本文主要是介绍【高等代数笔记】(18)N阶行列式,希望对大家解决编程问题提供一定的参考价值,需要的开发者们随着小编来一起学习吧!
2. N阶行列式
2.12 行列式按k行(列)展开
【拉普拉斯定理】 n n n阶矩阵 A = ( a i j ) \boldsymbol{A}=(a_{ij}) A=(aij),取定第 i 1 , i 2 , . . . , i k i_{1},i_{2},...,i_{k} i1,i2,...,ik行(其中 i 1 < i 2 < . . . < i k i_{1}<i_{2}<...<i_{k} i1<i2<...<ik),则 ∣ A ∣ |\boldsymbol{A}| ∣A∣等于这 k k k行形成的所有 k k k阶子式与它自己的代数余子式的乘积之和。
【证】(丘维声老师讲的那段证明实在没看懂,我自己上网查资料看懂了一个证明过程,写在下面,证明需要如下前置知识)
-
引理0:若排列 a 1 a 2 . . . a n a_{1}a_{2}...a_{n} a1a2...an经过 s s s次对换变为排列 c 1 c 2 . . . c n c_{1}c_{2}...c_{n} c1c2...cn,则 ( − 1 ) τ ( a 1 a 2 . . . a n ) + s = ( − 1 ) τ ( c 1 c 2 . . . c n ) (-1)^{\tau(a_{1}a_{2}...a_{n})+s}=(-1)^{\tau(c_{1}c_{2}...c_{n})} (−1)τ(a1a2...an)+s=(−1)τ(c1c2...cn)(显然成立,对换多少次,逆序数增加多少)
-
引理1:任意一个由 1 , 2 , 3 , . . . , n 1,2,3,...,n 1,2,3,...,n构成的排列 a 1 a 2 . . . a k b 1 b 2 . . . b n − k a_{1}a_{2}...a_{k}b_{1}b_{2}...b_{n-k} a1a2...akb1b2...bn−k,有 ( − 1 ) τ ( a 1 a 2 . . . a k b 1 b 2 . . . b n − k ) = ( − 1 ) τ ( a 1 a 2 . . . a k ) + τ ( b 1 b 2 . . . b n − k ) + ( a 1 + a 2 + . . . + a k ) + k ( k + 1 ) 2 (-1)^{\tau(a_{1}a_{2}...a_{k}b_{1}b_{2}...b_{n-k})}=(-1)^{\tau(a_{1}a_{2}...a_{k})+\tau(b_{1}b_{2}...b_{n-k})+(a_{1}+a_{2}+...+a_{k})+\frac{k(k+1)}{2}} (−1)τ(a1a2...akb1b2...bn−k)=(−1)τ(a1a2...ak)+τ(b1b2...bn−k)+(a1+a2+...+ak)+2k(k+1)
【证】将子排列 a 1 a 2 . . . a k a_{1}a_{2}...a_{k} a1a2...ak经过 s , s ∈ N + s,s\in\mathbb{N}^{+} s,s∈N+次对换变为排列 c 1 c 2 . . . c k c_{1}c_{2}...c_{k} c1c2...ck且 c 1 < c 2 < . . . < c k c_{1}<c_{2}<...<c_{k} c1<c2<...<ck,即 c 1 c 2 . . . c k c_{1}c_{2}...c_{k} c1c2...ck无逆序由引理0, ( − 1 ) τ ( a 1 a 2 . . . a k ) + s = ( − 1 ) τ ( c 1 c 2 . . . c k ) (-1)^{\tau(a_{1}a_{2}...a_{k})+s}=(-1)^{\tau(c_{1}c_{2}...c_{k})} (−1)τ(a1a2...ak)+s=(−1)τ(c1c2...ck)
由于 c 1 c 2 . . . c k c_{1}c_{2}...c_{k} c1c2...ck无逆序,故 τ ( c 1 c 2 . . . c k ) = 0 \tau(c_{1}c_{2}...c_{k})=0 τ(c1c2...ck)=0
从而 ( − 1 ) τ ( a 1 a 2 . . . a k ) + s = ( − 1 ) 0 = 1 (-1)^{\tau(a_{1}a_{2}...a_{k})+s}=(-1)^{0}=1 (−1)τ(a1a2...ak)+s=(−1)0=1
所以 ( − 1 ) τ ( a 1 a 2 . . . a k ) + s = 1 (-1)^{\tau(a_{1}a_{2}...a_{k})+s}=1 (−1)τ(a1a2...ak)+s=1
同时,由于子排列 a 1 a 2 . . . a k a_{1}a_{2}...a_{k} a1a2...ak经过 s s s次变换为 c 1 c 2 . . . c k c_{1}c_{2}...c_{k} c1c2...ck
故排列 a 1 a 2 . . . a k b 1 b 2 . . . b n − k a_{1}a_{2}...a_{k}b_{1}b_{2}...b_{n-k} a1a2...akb1b2...bn−k也经过 s s s次变换变为 c 1 c 2 . . . c k b 1 b 2 . . . b n − k c_{1}c_{2}...c_{k}b_{1}b_{2}...b_{n-k} c1c2...ckb1b2...bn−k
由引理0, ( − 1 ) τ ( a 1 a 2 . . . a k b 1 b 2 . . . b n − k ) + s = ( − 1 ) τ ( c 1 c 2 . . . c k b 1 b 2 . . . b n − k ) (-1)^{\tau(a_{1}a_{2}...a_{k}b_{1}b_{2}...b_{n-k})+s}=(-1)^{\tau(c_{1}c_{2}...c_{k}b_{1}b_{2}...b_{n-k})} (−1)τ(a1a2...akb1b2...bn−k)+s=(−1)τ(c1c2...ckb1b2...bn−k),
又因为 ( − 1 ) τ ( a 1 a 2 . . . a k ) + s = ( − 1 ) 0 = 1 (-1)^{\tau(a_{1}a_{2}...a_{k})+s}=(-1)^{0}=1 (−1)τ(a1a2...ak)+s=(−1)0=1即 ( − 1 ) − s = ( − 1 ) τ ( a 1 a 2 . . . a k ) (-1)^{-s}=(-1)^{\tau(a_{1}a_{2}...a_{k})} (−1)−s=(−1)τ(a1a2...ak)
