主要内容

高等代数教学辅导

2.3 向量组的秩

建设中!

子节 2.3.1 主要知识点

定义 2.3.1.

\(\mathbb{F}^n\)中有向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)和向量组\(\beta_1,\beta_2,\cdots ,\beta_t\)。 若向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)中的每个向量都可由\(\beta_1,\beta_2,\cdots ,\beta_t\)线性表出,则称向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)可由向量组\(\beta_1,\beta_2,\cdots ,\beta_t\)线性表出;如果向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)和向量组\(\beta_1,\beta_2,\cdots ,\beta_t\)可以互相线性表出, 则称向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)\(\beta_1,\beta_2,\cdots ,\beta_t\)等价。
  • 例:\({\alpha _1} = {(1,0,\cdots,0)^T},{\alpha _2} = {(1,1,0,\cdots,0)^T},\cdots,{\alpha _n} = {(1,1,\cdots,1)^T}\)\(\varepsilon_1,\varepsilon_2,\cdots ,\varepsilon_n\)
向量组等价关系满足:
  1. 反身性:\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)等价;
  2. 对称性:若\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)\(\beta_1,\beta_2,\cdots ,\beta_t\)等价,则\(\beta_1,\beta_2,\cdots ,\beta_t\)\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)等价;
  3. 传递性: 若\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)\(\beta_1,\beta_2,\cdots ,\beta_t\)等价,\(\beta_1,\beta_2,\cdots ,\beta_t\)\(\gamma_1,\gamma_2,\cdots ,\gamma_r\)等价,则\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)\(\gamma_1,\gamma_2,\cdots ,\gamma_r\)等价。

定义 2.3.6.

若向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)的部分组\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_r}\)满足:
  1. \(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_r}\)线性无关;
  2. 将向量组中任意向量添加到\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_r}\)得到的\(r +1\)个向量线性相关,
则称\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_r}\)是向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)的一个极大线性无关组,简称为极大无关组。
  • 等价定义:若向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)中任意向量可以由部分组\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_r}\)线性表出, 且表示法唯一,则称\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_r}\)是向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)的一个极大线性无关组。

2.3.7.

  1. 零向量组成的向量组没有极大线性无关组。
  2. 若向量组\(\alpha_1,\alpha_2,\cdots,\alpha_s\)线性无关,则\(\alpha_1,\alpha_2,\cdots,\alpha_s\)的极大无关组就是其自身。
  3. \(\mathbb{R}^2\)中,\(\alpha_1=(1,0)^T\)\(\alpha_2=(0,1)^T\)\(\alpha_3=(1,1)^T\), 则
    • \(\alpha_1,\alpha_2\)是向量组\(\alpha_1,\alpha_2,\alpha_3\)的一个极大无关组;
    • \(\alpha_1,\alpha_3\)是向量组\(\alpha_1,\alpha_2,\alpha_3\)的一个极大无关组;
    • \(\alpha_2,\alpha_3\)是向量组\(\alpha_1,\alpha_2,\alpha_3\)的一个极大无关组。

备注 2.3.8.

  1. 任意一个非零向量组必存在极大线性无关组。
  2. 一个向量组的极大线性无关组未必唯一。
  • 问题:同一向量组的不同极大无关组有哪些共同点?

定义 2.3.12.

向量组\(\alpha_1,\alpha_2,\cdots,\alpha_s\)的极大无关组所含向量个数称为该向量组的秩,记为\(r(\alpha_1,\alpha_2,\cdots,\alpha_s)\)
  • 约定零向量组成的向量组的秩为零。
  • \(P\)可逆, 则列向量组\(\alpha_1,\alpha_2,\cdots,\alpha_s\)\(P\alpha_1,P\alpha_2,\cdots,P\alpha_s\)有相同的线性关系。
  • \(A=(\alpha_1,\alpha_2,\cdots,\alpha_s)\)经行初等变换化为\(B=(\beta_1,\beta_2,\cdots,\beta_s)\),则\(\alpha_{i_1},\alpha_{i_2},\cdots,\alpha_{i_r}\)是向量组\(\alpha_1,\alpha_2,\cdots,\alpha_s\)的极大线性无关组的充要条件是\(\beta_{i_1},\beta_{i_2},\cdots,\beta_{i_r}\)是向量组\(\beta_1,\beta_2,\cdots,\beta_s\)的极大线性无关组;
  • \(r(\alpha_1,\alpha_2,\cdots,\alpha_s)=r(\beta_1,\beta_2,\cdots,\beta_s)\)
求列向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)的秩与极大无关组的一般步骤:
  • 第一步:作矩阵\(A=(\alpha_1,\alpha_2,\cdots,\alpha_s)\)
  • 第二步:用初等行变换化矩阵\(A\)为阶梯阵\(J\)
  • \(J\)中有\(r\)个非零行,则\(r(\alpha_1,\alpha_2,\cdots,\alpha_s)=r\)
  • \(J\)中第\(i\)个非零行第一个非零元所在列标号为\(j_i,i=1,2,\cdots,r\),则\(\alpha_{j_1},\alpha_{j_2},\cdots,\alpha_{j_r}\)是向量组\(\alpha_1,\alpha_2,\cdots,\alpha_s\)的一个极大线性无关组。

