1
$\begingroup$

conjecture:Today I have no intention of thinking about this question. I have only got two solutions so far. I guess there are only two solutions, but I won't prove it.

Let $n$ be postive integers,and such $$n+\tau{(n)}=2\varphi{(n)}$$ where φ is the Euler's totient function and τ is the divisor function i.e. number of divisors of an integer.

it is clear $n=1$,and also I found $n=9$ is another answer, because $\tau{(9)}=3,\varphi(9)=9\left(1-\dfrac{1}{3}\right)=6$,so we have $$9+3=2\cdot 6\Longleftrightarrow 9+\tau{(9)}=2\varphi(9)$$

But How to find other? I tried a lot. I couldn't find any more.

$\endgroup$
  • $\begingroup$ Upto $n=10^8$, the only solutions are $1$ and $9$. $\endgroup$ – Peter Mar 20 '18 at 13:26
  • 2
    $\begingroup$ Crossposted to MO. $\endgroup$ – Dietrich Burde Mar 20 '18 at 20:02
  • 1
    $\begingroup$ I'm voting to close this question as off-topic because it is both unclearly written and was crossposted to MO on the same day it was posted here. $\endgroup$ – Carl Mummert Apr 1 '18 at 15:03
6
$\begingroup$

$n=1,9$ are the only solutions.

For $n$ even, the LHS is strictly larger than the RHS.

If $n$ is odd, then we get that $\tau(n)$ is odd, so $n$ is a perfect square.

Say $n= \prod p_i^{a_i}$, in particular all $a_i\geq2$. We have $\prod p_i^{a_i-1} \mid \tau(n)$ so $$n \geq 3^{\sum (a_i-1)(p_i-1)}$$ (In general, if $\tau(n)$ is a product of primes then $n$ is at least $q$ to the power the sum of those primes minus $1$, where $q$ is the smallest prime dividing $n$.)(*)

Yet $3^{(a-1)(p-1)}\geq(p^2)^{a-1}\geq p^a$ for prime $p\geq 3$ and $a\geq 2$.

Equality can only occur if $3$ is the only prime dividing $n$, with exponent $a$ satifsying $2(a-1)=a$, i.e. $a=2$, $n=3^2$.


Note how in the odd case the approach shows that $ \varphi(n)$ does not divide $n\pm2^k\tau(n)$ except possibly for $n=1,9$.


(*) Let $\tau(n) = \prod (a_i+1)= \prod r_j$ with the $r_i$ prime.

  1. Show by induction that $\left( \prod b_i \right)-1 \geq \sum( b_i-1)$ for $b_i \geq 1$. (Or geometrically: the LHS counts lattice points in a hyperrectangle and excludes the origin; the RHS counts only the points on the coordinate axes, minus the origin.)
  2. Use $$n = \prod p_i^{a_i} \geq q^{\sum ((a_i+1) - 1)}$$ and factor each $a_i+1$ into prime factors $r_j$, using the above inequality as a lower bound, to get: $$\geq q^{\sum (r_j-1)}$$
$\endgroup$
  • $\begingroup$ So, there are no further solutions to the given equation ? $\endgroup$ – Peter Mar 20 '18 at 14:55
  • 1
    $\begingroup$ indeed.${}{}{}{}$ $\endgroup$ – punctured dusk Mar 20 '18 at 14:56

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.