From e38229b65f35cf4735df3ecb15aee6e073209410 Mon Sep 17 00:00:00 2001 From: unlockable Date: Tue, 20 Dec 2022 19:37:56 +0800 Subject: [PATCH] =?UTF-8?q?=E7=AC=AC14=E8=AF=BE=E5=AE=8C=E7=BB=93=E3=80=82?= MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit --- 14FiniteGeometriesAndLatinSquares.tex | 168 +++++++++++++++++++++++++- 1 file changed, 167 insertions(+), 1 deletion(-) diff --git a/14FiniteGeometriesAndLatinSquares.tex b/14FiniteGeometriesAndLatinSquares.tex index a3f2bb3..116db90 100644 --- a/14FiniteGeometriesAndLatinSquares.tex +++ b/14FiniteGeometriesAndLatinSquares.tex @@ -549,4 +549,170 @@ Fisher给出了另外一个存在$(v,k,\lambda)$-BIBD的必要条件。 如果$A$和$B$是正交拉丁方,则它们的转置$A\trans$与$B\trans$也是一对正交拉丁方。 \end{proposition} -问题:给定$n$,在所有$n$阶拉丁方中,最多能有多少对两两互相正交的拉丁方组? \ No newline at end of file +\begin{definition} + 两两正交的一组拉丁方叫做正交拉丁方组。 +\end{definition} + +问题:给定$n$,在所有$n$阶拉丁方中,一个正交拉丁方组最多能有多少个? + +我们有 + +\begin{theorem} + 如果$A_1, A_2, \dots, A_r$是$r$个两两正交的$n$阶正交拉丁方,则$r \leq n - 1$。 +\end{theorem} + +\begin{proof} + 根据正交拉丁方的性质,我们总可以对每个拉丁方$A_k$的元素进行适当的置换$p_k$,把$A_k$的第一行变为$1, 2, \dots, n$。置换后的$r$个拉丁方仍旧两两正交。因此我们不妨就直接设原来的$r$个拉丁方就是如下形式: + \[A_k = \begin{bmatrix} + 1 & 2 & \cdots & n\\ + a_{21}^{(k)} & a_{22}^{(k)} & \cdots & a_{2n}^{(k)}\\ + \vdots & \vdots & \ddots & \vdots\\ + a_{n1}^{(k)} & a_{n2}^{(k)} & \cdots & a_{nn}^{(k)} + \end{bmatrix}, + k = 1, 2, \dots, r\] + + 下面我们考察$a_{21}^{(k)}$。每列不可能出现两个1,因此$a_{21}^{(k)} \neq 1, k = 1, 2, \dots, n$。当$h \neq k$时,$a_{21}^{(h)} \neq a_{21}^{(k)}$,否则$(a_{21}^{(h)}, a_{21}^{(k)})$与第一行的某一个是相同的,不符合正交拉丁方的定义。因此 + \[a_{21}^(1), a_{21}^{(2)}, \dots, a_{21}^{(r)}\] + 是两两不同且都不等于1的。这样的数最多有$n-1$个,因此$r \leq n - 1$。 +\end{proof} + +含有$n-1$个$n$阶正交拉丁方组称为$n$阶拉丁方的完备正交组。完备正交组的存在性问题是至今尚未解决的问题。 + +通过元素置换将拉丁方的第一行化为自然顺序的过程称为规范化过程。如果对一个正交拉丁方组进行这个过程,也成为正交拉丁方规范化过程。这样的过程有利于比较拉丁方是不是``本质不同''的。 + +1782年,欧拉提出了以下猜想: + +\begin{conjecture} + 如果$n = 4k + 2$,则不存在任何一对$n$阶正交拉丁方。 +\end{conjecture} + +在100多年后的1901年,法国人Tarry证明了$n = 6$时Euler猜想的正确性。然而,在1960年前后,Bose,Parker,Shrikhande相继证明了 +\begin{theorem} + 如果$n \neq 2, 6$,则必然存在一对$n$阶正交拉丁方。 +\end{theorem} + +\subsection{幻方} +观察这样一组正交拉丁方: +\begin{table}[H] + \centering + \begin{tabular}{|c|c|c|c|} + \hline + 0 & 1 & 2 & 3\\ + \hline + 1 & 0 & 3 & 2\\ + \hline + 2 & 3 & 0 & 1\\ + \hline + 3 & 2 & 1 & 0\\ + \hline + \end{tabular} + \hspace{2cm} + \begin{tabular}{|c|c|c|c|} + \hline + 0 & 1 & 2 & 3\\ + \hline + 3 & 2 & 1 & 0\\ + \hline + 0 & 1 & 2 & 3\\ + \hline + 2 & 3 & 0 & 1\\ + \hline + \end{tabular} +\end{table} +将其合在一起我们得到 +\begin{table}[H] + \centering + \begin{tabular}{|c|c|c|c|} + \hline + 00 & 11 & 22 & 33\\ + \hline + 13 & 02 & 31 & 20\\ + \hline + 21 & 31 & 03 & 12\\ + \hline + 32 & 23 & 10 & 01\\ + \hline + \end{tabular} +\end{table} + +如果我们把表格中的每个数都看作一个两位的四进制数,再将它们转写回十进制,就有 +将其合在一起我们得到 +\begin{table}[H] + \centering + \begin{tabular}{|c|c|c|c|} + \hline + 0 & 5 & 10 & 15\\ + \hline + 7 & 2 & 13 & 8\\ + \hline + 9 & 12 & 3 & 6\\ + \hline + 14 & 11 & 4 & 1\\ + \hline + \end{tabular} +\end{table} + +它每一行的和与每一列的和都是定值。这样的数表称为幻方(弱幻方)。如果每一列、每一行、对角线的和都相等,就是强幻方,例如下面这个幻方。 + +\begin{table}[H] + \centering + \begin{tabular}{|c|c|c|} + \hline + 4 & 9 & 2\\ + \hline + 3 & 5 & 7\\ + \hline + 8 & 1 & 6\\ + \hline + \end{tabular} +\end{table} + +有限几何与拉丁方也有着联系。例如: + +\begin{figure}[H] + \centering + \begin{tikzpicture} + \node[draw=black, fill=black, circle, inner sep=0, minimum size=8pt] (a1) at (0,0) {}; + \node[draw=black, fill=black, circle, inner sep=0, minimum size=8pt] (a2) at (2,0) {}; + \node[draw=black, fill=black, circle, inner sep=0, minimum size=8pt] (a3) at (4,0) {}; + \node[draw=black, fill=black, circle, inner sep=0, minimum size=8pt] (b1) at (0,2) {}; + \node[draw=black, fill=black, circle, inner sep=0, minimum size=8pt] (b2) at (2,2) {}; + \node[draw=black, fill=black, circle, inner sep=0, minimum size=8pt] (b3) at (4,2) {}; + \node[draw=black, fill=black, circle, inner sep=0, minimum size=8pt] (c1) at (0,4) {}; + \node[draw=black, fill=black, circle, inner sep=0, minimum size=8pt] (c2) at (2,4) {}; + \node[draw=black, fill=black, circle, inner sep=0, minimum size=8pt] (c3) at (4,4) {}; + + \node at (0,4.5) {0}; + \node at (2,4.5) {1}; + \node at (4,4.5) {2}; + \draw (a1)--(a3) (b1)--(b3) (c1)--(c3) (a1)--(c1) (a2)--(c2) (a3)--(c3); + \draw (c1)--(a3) (c2) ..controls (2.4,3.8) and (3.8,3)..(b3).. controls (3.8,1) and (2,0.6)..(a1); + \draw (c3) ..controls (2,3.4) and (0.2,3).. (b1) .. controls(0.2,1) and (1.6,0.2) ..(a2); + \draw[dashed] (c1)..controls (2,3.4) and (3.8,3) .. (b3) .. controls (3.8,1) and (2.4,0.2) .. (a2); + \draw[dashed] (c2)..controls (1.6, 3.8) and (0.2,3) .. (b1) .. controls (0.2,1) and (2,0.6) ..(a3); + \draw[dashed] (a1)--(c3); + \end{tikzpicture} +\end{figure} + +\begin{table}[H] + \centering + \begin{tabular}{|c|c|c|} + \hline + 0 & 1 & 2\\ + \hline + 2 & 0 & 1\\ + \hline + 1 & 2 & 0\\ + \hline + \end{tabular} + \hspace{1cm} + \begin{tabular}{|c|c|c|} + \hline + 0 & 1 & 2\\ + \hline + 1 & 2 & 0\\ + \hline + 2 & 0 & 1\\ + \hline + \end{tabular} +\end{table}