2.3.13.

\({\alpha _1} = {\left( {2,1,3,0,4} \right)^T},{\alpha _2} = {\left( { - 1,2,3,1,0} \right)^T},{\alpha _3} = {\left( {3, - 1,0, - 1,4} \right)^T}\)
  1. \({\alpha _1},{\alpha _2},{\alpha _3}\)的一个极大线性无关组;
  2. \(r({\alpha _1},{\alpha _2},{\alpha _3})\)
  3. 将其余的向量表为该极大无关组的线性组合。
  • 问题:推论的逆命题成立吗?即若\(r(\alpha_1,\alpha_2,\cdots,\alpha_s)= r(\beta_1,\beta_2,\cdots ,\beta_t)\) ,则\(\alpha_1,\alpha_2,\cdots,\alpha_s\)\(\beta_1,\beta_2,\cdots ,\beta_t\)是否等价?

备注 2.3.16.

  1. 向量组\(\alpha_1,\alpha_2,\cdots,\alpha_s\)可以由\(\beta_1,\beta_2,\cdots ,\beta_t\)线性表出的充要条件是
    \begin{equation*} r(\beta_1,\beta_2,\cdots ,\beta_t)=r(\alpha_1,\alpha_2,\cdots,\alpha_s,\beta_1,\beta_2,\cdots ,\beta_t). \end{equation*}
  2. 向量组\(\alpha_1,\alpha_2,\cdots,\alpha_s\)\(\beta_1,\beta_2,\cdots ,\beta_t\)等价的充要条件是
    \begin{equation*} r(\beta_1,\beta_2,\cdots ,\beta_t)=r(\alpha_1,\cdots,\alpha_s,\beta_1,\cdots ,\beta_t)=r(\alpha_1,\alpha_2,\cdots,\alpha_s). \end{equation*}

2.3.17.

已知向量组
\begin{equation*} I:{\alpha _1} = {\left( {0,1,1} \right)^T},{\alpha _2} = {\left( {1,1,0} \right)^T}; \end{equation*}
\begin{equation*} II:{\beta _1} = {\left( { - 1,0,1} \right)^T},{\beta _2} = {\left( {1,2,1} \right)^T},{\beta _3} = {\left( {3,2, - 1} \right)^T}. \end{equation*}
求证:向量组I和II等价。

定义 2.3.18.

\(m\times n\)矩阵\(A\)的行向量组的秩称为\(A\)的行秩,\(A\)的列向量组的秩称为\(A\)的列秩。

2.3.22.

\(A\in \mathbb{F}^{m\times n}\),证明:\(r(A)\leq\min\{m,n\}\)

2.3.23.

证明:\(\max\{r(A),r(B)\}\leq r(A,B)\leq r(A)+r(B)\)
  • 矩阵增加一列(行), 则秩不变或加一; 若矩阵减去一列(行), 则秩不变或减一。

2.3.24.

证明:\(r\left(\begin{array}{cc} A&0\\0&B \end{array}\right)=r(A)+r(B);\quad r\left(\begin{array}{cc} A&0\\ C &B \end{array}\right) \geq r(A)+r(B)\)

2.3.25.

证明:\(r(AB)\leq \min\{r(A),r(B)\}\)

2.3.26.

证明:\(r(A+B)\leq r(A)+r(B)\)

2.3.27.

\(A\)\(n\)阶方阵,证明:\(r(A)+r(E_n-A)\geq n\)
  • 注:\(r(A)+r(E_n-A)=n\)的充要条件是\(A^2=A\)

练习 2.3.2 练习

1.

