markdown
置換による行列式の定義md bd34096
lecture/math/linear-algebra/determinants-by-permutations.lecture.n.md
Download PDF

置換ちかんpermutationによる行列式ぎょうれつしきdeterminantの定義ていぎ

date2026-07-14document_iddoc_e71393960578c58ac4e3a17955dff07adescription置換と符号を用いて一般の行列式を定義し、多重線型性・交代性・三角行列の公式・積の公式がどこから出るかを証明する講義である。prerequisites正方行列 / 順列と置換の基本 / 線型性 / 行列式の直感type講義content_typelecturestatusactiverelateddata/lecture/math/linear-algebra/determinants.lecture.n.md / data/lecture/math/linear-algebra/determinant-computation-rules.lecture.n.md / data/lecture/math/linear-algebra/cofactor-expansion-and-invertibility.lecture.n.md / data/lecture/math/probability/counting-permutations-and-combinations.lecture.n.md / data/exercise/math/linear-algebra/determinants-and-invertibility.exercise.n.md
mathlinear-algebraundergraduatelecturedeterminant

Defining the determinant行列式ぎょうれつしき by permutations置換ちかん

1導入どうにゅう

この講義こうぎで重要じゅうようなのは、行列式ぎょうれつしきdeterminantを「n 個この成分せいぶんを重複ちょうふくなく選えらび、その選えらび方かたの向むきorientationに応おうじて符号ふごうを付つけて足たす量りょう」として定義ていぎすることである。

2×2 の公式こうしき det(abcd)=ad-bc では、ad と bc の 2 項こうしか見みえない。しかし一般いっぱんの n×n 行列ぎょうれつmatrixでは、各行かくぎょうから 1 個こずつ、しかも各列かくれつも 1 回かいずつ使つかう選択せんたくをすべて足たす必要ひつようがある。その選択せんたくを記録きろくする道具どうぐが置換ちかんpermutationである。

1Introduction

The key idea is to define the determinant行列式ぎょうれつしき as a signed sum over all ways to choose one entry from each row and one entry from each column. A permutation置換ちかん records such a choice, and its sign records whether the choice preserves or reverses orientation.

2用語ようごと定義ていぎ

2.1置換ちかんpermutation

置換ちかんpermutationとは、集合しゅうごう {1,…,n} から自分自身じぶんじしんへの全単射ぜんたんしゃbijectionである。全単射ぜんたんしゃbijectionとは、重複ちょうふくも抜ぬけもなく対応たいおうさせる写像しゃぞうmapである。

たとえば n=3 のとき、

σ(1)=2,σ(2)=3,σ(3)=1

は置換ちかんpermutationである。行ぎょう 1,2,3 に対たいして、列れつ 2,3,1 を選えらぶという意味いみである。

{1,…,n} のすべての置換ちかんpermutationの集合しゅうごうを Sn と書かく。この講義こうぎでは、Sn を「すべての並ならべ替かえを集あつめたもの」として使つかえばよい。

2.2反転数はんてんすうinversion numberと符号ふごうsign

置換ちかんpermutation σ に対たいして、i<j なのに σ(i)>σ(j) となる組くみ (i,j) を反転はんてんinversionと呼よぶ。反転数はんてんすうinversion numberは、そのような組くみの個数こすうである。

反転数はんてんすうinversion numberが偶数ぐうすうなら偶置換ぐうちかんeven permutation、奇数きすうなら奇置換きちかんodd permutationである。置換ちかんpermutationの符号ふごうsignを

sgn(σ)=(-1)反転数

で定義ていぎする。偶置換ぐうちかんなら +1、奇置換きちかんなら -1 である。

2.3置換ちかんpermutationによる行列式ぎょうれつしきdeterminantの定義ていぎ

A=(aij) を n×n 行列ぎょうれつmatrixとする。行列式ぎょうれつしきdeterminantを

detA=∑σ∈Snsgn(σ)a1σ(1)a2σ(2)⋯anσ(n)

で定義ていぎする。

この式しきで aiσ(i) は、第だい i 行ぎょうrowから第だい σ(i) 列れつcolumnの成分せいぶんを選えらぶことを表あらわす。σ が置換ちかんpermutationであるため、同おなじ列れつcolumnを 2 回かい使つかうことはない。

2Terms and definitions

A permutation置換ちかん of {1,…,n} is a bijection全単射ぜんたんしゃ from the set to itself. The set of all such permutations is denoted by Sn.

