目录

Linear Algebra Done Right习题的参考解答及思考(第1章)

前言

本书风格类似陶哲轩Analysis I的风格,在引入一个东西前,都会给你解释动机,这对读者来说是好事,你不用盲目记一个东西或者反复读一段内容只为明白作者的动机是什么、为什么要这样做,也很容易自己串起来整个理论的体系结构,等等。

本文的习题是第3版的习题,仅供参考。

由于所有习题都放一篇文章的话,渲染太卡了,故分成多篇文章。

第1章

练习1.A

练习1.A.1

题目:

Suppose \( a \) and \( b \) are real numbers, not both \( 0 \). Find real numbers \( c \) and \( d \) such that \( 1 / (a + bi) = c + di \).

解答:

令\( c := a / (a^2 + b^2), d := -b / (a^2 + b^2) \),可得\( (a + bi)(c + di) = 1 \),根据复数除法的定义,有\( 1 / (a + bi) = 1 * (c + di) = c + di \)。

练习1.A.2

题目:

Show that \( \dfrac{-1 + \sqrt{3}i}{2} \) is a cube root of \( 1 \) (meaning that its cube equals \( 1 \)).

证明:

\( (\dfrac{-1 + \sqrt{3}i}{2})^2 = \dfrac{-1 - \sqrt{3}i}{2} \),故 \( (\dfrac{-1 + \sqrt{3}i}{2})^3 = \dfrac{-1 - \sqrt{3}i}{2} \times \dfrac{-1 + \sqrt{3}i}{2} = 1 \)。

证毕。

练习1.A.3

题目:

Find two distinct square roots of \( i \).

解答:

设该复数为\( a + bi \),其中\( a \in \mathbf{R}, b \in \mathbf{R} \), \( (a + bi)^2 = i \),得\( (a^2 - b^2) + 2abi = i \),从而可得\( a^2 - b^2 = 0 \)且\( 2ab = 1 \),由\( 2ab = 1 \)得到\( b = \dfrac{1}{2a} \)代入\( a^2 - b^2 = 0 \),得到\( a = \pm \dfrac{\sqrt{2}}{2} \),回代到\( b = \dfrac{1}{2a} \),得到两个平方为\( i \)的复数:\( \dfrac{\sqrt{2}}{2} + \dfrac{\sqrt{2}}{2}i \) 以及\( -\dfrac{\sqrt{2}}{2} - \dfrac{\sqrt{2}}{2}i \)。

练习1.A.4

题目:

Show that \( \alpha + \beta = \beta + \alpha \) for all \( \alpha, \beta \in \mathbf{C} \).

证明:

\( \forall \alpha, \beta \in \mathbf{C} \),有\( \exists a, b, c, d \in \mathbf{R} \),使得\( \alpha = a + bi, \beta = c + di \),可得\( \alpha + \beta = (a + c) + (b + d)i = (c + a) + (d + b)i = \beta + \alpha \)。

证毕。

练习1.A.5

题目:

Show that \( (\alpha + \beta) + \lambda = \alpha + (\beta + \lambda) \) forall \( \alpha, \beta, \alpha \in \mathbf{C} \).

证明:

\( \forall \alpha, \beta, \lambda \in \mathbf{C} \),有\( \exists a, b, c, d, e, f \in \mathbf{R} \),使得\( \alpha = a + bi, \beta = c + di, \lambda = e + fi \),可得\( (\alpha + \beta) + \lambda = ((a + c) + e) + ((b + d) + f)i = (a + (c + e)) + (b + (d + f))i = \alpha + (\beta + \lambda) \)。

证毕。

练习1.A.6

题目:

Show that \( (\alpha\beta)\lambda = \alpha(\beta\lambda) \) for all \( \alpha, \beta, \alpha \in \mathbf{C} \).

证明:

\( \forall \alpha, \beta, \lambda \in \mathbf{C} \),有\( \exists a, b, c, d, e, f \in \mathbf{R} \),使得\( \alpha = a + bi, \beta = c + di, \lambda = e + fi \)。

\( (\alpha\beta)\lambda = ((ac - bd) + (ad + bc)i)(e + fi) = ((ac - bd)e - (ad + bc)f)) + ((ac - bd)f + (ad + bc)e)i = ((ace - bde) - (adf + bcf)) + ((acf - bdf) + (ade + bce))i = (ace - bde - adf - bcf) + (acf - bdf + ade + bce)i \)。

\( \alpha(\beta\lambda) = (a + bi)((ce - df) + (cf + de)i) = (a(ce - df) - b(cf + de)) + (a(cf + de) + b(ce - df))i = ((ace - adf) - (bcf + bde)) + ((acf + ade) + (bce - bdf))i = (ace - adf - bcf - bde) + (acf + ade + bce - bdf)i = (\alpha\beta)\lambda \)。

证毕。

练习1.A.7

题目:

Show that for every \( \alpha \in \mathbf{C} \), there exists a unique \( \beta \in \mathbf{C} \) such that \( \alpha + \beta = 0 \).

证明:

\( \forall \alpha \in \mathbf{C} \),有\( \exists a, b \in \mathbf{R} \),使得\( \alpha = a + bi \),此时令\( \beta := -a + (-b)i \),有\( \alpha + \beta = 0 \)。

假设\( \exists \beta' \in \mathbf{C}, \alpha + \beta' = 0 \),则\( \beta = \beta + 0 = \beta + (\alpha + \beta') = (\beta + \alpha) + \beta' = 0 + \beta' = \beta' \),也就是\( \beta = \beta' \),即\( \beta \)是唯一的。

证毕。

练习1.A.8

题目:

Show that for every \( \alpha \in \mathbf{C} \) with \( \alpha \neq 0 \), there exists a unique \( \beta \in \mathbf{C} \) such that \( \alpha\beta = 1 \).

证明:

\( \forall \alpha \in \mathbf{C} \),有\( \exists a, b \in \mathbf{R} \),使得\( \alpha = a + bi \),此时令\( \beta := \dfrac{a}{a^2 + b^2} + \dfrac{-b}{a^2 + b^2}i \),可得\( \alpha\beta = 1 \)。

假设\( \exists \beta' \in \mathbf{C}, \alpha\beta' = 1 \),则\( \beta = \beta1 = \beta(\alpha\beta') = (\beta\alpha)\beta' = 1\beta' = \beta' \),也就是\( \beta = \beta' \),即\( \beta \)是唯一的。

证毕。

练习1.A.9

题目:

Show that \( \lambda(\alpha + \beta) = \lambda\beta + \lambda\alpha \) for all \( \lambda, \alpha, \beta \in \mathbf{C} \).

证明:

\( \forall \alpha, \beta, \lambda \in \mathbf{C} \),有\( \exists a, b, c, d, e, f \in \mathbf{R} \),使得\( \alpha = a + bi, \beta = c + di, \lambda = e + fi \)。

\( \lambda(\alpha + \beta) = (e + fi)((a + c) + (b + d)i) = (e(a + c) - f(b + d)) + (e(b + d) + f(a + c))i = (ea + ec - fb - fd) + (eb + ed + fa + fc)i \)。

\( \lambda\beta + \lambda\alpha = (e + fi)(c + di) + (e + fi)(a + bi) = ((ec - fd) + (ed + fc)i) + ((ea - fb) + (eb + fa)i) = (ec - fd + ea - fb) + (ed + fc + eb + fa)i = \lambda(\alpha + \beta) \)。

证毕。

练习1.A.10

题目:

Find \( x \in \mathbf{R}^4 \) such that \( (4, -3, 1, 7) + 2x = (5, 9, -6, 8) \).

解答:

设\( x := (x_1, x_2, x_3, x_4) \in \mathbf{R}^4 \),则 \( (4, -3, 1, 7) + 2x = (4 + 2x_1, -3 + 2x_2, 1 + 2x_3, 7 + 2x_4) = (5, 9, -6, 8) \),解得\( x = (\dfrac{1}{2}, 6, -\dfrac{7}{2}, \dfrac{1}{2}) \)。

练习1.A.11

题目:

Explain why there does not exist \( \lambda \in \mathbf{C} \) such that \( \lambda(2 - 3i, 5 + 4i, -6 + 7i) = (12 - 5i, 7 + 22i, -32 - 9i) \).

证明:

\( \lambda(2 - 3i, 5 + 4i, -6 + 7i) = (\lambda(2 - 3i), \lambda(5 + 4i), \lambda(-6 + 7i)) \),如果\( (\lambda(2 - 3i), \lambda(5 + 4i), \lambda(-6 + 7i)) = (12 - 5i, 7 + 22i, -32 - 9i) \),则列表的每个坐标分别相等,从\( \lambda(2 - 3i) = 12 - 5i \),解得\( \lambda = 3 + 2i \),从\( \lambda(-6 + 7i) = -32 - 9i \),解得\( \lambda = \dfrac{129}{85} + \dfrac{278}{85}i \),这和\( \lambda = 3 + 2i \)矛盾,故两者无法同时满足,即满足条件的\( \lambda \)不存在。

证毕。

练习1.A.12

题目:

Show that \( (x + y) + z = x + (y + z) \) for all \( x, y, z \in \mathbf{F}^n \).

证明:

记\( x = (x_1, x_2, \dots, x_n), y = (y_1, y_2, \dots, y_n), z = (z_1, z_2, \dots, z_n) \),则\( (x + y) + z = ((x_1 + y_1) + z_1, (x_2 + y_2) + z_2, \dots , (x_n + y_n) + z_n) = (x_1 + (y_1 + z_1), x_2 + (y_2 + z_2), \dots, x_n + (y_n + z_n)) = x + (y + z) \)。

证毕。

练习1.A.13

题目:

Show that \( (ab)x = a(bx) \) for all \( x \in \mathbf{F}^n \) and all \( a, b \in \mathbf{F} \).

证明:

记\( x = (x_1, x_2, \dots, x_n) \),则\( (ab)x = (ab)(x_1, x_2, \dots, x_n) = ((ab)x_1, (ab)x_2, \dots, (ab)x_n) = (a(bx_1), a(bx_2), \dots, a(bx_n)) = a(bx) \)。

证毕。

练习1.A.14

题目:

Show that \( 1x = x \) for all \( x \in \mathbf{F}^n \).

证明:

记\( x = (x_1, x_2, \dots, x_n) \),则\( 1x = 1(x_1, x_2, \dots, x_n) = (1 \times x_1, 1 \times x_2, \dots, 1 \times x_n) = (x_1, x_2, \dots, x_n) = x \)。

证毕。

练习1.A.15

题目:

Show that \( \lambda(x + y) = \lambda x + \lambda y \) for all \( \lambda \in \mathbf{F} \) and all \( x, y \in \mathbf{F}^n \).

证明:

记\( x = (x_1, x_2, \dots, x_n), y = (y_1, y_2, \dots, y_n) \),则\( \lambda(x + y) = \lambda(x_1 + y_1, x_2 + y_2, \dots, x_n + y_n) = (\lambda(x_1 + y_1), \lambda(x_2 + y_2), \dots, \lambda(x_n + y_n)) = (\lambda x_1 + \lambda y_1, \lambda x_2 + \lambda y_2, \dots, \lambda x_n + \lambda y_n)) = (\lambda x_1, \lambda x_2, \dots, \lambda x_n) + (\lambda y_1, \lambda y_2, \dots, \lambda y_n) = \lambda x + \lambda y \)。

证毕。

练习1.A.16

题目:

Show that \( (a + b)x = ax + bx \) for all \( a, b \in \mathbf{F} \) and all \( x \in \mathbf{F}^n \).

证明:

记\( x = (x_1, x_2, \dots, x_n) \),则\( (a + b)x = ((a + b)x_1, (a + b)x_2, \dots, (a + b)x_n) = (ax_1 + bx_1, ax_2 + bx_2, \dots, ax_n + bx_n) = (ax_1, ax_2, \dots, ax_n) + (bx_1, bx_2, \dots, bx_n) = ax + bx \)。

证毕。

练习1.B

练习1.B.1

题目:

Prove that \( -(-v) = v \) for every \( v \in V \).

证明:

\( (-v) + v = v + (-v) = 0 \),故\( v \)就是\( -v \)的加法逆元,即\( -(-v) = v \)。

证毕。

练习1.B.2

题目:

Support \( a \in \mathbf{F}, v \in V \), and \( av = 0 \). Prove that \( a = 0 \) or \( v = 0 \).

证明:

假设\( a \neq 0 \)且\( v \neq 0 \),则\( \dfrac{1}{a} \in \mathbf{F} \),\( av = 0 \)两边同乘\( \dfrac{1}{a} \),得\( v = \dfrac{1}{a}0 = 0 \),这和\( v \neq 0 \)矛盾,故假设不成立,有\( a = 0 \)或\( v = 0 \)。

证毕。

练习1.B.3

题目:

Support \( v, w \in V \). Explain why there exists a unique \( x \in V \) such that \( v + 3x = w \).

证明:

先证明存在性:\( v + 3x = w \),两边同加\( -v \),得\( 3x = -v + w \),再同乘\( \dfrac{1}{3} \),得\( x = \dfrac{1}{3}(-v + w) \in V \)。

再证唯一性:假设存在\( x, x' \)满足\( v + 3x = w, v + 3x' = w \),后式减前式得 \( 3(x' - x) = 0 \),根据练习1.B.2,有\( 3 = 0 \)或\( x' - x = 0 \),又\( 3 \neq 0 \),故\( x' - x = 0 \),即\( x' = x \)。

证毕。

练习1.B.4

题目:

The empty set is not a vector space. The empty set fails to satisfy only one of the requirements listed in 1.19. Which one?

解答:

空集不满足加法单位元存在的那条。

其他需求都有“for all”修饰,而空集中不包含任何元素,故这些需求直接满足了。

练习1.B.5

题目:

Show that in the definition of a vector space (1.19), the additive inverse condition can be replaced with the condition that \( 0v = 0 \) for all \( v \in V \) Here the \( 0 \) on the left side is the number \( 0 \), and the \( 0 \) on the right side is the additive identity of \( V \). (The phrase “a condition can be replaced” in a definition means that the collection of objects satisfying the definition is unchanged if the original condition is replaced with the new condition.)

证明:

必要性:

如果\( \forall v \in V \),均有\( 0v = 0 \),则\( \forall v \in V \), \( v + (-1)v = 1v + (-1)v = (1 + (-1))v = 0v = 0 \),这里\( (-1)v \in V \)就是\( v \)的加法逆元。

充分性:

见本书对1.29的证明。

证毕。

练习1.B.6

题目:

Let \( \infty \) and \( -\infty \) denote two distinct objects, neither of which is in \( \mathbf{R} \). Define an addition and scalar multiplication on \( \mathbf{R} \cup \{ \infty \} \cup \{ -\infty \} \) as you could guess from the notation. Specifically, the sum and product of two real numbers is as usual, and for \( t \in \mathbf{R} \) define

\[ t \infty = \begin{cases} -\infty, \text{ if } t < 0 \\ 0, \text{ if } t = 0 \\ \infty, \text{ if } t > 0 \end{cases}, \]

\[ t (-\infty) = \begin{cases} \infty, \text{ if } t < 0 \\ 0, \text{ if } t = 0 \\ -\infty, \text{ if } t > 0 \end{cases}, \]

\[ t + \infty = \infty + t = \infty, t + (-\infty) = (-\infty) + t = -\infty, \]

\[ \infty + \infty = \infty, (-\infty) + (-\infty) = -\infty, \infty + (-\infty) = 0. \]

Is \( R \cup \{ \infty \} \cup \{ -\infty \} \) a vector space over \( \mathbf{R} \)? Explain.

解答:

需要注意的是:题目说的是\( \mathbf{R} \)里面的元素作为数, \( R \cup \{ \infty \} \cup \{ -\infty \} \)里面的元素作为对象,故数乘的数不用考虑\( \{ \infty, -\infty \} \)。

\( R \cup \{ \infty \} \cup \{ -\infty \} \)不是向量空间,理由如下:

不满足加法交换律:只定义了\( \infty + (-\infty) = 0 \),没有定义\( (-\infty) + \infty = 0 \)。

不满足加法结合律:\( (t + \infty) + (-\infty) \neq t + (\infty + (-\infty)) \)。

不满足分配律:\( \infty = (2 + (-1))\infty = 2\infty + (-1)\infty = \infty + (-\infty) = 0 \),推出\( \infty = 0 \),但此时有\( \forall t_1 \in R, t_1 + 0 = t_1 + \infty = \infty = 0 \),即每个实数都是零向量,零向量不是唯一的,这与1.25矛盾。

练习1.C

练习1.C.1

题目:

For each of the following subsets of \( \mathbf{F}^3 \), determine whether it is a subspace of \( \mathbf{F}^3 \):

  1. \( \{ (x_1, x_2, x_3) \in \mathbf{F}^3 : x_1 + 2x_2 + 3x_3 = 0 \} \);
  2. \( \{ (x_1, x_2, x_3) \in \mathbf{F}^3 : x_1 + 2x_2 + 3x_3 = 4 \} \);
  3. \( \{ (x_1, x_2, x_3) \in \mathbf{F}^3 : x_1x_2x_3 = 0 \} \);
  4. \( \{ (x_1, x_2, x_3) \in \mathbf{F}^3 : x_1 = 5x_3 \} \);

证明1:

令\( U := \{ (x_1, x_2, x_3) \in \mathbf{F}^3 : x_1 + 2x_2 + 3x_3 = 0 \} \)。

\( 0 = (0, 0, 0) \in \mathbf{F}^3 \),且\( 0 + 2 \times 0 + 3 \times 0 = 0 \),即\( 0 \in U \)。

\( \forall (x_1, x_2, x_3), (y_1, y_2, y_3) \in U \),有\( x_1 + 2x_2 + 3x_3 = 0, y_1 + 2y_2 + 3y_2 = 0 \),两式相加可得\( (x_1 + y_1) + 2(x_2 + y_2) + 3(x_3 + y_3) = 0 \),即\( (x_1, x_2, x_3) + (y_1, y_2, y_3) \in U \)。

\( \forall (x_1, x_2, x_3) \in U, \forall \lambda \in \mathbf{F} \),有\( \lambda(x_1, x_2, x_3) = (\lambda x_1, \lambda x_2, \lambda x_3) \),因为\( x_1 + 2x_2 + 3x_3 = 0 \),故\( \lambda x_1 + 2 \lambda x_2 + 3 \lambda x_3 = \lambda(x_1 + 2x_2 + 3x_3) = \lambda 0 = 0 \),即\( \lambda(x_1, x_2, x_3) \in U \)。

根据1.34,有\( U \)是\( \mathbf{F}^3 \)的子空间。

证毕。

证明2:

