You are on page 1of 36

B.Sc.

in Computer Science and Engineering Thesis

A New Sufficient Condition for Hamitonian Cycle

Submitted by

Ismat Tarik
201314049

Fariha Chowdhury
201314056

Fazle Rabby
201214047

Supervised by

Dr. Mohammad Kaykobad

Professor

Department Of Computer Science & Engineering

Bangladesh University Of Engineering and Technology

Department of Computer Science and Engineering


Military Institute of Science and Technology

January 2017
CERTIFICATION

This thesis paper titled “A New Sufficient Condition for Hamitonian Cycle”, submitted
by the group as mentioned below has been accepted as satisfactory in partial fulfillment
of the requirements for the degree B.Sc. in Computer Science and Engineering in January
2017.

Group Members:

Ismat Tarik

Fariha Chowdhury

Fazle Rabby

Supervisor:

Dr. Mohammad Kaykobad


Professor
Department Of Computer Science &
Engineering
Bangladesh University Of Engineer-
ing and Technology

ii
CANDIDATES’ DECLARATION

This is to certify that the work presented in this thesis paper, titled, “A New Sufficient
Condition for Hamitonian Cycle”, is the outcome of the investigation and research carried
out by the following students under the supervision of Dr. Mohammad Kaykobad, Professor,
Department Of Computer Science & Engineering, Bangladesh University Of Engineering
and Technology.

It is also declared that neither this thesis paper nor any part thereof has been submitted
anywhere else for the award of any degree, diploma or other qualifications.

Ismat Tarik
201314049

Fariha Chowdhury
201314056

Fazle Rabby
201214047

iii
ACKNOWLEDGEMENT

We are thankful to Almighty Allah for his blessings for the successful completion of our
thesis. Our heartiest gratitude, profound indebtedness and deep respect go to our supervisor,
Dr. Mohammad Kaykobad, Professor, Department Of Computer Science & Engineering,
Bangladesh University Of Engineering and Technology, for his constant supervision, affec-
tionate guidance and great encouragement and motivation. His keen interest on the topic
and valuable advices throughout the study was of great help in completing thesis.

We are especially grateful to the Department of Computer Science and Engineering (CSE)
of Military Institute of Science and Technology (MIST) for providing their all out support
during the thesis work.

Finally, we would like to thank our families and our course mates for their appreciable
assistance, patience and suggestions during the course of our thesis.

Dhaka
January 2017 Ismat Tarik
.
Fariha Chowdhury

Fazle Rabby

iv
ABSTRACT
The subject of the thesis is Hamiltonian cycle problem. Hamiltonian cycle is a
simple graph cycle in which every vertex is visited only one time. Hamiltonian path and
cycle are named after William Rowan Hamilton. Hamiltonian cycle is one of the most
important topic in the field of graph theory, computational complexity and algorithm
design. Our goal is to find a sufficient condition for a simple 2-connected graph to be
Hamiltonian. Though there are already theorem for a graph to be Hamiltonian for all
pair of non-adjacent vertices with distance 3. Here, we are trying to propose a theorem
for a simple 2-connected graph to be Hamiltonian for all pair of non-adjacent vertices
with distance 4.

1
Contents

CERTIFICATION ii

CANDIDATES’ DECLARATION iii

ACKNOWLEDGEMENT iv

ABSTRACT 1

List of Abbreviation 5

1 Introduction 6

1.1 Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6

1.2 Basic Definitions and Notations . . . . . . . . . . . . . . . . . . . . . . . 7

1.3 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

1.4 The Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10

1.5 Scopes and objectives of the thesis . . . . . . . . . . . . . . . . . . . . . . 11

1.6 Thesis Organization . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13

2 Hamiltonicity 14

2.1 Complexity of the Recognition Problem . . . . . . . . . . . . . . . . . . . 14

2.2 Sufficiency Condition . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

2.2.1 Dirac’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 16

2.2.2 Ore’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

2.2.3 Nash-Williams Theorem . . . . . . . . . . . . . . . . . . . . . . . 18

2
2.2.4 Goodman and Hedetniemi Theorem . . . . . . . . . . . . . . . . . 19

2.2.5 Chvátal and Erdos Theorem . . . . . . . . . . . . . . . . . . . . . 19

2.2.6 Bondy and Chvátal’s Theorem . . . . . . . . . . . . . . . . . . . . 20

2.2.7 Chvátal’s Theorem . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2.2.8 Rahman-Kaykobad Theorem . . . . . . . . . . . . . . . . . . . . . 21

2.2.9 Improvement of Rahman-Kaykobad Theorem . . . . . . . . . . . . 23

3 A New Sufficiency Condition for Hamiltonicity 25

3.1 Two Crossover . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

3.2 Overview of the theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

3.2.1 Explanation with example . . . . . . . . . . . . . . . . . . . . . . 26

3.3 Proposal of new theorem . . . . . . . . . . . . . . . . . . . . . . . . . . . 27

3.3.1 Explanation with example . . . . . . . . . . . . . . . . . . . . . . 28

3.4 Findings . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 29

4 Conclusion and Recommendation 30

4.1 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

4.2 Recommendation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

3
List of Figures

1.1 A Hamiltonian Cycle in Dodecahedron . . . . . . . . . . . . . . . . . . . . 6

1.2 Simple Graph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7

1.3 Travelling Salesman Problem . . . . . . . . . . . . . . . . . . . . . . . . . 12

2.1 (a) An example of hamiltonian cycle shown in bold (b) A cover-enforcer


subgraph He . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14

2.2 Possible ways that a hamiltonian cycle can visit the edges in a cover-enforcer
He . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15

2.3 (a) A vertex v in G and its set of incident edges (b) The connections made
between the Hei ’s in H for the edges incident upon v. . . . . . . . . . . . . 16

2.4 Wheel Graph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17

2.5 Hamiltonian Graph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

2.6 k1,3 , k2,3 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18

2.7 k1,3 , k1,3 +x . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

2.8 Hamiltonian Graph . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19

2.9 Hamiltonian Closure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

3.1 Two Crossover Graph(a) . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

3.2 Two Crossover Graph(b) . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

3.3 Non-Hamiltonian graph with distance=3 . . . . . . . . . . . . . . . . . . . 27

3.4 Hamiltonian graph with distance=3 . . . . . . . . . . . . . . . . . . . . . . 27

3.5 Non-Hamiltonian graph with distance=4 . . . . . . . . . . . . . . . . . . . 28

3.6 Hamiltonian graph with distance=4 . . . . . . . . . . . . . . . . . . . . . . 28

4
LIST OF ABBREVIATION

SG : Simple Graph
HG : Hamiltonian Graph
VS : Vertex Set
EV : End Vertices
ES : Edge Set
CG : Complete Graph
DG : Directed Graph
TCP : Travelling Salesman Problem
VC : Vertex Connectivity
EC : Edge Connectivity

5
CHAPTER 1
INTRODUCTION

1.1 Overview

The objective of this thesis paper is to find a new sufficient condition for hamiltonian cycle.
A path is hamiltonian if all the vertices of the graph can be visited only once. If the hamil-
tonian path returns to its starting vertex then it is a hamiltonian cycle. The graph containing
hamiltonian cycle is called hamiltonian graph. Hamiltonian paths and cycles are named after
William Rowan Hamilton.

