1
$\begingroup$

I have a question for my class that asks to find the product of two infinite series, namely $a_n = b_n = \sum\limits_{k=0}^\infty r^n$ for $r\in(0,1)$. That is find $c_n = \sum\limits_{k=0}^n a_k b_{n-k}$. Then it asks to show without appealing to the "product rule" that the product is absolutely convergent. Lastly, find its sum.

My thought process is the following. Let, $a_n = b_n = \sum\limits_{k=0}^\infty r^n$. Then $c_n = \sum\limits_{k=0}^nr^kr^{n-k} = \sum\limits_{k=0}^nr^n.$

Now, if that is the cauchy product, then I figure that if we look at the partial sums of $c_n$, then as $n$ approaches $+\infty$ then $c_n$ approaches zero. Since we're dealing with strictly positive terms then it's also absolutely convergent.

Last thing is to find a sum of the series, so if $c_n = \sum\limits_{k=0}^nr^n$, then I believe that the sum would be like any other geometric sum, and since $a=1$, then $c_n$ converges to $\frac{1}{1-r}$.

$\endgroup$
  • 1
    $\begingroup$ What do you mean by $a_n=b_n=\sum r^n$? maybe $a_n=r^n$ and you are looking for $\sum a_n\sum_b_n$ ? $\endgroup$ – Dimitris Apr 24 '13 at 14:37
  • $\begingroup$ Both $a_n$ and $b_n$ are the same geometric series $\sum\limits_{n=0}^\infty r^n$. So I'm looking for the Cauchy Product, which I think I've interpreted correctly, but I could definitely be wrong. $\endgroup$ – Neurax Apr 24 '13 at 14:40
  • $\begingroup$ you can't have $n$ in $a_n$ out of the sum and inside the sum in r^n... $\endgroup$ – Dimitris Apr 24 '13 at 14:41
2
$\begingroup$

You have, correctly, that $$c_n=\sum_{k=0}^n r^n.$$ That's like $\sum_{k=0}^n b$, the sum of $n+1$ copies of $b$, since $r^n$ does not depend on $k$. That means that $$c_n=(n+1)r^n.$$

Now that you have the correct $c_n$, perhaps you can complete things.

Edit: So we want $\sum_0^\infty (n+1)r^n$. Let's change letters to $\sum_0^\infty (n+1)x^n$. By the Ratio Test, or the Root Test, you can prove that we have convergence if $|x|\lt 1$.

There are many ways to show that the sum is $\frac{1}{(1-r)^2}$. It depends on how much marchinery you want to use. For minimal machinery (but not much fun) one can find an explicit formula for $\sum_0^n (k+1)r^k$. For freewheeling use of machinery, note that $\frac{1}{1-x}=1+x+x^2+x^3+\cdots$, and differentiate term by term.

$\endgroup$

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.