主要内容\(\newcommand{\Ima}{\rm Im }
\newcommand{\N}{\mathbb N}
\newcommand{\Z}{\mathbb Z}
\newcommand{\Q}{\mathbb Q}
\newcommand{\R}{\mathbb R}
\newcommand{\F}{\mathbb F}
\newcommand{\C}{\mathbb C}
\newcommand{\K}{\mathbb K}
\newcommand{\myunit}{1 cm}
\newcommand{\blue}[1]{{\color{blue}#1}}
\newcommand\iddots{\mathinner{
\kern1mu\raise1pt{.}
\kern2mu\raise4pt{.}
\kern2mu\raise7pt{\Rule{0pt}{7pt}{0pt}.}
\kern1mu
}}
\tikzset{
node style sp/.style={draw,circle,minimum size=\myunit},
node style ge/.style={circle,minimum size=\myunit},
arrow style mul/.style={draw,sloped,midway,fill=white},
arrow style plus/.style={midway,sloped,fill=white},
}
\newcommand{\lt}{<}
\newcommand{\gt}{>}
\newcommand{\amp}{&}
\definecolor{fillinmathshade}{gray}{0.9}
\newcommand{\fillinmath}[1]{\mathchoice{\colorbox{fillinmathshade}{$\displaystyle \phantom{\,#1\,}$}}{\colorbox{fillinmathshade}{$\textstyle \phantom{\,#1\,}$}}{\colorbox{fillinmathshade}{$\scriptstyle \phantom{\,#1\,}$}}{\colorbox{fillinmathshade}{$\scriptscriptstyle\phantom{\,#1\,}$}}}
\)
节 2.3 分块矩阵
练习 练习
基础题.
1.
设\(M_1\)是\(m_1\)阶方阵,\(N_1\)是\(n_1\)阶方阵,\(A_{ij}\)为\(m_i\times m_j\)矩阵,\(B_{kl}\)为\(m_k\times n_l\)矩阵,\(K\)是\(m_1\times m_2\)矩阵,\(L\)是 \(n_1\times n_2\)矩阵,计算:
-
\(\begin{pmatrix}
M_1 &{\bf 0}\\
{\bf 0} & E_{m_2}
\end{pmatrix}\begin{pmatrix}
B_{11} & B_{12}\\
B_{21} & B_{22}
\end{pmatrix}, \begin{pmatrix}
B_{11} & B_{12}\\
B_{21} & B_{22}
\end{pmatrix}\begin{pmatrix}
N_1 & {\bf 0}\\
{\bf 0} & E_{n_2}
\end{pmatrix}\);
-
\(\begin{pmatrix}
{\bf 0} & E_{m_2}\\
E_{m_1} & {\bf 0}
\end{pmatrix}\begin{pmatrix}
B_{11} & B_{12}\\
B_{21} & B_{22}
\end{pmatrix}, \begin{pmatrix}
B_{11} & B_{12}\\
B_{21} & B_{22}
\end{pmatrix}\begin{pmatrix}
{\bf 0} & E_{n_1}\\
E_{n_2} & {\bf 0}
\end{pmatrix}\);
-
\(\begin{pmatrix}
E_{m_1} & K\\
{\bf 0} & E_{m_2}
\end{pmatrix}\begin{pmatrix}
B_{11} & B_{12}\\
B_{21} & B_{22}
\end{pmatrix},\ \begin{pmatrix}
B_{11} & B_{12}\\
B_{21} & B_{22}
\end{pmatrix}\begin{pmatrix}
E_{n_1} & L\\
{\bf 0} & E_{n_2}
\end{pmatrix}\);
-
\(\begin{pmatrix}
A_{11} & A_{12} & A_{13}\\
{\bf 0} & A_{22} & A_{23}\\
{\bf 0} & {\bf 0} & A_{33}
\end{pmatrix}\begin{pmatrix}
B_{11} & B_{12} & B_{13}\\
{\bf 0} & B_{22} & B_{23}\\
{\bf 0} & {\bf 0} & B_{33}
\end{pmatrix}\)。
解答.
