Loading [MathJax]/jax/output/HTML-CSS/jax.js
Skip to the content.

Finite-Dimensional Vector Spaces

Section 2.A. Span and Linear Independence

向量的组的记号,一般不在最外面加括号,比如:(4,1,6),(9,5,7)

Linear Combinations and Span

向量组 v1,,vm 的一个线性组合是形式为 a1v1++amvma1,,amF)的一个向量。

V 中一个向量组 v1,,vm张成空间(span,或者 linear span)是该向量组所有线性组合的集合,记为 span(v1,,vm)

​ 规定空组的张成空间为 0

张成空间是包含这组向量的最小子空间。

证明:

Suppose v1,,vm is a vector list in V.

First to prove: span(v1,,vm) is a subspace.

$0 = 0v_1+…+0v_m\in span(v_1,…,v_m)$

(a1v1++amvm)+(c1v1++cmvm)=(a1+c1)v1++(am+cm)vm, so span(v1,,vm) is closed under addition.

λ(a1v1++amvm)=λa1v1++λamvm, so span(v1,,vm) is closed under scalar multiplication.

Thus span(v!,,vm) is a subspace of V.

Then to prove that is smallest.

As each vj is a linear combination of v1,,vm, we have vjspan(v1,,vm).

As subspaces are closed under addition and scalar multiplication, so each subspace containing all v1,,vm also contains span(v1,,vm).

Thus span(v!,,vm) is the smallest containing subspace.

如果 span(v1,,vm) 等价于 V,那么我们说 v1,vm 张成(spans)V

如果一个向量空间可以由该空间内的某个向量组张成,那么该向量空间是有限维的。(根据定义,组的长度是有限的)

形如 p(z)=a0+a1z+a2z2++amzm 的函数称为多项式

所有系数都属于 F 的所有多项式的集合记为 P(F)

对于多项式 pP(F),如果存在标量 a1,,amFam0) 使得对于任意的 zF 都有:

p(z)=a0+a1z++amzm

那么我们说 p次数m,记作 deg p=m

对于非负整数 m,所有系数在 F 中且次数不超过 m 的多项式的集合记作 Pm(F)

如果一个向量空间不是有限维的,则称为无限维的

证明:P(F) 是无限维向量空间。

Consider any vector list in P(F), suppose it has a maximum degree of m, so the span of this vector list also have a maximum degree of m, thus p(z)=zm+1 is not in the span, so P(F) cannot been spanned from a vector list in it, so it is not finite-dimensional, so it is infinite-dimensional.

Linear Independence

对于 V 中的一个向量组 v1,,vm,如果使 a1v1++amvm0 的唯一方式是 a1==am=0,那么称该向量组线性无关

​ 规定空组线性无关。

V 中的一个向量组如果不是线性无关的,那么该向量组是线性相关的。

线性相关性引理(2.21):如果从 v1,,vm 是一个线性相关的向量组,那么存在 j1,,m,使得:

证明:

Since v1,,vm is linear dependent, there exist a list (a1,,am), which is not equal to (0,…,0), such that:

a1v1++amvm=0

Let j be the largest number in 1,,m such that aj0.

Then we can write: a1v1++aj1vj1=ajvj, then, vj=a1ajv1aj1ajvj1. (as equation 2.22)

Thus vjspan(v1,,vj1)

For any uspan(v1,,vm), we can write as u=c1v1++cmvm.

If we remove vj from the vector list, we can still rewrite u using the equation 2.22 without involving vj. So u is still in span(v1,,vm).

(此证明没有考虑 j 取 1 时的特例)

(2.23)在有限维向量空间中,线性无关组的长度小于等于向量空间的张成组的长度。

证明:

Suppose u1,,un is a linear independent list in V, and w1,,wm spans V. We need to prove nm.

Let list B be the list w1,,wm.

Step 1:

​ Add u1 in the beginning of B, we get B=u1,w1,,wm, which is linearly independent as u1 is a linear combination of the rest vectors.

​ According to 2.21, we can find a vector wi in B, with removing that vector, span(B) remains the same, equals to V.

Step j:

​ As the B list produced in the step j1 spans V, we add vj in the beginning of B and get vj,vj1,,v1,, which is a linearly dependent list.

​ According to 2.21, we can find a vector in B, with removing that vector, span(B) remains the same, equals to V. As vj,vj1,,v1 is a linearly independent list, the removing vector in B must be a vector o w.

As 2.21 imply, at each step j, we can successfully add a vj and remove a w. At last, we have B with at most m vs in it. So nm.

(2.26)所有有限维向量空间的子空间也是有限维的。

证明:

Suppose U is a subspace of V.

Step 1:

​ If U=0, it’s finite-dimensional, we are done. If not, we choose v1U and form a linear independent list v1.

Step j:

​ If U=span(v1,,vj1), it’s finite-dimensional, we are done. If not, we choose vjU but vjspan(v1,,vj1), add to the list.

As this steps not done, we constructed a list of vectors such that, no vector is in the span of the previous vectors. (2.21) According to 2.23, the length of the current constructed list of vectors is less than or equal to the length of the spanning vector list of V, so these steps can terminate at last.

Section 2.B. Bases

如果 V 中的一个向量组既线性无关又张成 V,则称为 V

V 中的一个向量组 v1,,vnV 的基当且仅当 V 中的每个向量 v 都可以被唯一表示为 v=a1v1++anvnaiF)。(2.30)

证明:

在向量空间中,每个张成组都可以化简为一个基。(2.31)

证明:

每个有限维向量空间都有基。(2.32)

证明:

在有限维向量空间中,每个线性无关向量组都可以扩充为一组基。(2.33)

证明:

对于有限维向量空间 V,如果 UV 的子空间,那么存在 V 的子空间 W 使得 V=UW。(2.34)

证明:

Section 2.C. Dimension

有限维向量空间的任意基的长度相同。(2.35)

证明:

有限维向量空间 V维度定义为该向量空间的基的长度,记作 dim V

如果 V 是有限维向量空间,UV 的子空间,那么 dim Udim V。(2.38)

证明:

在有限维向量空间 V 中,所有长度为 dim V 的线性无关向量组都是 V 的基。(2.39)

证明:

在有限维向量空间 V 中,所有长度为 dim V 的张成组都是 V 的基。(2.42)

证明:

如果 U1U2 是一个有限维向量空间的两个子空间,那么:dim(U1+U2)=dim U1+dim U2dim(U1U2)。(2.43)

证明: