HENSTOCK- KURZWEIL INTEGRAL ON [a,b]

Siti Nurul Afiyah

Dosen STMIK STIE Asia Malang
e-mail: noeroel_afy@yahoo.com

ABSTRACT
The theory of the Riemann integral was not fully satisfactory. Many important functions do not have a
Riemann integral. So, Henstock and Kurzweil make the new theory of integral. From the background, the
writer will be research about Henstock-Kurzweil integral and also theorems of Henstock- Kurzweil
Integral. Henstock- Kurzweil Integral is generalized from Riemann integral. In this case the writer uses
research methods literature or literature study carried out by way explore, observe, examine and identify
the existing knowledge in the literature. In this thesis explain about partition which used in Henstock-
Kurzweil Integral, definition and some property of Henstock- Kurzweil Integral. And some properties of
Henstock- Kurzweil integral as follows: value of the Henstock- Kurzweil integral is unique, linearity of the
Henstock-Kurzweil integral, Additivity of the Henstock-Kurzweil integral, Cauchy criteria, nonnegativity of
Henstock-Kurzweil integral and primitive function.

Keywords: Riemann Integral,
  partition, Henstock-Kurzweil Integral.

INTRODUCTION viewpoint of Riemann led others to invent other

We have already mentioned the
developments, during the 1630’s, by Fermat and
Descrates leading to analytic geometry and the
theory of the derivatives. However, the subject
we know as calculus did not begin to take shape
until the late 1660’s when Issac Newton (1642-
1727) created his theory of fluxions and invented
the method of inverse tangents to find areas
under curves. The reversal of the process for
finding tangent lines to find areas was also
discovered in the 1680’s by Leibniz (1646-1716),
who was unaware of Newton unpublished work
and who arrived at the discovery by a very
different route. Leibniz introduced the
terminology calculus differential and calculus
integral, since finding tangents lines involved
differences and finding areas involved
summations. Thus they had discovered that
integration, being a process of summation, was
inverse to the operation of differentiation.
During a century and a half of development
and refinement of techniques, calculus consisted
of these paired operations and their applications,
primarily to physical problems. In the 1850s,
Bernhard Riemann (1826-1866) adopted a new
and different viewpoint. He separated the
concept of integration from its companion,
differentiation, and examined the motivating
summation and limit process of finding areas by
itself. He broadened the scope by considering all
functions on an interval for which this process of
integration could be defined: the class of
integrable functions. The fundamental Theorem
of calculus became a result that held only for a
restricted set of integrable functions. The
integration theories, the most significant being
Lebesgue’s theory of integration.
The theory of the Riemann integral was not
fully satisfactory. Many important functions do
not have a Riemann integral even after we extend
the class of integrable functions slightly by
allowing "improper" Riemann integrals. For
example Characteristic function.
In 1957, the Czech mathematician Jaroslav
Kurzweil discovered a new definition of this
integral elegantly similar in nature to Riemann's
original definition which he named the gauge
integral; the theory was developed by Ralph
Henstock. Due to these two important
mathematicians, it is now commonly known as
Henstock-Kurzweil integral. The simplicity of
Kurzweil's definition made some educators
advocate that this integral should replace the
Riemann integral in introductory calculus
courses, but this idea has not gained traction.
Concerning the background of the study,
the writer formulates the statement of the
problems as follows:
1. How does the concept δ-fine partition of
Henstock-Kurzweil Integral?
2. How does the definition of Henstock-
Kurzweil Integral?
3. How does the fundamental properties of
Henstock-Kurzweil Integral?
REVIEW OF THE RELATED LITERATURE
1. Supremum and Infimum
We start with a straightforward definition
similar to many others in this course. Read the
 Henstock-Kurzweil Integral on [a,b]

definitions carefully, and note the use of ≤ and ≥ 3. Compact Sets

here rather than < and > .
Definition 4
Definition 1
A subset K of R is said to be compact if every
Let S be a subset of ℜ open cover of K has a finite sub cover.
1. A number u ∈ ℜ is said to be an upper In other words, a set K is compact if, whenever
bound of S if s ≤ u for all s ∈ S it is contained in the union of a collection G
2. A number w ∈ ℜ is said to be a lower = {Gα } of open sets in R , then it is contained
bound of S if w ≤ s for all s ∈ S in the union of some finite number of sets in G.
Definition 2 Theorem 1
Let S be a subset of ℜ . If K is a compact subset of R, then K is closed
1. If S is bounded above, then an upper bound and bounded.
u is said to be supremum (or a least upper Proof:
bound) of S if no number smaller than u is We shall first show that K is bounded. For
an upper bound of S. each   , let
 : , . Since each
2. If S is bounded below, then a lower bound
 is open and since    
 , we
w is said to be infimum (or a greatest lower see that the collection 
 :    is an open
bound) of S if no number greater than w is cover of K. since K is compact, this collection
a lower bound of S. has a finite sub cover , so there exists  
such that,
2. Limit of Function 
The essence of the concept of limit for real   

 , 
valued functions of a real variable is this: if L is a 
real number, then lim f ( x ) = L means that the Therefore K is bounded, since it is contained
in the bounded interval , .
x → x0
value f(x) can be made as close to L as we wish by We show that K is bounded, by showing that
taking x sufficiently close to x0. This is made its complement   is open. To do so,
precise in the following definition. let   be arbitrary and for each   ,
we let  ! "#  : |#  | % 1'(. Since
 ) , we have   ~  . Since K is
compact, there exists   such that
~

    

Now it follows from this that
 + ,  1' ,  - 1'. /,
so that ,  1' ,  - 1'.   . but since
u was an arbitrary point in , we infer that
 is open.
4. Continuity
Definition 5
Figure 1. The limit of f at x0 is L a) We say that f is continuous at x0 if f is
defined on an open interval (a, b )
Definition 3
We say that f(x) approaches the limit L as x containing x0 and lim f ( x) = f ( x0 ).
x→ x0
approaches x0 , and write b) We say that f is continuous from the left at
lim f ( x ) = L x0 if f is defined on an open interval
x → x0
(a, x0 ) and f ( x0 −) = f ( x0 ).
If f is defined on some deleted neighborhood of
x0 , and for every ε > 0 , there is a δ > 0 such
c) We say that f is continuous from the right
at x0 if f is defined on an open interval
that
f (x ) − L < ε (x0 , b) and f ( x0 +) = f ( x0 )
If
0 < x − x0 < δ

Jurnal CAUCHY – ISSN: 2086-0382 25

Siti Nurul Afiyah

Theorem 2 is a δ (ε ) > 0 such that if x, u ∈ A are any

f ( x ) − f (u ) < ε .
a) A function f is continuous at x0 if and only if f
number satisfying , then
is defined on an open interval (a, b ) Theorem 3
containing x0 and for each ε > 0 , there is a If f is continuous on a closed interval [a, b] ,
δ > 0 such that f ( x) − f ( x0 ) < ε , then f is uniformly continuous on a, b . [ ]
whenever x − x0 < δ Proof:
Suppose that ε > 0 . Since f is continuous on
b) A function f is continuous from the right at
x0 if and only if f is defined on an open
[a, b] , for each t in [a, b] there is a positive

interval [x0 , b) and for each ε > 0 , there is a number δt such that
ε
δ > 0 such that f ( x) − f ( x0 ) < ε holds f ( x ) − f (t ) <
whenever x0 ≤ x ≤ x0 + δ 2
if
c) A function f is continuous from the left at x0 x − t < 2δ t
if and only if f is defined on an open interval and
(a, x0 ] and for each ε > 0 , there is a δ > 0 x ∈ [a, b]
such that f ( x) − f ( x0 ) < ε holds whenever If
x0 − δ ≤ x ≤ x0 I t = (t − δ t , t + δ t ) ,
the collection
H = {I t t ∈ [a, b]}
Definition 6
A function f is continuous on an open interval
( )
a, b if it is continuous at every point in a, b . ( ) Is an open covering of a, b . Since a, b is [ ] [ ]
if, in addition. compact, the Heine-Borel theorem implies that
f (b − ) = f (b ) or f (a + ) = f (a ) there are finitely many points t1 , t 2 ,..., t n an
Then f is continuous on (a, b] or [a, b) , [a, b] such that I t , I t ,..., I t cover [a, b] .
respectively, if f continuous on (a, b ) and Now define
1 2 n

f (b − ) = f (b ) or f (a + ) = f (a ) both hold, then f δ = min {δ t , δ t ,..., δ t }

is continuous on [a, b] .
1 2 n

We will show that if

x − x' < δ
Definition 7
[a, b] if
and
A function f is piecewise continuous on
x, x'∈ [a, b]
a) f ( x0 + ). exist for all x0[a, b) ;
in Then
b) f ( x0 − ). exist for all x0 in (a, b] ; f ( x ) − f ( x ') < ε
c) f ( x0 + ) = f ( x0 −) = f ( x0 ) for all but From the triangle inequality.
f ( x ) − f ( x') = ( f ( x ) − f (t r )) + ( f (t r ) − f ( x'))
finitely many points x0 in (a, b ) .
If c) fails to hold at some x0 in (a, b ) , f has a
≤ ( f ( x ) − f (t r )) + ( f (t r ) − f (x'))
x0 . Also, f has a jump
[a, b] , x
jump discontinuity at
Since I t1 , I t 2 ,..., I t n cover
f (a + ) ≠= f (a ) or at b if
must be in
discontinuity at a if
one of these intervals. Suppose that x ∈ I t r ;
f (b − ) ≠ f (b ) .
That is,
5. Uniform Continuity x − t r < δ tr
With
Definition 8
t = tr ,
Let A ⊆ R , let f : A → R ,we say that f is
ε
uniformly continuous on A if for each ε > 0 there f ( x ) − f (t r ) < .
2

26 Volume 2 No. 1 November 2011

 Henstock-Kurzweil Integral on [a,b]

