高等代数

高等代数理论基础76:A-矩阵

2019-04-26  本文已影响2人  溺于恐

\mathscr{A}-矩阵

任给数域P上n维空间V上线性变换\mathscr{A},已定义过P上\mathscr{A}的多项式,即\forall f(\lambda)\in P[\lambda],f(\lambda)=a_k\lambda^k+a_{k-1}\lambda^{k-1}+\cdots+a_1\lambda+a_0,称f(\mathscr{A})=a_k\mathscr{A}^k+a_{k-1}\mathscr{A}^{k-1}+\cdots+a_1\mathscr{A}+a_0\mathscr{E}为P上\mathscr{A}的多项式,其中\mathscr{E}为V上恒等变换,f(\mathscr{A})仍为V上线性变换

定义:对P上任意\lambda-矩阵K(\lambda)=(a_{ij}(\lambda))_{m\times l},a_{ij}(\lambda)\in P[\lambda],令K(\mathscr{A})=(a_{ij}(\mathscr{A}))_{m\times l},称为P上的一个\mathscr{A}-矩阵

例:

K(\lambda)=\begin{pmatrix}\lambda^2+1&\lambda-1\\3&\lambda^3+\lambda\end{pmatrix},K(\mathscr{A})=\begin{pmatrix}\mathscr{A}^2+\mathscr{E}&\mathscr{A}-\mathscr{E}\\3\mathscr{E}&\mathscr{A}^3+\mathscr{A}\end{pmatrix}

若设F(\lambda)=\lambda E-B,称为B的特征矩阵,设B=(b_{ij})_{n\times n},则

F(\lambda)=\lambda E-B=\begin{pmatrix}\lambda-b_{11}&-b_{12}&\cdots&-b_{1n}\\ -b_{21}&\lambda-b_{22}&\cdots&-b_{2n}\\ \vdots&\vdots& &\vdots\\ -b_{n1}&-b_{n2}&\cdots&\lambda-b_{nn}\end{pmatrix}

F(\mathscr{A})=\begin{pmatrix}\mathscr{A}-b_{11}\mathscr{E}&-b_{12}\mathscr{E}&\cdots&-b_{1n}\mathscr{E}\\ -b_{21}\mathscr{E}&\mathscr{A}-b_{22}\mathscr{E}&\cdots&-b_{2n}\mathscr{E}\\ \vdots&\vdots& &\vdots\\ -b_{n1}\mathscr{E}&-b_{n2}\mathscr{A}&\cdots&\mathscr{A}-b_{1n}\mathscr{E}\end{pmatrix}​

矩阵运算及行列式定义中只用到元素的加法和乘法,而\mathscr{A}​的多项式有加法和乘法

\lambda-矩阵一样,\mathscr{A}-矩阵也有加法、乘法及"数量"乘法(用\mathscr{A}的多项式作为元素与\mathscr{A}-矩阵,或与数字矩阵作"数量"乘法),也有与数字矩阵、\lambda-矩阵类似的运算性质,也可定义\mathscr{A}-矩阵的行列式,也有与数字行列式相同的性质,如行列式乘法定理、伴随矩阵的存在性等

\mathscr{A}-矩阵运算可将F(\mathscr{A})表为

F(\mathscr{A})=\begin{pmatrix}\mathscr{A}-b_{11}\mathscr{E}&-b_{12}\mathscr{E}&\cdots&-b_{1n}\mathscr{E}\\ -b_{21}\mathscr{E}&\mathscr{A}-b_{22}\mathscr{E}&\cdots&-b_{2n}\mathscr{E}\\ \vdots&\vdots& &\vdots\\ -b_{n1}\mathscr{E}&-b_{n2}\mathscr{A}&\cdots&\mathscr{A}-b_{1n}\mathscr{E}\end{pmatrix}=\mathscr{A}E=\mathscr{E}B

f(\lambda),g(\lambda)\in P[x],令h(\lambda)=f(\lambda)+g(\lambda),p(\lambda)=f(\lambda)g(\lambda)

h(\mathscr{A})=f(\mathscr{A})+g(\mathscr{A}),p(\mathscr{A})=f(\mathscr{A})g(\mathscr{A})

同时,对\lambda-矩阵,K(\lambda)_{m\times k},L(\lambda)_{m\times k},若H(\lambda)=K(\lambda)+L(\lambda)

