Chapter 6 – Riemann-Stieltjes Integral

Course Title: Real Analysis 1 Course Code: MTH321

Course instructor: Dr. Atiq ur Rehman Class: MSc-II

Partition
Let [a, b] be a given interval. A finite set P = {a = x0 , x1 , x2 ,…, xk ,…, xn = b} is said
to be a partition of [a, b] which divides it into n such intervals
[ x0 , x 1 ] , [ x1, x 2 ] , [ x2 , x 3 ] ,..., [ xn−1, x n ] .
Each sub-interval is called a component of the partition.
Obviously, corresponding to different choices of the points xi we shall have
different partition.
The maximum of the length of the components is defined as the norm of the
partition.

Riemann Integral
Let f be a real-valued function defined and bounded on [a, b] . Corresponding to
each partition P of [a, b] , we put
M i = sup f ( x) ( xi −1 ≤ x ≤ xi )
mi = inf f ( x) ( xi −1 ≤ x ≤ xi )
We define upper and lower sums as Mi
n
U ( P, f ) = ∑ M i ∆xi mi
i =1 xi – 1 xi
n
and L ( P, f ) = ∑ mi ∆xi .
i =1
where ∆xi = xi − xi −1 (i = 1,2,…, n)
b
and finally ∫ f dx = inf U ( P, f ) ………….… (i)
a
b

∫ f dx = sup L( P, f ) ……………..(ii)
a

Where the infimum and the supremum are taken over all partitions P of [a, b] .
b b
Then ∫ f dx
a
and ∫ f dx
a
are called the upper and lower Riemann Integrals of f

over [a, b] respectively.

In case the upper and lower integrals are equal, we say that f is Riemann-
Integrable on [a, b] and we write f ∈ R , where R denotes the set of Riemann
integrable functions.
b b
The common value of (i ) and (ii ) is denoted by ∫ f dx
a
or by ∫ f ( x) dx .
a
Which is known as the Riemann integral of f over [a, b] .
2 Riemann-Stieltjes Integral

Theorem
The upper and lower integrals are defined for every bounded function f .
Proof
Take M and m to be the upper and lower bounds of f ( x) in [a, b] .
⇒ m ≤ f ( x) ≤ M ( a ≤ x ≤ b)
Then M i ≤ M and mi ≥ m (i = 1, 2,....., n)
Where M i and mi denote the supremum and infimum of f ( x) in ( xi −1 , xi ) for certain
partition P of [ a, b ] .
n n
⇒ L ( P, f ) = ∑ mi ∆xi ≥ ∑ m ∆xi (∆xi = xi −1 − xi )
i =1 i =1
n
⇒ L ( P, f ) ≥ m∑ ∆xi
i =1
n
But ∑ ∆x
i =1
i = ( x1 − x0 ) + ( x2 − x1 ) + ( x3 − x2 ) + .... + ( xn − xn−1 )

= xn − x0 = b − a
⇒ L ( P, f ) ≥ m(b − a )
Similarity U ( P, f ) ≤ M (b − a )
⇒ m(b − a ) ≤ L ( P, f ) ≤ U ( P, f ) ≤ M (b − a )
Which shows that the numbers L ( P, f ) and U ( P, f ) form a bounded set.
⇒ The upper and lower integrals are defined for every bounded function f . 

Riemann-Stieltjes Integral
It is a generalization of the Riemann Integral. Let α ( x) be a monotonically
increasing function on [a, b] . α (a) and α (b) being finite, it follows that α ( x) is
bounded on [a, b] . Corresponding to each partition P of [a, b] , we write
∆α i = α ( xi ) − α ( xi−1 )
( Difference of values of α at xi & xi −1 )
∵ α ( x) is monotonically increasing.
∴ ∆α i ≥ 0
Let f be a real function which is bounded on [a, b] .
n
Put U ( P, f ,α ) = ∑ M i ∆α i
i =1
n
L ( P, f ,α ) = ∑ mi ∆α i ,
i =1
where M i and mi have their usual meanings.
Define
b

∫ f dα = inf U ( P, f ,α ) ……………. (i)

a
 Riemann-Stieltjes Integral 3
b

∫ f dα = sup L ( P, f ,α ) ……………. (ii)

a

Where the infimum and supremum are taken over all partitions of [a, b] .
b b b b
If ∫ f dα = ∫ f dα ,
a a
we denote their common value by ∫ f dα
a
or ∫ f ( x) dα ( x) .
a

This is the Riemann-Stieltjes integral or simply the Stieltjes Integral of f w.r.t. α

over [a, b] .
b
If ∫ a
f dα exists, we say that f is integrable w.r.t. α , in the Riemann sense, and
write f ∈ R (α ) .