An inversion反転はんてん of a permutation σ is a pair i<j with σ(i)>σ(j). The sign is

sgn(σ)=(-1)numberofinversions.

For an n×n matrix A=(aij), define

detA=∑σ∈Snsgn(σ)a1σ(1)a2σ(2)⋯anσ(n).

Each term chooses one entry from each row and one entry from each column.

3方針ほうしん

この定義ていぎから証明しょうめいしたい性質せいしつは次つぎの 5 点てんである。

  1. 単位行列たんいぎょうれつidentity matrixでは detI=1 である。
  2. 行列式行列式ぎょうれつしきは各行かくぎょうまたは各列かくれつについて線型せんけいlinearである。
  3. 2 本ほんの行ぎょうrowまたは列れつcolumnを交換こうかんすると符号ふごうsignが反転はんてんする。
  4. 2 本ほんの行ぎょうrowまたは列れつcolumnが等ひとしいと行列式ぎょうれつしきdeterminantは 0 である。
  5. この 3 原理げんり、すなわち多重線型性たじゅうせんけいせいmultilinearity・交代性こうたいせいalternation・detI=1 は、行列式ぎょうれつしきdeterminantを一意いちいに決定けっていする。

先さきに置換ちかんpermutationの式しきを定義ていぎとして置おき、そこから計算規則けいさんきそくを導みちびく。計算規則けいさんきそくを暗記あんきするのではなく、「選えらび方かたを入いれ替かえると符号ふごうsignが変かわる」ことを根拠こんきょにする。

3Plan

We first take the permutation formula as the definition. Then we prove the main properties: normalization, multilinearity, sign change under row or column swaps, vanishing when two rows or columns coincide, uniqueness from these principles, triangular formulas, and multiplicativity.

4直感的ちょっかんてきな説明せつめい

行列式ぎょうれつしきdeterminantの各項かくこうは、行列ぎょうれつの中なかから「各行かくぎょう 1 個こ、各列かくれつ 1 個こ」を選えらぶ方法ほうほうである。同おなじ列れつcolumnを 2 回かい選えらぶと、体積たいせきを作つくるための独立どくりつした方向ほうこうが不足ふそくする。そのため、置換ちかんpermutationだけを使つかって選択せんたくする。

符号ふごうは、選択せんたくした列れつcolumnの順序じゅんじょが標準ひょうじゅんの順序じゅんじょからどれだけ入いれ替かわったかを記録きろくする。入いれ替かえが偶数回ぐうすうかいなら向むきは保たもたれ、奇数回きすうかいなら向むきは反転はんてんする。この「向むきの情報じょうほう」がなければ、面積めんせきや体積たいせきの大おおきさだけは読よめても、行列式ぎょうれつしきdeterminantの正負せいふが説明せつめいできない。

2×2 では、選えらび方かたは 2 通とおりだけである。列れつを 1,2 の順じゅんに選えらぶと ad、列れつを 2,1 の順じゅんに選えらぶと bc であり、後者こうしゃは 1 回かいの交換こうかんなので負ふの符号ふごうsignを持もつ。したがって ad-bc が現あらわれる。

4Intuitive explanation

Each determinant term chooses one entry from each row and each column. The permutation makes the no-repetition rule explicit. The sign records whether the chosen column order is an even or odd reordering of the standard order.

For a 2×2 matrix there are only two choices. The identity choice gives ad, and the swapped choice gives bc with negative sign, producing ad-bc.

