1
$\begingroup$

I'm asked to estimate the integral $\int_0^{0.2} \frac {1}{1+x^4}$ to six decimal places.

I know that this can be represented by the power series $\sum_0^\infty (-1)^n x^{4n}$. I also know that the alternating series estimation theorem states that the |error| for a particular estimation will always be less than the term we did not include, which is $b_{n+1}$. In this case, $b_{n+1}$ is $(-1)^{n+1} x^{4n+4}$.

However, I'm having difficulty actually estimating this integral. I was told by my professor that if you want to estimate something to $n$ decimal places, such as 2, your $b_{n+1}$ should have $n$ zeros. For example, $b_{n+1} = 0.001$ if you wish to estimate to 2 decimal places.

This means that in our case, $b_{n+1} = 0.0000001$, with 6 zeros after the decimal point.

If this were not a power series, I would estimate this by simply trying arbitrary $n$ values until I achieved the desired $b_{n+1}$, but there is an $x$ involved because it is a power series. Thus, if I try n = 5, for instance, I will have $b_{n+1} = x^{24}$, which has a variable and is not 0.000something.


How do I find the value of n for which this estimation is accurate within 6 decimal places so that I can use the n terms to then calculate the definite integral?

$\endgroup$
  • $\begingroup$ Or do I have to calculate the definite integral for each "n" I select and see if that is less than 0.0000001? $\endgroup$ – AleksandrH Oct 29 '16 at 22:32
2
$\begingroup$

If $z\in(0,1)$, $$ \sum_{n\geq 0}(-1)^n \frac{z^{4n}}{4n+1} $$ is a converging series with alternating signs. In particular, $$ \left|\sum_{n\geq N}(-1)^n \frac{z^{4n}}{4n+1}\right|<\frac{z^{4N}}{4N+1} $$ and if $z=\frac{1}{5}$ and $N=2$, the RHS is less than $8\cdot 10^{-8}$, so the LHS is less than $8\cdot 10^{-8}$ as well.

$\endgroup$
  • $\begingroup$ A bit confusing, but I eventually figured it out. I'll give you a +1 and accept anyway, but I personally don't like all the "z" variables and whatnot :P $\endgroup$ – AleksandrH Oct 30 '16 at 17:36
  • $\begingroup$ Actually, no, I got n = 1, and the correct answer is 0.199936 $\endgroup$ – AleksandrH Oct 30 '16 at 18:18

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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