This action might not be possible to undo. Are you sure you want to continue?

Ambar N. Sengupta

November, 2008

2 Ambar N. Sengupta

Contents

Introductory Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . 5

1 Ordered Fields and The Real Number System 7

1.1 Ordered Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . 8

1.1.1 Fields . . . . . . . . . . . . . . . . . . . . . . . . . . . . 9

1.1.2 Order Relations . . . . . . . . . . . . . . . . . . . . . . . 13

1.1.3 Ordered Fields . . . . . . . . . . . . . . . . . . . . . . . 14

1.1.4 The absolute value function . . . . . . . . . . . . . . . . 16

1.1.5 The Archimedean Property . . . . . . . . . . . . . . . . . 18

1.2 The Real Number System R . . . . . . . . . . . . . . . . . . . . 18

1.2.1 Hilbert Maximality and the Completeness Property . . . . 19

1.2.2 Completeness of R and measurement . . . . . . . . . . . 20

Problem Set 1 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 20

2 The Extended Real Line and Its Topology 23

2.1 The extended real line . . . . . . . . . . . . . . . . . . . . . . . . 23

2.2 Neighborhoods . . . . . . . . . . . . . . . . . . . . . . . . . . . 24

2.3 Types of points for a set . . . . . . . . . . . . . . . . . . . . . . . 25

2.4 Interior, Exterior, and Boundary of a Set . . . . . . . . . . . . . . 27

2.5 Open Sets and Topology . . . . . . . . . . . . . . . . . . . . . . 28

2.6 Closed Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 30

2.7 Open Sets and Closed Sets . . . . . . . . . . . . . . . . . . . . . 30

2.8 Closed sets in R and in R

∗

. . . . . . . . . . . . . . . . . . . . . 31

2.9 Closure of a set . . . . . . . . . . . . . . . . . . . . . . . . . . . 31

2.10 The closure of a set is closed . . . . . . . . . . . . . . . . . . . . 32

2.11 S is the smallest closed set containing S . . . . . . . . . . . . . . 32

2.12 R

∗

is compact . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32

2.13 Compactness of closed subsets of R

∗

. . . . . . . . . . . . . . . . 35

2.14 The Heine-Borel Theorem: Compact subsets of R . . . . . . . . . 36

3

4 Ambar N. Sengupta

2.15 Sequences . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 37

2.16 Limits points of a sequence . . . . . . . . . . . . . . . . . . . . . 38

2.17 Bolzano-Weierstrass Theorem . . . . . . . . . . . . . . . . . . . 39

2.18 Limit points and suprema and inﬁma . . . . . . . . . . . . . . . . 39

2.19 Limit of a sequence . . . . . . . . . . . . . . . . . . . . . . . . . 41

2.20 Simple Examples of limits . . . . . . . . . . . . . . . . . . . . . 42

2.21 The sequence 1/n . . . . . . . . . . . . . . . . . . . . . . . . . . 43

2.22 The sequence R

n

. . . . . . . . . . . . . . . . . . . . . . . . . . 44

2.23 Monotone Sequences . . . . . . . . . . . . . . . . . . . . . . . . 48

2.24 The limit of a sequence is unique . . . . . . . . . . . . . . . . . . 49

2.25 Convergent sequences and Cauchy sequences . . . . . . . . . . . 49

2.26 Every Cauchy sequence is bounded . . . . . . . . . . . . . . . . . 50

2.27 Every Cauchy sequence is convergent . . . . . . . . . . . . . . . 51

2.28 The rationals are countable . . . . . . . . . . . . . . . . . . . . . 52

2.29 The real numbers are uncountable . . . . . . . . . . . . . . . . . 54

2.30 Connected Sets . . . . . . . . . . . . . . . . . . . . . . . . . . . 55

3 Integration 57

3.1 Approaching the Riemann Integral . . . . . . . . . . . . . . . . . 58

3.2 Riemann Sums . . . . . . . . . . . . . . . . . . . . . . . . . . . 59

3.3 Deﬁnition of the Riemann Integral . . . . . . . . . . . . . . . . . 61

3.4 Reﬁning Partitions . . . . . . . . . . . . . . . . . . . . . . . . . 62

3.5 The Darboux Criterion . . . . . . . . . . . . . . . . . . . . . . . 65

3.6 Integrable functions are bounded . . . . . . . . . . . . . . . . . . 66

3.7 Variation of a Function . . . . . . . . . . . . . . . . . . . . . . . 67

3.8 The algebra R [a, b] . . . . . . . . . . . . . . . . . . . . . . . . . 70

3.9 C[a, b] ⊂R [a, b] . . . . . . . . . . . . . . . . . . . . . . . . . . . 72

3.10 The Integral as a Non-negative Linear Functional . . . . . . . . . 73

3.11 Additivity of the Integral . . . . . . . . . . . . . . . . . . . . . . 79

3.12 Monotone Functions are Riemann Integrable . . . . . . . . . . . . 81

3.13 Riemann Sums and the Riemann Integral . . . . . . . . . . . . . 83

Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . 86

A Question Bank 89

Notes in Introductory Real Analysis 5

Introductory Remarks

These notes were written for an introductory real analysis class, Math 4031, at

