μ‹€λ³€μˆ˜ ν•¨μˆ˜ $f(x)$λ₯Ό λ‹€ν•­ ν•¨μˆ˜μ˜ λ©±κΈ‰μˆ˜λ‘œ ν‘œν˜„ν•˜κΈ°. $n$μ°¨ 근사λ₯Ό λ¬΄ν•œλ²ˆ μˆ˜ν–‰ν•œ 것과 κ°™λ‹€.

6 minute read

β€œλ―Έμ λΆ„ν•™β€ μˆ˜μ—…μ—μ„œ 배운 것과 κ³΅λΆ€ν•œ 것을 μ •λ¦¬ν•œ ν¬μŠ€νŠΈμž…λ‹ˆλ‹€. 전체 ν¬μŠ€νŠΈλŠ” Calculus νŽ˜μ΄μ§€μ—μ„œ ν™•μΈν•˜μ‹€ 수 μžˆμŠ΅λ‹ˆλ‹€ πŸ“

ν•¨μˆ˜μ˜ λ©±κΈ‰μˆ˜ ν‘œν˜„

일반적으둜 μš°λ¦¬κ°€ λ§Œλ‚˜λŠ” ν•¨μˆ˜λ“€μ€ μ•„λž˜μ™€ 같은 ν˜•νƒœλ‹€.

\[f(x) = 1 + 3x + 4x^5\]

λ‹€ν•­μ‹μ˜ ν•©μœΌλ‘œ κ΅¬μ„±λ˜κ±°λ‚˜

\[f(x) = \frac{1}{1 + x}\]

λΆ„λͺ¨μ˜ ν˜•νƒœμ΄κ±°λ‚˜

\[f(x) = \sqrt{x + 5}\]

무리수 ν•¨μˆ˜μΌ 것이닀. 이 ν•¨μˆ˜λ“€μ˜ 곡톡점은 λͺ¨λ‘ μ‹€λ³€μˆ˜ ν•¨μˆ˜λΌλŠ” 것이닀.

λ§Œμ•½ μ‹€λ³€μˆ˜ ν•¨μˆ˜ $f(x)$κ°€ $x = x_0$μ—μ„œ λ¬΄ν•œλ²ˆ λ―ΈλΆ„ κ°€λŠ₯ν•œ ν•¨μˆ˜λΌλ©΄, 이 ν•¨μˆ˜λ₯Ό λ‹€ν•­μ‹μ˜ λ¬΄ν•œν•©μœΌλ‘œ ν‘œν˜„ν•  수 μžˆλ‹€. 예λ₯Ό λ“€μ–΄, ν•¨μˆ˜ $f(x) = e^x$λŠ” μ•„λž˜μ™€ 같이 ν‘œν˜„ν•  μˆ˜λ„ μžˆλ‹€!

\[f(x) = e^x = \sum_{k=0}^{\infty} \frac{x^k}{k!}\]

μ΄λ ‡κ²Œ ν•¨μˆ˜ $f(x)$λ₯Ό $x^k$의 λ©±κΈ‰μˆ˜ ν˜•νƒœλ‘œ ν‘œν˜„ν•˜λŠ” 것을 <ν…ŒμΌλŸ¬ κΈ‰μˆ˜; Taylor Series>라고 ν•œλ‹€.


Maclaurin Series

<ν…ŒμΌλŸ¬ κΈ‰μˆ˜; Taylor Series>λ₯Ό μ‚΄νŽ΄λ³΄κΈ° 전에, μ‘°κΈˆλ” μ‰¬μš΄ κ°œλ…μΈ <맀크둜린 κΈ‰μˆ˜; Maclaurin Series>λ₯Ό λ¨Όμ € μ‚΄νŽ΄λ³΄μž.

기본적인 μ•„μ΄λ””μ–΄λŠ” β€˜μ ‘μ„ β€™μ„ 톡해 ν•¨μˆ˜λ₯Ό κ·Όμ‚¬ν•˜λŠ” μ„ ν˜• 근사(Linear Approximation)이닀.

ν•¨μˆ˜ $f(x)$λ₯Ό μ„ ν˜• κ·Όμ‚¬ν•œ ν•¨μˆ˜ $L(x)$λŠ” μ•„λž˜μ™€ κ°™λ‹€. ν‘œκΈ°μ˜ 편의λ₯Ό κ·€ν•΄ $x = 0$ μ§€μ μ—μ„œ μ„ ν˜• 근사λ₯Ό μˆ˜ν–‰ν•˜κ² λ‹€.

\[L(x) = f(0) + f'(0) x\]

2μ°¨μ‹μœΌλ‘œ κ·Όμ‚¬ν•˜λŠ” 2μ°¨ 근사(Quadratic Approximation)도 ν•΄λ³΄μž.

\[Q(x) = f(0) + f'(0) x + \frac{f''(0)}{2!} x^2\]

$x^2$의 κ³„μˆ˜μ—μ„œ λΆ„λͺ¨μ— $2!$κ°€ λΆ™λŠ” μ΄μœ λŠ” 근사식 $Q(x)$λ₯Ό 2μ°¨ λ―ΈλΆ„ ν–ˆμ„ λ•Œ, $Q^{\prime\prime}(0) = f^{\prime\prime}(0)$κ°€ λ˜μ–΄μ•Ό ν•˜κΈ° λ•Œλ¬Έμ΄λ‹€.

