Boolean-Cayley-graphs / papers-talks / Twin-bent-Hurwitz-Radon / Leopardi-Twin-bent-Hurwitz-Radon.tex
22144 views\documentclass[11pt,a4paper]{jacodesmath}1%2\usepackage{amsmath}3\usepackage{amssymb}4\usepackage{amsthm}5\usepackage{graphicx}6%7%\usepackage[colorlinks=true,allcolors=blue]{hyperref}8%\usepackage{url}9%10\newcommand{\mb}[1]{\mathbb{#1}}11\newcommand{\mf}[1]{\mathbf{#1}}12\newcommand{\Cay}{\operatorname{Cay}}13\newcommand{\oE}{\mf{\operatorname{E}}}14\newcommand{\G}{\mb{G}}15\newcommand{\R}{\mb{R}}16\newcommand{\Z}{\mb{Z}}17\newcommand{\Rep}{P}18\newcommand{\V}[1]{\underline{#1}}19\newcommand{\abs}[1]{\left| #1 \right|}20\newcommand{\isomorphic}{\simeq}21\newcommand{\To}{\rightarrow}22%23\newtheorem{question}{Question}24%25\newenvironment{proofof}[1]{\noindent\emph{Proof of #1.}}{\qed}2627\rhead{ \fancyplain{}{P. Leopardi} }28\begin{document}2930\title{ \textbf{Twin bent functions, strongly regular Cayley graphs, and Hurwitz-Radon theory}}3132\author[1]{Paul Leopardi\thanks{paul.leopardi@gmail.com}}33\affil[1]{University of Melbourne, Australian Government - Bureau of Meteorology}34%3536%%37\maketitle38\begin{textblock*}{10cm} (8cm,3cm)39\begin{center}JACODESMATH \\40\footnotesize Twin bent functions, strongly regular Cayley graphs, and Hurwitz-Radon theory41\end{center}42\end{textblock*}43%%%4445\begin{abstract}46%47The real monomial representations of Clifford algebras48give rise to two sequences of bent functions.49For each of these sequences, the corresponding Cayley graphs are50strongly regular graphs, and the corresponding sequences of strongly regular graph parameters51coincide.52Even so, the corresponding graphs in the two sequences are not isomorphic, except in the first 353cases.54The proof of this non-isomorphism is a simple consequence of a theorem of Radon.55%56\textbf{Keywords.} Bent functions, strongly regular graphs, Clifford algebras, Hurwitz-Radon.57\\58\textbf{2010 AMS Classification. 05E30 (11T71 15A24 15B34) }5960\end{abstract}6162\section{Introduction}63\label{sec-Introduction}64Two recent papers \cite{Leo14Constructions,Leo15Twin} describe and investigate two infinite65sequences of bent functions and their Cayley graphs.66The bent function $\sigma_m$ on $\Z_2^{2 m}$ is described in the first paper67\cite{Leo14Constructions}, on68generalizations of Williamson's construction for Hada\-mard matrices.69The bent function $\tau_m$ on $\Z_2^{2 m}$ is described in the second paper \cite{Leo15Twin},70which investigates some of the properties of the two sequences of bent functions.71In this second paper it is shown that the bent functions $\sigma_m$ and $\tau_m$ both correspond to72Hada\-mard difference sets with the same parameters73\begin{align*}74(v_m,k_m,\lambda_m,n_m) &= (4^m, 2^{2 m - 1} - 2^{m-1}, 2^{2 m - 2} - 2^{m-1}, 2^{2 m - 2}),75\end{align*}76and that their corresponding Cayley graphs are both strongly regular with the same parameters77$(v_m,k_m,\lambda_m,\lambda_m)$.7879The main result of the current paper is the following.80\begin{theorem}\label{HR-non-imomorphic-theorem}81The Cayley graphs of the bent functions $\sigma_m$ and $\tau_m$ are isomorphic only when $m=1, 2,$82or $3.$83\end{theorem}8485The remainder of the paper is organized as follows.86Section \ref{sec-Background} outlines some of the background of this investigation.87Section \ref{sec-Preliminaries} includes further definitions used in the subsequent sections.88Section~\ref{sec-Results} proves the main result, and resolves the conjectures and the question89raised by the previous papers.90Section~\ref{sec-Discussion} puts these results in context, and suggests future research.9192\section{Background}\label{sec-Background}93A recent paper of the author \cite{Leo14Constructions} describes a generalization of94Williamson's construction for Hada\-mard matrices \cite{Wil44}95using the real monomial representation of the basis elements of the Clifford algebras $\R_{m,m}$.9697Briefly, the general construction uses some98\begin{align*}99%100\quad &A_k \in \{-1,0,1\}^{n \times n}, \quad B_k \in \{-1,1\}^{b \times b},101\quad k \in \{1,\ldots,n\},102%103\end{align*}104%105where the $A_k$ are \emph{monomial} matrices,106and constructs107%108\begin{align}109%110H &:= \sum_{k=1}^n A_k \otimes B_k,111\tag{H0}112%113\end{align}114%115such that116%117\begin{align}118%119H \in \{-1,1\}^{n b \times n b}120\quad121\text{and}122\quad123H H^T &= n b I_{(n b)},124\tag{H1}125%126\end{align}127%128i.e. $H$ is a Hada\-mard matrix of order $n b$.129The paper \cite{Leo14Constructions} focuses on a special case of the construction,130satisfying the conditions131%132\begin{align}133%134A_j \ast A_k = 0 \quad (j \neq k)&, \quad \sum_{k=1}^n A_k \in \{-1,1\}^{n \times n},135\notag136\\137A_k A_k^T &= I_{(n)},138\notag139\\140A_j A_k^T + \lambda_{j,k} A_k A_j^T &= 0 \quad (j \neq k),141\label{constructions-4}142\\143B_j B_k^T - \lambda_{j,k} B_k B_j^T &= 0 \quad (j \neq k),144\notag145\\146\lambda_{j,k} &\in \{-1,1\},147\notag148\\149\sum_{k=1}^n B_k B_k^T &= n b I_{(b)},150\notag151%152\end{align}153%154where $\ast$ is the Hada\-mard matrix product.155156In Section 3 of the paper \cite{Leo14Constructions},157it is noted that the Clifford algebra $\R^{2^m \times 2^m}$ has a canonical basis consisting of158$4^m$ real monomial matrices,159corresponding to the basis of the algebra $\R_{m,m}$, with the following properties:160161Pairs of basis matrices either commute or anticommute.162Basis matrices are either symmetric or skew,163and so the basis matrices $A_j, A_k$ satisfy164%165\begin{align}166%167A_k A_k^T &= I_{(2^m)},168\quad169A_j A_k^T + \lambda_{j,k} A_k A_j^T = 0 \quad (j \neq k),170\quad171\lambda_{j,k} \in \{-1,1\}.172%173\label{A-property-1}174\end{align}175176Additionally, for $n=2^m$, we can choose a transversal of $n$ canonical basis matrices that177satisfies conditions \eqref{constructions-4} on the $A$ matrices,178%179\begin{align}180%181A_j \ast A_k = 0 \quad (j \neq k)&,182\quad183\sum_{k=1}^n A_k \in \{-1,1\}^{n \times n}.184%185\label{A-property-2}186\end{align}187188Section 3 also contains the definition of $\varDelta_m$, the restricted amicability /189anti-amicability graph of $\R_{m,m}$,190and the subgraphs $\varDelta_m[-1]$ and $\varDelta_m[1]$, as well as the term ``transversal graph''.191These definitions are repeated here since they are used in the conjectures and question below.192\begin{definition}\label{definition-delta}193\cite[p. 225]{Leo14Constructions}194195Let $\varDelta_m$ be the graph whose vertices are the $n^2=4^m$196positive signed basis matrices of the real representation197of the Clifford algebra $\R_{m,m}$,198with each edge having one of two labels, $-1$ or $1$:199\begin{itemize}200\item201Matrices $A_j$ and $A_k$ are connected by an edge labelled by $-1$ (``red'') if they have disjoint202support and are anti-amicable,203that is, $A_j A_k^{-1}$ is skew.204\item205Matrices $A_j$ and $A_k$ are connected by an edge labelled by $1$ (``blue'') if they have disjoint206support and are amicable,207that is, $A_j A_k^{-1}$ is symmetric.208\item209Otherwise there is no edge between $A_j$ and $A_k$.210\end{itemize}211The subgraph $\varDelta_m[-1]$ consists of the vertices of $\varDelta_m$ and all edges in212$\varDelta_m$ labelled by $-1$.213Similarly, the subgraph $\varDelta_m[1]$ contains all of the edges of $\varDelta_m$ that are214labelled by $1$.215\end{definition}216217A \emph{transversal graph} for the Clifford algebra $\R_{m,m}$218is any induced subgraph of $\varDelta_m$ that is a complete graph on $2^m$ vertices.219That is, each pair of vertices in the transversal graph represents a pair of matrices,220$A_j$ and $A_k$ with disjoint support.221222The following three conjectures appear in Section 3 of the paper \cite{Leo14Constructions}:223224\begin{conjecture}\label{conjecture-1}225%226For all $m \geqslant 0$ there is a permutation $\pi$ of the set of $4^m$ canonical basis matrices,227that sends an amicable pair of basis matrices with disjoint support to an anti-amicable pair, and228vice-versa.229%230\end{conjecture}231232\begin{conjecture}\label{conjecture-2}233%234For all $m \geqslant 0,$235for the Clifford algebra $\R_{m,m},$ the subset of transversal graphs that are236not self-edge-colour complementary237can be arranged into a set of pairs of graphs with each member of the pair238being edge-colour complementary to the other member.239%240\end{conjecture}241242\begin{conjecture}\label{conjecture-3}243%244For all $m \geqslant 0,$245for the Clifford algebra $\R_{m,m},$ if a graph $T$ exists amongst the transversal graphs,246then so does at least one graph with edge colours complementary to those