PlanetMath (more info)
 Math for the people, by the people. Sponsor PlanetMath
Encyclopedia | Requests | Forums | Docs | Wiki | Random | RSS  
Login
create new user
name:
pass:
forget your password?
Main Menu
Owner confidence rating: Very high Entry average rating: Very high
perfect number (Definition)

An positive integer $n$ is called perfect if it is the sum of all positive divisors of $n$ less than $n$ itself. It is not known if there are any odd perfect numbers, but all even perfect numbers have been classified according to the following lemma:

Lemma   An even number is perfect if and only if it equals $2^{k-1}(2^k-1)$ for some integer $k>1$ and $2^k-1$ is prime.
Proof. Let $\sigma$ denote the sum of divisors function. Recall that this function is multiplicative.

Necessity: Let $p=2^k-1$ be prime and $n=2^{k-1}p$ We have that \begin{eqnarray*} \sigma(n) & = & \sigma(2^{k-1}p)\\ & = & \sigma(2^{k-1}) \sigma(p)\\ & = & (2^k-1)(p+1)\\ & = & (2^k-1)2^k\\ & = & 2n, \end{eqnarray*}which shows that $n$ is perfect.

Sufficiency: Assume $n$ is an even perfect number. Write $n=2^{k-1}m$ for some odd $m$ and some $k>1$ Then we have $\gcd(2^{k-1},m)=1$ Thus, $$ \sigma(n)=\sigma(2^{k-1}m)=\sigma(2^{k-1})\sigma(m)=(2^k-1)\sigma(m). $$ Since $n$ is perfect, $\sigma(n)=2n$ by definition. Therefore, $\sigma(n)=2n=2^km$ Piecing together the two formulas for $\sigma(n)$ yields $$ 2^km=(2^k-1)\sigma(m). $$ Thus, $(2^k-1)\mid 2^km$ which forces $(2^k-1)\mid m$ Write $m=(2^k-1)M$ Note that $1\le M<m$ From above, we have: \begin{eqnarray*} 2^km & = & (2^k-1)\sigma(m) \\ 2^k(2^k-1)M & = & (2^k-1)\sigma(m) \\ 2^kM & = & \sigma(m) \end{eqnarray*}Since $m\mid m$ by definition of divides and $M\mid m$ by assumption, we have $$ 2^kM=\sigma(m)\geq m+M=2^kM, $$ which forces $\sigma(m)=m+M$ Therefore, $m$ has only two positive divisors, $m$ and $M$ Hence, $m$ must be prime, $M=1$ and $m=(2^k-1)M=2^k-1$ from which the result follows. $ \qedsymbol$

The lemma can be used to produce examples of (even) perfect numbers:

  • If $k=2$ then $2^k-1=2^2-1=3$ which is prime. According to the lemma, $2^{k-1}(2^k-1)=2^{2-1} \cdot 3=6$ is perfect. Indeed, $1+2+3=6$
  • If $k=3$ then $2^k-1=2^3-1=7$ which is prime. According to the lemma, $2^{k-1}(2^k-1)=2^{3-1} \cdot 7=28$ is perfect. Indeed, $1+2+4+7+14=28$
  • If $k=5$ then $2^k-1=2^5-1=31$ which is prime. According to the lemma, $2^{k-1}(2^k-1)=2^{5-1} \cdot 31=496$ is perfect. Indeed, $1+2+4+8+16+31+62+124+248=496$

Note that $k=4$ yields that $2^k-1=2^4-1=15$ which is not prime.

The sequence of known perfect numbers appears in the OEIS as sequence A000396.




"perfect number" is owned by Wkbj79. [ full author list (2) | owner history (1) ]
(view preamble | get metadata)

View style:

Keywords:  number theory

Attachments:
multiply perfect number (Definition) by CompositeFan
Log in to rate this entry.
(view current ratings)

Cross-references: OEIS, sequence, sufficiency, necessity, multiplicative, function, sum of divisors function, prime, even number, even, odd, divisors, sum, integer, positive
There are 34 references to this entry.

This is version 19 of perfect number, born on 2001-10-15, modified 2007-06-26.
Object id is 206, canonical name is PerfectNumber.
Accessed 14002 times total.

Classification:
AMS MSC11A05 (Number theory :: Elementary number theory :: Multiplicative structure; Euclidean algorithm; greatest common divisors)

Pending Errata and Addenda
None.
[ View all 7 ]
Discussion
Style: Expand: Order:
forum policy
odd perfect numbers DO EXIST! by leavemsg2 on 2008-10-25 12:26:56
it takes a long, long time to digest the following argument:

i believe that (O)dd (P)erfect (N)umbers do NOT exist in the same
manner as (E)ven (P)erfect (N)umbers do!

if we allow the generic formula of (k^(m-1))*(k^m - (k-1)) to de-
fine the situation, then we will see that it encompasses both the
proven EPN(n) == (2^(n-1))*(2^n -1) and a closely-related possible
OPN(n) == (3^(n-1))*(3^n -2).

however, we can't be satisfied this approach, since these two in-
stances of a preferred formula don't have comparable 'pole' posi-
tions when 'n' is set equal to zero.

if we let n= 0, then the 'pole' for EPN(0) == (1/2) *(0) remains
unshifted from the origin, and the 'pole' of OPN(0) == (1/3) *(-1)
is shifted by a (-1). comparing the results with respect to per-
fection between these two instances is already suspect from the
onset.

after a closer examination, though, the formula (3^(n-1))*(3^(n+1)
-2) turns out to be a more appropriate function but with a little
twist involved; if we let n= 0, the 'pole' of OPN(0) == (1/3) *(1)
is shifted by a (+1). we'll use this fact to later offset our clo-
sest OPN formula by a (-1) to bring the two 'poles' into agreement.

now, a 'proper-incipient' EPN is generated when the exponent 'n' is
less than the value inside the portion of the second parentheses of
its generic formula; n= 1 implies that EPN(1) == (1)*(1), but it's
improper since n= 1 is not less than the second (1); n= 2 implies
that EPN(2) == (2)*(3) and since n= 2< (3), we quickly discover that
6 = 1 *2 *3 = 1 +2 +3 = 6 (is an unshifted 'proper-incipient' char-
acteristic that should be shared by the preferred OPN function).

if n= 1, then the preferred OPN formula gives an OPN(1) == (1)*(7)
and 1< (7). so, the OPN(1) would share the proper 'incipient' qua-
lity of its cousin EPN, but not without an adjustment to its value
with an offset of (-1) to combat the earlier mentioned (+1) shift;
allowing for this little twist on the OPN's existance, the OPN's be-
come not only odd, but a bit quirky!

if we subtract '1' from the summand and allow the use of an 'impro-
per' divisor, then we have OPN(1) == (1)*(7) = (-1) + ((1) +7) = 7
(offsetting it), and the incipient OPN presents itself; but it was
not created using the expected 'EPN' axiom. it's the only OPN attach
ed to k= 3 from the preferred OPN formula, but not the only OPN...

if k= 5 and n= 1, then 21 isn't the next OPN, but...

finally, if all OPN's are == (2^(n-1))*(2^(n+1) -1), or (3^(n-1))*(3
^(n+1) -2), or (5^(n-1)) *(5^(n+1) -4), or (7^(n-1))*(7^(n+1) -6), or
(11^(n-1))*(11^(n+1) -10), or (13^(n-1))* (13^(n+1) -12), etc., then
we can see that the sequence of OPN's intermittently emerge as 3, 7,
(not 21), 43, (not 111), 157, etc.

the OPN's are not governed by the same rules as their cousin EPN's &
are not the expected 'gigantic' numbers that contain a multitude of
factors, but actually a set of individual (must be) prime numbers
that are generated by a continually changing formula and by modify-
ing the calculations of both the product and summations due to a
shift from their respective 'pole' positions.

thus, OPN's could barely exist but not in the usual EPN sense... they
are quirky! at best, the term OPN is a fictional misnomer, because
all possible formulas for an OPN lack the offset that is necessary to
coincide with the zero 'pole' of an already Euler-proven EPN formula
which also share the 'proper-incipient' characteristic.

the definition has to be changed... re-read it!
Bill Bouris
[ reply | up ]
"Missing cached output" by CompositeFan on 2007-06-01 16:42:06
For HTML with images mode, this says "Missing cached output, contact an admin," but I can see it just fine in page images mode.
[ reply | up ]
wat's the 5/n problem called? by Lando47 on 2006-12-04 18:16:54
Or maybe its a conjecture. Taht for n > 3, their are 1 < a < b < c such that 5/n = 1/a + 1/b + 1/c. Do they need to be distinct (ie, can we have 1 <= a <= b <= c)? But most importantely, what is this problem called?
[ reply | up ]

Interact
post | correct | update request | add derivation | add example | add (any)