1μ°¨, 2μ°¨ κ·Όμ‚¬μ—μ„œ λ³΄μ΄λŠ” νŒ¨ν„΄μ„ λ°”νƒ•μœΌλ‘œ $k$μ°¨ 근사식을 μœ λ„ν•˜λ©΄ μ•„λž˜μ™€ κ°™λ‹€.

\[f(0) + f'(0) x + \frac{f''(0)}{2!} x^2 + \frac{f'''(0)}{3!} x^3 + \cdots \frac{f^{k}(0)}{k!} x^k\]

$k$μ°¨ κ·Όμ‚¬μ—μ„œ 더 λ‚˜μ•„κ°€ λ¬΄ν•œλ²ˆ 근사λ₯Ό ν•˜λ©΄ <Maclaurin Series>κ°€ λœλ‹€!

Definition. Maclaurin Series

For real-valued function $f(x)$, if it can be infinitely differential on $x = 0$, then we can represent $f(x)$ as a power series near the $x = 0$.

\[\begin{aligned} f(x) &= f(0) + f'(0) x + \frac{f''(0)}{2!} x^2 + \cdots \frac{f^{k}(0)}{k!} x^k \cdots \\ &= \sum^\infty_{n=0} \frac{f^{(n)}(x_0)}{n!} (x - x_0)^n \end{aligned}\]

Taylor Series

<ν…ŒμΌλŸ¬ κΈ‰μˆ˜: Taylor Series>λŠ” 이런 λ©±κΈ‰μˆ˜ 근사λ₯Ό $x = 0$κ°€ μ•„λ‹Œ $x = x_0$μ—μ„œ μˆ˜ν–‰ν•œ 것이닀. <Maclaurin Series>의 μΌλ°˜ν™” 버전이닀. <ν…ŒμΌλŸ¬ μ „κ°œ; Taylor Expansion>라고도 ν•œλ‹€.

Definition. Taylor Series

For real-valued function $f(x)$, if it can be infinitely differential on $x = x_0$, then we can represent $f(x)$ as a power series near the $x = x_0$.

\[\begin{aligned} T(x) &= f(x_0) + f'(x_0) (x-x_0) + \frac{f''(x_0)}{2!} (x-x_0)^2 + \cdots \frac{f^{k}(x_0)}{k!} (x-x_0)^k \cdots \\ &= \sum^\infty_{n=0} \frac{f^{(n)}(x_0)}{n!} (x - x_0)^n \end{aligned}\]

Examples

ν…ŒμΌλŸ¬ ν•¨μˆ˜λ‘œ ν‘œν˜„λ˜λŠ” λŒ€ν‘œμ μΈ ν•¨μˆ˜λ“€μ„ μ‚΄νŽ΄λ³΄μž.

  • $1 / (1-x)$
  • $e^x$
  • $\sin x$

Fractional Function

Fractional function $1 / (1-x)$에 λŒ€ν•œ 맀크둜린 κΈ‰μˆ˜λŠ” μ•„λž˜μ™€ κ°™λ‹€.

\[\left(\frac{1}{1-x}\right)^{(n)} = n! \cdot \frac{1}{(1-x)^n}\]

κ°€ 됨을 κΈ°μ–΅ν•˜μž.

\[\begin{aligned} \frac{1}{1-x} &= 1 + x + x^2 + x^3 + \cdots \\ &= \sum^{\infty}_{n=0} x^k \end{aligned}\]

단, 이 κ·Όμ‚¬λŠ” $-1 < x < 1$ λ²”μœ„μ—μ„œλ§Œ μœ νš¨ν•˜λ‹€.

Exponential Function

Exponential function $e^x$에 λŒ€ν•œ ν…ŒμΌλŸ¬ κΈ‰μˆ˜λŠ” μ•„λž˜μ™€ κ°™λ‹€. μ΄λ•Œ, $(e^x)’ = e^x$κ°€ 됨을 κΈ°μ–΅ν•˜μž.

\[\begin{aligned} e^x &= 1 + 1 x + \frac{1}{2!}x^2 + \cdots \frac{1}{k!} x^k + \cdots \\ &= \sum^{\infty}_{n=0} \frac{x^n}{n!} \end{aligned}\]

Sine Function

Sine function $\sin x$에 λŒ€ν•œ ν…ŒμΌλŸ¬ κΈ‰μˆ˜λŠ” μ•„λž˜μ™€ κ°™λ‹€.

\[\begin{aligned} \sin x &= 0 + 1 x + 0 - \frac{1}{3!}x^3 + \cdots \\ &= \sum^{\infty}_{n=0} (-1)^{n} \frac{x^{(2n + 1)}}{(2n + 1)!} \end{aligned}\]

Why Taylor Series?