LSU in the Fall of 2006. In addition to these notes, a set of notes by Professor L.

Richardson were used.

There are several different ideologies that would guide the presentation of

concepts and proofs in any course in real analysis:

(i) the historical way

(ii) the most natural way

(iii) the most efﬁcient way

(iv) a comprehensive way, explaining the insights from several different ap-

proaches

The reality of constraints of time makes (iii) the most convenient approach,

and perhaps the best example of this approach is Rudin’s Principles of Mathemat-

ical Analysis [5]. The downside is that there is little possibility of conveying any

insights or intuition.

In studying the notion of completeness, a choice has to be made whether to

treat the Cauchy sequence point of view or the existence of suprema as funda-

mental. I have chosen the latter; it conforms to the classical geometric notion of

a positive real number being speciﬁed by quantities greater than it and those less

than it. This point of view also guides the choice of approach in the treatment of

the Riemann integral; the Riemann integral of a function is the unique real number

lying between the upper Riemann sums and lower Riemann sums.

The notes here do not include a chapter on continuous functions, for which we

followed the Richardson notes.

These notes have not been proof read carefully. I will update them time to

time.

6 Ambar N. Sengupta

Chapter 1

Ordered Fields and The Real

Number System

In this chapter we go over the essential, foundational, facts about the real number

system.

Positive real numbers arose from geometry in Greek mathematics, as ratios of

magnitudes, such as segments or planar regions or even angles. In the discussion

below we focus on segments.

In Euclid’s Elements, a segment EF is taken to exceed a segment GH, symbol-

ically

EF>GH

if EF is congruent to a segment GK, where K is some point between E and F. An

important feature of this order relation is encapsulated in the archimedean axiom:

given any two segments, some multiple of any one of them exceeds the other.

Then aAa pair PQ and RS if for any positive numbers n and m, the segment

nAB (which is n copies of AB laid contiguously) exceeds the segment mCD if

and only if the segment nPQ exceeds the segment mRS. The ratio

AB

CD

is then essentially the equivalence class of all segment pairs which are in the same

ratio as AB is to CD. Euclid also deﬁnes the ratio XY/ZW to be greater than the

ratio PQ/RS :

XY

ZW

>

PQ

RS

if they are unequal and if whenever mZW>nXY then also mRS>nPQ.

7

8 Ambar N. Sengupta

A special case is that of commensurate segments: if a whole multiple of AB,

say nAB, is congruent to a whole multiple of CD, say mCD, then the ratio

AB

CD

is

rational, and is denoted by

m

n

.

It is readily checked that

m

n

=

p

q

if and only if

qm = pn.

Such ratios are the rational numbers. Other ratios are irrational. In either case,

Euclid’s considerations suggest that a ratio of segments may be understood in

terms of a set of rational numbers, for example the set of all those rationals which

exceed the given ratio.

The axioms of geometry, and the Euclidean construction procedures, show

that ratios of segments can be added and multiplied and, when 0 and negatives

are included, an algebraic structure called a ﬁeld emerges (this is discussed at

length by Hilbert [3]). The maximal such ﬁeld, respecting the axioms of geometry

pertaining to the order relation and congruence, constitutes the real number system

R.

Leaving aside the historical background, the real number system may be con-

structed by starting with the empty set, constructing the natural numbers, then the

rationals, and then the real numbers by Dedekind’s method of identifying a real

number with a splitting of the rationals into two disjoint classes with members of

one class exceeding those of the other.

Dedekind’s method has beed generalized in a striking, and vastly more power-

ful way, by Conway [1], who shows how the Dedekind cut method can be applied

to abstract sets leading to the construction of all real numbers as well as tran-

scendentals and inﬁnitesimals. Knuth’s novel [4] is an unusual and entertaining

presentation of this construction.

1.1 Ordered Fields

In this section we deﬁne and prove simple properties of ﬁelds, ordered ﬁelds, and

absolute values. The reader wishing to move on to properties of the real numbers

may skim the contents of the ﬁrst few subsections, and proceed to subsection

1.1.4.

