Sie sind auf Seite 1von 5

MATHEMATICS OF COMPUTATION

Volume 00, Number 0, Pages 000000


S 0025-5718(XX)0000-0

CRITERIA FOR ALMOST PERFECT AND DEFICIENT


NUMBERS
KENETH ADRIAN P. DAGAL, JOSE ARNALDO B. DRIS
Abstract. If (n) = 2n 1, then n is called an almost perfect number. In
this article, we give nontrivial lower and upper bounds for I(n), the abundancy
index of n, when n is almost perfect. We show that these bounds are both
necessary and sufficient in order for n to be almost perfect. Lastly, we give a
generalization of this criterion for the case of p-deficient integers m satisfying
(m) = 2m p, for some integer p > 1.

1. Introduction
If n is a positive integer, then we write (n) for the sum of the divisors of n.
A number n is almost perfect if (n) = 2n 1. It is currently unknown whether
there are any other almost perfect numbers apart from those of the form 2k , where
k 0. Antalan and Tagle [1] have shown that if there exist even almost perfect
numbers m other than the powers of two, then m takes the form m = 2r b2 where
b is an odd composite.
(z) 1
We denote the abundancy index I of the positive integer z as I(z) =
.
z
2. Preliminary Lemmata
We begin with some preliminary results.
Lemma 2.1. If n is a positive integer which satisfies the inequality
2n
I(n) < 2,
n+1
then n is almost perfect.
Proof. Let n be a positive integer, and suppose that
2n
I(n) < 2.
n+1
We then have
2n2
(n) < 2n.
n+1
But
2
2n2
= 2n 2 +
(n) < 2n.
n+1
n+1
Received by the editor October 12, 2015.
2010 Mathematics Subject Classification. Primary 11A05; Secondary 11J25, 11J99.
1Key Words and Phrases: almost perfect number, abundancy index
c
XXXX
American Mathematical Society

KENETH ADRIAN P. DAGAL, JOSE ARNALDO B. DRIS

Since (n) is an integer and 1 n, this last chain of inequalities forces


(n) = 2n 1,
and we are done.

It turns out that the inequality involving n, as given in Lemma 2.1, is also
necessary for n to be almost perfect.
Lemma 2.2. If n is almost perfect, then n satisfies the inequality
2n
I(n) < 2.
n+1
Proof. Let n be a positive integer, and suppose that (n) = 2n 1.
It follows that
1
(n)
= 2 < 2.
I(n) =
n
n
Now we want to show that
2n
I(n).
n+1
2n
. (Note that this forces n > 1.) MimAssume to the contrary that I(n) < n+1
icking the proof in Lemma 2.1, we have
(n) <

2n2
,
n+1

from which it follows that


2
< (2n 2) + 1 = 2n 1.
n+1
This contradicts our assumption that n is almost perfect. Hence the reverse
inequality
2n
I(n)
n+1
holds.

(n) < 2n 2 +

Remark 2.3. By their definition, all almost perfect numbers are automatically deficient. But of course, not all deficient numbers are almost perfect.
By Remark 2.3, it seems natural to try to establish an upper bound for the
abundancy index of an almost perfect number n, that is strictly less than 2, and
which (perhaps) can be expressed as a rational function of n (similar to the form
of the lower bound given in Lemma 2.1 and Lemma 2.2).
Lemma 2.4. If n is a positive integer which satisfies the inequality
2n
2n + 1
I(n) <
,
n+1
n+1
then n is almost perfect.
Proof. Let n be a positive integer, and suppose that
2n + 1
2n
I(n) <
.
n+1
n+1
Again, mimicking the proof in Lemma 2.1, we have
2n2 + n
2n2
(n) <
,
n+1
n+1

CRITERIA FOR ALMOST PERFECT AND DEFICIENT NUMBERS

from which it follows that


2
1
(n) < 2n 1 +
.
n+1
n+1
Since n 1, this last chain of inequalities forces the equation
2n 2 +

(n) = 2n 1
to be true. Consequently, n is almost perfect, and we are done.

We now show that the (nontrivial) upper bound obtained for the abundancy
index of n in Lemma 2.4 is also necessary for n to be almost perfect.
Lemma 2.5. If n is almost perfect, then n satisfies the inequality
2n + 1
2n
I(n) <
.
n+1
n+1
Proof. It suffices to prove that if n is almost perfect, then the inequality
2n + 1
I(n) <
n+1
holds. To this end, assume to the contrary that
2n + 1
I(n).
n+1
Mimicking the proof in Lemma 2.4, we obtain
2n 1 +

1
2n2 + n
=
(n),
n+1
n+1

from which it follows that


2n 1 < (n).
This contradicts our assumption that n is almost perfect, and we are done.

