1演習方針
線型写像かどうかは、いくつかの数値例ではなく、任意のベクトルとスカラーについて加法性と同次性を確認して判定する。
反例を示すときは、加法性または同次性のどちらかが破れる 1 例を出せばよい。証明するときは、任意の文字で計算する。
1Exercise plan
To decide whether a map is a linear map, do not rely on a few numerical examples. Check additivity加法性かほうせい and homogeneity同次性どうじせい for arbitrary vectorsベクトル and scalarsスカラー.
To give a counterexample反例はんれい, one example where additivity or homogeneity fails is enough. To prove linearity, compute with arbitrary symbols.
2問題もんだい 1
T:\mathbb R^2\to\mathbb R^2 を
T\begin{pmatrix}x\\y\end{pmatrix}
=
\begin{pmatrix}
x+y\\
2x-y
\end{pmatrix}
で定義ていぎする。T が線型写像せんけいしゃぞうlinear mapであることを示しめせ。
2.1解答例かいとうれい
u=(x_1,y_1)^T、v=(x_2,y_2)^T とする。
T(u+v)
=
\begin{pmatrix}
x_1+x_2+y_1+y_2\\
2(x_1+x_2)-(y_1+y_2)
\end{pmatrix}
であり、
T(u)+T(v)
=
\begin{pmatrix}
x_1+y_1\\
2x_1-y_1
\end{pmatrix}
+
\begin{pmatrix}
x_2+y_2\\
2x_2-y_2
\end{pmatrix}
=
\begin{pmatrix}
x_1+x_2+y_1+y_2\\
2(x_1+x_2)-(y_1+y_2)
\end{pmatrix}
なので加法性かほうせいadditivityが成立せいりつする。
また、スカラー c について
T(cu)
=
\begin{pmatrix}
cx_1+cy_1\\
2cx_1-cy_1
\end{pmatrix}
=
c
\begin{pmatrix}
x_1+y_1\\
2x_1-y_1
\end{pmatrix}
=
cT(u)
なので同次性どうじせいhomogeneityも成立せいりつする。したがって T は線型写像せんけいしゃぞうlinear mapである。
2.2解説かいせつ
この問題もんだいでは、線型性せんけいせいlinearityを定義ていぎdefinitionから確認かくにんしている。係数けいすうcoefficientつきの一次式いちじしきlinear expressionで作つくられていても、任意にんいの u,v,c について加法性かほうせいadditivityと同次性どうじせいhomogeneityを確認かくにんするまでは証明しょうめいproofにならない。
2Problem 1
Define T:\mathbb R^2\to\mathbb R^2 by
T\begin{pmatrix}x\\y\end{pmatrix}=\begin{pmatrix}x+y\\2x-y\end{pmatrix}.
Show that T is a linear map線型写像せんけいしゃぞう.
2.1Answer example
Let u=(x_1,y_1)^T and v=(x_2,y_2)^T. Then
T(u+v)=\begin{pmatrix}x_1+x_2+y_1+y_2\\2(x_1+x_2)-(y_1+y_2)\end{pmatrix},
and
T(u)+T(v)=\begin{pmatrix}x_1+y_1\\2x_1-y_1\end{pmatrix}+\begin{pmatrix}x_2+y_2\\2x_2-y_2\end{pmatrix}=\begin{pmatrix}x_1+x_2+y_1+y_2\\2(x_1+x_2)-(y_1+y_2)\end{pmatrix}.
Thus additivity holds. For a scalar c,
T(cu)=\begin{pmatrix}cx_1+cy_1\\2cx_1-cy_1\end{pmatrix}=c\begin{pmatrix}x_1+y_1\\2x_1-y_1\end{pmatrix}=cT(u),
so homogeneity holds. Therefore T is linear.
2.2Explanation
The proof returns to the definition定義ていぎ of linearity. A formula that looks linear is not proved linear until additivity and homogeneity are checked for arbitrary inputs.
