Binet-Cauchy公式

前置知识

∣ A ∣ \left| A\right| A是一个 n n n阶行列式, k < n k<n k<n
1 ≤ i 1 < i 2 < ⋯ < i k ≤ n 1 ≤ j 1 < j 2 < ⋯ < j k ≤ n 1\le i_1 < i_2<\cdots < i_k \le n\\ 1\le j_1 < j_2<\cdots < j_k \le n 1i1<i2<<ikn1j1<j2<<jkn
∣ A ∣ \left| \mathbf{A}\right| A中第 i 1 , … , i k i_1,\dots,i_k i1,,ik行以及第 j 1 , … , j k j_1,\dots,j_k j1,,jk列交点上的元素,按原来 ∣ A ∣ \left| \mathbf{A}\right| A中的相对位置构成一个 k k k阶行列式,
我们称为 A \mathbf{A} A k k k子式,记为
A ( i 1 , … , i k j 1 , … , j k ) \mathbf{A}\begin{pmatrix} i_1,\dots,i_k\\ j_1,\dots,j_k \end{pmatrix} A(i1,,ikj1,,jk)

∣ A ∣ \left| \mathbf{A}\right| A中去掉第 i 1 , … , i k i_1,\dots,i_k i1,,ik行以及第 j 1 , … , j k j_1,\dots,j_k j1,,jk列以后,
剩下的元素,按原来 ∣ A ∣ \left| \mathbf{A}\right| A中的相对位置构成一个 n − k n-k nk阶行列式。
这个行列式称为余子式,记为
M ( i 1 , … , i k j 1 , … , j k ) \mathbf{M}\begin{pmatrix} i_1,\dots,i_k\\ j_1,\dots,j_k \end{pmatrix} M(i1,,ikj1,,jk)
若令 p = i 1 + ⋯ + i k , q = j 1 + ⋯ + j k p=i_1+\dots+i_k,q=j_1+\dots+j_k p=i1++ik,q=j1++jk,记
A ^ ( i 1 , … , i k j 1 , … , j k ) = ( − 1 ) p + q M ( i 1 , … , i k j 1 , … , j k ) \widehat{\mathbf{A}}\begin{pmatrix} i_1,\dots,i_k\\ j_1,\dots,j_k \end{pmatrix}=(-1)^{p+q}\mathbf{M}\begin{pmatrix} i_1,\dots,i_k\\ j_1,\dots,j_k \end{pmatrix} A (i1,,ikj1,,jk)=(1)p+qM(i1,,ikj1,,jk)
称为代数余子式

Laplace展开

∣ A ∣ \left| \mathbf{A}\right| A是一个 n n n阶行列式,在 ∣ A ∣ \left|\mathbf{A}\right| A中任取 k k k行(列),
那么含于这 k k k行(列)的 k k k阶子式与它们对应的代数余子式的乘积之和等于 ∣ A ∣ \left| \mathbf{A}\right| A,
即若取定k个行: 1 ≤ i 1 < i 2 < ⋯ < i k ≤ n 1\le i_1 < i_2<\cdots < i_k \le n 1i1<i2<<ikn,则
∣ A ∣ = ∑ 1 ≤ j 1 < j 2 < ⋯ < j k ≤ n A ( i 1 , … , i k j 1 , … , j k ) A ^ ( i 1 , … , i k j 1 , … , j k ) \left| \mathbf{A}\right|=\sum_{1\le j_1 < j_2<\cdots < j_k \le n}\mathbf{A}\begin{pmatrix} i_1,\dots,i_k\\ j_1,\dots,j_k \end{pmatrix}\widehat{\mathbf{A}}\begin{pmatrix} i_1,\dots,i_k\\ j_1,\dots,j_k \end{pmatrix} A=1j1<j2<<jknA(i1,,ikj1,,jk)A (i1,,ikj1,,jk)
同样的,若取定 k k k个列: 1 ≤ j 1 < i 2 < ⋯ < j k ≤ n 1\le j_1 < i_2<\cdots < j_k \le n 1j1<i2<<jkn,则
∣ A ∣ = ∑ 1 ≤ i 1 < i 2 < ⋯ < i k ≤ n A ( i 1 , … , i k j 1 , … , j k ) A ^ ( i 1 , … , i k j 1 , … , j k ) \left| \mathbf{A}\right|=\sum_{1\le i_1 < i_2<\cdots < i_k \le n}\mathbf{A}\begin{pmatrix} i_1,\dots,i_k\\ j_1,\dots,j_k \end{pmatrix}\widehat{\mathbf{A}}\begin{pmatrix} i_1,\dots,i_k\\ j_1,\dots,j_k \end{pmatrix} A=1i1<i2<<iknA(i1,,ikj1,,jk)A (i1,,ikj1,,jk)