Notes in Introductory Real Analysis 9

1.1.1 Fields

A ﬁeld F is a set along with two binary operations

Addition : FF →F : (a, b) →a+b (1.1)

and

Multiplication : FF →F : (a, b) →ab (1.2)

satisfying the following axioms:

1. The associative law holds for addition

a+(b+c) = (a+b) +c for all a, b, c ∈ F (1.3)

2. There is an element 0 ∈ F, the zero or additive identity element, for which

a+0 = a = 0+a for all a ∈ F (1.4)

3. Every element a ∈ F has an additive inverse −a, called the negative of a:

a+(−a) = 0 = (−a) +a (1.5)

4. The commutative law holds for addition:

a+b = b+a for all a, b ∈ F (1.6)

5. The associative law holds for multiplication

6. The associative law holds for addition

a(bc) = (ab)c for all a, b, c ∈ F (1.7)

7. There is an element 1 ∈ F, the unit or multiplicative identity element, for

which

a1 = a = 1a for all a ∈ F (1.8)

8. Every non-zero element a ∈ F has an multiplicative inverse a

−1

, called the

reciprocal of a:

aa

−1

= 1 = a

−1

a for all non-zero a ∈ F (1.9)

10 Ambar N. Sengupta

9. The commutative law holds for multiplication:

ab = ba for all a, b ∈ F (1.10)

10. The distributive law holds:

a(b+c) = ab+ac, (b+c)a = ba+ca for all a, b, c ∈ F (1.11)

11. The element 1 is not equal to the element 0:

1 = 0

We have not attempted to provide a minimal axiomset, and some of the axioms

may be deduced from others. For instance, the commutativity of addition can be

deduced from the other axioms.

Because of the associative laws, we will just write

a+b+c

instead of a+(b+c), and

abc

instead of abc.

Let us note a few simple consequences:

Theorem 1 If x ∈ F is such that

a+x = a for some a ∈ F

and y ∈ F is such that

by = b for some non-zero b ∈ F

then

x = 0 and y = 1.

In particular, the additive identity and the multiplicative identity are unique. More-

over,

−0 = 0 and 1

−1

= 1

Notes in Introductory Real Analysis 11

Proof. Adding −a to

a+x = a

shows that x is 0. Multiplying

by = b

by b

−1

shows that y is 1. The other claims follow from

0+0 = 0 and 1 1 = 1. QED

Theorem 2 If a, b ∈ F, and b = 0, then

−(−a) = a, and (b

−1

)

−1

= b.

Moreover,

(−a)b =−ab, and (−a)(−b) = ab.

The multiplicative inverse a

−1

is best written as the reciprocal:

1

b

= b

−1

,

and the product ab

−1

as

a

b

= ab

−1

There are many other easy consequences of the axioms which we will use

without comment.

We denote the set of natural numbers by P:

P =¦1, 2, 3, ...¦, (1.12)

and the set of integers by

Z =¦0, 1, −1, 2, −2, 3, −3, ...¦, (1.13)

We can multiply any element a ∈ F by an integer as follows. First deﬁne

1a = a,

where now 1 is the number one in Z. Next,

2a

def

= a+a,

12 Ambar N. Sengupta

and, inductively,

(n+1)a

def

= na+a (1.14)

for all a ∈ F and all n ∈ P. Next deﬁne negative multiples by

(−n)a =−(na) (1.15)

for all n ∈ P and all a ∈ F. Finally,

0a = a (1.16)

where 0 is the integer 0. The following facts can be readily veriﬁed:

x(ya) = (xy)a for all x, y ∈ Z and all a ∈ F (1.17)

x(a+b) = xa+xb for all x ∈ Z and all a, b ∈ F (1.18)

(x +y)a = xa+ya for all x, y ∈ Z and all a ∈ F (1.19)

There is a multiplicative analog of this given by powers of elements. If m is a

positive integer and a ∈ F we deﬁne

a

1

= a

and

a

m+1

= a

m

a.

We also deﬁne

a

0

= 1 for non-zero a ∈ F

and, for any positive integer m and non-zero a ∈ F,

a

−m

= (a

−1

)

n

=

1

a

m

We will use without comment simple facts such as

(a

m

)

n

= a

mn

,

valid for suitable a ∈ F and m, n ∈ Z.

The simplest example of a ﬁeld is the two-element ﬁeld

Z

2

=¦0, 1¦,

with addition and multiplication deﬁned modulo 2; for example,

1+1 = 0 in Z

2

We will, however, be concerned with ﬁelds which permit a consistent ordering of

their elements.

Notes in Introductory Real Analysis 13

1.1.2 Order Relations

An order relation on a set S is a set O of ordered pairs (x, y) of elements of S sat-

isfying the conditions O1 and O2 below. It is convenient to adopt the convention

that

x < y means (x, y) ∈ O

We also write

y > x

to mean x < y. The axioms of order are:

O1. For any x, y ∈ F exactly one of the following hold:

x = y, or x < y, or y < x.

O2. If x < y and y < z then x < z.

It is also convenient to use the notation:

x ≤y means x = y or x < y

and, similarly,

x ≥y means x = y or x > y

If T is a subset of an ordered set S then an element u ∈ S is said to be an upper

bound of T if

t ≤u for all t ∈ T (1.20)

If there is a least such upper bound then that element is called the supremum of T:

supT = the least upper bound of T (1.21)

We deﬁne similarly lower bounds and inﬁmums:

if l ≤t for every t ∈ T then l is called a lower bound of T (1.22)

and

inf T = the greatest lower bound of T (1.23)

Of course, the sup or the inf might not exist.

14 Ambar N. Sengupta

1.1.3 Ordered Fields

An ordered ﬁeld is a ﬁeld F with an order relation in which, in addition to the ﬁeld

and order axioms stated above, the following hold:

OF1. If x, y, z ∈ F and x < y then x +z < y +z:

x < y implies x +z < y +z for all x, y, z ∈ F (1.24)

OF2. If x, y, z ∈ F and x < y, and if also z > 0, then xz < yz:

x < y and z > 0 imply xz < yz for all x, y, z ∈ F. (1.25)

If x > 0 we say that x is positive. If x < 0 we say that x is negative.

We have the following simple observations for an ordered ﬁeld:

Theorem 3 Let F be an ordered ﬁeld. Then

(i) x > 0 if and only if −x < 0

(ii) For any non-zero x ∈ F we have x

2

> 0

(iii) 1 > 0

(iv) For any r ∈ Z the element r1 ∈ F is > 0 if r is a positive integer and is < 0

if r is a negative integer

(v) x > y holds if and only if x −y > 0

(vi) If x ∈ F and x > 0 then 1/x > 0

(vii) The product of two positive elements is positive

(viii) The product of a positive and negative is negative

(ix) The product of two negative elements is negative

(x) If x > y and z < 0 then xz < yz.

(xi) If x > y then −x <−y

(xii) If x > y > 0 then 1/x < 1/y

Notes in Introductory Real Analysis 15

Proof. We prove some of the results.

(i) Suppose x > 0. Then we have

x +(−x) > 0+(−x),

and so

0 >−x.

Conversely, if −x < 0 then adding x to both sides shows that x > 0.

(ii) If x > 0 then, by OF2, we have

x

2

= xx > x0 = 0.

On the other hand, if x < 0 then we know that −x > 0 and so

x

2

= (−x)(−x) > (−x)0 = 0.

(iii) Since 1 is 1

2

, it follows that 1 > 0.

(vi) Suppose x > 0. Since x(1/x) = 1, and 1 = 0, it follows that 1/x cannot be

zero. If 1/x < 0 then, however, x(1/x) would have to be < 0, but we know that

1 > 0. Thus, 1/x > 0.

(xi) If x > y then, adding −x −y shows that −y <−x.

(xii) If x > y > 0 then xy > 0 and hence 1/(xy) > 0. Multiplying x > y by the

positive element 1/(xy) gives 1/y > 1/x. QED

Observe that if x > 0 then

2x = x +x > x +0 = x,

and

3x = 2x +x > 2x +0 = 2x,

and, proceeding inductively, we have

0 < x < 2x < 3x < < nx < (n+1)x < for all n ∈ P and x > 0 in F

(1.26)

In particular, inside the ordered ﬁeld F we have a copy of the natural numbers

1, 2, 3, ... on identifying n ∈ P with n1

F

, where 1

F

is the unit element in F. This

then leads to a copy of the integers inside F, and we can assume that

Z ⊂F (1.27)

16 Ambar N. Sengupta

Going further, if m, n ∈ Z, and n = 0, we have then the ratio

m

n

∈ F.

The set of all such ratios m/n is the set of rationals

Q⊂F (1.28)

and is an ordered subﬁeld of the ordered ﬁeld F.

1.1.4 The absolute value function

The absolute value [ [ function in an ordered ﬁeld F is deﬁned by

[x[ =

x if x ≥0

−x if x ≤0

(1.29)

For instance,

[1[ = 1, and [ −1[ = 1.