则两边同时乘 ( − 1 ) − s (-1)^{-s} (−1)−s得 ( − 1 ) τ ( a 1 a 2 . . . a k b 1 b 2 . . . b n − k ) + s ⋅ ( − 1 ) − s = ( − 1 ) τ ( c 1 c 2 . . . c k b 1 b 2 . . . b n − k ) ⋅ ( − 1 ) − s (-1)^{\tau(a_{1}a_{2}...a_{k}b_{1}b_{2}...b_{n-k})+s}\cdot(-1)^{-s}=(-1)^{\tau(c_{1}c_{2}...c_{k}b_{1}b_{2}...b_{n-k})}\cdot(-1)^{-s} (−1)τ(a1a2...akb1b2...bn−k)+s⋅(−1)−s=(−1)τ(c1c2...ckb1b2...bn−k)⋅(−1)−s
即 ( − 1 ) τ ( a 1 a 2 . . . a k b 1 b 2 . . . b n − k ) = ( − 1 ) τ ( c 1 c 2 . . . c k b 1 b 2 . . . b n − k ) ⋅ ( − 1 ) − s = ( − 1 ) τ ( c 1 c 2 . . . c k b 1 b 2 . . . b n − k ) ⋅ ( − 1 ) τ ( a 1 a 2 . . . a k ) = ( − 1 ) τ ( c 1 c 2 . . . c k b 1 b 2 . . . b n − k ) + τ ( a 1 a 2 . . . a k ) (-1)^{\tau(a_{1}a_{2}...a_{k}b_{1}b_{2}...b_{n-k})}=(-1)^{\tau(c_{1}c_{2}...c_{k}b_{1}b_{2}...b_{n-k})}\cdot(-1)^{-s}=(-1)^{\tau(c_{1}c_{2}...c_{k}b_{1}b_{2}...b_{n-k})}\cdot(-1)^{\tau(a_{1}a_{2}...a_{k})}=(-1)^{\tau(c_{1}c_{2}...c_{k}b_{1}b_{2}...b_{n-k})+\tau(a_{1}a_{2}...a_{k})} (−1)τ(a1a2...akb1b2...bn−k)=(−1)τ(c1c2...ckb1b2...bn−k)⋅(−1)−s=(−1)τ(c1c2...ckb1b2...bn−k)⋅(−1)τ(a1a2...ak)=(−1)τ(c1c2...ckb1b2...bn−k)+τ(a1a2...ak)
现在考虑 τ ( c 1 c 2 . . . c k b 1 b 2 . . . b n − k ) \tau(c_{1}c_{2}...c_{k}b_{1}b_{2}...b_{n-k}) τ(c1c2...ckb1b2...bn−k)
由于 c 1 c 2 . . . c k c_{1}c_{2}...c_{k} c1c2...ck本身无逆序,故 c i ∈ c 1 c 2 . . . c k c_{i}\in c_{1}c_{2}...c_{k} ci∈c1c2...ck只可能与 b i ∈ b 1 b 2 . . . b k b_{i}\in b_{1}b_{2}...b_{k} bi∈b1b2...bk形成逆序,现在考虑 c i c_{i} ci与 b 1 b 2 . . . b n − k b_{1}b_{2}...b_{n-k} b1b2...bn−k形成的逆序,即 b 1 b 2 . . . b n − k b_{1}b_{2}...b_{n-k} b1b2...bn−k中有几个数比 c i c_{i} ci小
a 1 a 2 . . . a k b 1 b 2 . . . b n − k a_{1}a_{2}...a_{k}b_{1}b_{2}...b_{n-k} a1a2...akb1b2...bn−k是由 1 , 2 , 3 , . . . , n 1,2,3,...,n 1,2,3,...,n构成的排列,且 c 1 c 2 . . . c k b 1 b 2 . . . b n − k c_{1}c_{2}...c_{k}b_{1}b_{2}...b_{n-k} c1c2...ckb1b2...bn−k是由 a 1 a 2 . . . a k b 1 b 2 . . . b n − k a_{1}a_{2}...a_{k}b_{1}b_{2}...b_{n-k} a1a2...akb1b2...bn−k经过对换构成的,则 a 1 a 2 . . . a k b 1 b 2 . . . b n − k a_{1}a_{2}...a_{k}b_{1}b_{2}...b_{n-k} a1a2...akb1b2...bn−k也是由 1 , 2 , 3 , . . . , n 1,2,3,...,n 1,2,3,...,n构成的排列
在 12... n 12...n 12...n这个自然序中,总共有 c i − 1 c_{i}-1 ci−1个数比 c i c_{i} ci小的数(比如12345中,有4-1=3个比4小的数,以此类推)
而 c 1 < c 2 < . . . < c i c_{1}<c_{2}<...<c_{i} c1<c2<...<ci,从而 c 1 c 2 . . . c i c_{1}c_{2}...c_{i} c1c2...ci中有 i − 1 i-1 i−1个比 c i c_{i} ci小的数
全部比 c i c_{i} ci小的数 − c i -c_{i} −ci前比 c i c_{i} ci小的数 = c i =c_{i} =ci后比 c i c_{i} ci小的数
即 ( c i − 1 ) − ( i − 1 ) = c i c i + 1 . . . c k (c_{i}-1)-(i-1)=c_{i}c_{i+1}...c_{k} (ci−1)−(i−1)=cici+1...ck中比 c i c_{i} ci小的数 + b 1 b 2 . . . b n − k +b_{1}b_{2}...b_{n-k} +b1b2...bn−k中比 c i c_{i} ci小的数 = c i − i =c_{i}-i =ci−i
