Unit-2 Degree Sequences and Graph Representation
Unit-2 Degree Sequences and Graph Representation
Unit-2 Degree Sequences and Graph Representation
UNIT 2
2.1 INTRODUCTION
In Unit 1, you saw graphs as models, and the basic concepts of graphs theory
including neighbourhood and degree; walks, paths and cycles; connectedness
and components. In this unit, we shall introduce some more essential
concepts to understand the structure of graphs.
In Section 2.2, you will see how to define the ‘distance’ between two vertices
of a graph. One particular application of distance, you will see is in
characterising bipartite graphs. In this section, you will also see, the related
concepts, such as radius and diameter of a graph. In Section 2.3, we shall
introduce two important graphs namely, Petersen graph and Hypercube graph,
and look at their properties. Section 2.4 deals with the representation of
graphs through matrices.
In Section 2.5, we shall discuss the concept of isomorphism, which will help
you to decide when two graphs are ‘isomorphic’. Finally, Section 2.6 deals with
the notion of degree sequence, and graphic sequence.
Objectives
43
Block 1 Fundamentals of Graph Theory
After studying this unit, you should be able to:
• define the distance between two vertices of a graph;
• find the radius and diameter of a graph;
• describe the Peterson and Hypercube graphs;
• represent a graph by adjacency matrix or incidence matrix;
• what isomorphism is, and when two graphs are isomorphic;
• differentiate between a degree sequence and a graphic sequence;
• describe and use the Havel-Hakimi Theorem to determine whether a
given sequence of non negative integers is a graphic sequence or not.
2.2 DISTANCE
In this section we shall study the concept of distance between two vertices of a
graph. We shall also see the related concepts such as the diameter and the
radius of a graph.
Definition: Let G be a graph, and u, v∈V (G) . The distance between u and
Note that v , denoted by d G (u, v) , is the length of a shortest (u , v) -path. If there is no
dG (u, v) = ∞ iff (u , v) -path in G then d G (u, v) = ∞ .
u and v lie in
different components. For example consider the graph given below.
When there is no
confusion, we shall a
write d (u , v) in place
of d G (u, v) . v
u w
x z
y
b c
Fig. 1
You can compute the distances between other pairs of vertices similarly. You
will find that the maximum distance between any two vertices of this graph is
3.
From other mathematics courses, such as real analysis or topology, you can
recall that a distance function d on a set S is a nonnegative function that
satisfies for all x, y, z ∈ S :
1) d ( x, y ) = 0 ⇔ x = y
2) d ( x, y ) = d ( y , x )
3) d ( x, z ) ≤ d ( x, y ) + d ( y, z )
44
Unit 2 Degree Sequences and Graph Representation
The first two properties hold trivially for the distance defined in graphs. Let us
verify the third property, which is called the triangle inequality, in the example
below.
Solution: Let P, P' and P' ' be three shortest ( x, y ) -path, ( y , z ) -path, and
( x, z ) -path, respectively. Then
length ( P) = d ( x, y), length ( P' ) = d ( y, z ), length ( P' ' ) = d ( x, z )
Recall, from Unit 1, that a bipartite graph can be written as a union of two
nonempty disjoint independent sets. The concept of distance helps us get
another characterisation of bipartite graphs in terms of even cycles.
Theorem 1: A graph is bipartite if and only if all its cycles are even.
Conversely, let us suppose that G contains only even cycles. We first assume
that G is connected. Let u be a vertex of G . Define
X = {v ∈ V (G ) : d (u , v) is even. } and Y = {v ∈ V (G ) : d (u , v) is odd. }.
u
x
v y
P′
Now look at the cycle C formed by the (v, x) -path along P , the edge xy , and
the ( y, v) -path along P' . We have
45
Block 1 Fundamentals of Graph Theory
length(C ) = [length( P ) − d (u, v )] + [length( P ' ) − d (u, v )] + 1
= length( P ) + length( P ' ) - 2d (u, v ) + 1
Recall that the diameter of a set in a metric space is the maximum distance
between any two points of the set. In the same spirit, we can define the
diameter of a graph.