≤ x'− x + x − t r < γ + δ tr ≤ 2δ tr . s(P ) = ∑ m j (x j − x j −1 )

Such that, n

x'−t r = ( x'− x ) + ( x − t r )
j =1

And the lower integral of f over [a, b] , denoted

Therefore with t = t r and x replaced by x’ implies by
that
f ( x )dx
b

f ( x ') − f (t r ) <
ε
.
∫ a
−−−
2 Is the supremum off all lower sums.
f ( x ) − f ( x ') < ε
This imply that .
7. Riemann Integral
Definition 9 (Lipschitz Functions)
Let A ⊆ R , let f : A → R . If there exist a Riemann integral, defined in 1854, was the
constant K > 0 such that first of the modern theories of integration and
f ( x ) − f (u ) ≤ K x − u
enjoys many of the desirable properties of an
integration theory. The groundwork for the
For all x, u ∈ A , then f is said to be a Lipschitz Riemann integral of a function f over the interval
Functions on A [ ]
a, b begins with dividing the interval into
smaller subintervals.
Theorem 4
If f : A → R is a Lipschitz Functions, then f is With infimum and suprimum taken include
uniform continuous on A. [ ]
all partitions P on a, b , if the upper integral and
Proof: lower integral same, then f can be said integrable
If the a Lipschitz conditions satisfied with
constant K, then given ε > 0 , we can take
[ ]
on a, b . And called Riemann function f on

δ := ε K . If x, u ∈ A satisfy x − u < δ , then

[a, b] and denoted by f ∈ a, b This same [ ]
value is called the Riemann integral function f on

f ( x ) − f (u ) < K ⋅
ε
=ε
[ ]
f ∈ a, b and written
K b

Therefore f is uniformly continuous on A. (R )∫ f ( x )dx

a
Definition 11
6. Upper And Lower Integral [ ]
Let a, b ⊂ ℜ . A partition of [a, b] is a finite set
Definition 10 of numbers P = {x0 , x1 ,..., xn } such that
x0 = a, xn = b xi −1 < xi for i = 1,2,..., n .
[a, b] and P{x0 , x1 ,..., xn } is
and
[xi−1 , xi ] , define its length
If f is bounded on
[a, b] , let
For each subinterval
ℓ([xi −1 , xi ]) = xi − xi −1 .
a partition of
f (x )
to be The mesh of the
M j = sup
x j −1 ≤ x ≤ x j partition is then the length of the largest
And subinterval, [x−1i , xi ] :
m j = inf f (x ) µ (P ) = max{xi − xi −1 : i = 1,2,..., n}
x j −1 ≤ x ≤ x j

The upper sum of f over P is Thus the point {x0 , x1 ,..., xn } form an increasing

S (P ) = ∑ M j (x j − x j −1 ) sequence of numbers in [a, b] that divides the

n

j =1 interval [a, b] into contiguous subintervals.

And the upper integral of f over [a, b] , denoted Let f : [a, b] → ℜ , P = {x0 , x1 ,..., x n } be a
partition of [a, b] , and t i ∈ [xi −1 , xi ] for each i.
by
−−−−

f ( x )dx
b
∫a
Riemann began by considering the approximating
(Riemann) sums
Is the infimum of all upper sums. The lower sum
( )
n
S f , P, {t i }i =1 = ∑ f (t i )( xi − xi −1 ),
of f over P is n

i =1

Jurnal CAUCHY – ISSN: 2086-0382 27

Siti Nurul Afiyah

Defined with respect to the partition P and the And this time choose the points
set of sampling points {t } n
i i =1 . Riemann ξ i = (xi + xi −1 ) , Clearly ξ i ∈ [xi −1 , xi ] For
1
considered the integral of f over [a, b] to be a 2
i = 1,2,⋅ ⋅ ⋅, n
S ( f , P, {t i }i =1 ),
n
“limit” of the sums in the
following sense.
Then,
Definition 12 n n
S( f , D) = ∑f (ξi )( xi − xi−1 ) = ∑ (xi + xi−1 )( xi − xi−1 )
1
A function f : [a, b] → ℜ is Riemann integrable
i=1 2
[a, b] if there is an
i=1
over A ∈ ℜ such that for all
ε > 0 there is a δ >0 so that if P is any =
1n 2 2
(
∑ xi − xi−1 )
partition of [a, b] with µ (P ) < δ and
2 i=1