-
\begin{equation*}
\begin{pmatrix}
M_1 & {\bf 0}\\
{\bf 0} & E_{m_2}
\end{pmatrix}\begin{pmatrix}
B_{11} & B_{12}\\
B_{21} & B_{22}
\end{pmatrix}= \begin{pmatrix}
MB_{11} & MB_{12}\\
B_{21} & B_{22}
\end{pmatrix},
\end{equation*}
\begin{equation*}
\begin{pmatrix}
B_{11} & B_{12}\\
B_{21} & B_{22}
\end{pmatrix}\begin{pmatrix}
N_1 & {\bf 0}\\
{\bf 0} & E_{n_2}
\end{pmatrix}=\begin{pmatrix}
B_{11}N & B_{12}\\
B_{21}N & B_{22}
\end{pmatrix};
\end{equation*}
-
\begin{equation*}
\begin{pmatrix}
{\bf 0} & E_{m_2}\\
E_{m_1} & {\bf 0}
\end{pmatrix}\begin{pmatrix}
B_{11} & B_{12}\\
B_{21} & B_{22}
\end{pmatrix}=\begin{pmatrix}
B_{21} & B_{22}\\
B_{11} & B_{12}
\end{pmatrix},
\end{equation*}
\begin{equation*}
\begin{pmatrix}
B_{11} & B_{12}\\
B_{21} & B_{22}
\end{pmatrix}\begin{pmatrix}
{\bf 0} & E_{n_1}\\
E_{n_2} & {\bf 0}
\end{pmatrix}=\begin{pmatrix}
B_{12} & B_{11}\\
B_{22} & B_{21}
\end{pmatrix};
\end{equation*}
-
\begin{equation*}
\begin{pmatrix}
E_{m_1} & K\\
{\bf 0} & E_{m_2}
\end{pmatrix}\begin{pmatrix}
B_{11} & B_{12}\\
B_{21} & B_{22}
\end{pmatrix}=\begin{pmatrix}
B_{11}+KB_{21} & B_{12}+KB_{22}\\
B_{21} & B_{22}
\end{pmatrix},
\end{equation*}
\begin{equation*}
\begin{pmatrix}
B_{11} & B_{12}\\
B_{21} & B_{22}
\end{pmatrix}\begin{pmatrix}
E_{n_1} & L\\
{\bf 0} & E_{n_2}
\end{pmatrix}=\begin{pmatrix}
B_{11} & B_{11}L+B_{12}\\
B_{21} & B_{21}L+B_{22}
\end{pmatrix};
\end{equation*}
-
\(\displaystyle \begin{pmatrix}
A_{11}B_{11} & A_{11}B_{12}+A_{12}B_{22} & A_{11}B_{13}+A_{12}B_{23}+A_{13}B_{33}\\
{\bf 0} & A_{22}B_{22} & A_{22}B_{23}+A_{23}B_{33}\\
{\bf 0} & {\bf 0} & A_{33}B_{33}
\end{pmatrix}.\)
2.
将矩阵
\(A=\begin{pmatrix}
1 & 0 & 1 & -1\\
0 & 1 & 0 & 1\\
0 & 0 & 0 & 1\\
0 & 0 & 2 & 0
\end{pmatrix}\)与
\(B=\begin{pmatrix}
1 & -3 & -1 & 2\\
-3 & 2 & -5 & -4\\
0 & 0 & 4 & 3\\
0 & 0 & 2 & 1
\end{pmatrix}\) 适当分块后,计算
\(AB\)。
解答.
令
\begin{equation*}
A_1=\begin{pmatrix}
1 & -1\\
0 & 1
\end{pmatrix},A_2=\begin{pmatrix}
0 & 1\\
2 & 0
\end{pmatrix},
\end{equation*}
\begin{equation*}
B_1=\begin{pmatrix}
1 & -3\\
-3 & 2
\end{pmatrix},B_2=\begin{pmatrix}
-1 & 2\\
-5 & -4
\end{pmatrix}, B_3=\begin{pmatrix}
4 & 3\\
2 & 1
\end{pmatrix},
\end{equation*}
则\(A=\begin{pmatrix}
E_2 & A_1\\
{\bf 0} & A_2
\end{pmatrix},\ B=\begin{pmatrix}
B_1 & B_2\\
{\bf 0} & B_3
\end{pmatrix}\),因此
\begin{equation*}
AB=\begin{pmatrix}
B_1 & B_2+A_1B_3\\
{\bf 0} & A_2B_3
\end{pmatrix}=\begin{pmatrix}
1 & -3 & 1 & 4\\
-3 & 2 & -3 & -3\\
0 & 0 & 2 & 1\\
0 & 0 & 8 & 6
\end{pmatrix}.
