\documentclass{article} \usepackage{amsmath,amssymb} \setlength{\oddsidemargin}{0.25 in} \setlength{\evensidemargin}{-0.25 in} \setlength{\topmargin}{-0.6 in} \setlength{\textwidth}{6.5 in} \setlength{\textheight}{8.5 in} \setlength{\headsep}{0.75 in} \setlength{\parindent}{0 in} \setlength{\parskip}{0.1 in} \newcounter{lecnum} % % The following macro is used to generate the header. % \newcommand{\lecture}[4]{ \pagestyle{myheadings} \thispagestyle{plain} \newpage \setcounter{lecnum}{#1} \setcounter{page}{1} \noindent \begin{center} \framebox{ \vbox{\vspace{2mm} \hbox to 6.28in { {{\bf 18.317~Combinatorics, Probability, and Computations on Groups} \hfill #2} } \vspace{4mm} \hbox to 6.28in { {\Large \hfill Lecture #1 \hfill} } \vspace{2mm} \hbox to 6.28in { {\it Lecturer: #3 \hfill Scribe: #4} } \vspace{2mm}}} \end{center} \markboth{Lecture #1: #2}{Lecture #1: #2} \vspace*{4mm} } \input{epsf} %usage: \fig{LABEL}{FIGURE-HEIGHT}{CAPTION}{FILENAME} \newcommand{\fig}[4]{ \begin{figure} \setlength{\epsfysize}{#2} \centerline{\epsfbox{#4}} \caption{#3} \label{#1} \end{figure} } \newtheorem{theorem}{Theorem} \newtheorem{lemma}[theorem]{Lemma} \newtheorem{proposition}[theorem]{Proposition} \newtheorem{claim}[theorem]{Claim} \newtheorem{corollary}[theorem]{Corollary} \newtheorem{definition}[theorem]{Definition} \newenvironment{proof}{{\bf Proof:}}{\hfill\rule{2mm}{2mm}} \newenvironment{mproof}[1]{{\bf #1}}{\hfill\rule{2mm}{2mm}} \def\E#1{{\rm E}[#1]} \def\i{\hspace*{5mm}} \newcounter{lineno} \setcounter{lineno}{0} \def\n{\addtocounter{lineno}{1} \hbox to 7mm{\hfill \arabic{lineno}:}\hspace{2mm}} \def\resetn{\setcounter{lineno}{0}} \def\tqbfk{\tqbf_k^\exists} \def\ep{\epsilon} \def\erd{Erd\H{o}s} \begin{document} \lecture{6}{21 September 2001}{Igor Pak}{C. Goddard} \section*{Probabilistic Generation} In this lecture, we will complete the classical proof for Dixon's theorem on the probabilistic generation of $S_n$, based on the work by \erd\ and Tur\'{a}n. Reiterating from the previous lectures, here is Dixon's theorem: \begin{theorem}{\bf (Dixon)} \label{Dixon} \[ \Pr(\langle\sigma_1, \sigma_2\rangle = A_n \mbox{ or } S_n) \rightarrow \mbox{1 as $n$} \rightarrow \infty. \] \end{theorem} We will use Lemma \ref{Lemma1}, proved last lecture, and Jordan's theorem (Theorem \ref{Jordan}) and combine them with Lemma \ref{Lemma2}, proved here, to prove Dixon's theorem (Theorem \ref{Dixon}) classically. Thus, Lemma \ref{Lemma1} and Jordan's theorem are merely stated here: \begin{lemma}{\bf (\erd-Tur\'{a}n)} \label{Lemma1} Let $1 \leq a_1 < a_2 < a_r \leq n$. Then \[ \Pr(\sigma\in S_n \mbox{ does not contain any cycles of length } a_i) \leq \sum_{i=1}^r\frac{1}{a_i} \,. \] \end{lemma} \begin{theorem}{\bf (Jordan 1873)} \label{Jordan} If $\ G \subset S_n$ is primitive and contains a cycle of length $p$ where $p$ is a prime less than $n-3$ the $G$ is equal to $A_n$ or $S_n$. \end{theorem} Now continuing from the previous lecture, we will prove the following lemma. \begin{lemma}{\bf (\erd-Tur\'{a}n)} \label{Lemma2} For a fixed prime $p$ (or prime power), \[ \Pr (\sigma \in S_n, p \nmid \mathrm{order}(\sigma)) = \prod^{\lfloor \frac{n}{p} \rfloor}_{i=1} \left(1 - \frac{1}{p \cdot i} \right) \] \end{lemma} \begin{proof} Let \[ z_{\lambda} = \frac{n!}{1^{m_1} \, m_1! \, 2^{m_2} \, m_2! \, \ldots} = \mbox{ \# elements in conjugacy class } (\lambda) \] where $\lambda = (1^{m_1} 2^{m_2} \ldots)$ and $\sum m_i \cdot i = n $. Now \begin{eqnarray*} 1 = \sum_{\lambda \vdash n} \frac{z_{\lambda}}{n!