A form of the Chebyshe_ inequalities [3] such that if a1 _ a2 _ : : : _ as and
b1 _ b2 _ : : : _ bs are real numbers, then
Xs
i=1
ai
!
Xs
i=1
bi
!
_s
Xs
i=1
aibi (12)
Let ai = dki
, bi = dti
(i = 1; s), where k; t > 0. Then
_k(n)_t(n) _ d(n)_k+t(n) (13)
For an application of (12) we note that by (2) this gives
_k+t(n)
_t(n)
_
_k(n)
d(n)
_ nk=2 (14)
We note that a simple proof of (2) follows by the application of the inequality
(a1 + a2 + : : : + as)
_
1
a1
+
1
a2
+:::+
1
as
_
_ s2
for ai = dki
, by taking into account of (7).
We now write the following extension of the Chebyshe_ inequality ([3]): If 0 _ a11
_
: : : _ a1
n, 0 _ a21
_ : : : _ a2
n; : : :, 0 _ am2
_ : : : _ amn
(where the superscripts are not
powers!) then
Xs
i=1
a1i
Xs
i=1
a2i
:::
Xs
i=1
ami
_ sm1
Xs
i=1
a1i
a2i
: : : ami
(15)
Let now in (14) a1i
= dk1
i ; : : : ; ami
= dkm
i , where di (i = 1; s) are the divisors of n, while
k1; : : : ; km > 0 are given real numbers. One obtains
_k1(n)_k2(n) : : : _km(n) _ (d(n))m1_k1+:::k2+:::+km(n) (16)
This is an extension of (12). Applying (2) this yields:
_(n)
k1+k2+:::+km
_ki(n)
_n
1
2
mP1
j=1
kj
; j 6= i (17)
For m = 2 one reobtains the weaker form of (13). Other relations can be found in [7].
Bibliography
1. G.H. Hardy and E.M. Wright, An introduction to the theory of numbers, Oxford,
1960.
191
2. E. Kr•atzel, Zahlentheorie, Berlin, 1981.
3. D.S. Mitrinovi_c, Analytic inequalities, Springer Verlag, 1970.
4. S. Philipp, Amer. Math. Monthly 72(1965).
5. G. P_olya and G. Szeg•o, Aufgaben und Lehrs•atze aus der Analysis, Berlin 1924, Band
I.
6. P. Schweitzer, On an inequality related to arithmetic mean (Hungarian), Math. Phys.
Lapok 23(1914), 257-261.
7. J. S_andor, On the sum of powers of divisors of a number (Hungarian), Mat. Lapok,
XCIV, 8/1989, 285-288.
192
21 Between totients and sum of divisors: the
arithmetical function
Introduction
Let n =
Yr
i=1
p_i
i (pi primes, _i _ 1, i = 1; r) be the canonical representation of
the natural number n > 1. Then it is well-known that the Euler function ' has the
representation
'(n) = n
Yr
i=1
(1 p1
i ); '(1) = 1:
This arithmetical function plays an important role in many problems of number theory,
algebra, geometry, etc. Another important function is the sum-of-divisors function _,
de_ned by
_(n) =
X
djn
d;
where the sum runs through all divisors d of n. We have
_(n) =
Yr
i=1
(p_i+1
i
1)=(pi 1)
(see [1], [3], [7], [9], [22]). These two functions have been studied extensively and many
interesting and important results have been proved. On the other hand there are some
open problems which are very di_cult to study at the present state of the science (e.g. the
study of the equation _(n) = 2n for odd n (odd perfect numbers), problems on _(_(n)),
_('(n)), etc.).
Dedekind's arithmetical function is de_ned by
(n) = n
Yr
i=1
(1 + p1
i ); (1) = 1:
This function proved to be useful in some problems of number theory ([2], [13], [14],
[18], [24], [25]) and has interesting connections with the above mentioned and other
arithmetical
functions. There are known some generalizations ([4], [18], [19]), one of which will
be used by us, namely the function
k(n) = nk
Yr
i=1
(1 + pk
i ); k _ 1:
193
Dedekind's
multiplicative, Pf is a multiplicative function.
Theorem. When f is multiplicative, we have the following relation:
Pf (n) =
Yr
i=1
X_i
k=0
f(pki
)(p_ik
i
p_ik1
i ) (4)
where n =
Yr
i=1
p_i
i is the prime factorization of n > 1 (and p1
i := 0).
Proof. Pf (n) =
Yr
i=1
Pf (p_i
i ), since Pf is multiplicative. By (3) one has
Pf (p_i
i) =
X
djP
_i
i
'(d)f(p_i
i =d) =
X_i
k=0
(pki
pk1
i )f(p_ik
i );
which is the same as (4).
Remark. Let f(n) = n_. Then we get
Pf (n) =
Yr
i=1
X_i
k=0
pk_
i (p_ik
i
p_ik1
i ) (5)
Various other relations for particular f can be obtained. For asymptotic relations and
open problems, see [1], [3], [7], [8].
Bibliography
1. K. Alladi, On generalized Euler functions and related totients, in New Concepts in
Arithmetic Functions, Mat. Sc. Report, no.83, Madras 1975.
