A Seminar on Graph Theory

THEORY

[1]

Some Theorems and Concepts of Graph Theory

The field of graph theory is even today being independently rediscovered by scholars in various disciplines as they require its techniques for the solution of their structural and combinatorial problems. In recent years, this subject has been receiving increasing attention and interest in mathematics and other sciences. The stimulus to the renaissance of graph theory stems from its wide applicability to many fields. In fact, the theory of graphs is now quite important not only in pure mathematics (combinatorial theory), but also in theoretical physics, electrical engineering, organic chemistry, social psychology, and operational research.

It is our hope that the presentation of some historical comments and illustrations, with the statements of selected theorems, will set the stage for the remaining lectures.

We will describe four independent approaches to graph theory:

(1) Euler: Königsberg Bridge Problem.

(2) Hamilton: Around the World.

(3) Cayley, Heawood: Four-color Conjecture.

(4) Kirchhoff, Cayley, Jordan, Sylvester: Trees.

Königsberg Bridge Problem

Euler (1707-1782) became the father of graph theory, as well as the rest of topology, when he settled a famous problem of his day called the Königsberg bridge problem [4]. The area in question contains two islands linked to each other and to the banks of the Pregel River by seven bridges, as in Fig. 1.1.

The problem was to begin at any of the four land areas and, without swimming or flying or travelling around the world, cross each bridge exactly once and return to the starting point. One can see immediately that there are many ways of attempting this problem without solving it.

The considerable contribution of Euler in this case was negative, for he proved that the problem is unsolvable. He replaced each land area by a point and each bridge by a line joining the corresponding points. The result, shown in Fig. 1.2, is a collection of points with lines joining them. The points, labeled a, b, c, and d, correspond to the land areas of Fig. 1.1.

Fig. 1.1. The Königsberg bridge problem.

Fig. 1.2. The multigraph of the Königsberg bridge problem.

Fig. 1.3. A (4, 5) graph with points and lines labeled.

Since almost any two graph theorists use different terminology, and since we wish this exposition to be self-contained, we will preface a precise formulation of the problem with a rather substantial list of definitions.

A graph G consists of a finite nonempty set V of points and a set X of lines, each of which joins two distinct points. We assume that distinct lines do not join the same pair of points; otherwise, as in Fig. 1.2, the configuration is a multigraph. Furthermore, if we permit loops, that is, lines joining a point with itself, the result is a general graph. A (p, q) graph has p points and q lines. The number of points in a graph is called its order. Fig. 1.3 shows a (4, 5) graph with points and lines labeled, i.e., designated by symbols.

The two points joined by a line are adjacent, and each is incident with the line. Two graphs are isomorphic if there is a one-to-one correspondence between their sets of points preserving adjacency. The degree d(v) of a point v is the number of lines incident with it. A point of degree 0 is isolated, and the graph consisting of such a point is called trivial. In a regular graph all points have the same degree. A regular graph in which every point has degree 3 is called cubic. The complete graph Kp of order p has every pair of its points adjacent and so is regular of degree p – 1. All the complete graphs with up to five points are shown in Fig. 1.4.

A walk in a graph G is an alternating sequence of points and lines of G, beginning and ending with a point, in which each line is incident with the point preceding it and the point following it. A walk of the form v1, x1, v2, x2, v3 · · ·, vn is said to join v1 with vn. The length of a walk is the number of occurrences of lines in it. A trail is a walk in which all lines are distinct. A path is a walk in which all points (and hence all lines) are distinct. A closed walk has the same first and last points. The walk v1, x1, v2, x2, · · ·,vn is often written v1v2· · ·vn, the lines being evident by context. A cycle is a closed walk v1v2· · ·vnv1, n ≥ 3, in which the n points vi are distinct. A spanning walk (spanning path, spanning cycle, and so on) contains all the points of G.

Some of these concepts are illustrated in Fig. 1.3. The walk v1v2v3v4. is a path; in fact, it is a spanning path. The cycle v1v2v4v1 has length 3, and v2v1v4v2v3vA, is a trail containing all lines.

In a connected graph, every pair of distinct points is joined by a path. A subgraph of G consists of subsets of V and X which themselves form a graph.

Fig. 1.4. The smallest complete graphs.

A component of G is a maximal connected subgraph. Obviously, G is connected if and only if it has exactly one component.

A multigraph is called eulerian if it has a spanning closed trail that contains all its lines; such a trail is also called eulerian. In these terms, the Königsberg bridge problem asks whether the multigraph in Fig. 1.2 is eulerian.

Several propositions may now be stated.

PROPOSITION 1. A walk joining two points contains a path joining them.

PROPOSITION 2. In a (p, q) graph, the sum of the degrees of the points is 2q.

