Vous êtes sur la page 1sur 10

CHROMATIC NUMBER

Desislava Kukova*
*30 Car Osvoboditel STR, ap. 7, 9000 Varna, Bulgaria, desi_kukova@abv.bg

ABSTRACT
The theme of the chromatic number is common mainly due to the Four Color Conjecture in Mathematics. The chromatic number is the most popular graphical invariant. Despite its fame, we know rather little about chromatic number because it is difficult to determine it. This project is dedicated to some theorems and problems on chromatic number in graph theory. Some basic characteristics and definitions from this theory are given. A lot of coloring problems have been exposed, which are good illustration to what chromatic number is. The chromatic number of a graph is the smallest number of colors needed to color the vertices of so that no two adjacent vertices share the same color. Convenient method is not known for determining the chromatic number of an arbitrary graph. But the chromatic number is known for some familiar classes of graphs: 1. Complete graph has . 2. Cycle graph has for even and for odd. 3. Star graph has . 4. Wheel graph has for even and for odd. The current work deals with an attempt to present some important ideas in this area. The main ideas developed in the project are: critical graphs; perfect graphs; bipartite graphs; planar graph; graphs with small degrees have small chromatic numbers; constructions for these graphs; appropriate coloring of vertices; planar maps and chromatic polynomials.

PART I DEFINITIONS The chromatic number of a graph is the smallest number of colors needed to color the vertices of so that no two adjacent vertices share the same color. Convenient method is not known for determining the chromatic number of an arbitrary graph. A good coloration of graph is an assignment of colors (usually one of the integers ) to the points such that every edge meets at least two different colors Some useful facts: graph complete graph graph complement cycle wheel star tree Some important observations: Every complete graph which has vertices Graph which has chromatic number Graph which has chromatic number
5

Chromatic number

, for , for , for , for

even odd even odd

is non-planar. is a null graph. is non-null bipartite.


1

Examples:

pic. 1

pic. 2

pic. 3 4

pic. 4

pic. 5

pic. 6

PART II PROBLEMS ASSOCIATED DIRECTLY WITH CHROMATIC NUMBER Problem1: If every point of a graph Solution: Let , by using induction we can has degree at most , then -color it. Vertex is adjacent to at most .

is a vertex of . Since every point of

has degree at most

points, therefore we can find a color, which does not occur among the neighbors of . If we give this color, we get a good coloration of . Problem2: Prove that for any graph such that Solution: Consider any good coloration of union of with induces a , of color classes and coloration of and a . On the other hand, , we can find a partition . with colors, let -coloration of be the : thus

the rest. The

-coloration we started cannot be colored by less than .

by less than , colors, but since such a coloration, together with the coloration we already have, would give a coloration of and similarly If Suppose is that not complete, such that a we can colors. Then Problem3: Solution:

find

a .

partition

contains and . Since .. We will show that

maximal clique Let . We will prove that is a clique, we have . Suppose that

admits a coloration, using colors . We can use colors +1,, to color . We will show that we can get rid of one of these colors (color ). Now let look at that set of those points of color which are neighbors of the point with color . Since is a maximal complete subgraph, each vertex is non-adjacent to a certain vertex Now re-color the graph giving to each the color of . This is another good coloration, because the color of does not occur anywhere else, and the new vertices of this color are independent. In this new coloration is the only point of color and has no neighbor of color . Then if we recolor by color , we get a good coloration by colors, which is a contradiction. Problem4: Suppose that that, for each then Solution: By and has a partition , there exist non-adjacent vertices: on , , such ,

we might assume that Let be a coloration of with colors , and extend to a coloration of graph by coloring the vertices of with colors . We want to re-color the graph, so as to get rid of one of those colors. Assume that each contains a vertex , which is non-adjacent to a certain vertex of There must be a color which occurs in the set only, then the number of points outside those is . Let is this color. Now re-color the points such that with the color of . Then the resulting coloration is good and uses colors only. Problem5: If is a simple graph on points then the following inequalities hold true: and . Solution: Consider a -coloration of graph with color classes Then any two color classes from are connected by an edge in , otherwise they could be merged. Thus graph satisfies the

induction

conditions in problem 4 and so first inequality .

. Hence the . The second is trivial because:

Problem6: If a digraph has no path of length , then . Solution: Suppose first that has no cycles. Let denote the maximum length of a path starting from vertex . Then is a coloration with colors For suppose that there were two adjacent vertices and which share the same color and let is a path staring from vertex with length , which does not go through , because has no cycles. Hence is a path of length starting from vertex . Hence , which is a contradiction. Observe that in this coloration, no path joins points of the same color. Now let is a minimum set of edges whose removal destroys all cycles. Then graph has no cycles and has the -coloration constructed above. Since is no longer acyclic, it has a cycle through . So has a path connecting the endpoints of . This implies that the endpoints of have different colors. Then, the -coloration of graph is a good coloration of graph . Problem7: Let be the maximum number of points in a clique of . If graph and graph both satisfy , then so does their strong product . Solution: Let be a maximum complete subgraph of . Then spans a complete spans a complete subgraph of of size . Hence . On the other hand, let be -coloration of . Define a coloration by for . It is easy to verify that is a good coloration of . stated: . and uses colors. Therefore Thus . We have equality