Since the maximum distance between any two vertices of the graph in Fig. 1
above is 3 , the diameter of the graph is 3 . Now let us consider the following
theorem that relates the diameter of a graph with its complement.
Case i): x, y ∈ N G (u )
In this case d G ( x, y ) ≤ d G ( x, u ) + d G (u, y) = 2.
Thus in all the cases, we have shown that d G ( x, y) ≤ 3. Since x and y are
arbitrary, we have max d G ( x, y ) ≤ 3. Therefore, diam (G ) ≤ 3.
x , y∈V ( G )
Just like the diameter, we can define the radius of a graph. See the definition
below.
Note that
v7 v8
diam (G ) = max ε G (v )
G v∈V ( G )
Here we can see that the eccentricity of v1 is 4 , i.e. ε (v1 ) = 4 (This is because
v6 is the farthest vertex from v1 at a distance of 4 ). Similarly we have
ε (v2 ) = 3, ε (v9 ) = 4, ε (v7 ) = 3, ε (v8 ) = 3, ε (v5 ) = 3, ε (v4 ) = 3, ε (v6 ) = 4, ε (v3 ) = 2.
Thus, rad (G ) = 2 and diam (G ) = 4 .
5 4
4 3 5
3 4
4 3 4
Here the minimum eccentricity is 3 , and the maximum eccentricity is 5 . So,
rad( G ) = 3 and diam( G ) = 5 .
***
Thus, unlike geometry or analysis, the diameter of a graph need not equal
twice the radius. Instead, we have the following relation between the radius
and the diameter of a graph.
Proof: The inequality rad (G ) ≤ diam (G ) follows from the definition. We have
to prove the second inequality diam (G ) ≤ 2 rad (G ) . So consider two vertices
u and v in G , with d (u, v) = diam (G ). Let w be a vertex in G with
ε (w) = rad (G ). Then
diam (G ) = d (u , v ) ≤ d (u , w) + d (v, w)
≤ ε ( w) + ε ( w) = 2 rad (G )
v12 v10 v6
v8 v5
v13
v7
v1 v2 v3 v4
E3) Give an example of a graph, different from any of the graphs given in
E2) such that
i) its radius and diameter are equal.
ii) its radius is half its diameter.
E4) Does there exist a graph with radius 2 and diameter 5 ? Justify.
Petersen graph
Definition: The Petersen graph is a graph whose vertices are the 2-element
subsets of a 5-element set {1, 2, 3, 4, 5}, and whose edges are pairs of disjoint
2-element subsets of {1, 2, 3, 4, 5} .
5
Note that a 2 -element subset of {1, 2, 3, 4, 5,} can be chosen in ways, i.e.,
2
in 10 ways. So there are 10 vertices in the Petersen graph. Now consider any
vertex {u , v} . There are three remaining elements in {1, 2, 3, 4, 5} . Of these
3
three elements, there are = 3 ways to pick a 2 -element subset. This
2
means, there are exactly three 2 -element subsets disjoint from {u, v}, and
hence {u , v} is adjacent to all of them. Thus every vertex in the Petersen graph
has degree 3 , that is, it is 3 -regular. Hence the number of edges in the
3 ×10
Petersen graph is = 15 .
2
35
13
45 34
25
14
24
23 15
Petersen Graph
Look at any two nonadjacent vertices in the graph above. How many
neighbours they have in common? Exactly one, right? This is guaranteed by
the following proposition.
Recall that the girth of a graph is the length of a shortest cycle in it.
Proof: By definition, the Petersen graph has no loops and parallel edges. This
means it has no 1-cycle and 2-cycle. A 3-cycle would require 3 pairwise-
disjoint 2-sets, which cannot occur among 5 elements. A 4-cycle in the
absence of 3-cycles would require nonadjacent vertices with 2 common
neighbours that is also not possible due to Proposition 1. The vertices 12, 34,
25, 14, 35 yield a 5-cycle, so the girth of the Petersen graph is 5.