令\( U := \{ (x_1, x_2, x_3) \in \mathbf{F}^3 : x_1 + 2x_2 + 3x_3 = 4 \} \),

\( \forall (x_1, x_2, x_3), (y_1, y_2, y_3) \in U \),有\( x_1 + 2x_2 + 3x_3 = 4, y_1 + 2y_2 + 3y_2 = 4 \),两式相加可得\( (x_1 + y_1) + 2(x_2 + y_2) + 3(x_3 + y_3) = 8 \neq 4 \),即\( (x_1, x_2, x_3) + (y_1, y_2, y_3) \notin U \)。

\( \forall (x_1, x_2, x_3) \in U, \forall \lambda \neq 1 \in \mathbf{F} \),有\( \lambda(x_1, x_2, x_3) = (\lambda x_1, \lambda x_2, \lambda x_3) \),因为\( x_1 + 2x_2 + 3x_3 = 4 \),故\( \lambda x_1 + 2 \lambda x_2 + 3 \lambda x_3 = \lambda(x_1 + 2x_2 + 3x_3) = \lambda 4 \neq 4 \),即\( \lambda(x_1, x_2, x_3) \notin U \)。

根据1.34,有\( U \)不是\( \mathbf{F}^3 \)的子空间。

证毕。

证明3:

令\( U := \{ (x_1, x_2, x_3) \in \mathbf{F}^3 : x_1x_2x_3 = 0 \} \), \( \forall x_1 \neq 0, x_2 \neq 0, y_1 \neq 0 \in \mathbf{F} \),有 \( (0, x_1, x_2) \in U, (y_1, 0, 0) \in U \),进而有\( (0, x_1, x_2) + (y_1, 0, 0) = (y_1, x_1, x_2), y_1x_1x_2 \neq 0 \),故\( (0, x_1, x_2) + (y_1, 0, 0) \notin U \),根据1.34,有\( U \)不是\( \mathbf{F}^3 \)的子空间。

证毕。

证明4:

令\( U := \{ (x_1, x_2, x_3) \in \mathbf{F}^3 : x_1 = 5x_3 \} \)。

\( 0 = (0, 0, 0) \in \mathbf{F}^3 \),且\( 0 = 5 \times 0 \),即\( 0 \in U \)。

\( \forall (x_1, x_2, x_3), (y_1, y_2, y_3) \in U \),有\( x_1 = 5x_3, y_1 = 5y_3 \),两式相加可得\( x_1 + y_1 = 5(x_3 + y_3) \),即\( (x_1, x_2, x_3) + (y_1, y_2, y_3) \in U \)。

\( \forall (x_1, x_2, x_3) \in U, \forall \lambda \in \mathbf{F} \),有\( \lambda(x_1, x_2, x_3) = (\lambda x_1, \lambda x_2, \lambda x_3) \),因为\( x_1 = 5x_3 \),故\( \lambda x_1 = 5 \lambda x_3 \),即\( \lambda(x_1, x_2, x_3) \in U \)。

根据1.34,有\( U \)是\( \mathbf{F}^3 \)的子空间。

证毕。

练习1.C.2

题目:

Verify all the assertions in Example 1.35.

Example 1.35的内容:

  1. If \( b \in \mathbf{F} \), then \( (x_1, x_2, x_3, x_4) \in \mathbf{F}^4 : x_3 = 5x_4 + b \) is a subspace of \( \mathbf{F}^4 \) if and only if b = 0.
  2. The set of continuous real-valued functions on the interval \( [0, 1] \) is a subspace of \( \mathbf{R}^{[0, 1]} \).
  3. The set of differentiable real-valued functions on \( \mathbf{R} \) is a subspace of \( \mathbf{R}^{\mathbf{R}} \).
  4. The set of differentiable real-valued functions \( f \) on the interval \( (0, 3) \) such that \( f'(2) = b \) is a subspace of \( \mathbf{R}^{(0, 3)} \) if and only if \( b = 0 \).
  5. The set of all sequences of complex numbers with limit \( 0 \) is a subspace of \( \mathbf{C}^{\infty} \).

证明1:

记\( \{ (x_1, x_2, x_3, x_4) \in \mathbf{F}^4 : x_3 = 5x_4 + b \} \)为\( U \)。

必要性:

如果\( U \)为\( \mathbf{F}^4 \)的子空间,则\( (0, 0, 0, 0) \in U \),可得\( 0 = 5 \times 0 + b = 0 + b = b \),即\( b = 0 \)。

充分性:

如果\( b = 0 \),则\( x_3 = 5x_4 \),下面验证\( U \)是\( \mathbf{F}^4 \)的子空间:

\( 0 = (0, 0, 0, 0) \in \mathbf{F}^4 \),有\( 0 = 5 \times 0 \),即\( 0 \in U \)。

\( \forall (x_1, x_2, x_3, x_4), (y_1, y_2, y_3, y_4) \in U \),有\( x_1 = 5x_4, y_1 = 5y_4 \),可得\( x_1 + y_1 = 5(x_4 + y_4) \),即\( (x_1, x_2, x_3, x_4) + (y_1, y_2, y_3, y_4) \in U \)。

\( \forall (x_1, x_2, x_3, x_4) \in U, \forall \lambda \in \mathbf{F} \),有\( x_1 = 5x_4 \),而\( \lambda(x_1, x_2, x_3, x_4) = (\lambda x_1, \lambda x_2, \lambda x_3, \lambda x_4) \),由\( x_1 = 5x_4 \)可得\( \lambda x_1 = 5 \lambda x_4 \),即\( \lambda(x_1, x_2, x_3, x_4) \in U \)。

根据1.34,有\( U \)是\( \mathbf{F}^4 \)的子空间。

证毕。

证明2:

记在\( [0, 1] \)上连续的实值函数的集合为\( U \),有\( U \subseteq \mathbf{R}^{[0, 1]} \)。

构造函数\( f_0: [0, 1] \to \mathbf{R}, \forall x \in [0, 1], f_0(x) = 0 \),有\( f_0 \)为在\( [0, 1] \)上连续的实值函数,即\( f_0 \in U \)。 \( \forall f \in U \),可得\( f + f_0 = f \),即\( 0 = f_0 \in U \)。

\( \forall f, g \in U \),有\( f + g \)为在\( [0, 1] \)上连续的实值函数,即\( f + g \in U \)。

\( \forall f \in U, \forall \lambda \in \mathbf{R} \),有\( \lambda f \)为在\( [0, 1] \)上连续的实值函数,即\( \lambda f \in U \)。

根据1.34,有\( U \)是\( \mathbf{R}^{[0, 1]} \)的子空间。

证毕。

证明3:

记在\( \mathbf{R} \)上可微的实值函数的集合为\( U \),有\( U \subseteq \mathbf{R}^{\mathbf{R}} \)。

构造函数\( f_0: \mathbf{R} \to \mathbf{R}, \forall x \in \mathbf{R}, f_0(x) = 0 \),有\( f_0 \)为在\( \mathbf{R} \)上可微的实值函数,即\( f_0 \in U \)。 \( \forall f \in U \),可得\( f + f_0 = f \),即\( 0 = f_0 \in U \)。

\( \forall f, g \in U \),有\( f + g \)为在\( \mathbf{R} \)上可微的实值函数,即\( f + g \in U \)。

\( \forall f \in U, \forall \lambda \in \mathbf{R} \),有\( \lambda f \)为在\( \mathbf{R} \)上可微的实值函数,即\( \lambda f \in U \)。

根据1.34,有\( U \)为\( \mathbf{R}^{\mathbf{R}} \)子空间。

证毕。

证明4:

记在\( (0, 3) \)上可微的实值函数的集合为\( A \),记集合\( \{ f \in A : f'(2) = b \} \)为\( U \)。

必要性:

如果\( U \)是\( \mathbf{R}^{(0, 3)} \)的子空间,则有\( 0 \in U \),而\( 0'(2) = 0 = b \),得\( b = 0 \)。

充分性:

如果\( b = 0 \),下面验证\( U \)为\( \mathbf{R}^{(0, 3)} \)的子空间。

构造函数\( f_0: (0, 3) \to \mathbf{R}, \forall x \in (0, 3), f_0(x) = 0 \),可得\( f_0 \)为在\( (0, 3) \)上可微的实值函数,且 \( f_0'(2) = 0 \),即\( f_0 \in U \)。 \( \forall f \in U \),有\( f + f_0 = f \),即\( 0 = f_0 \in U \)。

\( \forall f, g \in U \),有\( f + g \)为在\( (0, 3) \)上可微的实值函数,且\( (f + g)'(2) = f'(2) + g'(2) = 0 + 0 = 0 \),即 \( (f + g) \in U \)。

\( \forall f \in U, \forall \lambda \in \mathbf{R} \),有\( \lambda f \)为在\( (0, 3) \)上可微的实值函数,且\( (\lambda f)'(2) = \lambda f'(2) = 0 \),即\( \lambda f \in U \)。

根据1.34,有\( U \)为\( \mathbf{R}^{(0, 3)} \)的子空间。

证毕。

证明5:

记集合\( \{ (a_n)_{n = 1}^{\infty} \in \mathbf{C}^{\infty} : \lim_{n \to \infty} a_n = 0 \} \)为\( A \),有\( A \subseteq \mathbf{C}^{\infty} \)。