of $T$.247%248\end{conjecture}249Note that Conjecture \ref{conjecture-1} implies Conjecture \ref{conjecture-2},250which in turn implies Conjecture \ref{conjecture-3}.251252The significance of these conjectures can be seen in relation to the following result,253which is Part 1 of Theorem 10 of the paper \cite{Leo14Constructions}.254255\begin{lemma}\label{th-graph-image}256If $b$ is a power of 2, $b=2^m$, $m \geqslant 0$,257the amicability / anti-amicability graph $P_b$ of the matrices258$\{-1,1\}^{b \times b}$ contains a complete two-edge-coloured graph on $2 b^2$ vertices259with each vertex being a Hada\-mard matrix.260This graph is isomorphic to $\varGamma_{m,m}$, the amicability / anti-amicability graph of261the group $\G_{m,m}$.262\end{lemma}263The definitions of $\varGamma_{m,m}$ and $\G_{m,m}$264are given in Section 3 of the paper \cite{Leo14Constructions},265and the definition of $\G_{m,m}$ is repeated below.266For the current paper, it suffices to note that $\varDelta_m$ is a subgraph of $\varGamma_{m,m}$,267and so, therefore, are all of the transversal graphs.268269An $n$-tuple of $A$ matrices of order $n=2^m$ satisfying properties \eqref{A-property-1} and270\eqref{A-property-2}271yields a corresponding transversal graph $T$.272As noted in Section 5 of the paper \cite{Leo14Constructions},273if Conjecture \ref{conjecture-3} were true, this would guarantee the existence274of an edge-colour complementary transversal graph $\overline{T}$.275In turn, because Lemma \ref{th-graph-image} guarantees the existence of a complete two-edge-coloured276graph277isomorphic to $\varGamma_{m,m}$ within $P_b$, and because $\varDelta_m$ is a subgraph of278$\varGamma_{m,m}$,279the graph $P_b$ would have to contain a two-edge-coloured subgraph isomorphic to $\overline{T}$.280This would imply the existence of an $n$-tuple of $B$ matrices of order $n$ satisfying the condition281\eqref{constructions-4}282such that the construction (H0) would satisfy the Hada\-mard condition (H1), with a matrix of order283$n^2$.284285The author's subsequent paper on bent functions \cite{Leo15Twin}286refines Conjecture~\ref{conjecture-1} into the following question.287\begin{question}288\label{Question-1}289Consider the sequence of edge-coloured graphs $\varDelta_m$ for $m \geqslant 1$,290each with red subgraph $\varDelta_m[-1],$ and blue subgraph $\varDelta_m[1].$291For which $m \geqslant 1$ is there an automorphism of $\varDelta_m$292that swaps the subgraphs $\varDelta_m[-1]$ and $\varDelta_m[1]$?293\end{question}294295The main result of this paper, Theorem \ref{HR-non-imomorphic-theorem},296leads to the resolution of these conjectures and this question.297298\section{Further definitions and properties}299\label{sec-Preliminaries}300This section sets out the remainder of the definitions and properties used in this paper.301It is based on the previous papers \cite{Leo14Constructions, Leo15Twin} with additions.302303%\newpage304\paragraph*{Clifford algebras and their real monomial representations.}305\label{sec-Clifford}306307~308309The following definitions and results appear in the paper on Hada\-mard matrices and Clifford310algebras \cite{Leo14Constructions},311and are presented here for completeness, since they are used below.312Further details and proofs can be found in that paper, and in the paper on bent functions313\cite{Leo15Twin},314unless otherwise noted.315An earlier paper on representations of Clifford algebras \cite{Leo05} contains more background316material.317318The signed group \cite{Cra95}319$\G_{p,q}$ of order $2^{1+p+q}$320is extension of $\Z_2$ by $\Z_2^{p+q}$,321defined by the signed group presentation322%323\begin{align*}324\G_{p,q} := \bigg\langle \325&\mf{e}_{\{k\}}\ (k \in S_{p,q})\ \mid326\\327&\mf{e}_{\{k\}}^2 = -1\ (k < 0), \quad \mf{e}_{\{k\}}^2 = 1\ (k > 0),328\\329&\mf{e}_{\{j\}}\mf{e}_{\{k\}} = -\mf{e}_{\{k\}}\mf{e}_{\{j\}}\ (j \neq k) \bigg\rangle,330\end{align*}331%332where $S_{p,q} := \{-q,\ldots,-1,1,\ldots,p\}.$333%334% The following construction of the real monomial representation $\Rep(\G_{m,m})$335% of the group $\G_{m,m}$ is used below.336337The $2 \times 2$ orthogonal matrices338\begin{align*}339\oE_1 :=340\left[341\begin{array}{cc}3420 & -1 \\3431 & 0344\end{array}345\right],346\quad347\oE_2 :=348\left[349\begin{array}{cc}3500 & 1 \\3511 & 0352\end{array}353\right]354\end{align*}355generate $\Rep(\G_{1,1}),$ the real monomial representation of group $\G_{1,1}.