高等代数

高等代数理论基础23:矩阵的秩

2019-01-04  本文已影响63人  溺于恐

矩阵的秩

矩阵的秩

定义:矩阵的行向量组的秩称为矩阵的行秩,矩阵的列向量组的秩称为矩阵的列秩

引理:若齐次线性方程组\begin{cases}a_{11}x_1+a_{12}x_2+\cdots+a_{1n}x_n=0\\ a_{21}x_1+a_{22}x_2+\cdots+a_{2n}x_n=0\\ \cdots\\ a_{s1}x_1+a_{s2}x_2+\cdots+a_{sn}x_n=0\end{cases}\qquad (1)的系数矩阵

A=\begin{pmatrix}a_{11}&a_{12}&\cdots&a_{1n}\\ a_{21}&a_{22}&\cdots&a_{2n}\\ \vdots&\vdots& &\vdots\\ a_{s1}&a_{s2}&\cdots&a_{sn}\end{pmatrix}​的行秩r\lt n​则它有非零解

证明:

以\alpha_1,\alpha_2,\cdots,\alpha_s代表矩阵A的行向量组

\because 秩为r

\therefore 极大线性无关组由r个向量组成

不妨设\alpha_1,\cdots,\alpha_r为一个极大线性无关组

\alpha_1,\cdots,\alpha_r,\cdots,\alpha_s与\alpha_1,\cdots,\alpha_r等价

即方程组(1)(2)可互相线性表出

\begin{cases}a_{11}x_1+a_{12}x_2+\cdots+a_{1n}x_n=0\\ a_{21}x_1+a_{22}x_2+\cdots+a_{2n}x_n=0\\ \cdots\\ a_{r1}x_1+a_{r2}x_2+\cdots+a_{rn}x_n=0\end{cases}\qquad (2)

\therefore 方程组(1)(2)同解

\therefore 方程组(1)有非零解\qquad \mathcal{Q.E.D}

定理:矩阵的行秩与列秩相等

证明:

设矩阵A=\begin{pmatrix}a_{11}&a_{12}&\cdots&a_{1n}\\ a_{21}&a_{22}&\cdots&a_{2n}\\ \vdots&\vdots& &\vdots\\ a_{s1}&a_{s2}&\cdots&a_{sn}\end{pmatrix}

设A的行秩为r,列秩为r_1

下证r=r_1

以\alpha_1,\alpha_2,\cdots,\alpha_s代表矩阵A的行向量组

不妨设\alpha_1,\cdots,\alpha_r是它的一个极大无关组

\because \alpha_1,\cdots,\alpha_r线性无关

\therefore x_1\alpha_1+\cdots+x_r\alpha_r=0只有零解

即齐次线性方程组

\begin{cases}a_{11}x_1+a_{21}x_2+\cdots+a_{r1}x_r=0\\ a_{12}x_1+a_{22}x_2+\cdots+a_{r2}x_r=0\\ \cdots\\ a_{1n}x_1+a_{2n}x_2+\cdots+a_{rn}x_r=0\end{cases}

只有零解

\therefore 系数矩阵

\begin{pmatrix}a_{11}&a_{21}&\cdots&a_{r1}\\ a_{12}&a_{22}&\cdots&a_{r2}\\ \vdots&\vdots& &\vdots\\ a_{1n}&a_{2n}&\cdots&a_{rn}\end{pmatrix}

的行秩\ge r

\therefore 在它的行向量中可找到r个线性无关的

如(a_{11},a_{21},\cdots,a_{r1}),(a_{12},a_{22},\cdots,a_{r2}),

\cdots,(a_{1r},a_{2r},\cdots,a_{rr})线性无关

在上述向量中添上几个分量可得

(a_{11},a_{21},\cdots,a_{r1},\cdots,a_{s1}),(a_{12},a_{22},\cdots,a_{r2},\cdots,a_{s2}),

