主要内容

高等代数教学辅导

7.2 特征矩阵

建设中!

子节 7.2.1 主要知识点

备注 7.2.5.

\(A\in\mathbb{F}^{n\times n}\),且\(A\)的法式为\({\rm diag} (1,\cdots ,1,d_1(\lambda),\cdots ,d_k(\lambda))\),则以下几点等价:
  1. \(k=n\)
  2. \(\deg d_i(\lambda)=1,(i=1,2,\cdots ,n)\)
  3. \(d_1(\lambda)=d_2(\lambda)=\cdots =d_n(\lambda)=\lambda -\lambda_0\)
  4. \(\lambda E-A\simeq \lambda E-\lambda_0E\)
  5. \(A\)相似于数量矩阵\(\lambda_0E\)
  6. \(A=\lambda_0E\)

练习 7.2.2 练习

1.

\(A\)的特征矩阵的法式,其中
  1. \(A=\begin{pmatrix} -1&0&1\\3&2&-2\\-5&1&4 \end{pmatrix}\)
  2. \(A=\begin{pmatrix} 3&1&1\\0&4&0\\-1&1&5 \end{pmatrix}\)
解答.
  1. \begin{equation*} \begin{array}{rl} \lambda E-A=& \begin{pmatrix} \lambda+1&0&-1\\-3&\lambda-2&2\\5&-1&\lambda-4 \end{pmatrix}\xrightarrow[E(1,3)]{}\begin{pmatrix} -1&0&\lambda+1\\2&\lambda-2&-3\\\lambda-4&-1& 5 \end{pmatrix}\\ \xrightarrow[E(1,3(\lambda+1))]{{\begin{array}{c} E(2,1(2))\\E(3,1( \lambda -4)) \end{array}}}&\begin{pmatrix} -1&0&0\\0&\lambda-2&2 \lambda-1\\0&-1&\lambda^2-3 \lambda+1 \end{pmatrix}\xrightarrow{E(2,3)}\begin{pmatrix} -1&0&0\\0&-1&\lambda^2-3 \lambda+1\\0&\lambda-2&2 \lambda-1 \end{pmatrix}\\ \xrightarrow[E(2,3(\lambda^2-3 \lambda+1))]{{\begin{array}{c}E(3,2(\lambda-2))\\E(1(-1))\\E(2(-1))\end{array}}}&\begin{pmatrix} 1&0&0\\0&1&0\\0&0&\lambda^3-5 \lambda^2+9 \lambda-3 \end{pmatrix} \end{array} \end{equation*}
  2. \begin{equation*} \begin{array}{rl} \lambda E-A=& \begin{pmatrix} \lambda-3&-1&-1\\0&\lambda-4&0\\1&-1&\lambda-5 \end{pmatrix}\xrightarrow[E(1,3)]{}\begin{pmatrix} -1&-1&\lambda-3\\0&\lambda-4&0\\\lambda-5&-1&1 \end{pmatrix}\\ \xrightarrow[{\begin{array}{c} E(1,2(-1))\\E(1,3( \lambda -3)) \end{array}}]{E(3,1(\lambda-5))}&\begin{pmatrix} -1&0&0\\0&\lambda-4&0\\0&4-\lambda&(\lambda-4)^2 \end{pmatrix}\xrightarrow{E(3,2(1))}\begin{pmatrix} -1&0&0\\0&\lambda-4&0\\0&0&(\lambda-4)^2 \end{pmatrix}\\ \xrightarrow{E(1(-1))}&\begin{pmatrix} 1&0&0\\0&\lambda-4&0\\0&0&(\lambda-4)^2 \end{pmatrix}, \end{array} \end{equation*}
    所以\(A\)的特征矩阵的法式为\(\begin{pmatrix} 1&0&0\\0&\lambda-4&0\\0&0&(\lambda-4)^2 \end{pmatrix}\)

2.