t i ∈ [xi −1 , xi ] for all i then

1
(
= xn2 − x0n )
( )
2
S f , P, {t i }i =1 − A < ε
n

b b
1
(
= b2 − a2
2
)
We write A = ∫ f = ∫ f (t )dt or, if we set Then,

( )
n
S ( f , D ) = ∑ f (ξ i )( xi − xi −1 ) =
a a 1 2
b − a2
I = [a, b], ∫ f . i =1 2
I
This definition defines the integral as a limit of
sums as the mesh of the partition approaches 0. 2. Definition Of Henstock-Kurzweil Integral
Definition 13
DISCUSSION [ ]
A function f : a, b → ℜ is said to be
Henstock-Kurzweil integrable on [a, b] if there
1. Concept δ-fine Partition of Henstock-
Kurzweil Integral exists a real number A such that for every ε >0
[ ]
Let a, b be a compact interval in ℜ . Let D be a there exists δ : [a, b] → ℜ +
such that for every
finite collection of interval-point pairs δ-fine Henstock-Kurzweil partition
{([ui , vi ], ξ i )} {([ui , vi ])} are non- D = {([u i , vi ], ξ i )} of [a, b] , we have
n n n
i =1
, where i =1 i =1

overlapping subintervals of [a, b] . Let δ (ξ ) be a

n

∑ f (ξ i )(vi − u i ) − A < ε
positive function on [a, b] , i =1

i.e.δ (ξ ) : [a, b] → ℜ + . We say We denote the Henstock-KurzweilIntegral (also

write as HK-integral) A by (HK )∫ f ( x )dx .
b
D = {([u i , vi ], ξ i )}i =1 is δ-fine Henstock-
n
a

Kurzweil partition of [a, b] if Example 2

ξ i ∈ [ui , vi ] ⊂ B(ξ i , δ (ξ i )) = (ξ i − δ (ξ i ), ξ i + δ (ξ i )) Define f : [0,1] → ℜ the Dirichlet’s function (=
for all i = 1,2,3,..., n the characteristic function of the rational
Given an δ-fine Henstock-Kurzweil partition numbers in 0,1 ), by[ ]
D = {([u i , vi ], ξ i )}i =1 we write
n
1 if x ∈ Q
f (x ) = 
0 if x ∉ Q
n
S ( f , D ) = ∑ f (ξ i )(vi − u i )
i =1 Then f ( x ) is Henstock-Kurzweil integrable on
For integral sum over D , whenever 1
f : [a, b] → ℜ [0,1] . And ∫ f (x ) = 0
0
Example 1 To Prove this assertion, we will define an
Let f x ( )= x Consider a division appropriate gauge δ ε , First we enumerate these
a = x0 < x1 < ... < xn = b and {ξ1, ξ2 ,...,ξn } . rational numbers as r1 , r2 ,... . We define

28 Volume 2 No. 1 November 2011

 Henstock-Kurzweil Integral on [a,b]

δ (ri ) = ε 2 − i −1 for i = 1,2,... , and if x ∈ [0,1] is Proof: Let A and B denote respectively the

irrational we define δ (ξ ) = 1 ; clearly δ ε is a

[ ]
integrals of f and g on a, b . given ε > 0 , there
is a δ 1 (ξ ) > 0 such that for any δ1-fine division
gauge on [0,1] . If P is a δ-fine tagged partition, D = ([u, v ]; ξ ) we have
there can be at most two subintervals in P that ε
have the number ri as tag, and the length of each ∑ f (ξ )(v − u ) − A < 2
of those subintervals is ≤ ε 2 −i −1 . Hence the Similarly, there is a δ 2 (ξ ) > 0 such that for any δ2-
contribution to S ( f , P ) from subintervals with fine division D = ([u , v ]; ξ ) we have
ε
tag ri is ≤ ε 2 −i . Since the terms in S ( f , P) ∑ g (ξ )(v − u ) − B < 2
with tags at irrational points contribute 0, we Now put δ (ξ ) = min (δ 1 (ξ ), δ 2 (ξ )) . Note that any
readily see that
∞ δ-fine division is also δ1-fine and δ2-fine.
0 ≤ S ( f , P) < ∑ ε =ε Therefore for any δ-fine division D = ([u , v ]; ξ )
i =1
2i
we have
ε >0 is arbitrary, this shows that f ( x ) is
∑ ( f (ξ ) + g ( ξ ) ) ( v − u ) − ( A + B ) ≤
Since
Henstock-Kurzweil integrable on [0,1] . And
1 ∑ f (ξ )( v − u ) − A + ∑ g (ξ )( v − u ) − B < ε
∫ f (x) = 0 The proof is complete.
0

Example 3
3. Fundamental Properties Of Henstock- 1
Kurzweil Integral
valuate ∫ f +g where f (x ) = x 2 and
Theorem 5 (Unique Property) 0
if f is Henstock-Kurzweil integrable over a, b , [ ] g (x) = x .
then the value of the integral is unique. Solution:
Proof: By theorem 3.3.3,
Suppose that f is Henstock-Kurzweil integrable 1 1 1 1
[ ]
on a, b and both real number A and B satisfy ∫ f +g = ∫x
2
+ x dx = ∫ x dx + ∫ x dx 2