PROPOSITION 3. In any graph, the number of points of odd degree is even.

PROPOSITION 4. No cycle is properly contained in another cycle.

PROPOSITION 5. If two cycles Z1 and Z2 contain a common line x, and if y is a line of Z1 but not of Z2, then their union includes a cycle which contains y but not x.

PROPOSITION 6. A nontrivial graph G is connected if and only if, for any partition of V into subsets U and W, there is a line joining a point of U with a point of W.

PROPOSITION 7. Each component of a regular graph of degree 2 is a cycle. Each component of a regular graph of degree 1 is K2.

PROPOSITION 8. (Euler [4]). The following statements are equivalent for any connected graph G:

(1) G is eulerian.

(2) Every point has even degree.

(3) G has a set of cycles which partitions the set of lines of G.

COROLLARY 8. Every connected graph has a closed walk in which each line appears exactly twice.

Ore [9] studied a special class of eulerian graphs. An eulerian graph G is called arbitrarily traversable from point v, or briefly v-traversable, if the following procedure always results in an eulerian trail. Start at the given point v and traverse any line. On arriving at a point, choose any line which has not yet been traversed. When no new lines are available, the procedure terminates. For example, the graph of Fig. 1.5 is arbitrarily traversable only from point v.

PROPOSITION 9. (Ore [9]). An eulerian graph G is v-traversable if and only if every cycle of G contains the point v.

Fig. 1.5. A v-traversable graph.

COROLLARY 9a. If G is v-traversable, then v has maximum degree.

COROLLARY 9b. If H is an eulerian subgraph of a v-traversable graph, then H is also v-traversable.

Around the World

The hamiltonian game*, invented by Sir William Hamilton, uses a regular solid dodecahedron whose twenty vertices are labeled with the names of various important cities. The player is supposed to travel Around the World by finding a cycle that passes through each vertex exactly once. Hamilton sold this idea to a game manufacturer in Dublin for about twenty-five guineas, which was wise of him since it was not a commercial success. Of course, the object of the game is to find a spanning cycle in the graph of the dodecahedron shown in Fig. 1.6. The points of this graph are marked 1, 2, · · ·, 20 (rather than London, New York, Paris, Tokyo, . . . .), so that a spanning cycle is evident.

A graph is called hamiltonian if it has a spanning cycle. There are several necessary conditions known for a graph to be hamiltonian, as well as some sufficient conditions, but as yet there is no convenient combinatorial criterion for a graph to be hamiltonian.

The removal of a point v from a graph G results in that graph G – v which is the maximal subgraph of G not containing v. Similarly, the graph G – x obtained on removing a line x is the maximal subgraph of G not containing x. Thus G – x contains all the points of G and all the lines of G except x, and G – v contains all points of G except v and all the lines of G except those incident with v. If all graphs obtained from G by removing one line are isomorphic, we denote any such graph by G – x.

*See Ball and Coxeter [1] for an amusing and more complete description.

Fig. 1.6. Around the world.

A cutpoint of a connected graph G is a point whose removal results in a disconnected graph; line x is a bridge of G if G – x is disconnected. A block is a connected graph with no cutpoints. A block of a graph is a maximal connected subgraph having no cutpoints. A theta-graph is a block with two nonadjacent points of degree 3 and all other points of degree 2.

PROPOSITION 10. Every hamiltonian graph is a block.

PROPOSITION 11. Every nontrivial nonhamiltonian block has a theta-subgraph.

PROPOSITION 12. (Ore [10]). If G is a graph with p ≥ 3 points so that, for every pair of nonadjacent points u and v, the sum of the degrees d(u) + d(v) ≥ p, then G is hamiltonian.

COROLLARY 12. (Dirac [3], Newman [8]). A graph with p ≥ 3 points is hamiltonian if the degree of every point is at least p/2.

PROPOSITION 13. (Smith and Tutte [11]). In any cubic graph G, the symmetric difference of all the spanning cycles of G, regarded as sets of lines, is empty.

COROLLARY 13. If a cubic graph is hamiltonian, it has at least three spanning cycles.

The Four-color Conjecture

The most famous unsolved problem in graph theory, and probably in all of mathematics, is the four-color conjecture. This remarkable problem can be explained in five minutes by any mathematician to the so-called man in the street. At the end of the explanation, neither will be able to solve the problem, but both will understand its statement.

A most interesting account of the history of this conjecture is given in Coxeter [2]. At a meeting of the London Mathematical Society in 1878, the eminent mathematician A. Cayley asked, Has a solution been given of the statement that in colouring a map of a country, divided into counties, only four distinct colours are required, so that no two adjacent counties should be painted in the same colour? Its first