问题
在实数域上线性空间 C [ − π , π ] C[-\pi,\pi] C[−π,π]内判断下列向量组是否线性相关,并求它的秩: 1 , s i n x , s i n 2 x , s i n 3 x , ⋯ , s i n n x 1,sinx,sin2x,sin3x,\cdots,sinnx 1,sinx,sin2x,sin3x,⋯,sinnx
引理
设
f
i
(
x
)
f_i(x)
fi(x)是数域
K
K
K上的
i
i
i次多项式,其首项系数为
a
i
(
i
=
0
,
1
,
2
,
⋯
,
n
−
1
)
a_i(i=0,1,2,\cdots,n-1)
ai(i=0,1,2,⋯,n−1).又设
b
1
,
b
2
,
⋯
,
b
n
b_1,b_2,\cdots,b_n
b1,b2,⋯,bn是
K
K
K内一组两两互不相同的数。则下列
n
n
n阶矩阵:
[
f
0
(
b
1
)
f
0
(
b
2
)
⋯
f
0
(
b
n
)
f
1
(
b
1
)
f
1
(
b
2
)
⋯
f
1
(
b
n
)
⋮
⋮
⋮
f
n
−
1
(
b
1
)
f
n
−
1
(
b
2
)
⋯
f
n
−
1
(
b
n
)
]
\begin{bmatrix} f_0(b_1) & f_0(b_2) & \cdots & f_0(b_n) \\ f_1(b_1) & f_1(b_2) & \cdots & f_1(b_n) \\ \vdots & \vdots & & \vdots \\ f_{n-1}(b_1) & f_{n-1}(b_2) & \cdots & f_{n-1}(b_n) \\ \end{bmatrix}
⎣⎢⎢⎢⎡f0(b1)f1(b1)⋮fn−1(b1)f0(b2)f1(b2)⋮fn−1(b2)⋯⋯⋯f0(bn)f1(bn)⋮fn−1(bn)⎦⎥⎥⎥⎤
满秩。
引理证明
记
f
i
(
x
)
=
∑
k
=
0
n
−
1
a
i
k
x
i
−
k
f_i(x)=\sum_{k=0}^{n-1}a_{ik}x^{i-k}
fi(x)=∑k=0n−1aikxi−k,其中,
a
i
k
∈
K
,
a
i
0
=
a
i
a_{ik}\in K, a_{i0}=a_i
aik∈K,ai0=ai且
i
<
k
i<k
i<k时,
a
i
k
=
0
a_{ik}=0
aik=0。规定
f
i
(
0
)
=
0
f_i(0)=0
fi(0)=0。
有
∣
f
0
(
b
1
)
f
0
(
b
2
)
⋯
f
0
(
b
n
)
f
1
(
b
1
)
f
1
(
b
2
)
⋯
f
1
(
b
n
)
⋮
⋮
⋮
f
n
−
1
(
b
1
)
f
n
−
1
(
b
2
)
⋯
f
n
−
1
(
b
n
)
∣
=
∣
∑
k
0
=
0
n
−
1
a
0
k
0
b
1
0
−
k
0
∑
k
0
=
0
n
−
1
a
0
k
0
b
2
0
−
k
0
⋯
∑
k
0
=
0
n
−
1
a
0
k
0
b
n
0
−
k
0
∑
k
1
=
0
n
−
1
a
1
k
1
b
1
1
−
k
1
∑
k
1
=
0
n
−
1
a
1
k
1
b
2
1
−
k
1
⋯
∑
k
1
=
0
n
−
1
a
1
k
1
b
n
1
−
k
1
⋮
⋮
⋮
∑
k
n
−
1
=
0
n
−
1
a
n
−
1
k
n
−
1
b
1
n
−
1
−
k
n
−
1
∑
k
n
−
1
=
0
n
−
1
a
n
−
1
k
n
−
1
b
2
n
−
1
−
k
n
−
1
⋯
∑
k
n
−
1
=
0
n
−
1
a
n
−
1
k
n
−
1
b
n
n
−
1
−
k
n
−
1
∣
=
∑
k
0
=
0
n
−
1
∑
k
1
=
0
n
−
1
⋯
∑
k
n
−
1
=
0
n
−
1
∣
a
0
k
0
b
1
0
−
k
0
a
0
k
0
b
2
0
−
k
0
⋯
a
0
k
0
b
n
0
−
k
0
a
1
k
1
b
1
1
−
k
1
a
1
k
1
b
2
1
−
k
1
⋯
a
1
k
1
b
n
1
−
k
1
⋮
⋮
⋮
a
n
−
1
k
n
−
1
b
1
n
−
1
−
k
n
−
1
a
n
−
1
k
n
−
1
b
2
n
−
1
−
k
n
−
1
⋯
a
n
−
1
k
n
−
1
b
n
n
−
1
−
k
n
−
1
∣
=
∣
a
0
a
0
⋯
a
0
a
1
b
1
a
1
b
2
⋯
a
1
b
n