The deﬁnition of [x[ shows directly that

[ −x[ =[x[ ≥0 for all x ∈ F (1.30)

It is also useful to observe that

[x[ is the larger of x and −x (1.31)

We think of

[a−b[

as measuring the difference between a and b.

We have then

Theorem 4 For any a, b ∈ F we have:

(i) the triangle inequality

[a+b[ ≤[a[ +[b[ (1.32)

Equality holds if and only if a and b are both ≥0 or both ≤0.

Notes in Introductory Real Analysis 17

(ii) the absolute values differ by at most the difference in a and b:

[a[ −[b[

≤[a−b[ (1.33)

(iii) [ab[ =[a[[b[.

Proof. Recall that [x[ is the larger of x and −x. Therefore, [a[ +[b[ is greater or

equal to ±a+(±b); in particular, it is greater or equal to a+b and (−a) +(−b).

Consequently, [a[ +[b[ is greater or equal to a+b and −(a+b), and so

[a[ +[b[ ≥[a+b[.

Equality holds if and only if [a[ = a and [b[ = b or [a[ =−a and [b[ =−b. Thus,

equality holds if and only if a and b are either both ≥0 or both ≤0.

Next, using the triangle inequality we have

[a−b[ +[b[ ≥[a+b −b[ =[a[,

and so

[a−b[ ≥[a[ −[b[.

Interchanging a and b yields:

[b−a[ ≥[b[ −[a[.

Observe that

[b−a[ =[a−b[.

Thus,

[a−b[ is ≥ to both [a[ −[b[ and −([a[ −[b[).

Since the larger of the latter is

[a[ −[b[

, we have proved (1.33).

We know that [x[ is x or −x, whichever is ≥0. Consequently, the product [a[[b[

is one of the elements

ab, (−a)b, a(−b), (−a)(−b),

i.e. one of the elements ab and −ab. Thus, [a[[b[ is ab or −ab, whichever is ≥0,

and so it is [ab[. QED

18 Ambar N. Sengupta

1.1.5 The Archimedean Property

An ordered ﬁeld F is said to have the archimedean property if for any x, y ∈ F,

with x > 0, there exists a multiple of x which exceeds y:

nx > y for some n ∈ ¦1, 2, 3, ...¦ (1.34)

The ﬁeld Q of rationals is clearly archimedean:

Theorem 5 The ordered ﬁeld Q of rationals is acrhimedean.

Proof. Let x, y ∈ Q, with x > 0. If y ≥ 0 then we have 1x > y and we are done.

Now suppose x, y > 0. Then

x =

p

q

y =

r

s

,

where p, q, r, s ∈ P. Take

n = qr +1.

Then

nx = qr(p/q) + p/q > pr ≥

r

s

= y,

and we are done. QED

In an archimedean ordered ﬁeld there are no inﬁnities, and there are also no

inﬁnitesimals:

Theorem 6 If F is an archimedean ﬁeld then for any w > 0 in F there is an n ∈ P

such that

1

n

w < x.

Proof. Simply choose n for which nx is > w. QED

1.2 The Real Number System R

We shall work with the real number system in an axiomatic way. We will assume

that it is an ordered ﬁeld in which the completeness axiom of completeness holds.

Needless to say, it is essential to actually construct such a system so as to be

sure that there is no hidden contradiction between the axioms, but we shall not

describe a construction in these notes.

Notes in Introductory Real Analysis 19

1.2.1 Hilbert Maximality and the Completeness Property

As we have mentioned before, the structure of Euclidean geometry, as formalized

through the axioms of Hilbert, produces an archimedean ordered ﬁeld. To com-

plete the story, one can add to these axioms the further requirement that this ﬁeld

is maximal in the sense that it cannot be embedded inside any larger archimedean

ordered ﬁeld. It turns out then that any such ordered ﬁeld is isomorphic to any

other, and thus there is essentially one such ordered ﬁeld. This ordered ﬁeld is the

real number system R.

A crucial fact about R is the completeness property:

If L and U are non-empty subsets of R such that every element of L

is ≤ every element of U, then there is a real number m which lies

between L and U:

l ≤m ≤u for all l ∈ L and all u ∈U. (1.35)

This property is also often expressed as:

If R is partitioned into two disjoint subsets L and U whose union is

R, and if every element of L is ≤ every element of U then there is a

unique element x ∈ R which lies between L and U:

l ≤x ≤u for all l ∈ L and all u ∈U. (1.36)

It is useful to view the real numbers geometrically. Consider a line, with two

special points O and I marked on it. For any point P on the line on the same side of

O as P we think of the ratio OP/OI as a positive real number. Points on the other

side from I correspond to negative real numbers, and the point O itself should be

thought of as 0. The completebess property says that there are no ‘gaps’ in the

line.

