• 検索結果がありません。

1Introduction Aboutk-perfectnumbers

N/A
N/A
Protected

Academic year: 2022

シェア "1Introduction Aboutk-perfectnumbers"

Copied!
6
0
0

読み込み中.... (全文を見る)

全文

(1)

About k-perfect numbers

Mih´aly Bencze

Abstract

ABSTRACT. In this paper we present some results aboutk-perfect numbers, and generalize two inequalities due to M. Perisastri (see [6]).

1 Introduction

Definition. A positive integer n is k-perfect if σ(n) = kn, when k > 1, k ∈ Q. The special case k = 2 corresponds to perfect numbers, which are intimately connected with Mersenne primes. We have the following smallest k-perfect numbers. For k = 2 (6,28,496,8128, ...), for k = 3 (120,672,523776,459818240, ...),fork= 4 (30240,32760,2178540, ...),fork= 5 (14182439040,31998395520, ...),fork= 6 (154345556085770649600, ...).

For a given prime numberp, ifnisp-perfect andpdoes not dividen, then pn id (p+ 1)−perfect. This imples that an integernis a 3−perfect number divisible by 2 but not by 4, if and only if n2 is an odd perfect number, of which none are known. If 3nis 4k−perfect and 3 does not dividen, thenn is 3k−perfect.

Ak−perfect number is a positive integernsuch that its harmonic sum of divisors isk.

For the perfect numbers we have the followings: 28 = 13 + 33, 496 = 13+ 33+ 53+ 73,8128 = 13+ 33+ 53+ 73+ 93+ 113+ 133+ 153 etc. We posted the following conjecture:

Conjecture. (Bencze, M., 1978) Ifnisk-perfect, then exist odd positive integers ui (i= 1,2, ..., r) such that

Key Words: perfect numbers

2010 Mathematics Subject Classification: 11A25 Received: May, 2013.

Revised: September, 2013.

Accepted: November, 2013.

45

(2)

n=

r

X

i=1

uk+1i

MAIN RESULTS

Theorem 1. If f : R → R is convex and increasing, N =pα11pα22...pαnn written in cannonical form isk-perfect, then:

n

X

i=1

f 1

pi

 nf

n

q3 2−1

if N is even nf

3n√ k2−1

if N is odd Proof. IfN is even then it follows

n

Y

i=1

pi+ 1 pi > 3

2 Forx≥3 holdsx+1x3

r x

x−1

2

(see [9]), therefore ifNis odd then yields

n

Y

i=1

pi+ 1 pi > 3

v u u t

n

Y

i=1

pi

pi−1 2

> 3

√ k2

because

n

Y

i=1

pi

pi−1 =k

n

Y

i=1

pαii+1 pαii+1−1 > k Using the AM-GM inequality we obtain:

n

Y

i=1

pi+ 1

pi ≤ 1

n

n

X

i=1

pi+ 1 pi

!n

= 1 + 1 n

n

X

i=1

1 pi

!n

Finally

n

X

i=1

1 pi

>

 n

n

q3 2−1

if N is even n3n

k2−1

if N is odd

Because f is convex and increasing from Jensen’s inequality we get

(3)

n

X

i=1

f 1

pi

≥nf 1 n

n

X

i=1

1 pi

!

 nf

n

q3 2−1

if N is even nf3n

k2−1

if N is odd (1)

Theorem 2. If g : R →R is convex and increasing, N = pα11pα22...pαnn written in cannonical form isk-perfect, then:

n

X

i=1

g 1

pi

 ng

1−qn

6 2

if N is even ng

1−qn

8 2

if N is odd Proof. We have the following:

n

Y

i=1

pi

pi−1 =

n

Y

i=1

pαii+1−1 (pi−1)pαii

n

Y

i=1

pαii+1 pαii+1−1 =k

n

Y

i=1

1 1− 1

pαii +1

=k

n

Y

i=1

X

j=0

1 pi

j

≤

≤k

n

Y

i=1

X

j=0

1 p2ji

<



 k

P

n=1 1

n2 if N is even k

P

n=0 1

(2n+1)2 if N is odd

= ( 2

6 if N is even

2

8 if N is odd From AM-GM inequality yields

n

Y

i=1

pi pi−1 ≥

 n

n

P

i=1 pi−1

pi

n

=

 n n−

n

P

i=1 1 pi

n

therefore

n

X

i=1

1 pi

<

 n

1− qn

6 2

if N is even n

1− qn

8 2

if N is odd According to Jensen’s inequality yields

n

X

i=1

g 1

pi

≤ng 1 n

n

X

i=1

1 pi

!

 ng

1−qn

6 2

if N is even ng

1−qn

8 2

if N is odd

(4)

Corolloary 1. IfN =p1 p2 ...pnnwritten in cannonical form isk-perfect then:

 n

n

q3 2−1

if N is even n3n

k2−1

if N is odd

<

n

X

i=1

1 pi

<

 n

1− qn

6 2

if N is even n

1− qn

8 2

if N is odd Theorem 3. Ifx, t >0 then

(x+ 1)tx+11 −xt1x ≤1

Proof. Fort= 1 we have the equality. Let 0 < t <1. Since the function u(x) = xt1x is continuous and differentiable we can apply the Lagrange’s theorem and we obtain

(x+ 1)tx+11 −xtx1

(x+ 1)−x =u(x+ 1)−u(x)

(x+ 1)−x =u0(z) whenx < z < x+ 1 hence we have the inequality

tz1

1−1 zlnt

<1 or 1−1

zlnt < t1z. Developingt1z into McLauren’s series it results

1−1

zlnt <1− 1

1!zlnt+ 1

2!z2ln2t− 1

3!z3ln3t+...

or

X

r=2

(−1)rlnrt r!zr >0 or

X

r=2

lnr1t r!zt >0

that is obvious because ln1t > 0 due to 1t > 1. Let be t > 1. Then is enough to show that the functionV (x) =x

tx1 −1

is decreasing.