a
2
b
1
2
a
2
b
2
2
⋯
a
2
b
n
2
⋮
⋮
⋮
a
n
−
1
b
1
n
−
1
a
n
−
1
b
2
n
−
1
⋯
a
n
−
1
b
n
n
−
1
∣
=
∏
i
=
0
n
−
1
a
i
∣
1
1
⋯
1
b
1
b
2
⋯
b
n
b
1
2
b
2
2
⋯
b
n
2
⋮
⋮
⋮
b
1
n
−
1
b
2
n
−
1
⋯
b
n
n
−
1
∣
=
∏
i
=
0
n
−
1
a
i
∏
1
≤
j
<
k
≤
n
(
b
k
−
b
j
)
\begin{vmatrix} f_0(b_1) & f_0(b_2) & \cdots & f_0(b_n) \\ f_1(b_1) & f_1(b_2) & \cdots & f_1(b_n) \\ \vdots & \vdots & & \vdots \\ f_{n-1}(b_1) & f_{n-1}(b_2) & \cdots & f_{n-1}(b_n) \\ \end{vmatrix}\\ =\begin{vmatrix} \sum_{k_0=0}^{n-1}a_{0k_0}b_1^{0-k_0} & \sum_{k_0=0}^{n-1}a_{0k_0}b_2^{0-k_0} & \cdots & \sum_{k_0=0}^{n-1}a_{0k_0}b_n^{0-k_0} \\ \sum_{k_1=0}^{n-1}a_{1k_1}b_1^{1-k_1} & \sum_{k_1=0}^{n-1}a_{1k_1}b_2^{1-k_1}& \cdots &\sum_{k_1=0}^{n-1}a_{1k_1}b_n^{1-k_1} \\ \vdots & \vdots & & \vdots \\ \sum_{k_{n-1}=0}^{n-1}a_{n-1k_{n-1}}b_1^{n-1-k_{n-1}} & \sum_{k_{n-1}=0}^{n-1}a_{n-1k_{n-1}}b_2^{n-1-k_{n-1}} & \cdots & \sum_{k_{n-1}=0}^{n-1}a_{n-1k_{n-1}}b_n^{n-1-k_{n-1}} \\ \end{vmatrix}\\ =\sum_{k_0=0}^{n-1}\sum_{k_1=0}^{n-1}\cdots\sum_{k_{n-1}=0}^{n-1}\begin{vmatrix} a_{0k_0}b_1^{0-k_0} & a_{0k_0}b_2^{0-k_0} & \cdots & a_{0k_0}b_n^{0-k_0} \\ a_{1k_1}b_1^{1-k_1} &a_{1k_1}b_2^{1-k_1}& \cdots &a_{1k_1}b_n^{1-k_1} \\ \vdots & \vdots & & \vdots \\ a_{n-1k_{n-1}}b_1^{n-1-k_{n-1}} &a_{n-1k_{n-1}}b_2^{n-1-k_{n-1}} & \cdots & a_{n-1k_{n-1}}b_n^{n-1-k_{n-1}} \\ \end{vmatrix}\\ =\begin{vmatrix} a_0 & a_0 & \cdots & a_0\\ a_1b_1 & a_1b_2 & \cdots & a_1b_n\\ a_2b_1^2 & a_2b_2^2 & \cdots & a_2b_n^2\\ \vdots & \vdots & & \vdots\\ a_{n-1}b_1^{n-1} & a_{n-1}b_2^{n-1} & \cdots & a_{n-1}b_n^{n-1} \end{vmatrix}\\ =\prod_{i=0}^{n-1}a_i\begin{vmatrix} 1 & 1 & \cdots & 1\\ b_1 & b_2 & \cdots & b_n\\ b_1^2 & b_2^2 & \cdots & b_n^2\\ \vdots & \vdots & & \vdots\\ b_1^{n-1} & b_2^{n-1} & \cdots & b_n^{n-1} \end{vmatrix}\\ =\prod_{i=0}^{n-1}a_i\prod_{1\le j\lt k\le n}(b_k-b_j)
∣∣∣∣∣∣∣∣∣f0(b1)f1(b1)⋮fn−1(b1)f0(b2)f1(b2)⋮fn−1(b2)⋯⋯⋯f0(bn)f1(bn)⋮fn−1(bn)∣∣∣∣∣∣∣∣∣=∣∣∣∣∣∣∣∣∣∣∑k0=0n−1a0k0b10−k0∑k1=0n−1a1k1b11−k1⋮∑kn−1=0n−1an−1kn−1b1n−1−kn−1∑k0=0n−1a0k0b20−k0∑k1=0n−1a1k1b21−k1⋮∑kn−1=0n−1an−1kn−1b2n−1−kn−1⋯⋯⋯∑k0=0n−1a0k0bn0−k0∑k1=0n−1a1k1bn1−k1⋮∑kn−1=0n−1an−1kn−1bnn−1−kn−1∣∣∣∣∣∣∣∣∣∣=k0=0∑n−1k1=0∑n−1⋯kn−1=0∑n−1∣∣∣∣∣∣∣∣∣a0k0b10−k0a1k1b11−k1⋮an−1kn−1b1n−1−kn−1a0k0b20−k0a1k1b21−k1⋮an−1kn−1b2n−1−kn−1⋯⋯⋯a0k0bn0−k0a1k1bn1−k1⋮an−1kn−1bnn−1−kn−1∣∣∣∣∣∣∣∣∣=∣∣∣∣∣∣∣∣∣∣∣a0a1b1a2b12⋮an−1b1n−1a0a1b2a2b22⋮an−1b2n−1⋯⋯⋯⋯a0a1bna2bn2⋮an−1bnn−1∣∣∣∣∣∣∣∣∣∣∣=i=0∏n−1ai∣∣∣∣∣∣∣∣∣∣∣1b1b12⋮b1n−11b2b22⋮b2n−1⋯⋯⋯⋯1bnbn2⋮bnn−1∣∣∣∣∣∣∣∣∣∣∣=i=0∏n−1ai1≤j<k≤n∏(bk−bj)
由题
a
i
a_i
ai为多项式的首项系数,故不为零。又
b
j
≠
b
k
b_j\neq b_k
bj=bk,所以上式不为0,引理得证。
上述证明用到了行列式的行线性。在对每一行根据多项式的不同次项分拆行列式时,如果该行留下的不是最高次项,不妨记该行为第i行,那么该行列式的前i行中最多有(i-1)个不同次项,即必有两行留下了相同次项,将这两行的系数提出后会得到相同的两行,即该拆分后的行列式为0.
解答
判断实数域上线性空间中向量组是否线性相关,即判断是否存在一组不全为零的实数 k 0 , k 1 , k 2 , ⋯ , k n k_0,k_1,k_2,\cdots,k_n k0,k1,k2,⋯,kn,满足 k 0 + ∑ i = 1 n k i s i n i x = 0 k_0 + \sum_{i=1}^{n}k_isinix=0 k0+∑i=1nkisinix=0,记为(1)式,该式对 ∀ x ∈ [ − π , π ] \forall x\in[-\pi,\pi] ∀x∈[−π,π]都成立。
将x=0代入即得 k 0 = 0 k_0=0 k0=0。于是只要判断对 ∀ x ∈ [ − π , π ] \forall x\in[-\pi,\pi] ∀x∈[−π,π],是否存在一组不全为零的实数 k 1 , k 2 , ⋯ , k n k_1,k_2,\cdots,k_n k1,k2,⋯,kn,满足 ∑ i = 1 n k i s i n i x = 0 \sum_{i=1}^{n}k_isinix=0 ∑i=1nkisinix=0。
不难找到 [ − π , π ] [-\pi,\pi] [−π,π]上(n+1)个数 x 1 , x 2 , ⋯ , x n + 1 x_1,x_2,\cdots,x_{n+1} x1,x2,⋯,xn+1,使得 i ≠ j ( i , j = 1 , 2 , ⋯ , n + 1 ) i\neq j(i,j=1,2,\cdots,n+1) i=j(i,j=1,2,⋯,n+1)时, s i n x i ≠ s i n x j sinx_i\neq sinx_j sinxi=sinxj。
将上述(n+1)个数分别代入(1)式,即得一个实数域上由(n+1)个方程组成的(n+1)元线性齐次方程组, k 1 , k 2 , ⋯ , k n + 1 k_1,k_2,\cdots,k_{n+1} k1,k2,⋯,kn+1为其变元。该方程组的系数矩阵为 A = [ 1 s i n x 1 s i n 2 x 1 ⋯ s i n n x 1 1 s i n x 2 s i n 2 x 2 ⋯ s i n n x 2 ⋮ ⋮ ⋮ ⋮ 1 s i n x n + 1 s i n 2 x n + 1 ⋯ s i n n x n + 1 ] A=\begin{bmatrix} 1 & sinx_1 & sin2x_1 & \cdots & sinnx_1\\ 1 & sinx_2 & sin2x_2 & \cdots & sinnx_2\\ \vdots & \vdots & \vdots & & \vdots\\ 1 & sinx_{n+1} & sin2x_{n+1}& \cdots & sinnx_{n+1}\\ \end{bmatrix} A=⎣⎢⎢⎢⎡11⋮1sinx1sinx2⋮sinxn+1sin2x1sin2x2⋮sin2xn+1⋯⋯⋯sinnx1sinnx2⋮sinnxn+1⎦⎥⎥⎥⎤由于上述线性齐次方程组有非零解等价于矩阵A不满秩,从而问题转变为研究矩阵A是否满秩。
注意到
s
i
n
n
x
=
(
c
o
s
x
+
i
s
i
n
x
)
n
−
(
c
o
s
x
−
i
s
i
n
x
)
n
2
i
=
∑
k
=
0
n
(
c
o
s
x
)
n
−
k
[
(
i
s
i
n
x
)
k
−
(
−
i
s
i
n
x
)
k
]
2
i
=
−
i
∑
k
=
0
2
∤
k
n
(
c
o
s
x
)
n
−
k
(
i
s
i
n
x
)
k
,
sinnx=\frac {(cosx+isinx)^n-(cosx-isinx)^n}{2i}\\=\frac{\sum_{k=0}^n(cosx)^{n-k}[(isinx)^k-(-isinx)^k]}{2i}\\ =-i\sum_{\begin{array}{c} k=0\\ 2\not|k \end{array}}^n(cosx)^{n-k}(isinx)^k,
sinnx=2i(cosx+isinx)n−(cosx−isinx)n=2i∑k=0n(cosx)n−k[(isinx)k−(−isinx)k]=−ik=02∣k∑n(cosx)n−k(isinx)k,当n为奇数时,
(
n
−
k
)
(n-k)
(n−k)为偶数,
(
c
o
s
x
)
n
−
k
=
[
1
−
(
s
i
n
x
)
2
]
n
−
k
2
(cosx)^{n-k}=[1-(sinx)^2]^{\frac{n-k}2}
(cosx)n−k=[1−(sinx)2]2n−k,sinnx为复数域上关于sinx的n次多项式。
从而可令
f
i
(
x
)
=
s
i
n
i
x
f_i(x)=sinix
fi(x)=sinix,令
f
0
(
x
)
=
1
f_0(x)=1
f0(x)=1,则
A
′
=
[
f
0
(
x
1
)
f
0
(
x
2
)
⋯
f
0
(
x
n
+
1
)
f
1
(
x
1
)
f
1
(
x
2
)
⋯
f
1
(
x
n
+
1
)
⋮
⋮
⋮
f
n
(
x
1
)
f
n
(
x
2
)
⋯
f
n
(
x
n
+
1
)
]
,
A'=\begin{bmatrix} f_0(x_1) & f_0(x_2) & \cdots & f_0(x_{n+1}) \\ f_1(x_1) & f_1(x_2) & \cdots & f_1(x_{n+1}) \\ \vdots & \vdots & & \vdots \\ f_n(x_1) & f_n(x_2) & \cdots & f_n(x_{n+1}) \\ \end{bmatrix},
A′=⎣⎢⎢⎢⎡f0(x1)f1(x1)⋮fn(x1)f0(x2)f1(x2)⋮fn(x2)⋯⋯⋯f0(xn+1)f1(xn+1)⋮fn(xn+1)⎦⎥⎥⎥⎤,A’表示A的转置。
由引理可知A满秩。
从而向量组
1
,
s
i
n
x
,
s
i
n
2
x
,
s
i
n
3
x
,
⋯
,
s
i
n
n
x
1,sinx,sin2x,sin3x,\cdots,sinnx
1,sinx,sin2x,sin3x,⋯,sinnx线性无关。
当n为偶数时,(n+1)为奇数,向量组
1
,
s
i
n
x
,
s
i
n
2
x
,
s
i
n
3
x
,
⋯
,
s
i
n
(
n
+
1
)
x
1,sinx,sin2x,sin3x,\cdots,sin(n+1)x
1,sinx,sin2x,sin3x,⋯,sin(n+1)x线性无关,从而向量组
1
,
s
i
n
x
,
s
i
n
2
x
,
s
i
n
3
x
,
⋯
,
s
i
n
n
x
1,sinx,sin2x,sin3x,\cdots,sinnx
1,sinx,sin2x,sin3x,⋯,sinnx线性无关。
所以对任意正整数n,向量组
1
,
s
i
n
x
,
s
i
n
2
x
,
s
i
n
3
x
,
⋯
,
s
i
n
n
x
1,sinx,sin2x,sin3x,\cdots,sinnx
1,sinx,sin2x,sin3x,⋯,sinnx线性无关。
该线性无关向量组包含(n+1)个向量,所以它的秩为(n+1)。
出处
本题出自蓝以中《高等代数简明教程》第二版上册第四章习题一第6题(5)。引理出自同一本书第三章习题一第27题。