Figure 1.1: A Hamiltonian Cycle in Dodecahedron

He invented an icosian game currently known as Hamilton’s puzzle, which involves finding a
Hamiltonian cycle in the edge graph of the dodecahedron. Hamilton solved this problem us-
ing the icosian calculus, an algebraic structure based on roots of unity with many similarities
to the quaternions. This solution does not generalize to arbitrary graphs. Though hamilto-
nian cycle is named after Hamilton, Thomas Kirkman [3] studied hamiltonian cycles in
polyheydra a year before and set example of a polyheydron without hamiltonian cycle.

Tremendous amount of research have been done to find necessary and sufficient conditions
for a graph to be Hamiltonian [13]. Hamiltonicity of graphs have been studied in the 2
perspective of independent set [21], dominating circuit [14], k-ordered Hamiltonicity[15],
2-trail (a trail that uses every vertex at most twice) [10], density (size, degrees, neighborhood
conditions of a graph etc.)[13], [23], toughness [2], [4], forbidden subgraphs [9], multiple
Hamiltonian cycles [17], alternating Hamiltonian cycle (a properly colored Hamiltonian
cycle in which adjacent edges have different colors in a graph Gc with colored edges) [1]

6
etc. Rahman and Kaykobad [22] considered the shortest distance between the pair of non-
adjacent vertices along with their degree sum to give sufficient conditions for Hamiltonian
path in a graph.

1.2 Basic Definitions and Notations

Before presenting some of the relevant conditions of Hamiltonicity existing in the literature,
we need to introduce and define some of the notations that will be used throughout this
paper.

Definition. A graph G consists of a set of vertices V and a set of edges E. So it can easily
be said that G = (V,E).

Vertices of a graph will be represented by lower case English letters possibly with subscripts
like the following: v, u3 , or vi . Edges will be represented as the concatenation of the names
of the two vertices it connects. For example vu, s1 v would all represent edges in a graph. It
should be noted that since we are not dealing with directed graphs the order of the concate-
nation does not matter, therefore vu = uv.

Definition. A graph is simple if it has no loop or multiple edges.

Figure 1.2: Simple Graph

Definition. The degree of vertex v, deg(v), is equal to the number of distinct edges to which
it belongs.
Definition. The order of a graph G, n(G), is equal to the number of vertices present in G.

When only one graph is being discussed n(G) may simply be written as n for ease of notation.
This shorthand is utilized for many graph invariants.

Definition. The connectivity of a graph G, κ(G), is equal to the minimum size vertex set S
such that G - S is disconnected or has only one vertex.
Definition. Suppose G is a graph and let n = n(G). The degree sequence of the graph G,
(d1 , d2 , . . . , dn ), is a nondecreasing list of the degrees of the vertices in G. For notational
P
ease δ(G) = d1 , σ(G) = d2 , (G) = dn−1 , and ∆(G) = dn .
Definition. A vertex cover V0 of an undirected graph G = (V,E) is a subset of V such that

7
uv ∈ E ⇒ u ∈ V’ ∨ v ∈ V0 , that is to say it is a set of vertices V0 where every edge has at
least one endpoint in the vertex cover V0 . Such a set is said to cover the edges of G.

The next two definitions are for the graph substructures with which this paper is most con-
cerned.

Definition. A path P of order k in a graph G is a a sequence of vertices [pi ]k1 = [p1 , p2 , . . .


, pk ] such that each pi ∈ V(G), pi pi+1 ∈ E(G) and pi = pj if and only if i = j.
Definition. A Hamiltonian path of a graph G is a path that contains every vertex of G.

Other sources may refer to a Hamiltonian path simply as a spanning path or say that the
graph is traceable.

Definition. The (Hamiltonian) closure of a graph G, denoted C(G), is the graph with vertex
set V(G) obtained from G by iteratively adding edges joining pairs of non-adjacent vertices
whose degree sum is at least n until no such pair remains.

1.3 Literature Review

From history ,we acknowledged that first good sufficient condition for Hamiltonian cycle is
proposed by Dirac in 1952 where he works on simple graph and here he proposed that if
there exist crossover edge then the graph will be Hamiltonian.

Theorem 1(Dirac,[8]): If G is a simple graph with at least three vertices and δ(G)≥ n(G)/2,
then G is Hamiltonian.

Later Ore gave a improved version of Dirac. In 1960,Norwegian mathematician Oystein


Ore proved Ore’s theorem. Here, he states that a graph with sufficiently many edges must
contain a Hamiltonian cycle.

Theorem 2(Ore,[20]): Let G be a simple graph. If u,v are distinct non-adjacent vertices of
G with d(u) + d(v) ≥ n(G), then G is Hamiltonian if and only if G + uv is Hamiltonian.

The best vertex degree characterization of Hamiltonian graphs was provided in 1972 by the
Bondy-Chvátal theorem, which generalizes earlier results by G. A. Dirac (1952) and Oystein
Ore.

Theorem 3(Bondy-Chvátal,[5]): A simple n-vertex graph is Hamiltonian if and only if its


closure is Hamiltonian.

Then in 1972, the best sufficient condition is given by Chvátal. In this theorem he works on
complete graph.

Theorem 4(Chvátal,[6]): Let G be a simple graph with vertex degree d1 ≤....≤ dn , where n
≥ 3. If i< n/2 implies that di > i or dn−i ≥ n-i, then G is Hamiltonian.

8
In 2005 Rahman-Kaykobad gave a improved theorem of Ore’s Theorem. This theorem is
significant with respect to the related degree conditions like Dirac’s theorem, Ore’s theorem
for Hamiltonian paths and cycles in graphs.

Theorem 4(Rahman-Kaykobad,[22]): Let G=(V,E) be a connected graph with n vertices


such that for all pairs of distinct non-adjacent vertices u,v ∈ V we have du + dv + δ(u,v) ≥
n+1. Then G has a Hamiltonian path.

After this Rahman-Kaykobad again improved their theorem. Here they worked on for all
pairs of non-adjacent vertices.

Theorem 5,[18]: Suppose G = (V,E) is a 2-connected simple graph on n vertices. If for


all pairs of non-adjacent vertices di +dj +d(i,j)≥n+1 with equality holding for less than n-1/2
more pairs than number of pairs with inequality then G is Hamiltonian.

Sufficient conditions for a graph, or digraph, to have a Hamilton cycle usually take the form
of implicitly requiring many edges. One such digraph result was conjectured by C. Berge
and was proved by M. Alain Ghouila-Houri. We shall refer to this as the G-H theorem.