DifferentiableV we get V0(x) =tx1 −t1x · 1

xlnt−1 =−

X

r=2

lnrt xr(r−1)!

1−1

r

<0

Since V is decreasing and we may say that V(x+ 1) < V (x) hence and from it follows the inequality of the ennunciation.

Corollary 2. IfN =pα11pα22...pαnn is a k-perfect number written in can- nonical form, then:

(5)

ln32 if N is even

2

3lnk if N is odd <

n

X

i=1

1 pi

<

( ln62 if N is even ln82 if N is odd Proof. Using the Theorem 3 it is proved that the series

n n

r3 2 −1

!!

n∈N

and n3n

k2−1

n∈N

are decreasing, and the series n 1− n

r 6 kπ2

!!

n∈N

and n 1− n r 8

2

!!

n∈N

are increasing. It means that the minimum and maximum are reached only thenn→ ∞.

Sincen → ∞we have 0· ∞. That is why L’Hospital rule and so we find the results of the enunciation.

Remark 1. For k= 2 we reobtain the M.Perisastri’s inequality

n

X

i=1

1 pi

<2 lnπ 2 (see [6]).

Corollary 3. Let N = pα11pα22...pαnn be a k−perfect number written in cannonical form and Pmax = {p1, p2, ..., pn} and Pmin = min{p1, p2, ..., pn}, then

Pmin<

( 1

n

3

2−1 if N is even

1

3n

k2−1 if N is odd and

Pmax>

1 1−nq

6 2

if N is even

1 1−nq

8 2

if N is odd Proof. Considering that

n Pmax

<

n

X

i=1

1 pi

respective

n

X

i=1

1 pi

< n Pmin

from the theorem if follows the affirmation.

(6)

Remark 2. Let N =p1 p2 ...pnn be a k-perfect number written in can- nonical form, then

Pmin< 2n k2−1 + 2 (see the method of M. Perisastri’s)

Acknowledgements. The author wishes to express his gratitude to the Or- ganizing Committee of the workshop ”Workshop on Algebraic and Analytic Number Theory and their Applications” (PN-II-ID-WE - 2012 - 4 -161). The publication of this paper is supported by the grant of CNCS-UEFISCDI (Ro- manian National Authority for Scientific Research): PN-II-ID-WE - 2012 - 4 -161.

References

[1] B. Apostol, Extremal orders of some functions connected to regular in- tegers modulon,An. St. Univ. Ovidius Constanta, Vol. 21(2),2013, 5-19.

[2] M. Bencze, On perfect numbers, Studia Mathematica, Univ. Babes- Bolyai, Nr. 4, 1981, 14-18.

[3] G. Hardy, D.E. Littlewood, G. Polya,Inequalities, Cambridge, University Press, 1964.

[4] H.J. Kanold, Uber mehrfach volkommene Zahlen, II. J. Reine Angew.¨ Math., 1957, 197, 82-96.

[5] Octogon Mathematical Magazine (1993-2013).

[6] M. Perisastri,A note on odd perfect numbers, The Mathematics Student 26(1958), 179-181.

[7] J. S´andor, B. Cristici, eds.: Handbook of number theory II, Dordrecht, Kluwer Academic, 2004.

[8] W. Sierpinski,Elementary theory of numbers,Warsawa, 1964.

[9] The American Mathematical Monthly, E.2308(1971), E.2162(1969) (Simeon Reich’s note).

Mih´aly Bencze, Str. H˘armanului 6,,

505600 S˘acele-N´egyfalu, Jud. Bra¸sov, Romania Email: [email protected]

参照

関連したドキュメント

On the essential logical structure of inter-universal Teichm¨ uller theory I 19:10 – 20:10 Shinichi Mochizuki (RIMS, Kyoto University).. On the essential logical structure

The author would like to express his thanks to the editor for his kind help and invaluable suggestions in the formatting and writing of this

Rigorous numerical methods for infinite-dimensional discrete dynamical systems 16:30-17:00 Yutaka Ishii (Kyusyu University, Japan). Holomorphic dynamics and interval

Keywords: continuous time random walk, Brownian motion, collision time, skew Young tableaux, tandem queue.. AMS 2000 Subject Classification: Primary:

n , 1) maps the space of all homogeneous elements of degree n of an arbitrary free associative algebra onto its subspace of homogeneous Lie elements of degree n. A second

Rhodes, “A matrix inequality associated with bounds on solutions of algebraic Riccati and Lyapunov equations,” IEEE Transactions on Automatic Control, vol. Mori, “Comments on

In section 2, applying these results for a selfmapping T of a nonempty bounded and closed subset C of a uniformly convex Banach space, we present a (for- mal) generalization of

3AG/2G A becomes, after elementary transformations, exactly inequality 1.10, while the second inequality of 2.2, that is, 2PG/P G &gt; 3AG/2A G becomes inequality 1.7..