Note
The Riemann-integral is a special case of the Riemann-Stieltjes integral when we
take α ( x) = x .
∵ The integral depends upon f ,α , a and b but not on the variable of integration.
b b
∴ We can omit the variable and prefer to write ∫ f dα
a
instead of ∫ f ( x) dα ( x) .
a
In the following discussion f will be assume to be real and bounded, and α
monotonically increasing on [a, b] .

Refinement of a Partition
Let P and P∗ be two partitions of an interval [ a, b ] such that P ⊂ P∗ i.e. every
point of P is a point of P∗ , then P∗ is said to be a refinement of P .

Common Refinement
Let P1 and P2 be two partitions of [a, b] . Then a partition P∗ is said to be their
common refinement if P∗ = P1 ∪ P2 .

Theorem
If P∗ is a refinement of P , then
( )
L ( P, f ,α ) ≤ L P∗ , f ,α ……..………… (i)
and U ( P, f ,α ) ≥ U ( P , f ,α ) ………………. (ii)
∗

Proof
Let us suppose that P* contains just one point x∗ more than P such that
xi −1 < x∗ < xi where xi −1 and xi are two consecutive points of P .
Put
w1 = inf f ( x) (
xi −1 ≤ x ≤ x∗ )
xi – 1 x* xi
w2 = inf f ( x) (x ∗
≤ x ≤ xi )
4 Riemann-Stieltjes Integral

It is clear that w1 ≥ mi & w2 ≥ mi where mi = inf f ( x) , ( xi −1 ≤ x ≤ xi ) .

Hence
( )
L P ∗ , f ,α − L ( P, f ,α ) = w1 α ( x∗ ) − α ( xi −1 )  + w2 α ( xi ) − α ( x∗ ) 
−mi [α ( xi ) − α ( xi −1 )]
= w1 α ( x∗ ) − α ( xi −1 )  + w2 α ( xi ) − α ( x∗ ) 
−mi α ( xi ) − α ( x∗ ) + α ( x∗ ) − α ( xi −1 ) 
= ( w1 − mi ) α ( x∗ ) − α ( xi −1 )  + ( w2 − mi ) α ( xi ) − α ( x* ) 
∵ α is a monotonically increasing function.
∴ α ( x∗ ) − α ( xi −1 ) ≥ 0 , α ( xi ) − α ( x∗ ) ≥ 0
⇒ L ( P∗ , f ,α ) − L ( P, f ,α ) ≥ 0
(
⇒ L ( P , f ,α ) ≤ L P ∗ , f ,α ) which is (i)
If P∗ contains k points more than P , we repeat this reasoning k times and arrive
at (i).
Now put
W1 = sup f ( x) ( xi −1 ≤ x ≤ x∗ )
and W2 = sup f ( x) ( x∗ ≤ x ≤ xi )
Clearly M i ≥ W1 & M i ≥ W2
Consider
( )
U ( P, f ,α ) − U P∗ , f ,α = M i [α ( xi ) − α ( xi −1 )]
−W1 α ( x∗ ) − α ( xi −1 )  − W2 α ( xi ) − α ( x∗ ) 
= M i α ( xi ) − α ( x∗ ) + α ( x∗ ) − α ( xi −1 ) 
−W1 α ( x∗ ) − α ( xi −1 )  − W2 α ( xi ) − α ( x∗ ) 
= ( M i − W1 ) α ( x∗ ) − α ( xi −1 )  + ( M i − W2 ) α ( xi ) − α ( x∗ )  ≥ 0
(∵ α is ↑ )
⇒ U ( P , f ,α ) ≥ U P ∗ , f ,α ( ) which is (ii)


Theorem
 Riemann-Stieltjes Integral 5

Let f be a real valued function defined on [a, b] and α be a monotonically

increasing function on [a, b] . Then
sup L ( P, f ,α ) ≤ inf U ( P, f ,α )
b b
i.e. ∫ f dα
a
≤ ∫ f dα
a

