Primitive weird numbers having more than three

distinct prime factors

Gianluca Amato, Maximilian F. Hasler, Giuseppe Melfi, Maurizio Parton

To cite this version:

Gianluca Amato, Maximilian F. Hasler, Giuseppe Melfi, Maurizio Parton. Primitive weird numbers
having more than three distinct prime factors . Rivista di Matematica della Universita’ degli studi di
Parma, 2016, 7 (1), pp.153 - 163. �hal-01684543�

HAL Id: hal-01684543

https://hal.univ-antilles.fr/hal-01684543
Submitted on 15 Jan 2018

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci- destinée au dépôt et à la diffusion de documents
entific research documents, whether they are pub- scientifiques de niveau recherche, publiés ou non,
lished or not. The documents may come from émanant des établissements d’enseignement et de
teaching and research institutions in France or recherche français ou étrangers, des laboratoires
abroad, or from public or private research centers. publics ou privés.
 Riv. Mat. Univ. Parma, Vol. x (2014), 000-000

Gianluca Amato, Maximilian Hasler, Giuseppe Melfi,

and Maurizio Parton

Primitive weird numbers having more than three distinct prime factors

Abstract. In this paper we study some structure properties of primi-

tive weird numbers in terms of their factorization. We give sufficient
conditions to ensure that a positive integer is weird. Two algorithms for
generating weird numbers having a given number of distinct prime fac-
tors are presented. These algorithms yield primitive weird numbers of
the form mp1 . . . pk for a suitable deficient positive integer m and primes
p1 , . . . , pk and generalize a recent technique developed for generating
primitive weird numbers of the form 2n p1 p2 . The same techniques can
be used to search for odd weird numbers, whose existence is still an
open question.
Keywords. Abundant numbers, semiperfect numbers, almost perfect
numbers, sum-of-divisor function, Erdős problems.
Mathematics Subject Classification (2010): 11A25, 11B83.

1 - Introduction
P
Let n ∈ N be a natural number, and let σ(n) = d|n d be the sum of
its divisors. If σ(n) > 2n, then n is called abundant, whereas if σ(n) < 2n,
then n is called deficient. Perfect numbers are those for which σ(n) = 2n.
According to [5], we will refer to ∆(n) = σ(n) − 2n as the abundance of n, and
to d(n) = 2n − σ(n) = −∆(n) as the deficience of n. If n can be expressed
as a sum of distinct proper divisors, then n is called semiperfect, or sometimes
also pseudoperfect. Slightly abundant numbers with ∆(n) = 1 are called quasi-
perfect, and slightly deficient numbers with d(n) = 1 are called almost perfect.
A weird number is a number which is abundant but not semiperfect. In
other words, n ∈ N is weird if it is abundant and it cannot be written as the
sum of some of its proper divisors.
Weird numbers have been defined in 1972 by Benkoski [1], and appear to be
rare: for instance, up to 104 we have only 7 of them [10]. Despite this apparent
rarity, which is the reason of the name, weird numbers are easily proven to be
infinite: if n is weird and p is a prime larger than σ(n), then np is weird (see
for example [4, page 332]). But a much stronger property is true: Benkoski and
Erdős, in their joint 1974 paper [2], proved that the set of weird numbers has
positive asymptotic density.
Several questions on weird numbers have not been settled yet. For instance,
if we look for primitive weird numbers, that is, that are not multiple of other
weird numbers, we don’t know whether they are infinite or not:
C o n j e c t u r e 1.1. [2, end of page 621] There exist infinitely many primitive
weird numbers.
In this respect, the third author recently proved in [8] that the infiniteness
of primitive weird numbers follows by assuming the classic Cramér conjecture
on gaps between consecutive primes [3].
Another open question is the existence of odd weird numbers. Erdős offered
$ 10 for an example of odd weird number, and $ 25 for a proof that none can
exist [1]. Recently Wenjie Fang [10, Sequence A006037] claimed that there are
no odd weird numbers up to 1021 , and no odd weird numbers up to 1028 with
abundance not exceeding 1014 .
Moreover, very little is known about a pattern in the prime factorization of
primitive weird numbers. As of now, most of the known primitive weird numbers
are of the form 2n pq with p and q primes, and all the papers on primitive weird
numbers deal with numbers of this form [6, 7, 8, 9]. Relatively few examples of
primitive weird numbers with more than three distinct prime factors are known
up to now: for instance among the 657 primitive weird numbers not exceeding
1.8 · 1011 there are 531 primitive weird numbers having three distinct prime
factors; 69 having four distinct prime factors; 54 having five distinct prime
factors and only 3 having six distinct prime factors.
This paper considers primitive weird numbers that have several distinct
prime factors. In particular, we give sufficient conditions in order to ensure
that a positive integer of the form mp1 . . . pk is weird, where m is a deficient
number and p1 , . . . , pk are primes (see Theorem 3.1).
We then apply Theorem 3.1 to search for new primitive weird numbers, look-
ing in particular at four or more prime factors. We find hundreds of primitive
weird numbers with four distinct prime factors of the form 2m p1 p2 p3 , 75 primi-
tive weird numbers with five distinct prime factors of the form 2m p1 p2 p3 p4 , and
9 primitive weird numbers with six distinct prime factors (see Section 4).
This paper generalizes to several factors a technique developed in [8]. This
approach, as far as we know, is the first that can be used to generate primitive

2
weird numbers with several distinct prime factors. Moreover, since there are
many odd deficient numbers, Theorem 3.1 can be used to hunt for the first
example of an odd weird number (see Section 5).

2 - Basic ideas

We recall a fundamental lemma that will be extensively used, and that cor-
responds to an equivalent definition of weird number.
L e m m a 2.1. An abundant number w is weird if and only if ∆(w) cannot
be expressed as a sum of distinct divisors of w.
P r o o f. For a proof one can see [8, Lemma 2]. 

We will need another technical lemma, which will be used in the proof of
the main theorems.
L e m m a 2.2. If w = mq is an abundant number, m is deficient, q is prime
and q ≥ σ(pα ) − 1 for each pα ||m, then w is primitive abundant.
P r o o f. Since a multiple of an abundant number is abundant, in order to
prove that m is a primitive abundant number, (i.e., an abundant number whose
proper divisors are all deficient), it suffices to prove that w/p is deficient for each
p|w. If p = q, then w/q = m, and we assumed that m is deficient. Otherwise,
if pα ||m, then

σ(w)(pα − 1)p
 
σ(w/p) σ(w) p−1
= = · 1 − α+1
w/p w(pα+1 − 1) w p −1
 
σ(w) 1 σ(w) q σ(w/q)
≤ · 1− = · = < 2.
w q+1 w q+1 w/q


3 - Main result

In this section we provide two ways for generating primitive weird numbers.
T h e o r e m 3.1. Let m > 1 be a deficient number and k > 1. Let p1 , . . . , pk
be primes with σ(m) + 1 < p1 < · · · < pk . Let
 
∗ p1 − σ(m)
h = .
pk − p1
Let w̃ = mp1 . . . pk , and
∗
h
[
(1) Um,p1 ,pk := {n ∈ N | jpk + σ(m) < n < (j + 1)p1 }.
j=0

3
3.1.1 If w̃ is abundant and ∆(w̃) ∈ Um,p1 ,pk , then w = w̃ is a primitive weird
number.
3.1.2 If w̃ is deficient, let
2w̃
p< −1
d(w̃)
a prime with p > pk . Then w = w̃p is abundant. Furthermore, if ∆(w) ∈
Um,p1 ,pk , then
2w̃ (1 + h∗ )p1
p> −1−
d(w̃) d(w̃)
and moreover, if p > ∆(w) then w is a primitive weird number.

P r o o f o f 3.1.1 Since σ(m) + 2 ≤ p1 , the set union in the right side of (1)
is not empty. The sets of consecutive integers involved in the union in the right
side of (1) are pairwise disjoint. If

p1 − σ(m)
h< ,
pk − p1
then
σ(m) + hpk < (h + 1)p1 .
Now, let j ≤ h∗ and let n be an integer with jpk + σ(m) < n < (j + 1)p1 .
We will prove that n cannot be expressible as a sum of distinct divisors of w.
Note that n < p21 . This is because n < (j + 1)p1 ≤ (h∗ + 1)p1 < (p1 /2 + 1)p1
and p1 /2+1 < p1 . This means that if n is expressible as a sum of distinct divisors
of w, these divisors must be of the form dp with d | m and p ∈ {p1 , . . . , pk },
or simply of the form d, with d | m. Let’s say n = d1 p1 + · · · + dN pN + d01 +
d02 + · · · + d0M , where d01 , . . . , d0M are distinct divisors of m. Then, necessarily
d1 + d2 + · · · + dN ≤ j, since n < (j + 1)p1 . As a consequence, we have:

d01 + · · · + d0M = n − (d1 p1 + . . . dN pN ) ≥ n − jpk > σ(m)

and this is in contradiction with the assumption on d01 , . . . , d0M . So n cannot be

expressible as a sum of distinct divisors of w.
No elements in Um,p1 ,pk can be expressed as a sum of distinct divisors of w.
Since ∆(w) ∈ Um,p1 ,pk , by Lemma 2.1 this implies that w is weird.
In order to prove that w is a primitive weird number, by Lemma 2.2 it
suffices to prove that w/pk is deficient. If k = 2, then ∆(w/pk ) = ∆(mp1 ) =
∆(m)p1 + σ(m). Since ∆(m) ≤ −1 and p1 > σ(m), then ∆(w/p2 ) < 0. We may
assume that k ≥ 3. Then we have:
   
w w w σ(w) 2w
∆ = σ −2 = −
pk pk pk pk + 1 pk
2w
∆(w) −
pk (σ(w) − 2w) − 2w pk
= =
pk (pk + 1) pk + 1

4
Now, ∆(w) ∈ Um,p1 ,pk . Since k ≥ 3, between p1 and pk there is at least an odd
integer that is not prime, and therefore h∗ < p1 /2k. This means that
 p1 
∆(w) < (h∗ + 1)p1 < 1 + p1 .
2k
On the other hand 2w/pk ≥ 2mp21 , and since 2mp1 > p1 + p1 > 1 + p1 /2k,
this means that ∆(w) < 2w/pk and therefore ∆(w/pk ) < 0. In particular, since
w/pk is deficient, by Lemma 2.2, w is a primitive weird number. 

P r o o f o f 3.1.2 Note that p is the largest prime that divides w, and that
w/p = w̃ is deficient. So by Lemma 2.2, in order to prove that w is a primitive
weird number, it suffices to prove that w is indeed abundant and weird.
Since (w̃, p) = 1 and 2w̃ − d(w̃)p − d(w̃) > 0 by hypothesis, we have

∆(w) = σ(w) − 2w
= σ(w̃)(p + 1) − 2pw̃
= (σ(w̃) − 2w̃)p + σ(w̃)
= −d(w̃)p + 2w̃ − d(w̃) > 0

This proves that w is abundant.

Now assume that ∆(w) ∈ Um,p1 ,pk . Since max Um,p1 ,pk < (1 + h∗ )p1 , then
∆(w) = −d(w̃)p + 2w̃ − d(w̃) < (1 + h∗ )p1 and therefore

2w̃ (1 + h∗ )p1
p> −1−
d(w̃) d(w̃)

Let p > ∆(w). In order to prove that w is weird, by Lemma 2.1, we have to
prove that ∆(w) is not a sum of proper divisors of w.
Since ∆(w) < p, if ∆(w) is a sum of proper divisors of w, then all divisors
involved must be divisors of w̃. On the other hand ∆(w) ∈ Um,p1 ,pk and as seen
above, no element in Um,p1 ,pk can be expressed as a sum of distinct divisors of
w̃. This completes the proof. 

R e m a r k 3.1. Very often, when conditions of Theorem 3.1.2 hold, it is

(1 + h∗ )p1 < d(w̃). This means that in these cases, p = [2w̃/d(w̃) − 1].
R e m a r k 3.2. If m, p1 , . . . , pk is a sequence such that w = mp1 . . . pk is
primitive weird according to Theorem 3.1.1, k > 2 and ∆(w) < pk , then w̃ =
p1 . . . pk−1 and p = pk verify the conditions of Theorem 3.1.2
Indeed, if w satisfies the condition of Theorem 3.1.1, then w is abundant and
w̃ = w/pk is deficient. This implies hthat pk < i 2w̃/d(w̃) − 1. Moreover, since
p1 −σ(m)
∆(w) ∈ Um,p1 ,pk , there is a j ≤ h∗ = pk −p1 such that jpk +σ(m) < ∆(w) <
h i
−σ(m)
(j + 1)p1 . Since pk > pk−1 , then jpk−1 + σ(m) < ∆(w), with j ≤ pp1k−1 −p1 .
This means ∆(w) ∈ Um,p1 ,pk−1 as for Theorem 3.1.2. Finally, if ∆(w) < pk all
requirements of Theorem 3.1.2 are satisfied.

