a glimpse to topological graph theory

41
1 A GLIMPSE TO TOPOLOGICAL GRAPH THEORY BY ANJU. M M.SC MATHEMATICS

Upload: anju123mohanan

Post on 19-Jan-2017

221 views

Category:

Education


0 download

TRANSCRIPT

Page 1: A glimpse to topological graph theory

1

A GLIMPSE TO TOPOLOGICAL

GRAPH THEORY

BY

ANJU. M

M.SC MATHEMATICS

Page 2: A glimpse to topological graph theory

2

Preface

The book “ A glimpse to topological graph theory” is intended for the college students and

teachers in Kerala syllabus. In this book all the topic have been deal with in a simple and lucid

manner. A sufficiently large number of problems have been solved. By studying this book , the

student is expected to understand the concept of graph theory in practical situations.

Suggestion for the further improvement of this book will be highly appreciated.

Anju M

Page 3: A glimpse to topological graph theory

3

CONTENTS

Title Page No:

Preface

Chapter 1 Preliminaries 5

Chapter 2. New graphs from old 17

Chapter 3. Surfaces and imbedding 26

Reference 41

Page 4: A glimpse to topological graph theory

4

INTRODUCTION

Graph Theory is regarded as one of the areas of Applied Mathematics.

Graph Theory has been independently discovered many times. Leonhard Euler is

known as “Father of Graph Theory”. Subsequent rediscoveries of Graph Theory

has been made by Gustav Kirchhoff and Arthur Cayley. Another approach was of

Hamilton‟s. The origin of graph theory can be traced back to Euler's work on the

Konigsberg bridges problem (1735), which subsequently led to the concept of an

Eulerian graph. The study of cycles on polyhedra by the Thomas P. Kirkman (1806

- 95) and William R. Hamilton (1805-65) led to the concept of a Hamiltonian

graph. The concept of a tree, a connected graph without cycles, appeared implicitly

in the work of Gustav Kirchhoff (1824-87), who employed graph-theoretical ideas

in the calculation of currents in electrical networks or circuits. Later, Arthur

Cayley (1821-95), James J. Sylvester (1806-97), George Polya (1887-1985), and

others use 'tree' to enumerate chemical molecules. The study of planar graphs

originated in two recreational problems involving the complete graph 5K and the

complete bipartite graph 3,3K . These graphs proved to be planarity, as was

subsequently demonstrated by Kuratowski. First problem was presented by A. F.

Mobius around the year 1840.

In the past few years, Graph Theory has been established as an

important mathematical tool in wide variety of subjects. Graph Theory has

application to some areas of Physics, Chemistry, Communicative Science,

Computer Technology, Electrical and Civil Engineering, Architecture, Operation

Research, Genetics, Psychology, Sociology, Economics, Anthropology and

Linguistics. The theory had also emerged as a worthwhile mathematical discipline

in its own right. Graph Theory is intimately related to many branches of

Mathematics including Graph Theory, Matrix Theory, Numerical Analysis,

Probability and Topology. In fact Graph Theory serves as a mathematical model

for any system involving a binary relation.

Page 5: A glimpse to topological graph theory

5

CHAPTER – 1

PRELIMINARIES

Topological graph theory is a branch of graph theory which connects both graph

theory and topology. We can move on to some preliminary things of toplogical

graph theory.

REPRESENTATION OF GRAPHS

In topological graph theory, a graph is a network of nodes and curved arcs starting

from one node to the other or to that node itself. This is in the sense of topological

definition. In the combinatorial viewpoint, a graph G consists of a vertex set VG

and an edge set EG such that every edge e EG has endpoints from the set VG. The

endpoints of an edge e is represented by the set VG(e). Considering the set VG(e)

for every edge e EG, we get the incidence structure of the graph G. That is, the

incidence structure of a graph denoted by IG is

IG = { VG(e): e EG}. If there is only one graph under consideration, then we omit

the subscripts used above.

Page 6: A glimpse to topological graph theory

6

Note:

If the vertex set VG and edge set EG of a graph are finite, then we say that the graph

G is finite.

Note:

In graph theory, we know that a simple graph is a graph having no self loops and

parallel edges. Here in topological graph theory, that type of graphs are known as

simplicial graphs or simplicial 1-complex. This is because, in topology the word

simple have other meaning. A simple loop is a continuous non-singular image of a

circle. That is, a loop without self intersections.