Hypercube Graph
Q1 Q2 Q3
49
Block 1 Fundamentals of Graph Theory
Since each entry in a vertex (a1 , a2 ,Kak ) can be 0 or 1 , Qk has 2 k vertices.
Solution: Let X be the set of all vertices that contain an odd number of
1s, and Y the set of all those vertices that contain an even number of 1s.
Then ( X , Y ) is a bipartition of Qk .
***
v1 v2 A(G )
G
v1
v5 v2
v4 v3
***
v3
v1 v7
v4
v5 v6
i) Find A(G ) .
Solution:
i) We have
0 1 0 0 0 0 0
1 0 3 0 0 0 0
0 3 0 1 1 0 1
A(G ) = 0 0 1 0 0 0 1
0 0 1 0 0 1 0
0 0 0 0 1 0 1
0 0 1 1 0 1 2
ii) We can see that aij = a ji for all i and j . Hence A(G ) is symmetric.
iii) The sum of the entries of A(G ) in each row is given below:
row 1 2 3 4 5 6 7
sum 1 4 6 2 2 2 5
The column sums are also the same.
iv) Let us compare the degree of vertex vi with the sum in row i , for each
i = 1,2,K7, what we find is that
7
d (vi ) = sum of the entries of A(G ) in row i = aij
j =1
***
In the example above, you saw that the sum of the entries in a row equals the
degree of the corresponding vertex. This is indeed the case for all multigraphs.
To see, consider a vertex v in a multigraph G . Then the quantity auv
u∈V (G )
represents the number of edges incident on v , with each loop counted twice.
This is exactly the degree of v . That is,
d (v ) = a
u∈V ( G )
uv
Now in the equation above, let us take the sum over all v ∈V (G ) . We get
v∈V ( G )
d (v ) = a
v∈V ( G ) u∈V ( G )
uv ,
which implies 2m (G ) = a
u ,v∈V ( G )
uv , or
1
m (G ) = auv .
2 u ,v∈V (G )
This shows that we can compute the number of edges in a multigraph using its
adjacency matrix.
auv = 1 ⇔ u ↔ v ⇔ gcd(u, v) = 1
Hence the adjacency matrix of G is
0 1 0 1 1 0 1 0
1 0 1 1 1 0 1 1
0 1 0 1 1 0 1 0
1 1 1 0 1 1 1 1
A(G ) =
1 1 1 1 0 1 1 1
0 0 0 1 1 0 1 0
1 1 1 1 1 1 0 1
0 1 0 1 1 0 1 0
1 42
Therefore, we have m (G ) = auv =
2 u ,v∈V (G ) 2
= 21 .
***
1, if e j is incident on vi
bij =
0, othewise.
v5 e4 v4
e8 e3
v6 e6 e9 e2 e5 v3
e7
v1 v2
e1
In the incidence matrix B(G ) of a graph G , you can see that the sum of all
the entries in a row represents the degree of the corresponding vertex.
However, each column has exactly two 1s corresponding to the endpoints of
the edge representing the column.
v1 v2 v3
v5
is given below:
0 1 0 0 0
1 0 1 0 0
A = 0 1 0 1 1
0 0 1 0 1
0 0 1 1 0
1 0 1 0 0 0 2 0 1 1
0 2 0 1 1 2 0 4 1 1
A2 = 1 0 3 1 1, A3 = 0 4 2 4 4
0 1 1 2 1 1 1 4 2 3
0 1 1 1 2 1 1 4 3 2
Now let us see what we can infer from the (i, j ) th entry of A2 . For example,
the (1,1) th entry is 1 , which is the number of (v1 , v1 ) -walks of length 2 in the
graph. You can see there is exactly one such walk, that is, (v1 , v2 , v1 ) .
Now look at the (1,2) th entry which is 0 . In the graph there are no (v1 , v2 ) -walk
of length 2 . Similarly, (3,3) th entry is 3 , and all the (v3 , v3 ) -walks of length
2 are
(v3 , v2 , v3 ), (v3 , v5 , v3 ), (v3 , v4 , v3 ).