3. Generalization
In Lemma 2.4 and Lemma 2.5, if we denote the deficiency of n by
D(n) = 2n (n) = 1,
then the criterion takes the form
2n
2n + D(n)
I(n) <
.
n + D(n)
n + D(n)
We use this observation in generalizing our criterion to the case of p-deficient integers, where p > 1.
If an integer m satisfies (m) = 2m p for some integer p > 1, then m is said
to be p-deficient.
The following result gives both necessary and sufficient conditions in order for
m to be p-deficient.
Lemma 3.1. Let m be a positive integer. Then m is p-deficient if and only if the
following chain of inequalities hold:
2m + D(m)
2m
< I(m) <
,
m + D(m)
m + D(m)
where 1 < p = D(m) = 2m (m).

KENETH ADRIAN P. DAGAL, JOSE ARNALDO B. DRIS

Proof. Suppose that m is p-deficient, so that (m) = 2m p (where 1 < p). First,
we show that p m. This readily follows from the inequality m (m) = 2m p.
Next, we show that p 6= m. Indeed, if p = m were true, then
(m) = 2m p = m m = 1 = p,
contradicting 1 < p. Consequently, we now know that p < m. We now show that
m satisfies the inequalities
(m)
2m + p
2m
< I(m) =
<
.
m+p
m
m+p
Multiplying through by m(m + p), we obtain
2m2 < (m + p)(m) < 2m2 + mp.
Using the fact that m is p-deficient, we get
2m2 < (m + p)(2m p) < 2m2 + mp.
Collecting like terms and simplifying, we derive
0 < mp p2 < mp.
Since 1 < p, the left-hand inequality is just p < m, which is true. The right-hand
inequality is just 0 < p2 , which is also true.
Now, suppose that the inequalities
2m
(m)
2m + p
< I(m) =
<
m+p
m
m+p
are true, for some integer p > 1. As before, we have
2m2
2m2 + mp
< (m) <
.
m+p
m+p
(The proof that follows is taken from [5].)
We want to show that (m) = 2m p. To this end, we compute
LHS =

p2 mp
p(p m)
2m2
= (2m p) +
= (2m p) +
,
m+p
m+p
m+p

and
RHS =

2m2 + mp
mp
= (2m p) +
.
m+p
m+p

Since p < m, we have


LHS = (2m p) +

p(p m)
1
> (2m p) ,
m+p
2

and
RHS = (2m p) +

1
mp
< (2m p) + .
m+p
2

Consequently, we have
(2m p)

1
2m2
2m2 + mp
1
< LHS =
< (m) <
= RHS < (2m p) + .
2
m+p
m+p
2

This forces
(m) = 2m p.


CRITERIA FOR ALMOST PERFECT AND DEFICIENT NUMBERS

4. Conclusion
Collecting all the results from Lemma 2.1, Lemma 2.2, Lemma 2.4, Lemma 2.5,
and Lemma 3.1, we now have the following theorems.
Theorem 4.1. Let M be a positive integer. Then (M ) = 2M 1 if and only if
the following chain of inequalities hold:
2M
2M + 1
I(M ) <
.
M +1
M +1
Theorem 4.2. Let N be a positive integer. Then (N ) = 2N p for some integer
p > 1 if and only if the following chain of inequalities hold:
2N
2N + p
< I(N ) <
.
N +p
N +p
Remark 4.3. Note that equality holds in
2M
I(M )
M +1
if and only if M = 1.
Remark 4.4. It is trivial to verify that the known almost perfect numbers n = 2k
(for integers k 0) satisfy the inequalities in Theorem 4.1. In fact, Theorem 4.1
can be used to rule out particular families of integers from being almost perfect
numbers.
This can, of course, be attempted for the case N = 2r b2 , where b is an odd
composite [1].
In a similar vein, Theorem 4.2 can be used to show that certain integers are
deficient.
Finally, we end this section with an open problem for other researchers to solve.
Conjecture 4.5. The bounds in Theorem 4.1 and Theorem 4.2 are best-possible.
5. Acknowledgment
The authors express their gratitude to Will Jagy for having answered one of the
second authors questions at Math@StackExchange (see [2], and also the related
questions [3] and [4]), from which the proofs of Lemmas 2.1, 2.2, 2.4 and 2.5 were
patterned after. The authors would also like to thank John Rafael M. Antalan from
the Central Luzon State University, for helpful chat exchanges on this topic.
References
1. J. R. M. Antalan and R. P. Tagle, Revisiting forms of almost perfect numbers, preprint (2014).
2. J. A. B. Dris, Does the following inequality hold if and only if N is an odd deficient number?,
http://math.stackexchange.com/questions/462307.
3. J. A. B. Dris, Does this inequality hold true, in general?, http://math.stackexchange.com/
questions/445467.
4. J. A. B. Dris, Does this inequality have any solutions in N?, http://math.stackexchange.com/
questions/414663.
5. J. A. B. Dris, An elementary question about integers and fractions - Part 2, http://math.
stackexchange.com/questions/1473018.
Far Eastern University, Nicanor Reyes Street, Sampaloc, Manila, Philippines
E-mail address: kdagal@feu.edu.ph; josearnaldobdris@gmail.com