The completeness property can be formulated equivalently as:

Every non-empty subset of R which has an upper bound has a least upper bound.

(1.37)

The completeness property implies the archimedean peoperty:

Theorem 7 If in an ordered ﬁeld the property (1.37) holds then the archimedean

property also holds.

20 Ambar N. Sengupta

A proof of this is outlined in an exercise below.

The modern understanding and point of view on completeness grew out of the

work of Richard Dedekind [2].

1.2.2 Completeness of R and measurement

Even in Euclid’s geometry, a real number was, implicitly, understood in terms of

all rationals which exceeded it and all rationals below it. However, the traditional

axioms of Euclidean geometry, with requirements on intersections of lines and

circles, can work with a ﬁeld which is larger than the rationals but smaller than R,

and completeness is not essential.

The simplest measurement problem is to devise a measure of sets of points

in the plane which are made up of a ﬁnite collection of segments constructed by

Euclidean geometry. Two such sets should have the same measure if they can be

decomposed into a ﬁnite collection of congruent segments. For this we would

not need the full complete system R of real numbers. Moving up a dimension,

with the task of measuring areas of polygonal regions constructed by Euclidean

geometry, one could still get away with a less-than-complete system of numbers.

However, it was shown by Max Dehn in 1900, in resolving Hilbert’s Third Prob-

lem, that there are polyedra in three dimensions which have equal volumes (as

deﬁned by requirements of ‘upper’ and ‘lower’ approximations) which cannot be

decomposed into congruent pieces. This, along with, of course, the utility of

measuring areas of curved regions even in two dimensions, makes it absolutely

essential to work with a notion of measure that goes beyond simply decomposing

into geometrically congruent ﬁgures. For a truly useful theory of measure, the

completeness of the number system is essential.

Capturing a real number between upper approximations and lower approxima-

tions proves to be very useful. Archimedes and others computed areas of curved

regions by such upper and lower approximations. In modern calculus, this method

lives on in the Riemann integral, as we shall see later.

Problem Set 1

1. Prove that in any ordered ﬁeld, between any two distinct elements there is

at least one other element.

2. Prove that in any ordered ﬁeld, between any two distinct element lie in-

ﬁnitely many elements.

Notes in Introductory Real Analysis 21

3. If F is an archimedean ordered ﬁeld, such as R, show that between any two

distinct elements there is a rational element.

4. Let F be an archimedean ordered ﬁeld, and let x, y ∈ F with x < y. If z ∈ F

show that there is a rational multiple qz of z for which x < qz < y.

5. Suppose x ∈ R is ≥all elements of a non-empty subset S ⊂R. Explain why

this implies x ≥supS.

6. Suppose S is a non-empty subset of R and x ∈ R is not an upper bound of

S. Show that there exists y > x which is also not an upper bound of S.

7. Prove that the completeness property (1.37) implies the property (1.36).

[Hint: Suppose (1.37) holds. If S ⊂R is a non-empty set which is bounded

above, then we can take U to be the set of all upper bounds of S and L to be

the complement, i.e. all real numbers which are < some element of S. Let

x ∈ F be provided by (1.36). Show that x is the least upper bound of S.]

8. Prove that (1.36) implies the completeness property (1.37).

9. Prove that the completeness property implies the archimdean propery. [Hint:

Suppose F is an ordered ﬁeld in which (1.37) holds. Let x, y ∈ F, with

x > 0. Suppose that no positive integral multiple of x exceeds y, and let

S = ¦nx : n ∈ P¦. Then y is, by assumption, an upper bound for S. Let

u = supS, the least upper bound of S. Consider the element u −

1

2

x. This,

being < u, is not an upper bound. Use this to produce an element of S

greater than u, thus reaching a contradiction.]

10. Suppose L and U are non-empty subsets of R such that (i) every element

of L is ≤ every element of R, and (ii) for any ε > 0 there is an element

l ∈ L and an element u ∈U with u−l < ε. Prove that there is a unique real

number which is ≤ all elements of U and ≥ all elements of L.

22 Ambar N. Sengupta

Chapter 2

The Extended Real Line and Its

Topology

In this chapter we study topological concepts in the context of the real line. For

technical purposes, it will be convenient to extend the real line R by adjoining

to it a largest element ∞ and a smallest element −∞. No metaphysical meaning

need be attached to these inﬁnities. The primary reason for introducing them is to

simplify the statements of several theorems.

