logo

대각합 📂행렬대수

대각합

정의

$n\times n$ 행렬 $A$가 다음과 같이 주어졌다고 하자.

$$ A= \begin{bmatrix} a_{11} & a_{12} & \cdots & a_{1n} \\ a_{21} & a_{22} & \cdots & a_{2n} \\ \vdots & \vdots & \ddots & \vdots \\ a_{n1} & a_{n2} & \cdots & a_{nn} \end{bmatrix} $$

$A$의 대각 성분들의 합을 $A$의 대각합trace 이라 정의하고 다음과 같이 표기한다.

$$ \text{tr}(A)=\text{Tr}(A)=a_{11}+a_{22}+\cdots + a_{nn}=\sum \limits_{i=1}^{n} a_{ii} $$

설명

다음과 같이 대각합을 함수로 생각할 수도 있다. $M_{n\times n}(\mathbb{R})$을 실수를 성분으로 가지는 $n\times n$행렬들의 집합이라고 하자. 그러면 $\text{Tr}$은 다음과 같이 정의되는 함수이다.

$$ \text{Tr} : M_{n\times n} (\mathbb{R}) \to \mathbb{R},\quad \text{Tr}(A)=\sum \limits_{i=1}^{n} a_{ii} $$

트레이스가 함수라면 미분하지 못할 것도 없다.

성질

$A = [a_{ij}]$, $B=[b_{ij}]$, $C$가 $n \times n$ 행렬, $k$가 상수라고 하자.

(a) 상수배의 트레이스와 트레이스의 상수배가 같다.

$$ \text{Tr}(kA)= k\text{Tr}(A) $$

(b) 합의 트레이스와 트레이스의 합이 같다.

$$ \text{Tr}(A+B)=\text{Tr}(A)+\text{Tr}(B) $$

(a)+(b) 트레이스는 선형이다.

$$ \text{Tr}(kA+B)=k\text{Tr}(A)+\text{Tr}(B) $$

(c) $AB$와 $BA$의 트레이스는 같다.

$$ \text{Tr}(AB) = \text{Tr}(BA) $$

  • (c') 순환 성질Cyclic Property: 위 사실로부터 다음의 식이 성립함을 알 수 있다.

$$ \text{Tr}(ABC) = \text{Tr}(BCA) = \text{Tr}(CAB) $$

  • (c") 사실 임의의 $A \in \mathbb{R}^{n \times m}$와 $B \in \mathbb{R}^{m \times \ell}$, $C \in \mathbb{R}^{\ell \times n}$에 대해서도 성립한다.

(d) $A$와 전치행렬 $A^{\mathsf{T}}$의 트레이스가 같다.

$$ \text{Tr}(A) = \text{Tr}(A^{\mathsf{T}}) $$

(e) 다음이 성립한다.

$$ \sum\limits_{i,k} a_{ik}b_{ki} = \text{Tr}(AB) = \text{Tr}(B^{\mathsf{T}} A^{\mathsf{T}}) $$

$$ \sum\limits_{i,k} a_{ik}b_{ik} = \text{Tr}(AB^{\mathsf{T}}) = \text{Tr}(A^{\mathsf{T}} B) $$

(f) $A^{\mathsf{T}}A$의 고유값 $\lambda_{i}$에 대해서, $\Tr(A^{\mathsf{T}} A) = \sum_{i} \lambda_{i}$

증명

(e)

두 행렬 $A$, $B$의 곱의 $(i,j)$ 성분은 다음과 같이 나타낼 수 있다.

$$ [AB]_{ij} = \sum\limits_{k} a_{ik}b_{kj} $$

트레이스란 $i=j$인 성분을 모두 더한 것이므로,

$$ \Tr(AB) = \sum\limits_{i,k} a_{ik}b_{ki} $$

(f)

$A^{\mathsf{T}}A$는 대칭행렬이므로 직교대각화가 가능하다.

$$ A^{\mathsf{T}}A = Q^{\mathsf{T}} \Lambda Q $$

이때 $Q$는 직교행렬이고, $\Lambda$는 $A^{\mathsf{T}}A$의 고유값을 대각성분으로 갖는 대각행렬이다. 순환성질에 의해 다음이 성립한다.

$$ \begin{align*} \Tr (A^{\mathsf{T}}A) &= \Tr \left( Q^{\mathsf{T}} \Lambda Q \right) \\ &= \Tr \left( \Lambda QQ^{\mathsf{T}} \right) \\ &= \Tr \left( \Lambda \right) \\ &= \sum\limits_{i} \lambda_{i} \end{align*} $$