vertex
vi does not have this property. Then there is an arc from vi to v and an arc from v to
vi+1, as shown in Figure 7.1.5. Thus Q = v1v2 .....viv vi+1 vi+2 .....vn-1 is a directed
Hamiltonian path in T.

Figure 7.1.5

35
Definition : A directed Hamiltonian cycle in a digraph D is a directed cycle which
includes every vertex of D. If D contains such a cycle, then D is called
Hamiltonian.

The next two results are due to Camion.

THEOREM 7.3 (Camion)

A strongly connected tournament T on n vertices contains cycles of length 3,4,...,n.

Proof :

First we show that T contains a cycle of length three. Let v be any vertex of T. Let
W denote the set of all vertices w of T for which there is an arc from v to w. Let Z
denote the set of all vertices z of T for which there is an arc from z to v. We note
that W Z = ∅ , since T is a tournament.

Since T is strongly connected, W and Z are both nonempty. For, if W is empty, then
there is no arc going out of v, which is impossible because T is strongly connected
and the same argument can be used for Z. Again, because T is strongly connected,
there is an arc in T going from some w in W to some z in Z. This gives the directed
cycle vwzv of length 3.(figure 7.1.6)

Figure 7.1.6

Now induct on n. Assume T has a cycle C of length k, where k < n and k ≥ 3 and let
this cycle be v1 v2 ... vk v1. We show that T has a cycle of length k+1.

36
Let there be a vertex v not on the cycle C, with the property that there is an arc
from v to vi and an arc from vj to v for some vi, vj on C. Then there is a vertex vi on
C with an arc from vi-1 to v and an arc from v to vi. Therefore, C1 = v1v2 ...vi-1 vvi
vi+1 ...vk v1 is a cycle of length k+1.(figure 7.1.7)

Figure 7.1.7

If no vertex exists with the above property, then the set of vertices not contained in
the cycle can be divided into two distinct sets W and Z, where W is the set of
vertices w such that for each i, 1 ≤ i ≤ k, there is an arc from vi to w and Z is the set
of vertices z such that for each i, 1 ≤ i ≤ k, there is an arc from z to vi. If W is empty
then the vertices of C, and the vertices of Z together make up all the vertices in T.
But, by definition of Z, there is no arc from a vertex on C to a vertex in Z, a
contradiction, because T is strongly connected. Thus W is nonempty. A similar
argument shows that Z is nonempty. Again, since T is strongly connected, there is
an arc from some w in W to some z in Z. Then C1 = v1 w z v3 v4 ...vk v1 is a cycle of
length k+1(see figure 7.1.8). This completes the proof.

Figure 7.1.8

37
THEOREM 7.4 (Camion)

A tournament T is Hamiltonian if and only if it is strongly connected.

Proof :

Let T have n vertices. If T is strongly connected, then by above theorem, T has a

cycle of length n. Such a cycle is a Hamiltonian cycle, since it includes every
vertex of T. Hence T is Hamiltonian.

Conversely, let T be Hamiltonian with Hamiltonian cycle C = v1 v2 ...vn v1. Then

given any vi, vj with i ≥ j, in the vertex set of T, vj vj+1...vi is a path P1 from vj to vi
while vi vi+1 ….....vn-1 vn v1 …....vj-1 vj is a path P2 from vi to vj (figure 7.1.9). Thus
each vertex is reachable from any other vertex and so T is strongly connected.

Figure 7.1.9

38
8. CONCLUSION

There is so much to digraph theory that it is possible to write an entire book on

the subject. For the most part of this dissertation we emphasized the properties
and concepts of digraphs which related to graphs. Thus we began with basic
definitions and isomorphism, continued by developing three different kinds of
connectedness: strong, unilateral, and weak. After presenting the Directional
Duality Principle, we study matrices related to digraphs and the properties of
that matrices. We closed with a brief description of tournaments.

39
 REFERENCE

1. S.Kumaravelu & Suseela Kumaravelu : Graph Theory

2. C.Vasudev : Graph Theory With Applications
3. Hararay : Graph Theory
4. Narsigh Deo : Graph Theory With Applications To Engineering And
Computer Science
5. S.Arumugam & S.Ramachandran : Invitation To Graph Theory
6. Jorgen Bang-Jensen & Gregory Gutin : Digraphs Theory ,
Algorithms And Applications

40