Kesar want 202C

Name Rishabh DO 23

ASSIGNMENT
GRAPH THEORY
Assignment
-2

Quest
I

Above statement is TRUE

pay
.
.

Tree is a
acyclic connected graph .

So vets take graph G cu , E) be a

great with
be the
E K) LVG ) and let G ,
fer .
- .
- - -

Gµ
components of G .

tree
Assumption No Gi is a

i
VIG : ) for all
I
(Gi )
k
implies E >
- -
- i

so this

ECI ) C VEI l is not connected

E)
-

A
graph I put m

ELI ) VEI I is a tree

8) A graph
-
=

I with

E) A I with ELI ) s = VE ) is not acyclic

graph
above I , II III statement for all El, z . .
-

k
for the

contradiction when E ( g) > UCG)

glues
.

Therefore assumption is
our
wrong
.

than needles has

Every Graph with fewer edges a

component that's tree

 Given that Graph G is a tree so E(G)=n-1 where n= number of vertices. We also know that Tree is
also acyclic. x,y are the two vertices of the Graph G. We are creating a copy of xy edge . So lets now
create new graph(G’) with copied edge xy

A
case I G G
Tssefore copying my After copying my
•-•→
& ••#f
n
y
I
case I
Te Y n
• •
Y
e• •

Casein

Y.IE/I
•
•

.
: .

n
•
.
•
R o
•

Let n be the number of vertices between x & y .So Graph(G’) has n vertices & n-1 edges at starting.
After copying the edge xy .

Total number of Vertices = n + n-2(since x and y would be common so subtrating 2 from n)= 2n-2
Total No. Of edges= n-2+n-2=2n-2

Since number Of vertices and number of edges are same in Graph(G’) hence our graph(G’) has
exactly one cycle
 We know that, the distance between any two vertices - x, y in a graph is the shortest path between
them. Now consider a connected graph G, let the x and y be two adjacent vertices of graph G. Let
z some other vertex of graph G and d(x,z) = l. Then while finding d(y,z) will be either l, if we don’t
go through edge xy, or it’ll be l+1, if we go through edge xy. These are the only two possibilities.
Hence, absolute difference between d(x,z) and d(y,z) is either 0 or 1. This proves the above
statement.

Consider a graph G and it’s square as below:

a b a b
• ⑧ B ⑧

i °

I
• •

O D ⑧ C O D ⑧ C
• op

g j
n
g
n
j
• O B @ B B
d d
f e
f e

a a

As we know that, diameter of a graph is the maximum distance between a pair of vertices in the
graph, so for graph G, the distance between vertices i and e, which is 4, is maximum amongst all
pair of vertices.
Therefore, d(i,e) = 4 is the diameter of graph G. Now coming to G’s square i.e. G2, the distance
between vertices i and e is 2 and it is maximum amongst all pair of vertices. Therefore, the
diameter of G2 is 2, which is diam(G)/2
 Let the graph G has n vertices and has no cycles. The number of edges in G is (n-1). Let G is
disconnected. Therefore there exist at least two components of G,say X and Y. Each of the components is
acyclic as G is acyclic. Now to make the graph G connected, we need to add one edge between one end
vertex of X and one end vertex of Y.
Now the number of edges in G = (n – 1)+1 = n.
Now, G is connected graph and has no cycle with n vertices and n edges. But this is impossible because
the connected acyclic graph is a tree with n vertices and (n-1) edges. Therefore our assumption is wrong.
So the graph G with n vertices, (n-1) edges and no cycle is connected. Hence the given statement is
proved.

We know that a tree T is a connected graph with no cycles. Therefore the tree has no odd cycles.
Moreover the tree is non-trivial, so it has at least one edge. Therefore we can denote this tree with the help
of a bipartite graph. Now we split the vertices V of the tree into two sets of vertices 𝑉1 and 𝑉2 such that
V=𝑉1 U 𝑉2. Since T has no odd cycle, so there is no edge between any two vertices of the same set 𝑉1 or
𝑉2. Therefore 𝑉1 and 𝑉2 are independent and while finding the independent sets, we can include all the
vertices of 𝑉1 and 𝑉2 Therefore 𝑉1 and 𝑉2 are maximal independent sets. Similarly if we aplit the vertices
into more sets, we will have more than two maximal independent sets.
Now let’s consider the star graph.
• •