3問題もんだい 2
S:\mathbb R^2\to\mathbb R^2 を
S\begin{pmatrix}x\\y\end{pmatrix}
=
\begin{pmatrix}
x+1\\
y
\end{pmatrix}
で定義ていぎする。S が線型写像せんけいしゃぞうlinear mapではないことを示しめせ。
3.1解答例かいとうれい
線型写像せんけいしゃぞうlinear mapなら S(0)=0 でなければならない。しかし
S\begin{pmatrix}0\\0\end{pmatrix}
=
\begin{pmatrix}1\\0\end{pmatrix}
\ne
\begin{pmatrix}0\\0\end{pmatrix}
である。したがって S は線型写像せんけいしゃぞうlinear mapではない。
3.2解説かいせつ
平行移動へいこういどうtranslationは直線ちょくせんlineを直線ちょくせんlineへ移うつすが、原点げんてんoriginを原点げんてんoriginへ保たもたないため線型せんけいlinearではない。線型性せんけいせいlinearityの反例はんれいcounterexampleでは、まず T(0)=0 を満みたすかを確認かくにんすると早はやい。
3Problem 2
Define S:\mathbb R^2\to\mathbb R^2 by
S\begin{pmatrix}x\\y\end{pmatrix}=\begin{pmatrix}x+1\\y\end{pmatrix}.
Show that S is not a linear map.
3.1Answer example
A linear map must satisfy S(0)=0, but
S\begin{pmatrix}0\\0\end{pmatrix}=\begin{pmatrix}1\\0\end{pmatrix}\ne\begin{pmatrix}0\\0\end{pmatrix}.
Therefore S is not linear.
3.2Explanation
A translation平行移動へいこういどう sends lines to lines, but it does not keep the origin原点げんてん fixed. Checking T(0)=0 is often the fastest way to find a counterexample.
4問題もんだい 3
R:\mathbb R^2\to\mathbb R^2 を
R\begin{pmatrix}x\\y\end{pmatrix}
=
\begin{pmatrix}
x^2\\
y
\end{pmatrix}
で定義ていぎする。R が線型写像せんけいしゃぞうlinear mapではないことを、同次性どうじせいhomogeneityの失敗しっぱいから示しめせ。
4.1解答例かいとうれい
u=(1,0)^T、c=2 とすると、
R(2u)=R\begin{pmatrix}2\\0\end{pmatrix}
=
\begin{pmatrix}4\\0\end{pmatrix}
である。一方いっぽう、
2R(u)
=
2\begin{pmatrix}1\\0\end{pmatrix}
=
\begin{pmatrix}2\\0\end{pmatrix}
である。したがって R(2u)\ne2R(u) であり、同次性どうじせいhomogeneityが成立せいりつしない。
4.2解説かいせつ
x^2 のような非線型ひせんけいnonlinearな項こうtermがあると、スカラー倍ばいscalar multiplicationの倍率ばいりつが保たもたれない。この問題もんだいは、原点げんてんoriginを保たもつだけでは線型性せんけいせいlinearityに十分じゅうぶんでないことを確認かくにんしている。
4Problem 3
Define R:\mathbb R^2\to\mathbb R^2 by
R\begin{pmatrix}x\\y\end{pmatrix}=\begin{pmatrix}x^2\\y\end{pmatrix}.
Show that R is not linear by showing failure of homogeneity同次性どうじせい.
4.1Answer example
Let u=(1,0)^T and c=2. Then
R(2u)=R\begin{pmatrix}2\\0\end{pmatrix}=\begin{pmatrix}4\\0\end{pmatrix},
but
2R(u)=2\begin{pmatrix}1\\0\end{pmatrix}=\begin{pmatrix}2\\0\end{pmatrix}.
Thus R(2u)\ne2R(u), so homogeneity fails.
4.2Explanation
A nonlinear非線型ひせんけい term such as x^2 does not preserve scalar multiplication. Preserving the origin alone is not enough for linearity.