\end{equation*}
3.
设\(A=\begin{pmatrix}
3 & 1 & 0 & 0\\
0 & 3 & 0 & 0\\
0 & 0 & 3 & 9\\
0 & 0 & 1 & 3
\end{pmatrix}\),计算
-
-
解答.
令\(A_1=\begin{pmatrix}
3&1\\0&3
\end{pmatrix}\),\(A_2=\begin{pmatrix}
3&9\\1&3
\end{pmatrix}\),则\(A=\begin{pmatrix}
A_1&{\bf 0}\\{\bf 0}&A_2
\end{pmatrix}\)。
-
注意到
\begin{equation*}
A_1=3E_2+J,
\end{equation*}
其中\(J=\begin{pmatrix}
0&1\\0&0
\end{pmatrix}\),且当\(i\geq 2\)时,\(J^i=0\),所以
\begin{equation*}
\begin{array}{ll}
A_1^n&=\sum\limits_{i=0}^n C_n^i\cdot (3E_2)^{n-i}\cdot J^i\\
&=(3E_2)^n+n(3E_2)^{n-1}J\\
&=\begin{pmatrix}
3^n&n\cdot 3^{n-1}\\0&3^n
\end{pmatrix}.\end{array}
\end{equation*}
由于\(A_2=\begin{pmatrix}
3\\1
\end{pmatrix}\begin{pmatrix}
1&3
\end{pmatrix}\),所以
\begin{equation*}
\begin{array}{ll}
A_2^n&=[\begin{pmatrix}
3\\1
\end{pmatrix}\begin{pmatrix}
1&3
\end{pmatrix}]^n\\
&=\begin{pmatrix}
3\\1
\end{pmatrix}[\begin{pmatrix}
1&3
\end{pmatrix}\begin{pmatrix}
3\\1
\end{pmatrix}]^{n-1}\begin{pmatrix}
1&3
\end{pmatrix}\\
&=6^{n-1} \begin{pmatrix}
3&9\\1&3
\end{pmatrix}.\end{array}
\end{equation*}
因此
\begin{equation*}
A^n=\begin{pmatrix}
A_1^n&{\bf 0}\\{\bf 0}&A_2^n
\end{pmatrix}=\begin{pmatrix}
3^n&n\cdot 3^{n-1}&0&0\\0&3^n&0&0\\0&0&3\cdot 6^{n-1}&9\cdot 6^{n-1}\\0&0&6^{n-1}&3\cdot 6^{n-1}
\end{pmatrix}.
\end{equation*}
-
由于\(A^T=\begin{pmatrix}
A_1^T & {\bf 0}\\
{\bf 0} & A_2^T
\end{pmatrix},\) 所以
\begin{equation*}
AA^T=\begin{pmatrix}
A_1A_1^T & {\bf 0}\\
{\bf 0} & A_2{A_2}^T
\end{pmatrix}=\begin{pmatrix}
10 & 3 & 0 & 0\\
3 & 9 & 0 & 0\\
0 & 0 & 90 & 30\\
0 & 0 & 30 & 10
\end{pmatrix},
\end{equation*}
\begin{equation*}
A^TA=\begin{pmatrix}
A_1^TA_1 & {\bf 0}\\
{\bf 0} & {A_2}^TA_2
\end{pmatrix}=\begin{pmatrix}
9 & 3 & 0 & 0\\
3 & 10 & 0 & 0\\
0 & 0 & 10 & 30\\
0 & 0 & 30 & 90
\end{pmatrix}.
\end{equation*}
提高题.