设向量\(\beta\)可以由\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)线性表示,但不能由\(\alpha_1,\alpha_2,\cdots ,\alpha_{s-1}\)线性表示。证明:向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_{s-1},\alpha_s\)与向量组 \(\alpha_1,\alpha_2,\cdots ,\alpha_{s-1},\beta\)等价。
解答.
因为\(\beta\)可以由\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)线性表示,所以存在\(a_1,a_2,\cdots ,a_s\in\mathbb{F}\),使得
\begin{equation*} \beta=a_1\alpha_1+a_2\alpha_s+\cdots +a_s\alpha_s. \end{equation*}
\(a_s=0\),则\(\beta=a_1\alpha_1+a_2\alpha_s+\cdots +a_{s-1}\alpha_{s-1}\),与条件\(\beta\)不能由向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_{s-1}\)线性表出相矛盾。故\(a_s\neq 0\)。于是
\begin{equation*} \alpha_s=-\frac{a_1}{a_s}\alpha_1-\frac{a_2}{a_s}\alpha_2-\cdots -\frac{a_{s-1}}{a_{s}}\alpha_{s-1}+\frac{1}{a_{s-1}}\beta, \end{equation*}
\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)可由向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_{s-1},\beta\)线性表出。又\(\beta\)可以由向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)线性表出,故\(\alpha_1,\alpha_2,\cdots ,\alpha_{s-1},\beta\)可由向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)线性表出。从而,向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_{s-1},\alpha_s\)与向量组 \(\alpha_1,\alpha_2,\cdots ,\alpha_{s-1},\beta\)等价。

2.

设向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_r\)与向量组\(\beta_1,\beta_2,\cdots ,\beta_s\)等价,且\(\alpha_1,\alpha_2,\cdots ,\alpha_r\)线性无关,试问\(\beta_1,\beta_2,\cdots ,\beta_s\)是否一定线性无关?如果结论成立,证明;如果不成立,举出反例。
解答.
不一定成立。比如,向量组
\begin{equation*} \alpha_1=(1,0)^T,\alpha_2=(0,1)^T \end{equation*}
与向量组
\begin{equation*} \beta_1=(1,0)^T,\beta_2=(0,1)^T,\beta_3=(1,1)^T \end{equation*}
等价,\(\alpha_1,\alpha_2\)线性无关,但\(\beta_1,\beta_2,\beta_3\)线性相关。

3.

证明:一个向量组的任何一个线性无关组都可以扩充成一极大无关组,即若\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_p}\)\(\alpha_1,\alpha_2,\cdots,\alpha_s\)中一个线性无关向量组,那么向量组\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_p}\)一定可以扩充为\(\alpha_1,\alpha_2,\cdots,\alpha_s\)的一个极大线性无关组\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_p},\alpha_{i_{p+1}},\cdots ,\alpha_{i_r}\)
解答.
  1. \(\alpha_1,\alpha_2,\cdots,\alpha_s\)中每个向量都可由向量组\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_p}\)线性表出,则\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_p}\)已是向量组\(\alpha_1,\alpha_2,\cdots,\alpha_s\)的一个极大线性无关组。
  2. 若存在\(1\leq i_{p+1}\leq s\)使得\(\alpha_{i_{p+1}}\)不能由\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_p}\)线性表出,则向量组\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_p},\alpha_{i_{p+1}}\)线性无关。
    1. \(\alpha_1,\alpha_2,\cdots,\alpha_s\)中每个向量都可由向量组\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_{p+1}}\)线性表出,则\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_{p+1}}\)是向量组\(\alpha_1,\alpha_2,\cdots,\alpha_s\)的一个极大线性无关组。
    2. 否则,存在\(1\leq i_{p+2}\leq s\)使得\(\alpha_{i_{p+2}}\)不能由\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_{p+1}}\)线性表出,则向量组\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_p},\alpha_{i_{p+2}}\)线性无关。
继续以上过程,总可找到一个包含\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_p}\)的线性无关组,使得\(\alpha_1,\alpha_2,\cdots,\alpha_s\)中每个向量都可由它线性表出,即找到包含\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_p}\)的一个极大线性无关组\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_p},\alpha_{i_{p+1}},\cdots ,\alpha_{i_r}\)

4.

