IOSR Journal of Mathematics (IOSR-JM)

e-ISSN: 2278-5728, p-ISSN: 2319-765X. Volume 11, Issue 2 Ver. VI (Mar - Apr. 2015), PP 43-50
www.iosrjournals.org

A Tauberian Theorem for (, , )- Convergence of Ceso Means of

Order k of Functions
Suyash Narayan Mishra, Piyush Kumar Tripathi
& Alok Agrawal
Email:@. . 1 ,@. . 2 , @. . 3
Amity School of Applied Sciences, Amity University, Uttar Pradesh ( Lucknow Campus)
Near Malhaur Railway Station , Gomti Nagar Lucknow (U. P. ) India

Abstract: The objective of this paper to generalize certain Tauberian results proved by Gehring [3] for
summability ( , ; ) of sequences to functions. In [1] A. V. Boyd generalized the Tauberian theorem for
convergence of Cesro means of sequences. In this paper ,we obtain some Tauberian theorems for (, , )
convergence of Cesro means of order k of functions and investigate some of its properties .
Keywords: Tauberian theorem, Absolute and Cesro summability , Lebesgue Integral, Convergence.

I.

Introduction

The notation is similar that are in [3],with the following additional definitions: If > 1 then ,

1
1
denote the n-th Cesro sums of order k for the series
=0 , =0 where = . , denote the
, . Summability ( , 1; ) of will b( , 0; ) of . Mishra and Srivastava [6] introduced the
Summability method (C , , ) for functions by generalizing (C , ) summability method. In this paper, we
discuss some Tauberian theorems for (, , ) convergence of Cesro means of order k of functions and
investigate some of its properties .

II.

Definitions and Some Preliminaries

We would like to first introduce Summability method. Summability method is more general than that of
ordinary convergence. If we are given a sequence , we can construct a generalized sequence
, the arithmetic mean of by this sequence . If is convergent in ordinary sense for all
> 0, then we say that is summable (, 1) to the sum . This (, 1) is called Cesaro mean of first order.
If =

0 +1 +.+
+1

, ie

if a sequence is convergent, it is summable by method of

1

arithmetic mean. Also a series 1 1 + 1 + 1 + is not convergent , but is summable to the sum 2 . The
space of summable sequences is larger than space of convergent sequences. If as , then we say
that sequence is summable by method of arithmetic mean.

=0

For example : Consider the series

0 +1 +.+

And let =

+1

= 0 + 1 + . .

(1)

, It may happen that whereas (1) diverges , the quantities ( the arithmetic mean

of partial sum of series) converges to a definite limit as . For example 1 1 + 1 + 1 + diverges,

but in this case 0 = 1, 1 = 1 1 = 0, 2 = 1 1 + 1 = 1, 3 = 0
( ) =
(1,0,1,0,1 . . ) . Since =
=
=
=

1+(1)
2

0 +1 +.+
+1
1+(1)0
2
(+1)
2

+
1
2

1+(1)1
2

1+(1)2
2

+ ..+

1+ 1
2

/( + 1)

1 1 + 1 + 1 /( + 1)

DOI: 10.9790/5728-11264350

www.iosrjournals.org

43 | Page

A Tauberian Theorem For (, , )- Convergence Of Cesro Means Of Order K Of Functions

1

1+(1)

4( +1)

= +

2(+1)

, If n is even then = +

as and if n is odd then = . So in either

case lim = 2 , . Therefore space of summable sequences is larger than thar of space
of convergent sequences .

f (x) be any function which is Lebesgue-measurable, and that f : [0, + ) R, and integrable in (0, x )

Let

for any finite x and which is bounded in some right hand neighbourhood of origin. Integrals of the form

are

0
x

throughout to be taken as

lim

x 0

Let k 0. If, for

being a Lebesgue integral.

t 0, the integral

x k 1
f ( x)dx
k 1
(
x

t
)
0

g (t ) g ( k ) (t ) kt

(2.1)

g (t ) s as t , we say that function f (x) is summable ( D, k ) to the sum s and we

write f ( x) s( D, k ) as x .
We note that, for any fixed t 0, k 0, it is necessary and sufficient for convergence of (2.1) that
exists and if

f ( x)
dx , should converge .
x2

The

(2.2)