5厳密げんみつな説明せつめい

5.11. 2×2 の公式こうしきが戻もどる

A=(abcd)

とする。S2 の置換ちかんpermutationは 2 個こである。

σsgn(σ)項こう
σ(1)=1,σ(2)=2+1ad
σ(1)=2,σ(2)=1-1bc

したがって

detA=ad-bc

である。置換ちかんによる定義ていぎは、2×2 の公式こうしきを含ふくんでいる。

5.22. 単位行列たんいぎょうれつidentity matrixの行列式ぎょうれつしきdeterminant

単位行列たんいぎょうれつidentity matrix I では、対角成分たいかくせいぶんだけが 1 で、非対角成分ひたいかくせいぶんは 0 である。置換ちかんpermutation σ が恒等置換こうとうちかんでないなら、ある i について σ(i)≠i となり、その項こうには 0 が含ふくまれる。

したがって非零ひぜろの項こうは恒等置換こうとうちかんだけであり、

detI=1

である。

5.33. 多重線型性たじゅうせんけいせいmultilinearity

行列式ぎょうれつしきdeterminantを行ぎょうrowの関数かんすうとして見みる。ある第だい r 行ぎょうrowだけを u+v に置おき換かえると、各項かくこうの第だい r 成分せいぶんだけが

(u+v)σ(r)=uσ(r)+vσ(r)

となる。したがって和わを分配ぶんぱいすると、行列式ぎょうれつしきdeterminantは「第だい r 行ぎょうrowを u にした行列ぎょうれつの行列式ぎょうれつしきdeterminant」と「v にした行列ぎょうれつの行列式ぎょうれつしきdeterminant」の和わになる。

また第だい r 行ぎょうrowを λu にすると、すべての項こうで λ が 1 個こだけ外そとへ出でる。よって行列式ぎょうれつしきdeterminantは第だい r 行ぎょうrowについて線型せんけいlinearである。任意にんいの行ぎょうrowで同おなじため、多重線型性たじゅうせんけいせいmultilinearityが成立せいりつする。

列れつcolumnについても同様どうようである。列れつの場合ばあいは、選えらばれる成分せいぶんが aiσ(i) の列番号れつばんごうによって決きまるだけで、各項かくこうにその列れつcolumnから来くる因子いんしが 1 個こだけ現あらわれる。

5.44. 行交換ぎょうこうかんrow swapで符号ふごうsignが反転はんてんする

τ を第だい p 行ぎょうrowと第だい q 行ぎょうrowを交換こうかんする置換ちかんpermutationとする。B を A の第だい p 行ぎょうrowと第だい q 行ぎょうrowを交換こうかんした行列ぎょうれつとする。

B の行列式ぎょうれつしきdeterminantの項こうを A の行ぎょうで書かき直なおすと、置換ちかんpermutationは σ から σ∘τ へ変かわる。行交換ぎょうこうかん τ は 1 回かいの交換こうかんなので、

sgn(σ∘τ)=-sgn(σ)

である。σ が Sn 全体ぜんたいを動うごくと、σ∘τ も Sn 全体ぜんたいを動うごく。したがって

detB=-detA

である。

列交換れつこうかんcolumn swapでも同おなじ結論けつろんになる。列れつを交換こうかんすることは、置換ちかんの値あたいを交換こうかんすることであり、やはり符号ふごうsignを 1 回かい反転はんてんさせる。

5.55. 2 本ほんの行ぎょうrowが等ひとしいと 0 になる

2 本ほんの行ぎょうrowが等ひとしい行列ぎょうれつを A とする。その 2 本ほんを交換こうかんしても行列ぎょうれつそのものは変かわらない。一方いっぽう、前節ぜんせつより行列式ぎょうれつしきdeterminantの符号ふごうsignは反転はんてんする。

したがって

detA=-detA

である。実数じっすうまたは複素数ふくそすうの上うえでは、これは detA=0 を意味いみする。2 本ほんの列れつcolumnが等ひとしい場合ばあいも同様どうようである。