\begin{equation*} \alpha_1=(1,0,2,-1)^T,\alpha_2=(-2,1,-4,6)^T,\alpha_3=(3,2,7,5)^T,\alpha_4=(1,-2,6,-9)^T, \end{equation*}
  1. 证明:\(\alpha_1,\alpha_2\)线性无关;
  2. \(r(\alpha_1,\alpha_2,\alpha_3,\alpha_4)\)
  3. \(\alpha_1,\alpha_2\) 扩充为\(\alpha_1,\alpha_2,\alpha_3,\alpha_4\)的一个极大无关组,并将其余向量表示 为这个极大无关组的线性组合。
解答.
  1. 因为\(\alpha_1,\alpha_2\)不成比例,所以\(\alpha_1,\alpha_2\)线性无关。
  2. \(A=(\alpha_1,\alpha_2,\alpha_3,\alpha_4)\),对\(A\)作行初等变换
    \begin{equation*} A=\begin{pmatrix} 1&-2&3&1\\ 0&1&2&-2\\ 2&-4&7&6\\ -1&6&5&-9 \end{pmatrix}\rightarrow\begin{pmatrix} 1&0&0&-31\\ 0&1&0&-10\\ 0&0&1&4\\ 0&0&0&0 \end{pmatrix}=B, \end{equation*}
    所以\(r(\alpha_1,\alpha_2,\alpha_3,\alpha_4)=3\)
  3. \(B=(\beta_1,\beta_2,\beta_3,\beta_4)\)。因为行初等变换不改变列向量组的线性关系,\(\beta_1,\beta_2,\beta_3\)(或\(\beta_1,\beta_2,\beta_4\))是\(\beta_1,\beta_2,\beta_3,\beta_4\)的一个极大线性无关组,所以\(\alpha_1,\alpha_2,\alpha_3\)(或\(\alpha_1,\alpha_2,\alpha_4\))是\(\alpha_1,\alpha_2,\alpha_3,\alpha_4\)的一个极大线性无关组。因此\(\alpha_1,\alpha_2\) 扩充为向量组\(\alpha_1,\alpha_2,\alpha_3,\alpha_4\)的一个极大无关组\(\alpha_1,\alpha_2,\alpha_3\)(或\(\alpha_1,\alpha_2,\alpha_4\))。

5.

设向量\(\beta\)可以由\(\alpha_1,\alpha_2,\cdots ,\alpha_s\)线性表示,但不能由\(\alpha_1,\alpha_2,\cdots ,\alpha_{s-1}\)线性表示。证明:\(r(\alpha_1,\alpha_2,\cdots ,\alpha_{s-1},\alpha_s)=r(\alpha_1,\alpha_2,\cdots ,\alpha_{s-1},\beta)\)
解答.
由作业 练习 2.3.2.1 知:\(\alpha_1,\alpha_2,\cdots ,\alpha_{s-1},\alpha_s\)\(\alpha_1,\alpha_2,\cdots ,\alpha_{s-1},\beta\)等价,因此
\begin{equation*} r(\alpha_1,\alpha_2,\cdots ,\alpha_{s-1},\alpha_s)=r(\alpha_1,\alpha_2,\cdots ,\alpha_{s-1},\beta). \end{equation*}

6.

\(\alpha_1,\alpha_2,\cdots ,\alpha_n\)是一组\(n\)维向量,已知单位向量\(\varepsilon_1,\varepsilon_2,\cdots ,\varepsilon_n\)可被它们线性表示,证明:\(\alpha_1,\alpha_2,\cdots ,\alpha_n\)线性无关。
解答.
因为\(n\)维列向量\(\alpha_1,\alpha_2,\cdots ,\alpha_n\)必可由\(n\)维标准单位列向量\(\varepsilon_1,\varepsilon_2,\cdots ,\varepsilon_n\)线性表出,而由题设\(\varepsilon_1,\varepsilon_2,\cdots ,\varepsilon_n\)可由\(\alpha_1,\alpha_2,\cdots ,\alpha_n\)线性表出,因而\(\alpha_1,\alpha_2,\cdots ,\alpha_n\)\(\varepsilon_1,\varepsilon_2,\cdots ,\varepsilon_n\)等价。由此得
\begin{equation*} r(\alpha_1,\alpha_2,\cdots ,\alpha_n)=r(\varepsilon_1,\varepsilon_2,\cdots ,\varepsilon_n). \end{equation*}
注意到\(r(\varepsilon_1,\varepsilon_2,\cdots ,\varepsilon_n)=n\),故 \(r(\alpha_1,\alpha_2,\cdots ,\alpha_n)=n\)。从而\(\alpha_1,\alpha_2,\cdots ,\alpha_n\)线性无关。

7.

\(\alpha_1,\alpha_2,\cdots ,\alpha_n\)是一组\(n\)维向量,证明:\(\alpha_1,\alpha_2,\cdots ,\alpha_n\)线性无关的充要条件是任一\(n\)维向量都可被它们线性表示。
解答.
充分性:由题设,\(n\)维标准单位列向量\(\varepsilon_1,\varepsilon_2,\cdots ,\varepsilon_n\)可由向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_n\)线性表出,由上题知,\(\alpha_1,\alpha_2,\cdots ,\alpha_n\)线性无关。
必要性:设\(\beta\)为任一\(n\)维列向量,则\(\alpha_1,\alpha_2,\cdots ,\alpha_s,\beta\)\(n+1\)\(n\)维列向量,必线性相关。而\(\alpha_1,\alpha_2,\cdots ,\alpha_n\)线性无关,故\(\beta\)可由向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_n\)线性表出。

8.