(C, , ) transform of f (x) , which we denote by , ( x) is given by

f (x) ( 0)

( 1) 1
( x y) 1 y f ( y )dy , ( 0, 1)

( )( 1) x
0
x

If this exists for x 0 and

(C, , ) to

s,

and

, ( x)
we

(2.3)

s as x , we say that f (x) is summable

f ( x) s(C, , ) .
We
also
write

tends to a limit
write

x k 1
( x)dx ,
k 1 ,
(
x

t
)
0

U k , , (t ) kt

(2.4)

if this exists, and tends to a limit s as t , we say that the function f (x) is summable ( D, k )(C, , )
to s .
When 0 , ( D, k )(C, , ) and ( D, k )(C, ) denote the same method. Here we give some Gehrings
generalized Tauberian theorems.
Theorem 2.1: Suppose that 0 1 and that () is summable (, ) to s, then is (, , ) convergent
to s if and only if the function , , () is (, , ) convergent to 0.
Theorem 2.2: Suppose that 0 1 and that () is (, , ) convergent. If the function , () is
(, , ) convergent to 0,then () is summable (, , ) to its sum for every > 1.

III.

Now we shall prove the following theorem

Theorem 3.1: Suppose that 0 1 and that () is summable (, ) to s. Then for 1, is

()
summable (, , ) to s if and only if the function () is (, , ) to 0.
,

Proof : Necessary Condition: If = 1,the theorem immediate follows from the summability of (, 1, ). If
> 1, then by consistency theorem for (, , ) summability ( Gehring [3,theorem 4.2.1]) it follows that both

the functions and , () are (, , ) convergent to s. By Hardy [1, Equation (6.1.6)], = +1

+

DOI: 10.9790/5728-11264350

www.iosrjournals.org

44 | Page

A Tauberian Theorem For (, , )- Convergence Of Cesro Means Of Order K Of Functions

1

()

, and the result follows since a linear combination of functions summable (, , ) to itself. The

+1 , ()

sufficient conditions to prove the theorem are :

If > 1, it may be shown as in Sz [ 4 (1) ] , that

1
+1 0

, () (1 +1) =

Where

, = ,

+1

( 0)

(3.1)

( 1) 1
( x y) 1 y f ( y )dy , ( 0, 1)
=

( )( 1) x
0
x

Case (a) : = 0, > 1 is obvious.

Case (b) : 0 1, > 1 , putting

1
1
= +1 0 , () (1 +1) .

We get from (3.1) that = + 1 0 , 1

,
Where , () now has bounded (, , )- variation over 0, . Let

, , 1

= ( + 1)
1
(1
0

1
(1
0

) ,

, (1 )

. Then by theorem 201 of [5], we have

+1
) = M. .
Where = , : 0 . Thus , () has bounded (, , )- variation over 0, . It is readily
seen from Minkowskis inequality that the sum of two (, , ) convergent sequences is also (, , )
convergent and we therefore deduce that f(x) is (, , ) convergent to s.
Case (c ) r=-1,when = 0,the result reduces to Taubers original theorem; when 0 1 it follows from
above theorem . For = 1, the result was proved by Hyslop [2] .
Theorem 3.2 : Let > 0, > 1
, ()
1
2

, and suppose that a(x) is summable (, , ) to s and that

converges . Then a(x) is summable , (, , ) to s . We first prove this theorem under

unreasonable definition (2.2). However ,if the result holds with (2.2), then it must also hold under the definition
of (2.3). This follows from the following Lemmas.

p 1, 1 .

Lemma 3.1: Let

Suppose that f ( x) L(0, x) for finite

x 0 .Suppose that f (x)

C , , p ,according to the definition (2.3).

for x T
for x T

f ( x)
f ( x)
0

Define

(3.2)

Let

, ( y ) denote the expression corresponding to , ( y ) but with f (x) replaced by f (x) .

Then

y
0

p 1

d
, ( y ) dy .
dy

(3.3)

Thus

f (x) is summable C , , p under the definition (2.3).