Binet-Cauchy公式

设矩阵 A n × m , B m × n \mathbf{A}_{n\times m},\mathbf{B}_{m\times n} An×m,Bm×n,则
∣ A B ∣ = { 0 , n > m ∣ A ∣ ∣ B ∣ , n = m ∑ 1 ≤ j 1 < j 2 < ⋯ < j n ≤ m A ( 1 , … , n j 1 , … , j n ) B ( j 1 , … , j n 1 , … , n ) , n < m \left| \mathbf{A}\mathbf{B} \right| = \begin{cases} 0,&n>m\\ \left| \mathbf{A}\right| \left| \mathbf{B} \right|, &n=m\\ \sum_{1\le j_1<j_2<\cdots < j_n\le m} \mathbf{A}\begin{pmatrix} 1,\dots,n\\ j_1,\dots,j_n \end{pmatrix} \mathbf{B}\begin{pmatrix} j_1,\dots,j_n\\ 1,\dots,n \end{pmatrix},&n<m \end{cases} AB= 0,AB,1j1<j2<<jnmA(1,,nj1,,jn)B(j1,,jn1,,n),n>mn=mn<m

证明:
( I n − A 0 I m ) ( A 0 I m B ) = ( 0 − A B I m B ) \begin{pmatrix} \mathbf{I}_n& -\mathbf{A}\\ \mathbf{0}& \mathbf{I}_m\\ \end{pmatrix}\begin{pmatrix} \mathbf{A}& \mathbf{0}\\ \mathbf{I}_m & \mathbf{B}\\ \end{pmatrix}=\begin{pmatrix} \mathbf{0}& -\mathbf{A}\mathbf{B}\\ \mathbf{I}_m & \mathbf{B}\\ \end{pmatrix} (In0AIm)(AIm0B)=(0ImABB)
两边同取行列式
∣ I n − A 0 I m ∣ ∣ A 0 I m B ∣ = ∣ 0 − A B I m B ∣ ∣ I n ∣ ∣ I m ∣ ∣ A 0 I m B ∣ = ( − 1 ) m n ∣ − A B ∣ ∣ I m ∣ ∣ A 0 I m B ∣ = ( − 1 ) m n + n ∣ A B ∣ \begin{aligned} \left| \begin{array}{cccc} \mathbf{I}_n& -\mathbf{A}\\ \mathbf{0}& \mathbf{I}_m\\ \end{array} \right| \left| \begin{array}{cccc} \mathbf{A}&\mathbf{0}\\ \mathbf{I}_m & \mathbf{B}\\ \end{array} \right| &=\left| \begin{array}{cccc} \mathbf{0}& -\mathbf{A}\mathbf{B}\\ \mathbf{I}_m & \mathbf{B}\\ \end{array} \right| \\ \left|\mathbf{I}_n\right| \left|\mathbf{I}_m\right| \left| \begin{array}{cccc} \mathbf{A}&\mathbf{0}\\ \mathbf{I}_m & \mathbf{B}\\ \end{array} \right| &= (-1)^{mn}\left|-\mathbf{A}\mathbf{B}\right| \left|\mathbf{I}_m\right| \\ \left| \begin{array}{cccc} \mathbf{A}&\mathbf{0}\\ \mathbf{I}_m & \mathbf{B}\\ \end{array} \right| &= (-1)^{mn+n} \left|\mathbf{A}\mathbf{B}\right| \end{aligned} In0AIm AIm0B InIm AIm0B AIm0B = 0ImABB =(1)mnABIm=(1)mn+nAB
⇒ ∣ A B ∣ = ( − 1 ) m n + n ∣ A 0 I m B ∣ \Rightarrow \left|\mathbf{A}\mathbf{B}\right|= (-1)^{mn+n}\left| \begin{array}{cccc} \mathbf{A}&\mathbf{0}\\ \mathbf{I}_m & \mathbf{B}\\ \end{array} \right| AB=(1)mn+n AIm0B
( A 0 I m B ) = ( a 11 a 1 2 ⋯ a 1 m 0 0 ⋯ 0 a 21 a 2 2 ⋯ a 2 m 0 0 ⋯ 0 ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ a n 1 a n 2 ⋯ a n m 0 0 ⋯ 0 1 0 ⋯ 0 b 11 b 12 ⋯ b 1 n 0 1 ⋯ 0 b 21 b 22 ⋯ b 2 n ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ 0 0 ⋯ 1 b m 1 b m 2 ⋯ b m n ) \begin{pmatrix} \mathbf{A}&\mathbf{0}\\ \mathbf{I}_m&\mathbf{B}\\ \end{pmatrix}=\begin{pmatrix} a_{11}& a_{1_2}&\cdots & a_{1_m}& 0& 0 & \cdots 0\\ a_{21}& a_{2_2}&\cdots & a_{2_m}& 0& 0 & \cdots 0\\ \cdots& \cdots&\cdots & \cdots& \cdots& \cdots & \cdots\\ a_{n1}& a_{n2}&\cdots & a_{nm}& 0& 0 & \cdots 0\\ 1& 0&\cdots & 0& b_{11}& b_{12} & \cdots b_{1_n}\\ 0& 1&\cdots & 0& b_{21}& b_{22} & \cdots b_{2_n}\\ \cdots& \cdots&\cdots & \cdots& \cdots& \cdots & \cdots\\ 0& 0&\cdots & 1& b_{m 1} &b_{m_2}& \cdots b_{m_n}\\ \end{pmatrix} (AIm0B)= a11a21an1100a12a22an2010a1ma2manm001000b11b21bm1000b12b22bm2000b1nb2nbmn
使用拉普拉斯
展开前 n n n行,只需要取前 m m m列,因为前 n n n行的其余的列均为 0 0 0
n > m n>m n>m时, n n n阶子式至少有一列 0 0 0,所以
∣ A 0 I m B ∣ = 0 ⇒ ∣ A B ∣ = 0 \left| \begin{array}{cccc} \mathbf{A}&\mathbf{0}\\ \mathbf{I}_m & \mathbf{B}\\ \end{array} \right| =0 \Rightarrow \left| \mathbf{A}\mathbf{B} \right| =0 AIm0B =0AB=0
n = m n=m n=m
n m + n = n ( n + 1 ) nm+n=n(n+1) nm+n=n(n+1)为偶数
∣ A 0 I m B ∣ = ∣ A ∣ ∣ B ∣ = ( − 1 ) m n + n ∣ A B ∣ ⇒ ∣ A B ∣ = ∣ A ∣ ∣ B ∣ \left| \begin{array}{cccc} \mathbf{A}&\mathbf{0}\\ \mathbf{I}_m & \mathbf{B}\\ \end{array} \right| =\left| \mathbf{A} \right| \left| \mathbf{B} \right| = (-1)^{mn+n} \left|\mathbf{A}\mathbf{B}\right|\Rightarrow \left| \mathbf{A}\mathbf{B} \right| =\left| \mathbf{A} \right| \left| \mathbf{B} \right| AIm0B =AB=(1)mn+nABAB=AB
n < m n<m n<m
( A 0 I m B ) = ( a 11 a 1 2 ⋯ a 1 m 0 0 ⋯ 0 a 21 a 2 2 ⋯ a 2 m 0 0 ⋯ 0 ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ a n 1 a n 2 ⋯ a n m 0 0 ⋯ 0 1 0 ⋯ 0 b 11 b 12 ⋯ b 1 n 0 1 ⋯ 0 b 21 b 22 ⋯ b 2 n ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ ⋯ 0 0 ⋯ 1 b m 1 b m 2 ⋯ b m n ) \begin{pmatrix} \mathbf{A}&\mathbf{0}\\ \mathbf{I}_m&\mathbf{B}\\ \end{pmatrix}=\begin{pmatrix} a_{11}& a_{1_2}&\cdots & a_{1_m}& 0& 0 & \cdots 0\\ a_{21}& a_{2_2}&\cdots & a_{2_m}& 0& 0 & \cdots 0\\ \cdots& \cdots&\cdots & \cdots& \cdots& \cdots & \cdots\\ a_{n1}& a_{n2}&\cdots & a_{nm}& 0& 0 & \cdots 0\\ 1& 0&\cdots & 0& b_{11}& b_{12} & \cdots b_{1_n}\\ 0& 1&\cdots & 0& b_{21}& b_{22} & \cdots b_{2_n}\\ \cdots& \cdots&\cdots & \cdots& \cdots& \cdots & \cdots\\ 0& 0&\cdots & 1& b_{m 1} &b_{m_2}& \cdots b_{m_n}\\ \end{pmatrix} (AIm0B)= a11a21an1100a12a22an2010a1ma2manm001000b11b21bm1000b12b22bm2000b1nb2nbmn
e i \mathbf{e}_i ei为第 i i i个元素为 1 1 1,其余元素为 0 0 0的向量
u 1 , u 2 , … , u m − n u_1,u_2,\dots,u_{m-n} u1,u2,,umn 1 , 2 , … , m 1,2,\dots,m 1,2,,m去掉 j 1 , j 2 , … , j n j_1,j_2,\dots,j_n j1,j2,,jn后剩下的列