Theorem 6(Ghouila-Houri): If D is a strongly connected simple digraph and if id(v)≥ p/2


and od(v)≥ p/2 for all vertices v in D, then D has a hamilton cycle.

Haggkvist and Nicoghossian sharpened Diracs Theorem by incorporating the connectivity


of the graph into the degree bound.

Theorem 7(Haggkvist-Nicoghossian,[16]): If G is a 2 connected graph of order n, con-


nectivity k and minimum degree δ(G)≥ 31 (n +k) , then G is Hamiltonian.

This result itself was recently generalized.

Theorem 8(Haggkvist-Nicoghossian): If G is a 2 connected graph of order n and connec-


tivity k such that δ 3 (G) ≥ (n + k), then G is Hamiltonian.

A natural direction, taken by Bondy, was to further increase the number of vertices involved
in the independent set.
1
Theorem 9: If G is a k connected graph of order n ≥ 3 such that δ k+1 (G) > 2
(k+1)(n-1) ,
then G is Hamiltonian.

Bondy and Fan provided an Ore-type result for finding a dominating cycle, that is, a cy-
cle that is incident to every edge of the graph. Harary and Nash-Williams [149] showed
that the existence of a dominating cycle in G is essentially equivalent to the existence of a
hamiltonian cycle in the line graph of G, denoted L(G).

Theorem 10(Harary and Nash-Williams,[19]): Let G be a k connected ( k ≥ 2 ) graph of


order n. If any k + 1 independent vertices x i ( 0 ≤ i ≤ k ) with N(x i ) ∩ N(x j ) = φ ( θ ≤ i 6=
j ≤ k ) satisfy δ k+1 (G) ≥ n 2k,then G contains a dominating cycle.

9
Another very interesting approach was introduced by Fan. He showed that we need not
consider “all pairs of nonadjacent vertices”, but only a particular subset of these pairs.

Theorem 11: If G is a 2-connected graph of order n such that min { max ( deg u, deg v )|
dist ( u, v ) = 2 } ≥ n2 , then G is hamiltonian.

The first use of the generalized degree condition was to provide another generalization of
Diracs Theorem.

Theorem 12: If G is a 2-connected graph of order n such that deg(S) ≥ 2n−1 3


for each S =
{x, y} where x and y are independent vertices of G, then G is Hamiltonian.

Fraisse extended this result to larger independent sets of vertices.

Theorem 13(Fraisse,[11]): Let G be a k-connected graph of order n. Suppose there exists


some t ≤ k, such that for every independent set S of vertices with cardinality t we have
deg(S) ≥ t(n−1)
t+1
, then G is Hamiltonian.

Very recently, Lindquester was able to show that a Fan-type restriction to vertices at distance
two could also be used with generalized degrees, providing an improvement to Theorem.

Theorem 14(Lindquester): If G is a 2-connected graph of order n satisfying deg (S) ≥


2n−1
3
for every set S = {x, y} of vertices at distance 2 in G, then G is Hamiltonian.

Independent sets are not the only ones that have been useful in conjunction with generalized
degrees.The collection of all pairs of vertices (or all t-sets of vertices) provides yet another
generalization of Diracs Theorem.One with a more combinatorial flavor.

Theorem 15: If G is a 2-connected graph of sufficiently large order n such that deg(S) ≥ n2
for every set S of two distinct vertices of G, then G is Hamiltonian.

1.4 The Problem

We consider only simple graphs and hence neither self-loop nor multi-edges are present.
Suppose we have a graph G = (V,E) with n vertices. We sometimes use the notations V
|G| = V andE|G| = E. Two vertices, (u, v) ∈ V are said to be adjacent/neighbours to
each other if (u, v) ; otherwise, they are nonadjacent. The set of neighbours of a vertex u in
G is denoted by N(u). If G0 is a subgraph of G and u|G|, then N(u|G0 |) denotes the set of
neighbours of u (confined) in G0 . Now, d(u) = N(u) | and d[G0 ] = |Nu[G0 ]|. We use G0 to
denote the complimentary graph of G; that is, V d G e = V - V d G 0 e and E d G 0 e = E -
E d G 0 e. Sometimes, P is referred to as a u,v -path and u and v are referred to as the end
vertices or endpoints of P. Also, sometimes we use the notation P to denote the length of P.
So, by our definition, d P e = k - 1. If we have (x1 , xk ), then the graph C = ( V [ P ], E 0 )

10
such that E 0 = E [P] ∪ (x1 , xk ) is called a cycle. In what follows, we only consider simple
paths and simple cycles. A path P (cycle C) is called a Hamiltonian path (cycle) if V[P] =
V ( V [C] = V ) . Given a path P of G as defined above, assume that d ( x1 [ P ] ) 6= 0. Two
edges ( x1 , xi ) , ( xk , xj ) ∈ E, 1 ≤ k, are said to be crossover edges if and only if j = i -
1.
In a simple 2 connected-graph if G = ( V, E ) we need to find a Hamiltonian cycle using
distance as a parameter. Here, we maintain distance of all pair of non-adjacent vertices dij
= 4.

1.5 Scopes and objectives of the thesis

When a graph have a finite set of objects and some sets of connections between pair of
objects then we starts to find out that if these connection visits all the objects or not in a
efficient way. The requirement for efficiency is to visit all the objects only once and then if
the graph has a Hamiltonian path or not. When we start to visit from one objects and visits
all the objects. This problem is first considered in 1856 by Thosmas Kirkman.

Now if we assign value to the connection then it will become a famous unsolved problem
Travelling Salesman Problem (TSP). The Travelling Salesman Problem describes a sales-
man who must travel between N cities. The order in which he does so is something he does
not care about, as long as he visits each one during his trip, and finishes where he was at first.
Each city is connected to other close by cities, or nodes, by airplanes, or by road or railway.
Each of those links between the cities has one or more weights (or the cost) attached. The
cost describes how “difficult” it is to traverse this edge on the graph, and may be given, for
example, by the cost of an airplane ticket or train ticket, or perhaps by the length of the edge,
or time required to complete the traversal. The salesman wants to keep both the travel costs,
as well as the distance he travels as low as possible.

The Traveling Salesman Problem is typical of a large class of “hard” optimization problems
that have intrigued mathematicians and computer scientists for years. Most important, it
has applications in science and engineering. For example, in the manufacture of a circuit
board, it is important to determine the best order in which a laser will drill thousands of
holes. An efficient solution to this problem reduces production costs for the manufacturer.
The travelling salesman problem is regarded as difficult to solve. If there is a way to break
this problem into smaller component problems, the components will be at least as complex
as the original one. This is what computer scientists call NP-hard problems. In the theory
of computational complexity, the decision version of the TSP (where, given a length L, the
task is to decide whether the graph has any tour shorter than L) belongs to the class of NP-
complete problems. Thus, it is possible that the worst-case running time for any algorithm

11
for the TSP increases super polynomially (but no more than exponentially) with the number
of cities.