$356The cosets of $\{\pm I\} \equiv \Z_2$ in $\Rep(\G_{1,1})$ are357ordered using a pair of bits, as follows.358\begin{align*}3590 &\leftrightarrow 00 \leftrightarrow \{ \pm I \},360\\3611 &\leftrightarrow 01 \leftrightarrow \{ \pm \oE_1 \},362\\3632 &\leftrightarrow 10 \leftrightarrow \{ \pm \oE_2 \},364\\3653 &\leftrightarrow 11 \leftrightarrow \{ \pm \oE_1 \oE_2 \}.366\end{align*}367368For $m > 1$,369the real monomial representation $\Rep(\G_{m,m})$ of the370group $\G_{m,m}$ consists of matrices of the form $G_1 \otimes G_{m-1}$371with $G_1$ in $\Rep(\G_{1,1})$ and $G_{m-1}$ in $\Rep(\G_{m-1,m-1}).$372The cosets of $\{\pm I\} \equiv \Z_2$ in $\Rep(\G_{m,m})$ are373ordered by concatenation of pairs of bits,374where each pair of bits uses the ordering as per $\Rep(\G_{1,1}),$375and the pairs are ordered as follows.376\begin{align*}3770 &\leftrightarrow 00 \ldots 00 \leftrightarrow \{ \pm I \},378\\3791 &\leftrightarrow 00 \ldots 01 \leftrightarrow \{ \pm I_{(2)}^{\otimes {(m-1)}} \otimes \oE_1 \},380\\3812 &\leftrightarrow 00 \ldots 10 \leftrightarrow \{ \pm I_{(2)}^{\otimes {(m-1)}} \otimes \oE_2 \},382\\383&\ldots384\\3852^{2m} - 1 &\leftrightarrow 11 \ldots 11 \leftrightarrow \{ \pm (\oE_1 \oE_2)^{\otimes {m}} \}.386\end{align*}387This ordering is called388the \emph{Kronecker product ordering} of the cosets of $\{\pm I\}$ in $\Rep(\G_{m,m}).$389390%Recall that391The group $\G_{m,m}$ and its real monomial representation $\Rep(\G_{m,m})$392satisfy the following properties.393\begin{enumerate}394\item395Pairs of elements of $\G_{m,m}$ (and therefore $\Rep(\G_{m,m})$) either commute or anti\-commute:396for $g, h \in \G_{m,m},$ either $h g = g h$ or $h g = - g h.$397\item398The matrices $E \in \Rep(\G_{m,m})$ are orthogonal: $E E^T = E^T E = I.$399\item400The matrices $E \in \Rep(\G_{m,m})$ are either symmetric and square to give $I$ or401skew and square to give $-I$: either $E^T = E$ and $E^2 =I$ or $E^T = -E$ and $E^2 = -I.$402\end{enumerate}403404Taking the positive signed element of each of the $2^{2m}$ cosets listed above405defines a transversal of $\{\pm I\}$ in $\Rep(\G_{m,m})$406which is also a monomial basis for the real representation of the Clifford algebra $\R_{m,m}$ in407Kronecker product order,408called this basis the \emph{positive signed basis} of $\Rep(\R_{m,m}).$409410%\begin{definition}\label{definition-gamma}411The function $\gamma_m : \Z_{2^{2 m}} \To \Rep(\G_{m,m})$412chooses the corresponding basis matrix from the positive signed basis of $\Rep(\R_{m,m}),$413using the Kronecker product ordering.414This ordering also defines a corresponding function on $\Z_2^{2 m},$415also called $\gamma_m.$416%\end{definition}417418\paragraph*{Hurwitz-Radon theory.}419\label{sec-Hurwitz-Radon}420The key concept used in the proof of Lemma~\ref{Red-clique-lemma} below is that of a421\emph{Hurwitz-Radon family} of matrices.422~423424A set of real orthogonal matrices $\{A_1,A_2,\ldots,A_s\}$ is called a Hurwitz-Radon family425\cite{GerP74a,Hur22,Rad22} if426\begin{enumerate}427\item428$A_j^T = -A_j$ for all $j=1,\ldots,s$, and429\item430$A_j A_k = -A_k A_j$ for all $j \neq k$.431\end{enumerate}432The Hurwitz-Radon function $\rho$ is defined by433\begin{align*}434\rho(2^{4 d + c}) &:= 2^c + 8 d, \quad \text{where~} 0 \leqslant c < 4.435\end{align*}436As stated by Geramita and Pullman \cite{GerP74a}, Radon \cite{Rad22}437proved the following result, which is used as a lemma in this paper.438\begin{lemma}\label{Hurwitz-Radon-lemma}439\cite[Theorem A]{GerP74a}440441Any Hurwitz-Radon family of order $n$ has at most $\rho(n)-1$ members.442\end{lemma}443444\paragraph*{The two sequences of bent functions.