I'm working through Spivak's Calculus (Fourth Edition).

In chapter 2, problem 6:

The formula for $1^2 + ... + n^2$ may be derived as follows. We begin with the formula $(k+1)^3 - k^3 = 3k^2 +3k +1$
Writing this formula for $k = 1, ... , n$ and adding, we obtain
$(n+1)^3 - 1= 3 (1^2 + ... + n^2 ) +3(1+...+n) +n$
Thus we can find $(1^2 + ... + n^2 )$ if we already know $(1+...+n)$ (which could have been found in a similar way).

This is clear for me. However, I'm stuck in this problem: Use this method to find $$\frac{3}{1^2\times2^2} + \frac{5}{2^2\times3^2} +\cdots+ \frac{(2n+1)}{n^2(n+1)^2}$$

I would like to know how could the one think in this and not just how to solve it. I want to know how to deal with "deriving summation formulas" in general.

  • $\begingroup$ i don't think that this method will work here $\endgroup$ – Dr. Sonnhard Graubner Jan 29 '17 at 13:11
  • $\begingroup$ @Dr.SonnhardGraubner It worked for $1/n(n+1)$ In a way that may be complex for me. So I think it's possible to do it the same way for this expression. $\endgroup$ – Ahmed99 Jan 29 '17 at 13:13
  • $\begingroup$ Do you mean $3/(1^2\cdot 2^2)$ and so on? $\endgroup$ – Wolfram Jan 29 '17 at 13:13
  • $\begingroup$ @Wolfram Yeah, could you edit the question to make it looks clearer. $\endgroup$ – Ahmed99 Jan 29 '17 at 13:14

I think what was meant here is that $\frac{2k+1}{k^2(k+1)^2}=\frac1{k^2}-\frac1{(k+1)^2}$. You can use this formula to replace each member of the sum by the difference and then cancel out similar terms.

  • 1
    $\begingroup$ Yes. I got the formula for the summation: $1-1/(n+1)^2$. $\endgroup$ – Ahmed99 Jan 29 '17 at 13:28
  • $\begingroup$ @Ahmed99 Then you should be good? $\endgroup$ – Simply Beautiful Art Jan 29 '17 at 13:35
  • $\begingroup$ @SimplyBeautifulArt Yeah. Asking about how could the one deal with such problems in general, may not be appropriate for the site. But I asked this because I spent almost an hour staring in such a simple question. $\endgroup$ – Ahmed99 Jan 29 '17 at 13:45
  • $\begingroup$ @Ahmed99 I'm afraid, there is no general method to search closed forms of different sums. There are only a bunch of tricks, and the one shown here is trying to write the term of the sum $f(n)$ as the difference $g(n+1)-g(n)$ for some function $g$. Then the sum $f(1)+\cdots+f(n)$ is obviously $g(n+1)-g(1)$. But it is neither easy to find $g$ in the general case, nor it always works. $\endgroup$ – Wolfram Jan 29 '17 at 13:51
  • $\begingroup$ @Wolfram I see. Thanks for your clarification. $\endgroup$ – Ahmed99 Jan 29 '17 at 15:57

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.