\documentclass{article} \usepackage[english]{babel} \usepackage[utf8x]{inputenc} \usepackage{amsmath} \usepackage{amsfonts} \usepackage{graphicx} \usepackage[colorinlistoftodos]{todonotes} \usepackage{amsthm} \usepackage{wrapfig} \usepackage{subcaption} \title{Graph Coloring} \author{Ken Coiteux}

\begin{document} \maketitle \newtheorem{definition}{Definition} \newtheorem{conjecture}{Conjecture} \newtheorem{theorem}{Theorem}

\section{Intoduction} When I was younger, I used to do this thing. It was a combination of art and a game. I would draw a series of interlocking figures and then start shading in areas. The rule was to start shading one area, then shade the next area that contained an opposite angle to the first area. I would continue this process until no more regions could be shaded. Adjoining regions could not be shaded. \\

\begin{wrapfigure}{l}{0.5\textwidth} \begin{center} \includegraphics[width=0.48\textwidth]{Untitled} \end{center} \caption{Successful coloring game} \end{wrapfigure} At the time, I did not realize I was working on any math. I thought I was just playing a game and creating some interesting art design. It was recently, when I learned about the math topic of graph coloring, that I realized that was what I was doing when I was younger. \\

With a simple mapping, the shaded regions in Figure 1 become points and the intersection become lines. With this new re-imaging, I can now start defining some terms in graph theory.

Points are called vertices and lines connecting vertices are called edges. Two vertices are said to be adjacent when there is an edge connecting them. The set of vertices and edges in a figure is called a graph.

\begin {definition} A "proper coloring" of a graph is the assigning of colors to a graph in such a way that adjacent vertices have different colors. \end {definition}

\begin {definition} The minimum number of colors needed for a proper coloring of a graph is called the chromatic number of a graph. \end {definition}

\begin{wrapfigure}{l}{0.5\textwidth} \begin{center} \includegraphics[width=0.46\textwidth]{graph1} \end{center} \end{wrapfigure}

This graph has two vertices, one edge and a chromatic number of 2. Given the two colors listed, the number of proper colorings is also 2. Why is this? Red and green can switch places, giving us two different ways to color the graph. \\

\begin {definition} $P(G;x)$, the number of proper colorings of a graph $G$ using $x$ colors is the chromatic polynomial of G \end {definition}

\begin{figure}[h] \centering \includegraphics[width=\textwidth]{graph3} \label{graph3} \end{figure} With this graph, there are still two vertices, one edge, and a chromatic number of 2. However, the $x$ represents the number of colors possible. The chromatic polynomial here is then $x(x-1)$. The chromatic polynomial increases with complexity as its corresponding graph increases with complexity and will be explored later. \\

There are some special types of graphs, which are defined below:

\begin {definition} A graph is said to be complete when each vertex is adjacent to each other. \end {definition}

\begin {definition} A graph is said to be circular when each vertex is connected by one edge in the following way: $v_1$ to $v_2$, $v_2$ to $v_3$, ..., $v_{n-1}$ to $v_n$, and $v_n$ to $v_1$. \end {definition}

\begin {definition} A path graph is a circular graph without $v_n$ connected by an edge to $v_1$.. \end {definition}

\subsection{Chromatic polynomial}

\begin{wrapfigure}{r}{0.5\textwidth} \begin{center}

\includegraphics[width=0.48\textwidth]{tmp_2a1_nq} \end{center} \caption{graph with 2 vertices} \label{fig:a1} \end{wrapfigure} Remember that the chromatic ploynbeomial returns the number of proper colorings of a graph. Suppose $v_1$ has x number of colors available. Then $v_2$ will have x-1 colors available. The chromatic polynomial for \ref{fig:a1} will be $x(x-1)$, the number of colors available for one vertex multiplied by the number of colors for the next vertex. \\

This idead works when graphs have a few number of vertices. However, finding the chromatic polynomial when the number of vertices are greater than 3. Luckily, there is an algorithm for determining chromatic polynomials for increasingly more complicated graphs.

\begin{definition} The Birkhoff-Lewis Reduction Algorithm is: $$P(G,x)=P(G-e,x)-P(G_e,x)$$. \end{definition} \begin{figure}[h] \centering \begin{subfigure}[h]{0.3\textwidth} \includegraphics[width=\textwidth]{example1} \caption{$P(G,x)$} \label{fig:gull} \end{subfigure}% \qquad %add desired spacing between images, e. g. ~, \quad, \qquad etc. %(or a blank line to force the subfigure onto a new line) \begin{subfigure}[h]{0.3\textwidth} \includegraphics[width=\textwidth]{example2} \caption{$P(G-e,x)$} \label{fig:tiger} \end{subfigure} \quad %add desired spacing between images, e. g. ~, \quad, \qquad etc. %(or a blank line to force the subfigure onto a new line) \begin{subfigure}[h]{0.3\textwidth} \includegraphics[width=\textwidth]{example3}