H(\mathscr{A})=K(\mathscr{A})+L(\mathscr{A})

K(\lambda)_{m\times s},L(\lambda)_{s\times k},若M(\lambda)=K(\lambda)L(\lambda)

M(\mathscr{A})=K(\mathscr{A})L(\mathscr{A})

对V上的一个线性变换\mathscr{A}​,设它在基\varepsilon_1,\varepsilon_2,\cdots,\varepsilon_n​下的矩阵为A'​

\mathscr{A}(\varepsilon_1,\varepsilon_2,\cdots,\varepsilon_n)=(\varepsilon_1,\varepsilon_2,\cdots,\varepsilon_n)A'

A=(a_{ij})

\mathscr{A}\varepsilon_1=a_{11}\varepsilon_1+a_{12}\varepsilon_2+\cdots+a_{1n}\varepsilon_n=a_{11}\mathscr{E}\varepsilon_1+a_{12}\mathscr{E}\varepsilon_2+\cdots+a_{1n}\mathscr{E}\varepsilon_n​

\mathscr{A}\varepsilon_2=a_{21}\varepsilon_1+a_{22}\varepsilon_2+\cdots+a_{2n}\varepsilon_n=a_{21}\mathscr{E}\varepsilon_1+a_{22}\mathscr{E}\varepsilon_2+\cdots+a_{2n}\mathscr{E}\varepsilon_n

\cdots

\mathscr{A}\varepsilon_1=a_{n1}\varepsilon_1+a_{n2}\varepsilon_2+\cdots+a_{nn}\varepsilon_n=a_{n1}\mathscr{E}\varepsilon_1+a_{n2}\mathscr{E}\varepsilon_2+\cdots+a_{nn}\mathscr{E}\varepsilon_n

\mathscr{A}\begin{pmatrix}\varepsilon_1\\\varepsilon_2\\\vdots\\\varepsilon_n\end{pmatrix}=A\begin{pmatrix}\varepsilon_1\\\varepsilon_2\\\vdots\\\varepsilon_n\end{pmatrix}

注:可用\mathscr{A}-矩阵的形式表达式写出

矩阵A称为\mathscr{A}在基\varepsilon_1,\varepsilon_2,\cdots,\varepsilon_n下的左矩阵

\mathscr{A}在同一基\varepsilon_1,\varepsilon_2,\cdots,\varepsilon_n下的矩阵A'和左矩阵A互为转置关系,有相同特征多项式

\mathscr{A}在不同基下的矩阵是相似的,它们对应的转置也相似,即\mathscr{A}在不同基下的左矩阵也相似,故\mathscr{A}在任何基下的矩阵及左矩阵的特征多项式都相等,都是\mathscr{A}的特征多项式

(\mathscr{A}E-\mathscr{E}A)=\begin{pmatrix}\varepsilon_1\\\varepsilon_2\\\vdots\\\varepsilon_n\end{pmatrix}=O

注:左端为形式写法,是\mathscr{A}-矩阵与向量元素的矩阵的"乘法",不是\mathscr{A}-矩阵的乘法

\forall \mathscr{A}-矩阵K(\mathscr{A})=(k_{ij}(\mathscr{A}))_{m\times n}及向量矩阵\begin{pmatrix}\xi_1\\\xi_2\\\vdots \\\xi_n\end{pmatrix},

K(\mathscr{A})\begin{pmatrix}\xi_1\\\xi_2\\\vdots\\\xi_n\end{pmatrix}=\begin{pmatrix}k_{11}\xi_1+k_{12}\xi_2+\cdots+k_{1n}\xi_n\\ k_{12}\xi_1+k_{22}\xi_2+\cdots+k_{2n}\xi_n\\ \vdots\\ k_{n1}\xi_1+k_{n2}\xi_2+\cdots+k_{nn}\xi_n\end{pmatrix}​

\mathscr{A}-矩阵的形式写法,与数字矩阵的形式写法一样,有性质:

1.(K(\mathscr{A})_{m\times s}L(\mathscr{A})_{s\times k})\begin{pmatrix}\xi_1\\\xi_2\\\vdots\\\xi_k\end{pmatrix}=K(\mathscr{A})\begin{bmatrix}L(\mathscr{A})\begin{pmatrix}\xi_1\\\xi_2\\\vdots\\\xi_n\end{pmatrix}\end{bmatrix}