• Be @

• *

Fig: star graph

We split the vaertices of the star graph into two vertex sets say 𝑉1 and 𝑉2 where 𝑉1 contains the center
vertex and 𝑉2 contains all other vertices. Clearly there is no edge between any two vertices of the same
set
𝑉1 or 𝑉2. No other combination of vertices would result this scenario. Therefor 𝑉1 and 𝑉2 are the only
possible vertex sets and they are maximal independent. Hence we can conclude that for a Star graph there
are exactly two maximal independent sets.
Let us consider a tree T with n vertices. For n = 1, T is a star graph. For n > 1, let the tree T has a leaf x
and let y be its neighbor. The independent sets in T consist of the independent sets in T - x and all sets
formed by adding x to an independent set in T - x - y. By the induction hypothesis, the first type is
maximized (only) when T - x is a star. The second type contributes at most 2n-2 sets, and this is achieved
only when T -x - y has no edges, which requires that T - x is a star with center at y. Thus both
contributions are maximized when (and only when) T is a star with center y. Hence the proof.

Suppose e is a bridge in G and T a spanning tree on G not containing e. Then, since T is a tree, it
must be connected, but since T is a subgraph of G - {e}, a disconnected graph, it must be
disconnected. Therefore, any spanning tree contains every bridge.
Now suppose e is not a bridge. Then G - {e} is still connected, and so has a spanning tree T.
However, since G - {e} has the same vertices as G, T is also a spanning tree of G that does not
contain e. Therefore, any edge in every spanning tree is a bridge.

By definition, a spanning subgraph is a subgraph of a graph consisting of the same vertex set and
a subset of the edge set of the graph, which is not necessarily a tree. A spanning tree is a spanning
subgraph that is also a tree.

Using induction on the number of vertices = 2k. For k = 1, we have 2 vertices, and so there is only
one possible tree, which is just two vertices connected by 1 edge -- the graph itself is a spanning
subgraph with odd vertices.

Suppose for k >= 1, every tree with 2k vertices has a unique spanning subgraph with all odd
vertices. Now since each leaf has degree 1 in the tree, any such spanning subgraph must include
all edges incident to leaves.

Consider a tree with 2(k+1) vertices. Consider a longest possible path in this tree. Suppose 1
endpoint of the path is a vertex v, which would have to be a leaf. Suppose u is the vertex adjacent
to u in the path.

Case 1: u has another adjacent leaf (say w).

T - {v, w} has a unique such spanning subgraph by our assumption (we assumed this for any tree
with 2k vertices), so use that spanning subgraph and add edges uv and uw -- the degree of u gains
2, so it stays odd, and v and w only have the 1 incident edge, so their degree will be 1, which is
odd, and we have the type of subgraph we need.

Case 2: u has no other adjacent leaf.5

Since p is the longest path, this means the degree of u is 2, or else there would be a cycle in the
graph. So, we know that T - {u, v} has a unique such spanning subgraph by assumption -- add uv to
that subgraph and we have the type of subgraph we need (u + uv + v will be a separate component).
So, by induction, the above statement holds for any number of vertices 2n for a positive integer n
 We have a connected graph G which has n vertices and n edges. Since G is connected,so it has a spanning
tree say T. Since T is a tree it has no cycle. T contains n vertices so T can have n-1 edges wgich belong to
G. But G has n edges,so there is one edge say e is not present in T. We assume that the end vertices of e is
u and v. Now in T there is a unique path between u and v. So combining this uv path and the missing edge
e, we can form a cycle in G. Let’s call this cycle 𝐶1.
Now we assume that there is another cycle 𝐶2 that exist in G. From our previous discussion,we can
confirm that this cycle will also contain that missing edge e. So excluding e,there must be another path
from u to v. But for a spanning tree there exist a unique path between any two vertices. So this uv path is
same as the previous one. Therefor 𝐶1 and 𝐶2 are the same cycle. Hence we can conclude that a connected
graph with n vertices has exactly one cycle if and only if it has exactly n edges.

Powered by TCPDF (www.tcpdf.org)