\cdots,(a_{1r},a_{2r},\cdots,a_{rr},\cdots,a_{sr})线性无关

它们正好是矩阵A的r个列向量

由它们的线性无关性可知

矩阵A的列秩r_1\ge r

类似可证r\ge r_1

\therefore r=r_1,即行秩与列秩相等\qquad\mathcal{Q.E.D}

定理:n\times n矩阵A=\begin{pmatrix}a_{11}&a_{12}&\cdots&a_{1n}\\ a_{21}&a_{22}&\cdots&a_{2n}\\ \vdots&\vdots& &\vdots\\ a_{n1}&a_{n2}&\cdots&a_{nn}\end{pmatrix}的行列式为零\LeftrightarrowA的秩小于n

证明:

充分性

\because A的秩小于n

\therefore A的n个行向量线性相关

n=1时,A只有一个数,即只有一个一维向量

且为线性相关的向量组

即为零向量

\therefore |A|=|0|=0

n\gt 1时,A中有一行是其余各行的线性组合

这一行依次减去各行相应倍数,则该行全变成零

\therefore |A|=0

必要性

n=1时,由|A|=0可知

A仅有的一个元素为零

\therefore A的秩为零

n\gt 1时,假设结论对n-1级矩阵成立

对n级矩阵

以\alpha_1,\alpha_2,\cdots,\alpha_n代表A的行向量

对于A的第一列元素a_{11},a_{21},\cdots,a_{n1}

若全为零,则A的列向量组中含零向量

显然秩小于n

若这n个元素有一个不为零

不妨设a_{11}\neq 0

则从第二行到第n行减去第一行的适当倍数可得

|A|=\begin{vmatrix}a_{11}&a_{12}&\cdots&a_{1n}\\ 0&a'_{22}&\cdots&a'_{2n}\\ \vdots&\vdots& &\vdots\\ 0&a'_{n2}&\cdots&a'_{nn}\end{vmatrix}

=a_{11}\begin{vmatrix}a'_{22}&\cdots&a'_{2n}\\ \vdots& &\vdots\\ a'_{n2}&\cdots&a'_{nn}\end{vmatrix}

其中(0,a'_{i2},\cdots,a'_{in})=\alpha_i-{a_{i1}\over a_{11}}\alpha_1,i=2,\cdots,n

由|A|=0可知n-1级矩阵

\begin{pmatrix}a'_{22}&\cdots&a'_{2n}\\ \vdots& &\vdots\\ a'_{n2}&\cdots&a'_{nn}\end{pmatrix}

的行列式为零

\therefore 这个矩阵的行向量线性相关

\therefore \alpha_2-{a_{21}\over a_{11}}\alpha_1,\cdots,\alpha_n-{a_{n1}\over a_{11}}\alpha_1线性相关

即有不全为零的数k_2,\cdots,k_n使得

k_2(\alpha_2-{a_{21}\over a_{11}}\alpha_1)+\cdots+k_n(\alpha_n-{a_{n1}\over a_{11}}\alpha_1)=0

即-({a_{21}\over a_{11}}+\cdots+{a_{n1}\over a_{11}})\alpha_1+k_2\alpha_2+\cdots+k_n\alpha_n=0

-({a_{21}\over a_{11}}+\cdots+{a_{n1}\over a_{11}}),k_2,\cdots,k_n显然不全为零

\therefore 向量组\alpha_1,\alpha_2,\cdots,\alpha_n线性相关

\therefore 它的秩小于n

\therefore A的秩小于n\qquad \mathcal{Q.E.D}

推论:齐次线性方程组\begin{cases}a_{11}x_1+a_{12}x_2+\cdots+a_{1n}x_n=0\\ a_{21}x_1+a_{22}x_2+\cdots+a_{2n}x_n=0\\ \cdots\\ a_{n1}x_1+a_{n2}x_2+\cdots+a_{nn}x_n=0\end{cases}有非零解的充要条件是它的系数矩阵A=\begin{pmatrix}a_{11}&a_{12}&\cdots&a_{1n}\\ a_{21}&a_{22}&\cdots&a_{2n}\\ \vdots&\vdots& &\vdots\\ a_{s1}&a_{s2}&\cdots&a_{sn}\end{pmatrix}的行列式等于零