令序列\( (z_n)_{n = 1}^{\infty} := (0)_{n = 1}^{\infty} \),有\( \forall (a_n)_{n = 1}^{\infty} \in \mathbf{C}^{\infty} \),有\( (a_n)_{n = 1}^{\infty} + (z_n)_{n = 1}^{\infty} = (a_n)_{n = 1}^{\infty} \),即序列\( (z_n)_{n = 1}^{\infty} \)是加法单位元,且我们还有\( \lim_{n \to \infty} z_n = 0 \),即\( (z_n)_{n = 1}^{\infty} \in A \)。

\( \forall (a_n)_{n = 1}^{\infty}, (b_n)_{n = 1}^{\infty} \in \mathbf{C}^{\infty} \),有\( (a_n)_{n = 1}^{\infty} + (b_n)_{n = 1}^{\infty} \in \mathbf{C}^{\infty} \),且\( \lim_{n \to \infty} (a_n + b_n) = \lim_{n \to \infty} a_n + \lim_{n \to \infty} (b_n) = 0 + 0 = 0 \),即\( (a_n)_{n = 1}^{\infty} + (b_n)_{n = 1}^{\infty} \in A \)。

\( \forall (a_n)_{n = 1}^{\infty} \in \mathbf{C}^{\infty}, \forall \lambda \in \mathbf{C} \),有\( (\lambda (a_n)_{n = 1}^{\infty} = (\lambda a_n)_{n = 1}^{\infty} ) \in \mathbf{C}^{\infty} \),且\( \lim_{n \to \infty} (\lambda a_n) = \lambda \lim_{n \to \infty} a_n = 0 \),即\( (\lambda a_n)_{n = 1}^{\infty} \in A \)。

根据1.34,有\( A \)为\( \mathbf{C}^{\infty} \)的子空间。

证毕。

练习1.C.3

题目:

Show that the set of differentiable real-valued functions \( f \) on the interval \( (-4, 4) \) such that \( f'(-1) = 3f(2) \) is a subspace of \( \mathbf{R}^{(-4, 4)} \).

证明:

记所有在\( (-4, 4) \)上可微的实值函数\( f \),且\( f'(-1) = 3f(2) \)组成的集合为\( U \),有\( U \subseteq \mathbf{R}^{(-4, 4)} \)。

构造函数\( f_0: (-4, 4) \to \mathbf{R}, \forall x \in (-4, 4), f_0(x) = 0 \),有\( f_0 \)为在\( (-4, 4) \)上可微的实值函数,且\( f_0'(-1) = 0 = 3 \times 0 = 3f(2) \),即\( f_0 \in U \)。 \( \forall f \in U \),有\( f + f_0 = f \),即\( 0 = f_0 \in U \)。

\( \forall f, g \in U \),有\( f + g \)为在\( (-4, 4) \)上可微的实值函数,且\( (f + g)'(-1) = f'(-1) + g'(-1) = 3f(2) + 3g(2) = 3(f(2) + g(2)) = 3((f + g)(2)) \),即\( f + g \in U \)。

\( \forall f \in U, \forall \lambda \in \mathbf{R} \),有\( \lambda f \)为在\( (-4, 4) \)上可微的实值函数,且\( (\lambda f)'(-1) = \lambda f'(-1) = \lambda(3f(2)) = 3(\lambda f(2)) = 3((\lambda f)(2)) \),即\( \lambda f \in U \)。

根据1.34,有\( U \)为\( \mathbf{R}^{(-4, 4)} \)的子空间。

证毕。

练习1.C.4

题目:

Suppose \( b \in \mathbf{R} \). Show that the set of continuous real-valued functions \( f \) on the interval \( [0, 1] \) such that \( \int_0^1 f = b \) is a subspace of \( \mathbf{R}^{[0, 1]} \) if and only if \( b = 0 \).

证明:

记所有在\( [0, 1] \)上连续的实值函数\( f \),且\( \int_0^1 f = b \)组成的集合为\( U \)。

必要性:

如果\( U \)是\( \mathbf{R}^{[0, 1]} \)的子空间,记\( U \)中的加法单位元为\( f_0 \),则\( \forall f \in U, \int_0^1 (f + f_0) = \int_0^1 f + \int_0^1 f_0 \),又\( f + f_0 = f \),可得\( \int_0^1 f + \int_0^1 f_0 = \int_0^1 f \),即\( b + b = b \),解得\( b = 0 \)。

充分性:

如果\( b = 0 \):

构造函数\( f_0: [0, 1] \to \mathbf{R}, \forall x \in [0, 1], f_0(x) = 0 \),有\( f_0 \)为在\( [0, 1] \)上连续的实值函数,且\( \int_0^1 f = 0 = b \),即\( f_0 \in U \)。 \( \forall f \in U \),有\( f + f_0 = f \),即\( 0 = f_0 \in U \)。

\( \forall f, g \in U \),有\( f + g \)为在\( [0, 1] \)上连续的实值函数,且\( \int_0^1 (f + g) = \int_0^1 f + \int_0^1 g = b + b = 0 + 0 = 0 = b \),即\( f + g \in U \)。

\( \forall f \in U, \forall \lambda \in \mathbf{R} \),有\( \lambda f \)为在\( [0, 1] \)上连续的实值函数,且\( \int_0^1 (\lambda f) = \lambda(\int_0^1 f) = \lambda b = \lambda 0 = 0 = b \),即\( \lambda f \in U \)。

根据1.34,有\( U \)是\( \mathbf{R}^{[0, 1]} \)的子空间。

证毕。

练习1.C.5

题目:

Is \( \mathbf{R}^2 \) a subspace of the complex vector space \( \mathbf{C}^2 \)?

解答:

\( \mathbf{R}^2 \)是\( \mathbf{C}^2 \)的子空间,以下是证明:

\( \mathbf{R}^2 \subseteq \mathbf{C}^2 \)。

\( (0, 0) \in \mathbf{R}^2 \), \( \forall (x_0, x_1) \in \mathbf{R}^2, (x_0, x_1) + (0, 0) = (x_0, x_1) \),即\( 0 = (0, 0) \in \mathbf{R}^2 \)。

\( \forall (x_0, x_1), (y_0, y_1) \in \mathbf{R}^2 \),有\( (x_0, x_1) + (y_0, y_1) = (x_0 + y_0, x_1 + y_1) \in \mathbf{R}^2 \)。

\( \forall (x_0, x_1) \in \mathbf{R}^2, \forall \lambda \in \mathbf{R} \),有\( \lambda (x_0, x_1) = (\lambda x_0, \lambda x_1) \in \mathbf{R}^2 \)。

根据1.34,有\( \mathbf{R}^2 \)是\( \mathbf{C}^2 \)的子空间。

证毕。

练习1.C.6

题目:

  1. Is \( \{ (a, b, c) \in \mathbf{R}^3 : a^3 = b^3 \} \) a subspace of \( \mathbf{R}^3 \)?
  2. Is \( \{ (a, b, c) \in \mathbf{C}^3 : a^3 = b^3 \} \) a subspace of \( \mathbf{C}^3 \)?

证明1:

我们证明\( \forall a, b \in \mathbf{R}, (a = b) \equiv (a^3 = b^3) \),必要性是明显的,我们证明充分性,如果\( a^3 = b^3 \),假设 \( a \neq b \):

  1. 如果两者一个等于\( 0 \),另外一个不等于\( 0 \),则明显有\( a^3 \neq b^3 \),矛盾。
  2. 如果两者中一个\( > 0 \),另外一个\( < 0 \),不妨设\( a > 0, b < 0 \),则有\( a^3 > 0, b^3 < 0 \),因此\( a^3 \neq b^3 \),矛盾。
  3. 如果两者均\( > 0 \),此时因为\( a \neq b \),因此有\( a > b \)或\( b > a \),不妨设\( a > b \),则\( a^3 > b^3 \),矛盾。
  4. 如果两者均\( < 0 \),此时因为\( a \neq b \),因此有\( a > b \)或\( b > a \),不妨设\( b > a \),则\( -a > -b \),于是有\( (-a)^3 > (-b)^3 \),进而\( a^3 < b^3 \),矛盾。

综上,每种情况都会得到矛盾,因此假设不成立,有\( a = b \)。

接下来我们回去证明命题:

令\( U := \{ (a, b, c) \in \mathbf{R}^3 : a^3 = b^3 \} \),有\( U \subseteq \mathbf{R}^3 \)。

\( 0^3 = 0^3 \),因此\( (0, 0, 0) \in U \),而\( \forall (a, b, c) \in U \),有\( (a, b, c) + (0, 0, 0) = (a, b, c) \),即\( 0 = (0, 0, 0) \in U \)。

\( \forall (a, b, c), (d, e, f) \in U \),有\( a^3 = b^3, d^3 = e^3 \),根据前面的证明,有\( a = b, d = e \),因此\( (a, b, c) + (d, e, f) = (a + d, b + e, c + f) = (b + d, b + d, c + f) \),这里\( (b + d)^3 = (b + d)^3 \),因此\( (a, b, c) + (d, e, f) \in U \)。

\( \forall (a, b, c) \in U \),有\( a^3 = b^3 \),此时\( \forall \lambda \in \mathbf{R}, \lambda(a, b, c) = (\lambda a, \lambda b, \lambda c) \),可得\( (\lambda a)^3 = \lambda^3 a^3 = \lambda^3 b^3 = (\lambda b)^3 \),因此\( \lambda(a, b, c) \in U \)。

根据1.34,有\( U \)是\( \mathbf{R}^3 \)的子空间。

证毕。

证明2:

令\( U := \{ (a, b, c) \in \mathbf{C}^3 : a^3 = b^3 \} \), \( U \)是否是\( \mathbf{C}^3 \)的子空间的关键在于是否有\( (a^3 = b^3) \Rightarrow (a = b) \),我们尝试解下方程, \( a^3 = b^3 \Rightarrow a^3 - b^3 = 0 \Rightarrow (a - b)(a^2 + b^2 + ab) = 0 \),进而有\( a = b \)或\( a^2 + b^2 + ab \),由后者可得\( a = (\dfrac{-1}{2} \pm \dfrac{\sqrt{-3}}{2}i)b \),明显,不一定会有\( a = b \),令\( x := (\dfrac{-1}{2} + \dfrac{\sqrt{-3}}{2}i, 1, 1), y:= (\dfrac{-1}{2} - \dfrac{\sqrt{-3}}{2}i, 1, 1) \),易证\( x, y \in U \),而\( x + y = (-1, 2, 2) \),这里\( (-1)^3 \neq 2^3 \),因此\( x + y \notin U \),根据1.34,有\( U \)不是\( \mathbf{C}^3 \)的子空间。

证毕。

练习1.C.7

题目:

Give an example of a nonempty subset \( U \) of \( \mathbf{R}^2 \) such that \( U \) is closed under addition and under taking additive inverses (meaning \( -u \in U \) whenever \( u \in U \) ), but \( U \) is not a subspace of \( \mathbf{R}^2 \).

例子:

令\( U := \{ (z, z) : z \in \mathbf{Z} \} \),有\( U \subseteq \mathbf{R}^2, U \neq \emptyset \),易证\( U \)针对加法以及取加法逆元均封闭,但是令\( \lambda_0 := \dfrac{1}{2} \),有\( \lambda_0 (1, 1) = (\dfrac{1}{2}, \dfrac{1}{2}) \notin U \),根据1.34,有\( U \)不是\( \mathbf{R}^2 \)的子空间。

练习1.C.8

题目:

Give an example of a nonempty subset \( U \) of \( \mathbf{R}^2 \) such that \( U \) is closed under scalar multiplication, but \( U \) is not a subspace of \( \mathbf{R}^2 \).

例子:

令\( U := \{ (x, y) : x, y \in \mathbf{R}, x \mid y \} \) (注:\( x \mid y \)即\( x \)能整除\( y \)),有\( U \subseteq \mathbf{R}^2, U \neq \emptyset \), \( \forall \lambda \in \mathbf{R}, (x, y) \in U \),因为\( x \mid y \),因此\( \exists k \in \mathbf{Z}, y = kx \),进而\( \lambda y = \lambda kx = k (\lambda x) \),即\( \lambda x \mid \lambda y \),因此\( (\lambda x, \lambda y) \in U \),也就是\( U \)针对数乘封闭,然而\( (1, 3) + (2, 4) = (3, 7) \)且\( 3 \nmid 7 \),有\( (1, 3) + (2, 4) \notin U \),根据1.34,有\( U \)不是\( \mathbf{R}^2 \)的子空间。

练习1.C.9

题目:

A function \( f: \mathbf{R} \to \mathbf{R} \) is called periodic if there exists a positive number \( p \) such that \( f(x) = f(x + p) \) for all \( x \in \mathbf{R} \). Is the set of periodic functions from \( \mathbf{R} \to \mathbf{R} \) a subspace of \( \mathbf{R}^{\mathbf{R}} \)? Explain.

证明:

令\( U \)为所有\( \mathbf{R} \to \mathbf{R} \)的周期函数组成的集合,有\( U \subseteq \mathbf{R}^{\mathbf{R}} \)。

构造函数\( f_0: \mathbf{R} \to \mathbf{R}, \forall x \in \mathbf{R} \),令\( f_0(x) := 0 \),取\( p = 1 \),有\( \forall x \in \mathbf{R}, f_0(x) = 0 = f_0(x + p) \),即\( f_0 \)是\( \mathbf{R} \to \mathbf{R} \)的周期函数,于是有\( f_0 \in U \),除此之外,\( \forall f \in U \),有 \( f + f_0 = f \),因此\( 0 = f_0 \in U \)。

\( \forall f, g \in U \),因为\( f, g \)是周期函数,因此\( \exists p_1, p_2 \in \mathbf{Z}^+ \), \( \forall x \in X, f(x) = f(x + p_1), g(x) = g(x + p_2) \),易证\( f(x) = f(x + p_1p_2), g(x) = g(x + p_1p_2) \),因此\( (f + g)(x) = f(x) + g(x) = f(x + p_1p_2) + g(x + p_1p_2) = (f + g)(x + p_1p_2) \),可得\( f + g \)也是\( \mathbf{R} \to \mathbf{R} \)的周期函数,因此\( f + g \in U \)。

\( \forall f \in U \),因为\( f \)是周期函数,因此\( \exists p \in \mathbf{Z}^+ \), \( \forall x \in X, f(x) = f(x + p) \),可得\( \forall \lambda \in \mathbf{R}, (\lambda f)(x) = \lambda f(x) = \lambda f(x + p) = (\lambda f)(x + p) \),可得\( \lambda f \)也是\( \mathbf{R} \to \mathbf{R} \)的周期函数,因此\( \lambda f \in U \)。

综上,有\( U \)是\( \mathbf{R}^{\mathbf{R}} \)的子空间。

证毕。

练习1.C.10

题目:

Suppose \( U_1 \) and \( U_2 \) are subspaces of \( V \). Prove that the intersection \( U_1 \cap U_2 \) is a subspace of \( V \).

证明:

因为\( U_1, U_2 \)是\( V \)的子空间,有\( U_1 \subseteq V, U_2 \subseteq V \),可得\( U_1 \cap U_2 \subseteq V \),而\( 0 \in V \),因此\( 0 \in U_1, 0 \in U_2 \),进而\( 0 \in U_1 \cap U_2 \), 易证\( 0 \)也是\( U_1, U_2 \)的加法单位元。

\( \forall x, y \in U_1 \cap U_2 \),有\( x, y \in U_1 \)且\( x, y \in U_2 \),因为\( x, y \in U_1 \)以及\( U_1 \)是\( V \)的子空间,因此有\( x + y \in U_1 \),因为\( x, y \in U_2 \)以及\( U_2 \)是\( V \)的子空间,因此有\( x + y \in U_2 \),至此有\( x + y \in U_1 \cap U_2 \)。

\( \forall x \in U_1 \cap U_2 \),有\( x \in U_1 \)且\( x \in U_2 \), \( \forall \lambda \in \mathbf{F} \),因为\( U_1 \)是\( V \)的子空间,因此有\( \lambda x \in U_1 \),因为\( U_2 \)是\( V \)的子空间,因此有\( \lambda x \in U_2 \),至此有\( \lambda x \in U_1 \cap U_2 \)。

综上,有\( U_1 \cap U_2 \)是\( V \)的子空间。

证毕。

练习1.C.11

题目:

Prove that the intersection of every nonempty collection of subspaces of \( V \) is a subspace of \( V \).

注:

原文题目中,没有要求集合非空,要加上非空才是正确的,这里加上了。

证明:

令\( A \)为任意全由\( V \)的子空间组成的非空集合,我们证明 \( \bigcap_{U \in A} U \)是\( V \)的子空间:

\( \forall U \in A \),\( U \)是\( V \)的子空间,于是有\( 0 \in U \),综合可得\( 0 \in \bigcap_{U \in A} U \),易证\( 0 \)是\( \bigcap_{U \in A} U \)的加法单位元。

\( \forall x, y \in \bigcap_{U \in A} U \),有 \( \forall U \in A, x, y \in U \),而\( U \)是\( V \)的子空间,因此\( x + y \in U \),综合可得,\( x + y \in \bigcap_{U \in A} U \)。

\( \forall x \in \bigcap_{U \in A} U \),有 \( \forall U \in A, x \in U \),而\( U \)是\( V \)的子空间,因此\( \forall \lambda \in \mathbf{F}, \lambda x \in U \),综合可得,\( \forall \lambda \in \mathbf{F}, \lambda x \in \bigcap_{U \in A} U \)。

综上,有\( \bigcap_{U \in A} U \)是\( V \)的子空间。

证毕。

练习1.C.12

题目:

Prove that the union of two subspaces of \( V \) is a subspace of \( V \) if and only if one of the subspaces is contained in the other.

证明:

令\( U_1, U_2 \)为\( V \)的子空间。

必要性:

如果\( U_1 \cup U_2 \)是\( V \)的子空间,我们要证明\( U_1 \subseteq U_2 \)或\( U_2 \subseteq U_1 \),假设\( U_1 \nsubseteq U_2 \)且\( U_2 \nsubseteq U_1 \),则\( \exists x_1 \in U_1, x_1 \notin U_2 \), \( \exists x_2 \in U_2, x_2 \notin U_1 \),有\( x_1, x_2 \in U_1 \cup U_2 \),因为\( U_1 \cup U_2 \)是\( V \)的子空间,因此 \( x_1 + x_2 \in U_1 \cup U_2 \),于是有\( x_1 + x_2 \in U_1 \)或\( x_1 + x_2 \in U_2 \),不妨设\( x_1 + x_2 \in U_1 \)(另一种情况类似),则由于\( U_1 \)是\( V \)的子空间以及\( x_1 \in U_1 \),可得\( -x_1 \in U_1 \),进而\( ((-x_1) + x_1) + x_2 = x_2 \in U_1 \),而这与\( x_2 \notin U_1 \)矛盾,因此假设不成立,有\( U_1 \subseteq U_2 \)或\( U_2 \subseteq U_1 \)。

充分性:

如果\( U_1 \subseteq U_2 \)或\( U_2 \subseteq U_1 \),不妨设\( U_1 \subseteq U_2 \),此时\( U_1 \cup U_2 = U_2 \)是\( V \)的子空间。

证毕。

练习1.C.13

题目:

Prove that the union of three subspaces of \( V \) is a subspace of \( V \) if and only if one of the subspaces contains the other two. [This exercise is surprisingly harder than the previous exercise, possibly because this exercise is not true if we replace \( \mathbf{F} \) with a field containing only two elements.]