} & = & \mbox{coeff } [t^n] \prod_{i=1}^{\infty} \left(1 + \frac{t^i}{1! \cdot i} + \frac{t^{2i}}{2! \cdot i^2} + \frac{t^{3i}}{3! \cdot i^3} + \ldots\right) \\ \Pr(\sigma \in S_n, p \nmid \mathrm{order}(\sigma)) & = & \mbox{coeff } [t^n] \prod_{i=1, \ p \nmid i}^\infty \left(1 + \frac{t^i}{1! \cdot i} + \frac{t^{2i}}{2! \cdot i^2} + \frac{t^{3i}}{3! \cdot i^3} + \ldots \right) \end{eqnarray*} $ \mbox{Now denote } \Pi = 1 + \frac{t^i}{1! \cdot i} + \frac{t^{2i}}{2! \cdot i^2} + \frac{t^{3i}}{3! \cdot i^3} + \ldots $ \begin{eqnarray*} \mbox{So } \Pi & = & \prod_{p \nmid i, i = 1}^\infty \exp \left( \frac{t^i}{i} \right) \; (\mbox{since } e^x = 1 + \frac{x}{1!} + \frac{x^2}{2!} + \ldots ) \\ & = & \frac{\prod_{i = 1}^{\infty} \exp \left( \frac{t^i}{i} \right)}{\prod_{\alpha = 1}^{\infty} \exp \left( \frac{t^{p \alpha}}{p \alpha} \right)} \\ & = & \exp \left(\sum_{i = 1}^{\infty} \frac{t^i}{i} - \sum_{\alpha = 1}^{\infty} \frac{t^{\alpha p}}{\alpha p} \right) \\ & = & \exp \left(-\log(1 - t) + \frac{1}{p} \log(1 - t^p) \right) \; (\mbox{using } -\log(1-x) = x + \frac{x^2}{2} + \frac{x^3}{3} + \ldots ) \\ & = & \frac{(1 - t^p)^{\frac{1}{p}}}{1 - t} \\ & = & \frac{1 - t^p}{1 - t} \cdot (1 - t^p)^{\frac{1}{p} - 1} \\ & = & \frac{1 - t^p}{1 - t} \cdot \left(\frac{1}{1 - t^p}\right)^{1 - \frac{1}{p}} \\ & = & (1 + t + \ldots + t^{p - 1}) \cdot \left\{ 1 + \sum_{m = 1}^{\infty} t^{mp} \cdot \left(1 - \frac{1}{p}\right)\left(1 - \frac{1}{2p}\right) \cdot \ldots \cdot \left(1 - \frac{1}{mp}\right) \right\} \\ \end{eqnarray*} (Note: $(1 + x)^{\alpha} = 1 + \frac{\alpha}{1!} x + \frac{\alpha (\alpha - 1)}{2!} x^2 + \ldots$ ) So to find coeff[$t^n$], take $m = \lfloor \frac{n}{p} \rfloor$ above, and we're done. \end{proof} Now we are ready to prove Dixon's theorem (Theorem \ref{Dixon}). \begin{proof} $\Pr (\sigma \mbox{ has } p - \mbox{cycle}) \longrightarrow 1 \mbox{ as } n \rightarrow \infty, p < n-2 \;.$ Let $A = \{ \log^2 n < p < n-2, p\ - \mbox{prime} \}$. Using Lemma \ref{Lemma1}, $\Rightarrow \Pr (\sigma \mbox{ has no } A\ - \mbox{cycles})$ \begin{eqnarray*} \leq \frac{1}{\sum_{p = \log^2 n}^{n - 2} \frac{1}{p}} & \sim & \frac{1}{\log \log (n - 2) - \log \log (\log^2 n)} \\ & & (\mbox{Using Euler's theorem: } \sum_{p < x} \frac{1}{p} \sim \log \log x ) \\ & \sim & c \cdot (\log \log n)^{-1} \mbox { where c is a constant } \end{eqnarray*} $\Rightarrow \mbox{ with } \Pr > 1 - \frac{c}{\log \log n} \ ,\; \exists\ p\ - \mbox{cycle with } p \in A \mbox{ for some prime } p\;. $ Now from Lemma \ref{Lemma2}, since $p > \log^2 n$, we have: \begin{eqnarray*} & & \Pr(\sigma \ \text{contains exactly one $p$-cycle} \, | \, \sigma \ \text{contains at least one $p$-cycle}) = \prod_{i = 1}^{n - p} \left( 1 - \frac{1}{p \cdot i} \right) \\ & \qquad & > \exp\left(\sum_{i=1}^{n} \log \left( 1 - \frac{1}{p \cdot i} \right)\right) = \exp\left(-\sum_{i=1}^{n} \frac{1}{p \cdot i} + \mathrm{o}(1) \right) = \exp \left( \frac{-\log n + \log p + \mathrm{o}(1)}{p} \right) \\ & \qquad & > \exp \left(\frac{-\log n}{p}\right) > \exp \left(\frac{-1}{\log n}\right) > 1-\frac{1}{\log n} \end{eqnarray*} \end{proof} Finally, for the $\Pr$ as in Theorem~1, we obtain $$\qquad \qquad \qquad \qquad \qquad \qquad \Pr > \left(1 - \frac{c}{\log \log n}\right) \left(1-\frac{1}{\log n}\right) \to 1 \ \ \text{as} \ \ n \to \infty \qquad \qquad \ \qquad \qquad \qquad \qquad \blacksquare$$ \end{document}