2.[K(\mathscr{A})_{m\times k}+L(\mathscr{A})_{m\times k}]\begin{pmatrix}\xi_1\\\xi_2\\\vdots\\\xi_n\end{pmatrix}=K(\mathscr{A})\begin{pmatrix}\xi_1\\\xi_2\\\vdots\\\xi_n\end{pmatrix}+L(\mathscr{A})\begin{pmatrix}\xi_1\\\xi_2\\\vdots\\\xi_n\end{pmatrix}

3.设f(\mathscr{A})是一个\mathscr{A}-多项式,则

[f(\mathscr{A})K(\mathscr{A})]\begin{pmatrix}\xi_1\\\xi_2\\\vdots\\\xi_n\end{pmatrix}=K(\mathscr{A})\begin{bmatrix}K(\mathscr{A})\begin{pmatrix}\xi_1\\\xi_2\\\vdots\\\xi_n\end{pmatrix}\end{bmatrix}

注:\mathscr{A}-矩阵的形式写法与数字矩阵的形式有不同性质

例:对可逆的数字矩阵T_{n\times n}及一组基\varepsilon_1,\varepsilon_2,\cdots,\varepsilon_n,

\begin{pmatrix}\eta_1\\\eta_2\\\vdots\\\eta_n\end{pmatrix}=T\begin{pmatrix}\varepsilon_1\\\varepsilon_2\\\vdots\\\varepsilon_n\end{pmatrix}​

\eta_1,\eta_2,\cdots,\eta_n仍是一组基

对可逆的\mathscr{A}-矩阵K(\mathscr{A})_{n\times n}

即存在\mathscr{A}-矩阵L(\mathscr{A})_{n\times n}使L(\mathscr{A})K(\mathscr{A})=K(\mathscr{A})L(\mathscr{A})=E

对基\varepsilon_1,\varepsilon_2,\cdots,\varepsilon_n作形式写法

K(\mathscr{A})\begin{pmatrix}\varepsilon_1\\\varepsilon_2\\\vdots\\\varepsilon_n\end{pmatrix}=\begin{pmatrix}\eta_1\\\eta_2\\\vdots\\\eta_n\end{pmatrix}​

可能\eta_1,\eta_2,\cdots,\eta_n​不是基,可能有\eta_i=0​

例:设\mathscr{A}\begin{pmatrix}\varepsilon_1\\\varepsilon_2\\\varepsilon_3\end{pmatrix}=\begin{pmatrix}4&6&-15\\1&3&-5\\1&2&4\end{pmatrix}\begin{pmatrix}\varepsilon_1\\\varepsilon_2\\\varepsilon_3\end{pmatrix}

Q(\lambda)=\begin{pmatrix}1&-\lambda+3&-5\\0&1&-1\\0&0&1\end{pmatrix}

|Q(\lambda)|=1,故Q(\lambda)是可逆\lambda-矩阵

Q(\lambda)L(\lambda)=L(\lambda)Q(\lambda)=E

Q(\mathscr{A})L(\mathscr{A})=L(\mathscr{A})Q(\mathscr{A})=\mathscr{E}

Q(\mathscr{A})是可逆矩阵

\begin{pmatrix}\eta_1\\\eta_2\\\eta_3\end{pmatrix}=Q(\mathscr{A})\begin{pmatrix}\varepsilon_1\\\varepsilon_2\\\varepsilon_3\end{pmatrix}

=\begin{pmatrix}\varepsilon_1+(-\mathscr{A}+3\mathscr{E})\varepsilon_2-5\varepsilon_3\\\varepsilon_2-\varepsilon_3\\\varepsilon_3\end{pmatrix}

=\begin{pmatrix}\varepsilon_1+(-(\varepsilon_1+3\varepsilon_2-5\varepsilon_3)+3\varepsilon_2)-5\varepsilon_3\\\varepsilon_2-\varepsilon_3\\\varepsilon_3\end{pmatrix}

=\begin{pmatrix}0\\\varepsilon_2-\varepsilon_3\\\varepsilon_3\end{pmatrix}

\eta_1=0,

\eta_1,\eta_2,\eta_3不能为基

上一篇下一篇

猜你喜欢

热点阅读