You can look at other entries. What you will observe is that in A 2 , (i, j ) th entry
represents the number of (vi , v j ) -walks of length 2 .
54
Unit 2 Degree Sequences and Graph Representation
3
Now let us look at the entries of A .
Thus in A3 , (i, j )th entry represents the number of (vi , v j ) -walks of length 3 .
Now can you guess what the (i, j ) th entry of Ak would represent? The answer
is given in the following theorem.
( k −1)
Note that air arj is nonzero only when arj ≠ 0, i.e. when arj = 1 . But this
( k −1)
means there is an edge between v r and v j . Thus air arj is the number of
(k )
(vi , v j ) -walks of length (k − 1) + 1 = k through v r . Hence aij is the number of
all (vi , v j ) -walks of length k .
By the Principle of Mathematical induction the result holds for all k ≥ 1 .
The following result tells us more about the usefulness of adjacency matrices.
M = A + A2 + K + An−1
is nonzero.
v4
v5
v1 v3
v2
56
Unit 2 Degree Sequences and Graph Representation
20 11 20 4 11 26 51 26 40 51
* * * * * * * * * *
A 4 = * * * * * , and A5 = * * * * *
* * * * * * * * * *
* * * * * * * * * *
( 4) ( 5)
We find that a13 = 20 and a13 = 26, which are the number of (v1 , v3 ) -walks
of length 4 and 5, respectively.
***
E8) Write down the adjacency and incidence matrices of the following graph:
e6 v6 e5
v1 v5
e7 e8
e1 e4
e9
v2 v4
e2 e3
v3
E12) Find the number of walks of length 4 between any two nonadjacent
vertices of the following graph:
57
Block 1 Fundamentals of Graph Theory
In the next section, we shall talk about the concept of isomorphism which tells
us whether two graphs are structurally same or not.
t r
u x
p q
y
G1 G2
Let us resume our discussion on the graphs G1 and G2 given above. Let
φ : V (G1 ) → V (G2 ) be defined as φ ( p ) = w, φ (q ) = v, φ (r ) = u, φ (s) = y, φ (t ) = x.
Now for each edge in G1 , we verify the statement (1).
If you have to prove that two graphs G1 and G2 are isomorphic you have to
use the definition, there is no alternative. An isomorphism from G1 to
G2 preserves the adjacency relation between G1 and G2 . This implies, G1
and G2 must have the same order, the same number of components, the
same number of triangles, the same girth, the same radius, the same
diameter, etc. In short, we can say that
59
Block 1 Fundamentals of Graph Theory
G1 ≅ G2 If G1 has a property P , then G2 also has the property P.
Can you write down the contrapositive of the statement above? The
contrapositive is
Here, P can be any property a graph may have. Let us look at some examples
of nonisomorphic graphs.
Example 10: Show that the graphs G1 and G2 given below are nonisomorphic.
5 y
1 2 3 4 u v w x
G1 G2
Solution: Since G1 has only one triangle and G2 has two triangles, G1 ≅/ G2 .
***
v e
q h
r u
a d
s t
b c
Solution: The graph on the left is bipartite, and so has no odd cycles.
However, the graph on the right has a 5 -cycle. Hence, they are not
isomorphic.
***
By now you must have understood that an isomorphism between two graphs
tells us that they are the same, a copy of each other, but labelled or drawn
differently. Recall that the complement of a graph is obtained by removing the
existing edges from the graph and adding the missing edges. This process
may produce a copy of the original graph. What will you call such a graph?
“Self-complementary”? Let us see the definition below.
3
5 3 5
4 4
60
C5 C5
Unit 2 Degree Sequences and Graph Representation
Note that C 5 is the cycle (1, 3, 5, 2, 4,1) . Here E (C5 ) = {12, 23, 34, 45, 51}, and
E (C 5 ) = {13, 35, 52, 24, 41} . We have one-one onto mapping Two unlabelled
φ : {1,2,3,4,5, } → {1,2,3,4,5} defined as graphs are
isomorphic if their
φ (1) = 1,φ (2) = 3,φ (3) = 5,φ (4) = 2,φ (5) = 4 . labelled versions are
isomorphic.