The problem was first formulated in 1930 and is one of the most intensively studied problems
in optimization. It is used as a benchmark for many optimization methods. Even though the
problem is computationally difficult, a large number of heuristics and exact algorithms are
known, so that some instances with tens of thousands of cities can be solved completely and
even problems with millions of cities can be approximated within a small fraction of 1.

Figure 1.3: Travelling Salesman Problem

The TSP has several applications even in its purest formulation, such as planning, logis-
tics, and the manufacture of microchips. Slightly modified, it appears as a sub-problem in
many areas, such as DNA sequencing. In these applications, the concept city represents, for
example, customers, soldering points, or DNA fragments, and the concept distance repre-
sents travelling times or cost, or a similarity measure between DNA fragments. The TSP
also appears in astronomy, as astronomers observing many sources will want to minimize
the time spent moving the telescope between the sources. In many applications, additional
constraints such as limited resources or time windows may be imposed. Hypohamiltonian
graphs arise in integer programming solutions to the traveling salesman problem: certain
kinds of hypohamiltonian graphs define facets of the traveling salesman polytope, a shape
defined as the convex hull of the set of possible solutions to the traveling salesman problem,
and these facets may be used in cutting-plane methods for solving the problem. Grötschel
(1980) observes that the computational complexity of determining whether a graph is hy-
pohamiltonian, although unknown, is likely to be high, making it difficult to find facets
of these types except for those defined by small hypohamiltonian graphs; fortunately, the
smallest graphs lead to the strongest inequalities for this application.

Concepts closely related to hypohamiltonicity have also been used by Park, Lim and Kim
(2007) to measure the fault tolerance of network topologies for parallel computing. There

12
are some other application of Hamiltonian cycle like-
1. Stripification of triangle meshes in computer graphics.
2. Electronic circuit design/construction.
3. Public transport
4. Tour planning
5. Design of microchips
6. Genome sequencing
7. Planning and decision making for aerial robots

1.6 Thesis Organization

In Chapter 2 the complexity of finding a hamiltonian cycle in a graph is discussed. Besides


that, existing sufficient conditions are described that helped to establish our new sufficient
condition.

In chapter 3 the new sufficient theorem that we established is discussed with proof. We
worked with the degree of non-adjacent vertices of a graph.

In chapter 4 conclusion and recommendation for future work is described. Here we de-
rived the findings of our established with and tried to figure out future possibilities with this
theorem.

13
CHAPTER 2
HAMILTONICITY

2.1 Complexity of the Recognition Problem

Finding a Hamiltonian path and Hamiltonian cycle in a graph is a NP-complete problem. To


show that it is a NP-complete problem, we will reduce vertex-cover to it, using a component
design type of reduction.

Figure 2.1: (a) An example of hamiltonian cycle shown in bold (b) A cover-enforcer sub-
graph He

Let G and k be a given instance of the Vertex-Cover problem. We will construct a graph
H that has a Hamiltonian Cycle if and only if G has a vertex cover of size k. We begin
by including a set of k initially disconnected vertices X = {x1 ,x2 ,....,xk } to H. This set of
vertices will serve as a “cover-choosing” component, for they will serve to identify which
nodes of G should be included in a vertex cover. In addition, for each edge e = (v,w) in G we
create a “cover-enforcer” subgraph He in H. This subgraph He has 12 vertices and 14 edges
as shown in figure 2.1b. Six of the vertices in the cover-enforcer He for e = (v,w) correspond
to v and the other six correspond to w. Moreover, we label two vertices in cover-enforcer
He corresponding to v as ve,top and ve,bot , and we label two vertices in He corresponding to
w as we,top and we,bot . These are the only vertices in He that will be connected to any other
vertices in H outside of He .

Thus a Hamiltonian cycle can visit the nodes of He in only one of three possible ways, as
shown in figure 2.2.

We join the important vertices in each cover-enforce He to other vertices in H in two ways,

14
Figure 2.2: Possible ways that a hamiltonian cycle can visit the edges in a cover-enforcer He

one that corresponds to the cover-choosing component and one that corresponds to the cover-
enforcing component. For the cover-choosing component, we add an edge from each vertex
in X to every vertex ve,top and every vertex ve,bot . That is, we add 2kn edges to H, where n is
the number of the vertices in G.
For the cover-enforcing component, we consider each vertex v in G in turn. For each such v,
let {e1 ,e2 ,....ed(v) } be a listing of the edges of G that are incident upon v. We use this listing
to create edges in H by joining vei ,bot in Hei to vei+1 ,top in Hei+1 , for i = 1,2,....,d-1. We refer
to the Hei components joined in this way as belonging to the covering thread for v. This
completes the construction of the graph H. Note that this computation runs in polynomial
time in the size of G.

We claim that G has a vertex cover of size k if and only if H has a hamiltonian cycle.
Suppose, first, that G has a vertex cover of size k. Let C = {vi1 ,vi2 ,....vik } be such a cover.
We construct a Hamiltonian cycle in H, by connecting a series of paths Pj , where each Pj
starts at xj and ends at xj+1 , for j = 1,2,.....,k-1, except for the last path Pk , which starts
at xk and ends at x1 . We form such a path Pj as follows. Start with xj , and then visit the
entire covering thread for vij in H, returning to xj+1 (or x1 if j =k). For each cover-enforcer
subgraph He in the covering thread for vij , which is visited in this Pj , we write without loss
of generality, e as (vij ,w). If w is not also in C, then we visit this Hei as in Figure 2.2a or
Figure 2.2c (with respect to vij ). Instead, if w is also in C, then we visit this Hei as in Figure
2.2b. In this way we will visit each vertex in H exactly once, since C is a vertex cover for
G. Thus, this cycle we construct is in fact a hamiltonian cycle.

Suppose, conversely, that H has a Hamiltonian cycle. Since this cycle must visit all the
vertices in X, we break this cycle up into k paths, P1 ,P2 ,....,Pk , each of which starts and
ends at a vertex in X. Moreover, by the structure of the cover-enforcer subgraphs He and the
way that we connected them, each Pj must traverse a portion (possibly all) of a covering
thread for a vertex v in G. Let C be the set of all such vertices in G. Since Hamiltonian cycle
must include the vertices from every cover-enforcer He and every such subgraph must be

15
Figure 2.3: (a) A vertex v in G and its set of incident edges (b) The connections made
between the Hei ’s in H for the edges incident upon v.

traversed in a way that corresponds to one of e’s endpoints, C must be a vertex cover in G.

Therefore, G has a vertex cover of size k if and only if H has a Hamiltonian cycle. This
gives us the theorem Hamiltonian cycle is NP-complete.