\caption{$P(G_e,x)$} \label{fig:mouse} \end{subfigure} \caption{Image of Birkhoff-Lewis}\label{fig:ex} \end{figure} Birkhoff-Lewis says to find the chromatic polynomial of a graph, $P(G,x)$, find the polynomial of the graph missing an edge, $P(G-e,x)$, and subtract the polynomial of the graph with two vertices contracted, $P(G_e,x)$. \\

Now that we have some definitions in place, we can now examine chromatic numbers and polynomials of different types of graphs.

\section{Chromatic Findings}

Using SAGE, a mathematical software program, the following chromatic numbers and chromatic polynomials were derived for path graphs. \\

\begin{table}[h] \centering \caption{Path Graph} \begin{tabular}{c|c|r} Vertices & Chromatic Number & Chromatic Polynomial\\\hline 1 & 1 & $x$\\ 2 & 2 & $x^2-x$\\ 3 & 2 & $x^3-2x^2+x$\\ 4 & 2 & $x^4-3x^3+3x^2-x$\\ 5 & 2 & $x^5-4x^4+6x^3-4x^2+x$\\ 6 & 2 & $x^6-5x^5+10x^4-10x^3+5x^2-x$\\ 7 & 2 & $x^7-6x^6+15x^5-20x^4+15x^3-6x^2+x$\\ 8 & 2 & $x^8-7x^7+21x^6-35x^5+35x^4-21x^3+7x^2-x$\\ 9 & 2 & $x^9-8x^8+28x^7-56x^6+70x^5-56x^4+28x^3-8x^2+x$ \end{tabular}

\label{tab:Pthgrph} \end{table} Let's look at the list of chromatic polynomials in Table \ref{tab:Pthgrph}. First, notice that the order of each polynomial is the same as the number of vertices in the path graphs. This will be noticed again in other graphs and we will discuss that later. Another item (to be discussed later) to be noticed is that the second coefficent is the same as the number of edges in the various path graphs. \\

Again, with the help of SAGE, each of the chromatic polynomials for the path graphs can be factored. New patterns will emerge.

\begin{table}[h] \centering \caption{Path Graph polynomial factorization} \begin{tabular}{c|c} Chromatic Polynomial & Factored\\\hline $x$ & $x$\\ $x^2-x$ & $x(x-1)$\\ $x^3-2x^2+x$ & $x(x-1)^2$\\ $x^4-3x^3+3x^2-x$ & $x(x-1)^3$\\ $x^5-4x^4+6x^3-4x^2+x$ & $x(x-1)^4$\\ $x^6-5x^5+10x^4-10x^3+5x^2-x$ & $x(x-1)^5$\\ $x^7-6x^6+15x^5-20x^4+15x^3-6x^2+x$ & $x(x-1)^6$\\ $x^8-7x^7+21x^6-35x^5+35x^4-21x^3+7x^2-x$ & $x(x-1)^7$\\ $x^9-8x^8+28x^7-56x^6+70x^5-56x^4+28x^3-8x^2+x$ & $x(x-1)^8$ \end{tabular} \label{tab:Pthgrphfct} \end{table} \begin{conjecture} Given a path graph with n vertices, $P(G_n,x)=x(x-1)^{n-1}$. \end{conjecture} \begin{conjecture} Given a path graph with n vertices, the coefficeints of $P(G_n,x)$ coincide with the n-1 row of Pascal's Triangle. \end{conjecture}

Looking back at Table \ref{tab:Pthgrph}, notice that the chromatic number for path graphs having more than one vertex is 2. Remember that a chromatic number is the smallest number of colors that can be used to properly color a graph. Also remember that the chromatic polynomial yields, when given x number of colors, yields the total number of proper colorings.

\begin{table}[h] \centering \caption{Chromatic Number for Path Graphs (n=number of vertices)} \begin{tabular}{c|c} Number of colors & Number of proper colorings\\\hline 0 & $0(0-1)^{n-1}=0$\\ 1 & $1(1-1)^{n-1}=0$\\ 2 & $2(2-1)^{n-1}=2$\\ 3 & $3(3-1)^{n-1}=(3)(2^{n-1})$ \end{tabular} \label{tab:Pthgrphno} \end{table} The general form for the chromatic polynomial for path graphs was used in Table \ref{tab:Pthgrphno}. As you can see, 2 is the smallest number of colors used that generates a number of colorings that is greater than 0.