Proof
Let P* be the common refinement of two partitions P1 and P2 . Then
(
L ( P1 , f ,α ) ≤ L P ∗ , f ,α ) (
≤ U P ∗ , f ,α ) ≤ U ( P2 , f ,α )
Hence L ( P1 , f ,α ) ≤ U ( P2 , f ,α ) …………. (i)
If P2 is fixed and the supremum is taken over all P1 then (i) gives
b

∫ f dα ≤ U ( P2 , f ,α )
a

Now take the infimum over all P2

b b
⇒ ∫ f dα
a
≤ ∫ f dα
a


Theorem (Condition of Integrability or Cauchy’s Criterion for

Integrability.)
f ∈ R (α ) on [a, b] iff for every ε > 0 there exists a partition P such that
U ( P, f ,α ) − L ( P, f ,α ) < ε
Proof
Let U ( P, f ,α ) − L ( P, f ,α ) < ε …………. (i)
b b
Then L ( P, f ,α ) ≤ ∫ f dα ≤ ∫ f dα ≤ U ( P , f ,α )
a a
b b
⇒ ∫ f dα − L ( P, f ,α ) ≥ 0 and U ( P, f ,α ) − ∫ f dα ≥ 0
a a

Adding these two results, we have

b b

∫ f dα − ∫ f dα − L ( P, f ,α ) + U ( P, f ,α ) ≥ 0
a a
b b
⇒ ∫ f dα − ∫ f dα ≤ U ( P, f ,α ) − L ( P, f ,α ) < ε from (i)
a a
b b
i.e. 0≤ ∫ f dα − ∫ f dα
a a
<ε for every ε > 0 .
6 Riemann-Stieltjes Integral

b b
⇒ ∫ f dα
a
= ∫ f dα
a
i.e. f ∈ R (α )

Conversely, let f ∈ R (α ) and let ε > 0

b b b
⇒ ∫ f dα = ∫ f dα = ∫ f dα
a a a
b b
Now ∫ f dα = inf U ( P, f ,α )
a
and ∫ f dα = sup L ( P, f ,α )
a

There exist partitions P1 and P2 such that

b
ε
U ( P2 , f ,α ) − ∫ f dα < ……..…… (ii) U ( P2 , f , α ) − ε < ∫ f dα
a
2 2
ε ∫ f dα < L ( P , f , α ) + ε 2
b
and ∫ f dα − L ( P1 , f ,α ) < ……..…… (iii) 1

a
2
We choose P to be the common refinement of P1 and P2 .
Then
b
ε
U ( P, f ,α ) ≤ U ( P2 , f ,α ) < ∫ f dα + < L ( P1 , f ,α ) + ε ≤ L ( P, f ,α ) + ε
a
2
So that
U ( P, f ,α ) − L ( P, f ,α ) < ε 

Theorem
a) If U ( P, f ,α ) − L ( P, f ,α ) < ε holds for some P and some ε , then it holds
(with the same ε ) for every refinement of P .
b) If U ( P, f ,α ) − L ( P, f ,α ) < ε holds for P = { x0 ,…, xn } and si , ti are arbitrary
points in [ xi −1 , xi ] , then
n

∑
i =1
f ( si ) − f (ti ) ∆α i < ε

c) If f ∈ R (α ) and the hypotheses of (b) holds, then

n b

∑ f (t )∆α − ∫ f dα
i =1
i i <ε
a

Proof
a) Let P∗ be a refinement of P . Then
L ( P, f ,α ) ≤ L P∗ , f ,α ( )
and (
U P ∗ , f ,α ) ≤ U ( P, f ,α )
(
⇒ L ( P , f ,α ) + U P ∗ , f ,α ) ( )
≤ L P ∗ , f ,α + U ( P , f ,α )
( ) (
⇒ U P∗ , f ,α − L P∗ , f ,α ) ≤ U ( P, f ,α ) − L ( P, f ,α )
 Riemann-Stieltjes Integral 7

∵ U ( P, f ,α ) − L ( P, f ,α ) < ε
( ) (
∴ U P ∗ , f ,α − L P ∗ , f ,α < ε )
b) P = { x0 ,...., xn } and si , ti are arbitrary points in [ xi −1 , xi ] .
⇒ f ( si ) and f (ti ) both lie in [ mi , M i ] .
⇒ f ( si ) − f (ti ) ≤ M i − mi xi – 1 si ti xi
⇒ f ( si ) − f (ti ) ∆α i ≤ M i ∆α i − mi ∆α i
n n n
⇒ ∑
i =1
f ( si ) − f (ti ) ∆α i ≤ ∑M
i =1
i ∆α i − ∑ mi ∆α i
i =1
n
⇒ ∑
i =1
f ( si ) − f (ti ) ∆α i ≤ U ( P, f ,α ) − L ( P, f ,α )