5
 Despite of the above remark, the conditions of Theorem 3.1.1. and 3.1.2 are
not equivalent. For example, m = 211 , p1 = 11321, p2 = 12583 and p3 = 13093
verify conditions 3.1.1 (so w = mp1 p2 p3 is a primitive weird number) but not
3.1.2, because p3 < ∆(w) = 43936. In the other sense, for m = 212 , p1 = 23143,
p2 = 24043, p3 = 27061 and p = 3077507, conditions 3.1.2 are satisfied, but
the weird number w̃ = mp1 p2 p3 p4 does not verify the conditions 3.1.1, because
∆(w̃) = 39680 6∈ Um,p1 ,p4 .

4 - The application of Theorem 3.1

Theorem 3.1 may be used to develop an algorithm which searches for primi-
tive weird numbers with many different prime factors. The sufficient conditions
are computationally much easier to check than the standard definition of weird
number. The question, however, is in which range the prime numbers p1 , . . . , pk
should be chosen. The following theorem gives a partial answer.
T h e o r e m 4.1. Let m be a deficient number and let d be its deficience. Let
p1 , . . . , pk be primes, with m < p1 < p2 < · · · < pk . Let w̃ = mp1 . . . pk .
(i) If pk < 2km/d − (k + 2)/2 then w̃ is abundant.
(ii) If p1 > 2km/d − k/2 then w̃ is deficient.
P r o o f. (i). We first prove that if p1 , p2 , . . . , pk are distinct primes with
m < p1 < · · · < pk < 2km/d − (k + 2)/2 then w = mp1 . . . pk is abundant.
Since (m, pi ) = 1, one has σ(w) = σ(m)σ(p1 ) . . . σ(pk ). This implies that:
 k 
Y     k
σ(w) d 1 d 1
= 2− 1+ > 2− · 1+ > 2.
w m ph m pk
h=1

The above inequalities hold because for positive x, the function x → 1 + 1/x
is decreasing, and because the equation
   k
d 1
2− · 1+ =2
m q

holds for  
1 2k k+1 d
q= s = m− +O > pk .
2 d 2 m
k
d
−1
2− m
Therefore w is abundant.
(ii). The proof is analogous. 

If one wants to generate weird numbers by means of Theorem 3.1 (3.1.1

or 3.1.2), i.e., abundant numbers w̃ with ∆(w̃) ∈ Um,p1 ,pk on a hand Um,p1 ,pk
have to be as large as possible. Good choices are with p1 − σ(m) as large as

6
possible. On the other hand, in order to get higher values of h∗ , pk − p1 have to
be as small as possible. This leads to consider k-tuples of primes p1 , . . . pk in an
interval that, by Theorem 4.1, includes 2km/d − (k + 1)/2. However, if k ≤ d
then 2km/d might be smaller than σ(m), and if p1 < σ(m) then h∗ < 0 and
Um,p1 ,pk is empty. Therefore, k > d is generally preferable, and all new weird
numbers we have found enjoy this property.
The primitive weird numbers generated with Theorem 1 in [8] are particular
cases of Theorem 3.1.1, with m = 2h , and k = 2. When p1 and p2 are chosen
according to that theorem, then conditions of Theorem 3.1.1 are fulfilled and
w = 2h p1 p2 is a primitive weird number.
However, Theorems 3.1.1 and 3.1.2 become more interesting when applied
to search weird numbers with several prime factors.
Theorem 3.1.1 yields primitive weird numbers of the form mp1 . . . pk where
m is a deficient number, k is an integer larger than the deficience of m and
the pi ’s are suitably chosen. It is relatively easy to generate primitive weird
numbers up to four distinct prime factors. The table below shows some of the
primitive weird numbers having at least five distinct prime factors we were able
to generate with Theorem 3.1.1

w prime factorization m ∆(w)