}445\label{sec-Bent}446447~448449The previous two papers \cite{Leo14Constructions,Leo15Twin}450define two binary functions on $\Z_2^{2 m}$, $\sigma_m$ and $\tau_m$, respectively.451Their key properties are repeated below.452See the two papers for the proofs and for more details and references on bent functions.453454The function $\sigma_m : \Z_2^{2 m} \To \Z_2$ has the following properties.455\begin{enumerate}456\item457For $i \in \Z_2^{2m},$ $\sigma_m(i) = 1$ if and only if the number of458digits equal to 1 in the base 4 representation of $i$ is odd.459\item460Since each matrix $\gamma_m(i)$ is orthogonal, $\sigma_m(i) = 1$ if and only if the matrix461$\gamma_m(i)$ is skew.462\item463The function $\sigma_m$ is bent.464\end{enumerate}465466The function $\tau_m : \Z_2^{2 m} \To \Z_2$ has the following properties.467\begin{enumerate}468\item469For $i \in \Z_2^{2m},$ $\tau_m(i) = 1$ if and only if the number of digits equal to 1 or 2 in the470base 4471representation of $i$ is non zero, and the number of digits equal to 1 is even.472\item473The value $\tau_m(i) = 1$ if and only if the matrix $\gamma_m(i)$ is symmetric but not diagonal.474\item475The function $\tau_m$ is bent.476\end{enumerate}477478\paragraph*{The relevant graphs.}479\label{sec-Graphs}480481~482483For a binary function $f : \Z_2^{2 m} \To \Z_2$, with $f(0)=0$ we consider the simple undirected484\emph{Cayley graph} $\Cay(f)$ \cite[3.1]{BerC99}485where the vertex set $V(\Cay(f)) = \Z_2^{2 m}$ and for $i,j \in \Z_2^{2 m}$, the edge $(i,j)$ is in486the edge set $E(\Cay(f))$ if and only if $f(i+j)=1$.487488In the paper on Hada\-mard matrices \cite{Leo14Constructions} it is shown that489since $\sigma_m(i)=1$ if and only if $\gamma_m(i)$ is skew,490the subgraph $\varDelta_m[-1]$ is isomorphic to the Cayley graph $\Cay(\sigma_m)$.491492The paper on bent functions \cite{Leo15Twin} notes that493since $\tau_m(i) = 1$ if and only if $\gamma_m(i)$ is symmetric but not diagonal,494the subgraph $\varDelta_m[1]$ is isomorphic to the Cayley graph $\Cay(\tau_m)$.495In that paper, these isomorphisms and the characterization of $\Cay(\sigma_m)$ and $\Cay(\tau_m)$496as Cayley graphs of bent functions are used to prove the following theorem.497\begin{theorem}\label{Twins-are-strongly-regular-theorem}498\cite[Theorem 5.2]{Leo15Twin}499500For all $m \geqslant 1,$501both graphs $\varDelta_m[-1]$ and $\varDelta_m[1]$ are strongly regular, with parameters502$v_m = 4^m,$ $k_m = 2^{2 m - 1} - 2^{m - 1},$ $\lambda_m=\mu_m=2^{2 m - 2} - 2^{m - 1}.$503%504\end{theorem}505506\section{Proof of Theorem \ref{HR-non-imomorphic-theorem} and related results}507\label{sec-Results}508Here we prove the main result, and examine its implications for Conjectures~\ref{conjecture-1}509to~\ref{conjecture-3} and Question~\ref{Question-1}.510511The proof of Theorem~\ref{HR-non-imomorphic-theorem} follows from the following two lemmas.512The first lemma puts an upper bound on the clique number of the graph $\Cay(\sigma_m) \isomorphic513\varDelta_m[-1]$.514\begin{lemma}515\label{Red-clique-lemma}516The clique number of the graph $\Cay(\sigma_m)$ is at most $\rho(2^m)$,517where $\rho$ is the Hurwitz-Radon function.518Therefore $\rho(2^m) < 2^m$ for $m \geqslant 4$.519\end{lemma}520\begin{proof}521If we label the vertices of the graph $\Cay(\sigma_m)$ with the elements of $Z_2^{2m}$,522then any clique in this graph is mapped to another clique if a constant is added to all of the523vertices.524Thus without loss of generality we can assume that we have a clique of order $s+1$ with one of the525vertices labelled by 0.526If we then use $\gamma_m$ to label the vertices with elements of $\R_{m,m}$ to obtain the isomorphic527graph $\varDelta_m[-1]$,528we have one vertex of the clique labelled with the identity matrix $I$ of order $2^m$.529Since the clique is in $\varDelta_m[-1]$, the other vertices $A_1$ to $A_s$ (say) must necessarily530be skew matrices531that are pairwise anti-amicable,532\begin{align*}533A_j A_k^T &= -A_k A_j^T\quad\text{for all~} j \neq k.534\intertext{But then}535A_j A_k &= -A_k A_j\quad\text{for all~} j \neq k,536\end{align*}537and therefore $\{A_1,\ldots,A_s\}$ is a Hurwitz-Radon family.538By Lemma~\ref{Hurwitz-Radon-lemma}, $s$ is at most $\rho(2^m)-1$ and therefore the size of the539clique is at most540$\rho(2^m)$.541\end{proof}542543The second lemma puts a lower bound on the clique number of the graph $\Cay(\tau_m) \isomorphic544\varDelta_m[1]$.545\begin{lemma}546\label{Blue-clique-lemma}547The clique number of the graph $\Cay(\tau_m)$ is at least $2^m$.548\end{lemma}549\begin{proof}550We construct a clique of order $2^m$ in $\Cay(\tau_m)$ with the vertices labelled in $\Z_2^{2m}$,551using the following set of vertices, denoted in base 4:552\begin{align*}553C_m &:= \{ 00 \ldots 00, 00 \ldots 02, 00 \ldots 20, \ldots, 22 \ldots 22 \}.554\end{align*}555The set $C_m$ is closed under addition in $\Z_2^{2 m}$,556and therefore forms a clique of order $2^m$ in $\Cay(\tau_m)$,557since the sum of any two distinct elements of $C_m$ is in the support of $\tau_m$.558\end{proof}559560With these two lemmas in hand, the proof of Theorem~\ref{HR-non-imomorphic-theorem} follows easily.561562\begin{proofof}{Theorem~\ref{HR-non-imomorphic-theorem}}563The result is a direct consequence of Lemmas~\ref{Red-clique-lemma} and~\ref{Blue-clique-lemma}.564For $m \geqslant 4$, the clique numbers of the graphs $\Cay(\sigma_m)$ and $\Cay(\tau_m)$ are565different,566and therefore these graphs cannot be isomorphic.567\end{proofof}568569Lemmas~\ref{Red-clique-lemma} and~\ref{Blue-clique-lemma}, along with570Theorem~\ref{HR-non-imomorphic-theorem}571imply the failure of the conjectures~\ref{conjecture-1} to~\ref{conjecture-3}, as well as the572resolution of Question~\ref{Question-1}, as follows.573\begin{theorem}574\label{Conjectures-are-false-theorem}575For $m \geqslant 4$ the following hold.576\begin{enumerate}577\item578There exist transversal graphs that do not have an edge-colour complement, and579therefore Conjecture~\ref{conjecture-3} does not hold.580\item581As a consequence, Conjectures~\ref{conjecture-1} and~\ref{conjecture-2} also do not hold.582\item583Question~\ref{Question-1} is resolved.584The only $m \geqslant 1$ for which there is an automorphism of $\varDelta_m$585that swaps the subgraphs $\varDelta_m[-1]$ and $\varDelta_m[1]$586are $m=1,2$ and $3$.587\end{enumerate}588589\end{theorem}590591\begin{proof}592Assume that $m \geqslant 4$.593A transversal graph is a subgraph of $\varDelta_m$ which is a complete graph of order $2^m$.594The edges of a transversal graph are labelled with the colour red (if the edge is contained in595$\varDelta_m[-1]$) or blue596(if the edge is contained in $\varDelta_m[1]$).597By Lemma~\ref{Red-clique-lemma}, the largest clique of $\varDelta_m[-1]$ is of order $\rho(2^m) <5982^m$,599and by Lemma~\ref{Blue-clique-lemma}, the largest clique of $\varDelta_m[1]$ is of order $2^m$.600If we take a blue clique of order $2^m$ as a transversal graph, this cannot have an edge-colour601complement602in $\varDelta_m$, because no red clique can be this large.603More generally, we need only take a transversal graph containing a blue clique with order larger604than $\rho(2^m)$ to have a clique with no edge-colour complement605in $\varDelta_m$.606This falsifies Conjecture~\ref{conjecture-3}.607608Since Conjecture~\ref{conjecture-3} fails for $m \geqslant 4$,609the pairing of graphs described in Conjecture~\ref{conjecture-2} is impossible for $m \geqslant 4$.610Thus Conjecture~\ref{conjecture-2} is also false.611612Finally, Conjecture~\ref{conjecture-1} fails as a direct consequence of613Theorem~\ref{HR-non-imomorphic-theorem}614since, for $m \geqslant 4$, the subgraphs $\varDelta_m[-1]$ and $\varDelta_m[1]$are not isomorphic.615Therefore, for $m \geqslant 4$, there can be no automorphism of $\varDelta_m$ that swaps these616subgraphs.617\end{proof}618619\section{Discussion}620\label{sec-Discussion}621The result of Lemma~\ref{Red-clique-lemma} is well known.622For example, the graph $\varDelta_m[-1]$ is the complement of the graph $V^+$ of Yiu \cite{Yiu90},623and the result for $V^+$ in his Theorem 2 is equivalent to Lemma \ref{Red-clique-lemma}.624625The main consequence of Theorem~\ref{Conjectures-are-false-theorem} is that for $m>3$ there is at626least one $n$-tuple of $A$ matrices,627with $n=2^m$ such that no $n$-tuple of $B$ matrices of order $n$ can be found to satisfy628construction (H0) under condition (H1).629The proof of Theorem 5 of the Hada\-mard construction paper \cite{Leo14Constructions} shows by630construction that for any $m$,631and any $n$-tuple of $A$ matrices satisfying \eqref{constructions-4}, there is an $n$-tuple of $B$632matrices of order $nc$ that633satisfies construction (H0) under condition (H1), where $c=M(n-1)$, with634\begin{align}635M(q) &:=636\begin{cases}637\lceil \frac{q}{2} \rceil + 1, \quad \text{if~} q \equiv 2,3,4 \quad (\operatorname{mod} 8),638\\639\lceil \frac{q}{2} \rceil \quad \text{otherwise.