Definition 3.2.1. Fix ε > 0 choose δ A and δ B 0 0 0 0

1
corresponding to A and B, respectively, in the
∫x
2
Now, we evaluate value dx
ε
definition with ε '= . Let δ = min(δ A , δ B ) 0
2 Consider a division 0 = x0 < x1 < ... < xn = 1 and
and suppose for every δ-fine Henstock-Kurzweil
partition D = {([u i , vi ], ξ i )}i =1 of [a, b] , Then
n {ξ1, ξ2 ,...,ξn } . And this time choose the
intermediate points
 n   n 
A − B =  ∑ f (ξ i )(v i − u i ) − B  −  ∑ f (ξ i )(vi − u i ) − A 
1

 i =1   i =1  1
(  2
ξ i =  xi2 + xi xi −1 + xi2−1  )
n n 3  ,
≤ A − ∑ f (ξ i )(vi − u i ) + ∑ f (ξ )(v i i − ui ) − B
i =1 i =1 then
< ε '+ε ' = ε
1

Since ε was arbitrary, it follows that A = B . ( )

0 ≤ xi−1 = x 2
i−1
1
21
( 2 1
<  xi2 + xi xi−1 + xi2−1  < xi2 2 = xi ) ( )
3 
For i = 1,2,⋅ ⋅ ⋅, n ; that is ξ i ∈ ( xi −1 , xi ) for each i .
Thus, the value of the integral is unique.

Theorem 6 (Linearity of the Henstock- So

1
Kurzweil integral) 1
∫x dx =
2
If f and g are Henstock-Kurzweil integrable on 3
[ ]
a, b , then so are f + g and α f where α is
0
1
real. Furthermore,
b b b b b
And then , we evaluate value ∫ xdx
∫( f + g) = ∫ f + ∫ g ∫ (αf ) = α ∫ f
0

a a a
Consider a division 0 = x0 < x1 < ... < xn = 1 and
and a a
{ξ1 , ξ2 ,...,ξn } . And this time choose the points

Jurnal CAUCHY – ISSN: 2086-0382 29

Siti Nurul Afiyah

ξi =
1
(xi + xi −1 ) , Clearly ξ i ∈ [xi −1 , xi ] For where
Σ1 is the partial sum over [a, c ] and Σ 2
2
i = 1,2,⋅ ⋅ ⋅, n over
[c, b] we have
Now
n n
∑f (ξ)( v−u) −( A+ B) ≤
S( f , D) = ∑ f (ξi )(xi − xi−1 ) = ∑ (xi + xi−1 )(xi − xi−1 )
1
i=1 i=1 2
∑ f (ξ)( v−u) − A + ∑ f (ξ)( v−u) −B <ε
1 2

Hence f is Henstock-Kurzweil integrable to A+B

=
1 n 2 2
(
∑ xi − xi−1
2 i=1
) [ ]
on a, b .
Alternatively, Let χ [a , c ] denote the characteristic
1
(
= xn2 − x0n
2
) function of [a, c] and f 1 = fχ [a ,c ] . Similarly, let
1 f 2 = fχ [c ,b ] . Then it follows from Theorem 3.3.1
= ⋅12
2 that
b b c b

∫ f =∫ ( f + f 2 ) =∫ f + ∫ f
1 1
= ⋅1 = 1
2 2 a a a c
1
1
∫0 xdx = 2 Example 7

f ( x ) = x and Let 0 <

So, that, 1
1 1 1 1 Let < 1 .and, If f is
1 1 5 2
∫ f +g = ∫x + x dx = ∫ x2 dx +∫ x dx = + =
2

 1
Henstock-Kurzweil integrable on [a, c ] = 0, 
0 0 0 0
3 2 6
Theorem 7 (Additivity of the henstock-  2
1 
and on [c,b ] =  ,1 and
Kurzweil Integral)
Let a < c < b . If f is Henstock-Kurzweil 2 
integrable on [a, c] and on [c, b] and b c b
b c b
∫ f =∫ f +∫ f
∫
a
f =∫ f +∫ f
a c Solution:
a a c

Proof: Consider a division 0 = x0 < x1 < ... < x n = 1 and

Let A denote the integral of f on [a, c] and B that {ξ1 , ξ2 ,...,ξn } . And this time choose the points
of f on [c, b] . Given ε > 0 , there is a δ 1 (ξ ) > 0 ,
ξ i = (xi + xi − ) , Clearly ξ i ∈ [xi −1 , xi ] For
1
defined on [a, c ] , such that for any δ -fine
1
1 2
i = 1,2,⋅ ⋅ ⋅, n
division D = ([u, v]; ξ ) of [a, c ] we have
Now
ε
∑ f (ξ )(v − u ) − A < 2
n n
S( f , D) = ∑ f (ξi )( xi − xi−1 ) = ∑ ( xi + xi−1 )(xi − xi−1 )
1
i=1 2
Similarly, there is a δ 2 (ξ ) > 0 defined on [c, b]
i=1

such that for any δ2-fine division 0 1, 23; 5 of

1 n 2 2
= ∑ xi − xi−1
2 i=1
( )
[c, b] we have
ε
1
(
= xn2 − x02 )
∑ f (ξ )(v − u ) − B < 2 2

