Calculus 2Power Series

Taylor's Inequality

Bound the error of a Taylor polynomial.

Not started Jump to practice

A Taylor polynomial is an approximation. How good is it? Taylor's inequality bounds the error using the next derivative, much as the Mean Value Theorem bounds how far a function can move.

Why it works

Write $f(x) = T_n(x) + R_n(x)$. Taylor's theorem says the remainder has the form $$R_n(x) = \frac{f^{(n+1)}(c)}{(n + 1)!}(x - a)^{n+1}$$ for some $c$ between $a$ and $x$: it looks just like the next term of the series, with the derivative taken at an unknown point. (For $n = 0$ this is the Mean Value Theorem.) You do not know $c$, but if you know a bound $M$ for $|f^{(n+1)}|$ on the whole interval, you get a guaranteed error bound.

Key idea

Taylor's inequality: if $|f^{(n+1)}(t)| \le M$ for all $t$ between $a$ and $x$, then $$|R_n(x)| \le \frac{M}{(n + 1)!}|x - a|^{n+1}.$$ If $R_n(x) \to 0$ as $n \to \infty$, the Taylor series converges to $f(x)$.

Worked example: A sine estimate

Bound the error when $T_3(x) = x - \frac{x^{3}}{6}$ approximates $\sin(0.5)$.
  1. Every derivative of $\sin$ is $\pm\sin$ or $\pm\cos$, so $M = 1$ works.
  2. $|R_3(0.5)| \le \frac{1}{4!}(0.5)^{4} = \frac{1}{384} \approx 0.0026$.
  3. (Actually $T_3(0.5) = 0.479167$ and $\sin 0.5 = 0.479426$: the true error is $0.00026$, well inside the bound.)

Answer: At most $\frac{1}{384}$

Worked example: How many terms for e?

How many terms of $e = \sum\frac{1}{n!}$ guarantee an error less than $0.001$?
  1. For $f(x) = e^{x}$ on $[0, 1]$, every derivative is at most $e < 3$, so $M = 3$.
  2. $|R_n(1)| \le \frac{3}{(n + 1)!}$. Need $(n + 1)! > 3000$: $6! = 720$ is too small, $7! = 5040$ works.
  3. So $n = 6$: $$\begin{aligned} T_6(1) &= 1 + 1 + \frac{1}{2} + \cdots + \frac{1}{720} \\ &\approx 2.71806, \end{aligned}$$ within $0.001$ of $e = 2.71828$.

Answer: $n = 6$ (seven terms)

Taylor polynomials of $e^{x}$ with a probe at $x = 1$. The readout shows the actual error; compare it with the bound $\frac{3}{(n + 1)!}$ as you raise $n$.

Common mistake

Choosing $M$ from the value at the center only. For $e^{x}$ on $[0, 2]$, $f^{(n+1)}(0) = 1$, but the bound must hold for every $t$ in the interval, and $e^{t}$ reaches $e^{2} \approx 7.39$. Using $M = 1$ would understate the error. Find the largest value of $|f^{(n+1)}|$ on the whole interval between $a$ and $x$ (or a convenient number above it).

Worked example: Why the series for e^x works

Show that $\sum\frac{x^{n}}{n!}$ equals $e^{x}$ for every $x$.
  1. On the interval between 0 and $x$, $|f^{(n+1)}| \le e^{|x|}$, so $|R_n(x)| \le e^{|x|}\frac{|x|^{n+1}}{(n + 1)!}$.
  2. For fixed $x$, $\frac{|x|^{n+1}}{(n + 1)!} \to 0$ (factorials beat powers).
  3. So $R_n(x) \to 0$, and the series converges to $e^{x}$.

Answer: $e^{x} = \sum\frac{x^{n}}{n!}$ for all $x$

Practice

Pass with 3 right answers, the last 2 in a row. Each right answer earns 3 XP.

Loading practice…

Builds on

Leads to

Sources and further reading

  • OpenStax Calculus Volume 2, Section 6.3: Taylor and Maclaurin Series (Taylor's theorem with remainder)
  • Paul’s Online Math Notes (Lamar University), Calculus II: Taylor Series (remainder)
@teddyrookbook GitHub