考虑完 c 1 c 2 . . . c k c_{1}c_{2}...c_{k} c1c2...ck形成的逆序后,只剩下 b 1 b 2 . . . b n − k b_{1}b_{2}...b_{n-k} b1b2...bn−k本身形成的逆序即 τ ( b 1 b 2 . . . b n − k ) \tau(b_{1}b_{2}...b_{n-k}) τ(b1b2...bn−k)
故 ( − 1 ) τ ( c 1 c 2 . . . c k b 1 b 2 . . . b n − k ) = ( − 1 ) ( c 1 − 1 ) + ( c 2 − 2 ) + . . . + ( c k − k ) + τ ( b 1 b 2 . . . b n − k ) = ( − 1 ) ( c 1 + c 2 + . . . + c k ) − ( 1 + 2 + . . . + k ) + τ ( b 1 b 2 . . . b n − k ) = ( − 1 ) ( c 1 + c 2 + . . . + c k ) − k ( k + 1 ) 2 + τ ( b 1 b 2 . . . b n − k ) (-1)^{\tau(c_{1}c_{2}...c_{k}b_{1}b_{2}...b_{n-k})}=(-1)^{(c_{1}-1)+(c_{2}-2)+...+(c_{k}-k)+\tau(b_{1}b_{2}...b_{n-k})}=(-1)^{(c_{1}+c_{2}+...+c_{k})-(1+2+...+k)+\tau(b_{1}b_{2}...b_{n-k})}=(-1)^{(c_{1}+c_{2}+...+c_{k})-\frac{k(k+1)}{2}+\tau(b_{1}b_{2}...b_{n-k})} (−1)τ(c1c2...ckb1b2...bn−k)=(−1)(c1−1)+(c2−2)+...+(ck−k)+τ(b1b2...bn−k)=(−1)(c1+c2+...+ck)−(1+2+...+k)+τ(b1b2...bn−k)=(−1)(c1+c2+...+ck)−2k(k+1)+τ(b1b2...bn−k)
所以 ( − 1 ) τ ( a 1 a 2 . . . a k b 1 b 2 . . . b n − k ) = ( − 1 ) τ ( c 1 c 2 . . . c k b 1 b 2 . . . b n − k ) + τ ( a 1 a 2 . . . a k ) = ( − 1 ) ( c 1 + c 2 + . . . + c k ) − k ( k + 1 ) 2 + τ ( b 1 b 2 . . . b n − k ) + τ ( a 1 a 2 . . . a k ) (-1)^{\tau(a_{1}a_{2}...a_{k}b_{1}b_{2}...b_{n-k})}=(-1)^{\tau(c_{1}c_{2}...c_{k}b_{1}b_{2}...b_{n-k})+\tau(a_{1}a_{2}...a_{k})}=(-1)^{(c_{1}+c_{2}+...+c_{k})-\frac{k(k+1)}{2}+\tau(b_{1}b_{2}...b_{n-k})+\tau(a_{1}a_{2}...a_{k})} (−1)τ(a1a2...akb1b2...bn−k)=(−1)τ(c1c2...ckb1b2...bn−k)+τ(a1a2...ak)=(−1)(c1+c2+...+ck)−2k(k+1)+τ(b1b2...bn−k)+τ(a1a2...ak)
又 a 1 a 2 . . . a k a_{1}a_{2}...a_{k} a1a2...ak与 c 1 c 2 . . . c k c_{1}c_{2}...c_{k} c1c2...ck的关系只是打乱顺序,其排列里的元素是一样的,所以 c 1 + c 2 + . . . + c k = a 1 + a 2 + . . . + a k c_{1}+c_{2}+...+c_{k}=a_{1}+a_{2}+...+a_{k} c1+c2+...+ck=a1+a2+...+ak由于 k ∈ N + k\in\mathbb{N}^{+} k∈N+,所以 k ( k + 1 ) 2 ≥ 1 , k ( k + 1 ) 2 ∈ N + \frac{k(k+1)}{2}\ge 1,\frac{k(k+1)}{2}\in\mathbb{N}^{+} 2k(k+1)≥1,2k(k+1)∈N+
则 ( − 1 ) k ( k + 1 ) 2 = ( − 1 ) − k ( k + 1 ) 2 (-1)^{\frac{k(k+1)}{2}}=(-1)^{-\frac{k(k+1)}{2}} (−1)2k(k+1)=(−1)−2k(k+1)
于是 ( − 1 ) τ ( a 1 a 2 . . . a k b 1 b 2 . . . b n − k ) = ( − 1 ) τ ( c 1 c 2 . . . c k b 1 b 2 . . . b n − k ) + τ ( a 1 a 2 . . . a k ) = ( − 1 ) ( c 1 + c 2 + . . . + c k ) − k ( k + 1 ) 2 + τ ( b 1 b 2 . . . b n − k ) + τ ( a 1 a 2 . . . a k ) = ( − 1 ) ( a 1 + a 2 + . . . + a k ) + k ( k + 1 ) 2 + τ ( b 1 b 2 . . . b n − k ) + τ ( a 1 a 2 . . . a k ) (-1)^{\tau(a_{1}a_{2}...a_{k}b_{1}b_{2}...b_{n-k})}=(-1)^{\tau(c_{1}c_{2}...c_{k}b_{1}b_{2}...b_{n-k})+\tau(a_{1}a_{2}...a_{k})}=(-1)^{(c_{1}+c_{2}+...+c_{k})-\frac{k(k+1)}{2}+\tau(b_{1}b_{2}...b_{n-k})+\tau(a_{1}a_{2}...a_{k})}=(-1)^{(a_{1}+a_{2}+...+a_{k})+\frac{k(k+1)}{2}+\tau(b_{1}b_{2}...b_{n-k})+\tau(a_{1}a_{2}...a_{k})} (−1)τ(a1a2...akb1b2...bn−k)=(−1)τ(c1c2...ckb1b2...bn−k)+τ(a1a2...ak)=(−1)(c1+c2+...+ck)−2k(k+1)+τ(b1b2...bn−k)+τ(a1a2...ak)=(−1)(a1+a2+...+ak)+2k(k+1)+τ(b1b2...bn−k)+τ(a1a2...ak).
证毕.
