5
$\begingroup$

Can all positive integers $k$, be written as a difference of two perfect powers $k=a^m-b^n$, with $m,n>1$ and $a,b$ positive integers?

A number is imperfect if it can not, which numbers are imperfect?

What is the asymptotics of the number of imperfect numbers less then $x$, as $x\rightarrow\infty$?

I have proved that all odd numbers are the difference of two squares. How to solve the other cases?

$\endgroup$
  • 7
    $\begingroup$ A number is the difference of two squares if and only if it is odd or divisible by 4. $\endgroup$ – N. S. Jun 3 '12 at 19:45
  • 1
    $\begingroup$ So we are left with k=2*odd, and m or n above 2 $\endgroup$ – Sigma4 Jun 3 '12 at 19:55
  • 6
    $\begingroup$ The terminology "perfect number" already has a standard meaning in number theory - you should probably choose a different word. $\endgroup$ – Greg Martin Jun 3 '12 at 20:22
  • 1
    $\begingroup$ One exception: $4$ is not the difference of two positive squares. It is, however, $2^3-2^2$. To supplement NS's comment: $$\begin{array}{c} (k+1)^2-(k-1)^2=4k; \\ (k+1)^2-k^2=2k+1. \end{array}$$ This covers integers congruent to $0$ or $\pm1$ modulo $4$. $\endgroup$ – anon Jun 3 '12 at 21:10
5
$\begingroup$

See http://oeis.org/A074981 and references there.

$\endgroup$

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.