4.
设
\(A\)为
\(n\)阶方阵,证明:若对任意
\(n\)维列向量
\(\alpha\)都有
\(A\alpha={\bf 0}\),那么
\(A={\bf 0}\)。
解答.
根据已知条件,对任意
\(1\leq i\leq n\),
\(A\varepsilon_i={\bf 0}\),而
\(A\varepsilon_i\)表示
\(A\)的第
\(i\)列,所以
\(A\)的每一列全为零,由此推出
\(A={\bf 0}\)。
5.
设
\(\varepsilon_i\)为
\(n\)维标准单位列向量,
\(E_{ij}\)是
\(n\)阶基础矩阵。证明:
-
\(\varepsilon_i^T\varepsilon_j= \delta_{ij}\),其中
\(\delta_{ij}\)是
Kronecker符号,即
\(\delta_{ij}=\left\{\begin{array}{cl}
1, & i=j,\\
0, & i\neq j;
\end{array}\right.\)
-
\(\varepsilon_i\varepsilon_j^T= E_{ij}\);
-
\(\displaystyle E_{ij}E_{kl}=\left\{\begin{array}{cl}
E_{il}, & j=k,\\
{\bf 0}, & j\neq k;
\end{array}\right.\)
-
设
\(A\)是
\(n\)阶方阵,则
\(E_{ij}A\)的第
\(i\)行是
\(A\)的第
\(j\)行元,其它元为
\(0\);
-
设
\(A\)是
\(n\)阶方阵,则
\(AE_{ij}\)第
\(j\)列是
\(A\)的第
\(i\)列元,其它元为
\(0\);
-
设
\(A\)是
\(n\)阶方阵,则
\(E_{ij}AE_{kl}=a_{jk}E_{il}\)。
解答.
-
方法二:分别将单位矩阵\(E_n\)按行和列分块,得
\begin{equation*}
E_n=\begin{pmatrix}
\varepsilon_1^T\\\vdots\\\varepsilon_n^T
\end{pmatrix},\ E_n=\begin{pmatrix}
\varepsilon_1&\varepsilon_2&\cdots&\varepsilon_n
\end{pmatrix}.
\end{equation*}
由于\(E_n=E_n^2=\begin{pmatrix}
\varepsilon_1^T\\\vdots\\\varepsilon_n^T
\end{pmatrix}\begin{pmatrix}
\varepsilon_1&\varepsilon_2&\cdots&\varepsilon_n
\end{pmatrix}=(\varepsilon_i^T\varepsilon_j)_{n\times n}\),比较两边矩阵的第\((i,j)\)元素,得\(\varepsilon_i^T\varepsilon_j= \delta_{ij}\)。
-
\begin{equation*}
\begin{array}{ccccccccc}
&&\mbox{第}j\mbox{列}&&&&\mbox{第}j\mbox{列}&&\\\varepsilon_i\varepsilon_j^T=\varepsilon_i (0&\cdots&1&\cdots&0)=(0&\cdots&\varepsilon_i&\cdots&0)=E_{ij}.
\end{array}
\end{equation*}
-
\begin{equation*}
E_{ij}=\varepsilon_i\varepsilon_j^T,\ E_{kl}=\varepsilon_k\varepsilon_l^T,
\end{equation*}
故
\begin{equation*}
E_{ij}E_{kl}=(\varepsilon_i\varepsilon_j^T)(\varepsilon_k\varepsilon_l^T)=\varepsilon_i(\varepsilon_j^T\varepsilon_k)\varepsilon_l^T.
\end{equation*}
又\(\varepsilon_j^T\varepsilon_k= \delta_{jk}\),因此
\begin{equation*}
E_{ij}E_{kl}=\delta_{jk}\varepsilon_i\varepsilon_l^T=\delta_{jk}E_{il}=\left\{\begin{array}{cc}
E_{il},&j=k,\\{\bf 0},&j\neq k.
\end{array}\right.
\end{equation*}
-
将\(E_{ij}\)按行分块,得\(E_{ij}=\begin{pmatrix}
{\bf 0}\\
\vdots\\
\varepsilon_j^T\\
\vdots\\
{\bf 0}
\end{pmatrix}\ (\mbox{第}i\mbox{行})\),则
\begin{equation*}
E_{ij}A=\begin{pmatrix}
{\bf 0}\\
\vdots\\
\varepsilon_j^TA\\
\vdots\\
{\bf 0}
\end{pmatrix}(\mbox{第}i\mbox{行})
\end{equation*}
而\(\varepsilon_j^TA\)表示\(A\)的第\(j\)行,因此\(E_{ij}A\)将\(A\)的第\(i\)行变为第\(j\)行元,其他元变为\(0\)。
-
将\(E_{ij}\)按列分块,得
\begin{equation*}
\begin{array}{ccccc}
&&\mbox{第}j\mbox{列}&&\\
E_{ij}=({\bf 0}&\cdots&\varepsilon_i&\cdots&{\bf 0}),
\end{array}
\end{equation*}
则
\begin{equation*}
\begin{array}{ccccc}
&&\mbox{第}j\mbox{列}&&\\
AE_{ij}=({\bf 0}&\cdots&A\varepsilon_i&\cdots&{\bf 0}).
\end{array}
\end{equation*}
而\(A\varepsilon_i\)表示\(A\)的第\(i\)列,因此\(AE_{ij}\)将\(A\)的第\(j\)列变为第\(i\)列元,其他元变为\(0\)。
-
将
\(A\)按列分块得
\(A=\begin{pmatrix}
A_1&A_2&\cdots&A_n
\end{pmatrix}\),由
项 2.3.5.e 知
\begin{equation*}
\begin{array}{ccccc}
&&\mbox{第}l\mbox{列}&&\\
AE_{kl}=(0&\cdots&A_k&\cdots&0),
\end{array}
\end{equation*}
再由
项 2.3.5.d,
\(E_{ij}AE_{kl}\)将
\(AE_{kl}\)的第
\(i\)行变为第
\(j\)行元,其他元变为
\(0\)。因此
\begin{equation*}
E_{ij}AE_{kl}=a_{jk}E_{il}.
\end{equation*}
6.
设
\(A\)为
\(n\)阶方阵,证明:若对任意
\(n\)阶方阵
\(B\)都有
\({\rm tr}(AB)=0\),那么
\(A={\bf 0}\)。
解答.
根据已知,对任意\(1\leq i,j \leq n\),都有\({\rm tr}(AE_{ij})=0\),而
\begin{equation*}
AE_{ij}=\begin{pmatrix}
& & & \mbox{第}j\mbox{列} & & & \\
0 & \cdots & 0 & a_{1i} & 0 & \cdots & 0\\
\vdots & & \vdots & \vdots & \vdots & & \vdots\\
0 & \cdots & 0 & a_{ji} & 0 & \cdots & 0\\
\vdots & & \vdots & \vdots & \vdots & & \vdots\\
0 & \cdots & 0 & a_{ni} & 0 & \cdots & 0\end{pmatrix},
\end{equation*}
因此\({\rm tr}(AE_{ij})=a_{ji}=0\),从而\(A={\bf 0}\)。
7.
设
\begin{equation*}
A=\begin{pmatrix}
a_1E_{n_1} & & &\\
& a_2E_{n_2} & &\\
& & \ddots & \\
& & & a_rE_{n_r}
\end{pmatrix}
\end{equation*}
其中\(a_i\neq a_j\)(当\(i\neq j\)时),\(E_{n_i}\)是\(n_i\)阶单位矩阵。证明:与\(A\)可交换的矩阵只能是分块对角矩阵 \({\rm diag}(B_1, \ldots,B_r)\),其中\(B_i\)为\(n_i\)阶方阵,\(i=1,\ldots,r\)。
解答.