子式

定义:在一个s\times n矩阵A中任意选定k行和k列,位于这些选定的行和列的交点上的k^2个元素按原来的次序所组成的k级行列式称为A的一个k级子式

注:k\le min(s,n)

定理:一矩阵的秩是r的充要条件为矩阵中有一个r级子式不为零,同时所有r+1级子式全为零

证明:

必要性

设矩阵A的秩为r

则A中任意r+1个行向量线性相关

A的任意r+1级子式的行向量也线性相关

\therefore A的所有r+1级子式全为零

下证A中至少有一个r级子式不为零

A=\begin{pmatrix}a_{11}&a_{12}&\cdots&a_{1n}\\ a_{21}&a_{22}&\cdots&a_{2n}\\ \vdots&\vdots& &\vdots\\ a_{s1}&a_{s2}&\cdots&a_{sn}\end{pmatrix}的秩为r

\therefore A中有r个行向量线性无关

不妨设前r个行向量线性无关

取出这r行作一新矩阵

A_1=\begin{pmatrix}a_{11}&a_{12}&\cdots&a_{1n}\\ a_{21}&a_{22}&\cdots&a_{2n}\\ \vdots&\vdots& &\vdots\\ a_{r1}&a_{r2}&\cdots&a_{rn}\end{pmatrix}

显然矩阵A_1的行秩为r​

\therefore A的列秩也为r

即A_1中有r列向量线性无关​

不妨假设前r列线性无关

\therefore \begin{vmatrix}a_{11}&\cdots&a_{1r}\\ \vdots& &\vdots\\ a_{r1}&\cdots&a_{rr}\end{vmatrix}\neq 0

它为A的一个r级子式,必要性得证

充分性

设在矩阵A中有一r级子式不为零,而所有r+1级子式全为零

下证A的秩为r

显然,若A的r+1级子式全为零

则A的r+2级子式也一定为零
\therefore A的所有级数大于r的子式全为零

设A的秩为t

t\ge r

否则A的r级子式全为零

t\le r

否则A就要有一个t(\ge r+1)级子式不为零,矛盾

\therefore t=r\qquad\mathcal{Q.E.D}

注:

1.矩阵A的秩\ger的充要条件为A有一个r级子式不为零

2.矩阵A的秩\ler的充要条件为A的所有r+1级子式全为零

3.在秩为r的矩阵中,不为零的r级子式所在的行正是它行向量组的一个极大线性无关组,所在的列正是它列向量组的一个极大线性无关组

计算矩阵的秩

注:初等行变换初等列变换不改变矩阵的秩

阶梯形矩阵的秩就等于其中非零行的数目

证明:

设A为一阶梯形矩阵,不为零的行数是r

初等列变换不改变矩阵的秩

适当变换列的顺序,不妨设

A=\begin{pmatrix}a_{11}&a_{12}&\cdots&a_{1r}&\cdots&a_{1n}\\ 0&a_{22}&\cdots&a_{2r}&\cdots&a_{2n}\\ \vdots&\vdots& &\vdots& &\vdots\\ 0&0&\cdots&a_{rr}&\cdots&a_{rn}\\ 0&0&\cdots&0&\cdots&0\\ \vdots&\vdots& &\vdots& &\vdots\\ 0&0&\cdots&0&\cdots&0\end{pmatrix}

其中a_{ii}\neq 0,i=1,2,\cdots,r

显然A的左上角的r级子式

\begin{vmatrix}a_{11}&a_{12}&\cdots&a_{1r}\\ 0&a_{22}&\cdots&a_{2r}\\ \vdots&\vdots& &\vdots\\ 0&0&\cdots&a_{rr}\end{vmatrix}

=a_{11}a_{22}\cdots a_{rr}\neq 0

\therefore A的秩为r\qquad \mathcal{Q.E.D}

上一篇 下一篇

猜你喜欢

热点阅读