DRAWINGS

Here a finite graph is drawn same as in the case of normal graph theory. First we

consider the sets VG and EG of the graph. Then we represent each v VG as dots.

Then we join them by using each e EG. Drawing a graph in a plane will

sometimes provide crossover between the edges. But fortunately, the interior of an

edge is not necessary to meet a vertex. Crossover of two edges does not always

represent a vertex.

Example of a graph drawing is given below:

Page 7: A glimpse to topological graph theory

7

Let V = {v1, v2, v3} and E = {e1, e2, e3, e4}. Also V(e1) = {v1}, V(e2) = {v1, v2},

V(e3) = {v2, v3} and V(e4) = {v2, v3}

e1 e2

v1 v2

e4 e3

v3

Note:

A graph having a self loop is known as a pseudograph. Clearly they are non-

simplicial. Harary brought the idea of pseudograph.

An essential non-simplicial graph used in topological graph theory is the bouquet

of circles which will be discussed later.

Page 8: A glimpse to topological graph theory

8

INCIDENCE MATRIX AND ADJACENCY MATRIX

The incidence matrix of a graph G has #VG rows and #EG columns where #X

represents the number of elements in the set X. The entries in an incidence matrix

will be 0, 1, or 2. An i,jth

entry becomes 0 if vi VG(ej). An i,j

th entry becomes 1 if

vi VG(ej) and #VG(ej) = 2. And an i,jth

entry becomes 2 if {vi} = VG(ej).

The adjacency matrix is drawn by considering the adjacency of vertices. It has

#VG rows and columns. Here the entries will be any number. An i,jth

entry is

written by considering the adjacency of vi and vj. If there is no edges between vi

and vj, i,jth

entry will be 0. Otherwise the number of edges between them

constitutes that entry. A self edge contributes only one to the corresponding entry.

VALENCE

In normal graph theory, degree of a vertex is defined as the number of edges

incident on that vertex. But in topological graph theory, we will not use the word

degree for defining this. We will only use the word valence. This is because in

topology, the word degree is there in another terminology. Degree of a continuous

function is defined there, so we will not use it here. In the combinatorial viewpoint,

Page 9: A glimpse to topological graph theory

9

the valence of a vertex is the sum of entries in its corresponding row of the

incidence matrix associated with that graph.

DIRECTIONS

Let us consider the direction of edges in a graph. A direction of an edge e is

defined to be an onto function f: {BEGIN, END} V(e). The image vertex of

BEGIN is the initial vertex and the image vertex of END is the terminal vertex.

Sometimes this mapping acts as a constant map and that is in the case of a self

loop. Moreover, in topological graph theory, the direction of an edge is not in the

proper sense. That is, an edge e will have 2 directions eventhough there is an arrow

marked on it. We indicates the edge as e+ or e

-. e

+ indication is for which the

direction of that edge is same as the arrow mark. Where the e- indication is for

which the direction of that edge is opposite of the arrow mark. Consider the

directed graph below:

v1 e-

v2

Page 10: A glimpse to topological graph theory

10

Here the direction of the edge e is from v1 to v2 if we consider the arrow mark. But

here the sign given for the edge e is -. So the direction of e is from v2 to v1. This is

how directions are given for the graphs in topological graph theory.

GRAPH MAPS, ISOMORPHISMS

Consider 2 graphs G1 and G2. A graph map is a mapping from VG1 to VG2 and EG1

to EG2. In this mapping, the incidence is preserved. For every edge e EG2, the

vertex mapping maps the endpoints of e that is VG1(e) to the endpoints of VG2(e).

A graph map is an isomorphism if it is one to one and onto. We say 2 graphs are

isomorphic if there exists an isomorphism between them.

Note:

Under a graph map, a proper edge may be mapped to a self loop. But a self loop

will never be mapped to a proper edge.

AUTOMORPHISM

An isomorphism from a graph to itself is called an automorphism. The set of all

automorphisms of a graph forms a group under the operation composition of

functions. This group is represented by Aut(G)

Page 11: A glimpse to topological graph theory

11

SOME IMPORTANT CLASSES OF GRAPHS

In a graph, a walk is the combinatorial analogue of the continuous image of a

closed line segment, which may cross itself or retrace upon itself backwards or

forwards, arbitrarily often. If the starting vertex and the ending vertex of a walk are

different, we say that it is an open walk. Whereas if the starting vertex and the

ending vertex are same, we say that it is a closed walk.