证明:

令\( U_1, U_2, U_3 \)为\( V \)的子空间,令\( U := U_1 \cup U_2 \cup U_3 \)。

必要性:

如果\( U \)是\( V \)的子空间,我们要证明三者中存在一个集合包含另外两个集合:

如果三者中存在一个集合包含另外两个集中的其中一个,不妨设\( U_1 \subseteq U_2 \),则\( U = U_1 \cup U_2 \cup U_3 = U_2 \cup U_3 \) 是\( V \)的子空间,此时根据练习1.C.12,有\( U_2 \subseteq U_3 \) 或\( U_3 \subseteq U_2 \):

  1. 如果\( U_2 \subseteq U_3 \),则由\( U_1 \subseteq U_2 \),有\( U_1 \subseteq U_3 \),于是有\( U_1, U_2 \subseteq U_3 \)。
  2. 如果\( U_3 \subseteq U_2 \),则\( U_1, U_3 \subseteq U_2 \)。

综上,如果三者中存在一个集合包含另外两个集中的其中一个,则有三者中存在一个集合包含另外两个集合。

如果上述情况不成立,即三者中不存在一个集合包含另外两个集中的其中一个,也就是三者互不包含,我们要证明这种情况是不可能的:

针对\( \forall x_{12} \in U_1 \setminus U_2 \), \( \forall x_{21} \in U_2 \setminus U_1 \):

由\( U \)是\( V \)的子空间以及\( x_{12}, x_{21} \in U \),可得\( x_{12} + x_{21} \in U \),于是有\( x_{12} + x_{21} \in U_1 \)或\( x_{12} + x_{21} \in U_2 \)或\( x_{12} + x_{21} \in U_3 \):

  1. 如果\( x_{12} + x_{21} \in U_1 \),则由\( U_1 \)是\( V \)的子空间以及\( x_{12} \in U_1 \),有\( -x_{12} \in U_1 \),进而有\( (-x_{12}) + (x_{12} + x_{21}) = (-x_{12} + x_{12}) + x_{21} = x_{21} \in U_1 \),然而这和\( x_{21} \notin U_1 \)矛盾,因此这种情况不可能。
  2. 如果\( x_{12} + x_{21} \in U_2 \),则由\( U_2 \)是\( V \)的子空间以及\( x_{21} \in U_2 \),有\( -x_{21} \in U_2 \),进而有\( (x_{12} + x_{21}) + (-x_{21}) = x_{12} + (x_{21} + (-x_{21})) = x_{12} \in U_2 \),然而这和\( x_{12} \notin U_2 \)矛盾,因此这种情况也不可能。

综上,\( x_{12} + x_{21} \in U_3 \)。

由\( U \)是\( V \)的子空间以及\( x_{12}, -x_{21} \in U \),可得\( x_{12} + (-x_{21}) \in U \),于是有\( x_{12} + (-x_{21}) \in U_1 \)或\( x_{12} + (-x_{21}) \in U_2 \)或\( x_{12} + (-x_{21}) \in U_3 \):

  1. 如果\( x_{12} + (-x_{21}) \in U_1 \),则由\( U_1 \)是\( V \)的子空间以及\( -x_{12} \in U_1 \),有\( (-x_{12}) + (x_{12} + (-x_{21})) = (-x_{12} + x_{12}) + (-x_{21}) = -x_{21} \in U_1 \),进一步可得\( x_{21} \in U_1 \),然而这和\( x_{21} \notin U_1 \)矛盾,因此这种情况不可能。
  2. 如果\( x_{12} + (-x_{21}) \in U_2 \),则由\( U_2 \)是\( V \)的子空间以及\( x_{21} \in U_2 \),进而有\( (x_{12} + (-x_{21})) + x_{21} = x_{12} + (-x_{21} + x_{21}) = x_{12} \in U_2 \),然而这和\( x_{12} \notin U_2 \)矛盾,因此这种情况也不可能。

综上,\( x_{12} + (-x_{21}) \in U_3 \),结合前面得到的\( x_{12} + x_{21} \in U_3 \),有\( (x_{12} + (-x_{21})) + (x_{12} + x_{21}) = x_{12} + x_{12} = 2x_{12} \in U_3 \) (注:\( 2x_{12} = (1 + 1)x_{12} = x_{12} + x_{12} \)),进而可得\( x_{12} \in U_3 \)。

综上,\( \forall x_{12} \in U_1 \setminus U_2 \),有\( x_{12} \in U_3 \),也就是\( (U_1 \setminus U_2) \subseteq U_3 \),同理易证\( (U_3 \setminus U_1) \subseteq U_2 \) 以及\( (U_3 \setminus U_2) \subseteq U_1 \) (1)

接着,我们证明\( (U_1 \cap U_2) \subseteq U_3 \),假设\( (U_1 \cap U_2) \nsubseteq U_3 \),即\( \exists x_{c12n3} \in U_1 \cap U_2, x_{c12n3} \notin U_3 \) (注:下标\( c12n3 \)表示 common element of \( U_1, U_2 \), not an element of \( U_3 \)),因为三个集合互不包含,因此有\( U_3 \nsubseteq U_1, U_3 \nsubseteq U_2 \),可得\( \exists x_{31} \in U_3 \setminus U_1, \exists x_{32} \in U_3 \setminus U_2 \),由\( U \)是\( V \)的子空间以及\( x_{c12n3}, x_{31}, x_{32} \in U \),可得\( x_{c12n3} + x_{31} + x_{32} \in U \),于是有 \( x_{c12n3} + x_{31} + x_{32} \in U_1 \)或\( x_{c12n3} + x_{31} + x_{32} \in U_2 \) 或\( x_{c12n3} + x_{31} + x_{32} \in U_3 \):

  1. 如果\( x_{c12n3} + x_{31} + x_{32} \in U_1 \),此时由\( U_1 \)是\( V \)的子空间以及\( x_{c12n3} \in U_1 \),可得\( -x_{c12n3} \in U_1 \),由(1)处得到的\( (U_3 \setminus U_2) \subseteq U_1 \),可得\( x_{32} \in U_1 \),进而可得\( -x_{32} \in U_1 \),于是有\( (x_{c12n3} + x_{31} + x_{32}) + (-x_{c12n3}) + (-x_{32}) = x_{31} \in U_1 \),然而这和\( x_{31} \notin U_1 \)矛盾。
  2. 如果\( x_{c12n3} + x_{31} + x_{32} \in U_2 \),此时由\( U_2 \)是\( V \)的子空间以及\( x_{c12n3} \in U_2 \),可得\( -x_{c12n3} \in U_2 \),由(1)处得到的\( (U_3 \setminus U_1) \subseteq U_2 \),可得\( x_{31} \in U_2 \),进而可得\( -x_{31} \in U_2 \),于是有\( (x_{c12n3} + x_{31} + x_{32}) + (-x_{c12n3}) + (-x_{31}) = x_{32} \in U_2 \),然而这和\( x_{32} \notin U_2 \)矛盾。
  3. 如果\( x_{c12n3} + x_{31} + x_{32} \in U_3 \),此时由\( U_3 \)是\( V \)的子空间以及\( x_{31}, x_{32} \in U_3 \),可得\( -x_{31}, -x_{32} \in U_3 \),于是有\( (x_{c12n3} + x_{31} + x_{32}) + (-x_{31}) + (-x_{32}) = x_{c12n3} \in U_3 \),然而这和\( x_{c12n3} \notin U_3 \)矛盾。

综上,所有情况均矛盾,因此假设不成立,有\( (U_1 \cap U_2) \subseteq U_3 \),加上(1)处我们得到的\( (U_1 \setminus U_2) \subseteq U_3 \),可得\( ((U_1 \setminus U_2) \cup (U_1 \cap U_2) = U_1) \subseteq U_3 \),即\( U_1 \)被\( U_3 \)包含,这和三个集合互不包含矛盾,因此情况“三者中不存在一个集合包含另外两个集中的其中一个,也就是三者互不包含”是不可能的,仅剩下“三者中存在一个集合包含另外两个集中的其中一个”这种情况,而前面我们证明了,该情况下能推出“三者中存在一个集合包含另外两个集合”,即命题成立。

充分性:

如果\( U_2, U_3 \subseteq U_1 \)或\( U_1, U_3 \subseteq U_2 \) 或\( U_1, U_2 \subseteq U_3 \),不妨设\( U_2, U_3 \subseteq U_1 \),则\( U_1 \cup U_2 \cup U_3 = U_1 \)是\( V \)的子空间。

证毕。

练习1.C.14

题目:

Verify the assertion in Example 1.38.

