線形代数学14:ケーリー・ハミルトンの定理と行列のn乗

こんにちは、ひかりです。

今回は線形代数学からケーリー・ハミルトンの定理と行列の \( n \) 乗について解説していきます。

この記事では以下のことを紹介します。

  • ケーリー・ハミルトンの定理について
  • 行列の \( n \) 乗の求め方について
目次

ケーリー・ハミルトンの定理

ここでは、2次正方行列 \( A \) の \( n \) 乗 \( A^n \) の計算に有用なケーリー・ハミルトンの定理について解説します。

定理1 (ケーリー・ハミルトンの定理)

2次正方行列 \( A=\begin{pmatrix} a & b \\ c & d \end{pmatrix} \) に対して、次が成り立つ。

$$ A^2-(a+d)A+(ad-bc)E=O $$

定理1の証明(気になる方だけクリックしてください)

$$ \begin{align} &A^2-(a+d)A+(ad-bc)E \\ =&\begin{pmatrix} a & b \\ c & d \end{pmatrix}\begin{pmatrix} a & b \\ c & d \end{pmatrix}-(a+d)\begin{pmatrix} a & b \\ c & d \end{pmatrix}+(ad-bc)\begin{pmatrix} 1 & 0 \\ 0 & 1 \end{pmatrix} \\ =&\begin{pmatrix} a^2+bc & ab+bd \\ ac+cd & bc+d^2 \end{pmatrix}-\begin{pmatrix} a^2+ad & ab+bd \\ ac+cd & ad+d^2 \end{pmatrix}+\begin{pmatrix} ad-bc & 0 \\ 0 & ad-bc \end{pmatrix} \\ =&\begin{pmatrix} a^2+bc-a^2-ad+ad-bc & ab+bd-ab-bd \\ ac+cd-ac-cd & bc+d^2-ad-d^2+ad-bc \end{pmatrix} \\ =&\begin{pmatrix} 0 & 0 \\ 0 & 0 \end{pmatrix}=O \end{align} $$

ケーリー・ハミルトンの定理を利用する例を2つ紹介します。

例1

(1) \( A=\begin{pmatrix} -1 & -1 \\ 1 & 0 \end{pmatrix} \) のとき、 \( A^3 \) と \( A^{10} \) を求める。

ケーリー・ハミルトンの定理より、

$$ A^2-(-1+0)A+(0+1)E=O $$

よって、 \( A^2+A+E=O \) より、 \( A^2=-A-E \)

したがって、

$$ A^3=A(-A-E)=-A^2-A=-(-A-E)-A=E=\begin{pmatrix} 1 & 0 \\ 0 & 1 \end{pmatrix} $$

また、

$$ A^{10}=(A^3)^3A=E^3A=EA=A=\begin{pmatrix} -1 & -1 \\ 1 & 0 \end{pmatrix} $$


(2) \( A=\begin{pmatrix} a & -b \\ b & a \end{pmatrix} \) が

$$ A^2-4A+4E=O \tag{1} $$

をみたすとき、実数 \( a,b \) の値を求める。

ケーリー・ハミルトンの定理より、

$$ A^2-2aA+(a^2+b^2)E=O \tag{2} $$

式(2)-式(1)をすると、

$$ 2(a-2)A=(a^2+b^2-4)E \tag{3} $$

まず、 \( a\not=2 \) とすると、

$$ A=kE \quad \left(k=\frac{a^2+b^2-4}{2(a-2)}\right) \tag{4} $$

式(4)を式(1)に代入すると、 \( (kE)^2-4kE+4E=O \)

よって、 \( (k^2-4k+4)E=O \) より、 \( k^2-4k+4=0 \)

したがって、 \( k=2 \)

式(4)より、 \( A=2E=\begin{pmatrix} 2 & 0 \\ 0 & 2 \end{pmatrix} \left( =\begin{pmatrix} a & -b \\ b & a \end{pmatrix} \right) \) となり、 \( a=2, \ b=0 \) となるが、 \( a\not=2 \) に矛盾する。

よって、 \( a=2 \)となる。式(3)に代入すると、 \( 0=4+b^2-4=b^2 \) より \( b=0 \)

したがって、 \( a=2, \ b=0 \)

2次正方行列のn乗の求め方

それでは、いよいよ2次正方行列 \( A \) の \( n \) 乗 \( A^n \) の求め方をみていきます。

\( A^n \) が簡単に求まる4つのケース

定理2 (\( A^n \) の求め方)

(1) \( A^2=kA \quad (k:\text{実数}) \) のとき、 \( A^n=k^{n-1}A \quad (n=1,2,\dots) \)