5問題もんだい 4
線型写像せんけいしゃぞうlinear map T:\mathbb R^3\to\mathbb R^2 が
T(e_1)=\begin{pmatrix}1\\0\end{pmatrix},\qquad
T(e_2)=\begin{pmatrix}-1\\2\end{pmatrix},\qquad
T(e_3)=\begin{pmatrix}3\\1\end{pmatrix}
を満みたすとする。T の標準基底ひょうじゅんきていstandard basisに関かんする行列ぎょうれつmatrixを求もとめ、T(2,-1,4)^T を計算けいさんせよ。
5.1解答例かいとうれい
行列ぎょうれつmatrixの列れつcolumnは標準基底ひょうじゅんきていstandard basisの像ぞうimageなので、
A=
\begin{pmatrix}
1&-1&3\\
0&2&1
\end{pmatrix}
である。また
(2,-1,4)^T=2e_1-e_2+4e_3
したがって、
T(2,-1,4)^T
=
2T(e_1)-T(e_2)+4T(e_3)
=
\begin{pmatrix}15\\2\end{pmatrix}
である。
5.2解説かいせつ
この問題もんだいは、行列ぎょうれつmatrixの列れつcolumnが基底きていベクトルbasis vectorの行ゆき先さきdestinationを記録きろくしていることを確認かくにんしている。線型性せんけいせいlinearityがあるため、任意にんいの入力にゅうりょくinputは基底きていbasisの像ぞうimageの同おなじ係数けいすうcoefficientによる線型結合せんけいけつごうlinear combinationで計算けいさんできる。
5Problem 4
A linear map T:\mathbb R^3\to\mathbb R^2 satisfies
T(e_1)=\begin{pmatrix}1\\0\end{pmatrix},\qquad T(e_2)=\begin{pmatrix}-1\\2\end{pmatrix},\qquad T(e_3)=\begin{pmatrix}3\\1\end{pmatrix}.
Find the matrix of T in the standard basis標準基底ひょうじゅんきてい, and compute T(2,-1,4)^T.
5.1Answer example
The columns are the images of the standard basis vectors:
A=\begin{pmatrix}1&-1&3\\0&2&1\end{pmatrix}.
Since (2,-1,4)^T=2e_1-e_2+4e_3,
T(2,-1,4)^T=2T(e_1)-T(e_2)+4T(e_3)=\begin{pmatrix}15\\2\end{pmatrix}.
5.2Explanation
The columns of a matrix record the destinations of the basis vectors基底きていベクトル. Linearity lets every input be computed using the same coefficients on the images.
6問題もんだい 5
線型写像せんけいしゃぞうlinear map T:V\to W とベクトルvector v_1,v_2 について、v_2=3v_1 が成立せいりつするとする。このとき T(v_2) と T(v_1) の関係かんけいrelationを述のべよ。
6.1解答例かいとうれい
同次性どうじせいhomogeneityより
T(v_2)=T(3v_1)=3T(v_1)
である。
6.2解説かいせつ
線型写像せんけいしゃぞうlinear mapは線型結合せんけいけつごうlinear combinationの関係かんけいrelationを保存ほぞんする。したがって、入力側にゅうりょくがわinput sideで従属じゅうぞくdependentしている関係かんけいrelationは、出力側しゅつりょくがわoutput sideでも同おなじ係数けいすうcoefficientの関係かんけいrelationとして残のこる。ただし、逆ぎゃくに出力側しゅつりょくがわoutput sideで従属じゅうぞくdependentしているからといって、入力側にゅうりょくがわinput sideで同おなじ関係かんけいrelationがあったとは限かぎらない。
6Problem 5
Let T:V\to W be linear and suppose v_2=3v_1. State the relation between T(v_2) and T(v_1).
6.1Answer example
By homogeneity,
T(v_2)=T(3v_1)=3T(v_1).
6.2Explanation
A linear map preserves linear-combination線型結合せんけいけつごう relations. Therefore, if the input side has a dependent relation, the output side keeps the same coefficients in the corresponding relation. The converse need not hold: even if the output vectors are dependent, the same relation need not have existed on the input side, because the map may collapse directions.
7問題もんだい 6
T:\mathbb C\to\mathbb C を
T(z)=\overline z
で定義ていぎする。\mathbb C を \mathbb R 上じょうのベクトル空間くうかんと見みると T が実線型じつせんけいreal-linearであり、\mathbb C 上じょうのベクトル空間くうかんと見みると複素線型ふくそせんけいcomplex-linearではないことを示しめせ。
7.1解答例かいとうれい
z,w\in\mathbb C について
T(z+w)=\overline{z+w}=\overline z+\overline w=T(z)+T(w)
である。また、a\in\mathbb R なら
T(az)=\overline{az}=a\overline z=aT(z)
なので、T は実線型じつせんけいreal-linearである。
一方いっぽう、複素ふくそcomplexのスカラーscalar i を使つかうと、
T(i\cdot 1)=\overline i=-i
だが、
iT(1)=i
である。したがって T(i\cdot1)\ne iT(1) であり、複素線型ふくそせんけいcomplex-linearではない。
7.2解説かいせつ
この問題もんだいは、線型性せんけいせいlinearityが「どの体たいをスカラーscalarとして使つかうか」に依存いぞんすることを確認かくにんしている。同おなじ写像しゃぞうmapでも、実数じっすうreal numberのスカラーscalarだけを許ゆるす場合ばあいと、複素数ふくそすうcomplex numberのスカラーscalarを許ゆるす場合ばあいで同次性どうじせいhomogeneityの意味いみが変かわる。
7Problem 6
Define T:\mathbb C\to\mathbb C by T(z)=\overline z. Show that T is real-linear実線型じつせんけい when \mathbb C is a vector space over \mathbb R, but not complex-linear複素線型ふくそせんけい over \mathbb C.
7.1Answer example
For z,w\in\mathbb C,
T(z+w)=\overline{z+w}=\overline z+\overline w=T(z)+T(w).
If a\in\mathbb R, then
T(az)=\overline{az}=a\overline z=aT(z),
so T is real-linear. But for the complex scalar i,
T(i\cdot1)=\overline i=-i,
while
iT(1)=i.
Thus T(i\cdot1)\ne iT(1), so T is not complex-linear.
7.2Explanation
Linearity depends on which field体たい supplies the scalars. The same map may be linear over \mathbb R but not over \mathbb C.
9証明しょうめい演習えんしゅう:核かくと像ぞうが部分空間ぶぶんくうかんであること
9.1問題もんだい
T:V\to W を線型写像せんけいしゃぞうとする。\ker T が V の部分空間ぶぶんくうかんであり、\operatorname{Im}T が W の部分空間ぶぶんくうかんであることを証明しょうめいせよ。
9.2解答かいとう
u,v\in\ker T、a,b をスカラーscalarとする。線型性せんけいせいより
T(au+bv)=aT(u)+bT(v)=0
である。よって au+bv\in\ker T である。
y_1,y_2\in\operatorname{Im}T なら、ある v_1,v_2 について y_1=T(v_1)、y_2=T(v_2) と書かける。このとき
ay_1+by_2=T(av_1+bv_2)
なので、ay_1+by_2\in\operatorname{Im}T である。
9.3解説かいせつ
核かくは 0 に潰つぶれる方向ほうこう、像ぞうは到達とうたつできる範囲はんいである。どちらも線型結合せんけいけつごうで閉とじることが本質ほんしつである。
9Proof exercise: the kernel核かく and image像ぞう are subspaces部分空間ぶぶんくうかん
9.1Problem
Let T:V\to W be a linear map. Prove that \ker T is a subspace of V and \operatorname{Im}T is a subspace of W.
9.2Answer
If u,v\in\ker T and a,b are scalars, then by linearity
T(au+bv)=aT(u)+bT(v)=0.
Thus au+bv\in\ker T.
If y_1,y_2\in\operatorname{Im}T, write y_1=T(v_1) and y_2=T(v_2). Then
ay_1+by_2=T(av_1+bv_2),
so ay_1+by_2\in\operatorname{Im}T.
9.3Explanation
The kernel is the set of directions collapsed to 0, and the image is the reachable range. Both are closed under linear combinations.