You are on page 1of 9

Notes on Number Theory and Discrete Mathematics

ISSN 13105132
Vol. XX, XXXX, No. X, XXXX

On a Conjecture of Dris Regarding Odd Perfect


Numbers
Keneth Adrian P. Dagal1 and Jose Arnaldo B. Dris2

arXiv:1312.6001v5 [math.NT] 16 Jun 2015

Department of Mathematics, Far Eastern University


Manila, Philippines
e-mail: kdagal@feu.edu.ph

Graduate of M. Sc. Mathematics, De La Salle University


Manila, Philippines
e-mail: josearnaldobdris@gmail.com

Abstract: Dris conjectured in [3] that the inequality q k < n always holds, if N = q k n2 is an odd
perfect number given in Eulerian form. In this note, we show that either of the two conditions
n < q k or (q)/n < (n)/q holds. This is achieved by first proving that (q)/n 6= (n)/q k ,
where (x) is the sum of the divisors of x. Using this analysis, we show that the condition
q < n < q k holds in four out of a total of six cases. By utilizing a separate analysis, we show

that the condition n < q < n 3 holds in four out of a total of five cases. We conclude with some
open problems related to Sorlis conjecture that k = 1.
Keywords: Odd perfect number, abundancy index, Sorlis conjecture.
AMS Classification: 11A25.

1 Introduction
Let N = q k n2 be an odd perfect number given in Eulerian form (i.e., q is prime with q k 1
(mod 4) and gcd(q, n) = 1).
Therefore, q 6= n. It follows that either q < n or n < q.
Dris [2] proved that n < q implies Sorlis conjecture that k = 1 [6]. By the contrapositive,
k > 1 implies that q < n.
Acquaah and Konyagin [1] showed that all the prime factors r of N satisfy r < (3N)1/3 . In
particular, if k = 1, then

q < (3N)1/3 = q 3 < 3N = 3qn2 = q < n 3.


1

Therefore, regardless of the status of Sorlis conjecture, we know that

q<n 3
must be true.
Notice that, if Acquaah and Konyagins estimate for the Euler prime q (under the assumption

k = 1) could be improved to q < (2N)1/3 , then it would follow that q < n 2, and this would
hold unconditionally.
Let (x) be the sum of the divisors of the positive integer x. Let
I(x) = (x)/x
be the abundancy index of x.

2 Main Results
Here, we examine this problem.
Determine the correct ordering for the following quantities:
(q) (n) (q k ) (n)
, k ,
,
n
q
n
q
Recall the following results:
1 < I(qn) =

(q) (n)
(q k ) (n)

I(q k n) =
k < 2,
n
q
n
q

and
n < q k = {k = 1 n < q}.
In general, since 1 < q k n is deficient (being a proper factor of the perfect number N = q k n2 ),
then we have
(n)
(q k )
6= k .
n
q
In a similar vein,
(q)
(n)
6=
.
n
q
Note that the following implications are true:

(n)
(q)
<
= q < n 2,
n
q
and

(q)
(n)
<
= n < q.
q
n
We want to show that

(n)
(q)
6= k .
n
q
2

Suppose to the contrary that


(n)
(q)
= k .
n
q
Since gcd(q, n) = 1 and q is prime, we have:
(n)
(q)
= k N.
n
q
This means that
1

(n)
(q)
= k .
n
q

But since (q) = q + 1 is even while n is odd, we then have:


2

(q)
(n)
= k .
n
q

From the inequality


2

(q)
n

we get
2

(q) (n)
(n)

= I(qn) < 2
q
n
q

from which we obtain

(n)
< 1.
q

But then we finally have


(n)
(q)
(n)
(n)
<1<2
= k
,
q
n
q
q
which is a contradiction.
Consequently, we obtain:
(q)
(n)
6= k .
n
q
We now consider two separate cases:
Case 1:

(n)
(q)
< k
n
q

Since k 1, this implies that

(q)
(n)
(n)
< k
.
n
q
q
Consequently, under Case 1, we have the condition:
(n)
(q)
<
.
n
q

From a previous remark, we know that this implies q < n 2.


Case 2:
(q)
(n)
<
qk
n
3

Again, since k 1, this implies that


(q)
(q k )
(n)
<

.
qk
n
n
This implies that, under Case 2, we have the condition:
(n)
(q k )
<
.
qk
n
But recall that we have the following inequality [2]:

(n)
(q k )
3
k
=
I(q
)
<
.
2 < I(n) =
k
q
n
Together, the last two inequalities imply that:
n < qk .
This implies that the biconditional k = 1 n < q is true.
Now, suppose that
(q k )
(n)
=
.
q
n
Again, since gcd(q, n) = 1 and q is prime, we have
(q k )
(n)
=
N.
q
n
It follows that

(n)
(q k )
1
=
.
q
n

But
n 6= (q k ) k + 1 2
since k 1 (mod 4), while n is odd.
Therefore,
2
We claim that

Assume that

(mod 4),

(q k )
(n)
=
.
q
n

(n)
(q k )
=
= k > 1.
q
n
(q k )
(n)
=
.
q
n

Since gcd(q, n) = 1 and q is prime, we have


(n)
(q k )
=
N.
q
n

Consequently, we obtain (as before)


(n)
(q k )
=
q
n

2
(since (q k ) 6= n).
From the inequality

(n)
q

we get
2

(q)
(q) (n)

= I(qn) < 2
n
n
q

from which we obtain

(q)
< 1.
n

This last inequality implies that q < n.