2.1 The extended real line

The extended real line is obtained by a largest element ∞, and a smallest element

−∞, to the real line R:

R

∗

=R∪¦−∞, ∞¦ (2.1)

Here ∞ and −∞ are abstract elements. We extend the order relation to R by

declaring that

−∞ < x < ∞ for all x ∈ R (2.2)

Much of our work will be on R

∗

, instead of just R.

We deﬁne addition on R

∗

as follows:

x +∞ = ∞ = ∞+x for all x ∈ R

∗

with x >−∞ (2.3)

y +(−∞) = −∞ = (−∞) +y for all y ∈ R

∗

with y < ∞. (2.4)

Note that

∞+(−∞) is not deﬁned,

23

24 Ambar N. Sengupta

i.e. there is no useful or consistent deﬁnition for it.

The following algebraic facts continue to hold in R

∗

:

x +y = y +x, x +(y +z) = (x +y) +z, (2.5)

whenever either side of these equations holds (i.e. if one side is deﬁned then so is

the other and the equality

2.2 Neighborhoods

A neighborhood of a point p ∈ R is an interval of the form

(p−δ, p+δ)

where δ > 0 is any positive real number. Thus, the neighborhood consists of all

points distance less than δ from p:

(p−δ, p+δ) =¦x ∈ R : [x −p[ < δ¦. (2.6)

For example,

(1.2, 1.8)

is a neighborhood of 2.

A typical neighborhood of 0 is of the form

(−ε, ε)

for any positive real number ε.

A neighborhood of ∞ in R

∗

=R∪¦−∞, ∞¦ is a ray of the form

(t, ∞] =¦x ∈ R

∗

: x >t¦

with t any real number. For example,

(5, ∞]

is a neigborhood of ∞.

A neighborhood of −∞ in R

∗

is a ray of the form

[−∞, s) =¦x ∈ R

∗

: x < s¦

Notes in Introductory Real Analysis 25

where s ∈ R. An example is

[−∞, 4)

Observe that if U and V are neighborhoods of p then U∩V is also a neighbor-

hood of p. In fact, either U contains V as a subset of vice versa, and so U ∩V is

just the smaller of the two neighborhoods.

Observe also that if N is a neighborhood of a point p, and if q ∈ N then q had

a neighborhood lying entirely inside N. For example, the neighborhood (2, 4) of 3

contains 2.5, and we can form the neighborhood (2, 3) of 2.5 lying entirely inside

(2, 4).

Here is a simple but fundamental observation:

Distinct points of R

∗

have disjoint neighborhoods. (2.7)

This is called the Hausdorff property of R

∗

.

For example, 3 and 5 have the neighborhoods

(2, 4) and (4.5, 5.5)

The points 2 and ∞ have disjoint neighborhoods, such as

(−1, 5) and (12, ∞]

Exerise Give examples of disjoint neighborhoods of

(i) 2 and −4

(ii) −∞ and 5

(iii) ∞ and −∞

(iv) 1 and −1

2.3 Types of points for a set

Consider a set

S ⊂R

∗

.

A point p ∈ R

∗

is said to be an interior point of S if it has a neighborhood U lying

entirely inside S, i.e. with

U ⊂S.

26 Ambar N. Sengupta

For example, for the set

E = (−4, 5] ∪¦6, 8¦∪[9, ∞],

the points −2, 3, 11 are interior points. The point ∞ is also an interior point of E.

A point p is an exterior point if it has a neighborhood U lying entirely outside

S, i.e.

U ⊂S

c

.

For example, for the set E above, points −5, 7, and −∞ are exterior to E.

A point which is neither interior to S nor exterior to S is a boundary point of

S. Thus p is a boundary point of S if every neighborhood of p intersects both S

and S

c

.

In the example above, the boundary points of E are

−4, 5, 6, 8, 9, ∞.

Next consider the set

¦3¦∪(5, ∞)

The boundary points are 3, 5, and ∞. It is important to observe that if we work

with the real line R instead of the extended line R

∗

then we must exclude ∞ as a

boundary point, because it doesn’t exist as far as R is concerned.

Example For the set A = [−∞, 4) ∪¦5, 9¦ ∪[6, 7), decide which of the following are

true and which false:

(i) −6 is an interior point (T)

(ii) 6 is an interior point (F)

(iii) 9 is a boundary point (T)

(iv) 5 is an interior point (F)

Exerise For the set B = [−∞, −5) ∪¦2, 5, 8¦∪[4, 7), decide which of the following

are true and which false:

(i) −6 is an interior point

(ii) −5 is an interior point

(iii) 5 is a boundary point

(iv) 4 is an interior point

(v) 7 is a boundary point.

