大部分国内的线性代数都是从行列式开始讲起,这里只列出基本性质方便回顾。

行列式的几何意义

  1. 行列式中的行或列向量所构成的超平行多面体的有向面积或体积
  2. 坐标系变换下的图形面积或体积的伸缩因子,即变换矩阵AA

行列式的一般性质

  1. In=det(In)=1|I_n| = det(I_n)=1
  2. A=(α1,...,αn),B=(α1,...αi1,kαi,αi+1,...αn)A = (\alpha_1,...,\alpha_n),B = (\alpha_1,...\alpha_{i-1},k\alpha_i,\alpha_{i+1},...\alpha_n),则det(B)=k×det(A)det(B) = k\times det(A)
  3. A=A=(α1,...,αn),A=(α1,...,αi,...αn),B=(α1,...αi+αi,..,αn)A = A = (\alpha_1,...,\alpha_n),A^{'} = (\alpha_1,...,\alpha_i^{'},...\alpha_n),B = (\alpha_1,...\alpha_{i}+\alpha_i^{'},..,\alpha_n),则det(B)=det(A)+det(A)det(B) = det(A)+det(A^{'})
  4. det(A)=det(AT)det(A) = det(A^T)
  5. 任意交换AA的两列得到AA^{'},则det(A)=det(A)det(A) = -det(A^{'})

推论

  1. AA的两行(列)成比例,则det(A)=0det(A) = 0

  2. AA的某一行(列)乘上一个倍数加到另外一列(行),得到矩阵AA^{'},则det(A)=det(A)det(A) = det(A^{'})

  3. AA是一个方阵,则det(A)0Adet(A) \neq 0 \Leftrightarrow A可逆

  4. A,BA,B是两个nn阶方阵,则AB=A×B|AB| = |A| \times |B|

  5. A+BA+B,kAkA|A+B| \neq |A|+|B|,|kA| \neq k|A|

  6. A=(aij)nA = (a_{ij})_nnn阶矩阵,AijA_{ij}A|A|中元素aija_{ij}的代数余子式,则称矩阵AA^{\star}AA的伴随矩阵:

    A=(A11A12...A1nA21A22...A2n............An1An2...Ann)TA^{\star} = \begin{pmatrix}A_{11} &A_{12} &... &A_{1n} \\\\ A_{21}&A_{22} &... &A_{2n} \\\\ ... &... &... &... \\\\ A_{n1} &A_{n2} &... &A_{nn} \end{pmatrix}^T

    不难得到: AA=AIAA^{\star} = |A|I

    因此: A1=1AAA^{-1} = \frac {1}{|A|}A^{\star}

  7. AA为奇异矩阵时,不难得到AA=AI=zero matrixAA^{\star} = |A|I =zero \space matrix。因此,AA^{\star}的每一列都在AA的零空间中。

行列式的计算

  1. 直接利用公式计算

  2. 使用代数余子式(algebraic complement)计算

    Cij=(1)i+jdet(Mij)C_{ij} = (-1)^{i+j}det(M_{ij}),则:det(A)=ai1Ci1+ai2Ci2+...+ainCindet(A) = a_{i1}C_{i1}+a_{i2}C_{i2}+...+a_{in}C_{in}

  3. 综合利用消元法和降阶法

    典型例题:计算Vandermonde行列式

    证明λImAB=λmnλInBA|\lambda I_m -AB| = \lambda^{m-n}|\lambda I_n - BA|

求逆矩阵

A=(aij)n×nA = (a_{ij})_{n\times n}可逆,构造如下矩阵,称为AA的伴随矩阵(adjoint of A):

adj(A)=(C11C12...C1nC21C22...C2n............Cn1Cn2...Cnn)Tadj(A) = \begin{pmatrix}C_{11} &C_{12} &... &C_{1n} \\\\ C_{21}&C_{22} &... &C_{2n} \\\\ ... &... &... &... \\\\ C_{n1} &C_{n2} &... &C_{nn} \end{pmatrix}^T

A1=adj(A)AA^{-1} = \frac {adj(A)}{|A|}

r(A)=nr(adj(A))=nr(A ) = n \Rightarrow r(adj(A)) = n

r(A)=n1A×(adj(A))=0r(A ) = n-1 \Rightarrow A\times (adj(A)) = 0,因此adj(A)adj(A)的列属于AA的零空间。而dimN(A)=1r(adj(A))=1dim N(A) = 1 \Rightarrow r(adj(A)) = 1

r(A)n2Ar(A) \le n-2 \Rightarrow A的任意n-1阶子矩阵都不可逆Cij=0adj(A)=0\Rightarrow C_{ij} = 0 \rightarrow adj(A) = 0

外积

给定两个向量u=(u1u2u3),v=(v1v2v3),u×v=(u2v3u3v2u3v1u1v3u1v2u2v1)u=\begin{pmatrix}u_1\\\\u_2 \\\\u_3 \end{pmatrix},v=\begin{pmatrix}v_1\\\\v_2 \\\\v_3 \end{pmatrix},u\times v = \begin{pmatrix}u_2v_3 - u_3v_2\\\\u_3v_1-u_1v_3 \\\\ u_1v_2-u_2v_1 \end{pmatrix}

i,j,ki,j,k为单位向量,则:

ijku1u2u3v1v2v3=(u2v3u3v2)i+(u3v1u1v3)j+(u1v2u2v1)k\begin{vmatrix}i & j & k\\\\ u_1 &u_2 &u_3 \\\\ v_1& v_2 &v_3 \end{vmatrix} = (u_2v_3 - u_3v_2)\textbf i + (u_3v_1 - u_1v_3)\textbf{j}+ (u_1v_2 - u_2v_1)\textbf k

性质

  1. u×v=v×uu×u=0u \times v = -v \times u \rightarrow u\times u = 0
  2. (u1+u2)×v=u1×v+u2×v(u_1 + u_2) \times v = u_1 \times v + u_2 \times v

results matching ""

    No results matching ""