Define δ (ξ ) = min (δ 1 (ξ ), c − ξ ) when ξ ∈ [a, c ) ,

1
= b2 − a2
2
( )
min (δ 2 (ξ ), ξ − c ) when ξ ∈ (c, b] , and
min (δ 1 (c ), δ 2 (c )) when ξ − c . Note for any δ-fine With same procedure we get; on
division D of 16, 73, c is always a division point of [a, c] = 0, 1 
D. therefore for any δ-fine division 0 1, 23; 5  2

of 16, 73 with Σ over D, writing

Σ = Σ1 + Σ 2

30 Volume 2 No. 1 November 2011

 Henstock-Kurzweil Integral on [a,b]

n n
1
S ( f , D) = ∑ f (ξi )( xi − xi−1 ) = ∑ ( xi + xi−1 )( xi − xi−1 ) ∑ f (ξ )(v − u ) − ∑ f (ξ ')(v'−u ') < ε
i=1 i=1 2 Where the first sum is over D and the second over
D’.
∑( xi − xi−1)
1n 2 2
= Proof
2 i=1 ( )
⇒ we will prove that if A function is Henstock-
kurzweil integrable on [a, b] Then for every
=
2
( xn − x0 )
1 2 2
ε > 0 , there is aδ (ξ ) > 0 such that for any δ-
fine division D = ([u, v]; ξ ) and D' = ([u ' , v']; ξ ')
= ( c2 −a2 )
1
we have
2

So ε > 0 , there is a δ 1 (ξ ) > 0 , defined on 10, 3, ∑ f (ξ )(v − u ) − ∑ f (ξ ')(v'−u ') < ε
:
such that for any δ1-fine division 0 1, 23; 5 of A function is Henstock-kurzweil integrable on

10, 3 we have [ ]
a, b Then for every ε > 0 , there is a
:
1 ε δ (ξ ) > 0 such that for any δ-fine
( )
1 2 division
c − a2 − <
2 8 2 D = ([u, v]; ξ ) and D' = ([u ' , v']; ξ ')

And on 1;, 73 1 , 13, we get ∑f (ξ)(v−u) −∑f (ξ')(v'−u') = ∑f (ξ)(v−u) −A+A−∑f (ξ')(v'−u')
:
n n
S ( f , D ) = ∑ f (ξ i )(xi − xi −1 ) = ∑ (xi + xi −1 )(xi − xi −1 )
1
≤ ∑f (ξ)( v−u) −A+ A−∑f (ξ)( v−u)
i =1 i =1 2

=
1 n 2
∑ xi − xi2−1
2 i =1
( ) <ε
Analogous to the situation for real-valued
1
= x n2 − x02 ( ) sequences, the condition that
2 ∑ f (ξ )(v − u ) − ∑ f (ξ ')(v'−u ') < ε
1 2
= b − c2
2
( ) (⇐) We have already proved that the
integrability of f implies the Cauchy criterion. So,
δ 2 (ξ ) > 0 ,  assume the Cauchy criterion holds. We will prove
ε > 0, there is a defined on 1 , 13,
: that f is Henstock-kurzweil integrable .
such that for any δ1-fine division 0 1, 23; 5 of if for every ε > 0 , there is a δ ξ > 0 such that ( )

1 , 13 we have
: for any δ-fine division D = ([u, v]; ξ ) and
3 ε
1 2
2
(
b − c2 − <
8 2
) D' = ([u ' , v']; ξ ') we have
therefore for any δ-fine division 0 1, 23; 5 of ∑ f (ξ )(v − u ) − ∑ f (ξ ')(v'−u ') < ε
10,13 Then A function is Henstock-Kurzweil integrable
with Σ over D, writing
Σ = Σ1 + Σ 2 where Σ1 on [a,b].

is the partial sum over

[a, c] and Σ 2 over 1 , 13 For each k ∈ Ν , choose a δ k > 0 so that for any
we have
:
([ ] )
two division D = u, v ; ξ and D' = u ' , v' ; ξ ' , ([ ] )
and corresponding sampling points, we have
(
1 2
b − a2 ) − +  ≤
 1 3

∑ f (ξ )(v − u ) − ∑ f (ξ ')(v'−u ') < k

2  8 8 1

min{δ1 , δ 2 ,...,δ k },
1 3
∑ f (ξ )( v −u) − 8 + ∑ f (ξ )( v −u) − 8 <ε
1 2
Replacing
δk by we may

Hence f is Henstock-Kurzweil integrable to
:
on δ ≥ δ k +1 .
assume that k
10,13 Next for each k, fix a partition

Lemma 8 (Cauchy Criteria)

Dk = ([u k , vk ]; ξ k ) and set of sampling point
A function is Henstock-kurzweil integrable on {ξ } n
. Note for j > k Thus
[ ]
a, b if and only if for every ε > 0 , there is a
i i =1

∑ f (ξ )(v − u k ) − ∑ f (ξ j )(v j − u j ) <

1
δ (ξ ) > 0 such that for any δ-fine division k k
min{ j , k }
D = ([u, v]; ξ ) and D' = ([u ' , v']; ξ ') we have ,

