선형대수학 입문/고유값과 고유벡터 문서 원본 보기
←
선형대수학 입문/고유값과 고유벡터
둘러보기로 이동
검색으로 이동
문서 편집 권한이 없습니다. 다음 이유를 확인해주세요:
요청한 명령은 다음 권한을 가진 사용자에게 제한됩니다:
사용자
.
문서의 원본을 보거나 복사할 수 있습니다.
== 모티브 == 대각화, 고유값(eigenvalue), 고유벡터(eigenvector)를 논의 하기 전에, 어떤 모티브가 있었는지 봅시다. {{인용문| '''예시 (대각행렬의 거듭제곱 공식)''' <math>D=\begin{pmatrix}3&0\\0&-5\end{pmatrix}</math>를 생각하자. <math display=block> \begin{pmatrix}3&0\\0&-5\end{pmatrix}\begin{pmatrix}3^k&0\\0&(-5)^k\end{pmatrix} =\begin{pmatrix}3^{k+1}&0\\0&(-5)^{k+1}\end{pmatrix}, </math> 이므로, 각각의 양의 정수 <math>n</math>에 대해서 <math>D=\begin{pmatrix}3^n&0\\0&(-5)^n\end{pmatrix}</math>이라는 대각행렬의 거듭제곱 공식을 귀납법으로 증명할 수 있다.}} {{인용문|'''예시''' <math>P=\begin{pmatrix}1&1\\2&3\\\end{pmatrix}</math>, <math>D=\begin{pmatrix}3&0\\0&-5\end{pmatrix}</math>라고 하자. 여기서 <math>P^{-1}=\begin{pmatrix}3&-1\\-2&1\\\end{pmatrix}</math>임을 계산할 수 있다. 행렬 <math>A</math>를 <math>A=PDP^{-1}=\begin{pmatrix}19&-8\\48&-21\\\end{pmatrix}</math>라고 하자. 그러면 <math display=block> \begin{align} A^n&=(PDP^{-1})^n=\underbrace{(PD{\color{blue}P^{-1}})({\color{blue}P}DP^{-1})\cdots(PD{\color{brown}P^{-1}})({\color{brown}P}DP^{-1})}_{n\;PDP^{-1}\text{'s}}\\ &=PD(\underbrace{\color{blue}P^{-1}P}_{\color{blue}I})DP^{-1}\cdots PD(\underbrace{\color{brown}P^{-1}P}_{\color{brown}I})DP^{-1}\\ &=PD\underbrace{{\color{blue}I}D\cdots {\color{brown}I}D}_{n-1\;ID\text{'s}}P^{-1}\\ &=P\underbrace{DD\cdots D}_{n\;D\text{'s}}P^{-1}\\ &=PD^nP^{-1}\\ &=P\begin{pmatrix}3^n&0\\0&(-5)^n\end{pmatrix}P^{-1}\qquad\text{(위 의 예 시 )}\\ &=\begin{pmatrix}1&1\\2&3\\\end{pmatrix}\begin{pmatrix}3^n&0\\0&(-5)^n\end{pmatrix}\begin{pmatrix}3&-1\\-2&1\\\end{pmatrix}\\ &=\begin{pmatrix} 3^{n+1}-2(-5)^n&(-5)^n-3^n\\ 6(3^n)-6(-5)^n&3(-5)^n-2(3^n)\\ \end{pmatrix} \end{align} </math>}} 우리는 가역행렬 <math>P</math>와 대각행렬 <math>D</math>의 곱인 <math>PDP^{-1}</math>라는 조금 특수한 행렬에 대한 예시를 통해서 거듭제곱이 편리하게 계산될 수 있음을 보았습니다. 당연하게도, 이렇게 편한 도구가 주어졌으니, 주어진 행렬이 <math>PDP^{-1}</math>로 표현될 수 있는지, 그리고 그게 가능하다면 <math>P</math>와 <math>D</math>가 무엇인지 알고 싶을 것입니다. 이 챕터의 주요 목표가 바로 그것입니다. == 고유값, 고유벡터 그리고 대각화 == 모티브에서의 관점으로 다음 정의를 얻을 수 있습니다. {{인용문|'''정의 5.1 (대각화 가능 행렬)''' 정사각행렬 <math>A</math>는 가역행렬이 <math>P</math>가 존재하여 <math>P^{-1}AP</math>가 대각행렬이면 대각화 가능하다.}} {{인용문|'''참고''' 동치인 조건은 모티브에서 보았던 꼴인 어떤 대각행렬 <math>D</math>와 가역행렬 <math>P</math>에 대해서 <math>A=PDP^{-1}</math>를 만족하는 것이다. 따라서 행렬이 대각화 가능하다면, 우리는 그 행렬의 거듭제곱도 편하게 계산할 수 있다.}} {{인용문| '''예시''' 항등행렬 <math>I_n</math>는 <math>P=I_n</math>가 존재해 <math>P^{-1}I_nP</math>가 대각행렬(<math>I_n</math>)이므로 대각화가능하다. 또, <math>P=I_n,D=I_n</math>가 존재하여 <math>I_n=PDP^{-1}</math>를 이룬다. }} {{예제| <quiz display=simple> {Select all {{colored em|diagonalizable}} matrix (matrices). |type="[]"} + A zero matrix. || <math>O_n^{-1}O_nO_n=O_n</math> is a diagonal matrix + <math>2I.</math> || <math>I_n^{-1}(2I_n)I_n=2I_n</math> is a diagonal matrix + A diagonal matrix. || let the diagonal matrix be <math>D</math> with size <math>n\times n</math> . <math>I_n^{-1}DI_n=D</math> is a diagonal matrix. </quiz> }} 다음 내용은 어느 부분에서는 대각가능성과 관련된 아주 중요하고 일반적인 개념입니다. {{인용문|'''정의 5.2. (고유벡터와 고유값)''' <math>A</math>를 정사각행렬이라고 하자. <math>0</math>이 아닌 벡터 <math>\mathbf v</math>에 대해서 어떤 스칼라 <math>\lambda</math>가 존재하여 <math>A\mathbf v=\lambda\mathbf v</math>가 성립하면 <math>\mathbf v</math>를 <math>A</math>의 고유벡터라고 한다. 그리고 <math>\lambda</math>는 <math>\mathbf v</math>에 대응하는 <math>A</math>의 고유값이라고 한다.}} {{인용문| '''참고''' * <math>A\mathbf v=\lambda\mathbf v</math>는 벡터 <math>\mathbf v</math>와 행렬 <math>A</math>의 곱이 벡터 <math>\mathbf v</math>와 스칼라의 곱과 같다는 것을 의미한다.(벡터 scaling) * [[:en:wikt:eigen-|eigen-]]라는 접두어는 '특징적인', '소유의', '고유한'을 나타내는 독일어다.}} {{인용문|'''예시 (항등행렬의 고유벡터)''' 각각의 벡터 <math>\mathbf v\in\mathbb R^n</math>에 대해서 <math display=block> I_n\mathbf v=\mathbf v=1\cdot\mathbf v </math> 이므로, 각각의 벡터 <math>\mathbf v\in\mathbb R^n</math>는 <math>I_n</math>의 고유벡터이고, 이것들에 대응되는 고유값은 모두 <math>1</math>이다. }} {{colored exercise| <quiz display=simple> {Select all correct statement(s). |type="[]"} + If <math>\mathbf v</math> is an eigenvector of invertible matrix <math>A</math>, then it is also an eigenvector of <math>A^{-1}</math>. || <math>A\mathbf v=\lambda\mathbf v\implies A^{-1}A\mathbf v=A^{-1}(\lambda \mathbf v)\implies \mathbf v=\lambda(A^{-1}\mathbf v)\implies A^{-1}=\lambda^{-1}\mathbf v</math> || the eigenvalue <math>\lambda\ne 0</math> since, if the eigenvalue is zero, then <math>A\mathbf v=\mathbf 0\implies A=O</math> since <math>\mathbf v\ne 0</math> by definition, which contradicts to the condition that <math>A</math> is invertible + If <math>\lambda</math> is an eigenvalue of <math>A</math>, then <math>\lambda^n</math> is an eigenvalue of <math>A^n</math>. || <math>A^n\mathbf v=A^{n-1}A\mathbf v=A^{n-1}\lambda\mathbf v=\lambda A^{n-1}\mathbf v=\cdots=\lambda^{n-1}A\mathbf v=\lambda^n\mathbf v</math> + Each vector <math>\mathbf v\in\mathbb R^n</math> is an eigenvctor of the zero matrix <math>O_{n\times n}</math>. || for each vector <math>\mathbf v\in\mathbb