\(n\)阶矩阵\(A\)的特征矩阵的法式为
\begin{equation*} {\rm diag} (1,\cdots ,1,d_1(\lambda),d_2(\lambda),\cdots ,d_k(\lambda)), \end{equation*}
证明:\(A\)的特征多项式
\begin{equation*} f_A(\lambda)=d_1(\lambda)d_2(\lambda)\cdots d_k(\lambda)\mbox{。} \end{equation*}
解答.
由于\(A\)的特征矩阵的法式为 \({\rm diag} (1,\cdots ,1,d_1(\lambda),d_2(\lambda),\cdots ,d_k(\lambda))\),所以存在可逆\(\lambda\)-矩阵\(P(\lambda)\)\(Q(\lambda)\),使得
\begin{equation*} \lambda E_n-A=P(\lambda){\rm diag} (1,\cdots ,1,d_1(\lambda),d_2(\lambda),\cdots ,d_k(\lambda))Q(\lambda), \end{equation*}
\begin{equation*} \det (\lambda E_n-A)=\det P(\lambda)\det Q(\lambda)d_1(\lambda)d_2(\lambda)\cdots d_k(\lambda), \end{equation*}
其中\(\det P(\lambda)\)\(\det Q(\lambda)\)为非零常数,即\(f_A(\lambda)=cd_1(\lambda)d_2(\lambda)\cdots d_k(\lambda),c\in\mathbb{F}\)。注意到\(f_A(\lambda), d_1(\lambda),d_2(\lambda),\cdots ,d_k(\lambda)\)均为首一多项式,故\(c=1\)。因此
\begin{equation*} f_A(\lambda)=d_1(\lambda)d_2(\lambda)\cdots d_k(\lambda)\mbox{。} \end{equation*}

3.

\(A,B\in\mathbb{F}^{n\times n},M(\lambda),N(\lambda)\)\(n\)\(\lambda\)-矩阵,且满足
\begin{equation} M(\lambda)(\lambda E-A)=(\lambda E-B)N(\lambda).\tag{7.1} \end{equation}
证明:
  1. 存在\(R\in\mathbb{F}^{n\times n}\)\(\lambda\)-矩阵\(Q(\lambda)\),使得
    \begin{equation*} M(\lambda)=(\lambda E-B)Q(\lambda)+R,\ N(\lambda)=Q(\lambda)(\lambda E-A)+R; \end{equation*}
  2. \(M(\lambda)\)可逆的充要条件是\(N(\lambda)\)可逆,此时\(R\)可逆,进而\(A\)相似于\(B\)
解答.
  1. 引理 7.2.3, 存在\(\lambda\)-矩阵\(Q(\lambda)\)以及数字矩阵\(R\in\mathbb{F}^{n\times n}\),使得
    \begin{equation*} M(\lambda)=(\lambda E-B)Q(\lambda)+R, \end{equation*}
    代入\(M(\lambda)(\lambda E-A)=(\lambda E-B)N(\lambda)\),整理得
    \begin{equation*} R(\lambda E-A)=(\lambda E-B)\left[N(\lambda)-Q(\lambda)(\lambda E-A)\right]. \end{equation*}
    上式左边是一次多项式,所以右式中\(N(\lambda)-Q(\lambda)(\lambda E-A)\)必须是零次的,即为数字矩阵,记为\(P\),则有\(R(\lambda E-A)=(\lambda E-B)P\),即
    \begin{equation} \lambda R-RA=\lambda P-BP,\tag{7.2} \end{equation}
    比较系数,有\(P=R\),即
    \begin{equation*} N(\lambda)-Q(\lambda)(\lambda E-A)=R, \end{equation*}
    \(N(\lambda)=Q(\lambda)(\lambda E-A)+R\)
  2. (7.1)两端同时取行列式得
    \begin{equation*} \det M(\lambda)\cdot f_A(\lambda)=f_B(\lambda)\cdot\det N(\lambda). \end{equation*}
    因为\(f_A(\lambda),f_B(\lambda)\)均为\(n\)次首一多项式,所以 \(\det M(\lambda)\)为非0常数的充要条件是\(\det N(\lambda)\)为非0常数,即\(M(\lambda)\)可逆的充要条件是\(N(\lambda)\)可逆。
    根据 定理 7.2.1的证明可知,当\(M(\lambda)\)可逆时,\(R\)可逆,进而\(A\)相似于\(B\)