2.2 Sufficiency Condition

A graph is said to be Hamiltonian if it contains a spanning cycle. The spanning cycle is


called a Hamiltonian cycle of G and G is said to be a Hamiltonian graph. A Hamiltonian
path is a path that contains all the nodes in V(G) but does not return to the node in which
it began. No characterization of Hamiltonian graphs exists, yet there are many sufficient
conditions. We begin our investigation of sufficient conditions for Hamiltonicity with two
early results. The first is due to Dirac and the second is a result of Ore. Both results consider
this intuitive fact: the more edges a graph has, the more likely it is that a Hamiltonian cycle
will exist. Many sources on Hamiltonian theory treat Ore’s Theorem as the main result that
began much of the study of Hamiltonian graphs, and Dirac’s result a corollary of that result.
Diracs result actually preceded it, however, and in keeping with the historical intent of this
paper, we will begin with him.

2.2.1 Dirac’s Theorem

Theorem(Dirac,1952,[8]): If G is a simple graph with at least three vertices and δ(G) ≥


n(G)/2, then G is Hamiltonian.

16
Figure 2.4: Wheel Graph

As an illustration of Dirac’s Theorem, consider the wheel on six nodes, W6 in the above
figure. In this graph, δ = 3 ≥ ( 62 ) , so it is Hamiltonian. Traversing the nodes in numerical
order 1-6 and back to 1 yields a Hamiltonian cycle.

2.2.2 Ore’s Theorem

Theorem(Ore,1960,[20]): Let G be a simple graph.If u,v are distinct non-adjacent vertices


of G with d(u)+d(v)≥n(G),then G is Hamiltonian if and only if G+uv is Hamiltonian.

The wheel, W6 , also satisfies Ores Theorem. The sum of the degrees of nonadjacent nodes
(i.e., deg(2) + deg (5), or deg(3) + deg (6), etc.) is always 6, which is the order of the graph.

Before we discuss the results of Nash-Williams and Chvatal and Erdos, we must first define
the notions of connectivity and independence.

The connectivity κ = κ(G) of a graph G is the minimum number of nodes whose removal
results in a disconnected graph. For κ ≥ k, we say that G is k-connected. We will be
concerned with 2-connected graphs, that is to say that the removal of fewer than 2 nodes will
not disconnect the graph. For κ = k, we say that G is strictly k-connected. For clarification
purposes, consider the following. Let G be any simple graph, κ=3. Then G is 3-connected,
2-connected, and strictly 3-connected.

A set of nodes in G is independent if no two of them are adjacent. The largest number of
nodes in such a set is called the independence number of G, and is denoted by β . The fol-
lowing result by Nash-Williams builds upon the two previous results by adding the condition
that G be 2-connected and using the notion of independence.

17
2.2.3 Nash-Williams Theorem

Theorem(Nash-Williams,1971,[19]):Let G be a 2-connected graph of order n with δ(G) ≥


max{ (n+2)
3
, β}. Then G is Hamiltonian.

Figure 2.5: Hamiltonian Graph

The graph in Figure 2.5 demonstrates the Nash-Williams result. In this 2- connected graph
on six nodes, δ = 3, β = 2, and δ ≥ max { 6+2
3
,2}, implying Hamiltonicity.

In the same paper, Nash-Williams presents another very useful result. It is to be noted that a
cycle C is a dominating cycle in G if V(G - C) forms an independent set.

Theorem(Nash-Williams,1971,[19]):Let G be a 2-connected graph on n vertices with δ ≥


(n+2)
3
. Then every longest cycle is a dominating cycle.

Another sufficient condition uses the notion of a forbidden subgraph, i.e., a graph that can-
not be a subgraph of any graph under consideration. A subgraph of a graph G is a graph
having all of its nodes and edges in G. The following result by Goodman and Hedetniemi
introduces the connection between certain subgraphs and the existence of Hamiltonian cy-
cles. A bipartite graph G is a graph whose node set V can be partitioned into two subsets
V1 and V2 such that every edge of G joins V1 with V2.

Figure 2.6: k1,3 , k2,3

If G contains every possible edge joining V1 and V2, then G is a complete bipartite graph.
If V1 and V2 have m and n nodes, we write G = Km,n (see Figure 2.6).

18
2.2.4 Goodman and Hedetniemi Theorem

Goodman and Hedetniemi connected k1,3 , k1,3 +x -free graphs and Hamiltonicity in 1974. A
{k1,3 , k1,3 + x} -free graph is a graph that does not contain a k1,3 or a k1,3 +x (see Figure 2.7)
as an induced subgraph. (i.e., the maximal subgraph of G with a given node set S of V(G).)

Figure 2.7: k1,3 , k1,3 +x

Theorem(Goodman and Hedefniemi, 1974,[12]): If G is a 2-connected {k1,3 , k1,3 +x}-free


graph, then G is Hamiltonian.

The wheel, W6 , in Figure 2.4, is an example of a graph that is k1,3 , k1,3 +x -free. The subgraph
formed by node 1 and any three consecutive nodes on the cycle is k1,3 plus 2 edges. A year
after Nash-Williamss result, Chvátal and Erdos proved a sufficient condition linking the
ideas of connectivity and independence.

2.2.5 Chvátal and Erdos Theorem

Theorem(Chvátal and Erdos, 1972,[7]): Every graph G with n ≥ 3 and κ ≥ b has a Hamil-
tonian cycle.

Figure 2.8: Hamiltonian Graph

Chvátal and Erdoss result can be demonstrated by the graph in Figure 2.8. In this graph,
κ=2 and β=2.

Bondy and Chvátal [1976] phrased the essence of Ore’s argument in a much more general
form that yields sufficient conditions for cycles of length l and other subgraphs. Using Ore’s

19
theorem we can test whether a graph is Hamiltonian by testing whether the larger graph is
Hamiltonian.

2.2.6 Bondy and Chvátal’s Theorem

Theorem(Bondy-Chvátal,1976,[5]): A simple n-vertex graph is Hamiltonian if and only if


its closure is Hamiltonian.

Figure 2.9: Hamiltonian Closure

Bondy-Chvátal theorem can be demonstrated from Figure 2.9. By following the closure rule
we get a complete graph of k6 . As a complete graph is Hamiltonian thus the graph in first
step is also Hamiltonian.

2.2.7 Chvátal’s Theorem

Theorem(Chvátal,[6]): Let G be a simple graph with vertex degree d1 ≤....≤ dn , where n ≥


3. If i < n/2 implies that di > i or dn−i ≥ n - i, then G is Hamiltonian.

Proof: Adding edges to form a closure reduces no entry in the degree sequence. Also, G is
Hamiltonian if and only if C(G) is hamiltonian. Thus it suffices to consider the cases where
C(G)=G, which we describe by saying that G is closed. In this case, we prove that Chvátal’s
condition implies that G = Kn .

We prove the contrapositive; we prove that G is a closed n-vertex graph that is not a complete
graph, then we construct a value of i less than n/2 for which Chvátal’s condition is violated.
Violation means that at least i vertices have degree at most i and at least n - i vertices have
degree less than n - i.

With G 6= Kn , we choose among the pairs of nonadjacent vertices a pair u,v with maximum
degree sum. Because G is closed, u ↔ v implies that d(u) + d(v) < n. We choose the labels
on u,v so that d(u) ≤ d(v). Since d(u) + d(v) < n, it can be said that d(u) ≤ n/2. Let i = d(u).