We leave it for you to check that for all u , v ∈ {1, 2, 3, 4, 5}
uv ∈ E (C5 ) if and only if φ (u )φ (v) ∈ E (C 5 ) .
i)
ii)
G1 G2
12
35
45 13
34
25
14
24
23 15
i) ii)
61
Block 1 Fundamentals of Graph Theory
E16) Let G be a self-complementary graph. Let H 1 be the graph obtained by
joining each vertex of G to the two endpoints of P4 . Let H 2 be the
graph obtained by joining each vertex of G to the two internal vertices of
P . Show that both H 1 and H 2 are self-complementary.
In this graph there are two vertices of degree 4 , four vertices of degree 3 ,
three vertices of degree 2 , and two vertices of degree 1 . Hence its degree
sequence is (4, 4, 3, 3, 3, 3, 2, 2, 2,1,1) . Of course, every multigraph has exactly
one degree sequence. However, two multigraphs may have the same degree
sequence. For example, the graph G1 and the multigraph G2 given below
have the same degree sequence (3, 2, 2,1) .
G1 G2
v2
v5 f
v3 b
e
v4 v6 c
v7
d
G1 G2
Did you observe that the sum of the terms in each of the degree sequences
found above is even? This is because of the following theorem.
n
Conversely, let d
i =1
i be even. We shall construct a multigraph with vertices
n
v1 , v2 K, vn such that d (vi ) = d i for all i ∈{1,2,K, n} . Note that d
i =1
i is even
implies that the odd values in (d1 , d 2 , K , d n ) occur even number of times. So
the set S = {vi : d i is odd.} contains an even number of vertices, which can be
paired in any order. Draw an edge between every such pair. The remaining
di −1
degree at each vertex in S is even, and hence we can draw loops on
2
each of these vertices. The vertices not in S already have even degree, so on
di
each of these vertices draw loops, to get a multigraph.
2
The proof of Theorem 6 not just gives a characterisation of degree sequences
of multigraphs, but provides a method as well for constructing a multigraph
with the given degree sequence. Let us consider the following example.
v3
v1 v2
v4
***
Let us now shift our attention to the degree sequences of graphs only. To
distinguish between the degree sequence of a graph with that of a multigraph,
you need the following definition.
For example, the sequences (1, 2, 3, 4, 5, 3) and (1, 2, 3,3, 4, 5) are graphic,
because both can be permutated in the descending order as
(5, 4, 3, 3,2,1), which is the degree sequence of the following graph.
Note that the sequence (4, 4,1,1,1,1) is a degree sequence of the following
multigraph.
But (4, 4, 1,1,1,1) is not graphic. That is, there exists no 6 -vertex graph with
(4, 4,1,1,1,1) as its degree sequence. To prove our claim, let us assume, for a
moment that there is a 6 -vertex graph with degree sequence (4, 4,1,1,1,1) . Let
v1 , v2 , v3 , v 4 , v5 , v6 be its vertices. We draw these vertices on a circle, as you
can see below.
64
Unit 2 Degree Sequences and Graph Representation
v1 v2
v6 v3
v5 v4
The graph has two vertices of degree 4 , say v1 and v 2 . If v1 and v 2 are
nonadjacent, then both must be adjacent to the rest four vertices, leaving no
vertex with degree 1 . This is not possible. On the other hand, if v1 and v 2 are
adjacent, then both must have at least two common neighbours. This means,
there would be at most two vertices of degree 1 , which again is not possible.
Thus (4, 4,1,1,1,1) is not a graphic sequence.
G1 G2
d
i =1
i is even. However, this is not sufficient. A necessary and sufficient
condition was given by Havel and Hakimi, which we state in the following
theorem.
We know that there can be several graphs with the same degree sequence.
Let G be a graph with vertices v1 , v2 ,K, vn such that d (vi ) = d i for i = 1,2,K, n .