A path is an open walk whose vertices are distinct. Thus a path is a combinatorial

analogue of a homeomorphic image of a closed line segment. The standard path

with n vertices is denoted by Pn and it is called the n-path.

A cycle in a graph is a closed walk in which every vertices are distinct except the

starting vertex and the ending vertex. So a cycle is a combinatorial counterpart to

the homeomorphic image of a circle. The standard cycle with n vertices is denoted

by Cn and its called as the n-cycle.

Note:

A graph is connected if and only if for every pair of vertices, there exists a path

between them. Otherwise the graph is said to be disconnected. A connected graph

has only one component whereas a disconnected graph has 2 or more components.

Page 12: A glimpse to topological graph theory

12

TREES

A tree is defined to be a connected graph with no cycles contained in it. This is

the most important kind of graphs in both applications and theory. Since the

definition of tree is as same as in the normal graph theory, the same theorems on

trees holds here also.

Theorem:

Let u and v be any vertices of a tree T. Then there is a unique path from u to v.

Theorem:

Let T be a tree. Then #VT = #ET + 1.

Theorem:

Every non-trivial tree T has atleast 2 vertices of valence 1.

COMPLETE GRAPHS

As in normal graph theory, here also a complete graph is defined as a graph in

which for every pair of vertices, there exists an edge between them. The standard

model of a complete graph with n vertices is denoted as Kn. Some complete graphs

are given below:

Page 13: A glimpse to topological graph theory

13

K1 K2 K3

Note:

K1 is sometimes called the trivial graph.

BIPARTITE GRAPHS

A graph G is said to be bipartite, if the vertex set of that graph can be partitioned

into two sets such that for each vertex in the first set, there must be edges going to

every vertices in the second set. Also the vertices in each sets must be mutually

non-adjacent. A standard model of a complete bipartite graph is denoted as Km,n.

Here m and n are the number of vertices in the partitioned sets.

Page 14: A glimpse to topological graph theory

14

Note:

A complete bipartite graph is a graph which is both complete and bipartite.

An example of a complete bipartite graph is given below:

K3,3

BOUQUET OF CIRCLES

This is a special category of graphs which is only defined in topological graph

theory. A bouquet of n circles is a graph having one vertex and n self loops

attached to that vertex. It is denoted by Bn.

An example B4 is shown below:

B4

Page 15: A glimpse to topological graph theory

15

CAYLEY GRAPHS

Cayley used graphs to represent groups. He used them for drawing the pictures of

the group. It is drawn by using the group elements and the generators of that group.

Simply saying, a group can be represented as a graph as follows:

First represent the group elements as the vertices of the graph. Then for each

generators, we start from a vertex and draw edges to the next element which the

generator generates. An example will make it more clear. The graphical

representation of the group Z5 with generating set {2} is shown below:

0

4 1

3 2

Here we first plot all the group elements which is Z5 = {0, 1, 2, 3, 4} as the vertices

of the graph. Then we start from the vertex 0. Since the generator here is 2, the

Page 16: A glimpse to topological graph theory

16

next element it generates is 2. So we draw an edge from 0 to 2. Like that, continue

the process until all the elements are over. When the generator changes, the

direction of edges also changes. Sometimes there will be more than one generators

given. Then we give special colors to the edges of each generators separately. This

way of drawing results the graph known as the Cayley color graph which is

denoted by C(A, X) where A is the group and X is the generating set. In this graph,

the vertex set is the element of A and the edge set is X A, the cross product of X

and A. Each edge (x, a) travels from a to ax in the plus direction. (x, a) is

represented as subscript form also by xa. Now if we remove the colors of edges and

all such modifications from a Cayley color graph, the resultant graph is known as

the Cayley graph and is denoted by C(A, X)0.

Page 17: A glimpse to topological graph theory

17

CHAPTER 2

NEW GRAPHS FROM OLD

SUBGRAPHS

A graph G is said to be a subgraph of another graph G if and only if the sets VG,

EG, IG are subsets of VG, EG, IG respectively. Moreover, if VG = VG, then we say

that G spans G. If such a spanning graph is a tree, then it is called a spanning

tree.

Let us look at the following example:

0 0

a b 0

1 f 2 3 1 f 2 3 a b

c d c d 1 2

4 4

G G G

e e

Page 18: A glimpse to topological graph theory

18

In the above figure, G and G are the subgraphs of G. Among them, G is the

spanning tree of G. Whereas, G is not the spanning tree of G since it does not

contains all the vertices of G.

The induced subgraph of a graph G for a vertex set V VG is the subgraph G

having vertex set V and edge set with all the edge e such that VG(e) VG.

The following example makes it clear for you:

G G G

In the above figure, G is an induced subgraph of G. But G is not an induced

subgraph of G since does not contain all the edges in G incident on that four

vertices.

Page 19: A glimpse to topological graph theory

19

Note:

Every induced subgraph of a graph G is a subgraph of that graph, but the converse

may not be true always.

TOPOLOGICAL REPRESENTATION

Now we move on to the topological representation of the graphs. Any graph G can

be represented as a topological space. It is done by representing each vertex by a

distinct point and each edge by a distinct arc homeomorphic to [0,1]. Then the

boundary points of each arc represents the endpoints of each edge. It is essential

that the interior of the boundary points must not touch any distinct points. Such a

space is called the topological representation of the graph G.

SUBDIVISION

Next is the term subdivision. Subdivision is nothing but an operation done in

graphs. As we perform operations like edge deletion, vertex deletion, etc

subdivision is also an operation which has a basic rule to perform. This operation is

done in the edges of a graph. In the sense of drawing, subdividing an edge e is

done by adding a new vertex in the interior of that edge. In the combinatorial

viewpoint, a graph G after subdividing an edge e will have vertex set

Page 20: A glimpse to topological graph theory

20

VG {v} and edge set (EG – {e}) {e1, e2}. The incidence structure will be

changed as

(IG – VG(e)) {VG(e1), VG(e2)}.

An example for subdivision is given below:

e1 e1 e1

v

G G

Here in G the edge e1 has performed subdivision.

Note:

The major use of subdivision is that it makes a non-simplicial graph simplicial.

Just we have to do is first subdivide all the self loops, then subdivide all but one

edge among all the multiple adjacencies.

Page 21: A glimpse to topological graph theory

21

HOMEOMORPHISM

Now we move on to homeomorphism. Two graphs are homeomorphic if they

have the respective subdivisions such that the resultant graphs are isomorphic to

each other. So, it may be observed that two graphs are homeomorphic if and only if

their topological representations are homeomorphic as topological spaces.

An example of two homeomorphic graphs is given below:

CARTESIAN PRODUCT

Consider two graph G and G. The Cartesian product of the graphs G and G

denoted by G G is taken as the graph with vertex set VG VG and edge set (VG

EG) (VG EG). If the edge (e, v) VG EG, and the endpoints of e are v1

and v2, then the endpoints of the edge (e, v) will be (v1, v) and (v2, v). Also if the

edge (v, e) VG EG and the endpoints of e are v1 and v2, then the endpoints of

Page 22: A glimpse to topological graph theory

22

the edge (v, e) will be (v, v1) and (v, v2). An example for the cartisian product of

two graphs is given below.

The Cartesian product of P2 and P3:

0 (0,2) (a,3) (0,3)

a b =

1 2 3

(1,2) (1,3)

EDGE COMPLEMENTS

The edge complement of a simplicial graph G is a graph G having vertex set

same as G but such that two vertices in G are adjacent only if they are not adjacent

in G. It is denoted by Gc.

The following is an example for the edge compliment of a graph:

(a,2) (a,3)

(1,b)

Page 23: A glimpse to topological graph theory

23

G Gc

Note:

The edge complement of a complete graph Kn has no edges in it but only n

vertices. Also the edge complement of the union of two complete graphs Km and

Kn forms Km,n.

SUSPENSION

The suspension of a graph G from another graph G denoted by G + G is defined

by the graph with vertex set VG VG and edge set (EG EG) (VG VG). That

is the suspension is made by adjoining every vertex in G to every vertex in G by

edges. The edges in G + G which were in EG and EG remains unchanged. But the

Page 24: A glimpse to topological graph theory

24

edges that arises from VG VG have endpoints from the sets VG and VG. An

example of a suspension is given below.