DOI: 10.9790/5728-11264350

www.iosrjournals.org

45 | Page

A Tauberian Theorem For (, , )- Convergence Of Cesro Means Of Order K Of Functions

Lemma

3.2:

Let

the

hypothesis

k 0, 1 and 0

be

as

in

Lemma

3.1,and

define

f (x)

( D, k )(C, , ) p summability of

.Then

f (x) as above. Let

and

are
f (x)

equivalent.
Proof of Lemma 3.1: We are given that , for some T>0,

p 1

d
, ( x) dx
dx

(3.3)

But since, if (3.3) holds for given T, it holds for any greater T, it must hold for all sufficiently large T. Now by
standard properties of fractional integrals, and since
T

(T u)

1 ,we have

u f (u ) du ,

(3.4)

Since

(3.3)

holds,

this

will

follow

from

Minkowskis

inequality

if

we

prove

that

p 1

, ( x) , ( x) dx
dx

(3.5)

Now , it follows at once from the definition that, for x T ,

, ( x) , ( x)

( 1) 1
( 1) 1
1
(
x

y
)
y
f ( y) dy
( x y) 1 y f ( y) dy


( )( 1) x
( )( 1) x
0
0
T

If

2,

then

for

x T ,

we

have

( x y) 2 (T y) 2 ,

so

that

d
( 1) ( ) x T

(
x
)

(
x
)

( x y ) 2 y f ( y) dy
,
dx
( )( 1) x 0
=

Const.
x

by (3.4).

Proof of Lemma 3.2: We use notations as in Lemma 3.1, and write further

U k , , ( y) for the expression

corresponding to

U k , , ( y) but with f (x) replaced by f (x) .

DOI: 10.9790/5728-11264350

www.iosrjournals.org

46 | Page

A Tauberian Theorem For (, , )- Convergence Of Cesro Means Of Order K Of Functions

We know that for any fixed

y 0, k 0, 1, 0

convergence of

x k 1
U k , , ( y ) ky
, ( x) dx , is equivalent to the convergence of
( x y ) k 1
0
x

, ( x)
x2

dx

.Then the conclusion will follow from Minkowskis inequality, if we show that

p 1

d
U k , , ( y) U k , , ( y) dy ,
dy

(3.6)

where we take (3.6) as including the assertion that the integral defined by U k , , ( y) U k , , ( y)
converges for all y 0 . For large y ,we have

, ( y) , ( y)

( 1) 1
( )( 1) y

( y x)

x f ( x) dx

(3.7)

x k 1
ky
( x) , ( x) , ( x)dx , follows at once by a
k 1 ,
(
x

y
)
0
x

Hence the convergence of

result due to ( [2] ) . Now (3.6) is equivalent to

x k 1
p 1
1 y dy c0 ( x y)k 2 ( x ky) , ( x) , ( x)dx .

(3.8)

Let T0 be any sufficiently large constant. Then (3.8) will follow from Minkowskis inequality, if we show

that

x
( x ky) , ( x) , ( x)dx . (3.9)
k 2
(
x

y
)
0

dy c

p 1

k 1

x
( x ky) , ( x) , ( x)dx .
k 2
(
x

y
)
T0

p 1
y dy c
1

k 1

T0

DOI: 10.9790/5728-11264350

www.iosrjournals.org

(3.10)

47 | Page

A Tauberian Theorem For (, , )- Convergence Of Cesro Means Of Order K Of Functions

By (3.9), we have

T0

p 1
y dy c

k 1

x
( x ky) , ( x) , ( x)dx
( x y)k 2

= O(1) y

kp p
1

O(1). Hence (3.9) follows .

By (3.7), the expression on the left of (3.10) does not exceed a constant. Thus

T0

p 1
y dy c

k 1

x
( x ky) , ( x) , ( x)dx
( x y)k 2

T0

o(1) y p1dy

( x y ) 2 x 1 dx

(3.11)

By an obvious change of variables the expression (3.11) is equal to

o(1) y p 1dy

t 2 (t y ) 1 dt

o(1) C C . The result follows.

Proof of Theorem 3.2 : We divide the proof into the following cases .
Case I .

,Case II .

,Case III .