}640\end{cases}641\label{M-def}642\end{align}643Thus Theorem 5 remains valid.644The question remains as to whether the the order $nc$ is tight or can be reduced.645In the special case where the $n$-tuple of $A$ matrices is mutually amicable, the answer is given by646Corollary 15 of the paper \cite{Leo14Constructions}:647The set of $\{-1,1\}$ matrices of order $c$ contains an $n$-tuple of mutually anti-amicable648Hada\-mard matrices.649So in this special case, the required order can be reduced from $nc$ to $c$.650This leads to the following question.651\begin{question}652In the general case, for any $m>1$, $n=2^m$, for any $n$-tuple of $A$ matrices satisfying653\eqref{constructions-4},654does there always exist an $n$-tuple of $B$ matrices of order $c$ that655satisfies construction (H0) under condition (H1), where $c=M(n-1)$, with $M$ defined by656\eqref{M-def}?657\end{question}658659As a result of Theorems~\ref{Twins-are-strongly-regular-theorem} and660\ref{Conjectures-are-false-theorem},661we see that we have two sequences of strongly regular graphs, $\varDelta_m[-1]$ and $\varDelta_m[1]$662($m \geqslant 1$),663sharing the same parameters,664$v_m = 4^m,$ $k_m = 2^{2 m - 1} - 2^{m - 1},$ $\lambda_m=\mu_m=2^{2 m - 2} - 2^{m - 1},$665but the graphs are isomorphic only for $m=1, 2, 3$.666For these three values of $m$, the existence of667automorphisms of $\varDelta_m$ that swap $\varDelta_m[-1]$ and $\varDelta_m[1]$668as subgraphs \cite[Table 1]{Leo14Constructions}669is remarkable in the light of Theorem~\ref{Conjectures-are-false-theorem}.670671A paper of Bernasconi and Codenotti describes the relationship between bent functions and672their Cayley graphs, implying that a bent function corresponding to a $(v,k,\lambda,n)$ Hada\-mard673difference set has a Cayley graph674that is strongly regular with parameters $(v,k,\lambda,\mu)$ where $\lambda=\mu$~\cite[Lemma67512]{BerC99}.676The current paper notes that for two specific sequences of bent functions, $\sigma_m$ and $\tau_m$,677the corresponding Cayley graphs are not necessarily isomorphic.678679This raises the subject of classifying bent functions via their Cayley graphs, raising the following680questions.681\begin{question}682Which strongly regular graphs with parameters $(v,k,\lambda,\lambda)$ occur as Cayley graphs of bent683functions?684\end{question}685\begin{question}686What is the relationship between other classifications of bent functions and the classification via687Cayley graphs?688\end{question}689This classification is the topic of a paper in preparation \cite{Leo16Classifying}.690691With respect to the specific bent functions $\sigma_m$ and $\tau_m$ investigated here,692one of the anonymous reviewers of an earlier draft of this paper has asked whether each of these693functions are part of a larger class of bent functions.694695The function $\sigma_m$ is a quadratic form, as can be seen from its definition and696its recursive identity \cite[Lemma 7]{Leo14Constructions}.697Specifically, $\sigma_m(0)=0$,698and, in terms of algebraic normal form,699using a particular convention for the mapping of bits to Boolean700variables, the identity is $\sigma_1(x_0,x_1)=x_0 x_1 + x_0$, and701\begin{align*}702\sigma_{m+1}(x_0,x_1,&\ldots,x_{2m},x_{2m+1})703=704\sigma_m(x_0,x_1) +705\sigma_m(x_2,x_3,\ldots,x_{2m},x_{2m+1})706\\707&= x_0 x_1 + x_0 + x_2 x_3 + x_2 + \ldots + x_{2m} x_{2m+1} + x_{2m}.708\end{align*}709710In a paper in preparation \cite{Leo16Classifying},711it is proven that all quadratic bent functions with712the same dimension and weight have isomorphic Cayley graphs.713714As for $\tau_m$, it is a \emph{bent iterative} function715\cite[Theorem~V.4]{CanCCP01cryptographic} \cite[Theorem~2]{CanC03decomposing} \cite{Tok11number},716as can be seen from its definition, and from the proof that it is a bent function717\cite[Theorem 3.1]{Leo15Twin}.718719Since the $\mathcal{PS}^{(-)}$ \emph{partial spread}720bent functions are formed using $m$-di\-mensional subspaces of $\Z^{2m}$ which are disjoint except721for the $0$ vector \cite[p. 95]{Dil74},722these bent functions also have Cayley graphs whose clique number is at least $2^m$.723It could therefore be speculated that $\tau_m$ is also a $\mathcal{PS}^{(-)}$ bent function,724but exhaustive search using SageMathCloud \cite{SageMathCloud} shows that $\tau_3$ cannot be in725$\mathcal{PS}^{(-)}$.726Each clique of size 8 in $\Cay(\tau_3)$ that contains the 0 vector intersects each other such727clique at two vectors, only one of which is