设\(B=(B_{ij})\)与矩阵\(A\)可交换,其中\(B_{ij}\)是\(n_i\times n_j\)矩阵。由
\begin{equation*}
\begin{array}{ll}
AB & =\begin{pmatrix}
a_1E_{n_1}B_{11}&a_1E_{n_1}B_{12}&\cdots&a_1E_{n_1}B_{1r}\\
a_2E_{n_2}B_{21}&a_2E_{n_2}B_{22}&\cdots&a_2E_{n_2}B_{2r}\\
\vdots&\vdots& &\vdots\\
a_rE_{n_r}B_{r1}&a_rE_{n_r}B_{r2}&\cdots&a_rE_{n_r}B_{rr}
\end{pmatrix}\\
&=\begin{pmatrix}
a_1B_{11}&a_1B_{12}&\cdots&a_1B_{1r}\\
a_2B_{21}&a_2B_{22}&\cdots&a_2B_{2r}\\
\vdots&\vdots&&\vdots\\
a_rB_{r1}&a_rB_{r2}&\cdots&a_rB_{rr}
\end{pmatrix}, \end{array}
\end{equation*}
\begin{equation*}
\begin{array}{ll}
BA & =\begin{pmatrix}
B_{11}a_1E_{n_1}&B_{12}a_2E_{n_2}&\cdots&B_{1r}a_rE_{n_r}\\
B_{21}a_1E_{n_1}&B_{22}a_2E_{n_2}&\cdots&B_{2r}a_rE_{n_r}\\
\vdots&\vdots& &\vdots\\
B_{r1}a_1E_{n_1}&B_{r2}a_2E_{n_2}&\cdots&B_{rr}a_rE_{n_r}
\end{pmatrix}\\
& =\begin{pmatrix}
a_1B_{11}&a_2B_{12}&\cdots&a_rB_{1r}\\
a_1B_{21}&a_2B_{22}&\cdots&a_rB_{2r}\\
\vdots&\vdots& &\vdots\\
a_1B_{r1}&a_2B_{r2}&\cdots&a_rB_{rr}
\end{pmatrix},\end{array}
\end{equation*}
可知:\(a_iB_{ij}=a_jB_{ij},\ \forall 1\leq i,j\leq n\),即
\begin{equation*}
(a_i-a_j)B_{ij}=0,\ \forall 1\leq i,j\leq n,
\end{equation*}
而当\(i\neq j\)时,\(a_i\neq a_j\),故\(B_{ij}=0(i\neq j)\)。因此
\begin{equation*}
B={\rm diag} (B_{11}, B_{22}, \cdots,B_{rr})
\end{equation*}
为分块对角矩阵。
8.
计算
\(\begin{pmatrix}
{\bf 0} & E_4\\
1 & {\bf 0}
\end{pmatrix}^n\),其中
\(n=2,3,4,5\)。
解答.
因为\(A=\begin{pmatrix}
\varepsilon_5& \varepsilon_1&\varepsilon_2&\varepsilon_3&\varepsilon_4
\end{pmatrix}\),所以
\begin{equation*}
\begin{array}{cl}
A^2 &= A\begin{pmatrix}
\varepsilon_5& \varepsilon_1&\varepsilon_2&\varepsilon_3&\varepsilon_4
\end{pmatrix}\\
& =\begin{pmatrix}
A\varepsilon_5& A\varepsilon_1&A\varepsilon_2&A\varepsilon_3&A\varepsilon_4
\end{pmatrix}\\
&= \begin{pmatrix}
\varepsilon_4&\varepsilon_5& \varepsilon_1&\varepsilon_2&\varepsilon_3
\end{pmatrix}\\
& =\begin{pmatrix}
0&E_3\\E_2&0
\end{pmatrix},\end{array}
\end{equation*}
\begin{equation*}
\begin{array}{ll}
A^3 & =AA^2\\
& =\begin{pmatrix}
A\varepsilon_4&A\varepsilon_5& A\varepsilon_1&A\varepsilon_2&A\varepsilon_3
\end{pmatrix}\\
& =\begin{pmatrix}
\varepsilon_3&\varepsilon_4&\varepsilon_5& \varepsilon_1&\varepsilon_2
\end{pmatrix}\\
& =\begin{pmatrix}
0&E_2\\E_3&0
\end{pmatrix},\end{array}
\end{equation*}
\begin{equation*}
\begin{array}{ll}
A^4 & =AA^3\\
& =\begin{pmatrix}
A\varepsilon_3&A\varepsilon_4&A\varepsilon_5& A\varepsilon_1&A\varepsilon_2
\end{pmatrix}\\
& =\begin{pmatrix}
\varepsilon_2&\varepsilon_3&\varepsilon_4&\varepsilon_5& \varepsilon_1
\end{pmatrix}\\
& =\begin{pmatrix}
0&1\\E_4&0
\end{pmatrix},\end{array}
\end{equation*}
\begin{equation*}
\begin{array}{ll}
A^5 & =AA^4\\
& =\begin{pmatrix}
A\varepsilon_2&A\varepsilon_3&A\varepsilon_4&A\varepsilon_5& A\varepsilon_1
\end{pmatrix}\\
& =\begin{pmatrix}
\varepsilon_1&\varepsilon_2&\varepsilon_3&\varepsilon_4&\varepsilon_5
\end{pmatrix}\\
& =E_5.\end{array}
\end{equation*}
9.