5.66. 特徴とくちょうづけの定理ていり

定理ていり
行ぎょうrowに関かんして多重線型たじゅうせんけいmultilinearで、2 本ほんの行ぎょうrowが等ひとしいと 0 になり、F(I)=1 を満みたす関数かんすう F は、行列式ぎょうれつしきdeterminantに等ひとしい。

証明しょうめい
A の第だい i 行ぎょうrowを、標準基底ひょうじゅんきてい e1,…,en を用もちいて

Ri=∑j=1naijej

と書かく。F の多重線型性たじゅうせんけいせいmultilinearityを使つかって展開てんかいすると、F(A) は

F(ej1,ej2,…,ejn)

の形かたちの項こうの和わになる。もし j1,…,jn に重複ちょうふくがあれば、2 本ほんの行ぎょうrowが等ひとしいため、その項こうは 0 である。

したがって残のこるのは、(j1,…,jn) が置換ちかんpermutationになっている場合ばあいだけである。このとき

F(eσ(1),…,eσ(n))=sgn(σ)F(e1,…,en)=sgn(σ)

である。ここで、行ぎょうの交換こうかんで符号ふごうsignが変かわることと F(I)=1 を使つかった。よって

F(A)=∑σ∈Snsgn(σ)a1σ(1)⋯anσ(n)=detA

である。これで一意性いちいせいuniquenessが示しめされた。

5.77. 三角行列さんかくぎょうれつtriangular matrixの行列式ぎょうれつしきdeterminant

A が上三角行列うえさんかくぎょうれつupper triangular matrixなら、i>j の成分せいぶん aij は 0 である。ある項こう

a1σ(1)a2σ(2)⋯anσ(n)

が非零ひぜろであるには、すべての i について σ(i)[PARSE ERROR: Undefined("Command(\"ge\")")]i でなければならない。しかし σ(1),…,σ(n) は 1,…,n の並ならべ替かえなので、和わは常つねに

1+2+⋯+n

である。すべての σ(i)[PARSE ERROR: Undefined("Command(\"ge\")")]i で和わが同おなじなら、σ(i)=i がすべての i で成立せいりつする。非零ひぜろに寄与きよするのは恒等置換こうとうちかんだけである。

したがって

detA=a11a22⋯ann

である。下三角行列したさんかくぎょうれつlower triangular matrixでも同様どうように対角成分たいかくせいぶんの積せきになる。

5.88. 積せきの公式こうしき

定理ていり
n×n 行列ぎょうれつmatrix A,B について、

det(AB)=detAdetB

である。

証明しょうめい
B を固定こていし、F(A)=det(AB) と置おく。AB の第だい i 行ぎょうrowは、A の第だい i 行ぎょうrowを係数けいすうとして、B の行ぎょうrowを線型結合せんけいけつごうしたものである。したがって F(A) は A の行ぎょうrowに関かんして多重線型たじゅうせんけいmultilinearである。

また A の 2 本ほんの行ぎょうrowが等ひとしいと、AB の対応たいおうする 2 本ほんの行ぎょうrowも等ひとしいため、F(A)=0 である。さらに

F(I)=det(IB)=detB

である。

特徴とくちょうづけの定理ていりより、F(A) は detA の定数倍ていすうばいであり、その定数ていすうは F(I)=detB である。よって

F(A)=detAdetB

すなわち det(AB)=detAdetB である。

この公式こうしきは、可逆性かぎゃくせいinvertibilityとの接続せつぞくで重要じゅうようである。A が可逆かぎゃくinvertibleなら AA-1=I なので、

detAdet(A-1)=1

となり、detA≠0 が従したがう。

5Precise explanation

The formula recovers the 2×2 determinant because the two permutations of {1,2} give ad and bc with opposite signs.

The identity matrix has determinant 1, since only the identity permutation contributes a nonzero term.

The determinant is multilinear in rows and columns because each term contains exactly one factor from each row and each column. Swapping two rows or columns changes the permutation sign, so the determinant changes sign. If two rows or columns are equal, the swap leaves the matrix unchanged but also changes the sign of the determinant, forcing the determinant to be zero over the real or complex numbers.