9210347984 24 · 83 · 89 · 149 · 523 24 32
9772585048 23 · 17 · 317 · 419 · 541 23 · 17 304
23941578736 24 · 73 · 103 · 127 · 1567 24 32
109170719992 23 · 19 · 79 · 2731 · 3329 23 16
359214428128 25 · 127 · 211 · 509 · 823 25 64
446615164768 25 · 163 · 167 · 331 · 1549 25 64
83701780710848 26 · 181 · 563 · 3407 · 3767 26 128
823548808494656 26 · 139 · 3631 · 4441 · 5741 26 128
31871420410521385088 27 · 257 · 90803 · 98221 · 108631 27 · 257 78464
32852586770937891968 27 · 257 · 87443 · 90803 · 125777 27 · 257 85184
32892333375893455232 27 · 257 · 79841 · 109943 · 113909 27 · 257 76736
33622208489084493184 27 · 257 · 76757 · 107873 · 123439 27 · 257 72832

Table 1: The above primitive weird numbers have five distinct prime factors
and have been found by means of Theorem 3.1.1 with k = 4 for m = 2h and
with k = 3 for m = 136 = 23 · 17 or m = 32896 = 27 · 257.

An implementation of Theorem 3.1.2 yields in minutes hundreds of primitive

weird numbers of the form 2h p1 p2 p. By going deeper in the implementation of
Theorem 3.1.2, we have been able to find 65 primitive weird numbers w having
five distinct prime factors of the form w = 2h p1 p2 p3 p; and nine primitive weird
numbers having six distinct prime factors, that are shown in the table below.
As a comparison, before our computations only three primitive weird numbers
having six distinct prime factors were known at the OEIS database [10].

7
 w prime factorization ∆(w)
44257207676 22 · 11 · 37 · 59 · 523 · 881 8
125258675788784 24 · 47 · 149 · 353 · 1307 · 2423 32
147578947676144 24 · 43 · 211 · 367 · 1091 · 2539 32
4289395775422432 25 · 127 · 211 · 401 · 1861 · 6703 64
5976833582079328 25 · 181 · 197 · 353 · 431 · 34429 64
1663944565537013728 25 · 131 · 223 · 311 · 2179 · 2626607 64
206177959637947617894769024 27 · 257 · 84691 · 101891 · 116041 · 6259139 74752
2996153601600440129026407808 27 · 257 · 89021 · 93239 · 118621 · 92505877 83584
48083019473926272314825065088 27 · 257 · 97213 · 97973 · 100957 · 1520132521 287264

Table 2: The above primitive weird numbers have six distinct prime factors
and have been found by means of Theorem 3.1.2 with k = 4 for m = 2h and
k = 3 for m = 32896 = 27 · 257

5 - Tracking weird numbers with several distinct prime factors and

eventual odd weird numbers

As we have seen, Theorems 3.1.1 and 3.1.2 provide two distinct strategies
to track primitive weird numbers with several distinct prime factors. The same
approaches could be applied to track odd weird numbers. If m > 1 is odd, the
integers w that one creates by means of Theorem 3.1.1 or 3.1.2 are odd, and if
the conditions are fulfilled, w would be a primitive odd weird number.
For tracking a weird number with several distinct prime factors (even or
odd) both strategies (3.1.1 and 3.1.2) start with choosing a deficient number
m with deficience d. As a general rule, since we want k > d, small values of d
have to be preferred in order to keep the computational complexity low. Indeed,
the case d = 1 corresponds to m = 2h assuming that no further almost perfect
numbers exist. This case has been largely discussed in [6, 8]. The only known
composite numbers with d = 2 are even. For example, m = 136, m = 32896
have deficience 2 [10, Sequence A191363] and, as shown in Table 1 and Table 2,
Theorem 3.1 allows to find several primitive weird numbers starting with such
values of m. We expect that primitive weird numbers could be generated also
from m = 2147516416, whose deficience is 2, both with Theorem 3.1.1 and 3.1.2.
Unfortunately the computation becomes dramatically longer.
As far as we know there are no known integers with deficience d = 3. The
only known integers with deficience d = 4 are even, and the only known integer
with d = 5 is 9 for which the above approaches are hard to apply.
An interesting case is d = 6. There are several integers whose deficience is 6,
some of which are odd. The list starts with 7, 15, 52, 315, 592, 1155, 2102272,
815634435, and no other terms are known [10, Sequence A141548].
So, if one wants to track odd weird numbers, the starting m in both ap-
proaches 3.1.1 and 3.1.2, could be m = 315, m = 1155 or m = 815634435.
Unfortunately our attempts to track an odd weird numbers from such a choice
of m have been unfruitful.

