A number $N \in \mathbb{N}$ is said to be perfect if $\sigma(N)=2N$, where $\sigma=\sigma_{1}$ is the classical sum-of-divisors function. For example, $\sigma(6)=1+2+3+6=2\cdot{6}$, so that $6$ is perfect. (Note that $6$ is even.)

It is currently unknown whether there are any odd perfect numbers. Euler proved that an odd perfect number, if any exists, must take the form $N = q^k n^2$ where $q$ is prime satisfying $q \equiv k \equiv 1 \pmod 4$ and $\gcd(q,n)=1$. We call $q$ the Euler prime of the odd perfect number $N$.

In what follows, we denote the abundancy index of $x \in \mathbb{N}$ as $I(x)=\sigma(x)/x$.


Since the non-Euler part $n^2$ is deficient but not almost perfect, we know (from this preprint) that $$I(n^2) < \frac{2n^2 + D(n^2)}{n^2 + D(n^2)},$$ where $D(n^2)={2n^2}-\sigma(n^2)$ is the deficiency of $n^2$.

When the Descartes-Frenicle-Sorli conjecture that $k=1$ is true, we also know that we have the lower bound $$\frac{5}{3} \leq I(n^2),$$ whereupon we obtain, from the last two inequalities, the following lower bound $$\frac{n^2}{D(n^2)} > 2.$$ (When $k=1$, ${n^2}/D(n^2)=(q+1)/2$, so that ideally we should have ${n^2}/D(n^2) \geq 3$ since $q \geq 5$.)

From the following Answer 1 and Answer 2 to a related MSE question, we obtain improved upper bounds for $I(n^2)$ by successively computing the mediant of $$I(n^2) = \frac{2n^2 - D(n^2)}{n^2}=I_0$$ and the upper bounds $u_i$ (for $I(n^2)$) given by $$u_0 = \frac{2n^2 + D(n^2)}{n^2 + D(n^2)}$$ $$u_1 = \frac{\bigg(2n^2 - D(n^2)\bigg)+\bigg(2n^2 + D(n^2)\bigg)}{n^2 + \bigg(n^2 + D(n^2)\bigg)} = \frac{4n^2}{2n^2 + D(n^2)}$$ $$u_2 = \frac{\bigg(2n^2 - D(n^2)\bigg)+4n^2}{n^2 + \bigg(2n^2 + D(n^2)\bigg)}=\frac{6n^2 - D(n^2)}{3n^2 + D(n^2)}$$ $$u_3 = \frac{\bigg(2n^2 - D(n^2)\bigg)+\bigg(6n^2 - D(n^2)\bigg)}{n^2 + \bigg(3n^2 + D(n^2)\bigg)} = \frac{8n^2 - 2D(n^2)}{4n^2 + D(n^2)}$$ $$u_4 = \frac{\bigg(2n^2 - D(n^2)\bigg)+\bigg(8n^2 - 2D(n^2)\bigg)}{n^2 + \bigg(4n^2 + D(n^2)\bigg)} = \frac{10n^2 - 3D(n^2)}{5n^2 + D(n^2)}$$ $$\ldots \ldots \ldots$$ $$\ldots \ldots \ldots$$ $$\ldots \ldots \ldots$$ (That is, $u_i = \text{mediant}(u_{i-1},I_0)$.)

Note that it is easy to prove that $u_i$ (for $i=0, 1, 2, \ldots$) is a strictly decreasing sequence.

The resulting lower bounds for ${n^2}/D(n^2)$ (on assuming $k=1$) are:

$$u_0 : \frac{n^2}{D(n^2)} > 2$$ $$u_1 : \frac{n^2}{D(n^2)} > \frac{5}{2}$$ $$u_2 : \frac{n^2}{D(n^2)} > \frac{8}{3}$$ $$u_3 : \frac{n^2}{D(n^2)} > \frac{11}{4}$$ $$u_4 : \frac{n^2}{D(n^2)} > \frac{14}{5}$$ $$\ldots : \ldots$$


(1) Do the resulting lower bounds for ${n^2}/D(n^2)$ (using this procedure) have a closed form?

(2) Does the lower bound ever reach $3$, or is it possible to prove that it does not?

  • 2
    $\begingroup$ (+1) for your effort about this unsolved problem. Recently I've seen a new paper published by an author in arXiv on a conjecture due to you. Isn't required a response, good week. $\endgroup$ – user243301 Jun 11 '17 at 10:02

It appears that the numerators of the lower bounds form OEIS sequence A016789. Hence, the lower bounds have the closed form $$\frac{3i + 2}{i + 1}.$$

Consequently, $$\frac{3i + 2}{i + 1}=3-\frac{1}{i + 1},$$ so that the lower bounds never reach $3$.


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.