The permutation formula is also the unique function that is multilinear, alternating, and normalized by detI=1. Expanding each row in the standard basis leaves only terms indexed by permutations, and the alternating property supplies the sign.

For triangular matrices, only the identity permutation can contribute a nonzero term, so the determinant is the product of the diagonal entries. The multiplicative formula

det(AB)=detAdetB

follows by applying the uniqueness characterization to the function F(A)=det(AB).

6具体例ぐたいれい

A=(120034506)

を置換ちかんpermutationの定義ていぎで計算けいさんする。S3 には 6 個この置換ちかんpermutationがある。

\begin{aligned} \det A &= a_{11}a_{22}a_{33} +a_{12}a_{23}a_{31} +a_{13}a_{21}a_{32}\\ &\quad -a_{13}a_{22}a_{31} -a_{12}a_{21}a_{33} -a_{11}a_{23}a_{32} \end{aligned}

である。ここへ値あたいを代入だいにゅうすると、

detA=1·3·6+2·4·5+0·0·0-0·3·5-2·0·6-1·4·0=58

となる。

この例れいで確認かくにんすべき点てんは、6 項こうを丸暗記まるあんきすることではない。各項かくこうは「各行かくぎょう 1 個こ、各列かくれつ 1 個こ」を満みたす選択せんたくであり、符号ふごうsignはその選択せんたくの置換ちかんpermutationの偶奇ぐうきparityから決きまっている。

6Worked example

For

A=(120034506),

the six permutation terms give

\begin{aligned} \det A &= a_{11}a_{22}a_{33} +a_{12}a_{23}a_{31} +a_{13}a_{21}a_{32}\\ &\quad -a_{13}a_{22}a_{31} -a_{12}a_{21}a_{33} -a_{11}a_{23}a_{32}\\ &=58. \end{aligned}

The point is not to memorize six terms mechanically; each term corresponds to one permutation choice.

7別べつの見方みかた

7.1代数的だいすうてきな見方みかた

置換ちかんpermutationの定義ていぎは、行列式ぎょうれつしきdeterminantを完全かんぜんに成分せいぶんで書かく方法ほうほうである。抽象的ちゅうしょうてきな「体積たいせきvolume」を使つかわずに、有限個ゆうげんこの積せきと和わだけで定義ていぎできる。証明しょうめいを成分せいぶんへ戻もどしたいときに有効ゆうこうである。

7.2幾何的きかてきな見方みかた

符号ふごうsignは、向むき付づけられた体積たいせきoriented volumeの向むきを表あらわす。列れつを 1 回かい交換こうかんすると、基準きじゅんの向むきが反転はんてんする。そのため行列式ぎょうれつしきdeterminantの符号ふごうも反転はんてんする。

7.3計算けいさんとしての見方みかた

置換ちかんpermutationの定義ていぎは概念がいねんの根拠こんきょとして重要じゅうようである。しかし n! 個この項こうを持もつため、大おおきい行列ぎょうれつmatrixの直接計算ちょくせつけいさんには向むかない。実際じっさいの計算けいさんでは、余因子展開よいんしてんかいcofactor expansionや行基本変形ぎょうきほんへんけいelementary row operationを使つかう。

data/lecture/math/linear-algebra/determinant-computation-rules.lecture.n.md data/lecture/math/linear-algebra/cofactor-expansion-and-invertibility.lecture.n.md

7Other viewpoints

Algebraically, the permutation formula gives a complete componentwise definition. Geometrically, the sign records oriented volume. Computationally, the formula is usually not efficient for large matrices because it has n! terms; row reduction and cofactor expansion are usually better for calculation.

data/lecture/math/linear-algebra/determinant-computation-rules.lecture.n.md data/lecture/math/linear-algebra/cofactor-expansion-and-invertibility.lecture.n.md