the 0 vector \cite{Leo16SMC}.728729%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%730\paragraph*{Acknowledgements:}731732Thanks to Christine Leopardi for her hospitality at Long Beach.733Thanks to Robert Craigen, Joanne Hall, William Martin,734Padraig {\'O} Cath{\'a}in and Judy-anne Osborn for valuable discussions.735This work was begun in 2014 while the author was a Visiting Fellow at the Australian National University,736continued while the author was a Visiting Fellow and a Casual Academic at the University of Newcastle, Australia,737and concluded while the author was an employee of the Bureau of Meteorology of the Australian738Government, and an Honorary Fellow of the University of Melbourne.739Thanks also to the anonymous reviewers of previous drafts of this paper.740741%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%742743\begin{thebibliography}{20}744745\bibitem{BerC99}746A.~Bernasconi and B.~Codenotti.747\newblock Spectral analysis of {Boolean} functions as a graph eigenvalue748problem.749\newblock {\em IEEE Transactions on Computers}, 48(3):345--351, (1999).750751\bibitem{CanCCP01cryptographic}752A.~Canteaut, C.~Carlet, P.~Charpin, and C.~Fontaine.753\newblock On cryptographic properties of the cosets of {R (1, m)}.754\newblock {\em Information Theory, IEEE Transactions on}, 47(4):1494--1513,755(2001).756757\bibitem{CanC03decomposing}758A.~Canteaut and P.~Charpin.759\newblock Decomposing bent functions.760\newblock {\em Information Theory, IEEE Transactions on}, 49(8):2004--2019,761(2003).762763\bibitem{Cra95}764R.~Craigen.765\newblock Signed groups, sequences, and the asymptotic existence of {Hadamard}766matrices.767\newblock {\em J. Combin. Theory Ser. A}, 71(2):241--254, (1995).768769\bibitem{Dil74}770J.~F. Dillon.771\newblock {\em Elementary {Hadamard} Difference Sets}.772\newblock PhD thesis, University of Maryland College Park, Ann Arbor, USA,773(1974).774775\bibitem{GerP74a}776A.~V. Geramita and N.~J. Pullman.777\newblock A theorem of {Hurwitz and Radon} and orthogonal projective modules.778\newblock {\em Proceedings of the American Mathematical Society}, 42(1):51--56,779(1974).780781\bibitem{Hur22}782A.~Hurwitz.783\newblock {\"Uber} die {Komposition} der quadratischen {Formen}.784\newblock {\em Math. Ann.}, 88(1-2):1--25, (1922).785786\bibitem{Leo16Classifying}787P.~Leopardi.788\newblock {Classifying bent functions by their Cayley graphs}.789\newblock In preparation.790791\bibitem{Leo05}792P.~Leopardi.793\newblock A generalized {FFT} for {Clifford} algebras.794\newblock {\em Bulletin of the Belgian Mathematical Society -- Simon Stevin},79511(5):663--688, (2004).796797\bibitem{Leo14Constructions}798P.~Leopardi.799\newblock Constructions for {Hadamard} matrices, {Clifford} algebras, and their800relation to amicability / anti-amicability graphs.801\newblock {\em Australasian Journal of Combinatorics}, 58(2):214--248, (2014).802803\bibitem{Leo15Twin}804P.~Leopardi.805\newblock Twin bent functions and {Clifford} algebras.806\newblock In {\em Algebraic Design Theory and Hadamard Matrices}, 189--199.807Springer, (2015).808809\bibitem{Leo16SMC}810P.~Leopardi.811\newblock Boolean-cayley-graphs, (2016).812\newblock \url{http://tinyurl.com/Boolean-Cayley-graphs} SageMathCloud public813folder. Last accessed 16 April 2017.814815\bibitem{Rad22}816J.~Radon.817\newblock Lineare {Scharen} orthogonaler {Matrizen}.818\newblock {\em Abhandlungen aus dem Mathematischen Seminar der Universit{\"a}t819Hamburg}, 1(1):1--14, (1922).820821\bibitem{SageMathCloud}822{SageMath, Inc.}823\newblock {\em SageMathCloud Online Computational Mathematics}, (2016).824\newblock \url{https://cloud.sagemath.com/}.825826\bibitem{Tok11number}827N.~Tokareva.828\newblock On the number of bent functions from iterative constructions: lower829bounds and hypotheses.830\newblock {\em Adv. in Math. of Comm.}, 5(4):609--621, (2011).831832\bibitem{Wil44}833J.~Williamson.834\newblock Hadamard's determinant theorem and the sum of four squares.835\newblock {\em Duke Math. J.}, 11:65--81, (1944).836837\bibitem{Yiu90}838P.~Y. Yiu.839\newblock Strongly regular graphs and {Hurwitz-Radon} numbers.840\newblock {\em Graphs and Combinatorics}, 6(1):61--69, (1990).841842\end{thebibliography}843844\end{document}845% ----------------------------------------------------------------846847848