Jurnal CAUCHY – ISSN: 2086-0382 31

Siti Nurul Afiyah

∞ b b

∑ f (ξ k )(v k − u k )
Which implies that sequence
k =1
∫ f ≤ ∫g
a a
is a Cauchy sequence in R, and hence converges. Proof:
Let A be a limit of this sequence. it follows from In view of theorem 3.3.9 we may assume that
the previous inequality that
() ()
f x ≤ g x for all x. Given ε > 0 , as in the
f (ξ )(v − u ) − A <
1
∑ k proof of theorem 3.2.1, there is a δ (ξ ) > 0 such
It remains to show that A satisfies Definition 3.2.1 that for any δ-fine division D = ([u, v]; ξ ) we
Fix ε > 0 and let division D = u, v ; ξ . Then ([ ] ) have

∑f (ξ)(v −u) − A
b

∑ f (ξ )(v − u ) − ∫ f <ε
a
n n n ,
= ∑ f (ξi )(vi −ui ) − ∑f (ξk )(vk −uk ) + ∑f (ξk )(vk −uk ) − A b

i=1 k=1 k=1 ∑ g (ξ )(v − u ) − ∫ g < ε

n n n a

≤ ∑ f (ξi )(vi −ui ) − ∑ f (ξk )(vk −uk ) + ∑ f (ξk )(vk −uk ) − A It follows that
i=1 k=1 k=1 b b

1 1 ∫ f − ε < ∑ f (ξ )(v − u) < ∑g(ξ )(v − u) < ∫ g + ε

< + =ε a a
k k Since ε is arbitrary, we have the required
inequality.
It now follows that f is Henstock-Kurzweil
integrable on a, b [ ] Theorem 11
If f is Henstock-Kurzweil integrable on [a, b] with
Theorem 9 the primitive F, then for every ε > 0 , there is a
If f is Henstock-Kurzweil integrable on [a, b] , δ (ξ ) > 0 such that for any δ-fine division
then so it is on a subinterval [c, d ] of [a, b] . D = ([u, v]; ξ ) we have
∑ F (v ) − F (u ) − f (ξ )(v − u ) < ε
Proof :
Since f is Henstock-Kurzweil integrable on [a, b] ,
We shall make a few remarks. Before proof, from
the Cauchy condition holds. Take any two δ-fine
[ ]
the computational point of view, we may regard
divisions of c, d . say D1 and D2, and denote by
( )( )
f ξ v−u as an approximation of
F (v ) − F (u ) .
s1 and s2 respectively the Riemann sums of f over
D1 and D2. similarly, take another δ-fine division Then the difference
[ ] [ ]
D3 of a, c ∪ d , b and denote by s3 the F (v ) − F (u ) − f (ξ )(v − u )
is an error. The
corresponding Riemann sums. Then the union definition of the Henstock-Kurzweil integral says
D1 ∪ D3 forms a δ-fine division of [a, b] . Here that the absolute error is also small, whereas
Henstock’Lemma. In fact, the two are equivalent
the division points and associated points of by theorem 3.3.8. Another way of putting it is that
D1 ∪ D3 are the union of those from D1 and D3. taking any partial sum Σ1 of Σ we still have
The Riemann sum of f over D1 ∪ D3 is s1 + s3 . Σ1 F ( v ) − F ( u ) − f (ξ )( v − u ) < ε
And similarly that over D2 ∪ D3 is s 2 + s3 That is to say, the selected error is again small,
.Therefore by the Cauchy condition we have and indeed it is equivalent to the above two.
s1 − s 2 ≤ (s1 + s3 ) − (s 2 + s3 ) < ε
Proof:
( )
Given ε > 0 , there is a δ ξ > 0 such that for
any δ-fine division D = ([u, v]; ξ ) we have
Hence the result follows from lemma 3.3.7 with
[ ]
a, b replaced by c, d [ ]
ΣF (v ) − F (u ) − f (ξ )(v − u ) < ε
Theorem 10 (Nonnegativity of The Henstock- 4
Kurzweil integral) Let Σ1 be a partial sum of Σ and E1 the union of
[u, v] from Σ1 . Suppose E2. thus we can choose
If f and g are Henstock-Kurzweil integrable on
[ ] () ()
a, b and if f x ≤ g x for almost all in x in
a δ-fine division D2 = ([u, v ]; ξ ) of E2 such that
[a, b] , then

32 Volume 2 No. 1 November 2011

 Henstock-Kurzweil Integral on [a,b]

Σ 2 F (v ) − F (u ) − f (ξ )(v − u ) < ε CONCLUSION

4
Where Σ 2 is over D2 . Now writing From the discussion we get conclusion that:
1. δ-fine Henstock-Kurzweil partition
Σ 3 = Σ1 + Σ 2 we have D = {([u i , vi ], ξ i )}i =1
n
we write
Σ1( F( v) −F( u) − f (ξ)( v−u) ) ≤ n
S ( f , D ) = ∑ f (ξ i )(vi − u i )
i =1