\begin{theorem} The chromatic number of all path graphs having more than 1 vertex is two. \end{theorem} \begin{table}[h] \centering \caption{Complete Graph} \begin{tabular}{c|c|l} Vertices & Chromatic Number & Chromatic Polynomial\\\hline 1 & 1 & $x$\\ 2 & 2 & $x(x-1)$\\ 3 & 3 & $x(x-1)(x-2)$\\ 4 & 4 & $x(x-1)(x-2)(x-3)$\\ 5 & 5 & $x(x-1)(x-2)(x-3)(x-4)$\\ 6 & 6 & $x(x-1)(x-2)(x-3)(x-4)(x-5)$\\

7 & 7 & $x(x-1)(x-2)(x-3)(x-4)(x-5)(x-6)$\\ 8 & 8 & $x(x-1)(x-2)(x-3)(x-4)(x-5)(x-6)(x-7)$\\ 9 & 9 & $x(x-1)(x-2)(x-3)(x-4)(x-5)(x-6)(x-7)(x-8)$ \end{tabular} \label{tab:Cmpgrph} \end{table} \begin{table}[h] \centering \caption{Cycle Graph} \begin{tabular}{c|c|r} Vertices & Chromatic Number & Chromatic Polynomial\\\hline 1 & 1 & $x$\\ 2 & 2 & $x^2-x$\\ 3 & 3 & $x^3-3x^2+2x$\\ 4 & 2 & $x^4-4x^3+6x^2-3x$\\ 5 & 3 & $x^5-5x^4+10x^3-10x^2+4x$\\ 6 & 2 & $x^6-6x^5+15x^4-20x^3+15x^2-5x$\\ 7 & 3 & $x^7-7x^6+21x^5-35x^4+35x^3-21x^2+6x$\\ 8 & 2 & $x^8-8x^7+28x^6-56x^5+70x^4-56x^3+28x^2-7x$\\ 9 & 3 & $x^9-9x^8+36x^7-84x^6+126x^5-126x^4+84x^3-36x^2+8x$ \end{tabular} \label{tab:Cycgrph} \end{table} Two more tables are shown. Similar to Table \ref{tab:Pthgrph}, they list the number of vertices, chromatic numbers, and chromatic polynomials for complete graphs and cycle graphs. Looking at these tables, more conjectures can be made.

\begin{conjecture} Given a complete graph with n vertices, the chromatic number is equal to the number of vertices. \end{conjecture} \begin{conjecture} Given a complete graph with n vertices, $$P(G_n,x)=\displaystyle \prod_{k=1}^{n}x(x-1)^{n-1}$$ \end{conjecture} \begin{conjecture} Given a cycle graph with $n>1$ vertices, the chromatic number is 2 when n is even and 3 when n is odd. \end{conjecture}

Earlier, there was mention that the number of vertices in a graph is also the order of the assosciated chromatic polynomial. It was also mentioned that the second coefficient of a chromatic polynomial is the number of edges in the associated graph. I will leave it to the reader to draw graphs to verify the following conjecture.

\begin{conjecture} Given a graph with n vertices, the second coefficient of the associated chromatic polynomial identifies the number of edges in the graph. \end{conjecture} Consider a graph having n vertices and one edge. By using Birkhoff-Lewis, that edge can be removed and it can be contracted. When the edge is removed, the number of vertices remain the same. When the edge is contracted, the number of vertices decrease by 1. Assuming the number of vertices is also the order of the chromatic polynomial,$$P(G-e,x)=x^n+\cdots$$ and $$P(G_e,x)=x^{n-1}+\cdots$$ Subtracting the edge-contracted from the edge-removed still leaves $P(G,x)=x^n+\cdots$ The order of the polynomial remains the same. \\

Now consider a graph with $m$ vertices and $h$ edges. Removing an edge leaves $m$ vertices and $h1$ edges. Contracting an edge leaves $m-1$ vertices and $h-1$ edges.

\begin{align*} P(G,x) &= P(G-e,x) - P(G_e,x)\\ &= x^m+\cdots - (x^{m-1}+\cdots)\\ &= x^m+\cdots \end{align*} The number of vertices remaines the same in the original graph and the order of the polynomial remained the same after Birkhoff-Lewis.

\begin{theorem} The order of a chromatic polynomial is equal to the number of vertices in the associated graph. \end{theorem}

\section{Conclusion}

This is just an overview of graph coloring. We proved that the chromatic polynomial of a path graph is 2 and we proved that the degree of a polynomial is the same as the number of vertices. Several conjectures were made on tables of data that still need to be formally proved. A note of interest is the ideas discussed above assumes that there can only be one edge between two vertices. Further questions include what happens when two or more edges can connect two vertices or when an edge begins at a vertex and loops on itself to end at the same vertex.

\end{document}