2. S.S. Pillai, On an arithmetic function, J. Annamalai Univ., 2(1933), 243-248.
3. J. S_andor, A.V. Kramer, • Uber eine Zahlentheoretische funktion, Mathematica
Moravica 3(1999), 53-62.
4. R. Sivaramakrishnan, On the three existension of Pillai's arithmetic function _(n),
Math. Student 39(1971), 187-190.
223
5. S. Tab^_rc_a, T. Tab^_rc_a, Two new functions in number theory, Octogon Mathematical
Magazine, vol.5, 1998, 74-75.
6. S. Tab^_rc_a, Some considerations on the sum
Xn
i=1
[i; n], Octogon Mathematical Magazine,
vol.9, 2001, 415-417.
7. E. Teu_el, Aufgabe 599, Elem. Math. 25(1970), p.65.
8. L. T_oth, The unitary analogue of Pillai's arithmetical function, Collect. Math.
40(1989), 19-30.
224
24 A note on semigroup valued multiplicative
functions
Let (G; _) be a commutative semigroup with unit element, where _ : G _ G ! G
denotes the semigroup operation in G. We say that a function f : N_ ! G is (G; _)-
multiplicative (or "G-multiplicative"), if
f(ab) = f(a) _ f(b) for all (a; b) = 1 (1)
where (a; b) denotes the g.c.d. of the natural numbers a and b.
The aim of this note is to prove the following result.
Theorem. Let f : N_ ! G be a G-multiplicative function and suppose that f(k) has
an inverse element f1(k) in G. Then the function Fk : N_ ! G, de_ned by
Fk(n) = f(nk) _ f1(k); n 2 N_ (2)
is G-multiplicative, too.
Proof. A proof can be obtained by generalizing the idea from [2], but here we shall
adopt another method. For n 2 N_ and p prime denote ap(n) = t 2 N if ptjn, pt+1 - n,
i.e. the greatest power of p which divides n. Then clearly,
n=
Y
pjk
pap(n) _
Y
pjn;p-k
pap(n); k =
Y
pjn
pap(k) _
Y
pjk;k-n
pap(k) (3)
by the prime factorization of n and k. Let us denote by
Y_
p2P
xp a product (with respect
to the operation _) of elements xp 2 G, when p runs over a set P. By using (1) and (3),
a simple calculation gives
Fk(n) =
Y_
pjn
[f(pap(n)+ap(k)) _ f1(pap(k))]; (4)
where we used the properties of G and the obvious fact that a _ b has an inverse (a _ b)1
in G if and only if a and b have inverses. Of course, (a _ b)1 = a1 _ b1. Now the
Gmultiplicativity
of Fk follows at once from (4) by remarking that ap(xy) = ap(x) for pjx,
pjy. Finally, we note that this proof is based on the explicit expression of (2).
225
Remarks. 1) The notion of G-multiplicative functions is a common generalization
of the classical multiplicative and additive functions. Let (G; _) = (R; _) the set of real
numbers, endowed with the usual product operation. Then each a 2 R, a 6= 0, has an
inverse, so we obtain (with a new proof) the result from [2]: If f is a multiplicative function
and f(k) 6= 0, then Fk(n) = f(kn)=f(k) is multiplicative, too. We note that this theorem
extends a result of [1].
2) Let (G; _) = (R; +). Then we get the theorem: If f is an additive function (i.e.
satisfying f(xy) = f(x) + f(y) for (x; y) = 1, then Fk(n) = f(kn) f(k) is additive,
too. As a consequence we can remark that if f and g are additive functions, then the
function h de_ned by h(n) = g(k)f(kn) f(k)g(kn) is additive, too. This follows from
the equality h(n) = g(k)[f(kn)f(k)]f(k)[g(kn)g(k)]. Another corollary says that if
f is additive, then for (k; r) = 1, the function s de_ned by s(n) = f(krn)f(kn)f(rn)
is additive, too.
Bibliography
1. D.H and E. Lehmer, Heuristics, Anyone?, Studies in mathematical analysis and
related topics, Stanford Univ. Press, Stanford, Calif., 1962, 202-210.
2. A. Makowski, Remark on multiplicative functions, Elem. Math., 1972, 27/6.
226
Chapter 4. Divisibility properties of
numbers and functions
"... one of the attractions of the subject-matter is that, whilst it is rich in ideas, it requires
little initiation..."
(H. Halberstam and K.F. Roth, Sequences, Springer, 1983)
227
1 On a divisibility property
Let p > 3 be an odd prime (i.e. p _ 5). Then Wolstenholme's theorem asserts that
the numerator of the fraction in 1 +
1
2
+:::+
1
p1
is divisible by p2, in other words, if
a
b
=
1
1
+
1
2
+:::+
1
p1
;
then
p2ja ([2]): (1)
By using properties of the _eld Zp, in [2] it is proved that:
1) If p _ 5 and
a
b
=
1
12 +
1
22 + : : : +
1
(p 1)2 ;
then
pja (2)
and
2) If p _ 3 and if
a
b
=