\(\alpha_1,\alpha_2,\cdots,\alpha_s\)的秩为\(r\),证明:\(\alpha_1,\alpha_2,\cdots,\alpha_s\)中任意\(r\)个线性无关的向量都构成它的一极大无关组。
解答.
\(\alpha_{i_1},\alpha_{i_2},\cdots,\alpha_{i_r}\)\(\alpha_1,\alpha_2,\cdots,\alpha_s\)中任意\(r\)个线性无关向量,则\(\forall 1\leq j\leq s\),向量组\(\alpha_{i_1},\alpha_{i_2},\cdots,\alpha_{i_r},\alpha_j\)必线性相关(否则\(\alpha_1,\alpha_2,\cdots,\alpha_s\)的秩不小于\(r+1\)。)因此,\(\alpha_{i_1},\alpha_{i_2},\cdots,\alpha_{i_r}\)\(\alpha_1,\alpha_2,\cdots,\alpha_s\)的一个极大无关组。

9.

\(\alpha_1,\alpha_2,\cdots,\alpha_s\)的秩为\(r\)\(\alpha_{i_1},\alpha_{i_2},\cdots,\alpha_{i_r}\)\(\alpha_1,\alpha_2,\cdots,\alpha_s\)\(r\)个向量,使得\(\alpha_1,\alpha_2,\cdots,\alpha_s\)中每个向量都可被它们线性表示,证明:\(\alpha_{i_1},\alpha_{i_2},\cdots,\alpha_{i_r}\)\(\alpha_1,\alpha_2,\cdots,\alpha_s\)的一个极大无关组。
解答.
由于向量组\(\alpha_1,\alpha_2,\cdots,\alpha_s\)可由\(\alpha_{i_1},\alpha_{i_2},\cdots,\alpha_{i_r}\)线性表出,而显然\(\alpha_{i_1},\alpha_{i_2},\cdots,\alpha_{i_r}\)可由\(\alpha_1,\alpha_2,\cdots,\alpha_s\)线性表出,所以向量组\(\alpha_1,\alpha_2,\cdots,\alpha_s\)与向量组\(\alpha_{i_1},\alpha_{i_2},\cdots,\alpha_{i_r}\)等价。从而\(r(\alpha_{i_1},\alpha_{i_2},\cdots,\alpha_{i_r})=r(\alpha_1,\alpha_2,\cdots,\alpha_s)=r \),则\(\alpha_{i_1},\alpha_{i_2},\cdots,\alpha_{i_r}\)线性无关。 故\(\alpha_{i_1},\alpha_{i_2},\cdots,\alpha_{i_r}\)\(\alpha_1,\alpha_2,\cdots,\alpha_s\)的一个极大无关组。

10.

已知两个向量组\(\alpha_1=(1,0,2)^T,\alpha_2=(1,1,3)^T,\alpha_3=(1,-1,a+2)^T\)\(\beta_1=(1,2,a+3)^T,\beta_2=(2,1,a+6)^T,\beta_3=(2,1,a+4)^T\),问\(a\)为何值时,两个向量组等价;当\(a\)为何值时,两向量组不等价。
解答.
\begin{equation*} (\alpha_1,\alpha_2,\alpha_3,\beta_1,\beta_2,\beta_3)\rightarrow\left(\begin{array}{cccccc} 1&1&1&1&2&2\\0&1&-1&2&1&1\\0&0&a+1&a-1&a+1&a-1 \end{array}\right) \end{equation*}
\begin{equation*} (\beta_1,\beta_2,\beta_3)=\left(\begin{array}{ccc} 1&2&2\\2&1&1\\a+3&a+6&a+4 \end{array}\right)\rightarrow\left(\begin{array}{ccc} 1&2&2\\0&1&1\\0&0&-2 \end{array}\right) \end{equation*}
\(r(\beta_1,\beta_2,\beta_3)=3\)。故两个向量组等价\(\Leftrightarrow r(\alpha_1,\alpha_2,\alpha_3,\beta_1,\beta_2,\beta_3)=3\Leftrightarrow a\neq -1\);两个向量组不等价\(\Leftrightarrow a=-1\)

11.