The suspension of P3 from P2:

+ =

Note:

The complete graph Kn+1 is isomorphic to the suspension of the complete graph Kn

from the graph K1.

Note:

The suspension of a cycle Cn from the trivial graph K1 is called a wheel.

Page 25: A glimpse to topological graph theory

25

AMALGAMATION

Let G1 and G2 be two graphs. And also let H1 and H2 be two subgraphs of G1 and

G2 respectively. Consider there exists an isomorphism that carries H1 onto H2.

Then the amalgamation of two graphs G1 and G2 denoted by G1 f G2 is done by

joining G1 and G1 by pasting together these two isomorphic graphs H1 and H2 by

pasting together them according to their isomorphism. An example of the

amalgamation of two graphs is given below.

The amalgamation of C3 and C4:

f =

Note:

The bouquet of circles Bn+k is an amalgamation of the bouquets Bn and Bk at a

vertex.

Page 26: A glimpse to topological graph theory

26

CHAPTER 3

SURFACES AND IMBEDDINGS

The central concern of topological graph theory is the placement of graphs in a

surface. Such a surface will be compact and having no boundary. That is, it is a

closed one. The plane is not a closed surface. But the plane differs from the sphere

by only a single point. That is, removing a point from a sphere results in a plane.

So the placement of graphs in a plane will be equal to placement of the same in a

sphere. Moreover, a graph can be placed in a plane if and only if it can be placed in

a sphere. Closed surfaces are classified into two categories, orientable surfaces and

non-orientable surfaces. Let us look at those two categories.

ORIENTABLE SURFACES

Orientable surfaces can be described as follows. If we place a small 2-

dimensional object in that surface and move it along the surface, when we reach

the starting point again, that 2-dimensional object must not be reversed. If it

happens, then that surface is not an orientable surface. The sphere, Torus, Double

Torus, etc are orientable surfaces. They are usually denoted by S0, S1, S2, ...

Page 27: A glimpse to topological graph theory

27

S0 S1 S2

A major characteristics of the closed orientable surfaces is that each of them can be

obtained from a single sphere in 3-space. It is by the technique of adding handles

to a sphere.

Procedure of adding a handle to the sphere:

A handle is added to the sphere in the following manner. First cut two holes in the

sphere. Then attach each of the holes by a cylindrical handle. The technique is

illustrated in the figure given below:

Page 28: A glimpse to topological graph theory

28

Adding a handle like this will make the sphere homeomorphic to a torus. In that

sense, a torus is obtained from a sphere. Like this, every closed orientable surfaces

can be made out from a sphere. Just attach the required number of handles to it and

it will become that surface. Adding two handles will make the double torus, adding

three handles will make the triple torus and so on.

Note:

Because of the characteristics of adding handles, every finite graph can be drawn

in a sphere without any edge crossings. This is done by adding a handle to every

edge crossing portions and reroute the crossed edge through that handle. The

technique is shown in a figure next:

e e

e e

e e

e e

Page 29: A glimpse to topological graph theory

29

Here e and e are the edges which are crossed over. We add a handle to the surface

of the sphere and reroute e through that handle. Then the crossover is gone. In this

way, any finite graphs can be drawn in a sphere without edge crossings. In general,

any finite graph can be drawn easily in a closed connected orientable surface

without any edge crossings. This is because, every closed connected orientable

surface is homeomorphic to one among the S0, S1, S2, …

NON-ORIENTABLE SURFACES

Non-orientable surfaces can be described as follows. If we place a 2-dimensional

object in that surface and move it along the surface, when we reach the starting

position again, the object must be reversed. Such a surface is termed as a non-

orientable surface. The mobius band is a perfect example for a non-orientable

surface.

A mobius band is made by the following manner:

First take a rectangular piece of paper. Then twist it 180 so that the top and the

bottom of its one side is reversed. Then paste the right and lefty side of that paper.