We need to find i vertices with degree at most i. Because we chose a nonadjacent pair with

20
maximum degree sum, every vertex of V - {v} that is not adjacent to v has degree at most
d(u), which equals i. There are n - 1 - d(v) such vertices, and d(u) + d(v) ≤ n - 1 yields n - 1
- d(v) ≥ i.

We also need n - i vertices with degree less than n - i. Every vertex of V - {u} that is not
adjacent to u has degree at most d(v), and we have d(v) < n - d(u) = n - i. There are n - 1
- d(u) such vertices. Since d(u) ≤ d(v), we can also add u itself to the set of vertices with
degree at most d(v). We thus obtain n - i vertices with degree less than n - i.

We have proved that di ≤ i and dn−i ≤ n - i for this specially chosen i, which contradicts the
hypothesis.

2.2.8 Rahman-Kaykobad Theorem

Theorem(Rahman-Kaykobad,[22]): Let G = (V ,E) be a connected graph with n vertices


and P be a longest path in G having length k and with end vertices u and v. Then the
following statements must hold:

(a) Either δ(u, v) > 1 or P is a Hamiltonian path contained in a Hamiltonian cycle.

(b) If δ(u, v) ≥ 3 then dp (u) + dp (v) ≤ k δ(u, v) + 2.

(c) If δ(u, v) = 2, then either dp (u) + dp (v) ≤ k or P is a Hamiltonian path and there is a
Hamiltonian cycle.

Proof:

(a) Assume δ(u, v) ≤ 1. Since the graph is connected, δ(u, v) = 1. Let P ≡ h u =


u0 ,u1 ,u2 ,.....uk =v i. Since δ(u, v) = 1, we in effect have a cycle C ≡ h u = u0 ,u1 ,u2 ,.....uk =v,u0 =ui.

(b) Assume δ(u, v) ≥ 3. In this case surely dp (u) + dp (v) ≤ k - δ(u, v) + 2 ≤ k - 3 + 2 =


k - 1, since otherwise we would get a path from u to v with length less than δ(u, v), a
contradiction.

(c) Assume that δ(u, v) = 2. Now note that we cannot claim that dp (u) + dp (v) ≤ k -
δ(u, v) + 2 = k by arguing contradiction on δ(u, v) as we did in (b) because now
there is a common vertex adjacent to both u and v. However, we argue in a different
way as follows. Assume that dp (u) + dp (v) ≥ k + 1 = |V (P)|. We rewrite the path
P as follows: P ≡ h v = w1 ,w2 , . . . ,w|V (P )|−1 ,w|V (P )| = u i. Now we will try
to find out two crossover edges (v,wi+1 ) and (wi ,u) such that we get the cycle C =
hw1 ,wi+1 ,wi+2 ,.....,w|V (P )|−1 ,w|V (P )| , wi ,wi−1 ,...,w2 ,w1 i. To see that this is possible, let
us consider S ={j | (v,wj+1 ) ∈ E} and T = {j | (wj ,u) ∈ E}. Since S ∪ T ⊆ 1, 2, . . . ,

21
|V (P)| 1, we have |S ∪ T | ≤ |V(P)| - 1. Again because, |S| = dp (v), |T | = dp (u), and
dp (u) + dp (v) ≥ |V(P)|, we must have,

|S ∩ T | = |S| + |T | − |S ∪ T |
= dp (u) + dp (v) − |S ∪ T |
≥ dp (u) + dp (v) − (|V (P )| − 1)
≥ |V (P )| − |V (P )| + 1
=1

Hence S and T must have a common subscript so that the two crossover edges (v,wi+1 )
and (wi ,u) exist and we get the cycle C. So, in effect, we get a cycle C, which contains
a Hamiltonian path P0 = hw1 ,wi+1 ,wi+2 ,.....,w|V (P )|−1 ,w|V (P )| , wi ,wi−1 ,...,w2 i.

Corollary: Let G = (V ,E) be a connected graph with n vertices, and P be a longest path in
G having length k < n - 1 and with end vertices u and v. Then we must have dp (u) + dp (v) ≤
k - δ(u, v) + 2.

Proof: Since k < n - 1, P is not a Hamiltonian path. Hence by Theorem 1.5(a) we have δ(u,
v) > 1. Noting that if δ(u, v) = 2, then k = k - δ(u, v) + 2, by Theorem 1.5(b) and (c) we thus
have dp (u) + dp (v) ≤ k - δ(u, v) + 2.

Theorem(Rahman-Kaykobad,[22]): Let G=(V,E) be a connected graph with n vertices such


that for all pairs of distinct non-adjacent vertices u,v ∈ V we have du + dv + δ(u,v) ≥ n+1.
Then G has a Hamiltonian path.

Proof: This theorem can be proved by contradiction. Assume that the condition holds but
there is no Hamiltonian path in G. Then let P= (u = u0 ,u1 , . . . , uk = v) be a longest path in
G. Surely, k ≤ n - 2 and δ(u, v) ≤ k. Then by Corollary we must have dp (u) + dp (v) ≤ k -
δ(u, v) + 2. Now we have,

d(u) + d(v) + δ(u, v) = dp (u) + dp̄ (u) + dp (v) + dp̄ (v) + δ(u, v)
= {dp (u) + dp (v)} + dp̄ (u) + dp̄ (v) + δ(u, v)
≤ {k − δ(u, v) + 2} + dp̄ (u) + dp̄ (v) + δ(u, v)
= k + 2 + dp̄ (u) + dp̄ (v)
≤ n − 2 + 2 + dp̄ (u) + dp̄ (v)
= n + dp̄ (u) + dp̄ (v)

22
Since P is not a Hamiltonian path, by the previous Theorem , δ(u, v) > 1, i.e., u, v are
nonadjacent and hence we have d(u) + d(v) + δ(u, v) ≥ n + 1 according to our assumption.
We thus have,

n + dp̄ (u) + dp̄ (v) ≥ n + 1


⇒ dp̄ (u) + dp̄ (v) ≥ n − n + 1
⇒ dp̄ (u) + dp̄ (v) ≥ 1.

Hence there is at least one edge of the form (x, y) such that x ∈ {u, v} and y ∈ V (G) - V
(P) which means that we get a longer path in G by adding the edge (x, y) to P which is a
contradiction and the result follows.

2.2.9 Improvement of Rahman-Kaykobad Theorem

Theorem,[18]: Suppose G = (V,E) is a 2-connected simple graph with every non-adjacent


pair of vertices (i,j) ∈ E satisfying di + dj ≥ n - 1. Then G is Hamiltonian if | Eq | < | Es |
+ n/2.