Example 1.38的内容:

Suppose that \( U = \{ (x, x, y, y) \in \mathbf{F}^4 : x, y \in \mathbf{F} \} \) and \( W = \{ (x, x, x, y) \in \mathbf{F}^4 : x, y \in \mathbf{F} \} \) Then \( U + W = \{ (x, x, y, z) \in \mathbf{F}^4 : x, y, z \in \mathbf{F} \} \) as you should verify.

验证:

\( \forall e \in U + W \),\( \exists (a_1, a_1, a_2, a_2) \in U, (b_1, b_1, b_1, b_2) \in W, e = (a_1, a_1, a_2, a_2) + (b_1, b_1, b_1, b_2) = (a_1 + b_1, a_1 + b_1, a_2 + b_1, a_2 + b_2) \),这里由于前两个坐标均为\( a_1 + b_1 \in \mathbf{F} \)且\( a_2 + b_1, a_2 + b_2 \in \mathbf{F} \),因此\( e = (a_1 + b_1, a_1 + b_1, a_2 + b_1, a_2 + b_2) \in \{ (x, x, y, z) \in \mathbf{F}^4 : x, y, z \in \mathbf{F} \} \)。

反之,\( \forall e \in \{ (x, x, y, z) \in \mathbf{F}^4 : x, y, z \in \mathbf{F} \} \),令\( e_1 := (x - y, x - y, 0, z), e_2 := (y, y, y, 0) \),则\( e_1 \in U, e_2 \in W, e_1 + e_2 = e \in U + W \)。

综上,有\( U + W = \{ (x, x, y, z) \in \mathbf{F}^4 : x, y, z \in \mathbf{F} \} \)。

练习1.C.15

题目:

Suppose \( U \) is a subspace of \( V \). What is \( U + U \)?

解答:

根据1.39,有\( U + U \)也为\( V \)的子空间,且\( U + U \)是同时包含\( U, U \)的子空间中最小的,即包含\( U \)的子空间中最小的,实际上我们知道,包含\( U \)的最小子空间为\( U \)本身,因此实际上 \( U + U = U \)。

练习1.C.16

题目:

Is the operation of addition on the subspaces of \( V \) commutative? In other words, if \( U \) and \( W \) are subspaces of \( V \), is \( U + W = W + U \)?

解答:

是的,两个\( V \)的子空间的和操作满足交换律,证明如下:

\( \forall e \in U + W, \exists u \in U, w \in W, e = u + w \),而\( U + W \)为\( V \)的子空间,因此\( e = u + w = w + u \),这意味着\( e \in W + U \),类似的,\( \forall e \in W + U \),易证\( e \in U + W \),综上,有\( U + W = W + U \)。

证毕。

练习1.C.17

题目:

Is the operation of addition on the subspaces of \( V \) associative? In other words, if \( U_1, U_2, U_3 \) are subspaces of V, is \( (U_1 + U_2) + U_3 = U_1 + (U_2 + U_3) \)?

解答:

是的,三个\( V \)的子空间的和操作满足结合律,证明如下:

\( \forall e \in (U_1 + U_2) + U_3, \exists u_1 \in U_1, u_2 \in U_2, u_3 \in U_3, e = (u_1 + u_2) + u_3 \),而\( (U_1 + U_2) + U_3 \)为\( V \)的子空间,因此\( e = (u_1 + u_2) + u_3 = u_1 + (u_2 + u_3) \),这意味着\( e \in U_1 + (U_2 + U_3) \),类似的,\( \forall e \in U_1 + (U_2 + U_3) \),易证\( e \in (U_1 + U_2) + U_3 \),综上,有\( (U_1 + U_2) + U_3 = U_1 + (U_2 + U_3) \)。

练习1.C.18

题目:

Does the operation of addition on the subspaces of \( V \) have an additive identity? Which subspaces have additive inverses?

子空间的加法操作是否有加法单位元?

题目的意思是说,\( V \)是否有子空间\( I \),满足\( \forall V \)的子空间\( U \),有\( U + I = U \),有的,\( \{ 0 \} \)就是。

哪些子空间针对子空间的加法操作有加法逆元?

根据上面的讨论,子空间的加法操作的加法单位元是\( 0 = I = \{ 0 \} \),因此一个子空间\( U \)的加法逆元就是满足\( U + W = I \)的子空间\( W \),那哪些子空间针对子空间的加法操作有加法逆元呢?针对这些子空间\( U \),假设它的加法逆元是\( W \),则固定任意一个\( u \in U \),\( U + W = I \)要求\( \forall w \in W, u + w = 0 \),唯一的可能就是\( U = \{ 0 \} \),此时它的加法逆元\( W = \{ 0 \} \),这点是很容易证明的,假设\( U \)中包含非\( 0 \)元素\( u \),则\( 0 \in W \),有\( u + 0 = u \in U + W \neq 0 \),进而\( U + W \neq \{ 0 \} = I \),这和\( U + W = I \)矛盾,因此假设不成立,有\( U = \{ 0 \} \)。

练习1.C.19

题目:

Prove or give a counterexample: if \( U_1, U_2, W \) are subspaces of \( V \) such that \( U_1 + W = U_2 + W \), then \( U_1 = U_2 \)。

注:

题目想证明子空间的加法操作不满足加法消去律。

反例:

令\( V := \mathbf{F}^2, U_1 := \{ (x, 0) : x \in \mathbf{F} \}, U_2 := \{ (0, y) : y \in \mathbf{F} \}, W := V \),这里\( U_1, U_2, W \)均是\( V \)的子空间,且\( U_1 + W = W = U_2 + W \),但\( U_1 \neq U_2 \)。

练习1.C.20

题目

Suppose \( U = \{ (x, x, y, y) \in \mathbf{F}^4 : x, y \in \mathbf{F} \} \). Find a subspace \( W \) of \( \mathbf{F}^4 \) such that \( \mathbf{F}^4 = U \oplus W \).

解答:

易证\( U \)为\( \mathbf{F}^4 \)的子空间,根据1.45,我们只要找到\( \mathbf{F}^4 \)的子空间\( W \)满足\( U \cap W = \{ (0, 0, 0, 0) \} \)且 \( \mathbf{F}^4 = U + W \)就行,由于要求\( U \cap W = \{ (0, 0, 0, 0) \} \),即\( U, W \)的共同元素只有\( (0, 0, 0, 0) \),这意味着\( W \)中所有非\( 0 \)元均不能第\( 1, 2 \)坐标相等且第\( 3, 4 \)坐标相等,又由于要求\( \mathbf{F}^4 = U + W \),这意味着我们得想办法打破\( U \)中元素第\( 1, 2 \)坐标相等、第\( 3, 4 \)坐标相等的限制。

令\( W := \{ (0, w_1, 0, w_2) : w_1, w_2 \in \mathbf{F} \} \),易证\( W \)为\( V \)的子空间,还得证明下\( U \cap W = \{ (0, 0, 0, 0) \} \): \( \forall e = (e_1, e_2, e_3, e_4) \in U \cap W \),由\( e \in U \),可得\( e_2 = e_1, e_4 = e_3 \),由\( e \in W \),可得\( e_1 = e_3 = 0 \),综合可得,\( e = (0, 0, 0, 0) \),至此,我们有\( U \cap W = \{ (0, 0, 0, 0) \} \),因此根据1.45,可得\( U + W \)为直和,也就是\( U + W = U \oplus W \)。最后就是证明\( \mathbf{F}^4 = U + W \)了,明显的,\( U + W \subseteq \mathbf{F}^4 \),因此我们只需要证明\( \mathbf{F}^4 \subseteq U + W \): \( \forall (x_1, x_2, x_3, x_4) \in \mathbf{F}^4 \),令\( u := (x_1, x_1, x_3, x_3) \in U, w := (0, x_2 - x_1, 0, x_4 - x_3) \in W \),则有\( u + w = (x_1, x_2, x_3, x_4) \in U + W \),也就是\( \mathbf{F}^4 \subseteq U + W \),结合\( U + W \subseteq \mathbf{F}^4 \)可得, \( \mathbf{F}^4 = U + W = U \oplus W \)。

练习1.C.21

题目:

Suppose \( U = \{ (x, y, x + y, x - y, 2x) \in \mathbf{F}^5 : x, y \in \mathbf{F} \} \). Find a subspace \( W \) of \( \mathbf{F}^5 \) such that \( \mathbf{F}^5 = U \oplus W \).