This will form a mobius band. The construction is illustrated in the figure given

below:

Page 30: A glimpse to topological graph theory

30

A projective plane is made from a sphere using a mobius band. First cut a hole in

the sphere. Then the boundary of that hole will be homeomorphic to a circle. The

boundary of a mobius band is also homeomorphic to a circle. So we can attach a

mobius band to the hole of the sphere which is cut out. This will result in a

projective plane. The kleine bottle, a non-orientable surface can be obtained from a

sphere by cutting two holes in a sphere and closing each of them with two mobius

bands. This is impossible to illustrate in a figure. The surface obtained by cutting k

holes in a sphere and thereafter closing each of them with mobius bands is denoted

by Nk for k = 0, 1, 2, … In this method, a mobius band is often called a crosscap.

Finally, a closed surface is non-orientable if and only if it has a surface

homeomorphic to a mobius band.

Page 31: A glimpse to topological graph theory

31

Note:

Every closed connected non-orientable surfaces can be made from a sphere by

cutting holes in it and closing off them by mobius bands.

IMBEDDING

An imbedding of a graph in a surface is defined as a one to one function from the

topological representation of the graph to the surface. It will be noted that do not

consider the image of this mapping as a graph.

If a connected graph is imbedded in a sphere, then the complement of its image

will form a set of regions or faces. These regions each one will be homeomorphic

to an open disk. This is in the case of sphere only. But in more complicated

surfaces, the regions will not be homeomorphic to an open disk. If all the regions

are homeomorphic to an open disk in an imbedding, then we say that the

imbedding is 2-cell imbedding or 2-cellular imbedding. If the boundary circuit of

an open disk region has one or more repeated vertices, then the closure of the

region is not a closed disk. Therefore, the imbedding is a 2-cell imbedding depends

only on whether all the regions are open disks, not on whether the closures of the

regions are closed disks.

Page 32: A glimpse to topological graph theory

32

Consider a graph imbedding i: G S. The set of all regions of an imbedding i is

denoted by FG. Here, F represents the term faces of a polyhedron. This is because

the regions are somewhat like the faces of a polyhedron. The number of sides of a

region is defined as the total number of edge sides one encounters while traversing

a simple circuit just inside the boundary of that region. It is denoted by sf. The total

number of edges in that boundary will not be equal to the number of sides. Let us

consider an example for this.

A spherical imbedding:

In this example the exterior region has seven sides. If we draw a simple cycle just

inside the boundary of the exterior region, we will encounter seven edges. So it has

seven sides. But the exterior region has only five edges in its boundary. So the

Page 33: A glimpse to topological graph theory

33

number of sides of a region will not be equal to the number of edges in the

boundary of that region.

Note:

In an imbedding, the region having only one side is known as a monogon or

unigon. Whereas the region having two sides is known as digon or bigon.

Note:

If all the regions of a graph imbedding are three sided, then that imbedding is

known as triangular imbedding. Also if all the regions of the graph imbedding

are four sided, then it is called a quadrilateral imbedding.

Imbedding of K5 in the mobius band:

a a

0

1

3

4

2

Page 34: A glimpse to topological graph theory

34

Imbedding of K3,3 in the mobius band:

a a

Imbedding of K5 in the torus:

a a

b

0 1

2

3

4

5

b

Page 35: A glimpse to topological graph theory

35

Imbedding of K3,3 in the torus:

a a

EULER’S EQUATION FOR THE SPHERE

We have Euler’s equation in the graph theory for every connected planar graphs.

The same equation was proved in topological graph theory also. But there are

certain conditions for satisfying the Euler‟s equation. The imbedding must be 2-

cell. And there must be no holes in the surface used for imbedding. In general,

Euler‟s equation holds for every closed orientable surfaces. We know that the

Euler‟s equation is #V - #E + #F = 2. We will prove this for a particular imbedding

now. Before that, we can look on some terms. In this equation, the left side is

known as the Euler formula and the right side is known as the Euler

characteristic of the surface. The equation #V - #E + #F = 2 holds for only

Page 36: A glimpse to topological graph theory

36

imbeddings in a sphere. So the Euler characteristic of the sphere is 2. Now let us

move on to the theorem.

Theorem:

Let i: G S0 be an imbedding of a connected graph on a sphere. Then, #V - #E +

#F = 2.