∵ U ( P, f ,α ) − L ( P, f ,α ) < ε
n
∴ ∑
i =1
f ( si ) − f (ti ) ∆α i < ε

c) ∵ mi ≤ f (ti ) ≤ M i
∴ ∑ m ∆α i ∑ f (t ) ∆α ≤ ∑ M ∆α
i ≤ i i i i

⇒ L ( P, f ,α ) ≤ ∑ f (t ) ∆α ≤ U ( P, f ,α )
i i
b
and also L ( P, f ,α ) ≤ ∫a f dα ≤ U ( P , f ,α )

Using (b), we have

b

∑ f (ti ) ∆αi − ∫ f dα < ε 

a

Lemma
If M & m are the supremum and infimum of f and M ′ , m′ are the supremum
& infimum of f on [a, b] then M ′ − m′ ≤ M − m .
Proof
Let x1 , x2 ∈ [a, b] , then
f ( x1 ) − f ( x2 ) ≤ f ( x1 ) − f ( x2 ) ………………(A)
∵ M and m denote the supremum and infimum of f ( x) on [a, b]
∴ f ( x) ≤ M & f ( x ) ≥ m ∀ x ∈ [ a, b]
∵ x1 , x2 ∈ [a, b]
∴ f ( x1 ) ≤ M and f ( x2 ) ≥ m
⇒ f ( x1 ) ≤ M and − f ( x2 ) ≤ − m
⇒ f ( x1 ) − f ( x2 ) ≤ M − m ……………... (i)
Interchanging x1 & x2 , we get
8 Riemann-Stieltjes Integral

− [ f ( x1 ) − f ( x2 ) ] ≤ M − m ………….. (ii)
(i) & (ii) ⇒ f ( x1 ) − f ( x2 ) ≤ M − m
⇒ f ( x1 ) − f ( x2 ) ≤ M − m by eq. (A) …….…..(I)
∵ M ′ and m′ denote the supremum and infimum of f ( x) on [a, b]
∴ f ( x) ≤ M ′ and f ( x) ≥ m′ ∀ x ∈ [a, b]
⇒ ∃ ε > 0 such that
f ( x1 ) > M ′ − ε ………….. (iii)
and f ( x2 ) < m′ + ε ⇒ − f ( x2 ) + ε > − m′ ………….. (iv)
From (iii) and (iv), we get
f ( x1 ) − f ( x2 ) + ε > M ′ − m′ − ε
⇒ 2ε + f ( x1 ) − f ( x2 ) > M ′ − m′
∵ ε is arbitrary ∴ M ′ − m′ ≤ f ( x1 ) − f ( x2 ) …….…..…. (v)
Interchanging x1 & x2 , we get
M ′ − m′ ≤ − ( f ( x1 ) − f ( x2 ) ) …………… (vi)
Combining (v) and (vi), we get
M ′ − m′ ≤ f ( x1 ) − f ( x2 ) ……………. (II)
From (I) and (II), we have the require result
M ′ − m′ ≤ M − m 

Theorem
b b
If f ∈ R (α ) on [a, b] , then f ∈ R (α ) on [a, b] and ∫a f dα ≤ ∫a f dα .

Proof
∵ f ∈ R (α )
∴ given ε > 0 ∃ a partition P of [a, b] such that
U ( P, f ,α ) − L ( P, f ,α ) < ε
i.e. ∑M i ∆α i − ∑ mi ∆α i = ∑ ( M i − mi ) ∆α i < ε
Where M i and mi are supremum and infimum of f on [ xi −1 , xi ]
Now if M i′ and mi′ are supremum and infimum of f on [ xi −1 , xi ] then
M i′ − mi′ ≤ M i − mi
⇒ ∑ ( M i′ − mi′ ) ∆α i ≤ ∑ ( M i − mi ) ∆α i
⇒ U ( P, f ,α ) − L ( P, f ,α ) ≤ U ( P, f ,α ) − L ( P, f ,α ) < ε
⇒ f ∈ R (α ) .
Take c = +1 or −1 to make c ∫ f dα ≥ 0
b b
Then ∫a f dα = c ∫ f dα …….……… (i)
a
 Riemann-Stieltjes Integral 9