-
引理2:设 n n n阶行列式 ∣ A ∣ = ∑ j 1 j 2 … j n ( − 1 ) τ ( j 1 j 2 … j n ) a 1 j 1 a 2 j 2 … a n j n |\boldsymbol{A}|=\sum\limits_{j_{1} j_{2} \ldots j_{n}}(-1)^{\tau\left(j_{1} j_{2} \ldots j_{n}\right)} a_{1 j_{1}} a_{2 j_{2}} \ldots a_{n j_{n}} ∣A∣=j1j2…jn∑(−1)τ(j1j2…jn)a1j1a2j2…anjn(行指标按自然序排好),给定一个 12... n 12...n 12...n组成的排列 i 1 i 2 . . . i n i_{1}i_{2}...i_{n} i1i2...in,则 ∣ A ∣ = ∑ j i 1 j i 2 … j i n ( − 1 ) τ ( j i 1 j i 2 … j i n ) + τ ( i 1 i 2 . . i n ) a i 1 j 1 1 a i 2 j i 2 … a i n j i n |\boldsymbol{A}|=\sum\limits_{j_{i_{1}} j_{i_{2}} \ldots j_{i_{n}}}(-1)^{\tau\left(j_{i_{1}} j_{i_{2}} \ldots j_{i_{n}}\right)+\tau\left(i_{1} i_{2} . . i_{n}\right)} a_{i_{1} j_{1_{1}}} a_{i_{2} j_{i_{2}}} \ldots a_{i_{n} j_{i_{n}}} ∣A∣=ji1ji2…jin∑(−1)τ(ji1ji2…jin)+τ(i1i2..in)ai1j11ai2ji2…ainjin
【证】设 12... n 12...n 12...n经过 s , s ∈ N + s,s\in\mathbb{N}^{+} s,s∈N+次对换变成 i 1 i 2 . . . i n i_{1}i_{2}...i_{n} i1i2...in即 a 1 j 1 a 2 j 2 … a n j n a_{1 j_{1}} a_{2 j_{2}} \ldots a_{n j_{n}} a1j1a2j2…anjn对换为 a i 1 j i 1 a i 2 j i 2 … a i n j i n a_{i_{1} j_{i_{1}}} a_{i_{2} j_{i_{2}}} \ldots a_{i_{n} j_{i_{n}}} ai1ji1ai2ji2…ainjin
由引理0, ( − 1 ) τ ( 12 … n ) + s = ( − 1 ) τ ( i 1 i 2 … i n ) (-1)^{\tau(12 \ldots n)+s}=(-1)^{\tau\left(i_{1} i_{2} \ldots i_{n}\right)} (−1)τ(12…n)+s=(−1)τ(i1i2…in)即 ( − 1 ) 0 + s = ( − 1 ) s = ( − 1 ) τ ( i 1 i 2 … i n ) (-1)^{0+s}=(-1)^{s}=(-1)^{\tau\left(i_{1} i_{2} \ldots i_{n}\right)} (−1)0+s=(−1)s=(−1)τ(i1i2…in)
因为 i i i与 j i j_{i} ji是由 a i j i a_{ij_{i}} aiji绑定在一起的,故 j 1 j 2 . . . j n j_{1}j_{2}...j_{n} j1j2...jn也经过相应的 s s s次对换变为 j i 1 j i 2 . . . j i n j_{i_{1}}j_{i_{2}}...j_{i_{n}} ji1ji2...jin
由引理0, ( − 1 ) τ ( j 1 j 2 … j n ) + s = ( − 1 ) τ ( j i 1 j i 2 … j i n ) (-1)^{\tau\left(j_{1} j_{2} \ldots j_{n}\right)+s}=(-1)^{\tau\left(j_{i_{1}} j_{i_{2}} \ldots j_{i_{n}}\right)} (−1)τ(j1j2…jn)+s=(−1)τ(ji1ji2…jin)
从而 ( − 1 ) τ ( j 1 j 2 … j n ) = ( − 1 ) τ ( j 1 j i 2 … j i n ) + s (-1)^{\tau\left(j_{1} j_{2} \ldots j_{n}\right)}=(-1)^{\tau\left(j_{1} j_{i_{2}} \ldots j_{i_{n}}\right)+s} (−1)τ(j1j2…jn)=(−1)τ(j1ji2…jin)+s
∣ A ∣ = ∑ j 1 j 2 … j n ( − 1 ) τ ( j 1 j 2 … j n ) a 1 j 1 a 2 j 2 … a n j n = ∑ j 1 j 2 … j n ( − 1 ) τ ( j i 1 j i 2 … j i n ) + s a i 1 j i 1 a i 2 j i 2 … a i n j i n |\boldsymbol{A}|=\sum\limits_{j_{1} j_{2} \ldots j_{n}}(-1)^{\tau\left(j_{1} j_{2} \ldots j_{n}\right)} a_{1 j_{1}} a_{2 j_{2}} \ldots a_{n j_{n}}=\sum\limits_{j_{1} j_{2} \ldots j_{n}}(-1)^{\tau\left(j_{i_{1}} j_{i_{2}} \ldots j_{i_{n}}\right)+s} a_{i_{1} j_{i_{1}}} a_{i_{2} j_{i_{2}}} \ldots a_{i_{n} j_{i_{n}}} ∣A∣=j1j2…jn∑(−1)τ(j1j2…jn)a1j1a2j2…anjn=j1j2…jn∑(−1)τ(ji1ji2…jin)+sai1ji1ai2ji2…ainjin
(注意到此时表达式中已经与排列 j 1 j 2 . . . j n j_{1}j_{2}...j_{n} j1j2...jn无直接关系,从而随机选取排列 j 1 j 2 . . . j n j_{1}j_{2}...j_{n} j1j2...jn可直接改为随机选取排列 j i 1 j i 2 … j i n j_{i_{1}} j_{i_{2}} \ldots j_{i_{n}} ji1ji2…jin)
则 ∣ A ∣ = ∑ j 1 j 2 … j n ( − 1 ) τ ( j i 1 j 2 … j i n ) + τ ( i 1 i 2 … i n ) a i i 1 j 1 a i 2 j i 2 … a i n j i n = ∑ j i 1 j i 2 … j i n ( − 1 ) τ ( j i 1 j i 2 … j i n ) + τ ( i 1 i 2 … i n ) a i 1 j i 1 a i 2 j i 2 … a i n j i n |\boldsymbol{A}|=\sum\limits_{j_{1} j_{2} \ldots j_{n}}(-1)^{\tau\left(j_{i_{1}} j_{2} \ldots j_{i_{n}}\right)+\tau\left(i_{1} i_{2} \ldots i_{n}\right)} a_{i_{i_{1} j_{1}}} a_{i_{2} j_{i_{2}}} \ldots a_{i_{n} j_{i_{n}}}=\sum\limits_{j_{i_{1}} j_{i_{2}} \ldots j_{i_{n}}}(-1)^{\tau\left(j_{i_{1}} j_{i_{2}} \ldots j_{i_{n}}\right)+\tau\left(i_{1} i_{2} \ldots i_{n}\right)} a_{i_{1} j_{i_{1}}} a_{i_{2} j_{i_{2}}} \ldots a_{i_{n} j_{i_{n}}} ∣A∣=j1j2…jn∑(−1)τ(ji1j2…jin)+τ(i1i2…in)aii1j1ai2ji2…ainjin=ji1ji2…jin∑(−1)τ(ji1ji2…jin)+τ(i1i2…in)ai1ji1ai2ji2…ainjin
证毕.