设\(n\)阶基础循环矩阵
\begin{equation*}
C=\begin{pmatrix}
0 & 1 & 0 & \cdots & 0\\
0 & 0 & 1 & \cdots & 0\\
\vdots & \vdots & \vdots & \ddots & \vdots\\
0 & 0 & 0 & \cdots & 1\\
1 & 0 & 0 & \cdots & 0
\end{pmatrix},
\end{equation*}
证明:对任意\(n\)阶方阵\(A\),
-
\(CA\)相当于把
\(A\)的每一行向上移一行,第1行换到最后一行;
-
\(AC\)相当于把
\(A\)的每一列向右移一列,最后一列换到第1列。
解答.
-
将矩阵\(C\)行分块,则
\begin{equation*}
CA=\begin{pmatrix}
\varepsilon_2^T\\
\varepsilon_3^T\\
\vdots\\
\varepsilon_n^T\\
\varepsilon_1^T
\end{pmatrix}A=\begin{pmatrix}
\varepsilon_2^TA\\
\varepsilon_3^TA\\
\vdots\\
\varepsilon_n^TA\\
\varepsilon_1^TA
\end{pmatrix}
\end{equation*}
相当于把\(A\)的每一行向上移一行,第1行换到最后一行。
-
将矩阵\(C\)列分块,则
\begin{equation*}
\begin{array}{ll}
AC & =A\begin{pmatrix}
\varepsilon_n &
\varepsilon_1 & \varepsilon_2 &
\cdots & \varepsilon_{n-1}
\end{pmatrix}\\
& =\begin{pmatrix}
A\varepsilon_n &
A\varepsilon_1 & A\varepsilon_2 &
\cdots & A\varepsilon_{n-1}
\end{pmatrix}
\end{array}
\end{equation*}
相当于把\(A\)的每一列向右移一列,最后一列换到第1列。
10.
设\(C\)为\(n\)阶基础循环矩阵, 证明:对任意\(1\leq k\leq n\),
\begin{equation*}
C^k=\begin{pmatrix}
{\bf 0} & E_{n-k}\\
E_k & {\bf 0}
\end{pmatrix}.
\end{equation*}
解答.
当
\(k=1\)时,结论显然成立。假设
\(C^{k-1}=\begin{pmatrix}
{\bf 0} & E_{n-k+1}\\
E_{k-1} & {\bf 0}
\end{pmatrix}\),由
项 2.3.9.a知
\begin{equation*}
C^k=CC^{k-1}
\end{equation*}
相当于把\(C^{k-1}\)的每一行向上移一行,第1行换到最后一行,所以
\begin{equation*}
C^k=\begin{pmatrix}
{\bf 0} & E_{n-k}\\
E_{k} & {\bf 0}
\end{pmatrix}.
\end{equation*}
11.
下列形式的矩阵称为循环矩阵
\begin{equation*}
\begin{pmatrix}
a_1 & a_2 & a_3 & \cdots & a_n\\
a_n & a_1 & a_2 & \cdots & a_{n-1}\\
a_{n-1} & a_n & a_1 & \cdots & a_{n-2}\\
\vdots & \vdots & \vdots & & \vdots\\
a_2 & a_3 & a_4 & \cdots & a_1
\end{pmatrix}.