Here we observe that Case I and II follow from case III, with the aid of Theorem 3.1 .
For, if

, Choose any ' , summability C , ,

implies summability

C , ' , by Theorem 3.1,

p

and it follows from Case III, that this implies ( D, k )(C, , ) p . Hence it is sufficient to consider the case III
only.
Proof of Case III : Since f ( x) s(C, , ) implies that f ( x) s(C, , ) for
'

loss of generality in considering the Case

' o , there is no

k , k is a positive integer.

d
x k 1
U k , , ( y) C
( x ky) , ( x) dx
We have ,
dy
( x y) k 2
T0
DOI: 10.9790/5728-11264350

www.iosrjournals.org

(3.12)

48 | Page

A Tauberian Theorem For (, , )- Convergence Of Cesro Means Of Order K Of Functions

Now, by definition

p , ( x )

( 1)
1
p
( p )( 1) y

t2

It is clear that, whenever

, ( x)

p 1

, (t )dt.

, (t )

We also write R , ( x)

(x t)

x
1, ( x) ( 1 1) t , (t )dt.

,we see that

Putting p=1 and

x2

(3.13)

dt.

dx converges, R , ( x) is defined for x >0, and that R , ( x) 0 as

x . It follows immediately from (3.13) that

( 1)
t t 2 dR , (t )dt
1, ( x)
1

x
0
x

o( x1 ) and hence that, for p 1 ,

1, ( x) o( x1 )

(3.14)

Integrating (3.14) by parts k times,we deduce with the help of (3.13) that

dk
d
U k , , ( y ) (1) k C x k k , ( x) k
dy
0
dx

x k 1

( x ky) dx.

k 2
( x y)

x k 1
.
It is verified that expression in (3.16) is o
k 1
(
x

y
)

Let R( x, y)

k
t
0

In fact, for fixed

(3.15)

(3.16)

d k t k 1
(t ky) dt.
k
k 2
dx (t y )

k 0 , we have uniformly in x 0, y 0,

xk
.
R( x, y) 0
k 1
( x y)
This may be proved by induction on k , if
DOI: 10.9790/5728-11264350

(3.17)

k 0 ,we have
www.iosrjournals.org

49 | Page

A Tauberian Theorem For (, , )- Convergence Of Cesro Means Of Order K Of Functions

x
t k 1

xk
(
t

ky
)
dt
=
,
R( x, y) t

k 2
k 1
(
t

y
)

hence the result is evident. Suppose that k 1, and assume the result true for k 1. Integrating by parts ,we
have

R( x, y) x k

x
k 1

t k 1

d k 1 x k 1
k 1

t kydt.
x

ky

k
t

k 1
k 2
k 1
k 2

dx x y
t t y

the first term is of required order by (3.17) (with k replaced by k-1), and the second by induction hypothesis.
Now integrating (3.16) by parts, we have

d
U k , , ( y ) =
dy

R
(
x
,
y
)
k , ( x) dx =
0
dx

R
(
x
,
y
)
, ( x) dx .
0
dx

Since the integrated term tends to 0 as , ( x) is bounded and R( x, y) 0 as x .

Using (3.17) and putting

x t y , we see that the expression in curly brackets

x k 1
C
t k 1
C
dx

dt
,
k 1
k 1

y
y
(
x

y
)
(
1

t
)
0
0
x

Again using (3.18) , the inner integral

C x

( x y)

k 1

dy ,

(3.18)

on putting

y x t , the expression on the right of (3.19) is equal to

(1 t )

k 1

dt C

(Since the integral converges) . Hence the result follows.

References
[1].
[2].
[3].
[4].
[5].

A.V. boyd, Some theorems on Summability, 1950.

J. M .Hyslop, A Tauberian theorem for absolute summability,J.London Math.Soc.Vol.12 (1937) pp.176-180.
F.W. Gehring, A study of , I, Trans. Amer. Math.Soc.vol.76 (1954) pp.420-443.
O Szasz, On products of summability methods,Proc.Amer.Math.Soc.vol.3(1952) pp. 257-263.
G. H. Hardy, J. E.Littlewood,and Polya, Inequalities,1934.

[6].

Mishra, B. P. and Srivastava A.P. Some remarks on absolute Summability of functions based on

(C, , ) Summability

methods .To appear in Jour. Nat. Acad. of Math.

DOI: 10.9790/5728-11264350

www.iosrjournals.org

50 | Page