8見分みわけ方かた

  • 行列式行列式ぎょうれつしきの性質せいしつを証明しょうめいしたいときは、多重線型性たじゅうせんけいせいmultilinearity・交代性こうたいせいalternation・detI=1 の 3 条件じょうけんから攻せめる。
  • 行列式行列式ぎょうれつしきの公式こうしきそのものの根拠こんきょを確認かくにんしたいときは、置換ちかんpermutationの定義ていぎへ戻もどる。
  • 具体的ぐたいてきな数値計算すうちけいさんでは、n! 項こうを直接ちょくせつ展開てんかいするより、零れいの多おおい行ぎょうrow・列れつcolumnでの余因子展開よいんしてんかいcofactor expansion、または行基本変形ぎょうきほんへんけいelementary row operationを選えらぶ。

8How to recognize when to use this

Use the permutation definition when proving determinant properties or checking where the signs come from. For numerical computation, use cofactor expansion or row operations unless the matrix is very small.

9どこまで成立せいりつするか

この講義こうぎでは、成分せいぶんが実数じっすうまたは複素数ふくそすうである n×n 行列ぎょうれつmatrixを扱あつかった。正方行列せいほうぎょうれつでない場合ばあい、各行かくぎょうと各列かくれつから 1 個こずつ選えらんで完全かんぜんな置換ちかんpermutationを作つくれないため、ここでの行列式ぎょうれつしきdeterminantは定義ていぎされない。

また、この講義こうぎの証明しょうめいは実数じっすう・複素数ふくそすうの通常つうじょうの計算けいさんを前提ぜんていにしている。抽象代数ちゅうしょうだいすうで扱あつかうより一般いっぱんの数体系すうたいけいでは同おなじ式しきを使つかえる場合ばあいもあるが、この時点じてんの線型代数せんけいだいすうlinear algebraでは必要ひつようない。

9Scope

Here the entries are real or complex numbers and the matrix is square. Rectangular matrices do not have this determinant because the permutation choice requires the same number of rows and columns. More general number systems are not needed at this point in the linear algebra sequence.

10最終形さいしゅうけい

[PARSE ERROR: Undefined("Command(\"boxed\")")]detA=∑σ∈Snsgn(σ)a1σ(1)a2σ(2)⋯anσ(n)
[PARSE ERROR: Undefined("Command(\"boxed\")")]detI=1,det(AB)=detAdetB
[PARSE ERROR: Undefined("Command(\"boxed\")")]multilinearityalternation+normalizationdeterminethedeterminant[PARSE ERROR: Undefined("RBrace")]

10Final forms

[PARSE ERROR: Undefined("Command(\"boxed\")")]detA=∑σ∈Snsgn(σ)a1σ(1)a2σ(2)⋯anσ(n)
[PARSE ERROR: Undefined("Command(\"boxed\")")]detI=1,det(AB)=detAdetB
[PARSE ERROR: Undefined("Command(\"boxed\")")]multilinearityalternation+normalizationdeterminethedeterminant[PARSE ERROR: Undefined("RBrace")]

11一言ひとことでいうと

  • 置換ちかんpermutationは、行列式ぎょうれつしきdeterminantの各項かくこうで「各行かくぎょう・各列かくれつから 1 個こずつ選えらぶ」ことを記録きろくする。
  • 符号ふごうsignは、その選択せんたくが向むきorientationを保たもつか反転はんてんするかを記録きろくする。
  • この定義ていぎから、多重線型性たじゅうせんけいせいmultilinearity、交代性こうたいせいalternation、計算規則けいさんきそく、積せきの公式こうしきが導みちびかれる。

12演習えんしゅうリンク

data/exercise/math/linear-algebra/determinants-and-invertibility.exercise.n.md

11In one sentence

  • A permutation records one entry chosen from each row and column.
  • Its sign records orientation.
  • The determinant rules follow from this signed sum.

12Exercise links

data/exercise/math/linear-algebra/determinants-and-invertibility.exercise.n.md
raw .n.md をコピー
loc をコピー (filepath:line ~ line)
copy share link
copy encoded share link
path をコピー
copy share link
copy encoded share link
copy share link
copy encoded share link
タブを全て閉じる