下面来证明拉普拉斯定理:
由引理2,选定排列 i 1 . . . i k μ 1 . . . μ n − k i_{1}...i_{k}\mu_{1}...\mu_{n-k} i1...ikμ1...μn−k,且该排列无逆序对,其中 μ 1 … μ n − k = { 1 , 2 , … , n } / { i 1 , i 2 , … , i k } \mu_{1} \ldots \mu_{n-k}=\{1,2, \ldots, n\} /\left\{i_{1}, i_{2}, \ldots, i_{k}\right\} μ1…μn−k={1,2,…,n}/{i1,i2,…,ik}相当于选定 i 1 , i 2 , … , i k i_{1}, i_{2}, \ldots, i_{k} i1,i2,…,ik这些行和其对应的列 j i 1 , j i 2 , . . . , j i n j_{i_{1}},j_{i_{2}},...,j_{i_{n}} ji1,ji2,...,jin,剩下的那些行列,与余子式和 k k k阶子式定义中的取法类似。
则 ∣ A ∣ = ∑ j i 1 … j k j μ 1 … j μ n − k ( − 1 ) τ ( i 1 … i k μ 1 … μ n − k ) + τ ( j i 1 ⋯ j i k j μ 1 ⋯ j μ n − k ) a i 1 j i 1 … a i k j i k a μ 1 j μ 1 … a μ n − k j μ n − k |\boldsymbol{A}|=\sum\limits_{j_{i_{1} \ldots j_{k}} j_{\mu_{1} \ldots j_{\mu_{n-k}}}}(-1)^{\tau\left(i_{1} \ldots i_{k} \mu_{1} \ldots \mu_{n-k}\right)+\tau\left(j_{i_{1}} \cdots j_{i_{k}} j_{\mu_{1}} \cdots j_{\mu_{n-k}}\right)} a_{i_{1} j_{i_{1}}} \ldots a_{i_{k} j_{i_{k}}} a_{\mu_{1} j_{\mu_{1}}} \ldots a_{\mu_{n-k} j_{\mu_{n-k}}} ∣A∣=ji1…jkjμ1…jμn−k∑(−1)τ(i1…ikμ1…μn−k)+τ(ji1⋯jikjμ1⋯jμn−k)ai1ji1…aikjikaμ1jμ1…aμn−kjμn−k
由引理1, ∣ A ∣ = ∑ j i 1 … j i k j μ 1 … j μ n − k ( − 1 ) τ ( i 1 … i k ) + τ ( μ 1 … μ n − k ) + k ( k + 1 ) 2 + ( i 1 + … + i k ) + ( τ ( j i 1 … j i k ) + τ ( j μ 1 … j μ n − k ) + k ( k + 1 ) 2 + ( j i 1 + … + j i k ) ) a i 1 j i 1 … a i k j j k a μ 1 j μ 1 … a μ n − k j μ n − k = ∑ i 1 … … i i , j n … j n ( − 1 ) ( i 1 + … + i k ) + ( j i 1 + … + j k k ) + τ ( j i 1 … j k k ) + τ ( j μ 1 … j μ n − k ) a i 1 j i 1 … a i k j i k a μ 1 j μ 1 … a μ n − k j n n − k = ∑ j i 1 … j i k j μ 1 … j μ n − k ( − 1 ) ( i 1 + … + i k ) + ( j i 1 + … + j i k ) + τ ( j i 1 … j i k ) + τ ( j μ 1 … j μ n − k ) a i 1 j i 1 … a i k j i k a μ 1 j μ 1 … a μ n − k j μ n − k |\boldsymbol{A}|=\sum\limits_{j_{i_{1}} \ldots j_{i_{k}} j_{\mu_{1}} \ldots j_{\mu_{n-k}}}(-1)^{\tau\left(i_{1} \ldots i_{k}\right)+\tau\left(\mu_{1} \ldots \mu_{n-k}\right)+\frac{k(k+1)}{2}+\left(i_{1}+\ldots+i_{k}\right)+\left(\tau\left(j_{i_{1}} \ldots j_{i_{k}}\right)+\tau\left(j_{\mu_{1}} \ldots j_{\mu_{n-k}}\right)+\frac{k(k+1)}{2}+\left(j_{i_{1}}+\ldots+j_{i_{k}}\right)\right)} a_{i_{1} j_{i_{1}}} \ldots a_{i_{k} j_{j_{k}}} a_{\mu_{1} j_{\mu_{1}}} \ldots a_{\mu_{n-k} j_{\mu_{n-k}}}=\sum\limits_{i_{1} \ldots \ldots i_{i}, j_{n} \ldots j_{n}}(-1)^{\left(i_{1}+\ldots+i_{k}\right)+\left(j_{i_{1}}+\ldots+j_{k_{k}}\right)+\tau\left(j_{i_{1}} \ldots j_{k_{k}}\right)+\tau\left(j_{\mu_{1}} \ldots j_{\mu_{n-k}}\right)} a_{i_{1} j_{i_{1}}} \ldots a_{i_{k} j_{i_{k}}} a_{\mu_{1} j_{\mu_{1}}} \ldots a_{\mu_{n-k} j_{n_{n-k}}}=\sum\limits_{j_{i_{1}} \ldots j_{i_{k}} j_{\mu_{1}} \ldots j_{\mu_{n-k}}}(-1)^{\left(i_{1}+\ldots+i_{k}\right)+\left(j_{i_{1}}+\ldots+j_{i_{k}}\right)+\tau\left(j_{i_{1}} \ldots