R^n</math>, <math>O_{n\times n}\mathbf v=\mathbf 0=0\cdot\mathbf v</math> - Zero vector is an eigenvector of each square matrix. || by definition, eigenvector is a nonzero vector + If there exists an eigenvector for a matrix, then there are infinitely many eigenvectors of that matrix. || suppose matrix <math>A</math> has an eigenvector <math>\mathbf v</math> || so <math>A\mathbf v=\lambda\mathbf v</math> for some scalar <math>\lambda</math> || this implies <math>kA\mathbf v=k\lambda\mathbf v\implies A(k\mathbf v)=\lambda(k\mathbf v)</math> for each scalar <math>k</math> || so <math>k\mathbf v</math>, for each scalar <math>k</math>, are also eigenvectors of <math>A</math> </quiz> }} 다음 정리는 대각화 가능 행렬에 고유벡터와 고유값과 관련이 있다. {{인용문|'''정리 (대각화)''' <math>A</math>를 <math>n\times n</math> 행렬이라 하자. 따라서 오로지 <math>A</math>가 has <math>n</math> 개의 선형독립 고유벡터를 가지고 있을 때만 <math>A</math>는 대각화가능하다. <math>\mathbf v_1,\ldots,\mathbf v_n</math>가 고유값 <math>\lambda_1,\ldots,\lambda_n</math>(고유값은 같을 수 있다.)에 대응되는 <math>A</math>의 선형독립 고유벡터이면, 우리는 각각의 열이 <math>\mathbf v_1,\ldots,\mathbf v_n</math>인 가역행렬 <math>P</math>와 대각성분이 <math>\lambda_1,\ldots,\lambda_n</math>인 대각행렬 <math>D</math>를 정의하여 <math display=block> A=PDP^{-1} </math> 얻을 수 있다. {{증명}} 이 증명에서, <math>\begin{pmatrix}\mathbf v_1&\cdots&\mathbf v_n\end{pmatrix}</math>를 각각의 열을 <math>\mathbf v_1,\ldots,\mathbf v_n</math>로 삼는 행렬으로 표기할 것이다. <math display=block> \begin{align} && A&=PDP^{-1}\\ &\Leftrightarrow& AP&=PD\underbrace{PP^{-1}}_I\\ &\Leftrightarrow& A\begin{pmatrix}\mathbf v_1&\cdots&\mathbf v_n\end{pmatrix}&=\begin{pmatrix}\mathbf v_1&\cdots&\mathbf v_n\end{pmatrix} \begin{pmatrix}\lambda_1&0&\cdots&0\\0&\lambda_2&\cdots&0\\\vdots&\vdots&\ddots&\vdots\\0&0&\cdots&\lambda_n\end{pmatrix}\\ &\Leftrightarrow& \begin{pmatrix}A\mathbf v_1&\cdots&A\mathbf v_n\end{pmatrix}&=\begin{pmatrix}\lambda_1\mathbf v_1&\cdots&\lambda_n\mathbf v_n\end{pmatrix}\\ &\Leftrightarrow& A\mathbf v_1&=\lambda_1\mathbf v_1,\ldots,A\mathbf v_n=\lambda_n\mathbf v_n. \end{align} </math> 우리는 <math>\mathbf v_1,\ldots,\mathbf v_n</math>가 고유벡터임을 증명했다. 이제 가역성과 선형독립 간의 관계의 명제에 의해 <math>P</math>가 가역행렬이려면 오로지 이 벡터들이 선형독립이어야 되기 때문에 벡터들이 선형독립임을 증명하는 것이 남아있다. {{증명끝}} }} {{편집중}}
이 문서에서 사용한 틀:
틀:Colored exercise
(
원본 보기
)
틀:예제
(
원본 보기
)
틀:인용문
(
원본 보기
)
틀:편집중
(
원본 보기
)
선형대수학 입문/고유값과 고유벡터
문서로 돌아갑니다.
둘러보기 메뉴
개인 도구
로그인
이름공간
문서
토론
한국어
보기
읽기
원본 보기
역사 보기
더 보기
검색
둘러보기
대문
최근 바뀜
임의의 문서로
미디어위키 도움말
특수 문서 목록
도구
여기를 가리키는 문서
가리키는 글의 최근 바뀜
문서 정보