graph theory
- Key People:
- Paul Erdős
- Endre Szemerédi
- Leonard M. Adleman
- Avraham Trahtman
- Related Topics:
- graph
- traveling salesman problem
- Hamilton circuit
- network theory
- linear graph
graph theory, branch of mathematics concerned with networks of points connected by lines. The subject of graph theory had its beginnings in recreational math problems (see number game), but it has grown into a significant area of mathematical research, with applications in chemistry, operations research, social sciences, and computer science.
The history of graph theory may be specifically traced to 1735, when the Swiss mathematician Leonhard Euler solved the Königsberg bridge problem. The Königsberg bridge problem was an old puzzle concerning the possibility of finding a path over every one of seven bridges that span a forked river flowing past an island—but without crossing any bridge twice. Euler argued that no such path exists. His proof involved only references to the physical arrangement of the bridges, but essentially he proved the first theorem in graph theory.
As used in graph theory, the term graph does not refer to data charts, such as line graphs or bar graphs. Instead, it refers to a set of vertices (that is, points or nodes) and of edges (or lines) that connect the vertices. When any two vertices are joined by more than one edge, the graph is called a multigraph. A graph without loops and with at most one edge between any two vertices is called a simple graph. Unless stated otherwise, graph is assumed to refer to a simple graph. When each vertex is connected by an edge to every other vertex, the graph is called a complete graph. When appropriate, a direction may be assigned to each edge to produce what is known as a directed graph, or digraph.
An important number associated with each vertex is its degree, which is defined as the number of edges that enter or exit from it. Thus, a loop contributes 2 to the degree of its vertex. For instance, the vertices of the simple graph shown in the diagram all have a degree of 2, whereas the vertices of the complete graph shown are all of degree 3. Knowing the number of vertices in a complete graph characterizes its essential nature. For this reason, complete graphs are commonly designated Kn, where n refers to the number of vertices, and all vertices of Kn have degree n − 1. (Translated into the terminology of modern graph theory, Euler’s theorem about the Königsberg bridge problem could be restated as follows: If there is a path along edges of a multigraph that traverses each edge once and only once, then there exist at most two vertices of odd degree; furthermore, if the path begins and ends at the same vertex, then no vertices will have odd degree.)
Another important concept in graph theory is the path, which is any route along the edges of a graph. A path may follow a single edge directly between two vertices, or it may follow multiple edges through multiple vertices. If there is a path linking any two vertices in a graph, that graph is said to be connected. A path that begins and ends at the same vertex without traversing any edge more than once is called a circuit, or a closed path. A circuit that follows each edge exactly once while visiting every vertex is known as an Eulerian circuit, and the graph is called an Eulerian graph. An Eulerian graph is connected and, in addition, all its vertices have even degree.
In 1857 the Irish mathematician William Rowan Hamilton invented a puzzle (the Icosian Game) that he later sold to a game manufacturer for £25. The puzzle involved finding a special type of path, later known as a Hamiltonian circuit, along the edges of a dodecahedron (a Platonic solid consisting of 12 pentagonal faces) that begins and ends at the same corner while passing through each corner exactly once. The knight’s tour (see number game: Chessboard problems) is another example of a recreational problem involving a Hamiltonian circuit. Hamiltonian graphs have been more challenging to characterize than Eulerian graphs, since the necessary and sufficient conditions for the existence of a Hamiltonian circuit in a connected graph are still unknown.
The histories of graph theory and topology are closely related, and the two areas share many common problems and techniques. Euler referred to his work on the Königsberg bridge problem as an example of geometria situs—the “geometry of position”—while the development of topological ideas during the second half of the 19th century became known as analysis situs—the “analysis of position.” In 1750 Euler discovered the polyhedral formula V – E + F = 2 relating the number of vertices (V), edges (E), and faces (F) of a polyhedron (a solid, like the dodecahedron mentioned above, whose faces are polygons). The vertices and edges of a polyhedron form a graph on its surface, and this notion led to consideration of graphs on other surfaces such as a torus (the surface of a solid doughnut) and how they divide the surface into disklike faces. Euler’s formula was soon generalized to surfaces as V – E + F = 2 – 2g, where g denotes the genus, or number of “doughnut holes,” of the surface (see Euler characteristic). Having considered a surface divided into polygons by an embedded graph, mathematicians began to study ways of constructing surfaces, and later more general spaces, by pasting polygons together. This was the beginning of the field of combinatorial topology, which later, through the work of the French mathematician Henri Poincaré and others, grew into what is known as algebraic topology.
The connection between graph theory and topology led to a subfield called topological graph theory. An important problem in this area concerns planar graphs. These are graphs that can be drawn as dot-and-line diagrams on a plane (or, equivalently, on a sphere) without any edges crossing except at the vertices where they meet. Complete graphs with four or fewer vertices are planar, but complete graphs with five vertices (K5) or more are not. Nonplanar graphs cannot be drawn on a plane or on the surface of a sphere without edges intersecting each other between the vertices. The use of diagrams of dots and lines to represent graphs actually grew out of 19th-century chemistry, where lettered vertices denoted individual atoms and connecting lines denoted chemical bonds (with degree corresponding to valence), in which planarity had important chemical consequences. The first use, in this context, of the word graph is attributed to the 19th-century Englishman James Sylvester, one of several mathematicians interested in counting special types of diagrams representing molecules.
Another class of graphs is the collection of the complete bipartite graphs Km,n, which consist of the simple graphs that can be partitioned into two independent sets of m and n vertices such that there are no edges between vertices within each set and every vertex in one set is connected by an edge to every vertex in the other set. Like K5, the bipartite graph K3,3 is not planar, disproving a claim made in 1913 by the English recreational problemist Henry Dudeney to a solution to the “gas-water-electricity” problem. In 1930 the Polish mathematician Kazimierz Kuratowski proved that any nonplanar graph must contain a certain type of copy of K5 or K3,3. While K5 and K3,3 cannot be embedded in a sphere, they can be embedded in a torus. The graph-embedding problem concerns the determination of surfaces in which a graph can be embedded and thereby generalizes the planarity problem. It was not until the late 1960s that the embedding problem for the complete graphs Kn was solved for all n.
Another problem of topological graph theory is the map-colouring problem. This problem is an outgrowth of the well-known four-colour map problem, which asks whether the countries on every map can be coloured by using just four colours in such a way that countries sharing an edge have different colours. Asked originally in the 1850s by Francis Guthrie, then a student at University College London, this problem has a rich history filled with incorrect attempts at its solution. In an equivalent graph-theoretic form, one may translate this problem to ask whether the vertices of a planar graph can always be coloured by using just four colours in such a way that vertices joined by an edge have different colours. The result was finally proved in 1976 by using computerized checking of nearly 2,000 special configurations. Interestingly, the corresponding colouring problem concerning the number of colours required to colour maps on surfaces of higher genus was completely solved a few years earlier; for example, maps on a torus may require as many as seven colours. This work confirmed that a formula of the English mathematician Percy Heawood from 1890 correctly gives these colouring numbers for all surfaces except the one-sided surface known as the Klein bottle, for which the correct colouring number had been determined in 1934.
Among the current interests in graph theory are problems concerning efficient algorithms for finding optimal paths (depending on different criteria) in graphs. Two well-known examples are the Chinese postman problem (the shortest path that visits each edge at least once), which was solved in the 1960s, and the traveling salesman problem (the shortest path that begins and ends at the same vertex and visits each edge exactly once), which continues to attract the attention of many researchers because of its applications in routing data, products, and people. Work on such problems is related to the field of linear programming, which was founded in the mid-20th century by the American mathematician George Dantzig.