We have to choose G in such a way that the set S = {v2 , v3 ,K, vd1+1} is the
65
Block 1 Fundamentals of Graph Theory
neighbourhood of v1 . But the question is - can we do so? We use
contradiction to show that this is possible. So, let us start with the assumption
that this is not possible. Let G be chosen among all graphs with
d (vi ) = d i such that v1 is adjacent to as many vertices in S as possible. By
definition, some vertex v j in S is not a neighbour of v1 . But since d (v1 ) = d1 ,
there must be some neighbour v k of v1 outside S . Since d j ≥ d k , there exists
a vertex v r such that v j vr ∈ E (G) but vk vr ∉ E (G ).
Now, form a new graph G ' from G by removing the edges v1vk and v j vr , and
adding two new edges v1 v j and vk vr . (See the picture below.)
vr
vr
S vj S vj
vk vk
v1
v1
G G′
Now let us relabel the vertices of G as {v1 ' , v2 ' ,K, v' n−1 } such that
d (v1 ' ) = d 2 − 1, d (v2 ' ) = d 3 − 1,K , d (v' d1 ) = d d1 +1 − 1,K, d (v 'n −1 ) = d n .
Add a vertex vn ' in G such that N (v'n ) = {v1 ' , v2 ' ,K, vd1 '}. Let G ' be the
graph so obtained .
Clearly d (v 'n ) = d1 and d (vi ' ) = d i +1 , ∀i = 1,K, n − 1 .
Thus d = (d1 , d 2 ,K d n ) is the degree sequence of G '. Hence d is graphic.
Note that d ' ' ' ' is not graphic, as it has negative entries. Using the Havel-
Hakimi Theorem, d ′′′ is not graphic, which implies d ' ' is not graphic, which
implies d ' is not graphic, which implies d is not graphic.
Remark 1: To show that a sequence is graphic or not, you can stop as soon
as you find that the current sequence is graphic or not.
67
Block 1 Fundamentals of Graph Theory
To obtain a graph corresponding to sequence d = (6, 5, 4, 4, 3, 2, 2, 2, 2 ) , add a
vertex adjacent to the six vertices with degrees 4,3,3,2,1,1 in the graph above.
The resulting graph is:
***
You must have noted that the process shown in Example 16 above does not
produce a unique graph.
E22) Draw two nonisomorphic graphs with degree sequence (2, 2, 2, 2, 2,2) .
E23) Draw a graph having the given properties or explain why no such graph
exists.
i) multigraph with four vertices of degree 1,1, 2 and 3 .
ii) multigraph with four vertices of degree 1,1, 3 and 3.
With this we come to the end of this section, and also to this unit. We shall
now summarise the important points discussed in this unit.
2.7 SUMMARY
In this unit, we have covered the following points:
7. Havel-Hakimi Theorem.
Case 1: n is odd
n −1 n −1
In this case we have for all 1 ≤ i ≤ , d (vi , vi +( n−1) / 2 ) = . Thus
2 2
n −1
for 1 ≤ i ≤ , the maximum distance of vi from any other
2
n −1 n −1
vertex is , i.e., ε (vi ) = . Since the distance function is
2 2
n −1 n −1
symmetric, for 1 ≤ i ≤ , we also have d (vi +( n−1) / 2 , vi ) = .
2 2
n −1 n −1
This implies for ≤ i ≤ n − 1, ε (vi ) = . Thus
2 2
n −1
diam (Cn ) = rad (Cn ) = .
2
69
Block 1 Fundamentals of Graph Theory
n −1
2 , if n is odd
diam (Cn ) = rad (C n ) =
n if n is even.
2 ,
n
Hence rad ( Pn ) = − 1.
2
iv) We know that the maximum distance between any two vertices of
K m ,n is 2. Hence diam ( K m , n ) = 2. Also we know that every vertex
has eccentricity 2. Therefore, rad ( K m , n ) = 2.
E3) i) One such graph is given below, with radius and diameter equal to
2.
ii) One such graph is given below, with radius 2 and diameter 4.