证明:数域\(\mathbb{F}\)上的\(n\)个方程的\(n\)元线性方程组
\begin{equation*} x_1\alpha_1+x_2\alpha_2+\cdots +x_n\alpha_n=\beta \end{equation*}
对任意\(\beta\in\mathbb{F}^n\)都有解的充分必要条件是\(\det (\alpha_1,\alpha_2,\cdots ,\alpha_n)\neq 0\)
解答.
线性方程组\(x_1\alpha_1+x_2\alpha_2+\cdots +x_n\alpha_n=\beta\)对任意\(\beta\in\mathbb{F}^n\)都有解当且仅当任一\(n\)维向量\(\beta\)都可由向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_n\)线性表出。由作业 练习 2.3.2.7 可知,其充分必要条件为\(\alpha_1,\alpha_2,\cdots ,\alpha_n\)线性无关,即\(\det (\alpha_1,\alpha_2,\cdots ,\alpha_n)\neq 0\)

12.

\(A\)是秩为\(r\)\(m\times n\)矩阵,从\(A\)中任意取\(s\)行作一个\(s\times n\)矩阵\(B\)。证明:\(r(B)\geq r+s-m\)
解答.
\(A\)的行向量组为\(\alpha_1,\alpha_2,\cdots ,\alpha_m\)\(B\)的行向量组为\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_s}\)。若\(r(B)=t\),则向量组\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_s}\)的秩为\(t\)。因为\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_s}\)的极大无关组为向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_m\)的线性无关组,故可将\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_s}\)的一个极大无关组(含\(t\)个向量)扩充为\(\alpha_1,\alpha_2,\cdots ,\alpha_m\)的一个极大线性无关组(扩充了\(r-t\)个向量)。注意到上述扩充过程中,扩充的向量均取自\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_s}\)以外的向量,而\(\alpha_1,\alpha_2,\cdots ,\alpha_m\)中除\(\alpha_{i_1},\alpha_{i_2},\cdots ,\alpha_{i_s}\)外的向量个数为\(m-s\),故\(r-t\leq m-s\)。因此\(r(B)=t\geq r+s-m\)

13.

\(A\)是秩为\(r\)\(m\times n\)矩阵,从\(A\)中任意划去\(m-s\)行与\(n-t\)列,其余元素按原来位置排成一个\(s\times t\)矩阵\(C\)。证明:\(r(C)\geq r+s+t-m-n\)
解答.
\(A\)划去\(m-s\)行后的矩阵为\(B\),则\(B\)划去\(n-t\)列后得到矩阵\(C\)。由上题知,\(r(B)\geq r+s-m\)。同理,\(r(C)\geq r(B)+t-n\)。因此
\begin{equation*} r(C)\geq r+s+t-m-n. \end{equation*}

14.

\(n\)维实列向量\(\alpha_1,\alpha_2,\cdots ,\alpha_r\)线性无关,其中\(\alpha_i=(a_{i1},a_{i2},\cdots ,a_{in})^T,i=1,2,\cdots ,r\)。已知\(\beta=(b_1,b_2,\cdots ,b_n)^T\)是齐次线性方程组
\begin{equation*} \left\{\begin{array}{c} 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,\\ \vdots\\ a_{r1}x_1+a_{r2}x_2+\cdots +a_{rn}x_n=0 \end{array}\right. \end{equation*}
的一个非零实解向量。试判断向量组\(\alpha_1,\alpha_2,\cdots ,\alpha_r,\beta\)的线性相关性。
解答.
\(A=\begin{pmatrix} \alpha_1^T\\\alpha_2^T\\\vdots\\\alpha_r^T \end{pmatrix}\),由题设\(\beta\)是齐次线性方程组\(AX=0\)的解,即\(A\beta=0\),得对任意\(1\leq i\leq r\),有\(\alpha_i^T\beta=0\)
假设
\begin{equation} a_1\alpha_1+a_2\alpha_2+\cdots +a_r\alpha_r+a_{r+1}\beta=0,\tag{2.5} \end{equation}
\begin{equation*} a_1\alpha_1^T+a_2\alpha_2^T+\cdots +a_r\alpha_r^T+a_{r+1}\beta^T=0, \end{equation*}
两边同时右乘\(\beta\)
\begin{equation*} a_1(\alpha_1^T\beta)+a_2(\alpha_2^T\beta)+\cdots +a_r(\alpha_r^T\beta)+a_{r+1}(\beta^T\beta)=0, \end{equation*}
\(a_{r+1}(\beta^T\beta)=0\)。注意到\(\beta=(b_1,b_2,\cdots ,b_n)^T\)为非零向量,
\begin{equation*} \beta^T\beta=b_1^2+b_2^2+\cdots+b_n^2\neq 0, \end{equation*}
\(a_{r+1}=0\)。代入 (2.5) ,得
\begin{equation*} a_1\alpha_1+a_2\alpha_2+\cdots +a_r\alpha_r=0, \end{equation*}
\(\alpha_1,\alpha_2,\cdots ,\alpha_r\)线性无关知\(a_1=a_2=\cdots =a_r=0\)。因此\(\alpha_1,\alpha_2,\cdots ,\alpha_r,\beta\)线性无关。

15.

