सभी नोट्स

Taylor series and approximation

गेट DA150 मिनट में पढ़ें

मुफ़्त प्रीव्यू: नोट का 29%

The Taylor polynomial

Idea. Near a point aa, a smooth function is close to the polynomial that matches its value and first nn derivatives at aa:

Tn(x)=f(a)+f′(a)(x−a)+f′′(a)2!(x−a)2+⋯+f(n)(a)n!(x−a)n.T_n(x) = f(a) + f'(a)(x - a) + \frac{f''(a)}{2!}(x - a)^2 + \dots + \frac{f^{(n)}(a)}{n!}(x - a)^n .

With a=0a = 0 this is the Maclaurin polynomial. If the remainder tends to 0 as n→∞n \to \infty, the infinite sum is the Taylor series, and it equals ff.

Why the factorials. Differentiating (x−a)k(x - a)^k exactly kk times gives k!k!. Dividing by k!k! makes the kk-th derivative of TnT_n at aa equal to f(k)(a)f^{(k)}(a).

Remainder (Lagrange form).

f(x)−Tn(x)=f(n+1)(c)(n+1)!(x−a)n+1for some c between a and x.f(x) - T_n(x) = \frac{f^{(n+1)}(c)}{(n+1)!}(x - a)^{n+1} \quad \text{for some } c \text{ between } a \text{ and } x.

So if ∣f(n+1)∣≤M\lvert f^{(n+1)}\rvert \le M on the interval, the error is at most M∣x−a∣n+1(n+1)!\dfrac{M\lvert x - a\rvert^{n+1}}{(n+1)!}.

Series to know by heart

FunctionMaclaurin seriesValid for
exe^x1+x+x22!+x33!+…1 + x + \frac{x^2}{2!} + \frac{x^3}{3!} + \dotsall xx
sin⁡x\sin xx−x33!+x55!−…x - \frac{x^3}{3!} + \frac{x^5}{5!} - \dotsall xx
cos⁡x\cos x1−x22!+x44!−…1 - \frac{x^2}{2!} + \frac{x^4}{4!} - \dotsall xx
11−x\frac{1}{1 - x}1+x+x2+x3+…1 + x + x^2 + x^3 + \dots∣x∣<1\lvert x\rvert < 1
ln⁡(1+x)\ln(1 + x)x−x22+x33−…x - \frac{x^2}{2} + \frac{x^3}{3} - \dots−1<x≤1-1 < x \le 1
(1+x)α(1 + x)^\alpha1+αx+α(α−1)2!x2+…1 + \alpha x + \frac{\alpha(\alpha - 1)}{2!}x^2 + \dots∣x∣<1\lvert x\rvert < 1
arctan⁡x\arctan xx−x33+x55−…x - \frac{x^3}{3} + \frac{x^5}{5} - \dots∣x∣≤1\lvert x\rvert \le 1
tan⁡x\tan xx+x33+2x515+…x + \frac{x^3}{3} + \frac{2x^5}{15} + \dots∣x∣<π2\lvert x\rvert < \frac{\pi}{2}
sinh⁡x\sinh x, cosh⁡x\cosh xlike sin⁡\sin and cos⁡\cos, but with all signs ++all xx

Patterns.

  • sin⁡\sin is odd and contains only odd powers. cos⁡\cos is even and contains only even powers.
  • The derivative of the exe^x series is itself. The derivative of the sin⁡\sin series is the cos⁡\cos series.
  • Substituting x→−xx \to -x in 11−x\frac{1}{1-x} gives 11+x=1−x+x2−…\frac{1}{1 + x} = 1 - x + x^2 - \dots. Integrating that gives ln⁡(1+x)\ln(1 + x).
  • Substituting x→x2x \to x^2 in 11+x\frac{1}{1+x} and integrating gives arctan⁡x\arctan x.

Computing Taylor expansions

Example 1: exe^x about a=1a = 1. Every derivative equals ee at 1, so ex=e[1+(x−1)+(x−1)22!+… ].e^x = e\left[1 + (x - 1) + \frac{(x - 1)^2}{2!} + \dots\right].

Example 2: ln⁡x\ln x about a=1a = 1. f(1)=0f(1) = 0, f′(1)=1f'(1) = 1, f′′(1)=−1f''(1) = -1 and f′′′(1)=2f'''(1) = 2. So ln⁡x=(x−1)−(x−1)22+(x−1)33−…\ln x = (x - 1) - \frac{(x - 1)^2}{2} + \frac{(x - 1)^3}{3} - \dots This is the ln⁡(1+t)\ln(1 + t) series with t=x−1t = x - 1.

Example 3: approximate 4.1\sqrt{4.1}. Take f(x)=xf(x) = \sqrt x and a=4a = 4.

  • f(4)=2f(4) = 2.
  • f′(x)=12xf'(x) = \frac{1}{2\sqrt x}, so f′(4)=14f'(4) = \frac14.
  • f′′(x)=−14x3/2f''(x) = -\frac{1}{4x^{3/2}}, so f′′(4)=−132f''(4) = -\frac{1}{32}.

4.1≈2+0.14−132⋅0.012=2+0.025−0.00015625=2.02484375.\sqrt{4.1} \approx 2 + \frac{0.1}{4} - \frac{1}{32}\cdot\frac{0.01}{2} = 2 + 0.025 - 0.00015625 = 2.02484375 .

The true value is 2.0248457…2.0248457\ldots. The first-order (linear) approximation 2.025 is already within 1.6×10−41.6 \times 10^{-4}.

Example 4: an error bound. Approximate e0.5e^{0.5} by T3=1+0.5+0.125+0.02083‾=1.64583‾T_3 = 1 + 0.5 + 0.125 + 0.0208\overline{3} = 1.6458\overline{3}.

  • Every derivative of exe^x on [0,0.5][0, 0.5] is at most e0.5<1.65e^{0.5} < 1.65.
  • So the error is at most 1.65×0.544!=1.65×0.062524≈0.0043\frac{1.65 \times 0.5^4}{4!} = \frac{1.65 \times 0.0625}{24} \approx 0.0043.

The true error is 1.64872−1.64583≈0.00291.64872 - 1.64583 \approx 0.0029, within the bound.

आपने इस नोट का मुफ़्त 29% हिस्सा पढ़ लिया

आगे पढ़ने के लिए Google से साइन इन करें। यह मुफ़्त है, एक टैप में होता है, और साथ में मिलता है:

  • इस नोट से 7 अभ्यास सवाल
  • हर नोट पूरा, साथ में PDF लाइब्रेरी
  • आधिकारिक पिछले पेपर समय-सीमा वाले टेस्ट की तरह, व्याख्या के साथ
  • डाउट के लिए AI ट्यूटर, हिंदी या अंग्रेज़ी में
  • आपकी प्रगति, गलतियाँ और स्टडी प्लान हर डिवाइस पर सेव
आगे पढ़ें, मुफ़्त