解答:

易证\( U \)为\( \mathbf{F}^5 \)的子空间,根据1.45,我们只要找到\( \mathbf{F}^5 \)的子空间\( W \)满足\( U \cap W = \{ (0, 0, 0, 0, 0) \} \)且 \( \mathbf{F}^5 = U + W \)就行。

令\( W := \{ (0, 0, w_1, w_2, w_3) : w_1, w_2, w_3 \in \mathbf{F} \} \),易证\( W \)为\( V \)的子空间,还得证明下\( U \cap W = \{ (0, 0, 0, 0, 0) \} \): \( \forall e = (e_1, e_2, e_3, e_4, e_5) \in U \cap W \),由\( e \in U \),可得\( e_3 = e_1 + e_2, e_4 = e_1 - e_2, e_5 = 2e_1 \),由\( e \in W \),可得\( e_1 = e_2 = 0 \),综合可得,\( e = (0, 0, 0, 0, 0) \),至此,我们有\( U \cap W = \{ (0, 0, 0, 0, 0) \} \),因此根据1.45,可得\( U + W \)为直和,也就是\( U + W = U \oplus W \)。最后就是证明\( \mathbf{F}^5 = U + W \)了:明显的,\( U + W \subseteq \mathbf{F}^5 \),因此我们只需要证明\( \mathbf{F}^5 \subseteq U + W \): \( \forall (x_1, x_2, x_3, x_4, x_5) \in \mathbf{F}^5 \),令\( u := (x_1, x_2, x_1 + x_2, x_1 - x_2, 2x_1) \in U, w := (0, 0, x_3 - x_1 - x_2, x_4 - x_1 + x_2, x_5 - 2x_1) \in W \),则有\( u + w = (x_1, x_2, x_3, x_4, x_5) \in U + W \),也就是\( \mathbf{F}^5 \subseteq U + W \),结合\( U + W \subseteq \mathbf{F}^5 \)可得, \( \mathbf{F}^5 = U + W = U \oplus W \)。

练习1.C.22

Suppose \( U = \{ (x, y, x + y, x - y, 2x) \in \mathbf{F}^5 : x, y \in \mathbf{F} \} \). Find a subspace \( W_1, W_2, W_3 \) of \( \mathbf{F}^5 \), none of which equals \( \{ 0 \} \), such that \( \mathbf{F}^5 = U \oplus W_1 \oplus W_2 \oplus W_3 \).

解答:

易证\( U \)为\( \mathbf{F}^5 \)的子空间,令\( W_1 := \{ (0, 0, w_1, 0, 0) : w_1 \in \mathbf{F} \}, W_2 := \{ (0, 0, 0, w_2, 0) : w_2 \in \mathbf{F} \}, W_3 := \{ (0, 0, 0, 0, w_3) : w_3 \in \mathbf{F} \} \),三者都不等于\( \{ 0 \} \),易证三者是\( V \)的子空间。

我们得证明下\( U + W_1 + W_2 + W_3 \)是直和,根据1.44,我们需要证明\( (0, 0, 0, 0, 0) \)在\( U + W_1 + W_2 + W_3 \) 有唯一的表示\( 0 + 0 + 0 + 0 \),这里第1个\( 0 = (0, 0, 0, 0, 0) \in U \),第2个\( 0 = (0, 0, 0, 0, 0) \in W_1 \),以此类推, \( \forall u = (u_1, u_2, u_3, u_4, u_5) \in U, w_1 = (w_{11}, w_{12}, w_{13}, w_{14}, w_{15}) \in W_1, w_2 = (w_{21}, w_{22}, w_{23}, w_{24}, w_{25}) \in W_2, w_3 = (w_{31}, w_{32}, w_{33}, w_{34}, w_{35}) \in W_3 \)满足\( u + w_1 + w_2 + w_3 = (0, 0, 0, 0, 0) \),假设\( u, w_1, w_2, w_3 \)中至少有一个\( \neq (0, 0, 0, 0, 0) \):

  1. 如果\( u \neq (0, 0, 0, 0, 0) \),则由\( u \in U \),于是有\( u_1 \neq 0 \)或\( u_2 \neq 0 \),不妨设\( u_1 \neq 0 \),由\( u + w_1 + w_2 + w_3 = 0 \),可得\( u_1 + w_{11} + w_{21} + w_{31} = u_1 + 0 + 0 + 0 = u_1 = 0 \),矛盾,因此该种情况不可能。
  2. 如果\( w_1 \neq (0, 0, 0, 0, 0) \),则由\( w_1 \in W_1 \),可得\( w_{13} \neq 0 \),由\( u + w_1 + w_2 + w_3 = 0 \),可得\( u_3 + w_{13} + w_{23} + w_{33} = u_3 + w_{13} + 0 + 0 = u_3 + w_{13} = 0 \),而\( u_3 = u_1 + u_2 \),由前一种情况的讨论,可得\( u_1 = u_2 = 0 \),因此\( u_3 = u_1 + u_2 = 0 + 0 = 0 \),进而有\( u_3 + w_{13} = 0 + w_{13} = w_{13} \neq 0 \),矛盾,因此这种情况也不可能。

同理可得,\( w_2 \neq (0, 0, 0, 0, 0) \)以及\( w_3 \neq (0, 0, 0, 0, 0) \)均不可能,因此有\( u = w_1 = w_2 = w_3 = (0, 0, 0, 0, 0) \),即\( (0, 0, 0, 0, 0) \)在\( U + W_1 + W_2 + W_3 \)有唯一的表示,根据1.44,有\( U + W_1 + W_2 + W_3 \)是直和。

最后就是证明\( \mathbf{F}^5 = U + W_1 + W_2 + W_3 \)了:明显的,\( U + W_1 + W_2 + W_3 \subseteq \mathbf{F}^5 \),因此我们只需要证明\( \mathbf{F}^5 \subseteq U + W_1 + W_2 + W_3 \): \( \forall (x_1, x_2, x_3, x_4, x_5) \in \mathbf{F}^5 \),令\( u := (x_1, x_2, x_1 + x_2, x_1 - x_2, 2x_1) \in U, w_1 := (0, 0, x_3 - x_1 - x_2, 0, 0) \in W_1, w_2 := (0, 0, 0, x_4 - x_1 + x_2, 0) \in W_2, w_3 := (0, 0, 0, 0, x_5 - 2x_1) \in W_3 \),则有\( u + w_1 + w_2 + w_3 = (x_1, x_2, x_3, x_4, x_5) \in U + W_1 + W_2 + W_3 \),也就是\( \mathbf{F}^5 \subseteq U + W_1 + W_2 + W_3 \),结合\( U + W_1 + W_2 + W_3 \subseteq \mathbf{F}^5 \)可得, \( \mathbf{F}^5 = U + W_1 + W_2 + W_3 = U \oplus W_1 \oplus W_2 \oplus W_3 \)。

练习1.C.23

题目:

Prove or give a counterexample: if \( U_1, U_2, W \) are subspaces of \( V \) such that \( V = U_1 \oplus W \) and \( V = U_2 \oplus W \), then \( U_1 = U_2 \).

注:

题目想探讨子空间的直和操作是否满足加法消去律,可以对比下练习1.C.19。

解答:

该命题是错误的,下面是反例,令\( V := \mathbf{F}^2, U_1 := \{ (0, x_1) : x_1 \in \mathbf{F} \}, U_2 := \{ (x_2, x_2) : x_2 \in \mathbf{F} \}, W := \{ (w_1, 0) : w_1 \in \mathbf{F} \} \),则使用1.45,易证\( U_1 + W, U_2 + W \)均为直和,除此之外,还易得\( V = U_1 + W = U_1 \oplus W = U_2 + W = U_2 \oplus W \),但\( U_1 \neq U_2 \)。

练习1.C.24

题目:

A function \( f: \mathbf{R} \to \mathbf{R} \) is called even if \( f(-x) = f(x) \) for all \( x \in \mathbf{R} \). A function \( f: \mathbf{R} \to \mathbf{R} \) is called odd if \( f(-x) = -f(x) \) for all \( x \in \mathbf{R} \). Let \( U_e \) denote the set of real-valued even functions on \( \mathbf{R} \) and let \( U_o \) denote the set of real-valued odd functions on \( \mathbf{R} \). Show that \( \mathbf{R^R} = U_e \oplus U_o \).

证明:

定义函数\( 0: \mathbf{R} \to \mathbf{R}, \forall x \in \mathbf{R}, 0(x) = 0 \),易得\( 0 \)是\( \mathbf{R^R} \)的加法单位元。

明显的,\( U_e, U_o \subseteq \mathbf{R^R} \),易证\( U_e, U_o \)为\( \mathbf{R^R} \)的子空间。

接着,我们证明\( U_e + U_o \)是直和,根据1.45,我们只需要证明\( U_e \cap U_o = \{ 0 \} \)即可: \( \forall f \in U_e \cap U_o \),由\( f \in U_e \),可得\( \forall x \in \mathbf{R}, f(-x) = f(x) \),由\( f \in U_o \),可得\( \forall x \in \mathbf{R}, f(-x) = -f(x) \),于是可得\( \forall x \in \mathbf{R}, f(x) = -f(x) \),解得\( f(x) = 0 \)。综上可得,\( \forall x \in \mathbf{R}, f(x) = 0 \),即\( f = 0 \),这意味着\( U_e \cap U_o = \{ 0 \} \)。

最后,我们需要证明\( \mathbf{R^R} = U_e \oplus U_o \),明显的,有\( U_e \oplus U_o \subseteq \mathbf{R^R} \),我们只需要证明 \( \mathbf{R^R} \subseteq U_e \oplus U_o \)即可: \( \forall f \in \mathbf{R^R} \),我们需要证明\( f \in U_e \oplus U_o \),这需要证明\( f \)能被表示成一个偶函数和一个奇函数的和,我们假设\( f(x) = g(x) + h(x) \),其中\( g \in U_e, h \in U_o \),则\( f(-x) = g(x) - h(x) \),进而有\( f(x) + f(-x) = 2g(x) \),解得\( g(x) = \dfrac{f(x) + f(-x)}{2}, h(x) = f(x) - g(x) = \dfrac{f(x) - f(-x)}{2} \),易验证\( g \in U_e, h \in U_o \),加上\( f(x) = g(x) + h(x) \),可得\( f \in U_e \oplus U_o \),综上,有\( \mathbf{R^R} \subseteq U_e \oplus U_o \),结合前面得到的\( U_e \oplus U_o \subseteq \mathbf{R^R} \),可得\( \mathbf{R^R} = U_e \oplus U_o \)。

证毕。