\(A\)\(n\)阶方阵,证明:\(A^2=E\)的充要条件是\(r(A+E)+r(A-E)=n\)
解答.
作分块矩阵的初等变换,
\begin{equation*} \begin{pmatrix} A+E_n&0\\0&A-E_n \end{pmatrix}\rightarrow\begin{pmatrix} A+E_n&A-E_n\\0&A-E_n \end{pmatrix}\rightarrow\begin{pmatrix} 2E_n&A-E_n\\E_n-A&A-E_n \end{pmatrix}\rightarrow\begin{pmatrix} 2E_n&0\\0&\frac{1}{2}(A^2-E_n) \end{pmatrix}, \end{equation*}
\begin{equation*} r(A+E_n)+r(A-E_n)=n+r(A^2-E). \end{equation*}
因此\(A^2=E\)的充要条件是\(r(A+E)+r(A-E)=n\)

16.

\(A\)\(m\times n\)矩阵,\(B\)\(n\times s\)矩阵。证明:\(r(AB)\geq r(A)+r(B)-n\)
解答.
证法一:作分块矩阵的初等变换,
\begin{equation*} \begin{pmatrix} E_n&0\\0&AB \end{pmatrix}\rightarrow\begin{pmatrix} E_n&0\\A&AB \end{pmatrix}\rightarrow\begin{pmatrix} E_n&-B\\A&0 \end{pmatrix}\rightarrow\begin{pmatrix} A&0\\E_n&-B \end{pmatrix}\rightarrow\begin{pmatrix} A&0\\-E_n&B \end{pmatrix}, \end{equation*}
所以
\begin{equation*} r(E_n)+r(AB)=r(\begin{pmatrix} A&0\\-E_n&B \end{pmatrix}), \end{equation*}
\begin{equation*} r(AB)=r(\begin{pmatrix} A&0\\-E_n&B \end{pmatrix})-n. \end{equation*}
注意到\(r(\begin{pmatrix} A&0\\-E_n&B \end{pmatrix})\geq r(A)+r(B)\),故
\begin{equation*} r(AB)\geq r(A)+r(B)-n. \end{equation*}
证法二:设\(r(A)=r\),则存在\(m\)阶可逆矩阵\(P\)\(n\)阶可逆矩阵\(Q\)使得
\begin{equation*} A=P \begin{pmatrix} E_r&0\\0&0 \end{pmatrix}Q. \end{equation*}
所以
\begin{equation*} r(AB)=r(P\begin{pmatrix} E_r&0\\0&0 \end{pmatrix}QB)=r(\begin{pmatrix} E_r&0\\0&0 \end{pmatrix}QB). \end{equation*}
\(QB=\begin{pmatrix} B_1\\B_2{} \end{pmatrix}\),其中\(B_1\in\mathbb{F}^{r\times s},B_2\in\mathbb{F}^{(n-r)\times s}\),则\(\begin{pmatrix} E_r&0\\0&0 \end{pmatrix}QB= \begin{pmatrix} B_1\\0 \end{pmatrix}\)。从而
\begin{equation*} r(AB)=r(\begin{pmatrix} B_1\\0 \end{pmatrix})=r(B_1). \end{equation*}
由作业 练习 2.3.2.12 知:\(r(B_1)\geq r(QB)+r-n\),故\(r(AB)\geq r(A)+r(B)-n\)
证法三: 记\(S = \{X|ABX = 0\}\),则\(r(S) = s - r(AB) \)
\(S_1 =\{X|BX= 0 \}\),取\(S_1\)的一个极大无关组为\(\eta_1,\ldots,\eta_k\),其中\(k = s - r(B)\)
注意到\(S\)也可以表示为\(S = \{X| BX=Y,\ AY=0\}\)。记 \(T = \{Y|AY = 0, \mbox{ 且}BX=Y\mbox{有解}\} \),取\(T\)的一个极大无关组为\(Y_1,\ldots, Y_t\)。因为 \(T\subseteq \{Y|AY=0\}\),所以\(t = r(T)\leq n - r(A)\)
\(\gamma_i\)为非齐次方程组\(BX = Y_i\)的一个特解,\(i=1,\ldots,t\)。下证向量组\(S\)可以由向量组\(\gamma_1,\ldots, \gamma_t,\eta_1,\ldots,\eta_k\)线性表出。任取\(X\in S\)。若\(BX=0\),根据齐次线性方程组解的结构定理,\(X\)可以由\(\eta_1,\ldots,\eta_k\)线性表出;若\(BX=Y_0\ne 0\),记\(Y_0 = \sum_{j=1}^t c_jY_j\),则\(B(\sum_{j=1}^t c_j \gamma_j) = Y_0\),即\(\sum_{j=1}^t c_j \gamma_j\)\(BX=Y_0\)的一个特解,根据非齐次线性方程组解的结构定理,\(X\)可以由\(\gamma_1,\ldots, \gamma_t,\eta_1,\ldots,\eta_k\)线性表出。综上,\(T\)可以由向量组\(\gamma_1,\ldots, \gamma_t,\eta_1,\ldots,\eta_k\)线性表出。
于是,我们有
\begin{equation*} s- r(AB) = r(S)\le r(\gamma_1,\ldots, \gamma_t,\eta_1,\ldots,\eta_k)\le t+k\le s-r(B)+n-r(A), \end{equation*}
整理可得\(r(AB)\geq r(A)+r(B)-n\)