E5) The Petersen graph is not bipartite, because it contains a 5-cycle, which
is odd.
E6) Note that any two nonadjacent vertices of the Petersen graph have a
common neighbour. Hence, its diameter is 2. To find the radius, let us
compute the eccentricity of a vertex u. There are 3 vertices adjacent to
it. The remaining 6 vertices are nonadjacent to u , and hence each of
them has a common neighbour with u . Therefore, ε (u ) = 2. Since u is
arbitrary, every vertex has eccentricity 2. Thus the radius of the graph is
70 2.
Unit 2 Degree Sequences and Graph Representation
E7) A drawing of Q4 is given below.
E8) Let us take the vertex and edge labellings in the order
V (G ) = {v1 , v2 , v3 , v4 , v5 , v6 }, and E (G ) = {e1 , e2 , e3 ,K , e9 }. Then the
adjacency and incidence matrices are as follows.
0 1 0 0 0 1 1 0 0 0 0 1 0 0 0
1 0 1 1 0 1 1 1 0 0 0 0 1 0 1
0 1 0 1 0 0 0 1 1 0 0 0 0 0 0
A(G ) = , B (G ) =
0 1 1 0 1 1 0 0 1 1 0 0 0 1 1
0 0 0 1 0 1 0 0 0 1 1 0 0 0 0
1 1 0 1 1 0 0 0 0 0 1 1 1 1 0
E9) Since A is the adjacency matrix, we know that aij = 1 iff vi ↔ v j iff
v j ↔ vi iff a ji = 1. Therefore, aij = a ji for all i, j. Hence A is
k
symmetric. Assume that A is symmetric for some k . Then aij( k ) = a (jik )
n n n
for all i, j. Now aij( k +1) = air( k ) arj = ari( k ) a jr = a jr ari( k ) = a (jik +1)
r =1 r =1 r =1
( k +1)
Thus A is also symmetric. Hence by the Principle of Mathematical
Induction, A k is symmetric for all k ≥ 1.
13 13 0 13
13 13 0 13
=
0 0 3 0
13 13 0 13
We note that the (1, 3) rd entry of M is 0. Hence the multigraph is
disconnected.
v1 v2
v3
v5 v4
v4 v3
v1
v2
v4
v3
8 9 v4
G2
G1
2 4
2 4
3 3
G G
Let us define φ : V (G ) → V (G ) by
φ (1) = 2, φ (2) = 5, φ (3) = 3, φ (4) = 1, φ (5) = 4
1 1
2
2
5
5
3 4 3 4
G G
G G
P4 P4
H1 H1
P4 P4
75
Block 1 Fundamentals of Graph Theory
d ′ = (4, 3, 3,1,1,1,1)
d ′′ = (2, 2,1,1, 0, 0)
d ′′′ = (1,1, 0,0, 0)
Since d ′′′ is graphic, d is graphic.
iv) Here we find that d ′′′ = (1, 0, 0, 0, − 1), which is not graphic. Hence d
is not graphic.
E20) False. Because, for example we know that (1,1) is graphic but (2, 2) is
not.
E21) Since (d1 , d 2 ,K, d n ) is graphic there exists an m -vertex graph, say G1 ,
with V (G1 ) = {v1 , v2 , K , vm } where d (vi ) = d i for 1 ≤ i ≤ m. Likewise,
since (d m+1 , d m+ 2 ,K, d n ) is graphic, there exists a graph, say G2 , with
V (G2 ) = {vm+1 , vm+ 2 ,K, vn }, where d (vi ) = d i for m + 1 ≤ i ≤ n. Then the
graph G1 ∪ G2 with V (G1 ∪ G2 ) = {v1 , v2 , K , vn } realises to the sequence
(d1 , d 2 ,K, d n ). Hence (d1 , d 2 ,K, d n ) is graphic.
G1 G2
iii) Here d = (3, 3,1,1). So, d ′ = (2, 0, 0), which is not graphic. Hence d
is not graphic, and so no such graph exists.
76