Σ3( F( v) −F( u) − f (ξ)( v−u) ) +Σ2( F( v) −F(u) − f (ξ)( v−u) ) <ε where D be a finite collection of interval-point
2
Consequently the result follows. pairs {([ui , vi ], ξ i )}in=1 ,
{([ui , vi ])}in=1
where
Example 8 are non-overlapping subintervals of [a, b] .
Let f ( x ) = 1 for 0 < x ≤ 1 . Given ε > 0 , Let δ (ξ ) be a positive function on [a, b] ,
i.e.δ (ξ ) : [a, b] → ℜ + .
x
we shall construct δ (ξ ) so that f is Henstock-
[0,1] . Consider a division
And if
Kurzweil integrable on ξi ∈ [ui , vi ] ⊂ B(ξi , δ (ξi )) = (ξi − δ (ξi ), ξi + δ (ξi ))
0 = x0 < x1 < ... < xn = 1 {ξ , ξ ,...,ξn }
and 1 2 for all i = 1,2,3,..., n .
With ξ1 = 0 xi −1 ≤ ξi ≤ xi for i = 2,..., n .
and
2. A function f : [a, b] → ℜ is said to be
Note that the primitive of 1
x
is 2 x . Then Henstock-Kurzweil integrable on [a, b] if
we can write there exists a real number S f such that for
ε >0there exists δ : [a, b] → ℜ + such
n
2 − ∑ f (ξi )( xi − xi −1 ) ≤ every
i =1
that for every δ-fine Henstock-Kurzweil
partition D = {([u i , vi ], ξ i )}i =1 of a, b , we
n
[ ]
( )
n
x − ∑( xi − xi −1 )
1
2 − 2 − x1 + ∫ dx ξi ≤
x1
i =2 have
S( f , D)− S f < ε.

( )(x − x
n
2 x1 + ∑ 1 xi −1 −1 xi i i −1 ). 3. And the fundamental properties of Henstock-
i =2 Kurzweil integral as follows: value of the
We shall prove that above is less than ε for Henstock- Kurzweil integral is unique,
suitable δ − fine divisions. Suppose δ (ξ ) − cξ for linearity of the Henstock-Kurzweil integral,
0 < ξ ≤ 1 and 0 < c < 1 2 so that ξ1 = 0 always.
Additivity of the Henstock-Kurzweil integral,
Cauchy criteria, nonnegativity of Henstock-
If the above division is δ − fine and [u, v] is a Kurzweil integral, and primitive function.
typical interval [xi −1 , xi ] in the division with
u ≠ 0 and u ≤ ξ ≤ v , then
BIBLIOGRAPHY
0 < v − u < 2δ (ξ ) ≤ 2cv ,
re-arranging we get v u ≤ 1 (1 − 2c ) , and finally [1] Ding and G.J. Ye., (2009). Generalized
Gronwall-Bellman Inequalities Using the
(v − u ) uv < 2cv u ≤ 2c (1 − 2c ). Henstock-Kurzweil Integral, Southeast Asian
Now choose c so that 0 < c <1 2 and Bulletin of Mathematic(2009)33: 103-713.
2c (1 − 2c ) ≤ ε 2 . In addition, put δ (0) ≤ ε 16 . 2
[2] Douglas S Kurtz, Charles W Swartz. (2004).
Then for the given δ − fine division the above Theories Of Integration; The integrals of
inequality is less than Riemann, Lebesgue, Henstock-Kurzweil and
Mc. Shane. London: World Scientific
2 δ (0 ) +
2c n
∑
1 − 2c i = 2
( x i − x i −1 < ε) Publishing Co.Pc.Ltd.
[3] G. Bartle, Robert. And Donald R. Sherbert.
For example, when 0 < ε ≤ 1 we may choose (1994). Introduction To Real Analysis, 2nd
ε Edition. New York: Wiley.
c= . Hence the function is Henstock-Kurzweil
6 [4] Gordon, Russel A. (1955). The integrals of
integrable on [0,1] . Lebesgue, Denjoy, Perron, and Henstock,
Lyrary of Congress Cataloging in Publication
Data. American Mathematical Society.

Jurnal CAUCHY – ISSN: 2086-0382 33

Siti Nurul Afiyah

[5] Hutahean, effendi. (1989). Fungsi Riil. ITB: [9] Slavik, Antonin and Stefan Schwabik, Praha.
Bandung (2006). Henstock-Kurzweil and McShane
product integration; Descriptive definitions.
[6] Lee, P.-Y. (1989). Lanzhou Lectures on
October 23, 2006
Henstock . Singapore: World Scientific.
[10] Rahman, Hairur. (2008). Pengantar
[7] Mark Bridger, Real Analysis, A Constructive
Analisis Real. Malang: UIN Malang.
Approach, Northeastern University.
Department of Mathematics Boston, M.A, A
John Willey and Sons, Inc., Publication
[8] Przynski, William and Philip. (1987).
Introduction to Mathematical Analysis.
mcGraww Hill International.

34 Volume 2 No. 1 November 2011