PART III CONSTRUCTIONS FOR GRAPHS A graph is called critical if on removing any edge, the resulting graph has smaller chromatic number: , for each edge . Problem8: Construct a critically 3-chromatic graph. Solution: Any cycle graphs with odd degrees have 3 for chromatic number. We can easily see that these graphs are critical. PART IV APPROPRIATE COLORING OF VERTICES CHROMATIC POLYNOMIALS The chromatic polynomial denoted by of a graph G is the number of good -colorations of . This turns out to be a polynomial in (for fixed ) so its definition can be extended to all real values of . Note that two colorations differing in the labeling of colors count as different. Relation between chromatic number and chromatic polynomial: Actually the chromatic number is the smallest positive integer , such that Problem9: Determine the chromatic polynomial of a tree.
Solution: Let is a tree and is a vertex from with degree one. Any -coloration of , can be extended to in ways. Hence Since the chromatic polynomial of the one-point tree is , we have . pic. 7

PLANAR GRAPH OR PLANAR MAPS In mathematics, a map can be consider as a planar map (planar graph): every capital of contiguous country
pic. 8

is a vertex and an edge connects capitals of adjacent countries. Two countries are called adjacent if they share a border segment, not just a point. From the definition of a map follows directly that the minimum number of colors required for coloring a map is actually the chromatic number of this map.

MAPS PROBLEMS

Problem10: Given a planar 3-regular 2-connected, we can 4-color the faces in such way that adjacent faces have different colors if and only if there is a good coloration of the edges with 3 colors. Solution: Suppose first that the faces are colored with 1, 2, 3 and 4 colors. We will give color to edges adjacent to faces of color 1 and 2, or 3 and 4; give color to edges adjacent to faces of color 1 and 3, or 2 and 4; give color to the rest edges. This coloration of the edges shoe that the edges can be colored with 3 colors. Now, let us consider a 3 coloration of the edges. Let be the map formed by red and blue edges and edges. Clearly, and be the map formed by red and green can can be colored consists of disjoint circuits. Hence the faces of

be colored with two colors: red and green; also the faces of

with two colors: light and dark. Now use light red, light green, dark red, dark green to color the faces of map . Each face of is contained in a face of and in a face red, if the map . Problem11: Assume that a planar map has a Hamiltonian circuit. Show that its chromatic number is 4. (Show that its faces have a good coloration with 4 colors.) Solution: Let the planar map is denoted by . The faces inside the Hamiltonian of . The principle of coloring is clear: we color in light is red and is light. Thus we get a good coloration of the faces of

circuit

can be good colored with two colors. To show this, will use the contains no circuit. For such a circuit , which is is a forest and so it can be colored from the Hamiltonian circuit , inside

corresponding piece of the dual graph would isolate a vertex of map impossible. Thus, the piece of

with two colors. Analogically, we prove that the faces outside the Hamiltonian circuit can be colored with two other colors.

PART V APPLICATIONS Application of the chromatic number in real life situations: Algorithm for graph coloring and finding the chromatic number of a simple graph:
1) Assign color 1 to the vertex with highest degree. 2) Assign color 1 to any other vertex that is non-adjacent to vertex which is

colored in 1. 3) Assign color 2 to the vertex with the next highest degree that is not already colored. 4) Assign color 2 to any other vertex that is non-adjacent to vertex which is colored in 2 and that is not already colored. 5) Proceed in this manner until all vertices are colored The Chromatic number is particularly useful in the preparation of lists and schedules for various programs and events. For example, when we plan a program for a number of presentations. We must always be very careful to avoid a participant to have more than one presentation at the same time. If presentations are over a certain period: time, at least how long does it take to conduct all presentations? How many different rooms will be needed if for a presentation we need one room? The answers to these questions are linked to the chromatic number. It is

1
1 2 +

2 +

3 +

4 +

5 5 + 5

most clearly demonstrated by the following example: Those applications use tables, because with them it is easier to see how we have to construct our graph (depends on the conditions in the problem). We have 5 presentations, denoted by numbers 1 to 5. Assign + to those presentations that have at least one common participant.

pic. 9

We construct a graph that has a presentation for vertex and edges are connected if in the table we have + for two presentations. Color the graph using the algorithm above. Assigned in one color presentations will be hold in one room. Therefore, we need two rooms (2 is the chromatic number of the graph). The duration of all presentations is the smallest if and only if we conduct the presentations colored in different colors simultaneously. Then we need at least 3 different times to conduct all the presentations. (follows from the maximum number of vertices which share the same color). On this principle we obtain that we need times to conduct all the program. Method for applying the chromatic number in real life situations: 1. Construct a table for the requested event 2. Construct a graph based on the table 3. Color the graph using the algorithm above 4. Reach a result for the event

Applications of the chromatic number using the method above: 1. 2. 3. 4. schedules (conferences and events) programs (school program) timetables (trains) distribution of items: a. animals which can and cannot live together (distribution of species: fishes, spiders, snakes) b. plants that can and cannot be kept together c. food which can and cannot be consumed together (preparation of a diet) d. people who can and cannot stay together (celebrations, wedding tables) e. determination of a radio frequencies so that they dont detect each other f. Compiler optimization of a computer program (Register allocation)

Vous aimerez peut-être aussi