j_{i_{k}}\right)+\tau\left(j_{\mu_{1}} \ldots j_{\mu_{n-k}}\right)} a_{i_{1} j_{i_{1}}} \ldots a_{i_{k} j_{i_{k}}} a_{\mu_{1} j_{\mu_{1}}} \ldots a_{\mu_{n-k} j_{\mu_{n-k}}} ∣A∣=ji1…jikjμ1…jμn−k∑(−1)τ(i1…ik)+τ(μ1…μn−k)+2k(k+1)+(i1+…+ik)+(τ(ji1…jik)+τ(jμ1…jμn−k)+2k(k+1)+(ji1+…+jik))ai1ji1…aikjjkaμ1jμ1…aμn−kjμn−k=i1……ii,jn…jn∑(−1)(i1+…+ik)+(ji1+…+jkk)+τ(ji1…jkk)+τ(jμ1…jμn−k)ai1ji1…aikjikaμ1jμ1…aμn−kjnn−k=ji1…jikjμ1…jμn−k∑(−1)(i1+…+ik)+(ji1+…+jik)+τ(ji1…jik)+τ(jμ1…jμn−k)ai1ji1…aikjikaμ1jμ1…aμn−kjμn−k
(上面这个式子是这么来的:由于排列 i 1 . . . i k μ 1 . . . μ n − k i_{1}...i_{k}\mu_{1}...\mu_{n-k} i1...ikμ1...μn−k无逆序对,则 τ ( i 1 … i k ) = 0 , τ ( μ 1 … μ n − k ) = 0 \tau\left(i_{1} \ldots i_{k}\right)=0,\tau\left(\mu_{1} \ldots \mu_{n-k}\right)=0 τ(i1…ik)=0,τ(μ1…μn−k)=0,又 k ( k + 1 ) 2 × 2 = k ( k + 1 ) = k 2 + k \frac{k(k+1)}{2}\times 2=k(k+1)=k^{2}+k 2k(k+1)×2=k(k+1)=k2+k,偶数+偶数=偶数,奇数+奇数=偶数,奇数的平方是奇数,偶数的平方是偶数,由这些规律,假如 k k k是奇数,则 k 2 + k k^{2}+k k2+k是偶数,若 k k k是偶数,则 k 2 + k k^{2}+k k2+k还是偶数,所以 ( − 1 ) k ( k + 1 ) 2 × 2 = 1 (-1)^{\frac{k(k+1)}{2}\times 2}=1 (−1)2k(k+1)×2=1)
然后对 n ! n! n!个项进行分组,分组方式(按从小到大)如下:
任取第 j 1 , j 2 , . . . , j k j_{1},j_{2},...,j_{k} j1,j2,...,jk,且 1 ≤ j 1 < … < j k ≤ n 1 \leq j_{1}<\ldots<j_{k} \leq n 1≤j1<…<jk≤n
则其对应的 n n n元排列有: η 1 … η k v 1 … v n − k \eta_{1} \ldots \eta_{k} v_{1} \ldots v_{n-k} η1…ηkv1…vn−k
其中 η 1 … η k \eta_{1} \ldots \eta_{k} η1…ηk是 j 1 . . . j k j_{1}...j_{k} j1...jk形成的排列, v 1 … v n − k = { 1 , 2 , … , n } / { j 1 , j 2 , … , j k } v_{1} \ldots v_{n-k}=\{1,2, \ldots, n\} /\left\{j_{1}, j_{2}, \ldots, j_{k}\right\} v1…vn−k={1,2,…,n}/{j1,j2,…,jk}
则每一个 j 1 , j 2 , . . . , j k j_{1},j_{2},...,j_{k} j1,j2,...,jk的选择都对应到所有前 k k k个元由 j 1 , . . . , j k j_{1},...,j_{k} j1,...,jk组成的排列,一共可以选取 C n k C_{n}^{k} Cnk组(相当于从 n n n个格子里选数字,取法共有 C n k C_{n}^{k} Cnk组,取完以后再从小到大排列)
于是 ∣ A ∣ = ∑ 1 ≤ j 1 < … < j k ≤ n ∑ η 1 … η k ∑ v 1 … v n − k ( − 1 ) ( i 1 + … + i k ) + ( j 1 + … + j k ) + τ ( j 1 1 … j i k ) + τ ( j μ 1 ⋯ j μ n − k ) a i 1 j i 1 … a i k j k a μ 1 j μ 1 … a μ n − k j μ n − k = |\boldsymbol{A}|=\sum\limits_{1 \leq j_{1}<\ldots<j_{k} \leq n} \sum\limits_{\eta_{1} \ldots \eta_{k}} \sum\limits_{v_{1} \ldots v_{n-k}}(-1)^{\left(i_{1}+\ldots+i_{k}\right)+\left(j_{1}+\ldots+j_{k}\right)+\tau\left(j_{1_{1}} \ldots j_{i_{k}}\right)+\tau\left(j_{\mu_{1}} \cdots j_{\mu_{n-k}}\right)} a_{i_{1} j_{i_{1}}} \ldots a_{i_{k} j_{k}} a_{\mu_{1} j_{\mu_{1}}} \ldots a_{\mu_{n-k}} j_{\mu_{n-k}}= ∣A∣=1≤j1<…<jk≤n∑η1…ηk∑v1…vn−k∑(−1)(i1+…+ik)+(j1+…+jk)+τ(j11…jik)+τ(jμ1⋯jμn−k)ai1ji1…aikjkaμ1jμ1…aμn−kjμn−k=
j i 1 , . . . , j i k j_{i_{1}},...,j_{i_{k}} ji1,...,jik与 j μ 1 , . . . , j μ n − k j_{\mu_{1}},...,j_{\mu_{n-k}} jμ1,...,jμn−k本就是随机取的,只要取遍 n ! n! n!个不同的排列即可,这与 η 1 … η k \eta_{1} \ldots \eta_{k} η1…ηk和 v 1 . . . v n − k v_{1}...v_{n-k} v1...vn−k相对应,只在于符号不同