Proof: This proof proceeds by induction on the number #F of regions. First,

observe that if #FG = 1, then G must be a tree, since the Jordan curve theorem

implies that any cycle would separate the sphere. Thus by a theorem, #VG - #EG =

1. From this it follows that #VG - #EG + #HG = 2. Now suppose that the Euler‟s

equation holds when the number of regions is at most n, and suppose that #FG = n

+ 1. Then some edge e lies in the boundary circuit of two distinct regions. Since

the two regions are distinct the subgraph G obtained by removing the edge e is

connected. Then #FG = #FG – 1 = n, so by induction, #VG - #EG + #FG = 2. Since

#VG = #VG, #EG = #EG – 1 and #FG = #FG – 1, it follows that #VG - #EG + #FG =

2.

Page 37: A glimpse to topological graph theory

37

KURATOWSKI’S GRAPHS

From the Euler equation, there arises an inequality which is used to prove certain

graphs cannot be imbedded on a sphere. This inequality relates the number of

edges and number of regions of an imbedding. So this inequality is also known as

the edge-region inequality. It is established in the theorem below.

Theorem:

Let i: G S be an imbedding of a connected, simplicial graph with atleast three

vertices into a surface. Then 2#E 3#F.

Proof: For proving this, we will consider a sum first. That is, we consider the sum

of numbers of sides of all regions of the imbedding i. The corresponding sum is f

FG (sf). We know that every edge contributes two to the above sum. Given that the

graph G is simplicial. That implies, there will be no monogons and digons in the

imbedding. So

sf 3 f FG (1).

Since every edge contributes two to the sum f FG (sf),

2#E = f FG (sf).

Therefore from (1), we get 2#E 3#F. Hence proved the inequality.

Page 38: A glimpse to topological graph theory

38

Note:

The complete graph K5 and the complete bipartite graph K3,3 are known as the

Kuratowski‟s graphs, because Kuratowski proved that they are the complete set of

obstructions to imbed some graphs on the sphere.

Kuratowski’s theorem:

A graph G has an imbedding in the sphere if it contains no homeomorph of K5 or

K3,3.

The proof is omitted in this section.

We can easily show that K5 and K3,3 are non-spherical by using the edge-region

inequality. Just apply the corresponding values, #E and #F of the K5 and K3,3 in

that inequality and we will get the wrong answer. So they cannot be imbedded in a

sphere. Next is a term named girth. The girth of a graph is defined as the number

of edges in its shortest cycle. If a graph has a loop in it, then its girth is 1. If a

graph does not have a loop in it and have a multiple edge in it, then its girth is 2. A

simplicial graph have always girth 3. The girth of a tree is defined as infinite.

The girth of a bipartite graph cannot be an odd number since it has no odd cycles.

The girth of a bipartite simplicial graph is at least 4 always.

Page 39: A glimpse to topological graph theory

39

DUALITY

Let us consider a connected graph G, a closed surface S, and a 2-cell imbedding i:

G S. The construction of a dual graph and dual imbedding can be discussed

now. First consider the imbedding. Then we construct a dual graph as follows.

Add a vertex f in each of the region‟s interior part of the imbedding. Then draw

edges e from a vertex in one region to the vertex in the another region such that,

these edges pass through the actual edges in the imbedding. If an edge is situated in

one region only, then the dual edge e will be a self loop along the dual vertex in

that region. After this drawing is done, the resultant graph got is the dual graph

denoted by Gi

or G. If „i‟ is the only imbedding under consideration, then we use

the notation G only.

The imbedding of the dual graph in a surface is known as the dual imbedding. In

this context the original graph and the original imbedding is known as the primal

graph and the primal imbedding respectively.

Two different duals of the same graph imbedded in the sphere is illustrated below:

Page 40: A glimpse to topological graph theory

40

In these graphs, the primal graph is drawn using solid lines and solid dots. Whereas

the dual graph is drawn using dashed lines and open dots. In the first graph, a

vertex has a maximum valence seven. But in the second graph the vertex having

maximum valence is just six. So the duals of a same graph imbedding need not be

isomorphic. For any region f of the primal imbedding, the valence of the dual

vertex f is equal to the number of sides of f.

Page 41: A glimpse to topological graph theory

41

REFERENCE

1.Diestel.R, Graph theory, Electronic edition, 2005

2. Graph Theory and Its Applications, Second Edition

by Jonathan L. Gross (Author), Jay Yellen (Author)