1 minute read

Linear Transformation

\[\begin{aligned} AX\ =\ b \end{aligned}\]

위의 식을 하나의 System equation이라고 생각하면 X는 연립방정식의 Solution이라고 생각할 수 있다. 즉, X가 System equation의 input이 되며 linear combination of column vector이고, b가 output이 된다는 뜻이다.

image-20220327174441038

즉, n차원의 벡터를 m차원의 공간으로 변형하는 것 같은 관계를 생각할 수 있다. 그때 A는 Linear transformation이다.

Differential of polynomial

다음과 같은 다항식을 벡터로 표시해보자. \(\begin{aligned} X {(t)}\ &=\ a_0\ +\ a_1 t +\ a_2 t^2 +\ \cdots\ +\ a_n t_n\\ y {(t)}\ &=\ \frac {d} {dt} X(t)\ =\ b_o\ +\ b_1 t^\ +\ b_2 t^2\ +\ \cdots\ +\ b_{n-1} t^{n-1} \end{aligned}\) image-20220327180607542

$\frac {d} {dt}$를 하면 차원이 하나 줄어든다고 생각할 수도 있다.

Integration

\[\begin{aligned} \int_{- \infty}^{t} X(t)\ dt \end{aligned}\]

image-20220327180823886

Differential과 반대로 차원이 늘어난다고 생각할 수 있다.

Multiplication by a polynomial

\[\begin{aligned} x(t)\ \cdot\ p(t)\ =\ y(t)\ ,\ x(t)\ =\ a_1x_1(t)\ +\ a_2x_2(t) \end{aligned}\]

라는 다항식의 곱으로 나타낸다면, \(\begin{aligned} (a_1 x_1(t)\ +\ a_2 x_2(t)) p(t)\ =\ a_1 x_1(t)p(t)\ +\ a_2 x_2(t)p(t) \end{aligned}\) 만약 p(t)가 이차다항식이면, \(\begin{aligned} p(t)\ &=\ p_0\ +\ p_1t\ +\ p_2t^2 \\ \\ x(t)\ &=\ a_0\ +\ a_1t\ +\ \cdots\ +\ a_nt^n \\ x(t)p(t)\ &=\ b_0\ +\ b_1t\ +\ \cdots\ +\ b_{n+2}t^{n+2}\\ \\ \mathbb{R}^{n+1}\ &\rightarrow\ \R^{n+3} \end{aligned}\)

다항함수를 미분해서 transformation을 할 때, 행렬식이 어떻게 나오는지 예를 들어 보자.

Ax=b라는 행렬이 주어져 있고 3 차원 다항식을 예를 들면, \(\begin{aligned} x(t)\ =\ a_0\ +\ a_1t\ +\ a_2t^2\ +\ a_3t^3 \end{aligned}\) X를 x(t)의 계수로 이루어진 column vector로 표현하고 각각 basis vector로 나타내보자. \(\begin{aligned} X\ =\ \begin{pmatrix} a_0 \\ a_1 \\ a_2 \\ a_3 \end{pmatrix} \end{aligned}\) image-20220327185243634

이를 통해 행렬 A를 transformation한 $A_diff$로 나타낼 수 있다. $$ \begin{aligned} A_{diff}\ &=\ \begin{pmatrix} 0\ 1\ 0\ 0
0\ 0\ 2\ 0
0\ 0\ 0\ 3
0\ 0\ 0\ 0
\end{pmatrix} \begin{pmatrix} a_0\ a_1\ a_2\ a_3 \end{pmatrix} \Rightarrow \mathbb{y}\ =\ b_0\ +\ b_1t\ +\ b_2t^2\

&= \begin{pmatrix} a_1\ 2a_2\ 3a_3\ 0\end{pmatrix} \rightarrow \mathbb{y}(t)\ =\ a_1\ +\ 2a_2t\ +\ 3a_3t^2 \end{aligned} $$