ONTHEM-POWERFREEPARTOF ANINTEGER

Liu Yanni
Department of Mathematics, Northwest University, Xian, Shaanxi, P.R.China
Gao Peng
School of Economics and Management,Northwest University, Xian, Shaanxi, P.R.China
Abstract The main purpose of this paper is using the elementary method to study the mean
value properties of a new arithmetical function involving the m-power free part
of an integer, and give an interesting asymptotic formula for it.
Keywords: Arithmetical function; Mean value; Asymptotic formula
1. Introduction
For any positive integer n, it is clear that we can assume n = u
m
v, where v
is a m-power free number. Let b
m
(n) = v be the m-power free part of n. For
example, b
3
(8) = 1, b
3
(24) = 3, b
2
(12) = 3, . Now for any positive
integer k > 1, we dene another function
k
(n) as following:

k
(n) = max{d : d | n, (d, k) = 1}.
From the denition of
k
(n), we can prove that
k
(n) is also a completely
multiplicative function. In reference [1], Professor F.Smarandache asked us
to study the properties of the sequence {b
m
(n)}. It seems that no one knows
the relations between sequence {b
m
(n)} and the arithmetical function
k
(n)
before. The main purpose of this paper is to study the mean value properties
of
k
(b
m
(n)), and obtain an interesting mean value formula for it. That is, we
shall prove the following conclusion:
Theorem. Let m and k be any xed positive integer. Then for any real
number x 1, we have the asymptotic formula

nx

k
(b
m
(n)) =
x
2
2
(2m)
(m)

p|k
p
m
+ 1
p
m1
(p + 1)
+ O
_
x
3
2
+
_
,
204 SCIENTIA MAGNA VOL.1, NO.1
where denotes any xed positive number, (s) is the Riemann zeta-function,
and

p|k
denotes the product over all different prime divisors of k.
Taking m = 2 in this Theorem, we may immediately obtain the following:
Corollary. For any real number x 1, we have the asymptotic formula

nx

k
(b
2
(n)) =

2
30
x
2

p|k
p
2
+ 1
p(p + 1)
+ O(x
3
2
+
).
2. Proof of the Theorem
In this section, we shall use the analytic method to complete the proof of
the theorem. In fact, we know that b
m
(n) is a completely multiplicative func-
tion, so we can use the properties of the Riemann zeta-function to obtain a
generating function. For any complex s, if Re(s) > 2, we dene the Dirichlet
series
f(s) =

n=1

k
(b
m
(n))
n
s
.
If positive integer n = p

, then from the denition of

k
(n) and b
m
(n) we
have:

k
(b
m
(n)) =
k
(b
m
(p

)) = 1, if p|k,
and

k
(b
m
(n)) =
k
(b
m
(p

)) = p

, if modm, 0 < m and pk.

From the above formula and the Euler product formula (See Theorem 11.6
of [3]) we can get
f(s) =

p
_
1 +

k
(b
m
(p))
p
s
+

k
(b
m
(p
2
))
p
2s
+

k
(b
m
(p
3
))
p
3s
+
_
=

p|k
_
1 +
1
p
s
+
1
p
2s
+ +
1
p
(m1)s
+
1
p
ms
+
1
p
(m+1)s
+
_

pk
_
1 +
p
p
s
+
p
2
p
2s
+ +
p
m1
p
(m1)s
+
1
p
ms
+
p
p
(m+1)s
+
_
=

p|k
1
1
1
p
s

pk
__
1 +
p
p
s
+ . . . +
p
m1
p
(m1)s
_
_
1 +
1
p
ms
+
1
p
2ms
+
_
_
=

p|k
1
1
1
p
s

pk
1
1
1
p
ms

pk
_
1 +
p
p
s
+ . . . +
p
m1
p
(m1)s
_
On the m-power free part of an integer 205
=

p|k
1
1
1
p
s

pk
1
1
p
m(s1)
1
1
p
s1

1
1
1
p
ms
=
(s 1)(ms)
(ms m)

p|k
(1
1
p
s1
)(1
1
p
ms
)
(1
1
p
m(s1)
)(1
1
p
s
)
.
Because the Riemann zeta-function (s) have a simple pole point at s = 1
with the residue 1, we know that f(s)
x
s
s
also have a simple pole point at s = 2
with the residue
(2m)
(m)

p|k
p
m
+ 1
p
m1
(p + 1)
x
2
2
. By Perron formula (See [2]),
taking s
0
= 0, b = 3, T > 1, then we have

nx

k
(b
m
(n)) =
1
2i
_
3+iT
3iT
f(s)
x
s
s
ds + O
_
x
3+
T
_
.
Now we move the integral line to Re s =
3
2
+ , then taking T = x
3
2
, we
can get

nx

k
(b
m
(n))
=
(2m)
(m)

p|k
p
m
+ 1
p
m1
(p + 1)
x
2
2
+
1
2i
_ 3
2
++iT
3
2
+iT
f(s)
x
s
s
ds + O
_
x
3
2
+
_
=
(2m)
(m)

p|k
p
m
+ 1
p
m1
(p + 1)
x
2
2
+ O
_
_
T
T

f(
3
2
+ + it)

x
3
2
+
1 + |t|
dt
_
+O
_
x
3
2
+
_
=
(2m)
(m)

p|k
p
m
+ 1
p
m1
(p + 1)
x
2
2
+ O
_
x
3
2
+
_
.
This completes the proof of Theorem.
Note that (2) =

2
6
and (4) =

4
90
, taking m = 2 in the theorem, we may
immediately obtain the Corollary.
References
[1] F.Smarandache, Only Problems,Not Solutions, Chicago, Xiquan Pub-
lishing House, 1993.
206 SCIENTIA MAGNA VOL.1, NO.1
[2] Pan Chengdong and Pan Chengbiao, Foundation of Analytic Number
Theory, Beijing, Science Press, 1991.
[3] Tom M.Apstol, Introduction to Analytic Number Theory, New York,
Springer-Verlag, 1976.