8
 6 - Conclusion

As one can argue from a table of primitive weird numbers, most of the
primitive weird numbers are of the form 2k pq for k ∈ N, and primes p and q.
This was already pointed out in [8] where the third author, among other things,
conjectured that there are infinitely many primitive weird numbers of the form
2k pq.
It seems that primitive weird numbers that are not of this form become
rarer. For example, between the 301st and the 400th, there are only 7 primitive
weird numbers that are not of the form 2k pq. Five of them have four distinct
prime factors and two of them have five distinct prime factors. The existence
of several weird numbers having six distinct prime factors leads to the following
conjecture.
C o n j e c t u r e 6.1. Given an integer k ≥ 3, there exists a primitive weird
number having at least k distinct prime factors.
Of course, a positive answer would settle the question of the infiniteness of
primitive weird numbers.
However, a proof that primitive weird numbers have a bounded number of
distinct prime factors would not settle neither the question of the infiniteness of
primitive weird numbers nor the question of the existence of odd weird numbers.

A c k n o w l e d g m e n t s. Part of this research was done while the first author

was visiting the Department of Mathematics and Computer Science at Wesleyan
University.

References

[1] Benkoski, S.T. Elementary problem E2308, Amer. Math. Monthly 79

(1972), 774.
[2] Benkoski, S.T. and Erdős, P., On weird and pseudoperfect numbers, Math-
ematics of Computation, 28 (1974), 617–623.
[3] Cramér, H., On the order of magnitude of the difference between consecutive
prime numbers, Acta Arithmetica 2 (1936), 23–46.
[4] Friedman, C.N., Sums of divisors and Egyptian Fractions, Journal of Num-
ber Theory 44 (1993), 328–339.
[5] Guy, R.K., “Unsolved Problems in Number Theory”, Third Edition,
Springer, 2004.
[6] Iannucci, D.E., On primitive weird numbers of the form 2k pq, (2015),
arXiv:1504.02761v1.
[7] Kravitz, S., A search for large weird numbers, Journal of Recreational
Mathematics 9 (1976), 82—85.

9
 [8] Melfi, G., On the conditional infiniteness of primitive weird numbers, Jour-
nal of Number Theory 147 (2015), 508-514.
[9] Pajunen, S., On primitive weird numbers, in “A Collection of manuscripts
related to the Fibonacci sequence”, V.E. Hoggatt, Jr. and M. Bicknell-
Johnson (Eds.) 1980, Fibonacci Association, 162–166.

[10] Sloane, N.J.A.., “The On-line Encyclopedia of Integer Sequences”,

www.oeis.org

Gianluca Amato
Università “G. D’Annunzio” di Chieti-Pescara,
Dipartimento di Economia
Viale della Pineta, 4
I-65129, Pescara, Italy
e-mail: gianluca.amato@unich.it

Maximilian Hasler
Université des Antilles,
Département Scientifique Interfacultaire
B.P. 7209 Campus de Schoelcher
F-97275 Schoelcher cedex, Martinique, French West Indies
e-mail: MHasler@Univ-AG.fr

Giuseppe Melfi (corresponding author)

University of Applied Sciences of Western Switzerland,
HEG-Arc
Espace de l’Europe, 21
CH-2000 Neuchâtel, Switzerland
e-mail: Giuseppe.Melfi@he-arc.ch

Maurizio Parton
Università “G. D’Annunzio” di Chieti-Pescara,
Dipartimento di Economia
Viale della Pineta, 4
I-65129, Pescara, Italy
e-mail: maurizio.parton@unich.it

10