Proof: Let us have a Hamiltonian path P(i,j). If di + dj > n - 1 then Ore’s condition asserts
existence of a Hamiltonian cycle. Without loss of generality let us assume that di + dj =
n-1. For simplicity of arguments let us call i left end vertex and j as right end vertex of the
Hamiltonian path. We note that if (i0 ,ik+1 ) ∈ E then (ik → i0 → ik+1 → in−1 ) is another
Hamiltonian path where ik is the left end vertex. So let L = {ik |(i0 , ik+1 ) ∈ E}. Similarly
let R = {ik |(i0 , ik−1 ) ∈ E}. Note that existence of edge (i,j) ∈ E with i ∈ L, and j ∈ R
implies that P(i,j) together with such an edge constitute a Hamiltonian Cycle. We note that
di is same for every i ∈ L. Similarly dj is same for every j ∈ R.
We consider the following two cases: (a) di < dj : Let us observe that L ∩ R = ∅. Moreover,
G(L) is a complete subgraph since otherwise for nonadjacent vertex pairs of G(L) degree
sum condition will be violated. Since |L| = di , degree of any vertex of G(L) must be di - 1.
So each vertex of L must be adjacent to a vertex of V\L ∩ R that contains a single vertex
v since |L ∪ R| = n - 1. Similarly since dj = |R|, G(R) is also a complete subgraph and
each vertex of R is adjacent to the same vertex v of V\R ∩ L. Now we have a Hamiltonian
Path with vertices of L ordered in any permutation then v followed by vertices of R in any
permutation. All vertices in L are adjacent to v and not to buy any vertex of R. In the same
way all vertices of R are adjacent to v and not to any vertex of L. Hence v is a cut-vertex that
contradicts the hypothesis That G is 2-connected. Hence there must be a vertex il of L that
is adjacent to a vertex of ir of R. Now since il and ir can be made respectively the left and
right end vertices of a Hamiltonian Path, together with (il ,ir ) this must form a Hamiltonian
cycle.

23
(b) Now in this case di = dj = d = |L| = |R|, and n = 2d + 1. Let us assume that
|L ∪ R| = d + k. So there are k vertices not common in L and R, and there are d - k vertices
common to both ends completes a Hamiltonian cycle together with hamiltonian Path with
those vertices as end vertices we assume that they form an independent set. They cannot
be adjacent to uncommon vertices as well. So |Eq | = d+1

2
. So only vertices in M can be
adjacent to these d - k + 1 common vertices. So degree of each of them cannot be more than
d - k + 1. But we know their degree is d. Hence k = 1, and each member of L ∪ R must
be adjacent to each member of M. So ∀i ∈ L ∪ R and j ∈ M, (i, j) ∈ E. Therefore, the
only other adjacent pairs can be in M. Since there are only d2 pairs of vertices |Es | = d2 .
 

Now |Ee | − |Es | = d+1 d


= d < n2 but hypothesis of the theorem asserts that
 
2
− 2
|Ee | < |Es | + n2 , Ee must be smaller indicating that there are some vertices in L ∪ R that are
adjacent. Such an edge together with an appropriate Hamiltonian Path ensures Hamiltonicity
of G.
This completes the proof.

24
CHAPTER 3
A NEW SUFFICIENCY CONDITION FOR
HAMILTONICITY

From Ore’s theorem, Dirac’s theorem, Chvatál’s theorem and through many other theorem
and preliminaries we get to know that in a 2-connected simple graph there cannot be any
cut vertex in Hamiltonian cycle because it contradicts the two connectedness of the graph.
Moreover, in a 2-connected simple graph if there exists any crossover edge it creates Hamil-
tonian cycles. We also get to know through a theorem that if we increase minimum distance
between two vertices then the degree sum will be minimum. Based on theorem, we intro-
duce our new theorem. Before this we need to discuss about crossover edge and already
established theorem.

3.1 Two Crossover

As our theorem is based on cross over edges, it is important to have a clear idea about
this. When two edges approaches from very most corner vertices and one falls on a vertex
immediate to the vertex where another falls two crossover formed. And from the Ores
theorem it is proved that when d(u) + d(v) ≥ n two crossover formed.

In Figure 3.1, the edges (u,u2) and (u1,v) together forms crossover edges. And this forces
the graph to be Hamiltonian.

Figure 3.1: Two Crossover Graph(a)

In Figure 3.2, even though another edge (u1,u4) is present but these does not affect the

25
crossover edge between (u,u3) and (u2,v). And this graph is also Hamiltonian.

Figure 3.2: Two Crossover Graph(b)

3.2 Overview of the theorem

Theorem: Let G = (V,E) be a 2-connected simple graph. Then G is Hamiltonian if for all
(i,j) ∈
/ E di + dj ≥ n - (dij - 1).

Proof: Let us assume that for this case the minimum distance between two nonadjacent
vertices, dij is 3. P(i,j) has p vertices. i,j cannot both be adjacent to the same vertex since
dij ≥ 3. Then di + dj ≤ p - 2. But di + dj ≥ n - 2. Then, we can say p=n . So, we always
have a Hamiltonian path. Even vertex of p must be adjacent to exactly once of i of d. Since
existence of cross over edge forces a Hamiltonian cycle.
We assume that there is no cross over edge. So, i must be adjacent to all neighbouring
vertices on p,so must be j. This forces existence of cut vertex which contradicts the 2-
connectedness of G. Hence there must be a crossover edge resulting in a Hamiltonian cycle.

3.2.1 Explanation with example

Let us assume a graph with n = 16 vertices. i and j are the left and right end vertices
respectively. These two vertices are non-adjacent and the minimum distance between them
is dij = 3.
As we prove the theorem by contradiction, we suppose that the graph is not hamiltonian. In
that case, there is no crossover edges. The vertex i can be connected to at most k vertices and
the vertex j can be connected to at most n - k vertices by maintaining distance = 3 (Figure
3.1).

Thus the degree of vertex i is di = k-1 and the degree of vertex j is dj = n - k -2. So, the
degree sum of i and j is di + dj ≥ (k - 1) + (n - k -1) ≥ n - 2 = n - (dij - 1). But this creates a
cut-vertex at (k+1). This contradicts the 2-connectedness of the theorem. So, there must be
a crossover edge to maintain the hypothesis of the theorem.

26
Figure 3.3: Non-Hamiltonian graph with distance=3

Here, the filled black circle is the cut-vertex.

Figure 3.4: Hamiltonian graph with distance=3

In Figure 3.4, an edge (i,k+2) is drawn which creates a crossover edge. To maintain distance
dij = 3 the edge (j,k+2) is removed. Because existence of this edge decreases the distance
to 2. So the degree sum of i and j remains same and that is di + dj ≥ n - 2.

3.3 Proposal of new theorem

Theorem: Let G=(V,E) be a 2-connected simple graph. Then G is Hamiltonian if for all
pair of non-adjacent vertices (i,j) ∈
/ E, di + dj ≥ n - 4.