(2) \( A=\begin{pmatrix} a & 0 \\ 0 & b \end{pmatrix} \) のとき、 \( A^n=\begin{pmatrix} a^n & 0 \\ 0 & b^n \end{pmatrix} \quad (n=1,2,\dots) \)

(3) \( A=\begin{pmatrix} 1 & a \\ 0 & 1 \end{pmatrix} \) のとき、 \( A^n=\begin{pmatrix} 1 & na \\ 0 & 1 \end{pmatrix} \quad (n=1,2,\dots) \)

(4) \( A=\begin{pmatrix} \cos\theta & -\sin\theta \\ \sin\theta & \cos\theta \end{pmatrix} \) のとき、 \( A^n=\begin{pmatrix} \cos n\theta & -\sin n\theta \\ \sin n\theta & \cos n\theta \end{pmatrix} \quad (n=1,2,\dots) \)

定理2の証明(気になる方だけクリックしてください)

すべて数学的帰納法で示すことができます。

(1) まず、 \( n=1 \) のときは、

$$ A^1=k^0\cdot A=1\cdot A=A $$

より、成り立ちます。

次に、 \( n=m \) で成り立つと仮定して、 \( n=m+1 \) で示します。

$$ A^{m+1}=A^m\cdot A=k^{m-1}A\cdot A=k^{m-1}A^2=k^{m-1}kA=k^mA $$

よって、数学的帰納法より、任意の自然数 \( n \) に対して(1)は成り立つ。


(2) まず、 \( n=1 \) のときは、

$$ A^1=\begin{pmatrix} a^1 & 0 \\ 0 & b^1 \end{pmatrix}=\begin{pmatrix} a & 0 \\ 0 & b \end{pmatrix} $$

より、成り立ちます。

次に、 \( n=m \) で成り立つと仮定して、 \( n=m+1 \) で示します。

$$ A^{m+1}=A^m\cdot A=\begin{pmatrix} a^m & 0 \\ 0 & b^m \end{pmatrix}\begin{pmatrix} a & 0 \\ 0 & b \end{pmatrix}=\begin{pmatrix} a^{m+1} & 0 \\ 0 & b^{m+1} \end{pmatrix} $$

よって、数学的帰納法より、任意の自然数 \( n \) に対して(2)は成り立つ。


(3) まず、 \( n=1 \) のときは、

$$ A^1=\begin{pmatrix} 1 & 1\cdot a \\ 0 & 1 \end{pmatrix}=\begin{pmatrix} 1 & a \\ 0 & 1 \end{pmatrix} $$

より、成り立ちます。

次に、 \( n=m \) で成り立つと仮定して、 \( n=m+1 \) で示します。

$$ A^{m+1}=A^m\cdot A=\begin{pmatrix} 1 & ma \\ 0 & 1 \end{pmatrix}\begin{pmatrix} 1 & a \\ 0 & 1 \end{pmatrix}=\begin{pmatrix} 1 & (m+1)a \\ 0 & 1 \end{pmatrix} $$

よって、数学的帰納法より、任意の自然数 \( n \) に対して(3)は成り立つ。


(4) まず、 \( n=1 \) のときは、

$$ A^1=\begin{pmatrix} \cos 1\cdot \theta & -\sin 1\cdot \theta \\ \sin 1\cdot \theta & \cos 1\cdot \theta \end{pmatrix}=\begin{pmatrix} \cos \theta & -\sin \theta \\ \sin \theta & \cos \theta \end{pmatrix} $$

より、成り立ちます。

次に、 \( n=m \) で成り立つと仮定して、 \( n=m+1 \) で示します。三角関数の加法定理より、

$$ \begin{align} A^{m+1}&=A^m\cdot A=\begin{pmatrix} \cos m\theta & -\sin m\theta \\ \sin m\theta & \cos m\theta \end{pmatrix}\begin{pmatrix} \cos \theta & -\sin \theta \\ \sin \theta & \cos \theta \end{pmatrix} \\ &=\begin{pmatrix} \cos m\theta\cos \theta-\sin m\theta\sin \theta & -(\sin m\theta\cos \theta+\cos m\theta\sin \theta) \\ \sin m\theta\cos \theta+\cos m\theta\sin \theta & \cos m\theta\cos\theta-\sin m\theta\sin \theta \end{pmatrix} \\ &=\begin{pmatrix} \cos (m+1)\theta & -\sin (m+1)\theta \\ \sin (m+1)\theta & \cos (m+1)\theta \end{pmatrix} \end{align} $$

よって、数学的帰納法より、任意の自然数 \( n \) に対して(4)は成り立つ。

例2

(1) \( A=\begin{pmatrix} 2 & 1 \\ 4 & 2 \end{pmatrix} \) とおくと、

\( |A|=\begin{vmatrix} 2 & 1 \\ 4 & 2 \end{vmatrix}=0 \) より、 \( A \) は正則ではない。

よって、ケーリー・ハミルトンの定理を用いると、 \( A^2-(2+2)A+0E=O \) より、 \( A^2=4A \)

したがって、

$$ A^n=4^{n-1}A=4^{n-1}\begin{pmatrix} 2 & 1 \\ 4 & 2 \end{pmatrix}=\begin{pmatrix} 2\cdot 4^{n-1} & 4^{n-1} \\ 4^n & 2\cdot 4^{n-1} \end{pmatrix} $$


(2) \( A=\begin{pmatrix} 2 & 0 \\ 0 & -3 \end{pmatrix} \) とおくと、 \( A^n=\begin{pmatrix} 2^n & 0 \\ 0 & (-3)^n \end{pmatrix} \)


(3) \( A=\begin{pmatrix} 1 & 2 \\ 0 & 1 \end{pmatrix} \) とおくと、 \( A^n=\begin{pmatrix} 1 & 2n \\ 0 & 1 \end{pmatrix} \)


(4) \( A=\begin{pmatrix} 1 & -\sqrt{3} \\ \sqrt{3} & 1 \end{pmatrix} \) とおくと、

$$ A=2\begin{pmatrix} \frac{1}{2} & -\frac{\sqrt{3}}{2} \\ \frac{\sqrt{3}}{2} & \frac{1}{2} \end{pmatrix}=2\begin{pmatrix} \cos\frac{\pi}{3} & -\sin\frac{\pi}{3} \\ \sin\frac{\pi}{3} & \cos\frac{\pi}{3} \end{pmatrix} $$

よって、 \( A^n=2^n\begin{pmatrix} \cos\frac{n\pi}{3} & -\sin\frac{n\pi}{3} \\ \sin\frac{n\pi}{3} & \cos\frac{n\pi}{3} \end{pmatrix} \)

その他のケース

上の4つのケースに当てはまらない行列 \( A \) に対して、ケーリー・ハミルトンの定理を用いて \( A^n \) を求めることを考えます。

例を通じて見ていきましょう。

例3

\( A=\begin{pmatrix} 4 & -5 \\ 1 & -2 \end{pmatrix} \) に対して、 \( A^n \) を求める。

ケーリー・ハミルトンの定理より、

$$ A^2-(4-2)A+(-8+5)E=A^2-2A-3E=O $$

ここで、 \( A \) を \( x \) 、 \( E \) は \( 1 \) 、 \( O \) は \( 0 \) にした方程式(特性方程式という)を考えると、

$$ x^2-2x-3=0 \tag{5} $$

\( x^n \) を(5)の左辺で割った商を \( Q(x) \) 、余りを \( ax+b \) とおくと、

$$ \frac{x^n}{x^2-2x-3}=Q(x)+ax+b $$

よって、

$$ \begin{align} x^n&=(x^2-2x-3)Q(x)+ax+b \\ &=(x-3)(x+1)Q(x)+ax+b \tag{6} \end{align} $$

(6)に \( x=-1,3 \) を代入すると、

$$ -a+b=(-1)^n \tag{7} $$

$$ 3a+b=3^n \tag{8} $$

(8)-(7)とすると、 \( 4a=3^n-(-1)^n \) となる。よって、

$$ a=\frac{1}{4}\{ 3^n-(-1)^n\} \tag{9} $$

また、(7)×3+(4)とすると、 \( 4b=3\cdot(-1)^n+3^n \) となる。よって、

$$ b=\frac{1}{4}\{ 3^n+3\cdot(-1)^n \} \tag{10} $$

ここで、 \( A^k \) や \( E \) の積は交換可能なので、 \( A^k \) についても(6)と同じ式が成り立つ。

$$ A^n=(A^2-2A-3E)\cdot Q(A)+aA+bE $$

ここで、 \( A^2-2A-3E=O \) より、

$$ A^n=aA+bE \tag{11} $$

(11)に(9)と(10)を代入すると、

$$ \begin{align} A^n&=\frac{1}{4}\{ 3^n-(-1)^n \} \begin{pmatrix} 4 & -5 \\ 1 & -2 \end{pmatrix}+\frac{1}{4}\{ 3^n+3\cdot(-1)^n\} \begin{pmatrix} 1 & 0 \\ 0 & 1 \end{pmatrix} \\ &=\frac{1}{4}\begin{pmatrix} 4\{ 3^n-(-1)^n \} +3^n+3\cdot(-1)^n & -5\{ 3^n-(-1)^n \} \\ 3^n-(-1)^n & -2\{ 3^n-(-1)^n\}+3^n+3\cdot(-1)^n \end{pmatrix} \\ &=\frac{1}{4}\begin{pmatrix} 5\cdot3^n-(-1)^n & -5\{ 3^n-(-1)^n \} \\ 3^n-(-1)^n & -3^n+5\cdot(-1)^n \end{pmatrix} \quad (n=1,2,\dots) \end{align} $$