\end{equation*}
证明:
-
\(n\)阶循环矩阵
\(A\)必可表示成
\(n\)阶基本循环矩阵
\(C\)的多项式;
-
解答.
-
由于
\begin{equation*}
A=a_1E_n+a_2\begin{pmatrix}
{\bf 0} & E_{n-1}\\
1 & 0\end{pmatrix}+a_3\begin{pmatrix}
{\bf 0} & E_{n-2}\\
E_2 & {\bf 0}\end{pmatrix}+\cdots +a_n\begin{pmatrix}
{\bf 0} & 1\\
E_{n-1} & {\bf 0}\end{pmatrix},
\end{equation*}
\begin{equation*}
A=a_1E_n+a_2C+a_3C^2+\cdots +a_n C^{n-1}
\end{equation*}
可表示成\(n\)阶基本循环矩阵\(C\)的多项式。
-
设
\(A,B\)是
\(n\)阶循环矩阵,则
\(A,B\)都可表成
\(n\)阶基本循环矩阵
\(C\)的多项式,其乘积矩阵
\(AB\)仍是
\(C\)的多项式,因此
\(AB\)仍为循环矩阵。
12.
每一行和每一列有且仅有一个1,其余元素均为0的\(n\)阶方阵称为\(n\)阶置换矩阵,证明:
-
若
\(P\)是
\(n\)阶置换矩阵,则
\(P^TP=E_n\);
-
若
\(P_1,P_2\)都是
\(n\)阶置换矩阵,则
\(P_1P_2\)也是
\(n\)阶置换矩阵。
解答.
-
因为\(P\)是\(n\)阶置换矩阵,所以按列分块
\begin{equation*}
P=\begin{pmatrix}
\varepsilon_{i_1} & \varepsilon_{i_2} & \cdots & \varepsilon_{i_n}
\end{pmatrix},
\end{equation*}
其中\((i_1,i_2,\ldots,i_n)\)是\((1,2,\ldots,n)\)的一个排列。
\begin{equation*}
P^TP=\begin{pmatrix}
\varepsilon_{i_1}^T\varepsilon_{i_1} & \varepsilon_{i_1}^T\varepsilon_{i_2} & \cdots & \varepsilon_{i_1}^T\varepsilon_{i_n}\\
\varepsilon_{i_2}^T\varepsilon_{i_1} & \varepsilon_{i_2}^T\varepsilon_{i_2} & \cdots & \varepsilon_{i_2}^T\varepsilon_{i_n}\\
\vdots & \vdots & & \vdots\\
\varepsilon_{i_n}^T\varepsilon_{i_1} & \varepsilon_{i_n}^T\varepsilon_{i_2} & \cdots & \varepsilon_{i_n}^T\varepsilon_{i_n}
\end{pmatrix},
\end{equation*}
注意到
\begin{equation*}
\varepsilon_{i_j}^T\varepsilon_{i_k}=\left\{\begin{array}{ll}
1, & j=k,\\
0, & j\neq k, \end{array}\right.
\end{equation*}
因此\(P^TP=E_n\)。
-
设
\begin{equation*}
P_2=\begin{pmatrix}
\varepsilon_{i_1} & \varepsilon_{i_2} & \cdots & \varepsilon_{i_n}
\end{pmatrix},
\end{equation*}
其中\((i_1,i_2,\ldots,i_n)\)是\((1,2,\ldots,n)\)的一个排列,则
\begin{equation*}
P_1P_2=\begin{pmatrix}
P_1\varepsilon_{i_1} & P_1\varepsilon_{i_2} & \cdots & P_1\varepsilon_{i_n}
\end{pmatrix}
\end{equation*}
的列向量是\(P_1\)列向量的重新排列。由于\(P_1\)是置换矩阵,其列向量是\(\varepsilon_1,\ldots,\varepsilon_n\)的一个排列,所以\(P_1P_2\)列向量也是\(\varepsilon_1,\ldots,\varepsilon_n\)的排列,由此可知\(P_1P_2\)也是置换矩阵。