Proof: Let us assume that for this case the minimum distance between two nonadjacent
vertices i and j, dij is 4. P(i,j) has p vertices. i,j cannot be both adjacent to the same vertex
as well as the neighbouring vertex since dij ≥ 4.
Here, P is a Hamiltonian path and we also assume that p(i,j) has 4 hops (h1 = i → h2 →
h3 → h4 → h5 = j). Since, we know existence of crossover edge causes Hamiltonian cycle
so we assume that there is no cross over edge. So i must be adjacent to all neighbouring
vertices and so must be j. This forces existence of cut vertex which contradicts the 2-
connectedness of G. So, there must be a cross over edge which results in a Hamiltonian

27
cycle.
This proves that the graph is Hamiltonian.

3.3.1 Explanation with example

Let us assume a graph with n = 16 vertices. i and j are the left and right end vertices
respectively. These two vertices are non-adjacent and the minimum distance between them
is dij = 4.
As we prove the theorem by contradiction, we suppose that the graph is not Hamiltonian. In
that case, there is no crossover edges. The vertex i can be connected to at most k vertices
and the vertex j can be connected to at most n - k - 1 vertices by maintaining distance = 4
(Figure 3.5).

Figure 3.5: Non-Hamiltonian graph with distance=4

Thus the degree of vertex i is at most di ≥ k-1 and the degree of vertex j is at most dj ≥ n -
k - 2. So, the degree sum of i and j is di + dj ≥ (k - 1) + (n - k -2) ≥ n - 3. But this creates a
cut-vertex at (k+1). This contradicts the 2-connectedness of the theorem. So, there must be
a crossover edge to maintain the hypothesis of the theorem.

Figure 3.6: Hamiltonian graph with distance=4

28
3.4 Findings

When we add an edge from i to k+2 then we need to remove the edge from k+2 to j and also
from k+3 to j to maintain dij = 4. So we can see that di is increase by 1 and dj is decrease by
2. When the distance of all pair of non adjacent vertices is dij = 3 then the degree sum is di
+ dj = n - 2 and when the distance of all pair of non adjacent vertices is dij = 4 then the total
degree sum is di + dj =n-4. This means that by increasing the value of distance of all pair of
non adjacent vertices total degree sum decreases. Condition of a graph to be Hamiltonian
can be more efficient by increasing distance of pair of non adjacent vertices.

29
CHAPTER 4
CONCLUSION AND RECOMMENDATION

4.1 Conclusion

Hamiltonian cycle is a important topic in the area of graph theory,algorithm and complexity
theory. Through theorem Hamiltonian cycle can be used to proof complexity. Our goal
of the research is to know about various condition of Hamiltonian cycle. In chapter 1, we
have seen the history of Hamiltonian cycle , some basic definition and notations , scope
and objective of the thesis. In chapter 2, we have seen the complexity of the recognition
problem and some sufficient condition of Hamiltonian cycle and its proof. In the chapter 3,
we propose a new condition of a Hamiltonian cycle for dij = 4 the total degree sum di + dj ≥
n-4. So, the theorem become more generalized and more efficient from previous condition.

4.2 Recommendation

Here we are trying to find a condition a graph will be Hamiltonian for distance dij =4. If it
is possible to find a sufficient condition for distance dij ≥5 then the condition will be more
efficient. In future we will try to increase the efficiency of the condition of Hamiltonian cycle
by increasing distance as through previous knowledge we get to know that by increasing
distance total degree sum will be decrease.

30
Bibliography

[1] Barr, O. (1998). Properly coloured hamiltonian paths in edge-colored complete graphs
without monochromatic triangles. Ars Combinatoria 50, pages 316–318.

[2] Bauer, D., Broersma, H., and Schmeichel, E. (July 2002). More progress on tough
graphs - the y2k report. Electronic Notes in Discrete Math. 11.

[3] Biggs, N. L. (1981). T.p. kirkman, mathematician. The Bulletin of the London Mathe-
matical Society, 13 (2), page 97120.

[4] Bȯhme, T., Harant, J., and Tkac, M. (1999). More than 1-tough chordal planar graphs
are hamiltonian. J. Graph Theory 32, pages 405–410.

[5] Bondy, J. and Chvátal, V. (1976). A method in graph theory. Discrete Math. 15, pages
111–136.

[6] Chvátal, V. (1972). On hamiltons ideals. J. Combinatorial Theory Ser. B 12, pages 163
– 168.

[7] Chvátal, V. and Erdos, P. (1972). A note on hamiltonian circuits. Discrete Math. 2,
pages 111–113.

[8] Dirac, G. (1952). Some theorems on abstract graphs. Proc. London Math. Soc. 2, pages
69–81.

[9] Duffus, D., Gould, R. J., and Jacobson, M. (1981). Forbidden subgraphs and the hamil-
tonian theme. The Theory and Applications of Graphs, ed. by Chartrand, Alavi, Gold-
smith, Lesniak and Lick, pages 297–306.

[10] Ellingham, M., Zha, X., and Zhang, Y. (2004). Spanning 2-trails from degree sum
conditions. J. Graph Theory, Vol. 45, pages 298–319.

[11] Fraisse, P. (1986). A new sufficient condition for hamiltonian graphs. J. of Graph
Theory Vol. 10, No. 3, pages 405–409.

[12] Goodman, S. and Hedetniemi, S. (1974). Sufficient conditions for a graph to be hamil-
tonian. J. Combin Theory.

31
[13] Gould, R. J. (2003). Advances on the hamiltonian problem - a survey. Graphs and
Combinatorics, Vol. 19, Num. 1, pages 7 – 52.

[14] Gould, R. J. and Hynds, E. (1999). A note on cycles in 2-factors of line graphs. Bull.
of the I.C.A., Vol.26, pages 46 – 48.

[15] H. Kierstead, G. S. and Selkow, S. (1999). On k-ordered hamiltonian graphs. J. Graph


Theory, 32, pages 17 – 25.

[16] Haggkvist, R. and Nicoghossian, C. G. (1981). A remark on hamiltonian cycles. J.


Combinat. Theory B 30, pages 118–120.

[17] Horak, P. and Stacho, L. (2000). A lower bound on the number of hamiltonian cycles.
Discrete Math.222, pages 275–280.

[18] Kaykobad, T. and Kaykobad, M. (2015). Hamiltonicity revisited.

[19] Nash-Williams, C. S. J. A. (1971). Edge-disjoint hamiltonian circuits in graphs with


vertices of high valency. Studies in Pure Mathematics.

[20] Ore, O. (1960). Note on hamiltonian circuits. Amer. Math. Monthly 67, page 55.

[21] Plotnikov, A. D. (1998). One criterion of existence of a hamiltonian cycle. Reliable


Comput. 4, pages 199 – 202.

[22] Rahman, M. S. and Kaykobad, M. (2005). On hamiltonian cycles and hamiltonian


paths. Information Processing Letters 94, pages 37–51.

[23] Rahman, M. S., Kaykobad, M., and Rahman, M. S. (2005). A new sufficient condition
for the existence of hamiltonian paths. 20th International Conference on Computers and
their Applications (CATA).

32

You might also like