Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

Is there a way to show that the Taylor series around 0 of $f(x) = \ln(1-x)$ converges towards $f$ on the interval $(-1,1)$, just by considering the remainder from the Taylor polynomial? I'm having a little trouble with this.

The series is

$ T_n(x) = - \sum_{k=1}^n \frac{x^k}{k}$

The convergence on $(-1,0]$ is not a problem, but on $[0,1)$ things start to get a little complicated. Let $0<x<1$. then $|f^{(n+1)}(t)| = \frac{n!}{(1-t)^{n+1}} \leq \frac{n!}{(1-x)^{n+1}}$ for all $t \in [0,x]$. The Lagrange form of the remainder then tells us that

$|R_n(x)| \leq \frac{n!}{(1-x)^{n+1} (n+1)!}x^{n+1} = \frac{1}{n+1} \left( \frac{x}{1-x} \right)^n$

But if $x > 1/2$, this goes to infinity. I have tried to use the integral form of the remainder as well, but with no luck.

Using the machinery of power series, it's easy to prove the convergence. But is there a way using only theory about Taylor polynomials?

share|improve this question
    
Have you tried using the Cauchy form of the remainder? –  David Mitra Feb 18 '13 at 13:13
    
No I haven't, and I'm going try that, thanks! However, it should be possible to solve the problem without using the cauchy remainder, because the problem is from a textbook that doesn't mention cauchy remainders at all. –  Svinepels Feb 18 '13 at 13:16
add comment

1 Answer

up vote 2 down vote accepted

The only proof of this that I know of uses the Cauchy form of the remainder. If $f^{(n+1)}(x)$ exists for all $x$ between $0$ and $h$ and is continuous on this region$^\dagger$, then this form of the remainder can be shown to hold by using the integral form of the remainder: $$R_{n+1}(x)= {1\over n!} \int_0^x f^{(n+1)}(t)(x-t)^n\,dt $$ and applying the Second Mean Value Theorem for Integrals to this expression (with $g(t)=1$).


So, let's use the Cauchy form of the remainder to show that the Taylor (Maclaurin) series of $\ln(1+x)$ converges to $\ln(1+x)$ for $-1<x<0$:

Let $f(x)=\ln(1+x)$. Using the Cauchy form of the remainder, one has $$ \ln(1+x) =x-{x^2\over2}+\cdots+{(-1)^{n-1} x^n\over n} + R_{n+1}(x), $$ where $$ R_{n+1}(x)={f^{(n+1)}(c)\over n!}(x-c)^n x $$ for some $c$ between $0$ and $x$. Note we may write $$ R_{n+1}(x)= {f^{(n+1)}(\theta x)\over n!}(1-\theta)^n x^{n+1} $$ for some $0\le\theta\le1$.

Evaluating the required derivative, we have $$ R_{n+1}(x)=(-1)^n{x^{n+1}(1-\theta)^n\over (1+\theta x)^{n+1}}. $$

Now, for $-1<x<0$, note that $$ 1+\theta x\ge 1+x $$ and $$ 0\le{1-\theta\over 1+\theta x}\le 1; $$ whence $$ | R_{n+1}(x)| =\biggl|{x^{n+1}(1-\theta)^n\over (1+\theta x)^{n+1}} \biggr| =\biggl|{1-\theta\over 1+\theta x} \biggr|^n \cdot{|x|^{n+1}\over |1+\theta x|^n} \le{|x|^{n+1}\over 1+x}\ \buildrel{n\rightarrow\infty}\over\longrightarrow\ 0. $$


$^\dagger$ The continuity of $f^{(n+1)}$ is not required here; but in this case, a different proof is required.

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.