则 ∣ A ∣ = ∑ 1 ≤ j 1 < … < j k ≤ n ∑ η 1 … η k ∑ v 1 … v n − k ( − 1 ) ( i 1 + … + i k ) + ( j 1 + … + j k ) + τ ( η 1 … η k ) + τ ( v 1 … v n − k ) a i 1 η 1 … a i k η k a μ 1 v 1 … a μ n − k v n − k = ∑ 1 ≤ j 1 < … < j k ≤ n ( ∑ η 1 … η k ( − 1 ) τ ( η 1 … η k ) a i 1 η 1 … a i k η k ) ( − 1 ) ( i 1 + … + i k ) + ( j 1 + … + j k ) ( ∑ v 1 … v n − k ( − 1 ) τ ( v 1 … v n − k ) a μ 1 v 1 … a μ n − k v n − k ) = ∑ 1 ≤ j 1 < … < j k ≤ n A [ i 1 … i k j 1 … j k ] ( − 1 ) ( i 1 + … + i k ) + ( j 1 + … + j k ) A [ i 1 … i k j 1 … j k ] ′ |\boldsymbol{A}|=\begin{array}{l} \sum\limits_{1 \leq j_{1}<\ldots<j_{k} \leq n} \sum\limits_{\eta_{1} \ldots \eta_{k}} \sum\limits_{v_{1} \ldots v_{n-k}}(-1)^{\left(i_{1}+\ldots+i_{k}\right)+\left(j_{1}+\ldots+j_{k}\right)+\tau\left(\eta_{1} \ldots \eta_{k}\right)+\tau\left(v_{1} \ldots v_{n-k}\right)} a_{i_{1} \eta_{1}} \ldots a_{i_{k} \eta_{k}} a_{\mu_{1} v_{1}} \ldots a_{\mu_{n-k} v_{n-k}} \\ =\sum\limits_{1 \leq j_{1}<\ldots<j_{k} \leq n}\left(\sum\limits_{\eta_{1} \ldots \eta_{k}}(-1)^{\tau\left(\eta_{1} \ldots \eta_{k}\right)} a_{i_{1} \eta_{1}} \ldots a_{i_{k} \eta_{k}}\right)(-1)^{\left(i_{1}+\ldots+i_{k}\right)+\left(j_{1}+\ldots+j_{k}\right)}\left(\sum\limits_{v_{1} \ldots v_{n-k}}(-1)^{\tau\left(v_{1} \ldots v_{n-k}\right)} a_{\mu_{1} v_{1}} \ldots a_{\mu_{n-k} v_{n-k}}\right) \\ =\sum\limits_{1 \leq j_{1}<\ldots<j_{k} \leq n} \boldsymbol{A}\left[\begin{array}{lll} i_{1} & \ldots & i_{k} \\ j_{1} & \ldots & j_{k} \end{array}\right](-1)^{\left(i_{1}+\ldots+i_{k}\right)+\left(j_{1}+\ldots+j_{k}\right)} \boldsymbol{A}\left[\begin{array}{ccc} i_{1} & \ldots & i_{k} \\ j_{1} & \ldots & j_{k} \end{array}\right]^{\prime} \end{array} ∣A∣=1≤j1<…<jk≤n∑η1…ηk∑v1…vn−k∑(−1)(i1+…+ik)+(j1+…+jk)+τ(η1…ηk)+τ(v1…vn−k)ai1η1…aikηkaμ1v1…aμn−kvn−k=1≤j1<…<jk≤n∑(η1…ηk∑(−1)τ(η1…ηk)ai1η1…aikηk)(−1)(i1+…+ik)+(j1+…+jk)(v1…vn−k∑(−1)τ(v1…vn−k)aμ1v1…aμn−kvn−k)=1≤j1<…<jk≤n∑A[i1j1……ikjk](−1)(i1+…+ik)+(j1+…+jk)A[i1j1……ikjk]′
证毕
我超了,这玩意我看了一天才看懂,证着证着就忘了前面要证明什么了,证明这玩意使我受到了极大的精神折磨,很不好的体验,恨来自非数专业人。
【推论】 ∣ a 11 ⋯ a 1 k 0 ⋯ 0 ⋮ ⋮ ⋮ ⋮ a k 1 ⋯ a k k 0 ⋯ 0 c 11 ⋯ c 1 k b 11 ⋯ b 1 r ⋮ ⋮ ⋮ ⋮ c r 1 ⋯ c r k b r 1 ⋯ b r r ∣ = ∣ a 11 ⋯ a 1 k ⋮ ⋮ a k 1 ⋯ a b k ∣ ⋅ ∣ b 11 ⋯ b 1 r ⋮ ⋮ b r 1 ⋯ b r r ∣ \left|\begin{array}{cccccc} a_{11} & \cdots & a_{1 k} & 0 & \cdots & 0 \\ \vdots & & \vdots & \vdots & & \vdots \\ a_{k 1} & \cdots & a_{k k} & 0 & \cdots & 0 \\ c_{11} & \cdots & c_{1 k} & b_{11} & \cdots & b_{1 r} \\ \vdots & & \vdots & \vdots & & \vdots \\ c_{r 1} & \cdots & c_{r k} & b_{r 1} & \cdots & b_{r r} \end{array}\right|=\left|\begin{array}{ccc} a_{11} & \cdots & a_{1 k} \\ \vdots & & \vdots \\ a_{k 1} & \cdots & a_{b k} \end{array}\right| \cdot\left|\begin{array}{ccc} b_{11} & \cdots & b_{1 r} \\ \vdots & & \vdots \\ b_{r 1} & \cdots & b_{r r} \end{array}\right| a11⋮ak1c11⋮cr1⋯⋯⋯⋯a1k⋮akkc1k⋮crk0⋮0b11⋮br1⋯⋯⋯⋯0⋮0b1r⋮brr = a11⋮ak1⋯⋯a1k⋮abk ⋅ b11⋮br1⋯⋯b1r⋮brr