μ‹€λ³€μˆ˜ ν•¨μˆ˜ $f(x)$λ₯Ό ν…ŒμΌλŸ¬ μ „κ°œν•˜λ©΄ κ·Έ ν•¨μˆ˜μ˜ 닀항식 ν‘œν˜„μ„ μ•Œ 수 μžˆμŠ΅λ‹ˆλ‹€. μš°λ¦¬κ°€ λ‹€ν•­ ν•¨μˆ˜μ— λŒ€ν•΄μ„  μ‰½κ²Œ λ‹€λ£° 수 있기 λ•Œλ¬Έμ— λ³΅μž‘ν•˜κ³  μ–΄λ €μš΄ ν•¨μˆ˜λ₯Ό μ‰¬μš΄ λ²„μ „μœΌλ‘œ λ°”κΏ€ 수 μžˆμŠ΅λ‹ˆλ‹€!

μ˜ˆμ‹œλŠ” β€˜λ‹€ν¬ ν”„λ‘œκ·Έλž˜λ¨Έβ€™λ‹˜μ˜ 포슀트λ₯Ό μ°Έκ³ ν–ˆμŒμ„ 미리 λ°νž™λ‹ˆλ‹€ πŸ™

적뢄 계산

적뢄 계산이 μ–΄λ €μš΄ ν•¨μˆ˜λ₯Ό ν…ŒμΌλŸ¬ μ „κ°œν•  수 μžˆλ‹€λ©΄, 적뢄이 훨씬 μ‰¬μ›Œμ§‘λ‹ˆλ‹€.

\[\int \sin (x^2) \; dx = \int \left( x^2 - \frac{x^6}{3!} + \frac{x^{10}}{5!} - \cdots \right) \; dx\]

ν•¨μˆ˜μ˜ 점근 νŠΉμ„± νŒŒμ•…

λ³΅μž‘ν•œ ν•¨μˆ˜λ“€μ˜ 점근적(asymptotic) νŠΉμ„±μ„ μ‰½κ²Œ νŒŒμ•…ν•  수 μžˆμŠ΅λ‹ˆλ‹€.

\[\lim_{x \rightarrow 0} \frac{\sin x}{x} = \lim_{x \rightarrow 0} \frac{x - x^3 / 3! + x^5 / 5! - \cdots }{x} = \lim_{x \rightarrow 0} \frac{x}{x} = 1\]

μ»΄ν“¨ν„°μ˜ μ΄ˆμ›”ν•¨μˆ˜ 계산

μ΄ˆμ›”ν•¨μˆ˜ $\sin x$, $e^x$ 등을 Bit 계산을 ν•˜λŠ” 컴퓨터가 κ³„μ‚°ν•˜λŠ” 것은 μ–΄λ ΅μŠ΅λ‹ˆλ‹€. κ·ΈλŸ¬λ‚˜ μ΄ˆμ›”ν•¨μˆ˜λ₯Ό ν…ŒμΌλŸ¬ μ „κ°œν•΄ λ‹€ν•­ ν•¨μˆ˜λ‘œ λ§Œλ“€λ©΄, 컴퓨터가 κ³„μ‚°ν•˜κΈ° μ‰¬μš΄ ν˜•νƒœκ°€ λ©λ‹ˆλ‹€. μ»΄ν“¨ν„°λ‘œ μ΄ˆμ›”ν•¨μˆ˜λ₯Ό $n$μ°¨κΉŒμ§€ κ·Όμ‚¬ν•œ 수 κ³„μ‚°ν•˜λ©΄ 맀우 μ •λ°€ν•œ κ²°κ³Όλ₯Ό 얻을 수 μžˆμŠ΅λ‹ˆλ‹€.


맺음말

이번 ν¬μŠ€νŠΈμ—μ„  <ν…ŒμΌλŸ¬ κΈ‰μˆ˜>와 <맀크둜린 κΈ‰μˆ˜>λ₯Ό μ†Œκ°œν•˜λŠ” μ„ μ—μ„œ λ§ˆλ¬΄λ¦¬ν•˜κ³ μž ν•œλ‹€. 이런 <ν…ŒμΌλŸ¬ κΈ‰μˆ˜>κ°€ μ‘΄μž¬ν•˜λŠ”μ§€μ— λŒ€ν•΄ λ…Όμ˜ν•˜λŠ” <ν…ŒμΌλŸ¬ 정리; Taylor Theorem>λŠ” λ³„λ„μ˜ ν¬μŠ€νŠΈμ—μ„œ 닀루겠닀.

μ£ΌκΈ° ν•¨μˆ˜λ₯Ό $\sin$, $\cos$의 λ¬΄ν•œν•©μœΌλ‘œ κ·Όμ‚¬ν•˜λŠ” <푸리에 κΈ‰μˆ˜; Fourier Series>도 μžˆλ‹€. <ν…ŒμΌλŸ¬ κΈ‰μˆ˜>와 λ§ˆμ°¬κ°€μ§€λ‘œ λ¬΄ν•œν•©μ˜ κ³„μˆ˜ $a_n$ 값을 μ°ΎλŠ” λ°©μ‹μœΌλ‘œ 근사λ₯Ό μˆ˜ν–‰ν•œλ‹€.

πŸ‘‰ Fourier Series


References

Categories:

Updated: