Scribd
Upload
48 views

Riemann-Stieltjes Integrals with α a Step Function: t→s t>s t→s t<s

The document proves that the Riemann-Stieltjes integral of a continuous function f with respect to a step function α exists and can be written as the sum of the values of f at the discontinuities of α times the jumps in α. It does this by showing the Riemann sum over partitions finer than one related to the discontinuities of α converges to this sum.

Uploaded by

Christopher Hitchens
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
48 views

Riemann-Stieltjes Integrals with α a Step Function: t→s t>s t→s t<s

The document proves that the Riemann-Stieltjes integral of a continuous function f with respect to a step function α exists and can be written as the sum of the values of f at the discontinuities of α times the jumps in α. It does this by showing the Riemann sum over partitions finer than one related to the discontinuities of α converges to this sum.

Uploaded by

Christopher Hitchens
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

RiemannStieltjes Integrals with a Step Function

Theorem. Let : [a, b] IR be a step function with discontinuities at s1 < < sn ,


where a s1 and sn b. Let f : [a, b] IR be continuous at each sj , 1 j n. Then
f R() on [a, b] and
Z

f d =

n
X
j=1



f (sj ) (sj +) (sj )

where
(s+)
(t)
(s+ ) = lim
ts

(s)
(s)

(t)
(s ) = lim
ts
t<s

t>s

and, by convention,
if s1 = a

(s1 ) = (a)

if sn = b

(sn +) = (b)



Proof: Let > 0. Choose the partition P = a = x
0 < x
1 < < x
m = b so that
(1) {s1 , , sn } P


i x
i1 < with
(2) the norm or mesh of P = kP k = max x
1im


= min |s2 s1 |, , |sn sn1 |, 0 and 0 is given by
Insert () here.
Now let P = {a = x0 , x1 , , xp = b} P be any partitition finer than P and T =
{t1 , , tp } be any choice for P and consider each term in
S(P, T, f, ) =

p
X
i=1



f (ti ) (xi ) (xi1 )

For each 1 i p, either


(1) neither xi nor xi1 is in {s1 , , sn }, in which case both xi and xi1 lie in a
subinterval of [a, b] (either [a, s1 ), or (sj1 , sj ) for some 2 j n, or (sn , b]) on
which is required to be constant. In this case (xi ) (xi1 ) = 0.
or (2) there is a 1 j n with xi = sj . In this case (xi ) (xi1 ) = (sj ) (sj ).
or (3) there is a 1 j n with xi1 = sj . In this case (xi )(xi1 ) = (sj +)(sj ).
These three possibilities are illustrated below, with the points of P indicated by hash
marks.
(1)
(2)
(3)
(1)
(2)
(3)
(2)
(3)
(1)
a
c Joel Feldman.

s1

2008. All rights reserved.

s2
January 9, 2008

s3

Step Function s

So

S(P, T, f, ) =
=

n n
X



case (2) terms


+




case (3) terms


f (tij ) (sj ) (sj ) + f (t ) (sj +) (sj )

j=1

ij

o

(1)

Here tij lies in the subinterval of P whose right hand end point is sj and tij lies in the
subinterval of P whose left hand end point is sj . Because kP k < , we have sj < tij sj
and sj tij < sj + . We may write the value of the integral given in the statement of
the theorem in a form quite like that of S(P, T, f, ):
n
X
j=1

n n



o

 X
f (sj ) (sj ) (sj ) + f (sj ) (sj +) (sj )
f (sj ) (sj +) (sj ) =
j=1

(2)
(The two f (sj )(sj ) terms cancel.) Subtracting (2) from (1) and using the triangle inequality gives
n

X



f (sj ) (sj +) (sj )
S(P, T, f, )

j=1
n
Xn
j=1

o




f (ti ) f (sj ) (sj ) (sj ) + f (ti ) f (sj ) (sj +) (sj )
j
j

Now
() for each 1 j n, f is continuous at sj so that there is a j > 0 such that



k=1 |(sk ) (sk )| + |(sk +) (sk )|



f (t) f (sj ) < Pn

for all t with |t sj | < j . Choose 0 = min{1 , j }.


Consequently
n

X



f (sj ) (sj +) (sj ) <
S(P, T, f, )
j=1

as desired.

c Joel Feldman.

2008. All rights reserved.

January 9, 2008

Step Function s

You might also like