Also c f ( x) ≤ f ( x) ∀ x ∈ [ a, b]
b b b b
⇒ ∫a c f dα ≤ ∫a f dα ⇒ c ∫ f dα ≤
a
∫a f dα ………. (ii)

From (i) and (ii), we have

b b

∫a f dα ≤ ∫a f dα 

Theorem (Ist Fundamental Theorem of Calculus)

x
Let f ∈ R on [a, b] . For a ≤ x ≤ b , put F ( x) = ∫ f (t ) dt , then F is continuous
a
on [a, b] ; furthermore, if f is continuous at point x0 of [a, b] , then F is
differentiable at x0 , and F ′( x0 ) = f ( x0 ) .
Proof
∵ f ∈R
∴ f is bounded.
Let f (t ) ≤ M for t ∈ [a, b]
If a ≤ x < y ≤ b , then a x y b
y x
F ( y) − F ( x) = ∫ f (t ) dt − ∫ f (t ) dt
a a
x y x
= ∫ f (t ) dt + ∫ f (t ) dt − ∫ f (t ) dt
a x a
y y y
= ∫ f (t ) dt
x
≤ ∫x f (t ) dt ≤ M ∫ dt = M ( y − x)
x

⇒ F ( y ) − F ( x) < ε for ε > 0 provided M y − x < ε

ε
i.e. F ( y) − F ( x) < ε whenever y−x <
M
This proves the continuity (and, in fact, uniform continuity) of F on [a, b] .
Next, we have to prove that if f is continuous at x0 ∈ [a, b] then F is
differentiable at x0 and F ′( x0 ) = f ( x0 )
F (t ) − F ( x0 )
i.e. lim = f ( x0 )
t → x0 t − x0
Suppose f is continuous at x0 . Given ε > 0 , ∃ δ > 0 such that
f (t ) − f ( x0 ) < ε if t − x0 < δ where t ∈ [a, b]
⇒ f ( x0 ) − ε < f (t ) < f ( x0 ) + ε if x0 − δ < t < x0 + δ
t t t t
⇒ ∫x ( f ( x0 ) − ε ) dt < ∫x f (t ) dt < ∫x ( f ( x0 ) + ε ) dt a x0 – δ x0 x0+δ
0 0 0
10 Riemann-Stieltjes Integral

t t t
⇒ ( f ( x0 ) − ε ) ∫ dt < ∫x f (t ) dt < ( f ( x0 ) + ε ) ∫ dt
x0 0 x0

⇒ ( f ( x0 ) − ε ) (t − x0 )
< F (t ) − F ( x0 ) < ( f ( x0 ) + ε ) (t − x0 )
F (t ) − F ( x0 )
⇒ f ( x0 ) − ε < < f ( x0 ) + ε
t − x0
F (t ) − F ( x0 )
⇒ − f ( x0 ) < ε
t − x0
F (t ) − F ( x0 )
⇒ lim = f ( x0 )
t → x0 t − x0
⇒ F ′( x0 ) = f ( x0 ) 

Theorem (IInd Fundamental Theorem of Calculus)

If f ∈ R on [a, b] and if there is a differentiable function F on [a, b] such that
F ′ = f , then
b

∫a f ( x) dx = F (b) − F (a)
Proof
∵ f ∈ R on [a, b]
∴ given ε > 0 , ∃ a partition P of [a, b] such that
U ( P, f ) − L ( P, f ) < ε
∵ F is differentiable on [a, b]
∴ ∃ ti ∈ [ xi −1 , xi ] such that
F ( xi ) − F ( xi −1 ) = F ′(ti )∆xi
⇒ F ( xi ) − F ( xi −1 ) = f (ti ) ∆xi for i = 1,2,...., n ∵ F′ = f
n
⇒ ∑=
i 1
f (ti ) ∆xi = F (b) − F (a )
∵ if f ∈ R (α ) then

∑ f (t ) ∆α − ∫
b b
f dα < ε
⇒ F (b) − F (a ) − ∫ f ( x) dx < ε i i a

a
∵ ε is arbitrary
b
∴ ∫a f ( x) dx = F (b) − F (a) 

-------------------------------------------------------