∣ A 0 I m B ∣ = ∑ 1 ≤ j 1 < j 2 < ⋯ < j n ≤ m A ( 1 , … , n j 1 , … , j n ) ( − 1 ) ∑ i = 1 n i + ∑ i = 1 n j i ∣ C ∣ \left| \begin{array}{cccc} \mathbf{A}&\mathbf{0}\\ \mathbf{I}_m & \mathbf{B}\\ \end{array} \right| =\sum_{1\le j_1<j_2<\cdots < j_n\le m} \mathbf{A}\begin{pmatrix} 1,\dots,n\\ j_1,\dots,j_n \end{pmatrix} (-1)^{\sum_{i=1}^{n} i+\sum_{i=1}^{n}j_{i}}\left| C\right| AIm0B =1j1<j2<<jnmA(1,,nj1,,jn)(1)i=1ni+i=1njiC
其中
C = ( e u 1 , e u 2 , … , e u m − n , B ) \mathbf{C}=(\mathbf{e}_{u_1},\mathbf{e}_{u_2},\dots,\mathbf{e}_{u_{m-n}},\mathbf{B}) C=(eu1,eu2,,eumn,B)
然后再展开前 m − n m-n mn
∣ C ∣ = ∣ I m − n ∣ ( − 1 ) ∑ i = 1 m − n i + ∑ i = 1 m − n u i B ( j 1 , … , j n 1 , … , n ) \left| \mathbf{C} \right| =\left| \mathbf{I}_{m-n}\right| (-1)^{\sum_{i=1}^{m-n} i+\sum_{i=1}^{m-n} u_{i}}\mathbf{B}\begin{pmatrix} j_1,\dots,j_n\\ 1,\dots,n \end{pmatrix} C=Imn(1)i=1mni+i=1mnuiB(j1,,jn1,,n)

至于为什么 B B B中也是 j 1 , ⋯   , j n j_1,\cdots,j_n j1,,jn
红色为 j j j,蓝色为 u u u
可以看出第二次展开 u u u之后,剩下的就是 j j j
在这里插入图片描述

两个式子合在一起
接着我们重点看 − 1 -1 1的次幂
( − 1 ) ∑ i = 1 m − n i + ∑ i = 1 m − n u i ( − 1 ) ∑ i = 1 n i + ∑ i = 1 n j i = ( − 1 ) ( 1 + n ) n 2 + ( 1 + m − n ) ( m − n ) 2 + ∑ i = 1 m − n u i + ∑ i = 1 n j i = ( − 1 ) ( 1 + n ) n 2 + ( 1 + m − n ) ( m − n ) 2 + ( 1 + m ) m 2 = ( − 1 ) n 2 + m 2 + m − m n \begin{aligned} &\quad (-1)^{\sum_{i=1}^{m-n} i+\sum_{i=1}^{m-n} u_{i}}(-1)^{\sum_{i=1}^{n} i+\sum_{i=1}^{n}j_{i}}\\ &=(-1)^{\frac{(1+n)n}{2}+\frac{(1+m-n)(m-n)}{2}+\sum_{i=1}^{m-n} u_{i}+\sum_{i=1}^{n}j_{i}}\\ &=(-1)^{\frac{(1+n)n}{2}+\frac{(1+m-n)(m-n)}{2}+\frac{(1+m)m}{2}}\\ &=(-1)^{n^2+m^2+m-mn} \end{aligned} (1)i=1mni+i=1mnui(1)i=1ni+i=1nji=(1)2(1+n)n+2(1+mn)(mn)+i=1mnui+i=1nji=(1)2(1+n)n+2(1+mn)(mn)+2(1+m)m=(1)n2+m2+mmn

∣ A B ∣ = ( − 1 ) m n + n ∣ A 0 I m B ∣ = ( − 1 ) m n + n ( − 1 ) n 2 + m 2 + m − m n ∑ 1 ≤ j 1 < j 2 < ⋯ < j n ≤ m A ( 1 , … , n j 1 , … , j n ) B ( j 1 , … , j n 1 , … , n ) = ( − 1 ) n ( n + 1 ) + m ( m + 1 ) ∑ 1 ≤ j 1 < j 2 < ⋯ < j n ≤ m A ( 1 , … , n j 1 , … , j n ) B ( j 1 , … , j n 1 , … , n ) = ∑ 1 ≤ j 1 < j 2 < ⋯ < j n ≤ m A ( 1 , … , n j 1 , … , j n ) B ( j 1 , … , j n 1 , … , n ) \begin{aligned} \left|\mathbf{A}\mathbf{B}\right| &= (-1)^{mn+n}\left| \begin{array}{cccc} \mathbf{A}&\mathbf{0}\\ \mathbf{I}_m & \mathbf{B}\\ \end{array} \right| \\ &=(-1)^{mn+n} (-1)^{n^2+m^2+m-mn} \\ &\quad \sum_{1\le j_1<j_2<\cdots < j_n\le m} \mathbf{A}\begin{pmatrix} 1,\dots,n\\ j_1,\dots,j_n \end{pmatrix} \mathbf{B}\begin{pmatrix} j_1,\dots,j_n\\ 1,\dots,n \end{pmatrix} \\ &=(-1)^{n(n+1)+m(m+1)} \\ &\quad \sum_{1\le j_1<j_2<\cdots < j_n\le m} \mathbf{A}\begin{pmatrix} 1,\dots,n\\ j_1,\dots,j_n \end{pmatrix} \mathbf{B}\begin{pmatrix} j_1,\dots,j_n\\ 1,\dots,n \end{pmatrix} \\ &=\sum_{1\le j_1<j_2<\cdots < j_n\le m} \mathbf{A}\begin{pmatrix} 1,\dots,n\\ j_1,\dots,j_n \end{pmatrix} \mathbf{B}\begin{pmatrix} j_1,\dots,j_n\\ 1,\dots,n \end{pmatrix} \end{aligned} AB=(1)mn+n AIm0B =(1)mn+n(1)n2+m2+mmn1j1<j2<<jnmA(1,,nj1,,jn)B(j1,,jn1,,n)=(1)n(n+1)+m(m+1)1j1<j2<<jnmA(1,,nj1,,jn)B(j1,,jn1,,n)=1j1<j2<<jnmA(1,,nj1,,jn)B(j1,,jn1,,n)
所以结论成立

评论 1
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包

打赏作者

Nightmare004

你的鼓励将是我创作的最大动力

¥1 ¥2 ¥4 ¥6 ¥10 ¥20
扫码支付:¥1
获取中
扫码支付

您的余额不足,请更换扫码支付或充值

打赏作者

实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值