17.

证明Frobenius不等式:\(r(ABC)\geq r(AB)+r(BC)-r(B)\)
解答.
证法一:作分块矩阵的初等变换,
\begin{equation*} \begin{pmatrix} ABC&0\\0&B \end{pmatrix}\rightarrow\begin{pmatrix} ABC&AB\\0&B \end{pmatrix}\rightarrow\begin{pmatrix} 0&AB\\-BC&B \end{pmatrix}\rightarrow\begin{pmatrix} -BC&B\\0&AB \end{pmatrix}\rightarrow\begin{pmatrix} BC&-B\\0&AB \end{pmatrix}, \end{equation*}
所以
\begin{equation*} r(ABC)+r(B)=r(\begin{pmatrix} BC&-B\\0&AB \end{pmatrix}), \end{equation*}
\begin{equation*} r(ABC)=r(\begin{pmatrix} BC&-B\\0&AB \end{pmatrix})-r(B). \end{equation*}
注意到\(r(\begin{pmatrix} BC&-B\\0&AB \end{pmatrix})\geq r(AB)+r(BC)\),故
\begin{equation*} r(ABC)\geq r(AB)+r(BC)-r(B). \end{equation*}
证法二:设\(A\in\mathbb{F}^{m\times n},B\in\mathbb{F}^{n\times s},C\in\mathbb{F}^{s\times t}\)\(r(B)=r\),则存在\(n\)阶可逆矩阵\(P\)\(s\)阶可逆矩阵\(Q\),使得\(B=P \begin{pmatrix} E_r&0\\0&0 \end{pmatrix}Q\)。所以
\begin{equation*} r(ABC)=r(AP\begin{pmatrix} E_r&0\\0&0 \end{pmatrix}QC)=r([AP \begin{pmatrix} E_r\\0 \end{pmatrix}]\cdot [\begin{pmatrix} E_r&0 \end{pmatrix}QC]), \end{equation*}
这里\(AP \begin{pmatrix} E_r\\0 \end{pmatrix}\)\(m\times r\)矩阵,\(\begin{pmatrix} E_r&0 \end{pmatrix}QC\)\(r\times t\)矩阵。根据上题结论,
\begin{equation*} r(ABC)\geq r(AP \begin{pmatrix} E_r\\0 \end{pmatrix})+r(\begin{pmatrix} E_r&0 \end{pmatrix}QC)-r. \end{equation*}
注意到
\begin{equation*} r(AP \begin{pmatrix} E_r\\0 \end{pmatrix})=r(AP \begin{pmatrix} E_r&0\\0&0 \end{pmatrix})=r(AP \begin{pmatrix} E_r&0\\0&0 \end{pmatrix}Q)=r(AB), \end{equation*}
\begin{equation*} r(\begin{pmatrix} E_r&0 \end{pmatrix}QC)=r(\begin{pmatrix} E_r&0\\0&0 \end{pmatrix}QC)=r(P\begin{pmatrix} E_r&0\\0&0 \end{pmatrix}QC)=r(BC), \end{equation*}
\(r(ABC)\geq r(AB)+r(BC)-r(B)\)