【证】按前 k k k行展开,只有左上角的 k k k阶子式 ∣ a 11 ⋯ a 1 k ⋮ ⋮ a k 1 ⋯ a b k ∣ \left|\begin{array}{ccc} a_{11} & \cdots & a_{1 k} \\ \vdots & & \vdots \\ a_{k 1} & \cdots & a_{b k} \end{array}\right| a11⋮ak1⋯⋯a1k⋮abk 不为零,其余 k k k阶子式肯定包含0列,从而其值为0,左上角的 k k k阶子式的余子式恰好是右下角的子式 ∣ b 11 ⋯ b 1 r ⋮ ⋮ b r 1 ⋯ b r r ∣ \left|\begin{array}{ccc} b_{11} & \cdots & b_{1 r} \\ \vdots & & \vdots \\ b_{r 1} & \cdots & b_{r r} \end{array}\right| b11⋮br1⋯⋯b1r⋮brr ,并且左上角的 k k k阶子式的代数余子式的系数为 ( − 1 ) ( 1 + 2 + ⋯ + k ) + ( 1 + 2 + ⋯ + k ) = 1 (-1)^{(1+2+\cdots+k)+(1+2+\cdots+k)}=1 (−1)(1+2+⋯+k)+(1+2+⋯+k)=1,由拉普拉斯定理可知 ∣ a 11 ⋯ a 1 k 0 ⋯ 0 ⋮ ⋮ ⋮ ⋮ a k 1 ⋯ a k k 0 ⋯ 0 c 11 ⋯ c 1 k b 11 ⋯ b 1 r ⋮ ⋮ ⋮ ⋮ c r 1 ⋯ c r k b r 1 ⋯ b r r ∣ = ∣ a 11 ⋯ a 1 k ⋮ ⋮ a k 1 ⋯ a b k ∣ ⋅ ∣ b 11 ⋯ b 1 r ⋮ ⋮ b r 1 ⋯ b r r ∣ \left|\begin{array}{cccccc} a_{11} & \cdots & a_{1 k} & 0 & \cdots & 0 \\ \vdots & & \vdots & \vdots & & \vdots \\ a_{k 1} & \cdots & a_{k k} & 0 & \cdots & 0 \\ c_{11} & \cdots & c_{1 k} & b_{11} & \cdots & b_{1 r} \\ \vdots & & \vdots & \vdots & & \vdots \\ c_{r 1} & \cdots & c_{r k} & b_{r 1} & \cdots & b_{r r} \end{array}\right|=\left|\begin{array}{ccc} a_{11} & \cdots & a_{1 k} \\ \vdots & & \vdots \\ a_{k 1} & \cdots & a_{b k} \end{array}\right| \cdot\left|\begin{array}{ccc} b_{11} & \cdots & b_{1 r} \\ \vdots & & \vdots \\ b_{r 1} & \cdots & b_{r r} \end{array}\right| a11⋮ak1c11⋮cr1⋯⋯⋯⋯a1k⋮akkc1k⋮crk0⋮0b11⋮br1⋯⋯⋯⋯0⋮0b1r⋮brr = a11⋮ak1⋯⋯a1k⋮abk ⋅ b11⋮br1⋯⋯b1r⋮brr
令:
A = ( a 11 ⋯ a 1 k ⋮ ⋮ a k 1 ⋯ a b k ) , B = ( b 11 ⋯ b 1 r ⋮ ⋮ b r 1 ⋯ b r ) , C = ( c 11 ⋯ c 1 k ⋮ ⋮ c r 1 ⋯ c r k ) , 0 = ( 0 ⋯ 0 ⋮ ⋮ 0 ⋯ 0 ) , \begin{array}{l} \boldsymbol{A}=\left(\begin{array}{ccc} a_{11} & \cdots & a_{1 k} \\ \vdots & & \vdots \\ a_{k 1} & \cdots & a_{b k} \end{array}\right), \boldsymbol{B}=\left(\begin{array}{ccc} b_{11} & \cdots & b_{1 r} \\ \vdots & & \vdots \\ b_{r 1} & \cdots & b_{r} \end{array}\right), \\ \boldsymbol{C}=\left(\begin{array}{ccc} c_{11} & \cdots & c_{1 k} \\ \vdots & & \vdots \\ c_{r 1} & \cdots & c_{r k} \end{array}\right), \mathbf{0}=\left(\begin{array}{ccc} 0 & \cdots & 0 \\ \vdots & & \vdots \\ 0 & \cdots & 0 \end{array}\right), \end{array} A= a11⋮ak1⋯⋯a1k⋮abk ,B= b11⋮br1⋯⋯b1r⋮br ,C= c11⋮cr1⋯⋯c1k⋮crk ,0= 0⋮0⋯⋯0⋮0 ,
则此推论简写为:
∣ A 0 C B ∣ = ∣ A ∣ ∣ B ∣ \left|\begin{array}{ll} \boldsymbol{A} & \mathbf{0} \\ \boldsymbol{C} & \boldsymbol{B} \end{array}\right|=|\boldsymbol{A}||\boldsymbol{B}| AC0B =∣A∣∣B∣
这篇关于【高等代数笔记】(18)N阶行列式的文章就介绍到这儿,希望我们推荐的文章对编程师们有所帮助!