Now, from the inequality I(q k ) < 3 2 < I(n) (see [2]), we get:
(q k )
qk
< .
(n)
n
Note that from the following inequality:
2

(q k )
n

we get
(n) < 2n (q k )
from which we obtain
1<

qk
(q k )
< .
(n)
n

Thereupon, we get n < q k , which as weve noted before, implies that the biconditional
k = 1 n < q
is true.
But we have already obtained q < n. Therefore, we know that k > 1.
A shorter way to prove the implication
(q k )
(n)
=
= k > 1
q
n
would be to note that, in general, the condition
(n)
(q)
6=
q
n
is true. In particular,
(q k )
(n)
(q)
=
6=
n
q
n
5

from which it follows that

(q)
(q k )
6=
.
n
n
It follows from the last inequation that k 6= 1. Since we know that k 1, it follows that k > 1.
We now summarize the results we have obtained so far:
(q)
(n)
6= k .
n
q
The following inequations are trivial:
(n)
(q k )
6= k
n
q
(q)
(n)
6=
.
n
q
Also, note that

and

(q k )
(q)

n
n
(n)
(n)

.
k
q
q

3 Synopsis
We now list all the possible orderings for:


(q) (n) (q k ) (n)
, k ,
,
n
q
n
q
A:

B:

C:

D:

E:

F:

(q)
(q k )
(n)
(n)

< k
n
n
q
q
(q)
(n)
(q k )
(n)
< k <

n
q
n
q
(n)
(n)
(q k )
(q)
< k <
<
n
q
q
n
(n)
(q)
(n)
(q k )
<
<

qk
n
q
n
(q)
(q k )
(n)
(n)
<
<
<
k
q
n
n
q
(n)
(q)
(q k )
(n)
=
<
=
qk
q
n
n
6

Note that, under cases B, C, D, E and F, we have the inequality


(q k )
(n)
<
qk
n
which implies that n < q k .
Furthermore, note that, under cases A, B, C, D and E, we have the condition
(q)
(n)
<
.
n
q
Lastly, notice that, under cases B, C, D and E, we actually have the inequalities
q < n < qk
since k > 1 in each of these cases.
For another case-to-case analysis, we first prove the following claim:

Claim: Either (q)/n < 2 or (n)/q < 2.


Since (q)/n 6= (n)/q and I(qn) = ((q)/n) ((n)/q) < 2, by symmetry, it suffices to
consider the case:
(q)/n < (n)/q.
Suppose to the contrary that

2 < (q)/n < (n)/q.

This leads to
2 < ((q)/n) ((n)/q) = I(qn) < 2,
which is a contradiction. (The case (n)/q < (q)/n is treated similarly.) This establishes the
claim.
Now, there are four cases to consider:
Case I: (q)/n <

2 < (n)/q

There are two further subcases here:


Sub-Case I-1: (q)/n < 1 <

2 < (n)/q

2 < (n)/q < 2

Under this subcase, q < n.


Sub-Case I-2: 1 < (q)/n <

Under this subcase, n < q < n 2.


Case II: (n)/q <

2 < (q)/n <

3 + 10375


Under this case, n < q < n 3.
Case III: 1 < (q)/n < (n)/q <

Under this case, n < q < n 2.


Case IV: 1 < (n)/q < (q)/n <

Under this case, n < q < n 2.


It therefore remains to improve the upper bound for q under Case II [5].

4 Open Problems
In the paper [4], a heuristic motivating the pursuit of a proof for the conjectures of Sorli (k = 1)
and Dris (q k < n) is presented. In particular, it seems fruitful to try to establish the following
predictions:
Conjecture 1:
q < n = k = 1
The truth of this conjecture will imply Sorlis conjecture.
Conjecture 2:
q<n
The truth of this conjecture, together with Sorlis conjecture, will imply Driss conjecture.
Note at this point that Conjectures 1 and 2 are consistent with those of Sorli and Dris.
Lastly, it might be feasible to try to come up with a proof for n < q (thereby proving k = 1
as a consequence), then coming up with a separate proof for k > 1 (from which it follows that
q < n), in order to finally show that, indeed, there are no odd perfect numbers.

Acknowledgements
The authors would like to thank the anonymous referees for their valuable feedback and suggestions which helped in improving the presentation and style of this manuscript.

References
[1] Acquaah, P., S. Konyagin, On prime factors of odd perfect numbers, Int. J. Number Theory,
08 (2012), 1537, doi:http://dx.doi.org/10.1142/S1793042112500935.
[2] Dris,
fect

J. A.
numbers,

B.,
J.

The abundancy index of


Integer
Seq.,
15
(Sep.
8

divisors
2012),

of odd perArticle
12.4.4,

https://cs.uwaterloo.ca/journals/JIS/VOL15/Dris/dris8.html,
ISSN 1530-7638.
[3] Dris, J. A. B., Solving the odd perfect number problem: some old and new
approaches, M. Sc. thesis, De La Salle University, Manila, Philippines, 2008,
http://arxiv.org/abs/1204.1450.
[4] Dris, J. A. B., Euclid-Euler heuristics for (odd) perfect numbers, preprint,
http://arxiv.org/abs/1310.5616.

[5] Dris, J. A. B., Improving the bound q < n 3 for an odd perfect number N = q k n2 given in
Eulerian form, http://mathoverflow.net/questions/188831, last viewed on
06/16/2015.
[6] Sorli, R. M., Algorithms in the study of multiperfect and odd perfect numbers, Ph. D. Thesis, University of Technology, Sydney, 2003,
http://epress.lib.uts.edu.au/research/handle/10453/20034.