\( P^{-1}AP \) 型に変換できるケース

2次正方行列 \( A \) に対して、もし、 \( P^{-1}AP=\begin{pmatrix} a & 0 \\ 0 & b \end{pmatrix} \) (対角行列) となる行列 \( P \) が存在したとすると、それを利用して \( A^n \) を計算することができます。

\( P^{-1}AP=\begin{pmatrix} a & 0 \\ 0 & b \end{pmatrix} \) とします。両辺を \( n \) 乗すると、

$$ (P^{-1}AP)^n=\begin{pmatrix} a & 0 \\ 0 & b \end{pmatrix}^n=\begin{pmatrix} a^n & 0 \\ 0 & b^n \end{pmatrix} \quad (\text{簡単に求まるケース(2)より}) $$

ここで、

$$ \begin{align} (P^{-1}AP)^n&=(P^{-1}A\color{red}{P})(\color{red}{P^{-1}}A\color{red}{P})(\color{red}{P^{-1}}A\color{red}{P})\cdots (\color{red}{P^{-1}}AP) \\ &=P^{-1}A\color{red}{E}A\color{red}{E}A\cdots \color{red}{E}AP \\ &=P^{-1}A^nP \end{align} $$

したがって、

$$ P^{-1}A^nP=\begin{pmatrix} a^n & 0 \\ 0 & b^n \end{pmatrix} \tag{12} $$

(12)の両辺に左から \( P \)、右から \( P^{-1} \) をかけると、 \( A^n=P\begin{pmatrix} a^n & 0 \\ 0 & b^n \end{pmatrix}P^{-1} \) となり、 \( A^n \) が求まります。

例4

$$ A=\begin{pmatrix} 2 & 1 \\ 2 & 3 \end{pmatrix}, \ P=\begin{pmatrix} 1 & 1 \\ -1 & 2 \end{pmatrix} $$

のとき、 \( A^n \) を求める。

\( |P|=\begin{vmatrix} 1 & 1 \\ -1 & 2 \end{vmatrix}=3 \) より、 \( P^{-1} \) が存在して、 \( P^{-1}=\frac{1}{3}\begin{pmatrix} 2 & -1 \\ 1 & 1 \end{pmatrix} \)

よって、

$$ \begin{align} P^{-1}AP&=\frac{1}{3}\begin{pmatrix} 2 & -1 \\ 1 & 1 \end{pmatrix}\begin{pmatrix} 2 & 1 \\ 2 & 3 \end{pmatrix}\begin{pmatrix} 1 & 1 \\ -1 & 2 \end{pmatrix}=\frac{1}{3}\begin{pmatrix} 2 & -1 \\ 4 & 4 \end{pmatrix}\begin{pmatrix} 1 & 1 \\ -1 & 2 \end{pmatrix} \\ &=\frac{1}{3}\begin{pmatrix} 3 & 0 \\ 0 & 12 \end{pmatrix}=\begin{pmatrix} 1 & 0 \\ 0 & 4 \end{pmatrix} \end{align} $$

最両辺を \( n \) 乗すると、

$$ P^{-1}AP=\begin{pmatrix} 1 & 0 \\ 0 & 4^n \end{pmatrix}=\begin{pmatrix} 1 & 0 \\ 0 & 2^{2n} \end{pmatrix} \tag{13} $$

(13)の両辺に左から \( P \)、右から \( P^{-1} \) をかけると、

$$ \begin{align} A^n&=\begin{pmatrix} 1 & 1 \\ -1 & 2 \end{pmatrix}\begin{pmatrix} 1 & 0 \\ 0 & 2^{2n} \end{pmatrix}\cdot\frac{1}{3}\begin{pmatrix} 2 & -1 \\ 1 & 1 \end{pmatrix}=\frac{1}{3}\begin{pmatrix} 1 & 2^{2n} \\ -1 & 2^{2n+1} \end{pmatrix}\begin{pmatrix} 2 & -1 \\ 1 & 1 \end{pmatrix} \\ &=\frac{1}{3}\begin{pmatrix} 2^{2n}+2 & 2^{2n}-1 \\ 2^{2n+1}-2 & 2^{2n+1}+1 \end{pmatrix} \quad (n=1,2,\dots) \end{align} $$

今回はここまでです。お疲れ様でした。また次回にお会いしましょう。

目次