Notes in Introductory Real Analysis 27

2.4 Interior, Exterior, and Boundary of a Set

The set of all interior points of a set S is denoted

S

0

and is called the interior of S.

The set of all boundary points of S is denoted

∂S

and is called the boundary of S.

The set of all points exterior to S is the exterior of S, and we shall denote it

S

ext

.

Thus, the whole extended line R

∗

is split up into three disjoint pieces:

R

∗

= S

0

∪∂S∪S

ext

(2.8)

Recall that a point p is on the boundary of S if every neighborhood of the

point intersects both S and S

c

. But this is exactly the condition for p to be on the

boundary of S

c

. Thus

∂S = ∂S

c

. (2.9)

The interior of the entire extended line R

∗

is all of R

∗

. So

∂R

∗

= / 0.

Example For the set A = [−∞, 4) ∪¦5, 9¦∪[6, 7),

(i) A

0

= [−∞, 4) ∪(6, 7)

(ii) ∂A =¦4, 5, 9, 7, 6¦

(iii) A

c

= [4, 5) ∪(5, 6) ∪[7, 9) ∪(9, ∞]

(iv) the interior of the complement A

c

is

(A

c

)

0

= (4, 5) ∪(5, 6) ∪(7, 9) ∪(9, ∞]

28 Ambar N. Sengupta

For the set

G = (3, ∞)

the boundary of G, when viewed as a subset of R

∗

, is

∂G =¦3, ∞¦.

But if we decide to work only inside R then the boundary of G is just ¦3¦.

Exerise For the set B =¦−4, 8¦∪[1, 7) ∪[9, ∞),

(i) B

0

=

(ii) ∂B =

(iii) B

c

=

(iv) the interior of the complement B

c

is

(B

c

)

0

=

2.5 Open Sets and Topology

We say that a set is open if it does not contain any of its boundary points. For

example,

(2, 3) ∪(5, 9)

is open.

Also

(4, ∞]

is open.

But

(3, 4]

is not open, because the point 4 is a boundary point.

The entire extended line R

∗

is open, because it has no boundary points.

Moreove, the empty set / 0 is open, because, again, it doesn’t have any boundary

points.

Notice then that every point of an open set is an interior point. Thus, a set S is

open means that

S

0

= S.

Notes in Introductory Real Analysis 29

Thus for an open set S each point has a neighborhood contained entirely inside

S. In other words, S is made up of a union of neighborhoods.

Viewed in this way, it becomes clear that the union of open sets is an open set.

Now consider two open sets A and B. We will show that A∩B is open. Take

any point

p ∈ A∩B.

Then p is in both A and B. Since p ∈A and since A is open, there is a neighborhood

U of p which is a subset of A:

U ⊂A.

Similarly there is a neighborhood V of p which is a subset of B:

V ⊂B

But then U ∩V is a neighborhood of p which is a subset of both A and B:

U ∩V ⊂A∩B.

Thus every point in A∩B has a neighborhood lying inside A∩B. Consequently,

A∩B is open.

Now consider three open sets A, B,C. The intersection

A∩B∩C

can be viewed as

(A∩B) ∩C

But here both A∩B and C are open, and hence so is their intersection. Thus,

A∩B∩C

is open. This type of argument works for any ﬁnite number of open sets. Thus:

The intersection of a ﬁnite number of open sets is open.

Exerise Check that the intersection of the sets (4, ∞) and (−35) and (2, 6) is open.

The collection of all open subsets of R is called the topology of R.

The set of all open subsets of R

∗

is called the topology of R

∗

.

30 Ambar N. Sengupta

2.6 Closed Sets

A set S is said to be closed if it contains all its boundary points.

In other words, S is closed if

∂S ⊂S

Thus,

[4, 8] ∪[9, ∞]

is closed.

But

[4, 5)

is not closed because the boundary point 5 is not in this set.

The set

[3, ∞)

is not closed (as a subset of R

∗

) because the boundary point ∞ is not inside the

set. But, viewed as a subset of R it is closed. So we need to be careful in deciding

what is close and what isn’t: a set may be closed viewed as a subset of R but not

as a subset of R

∗

.

The full extended line R

∗

is closed.

The empty set / 0 is also closed.

Note that the sets R

∗

and / 0 are both open and closed.

2.7 Open Sets and Closed Sets

Consider a set S ⊂R

∗

.

If S is open then its boundary points are all outside S:

∂S ⊂S

c

.

But recall that the boundary of S is the same as the boundary of the complement

S

c

. Thus, for S to be open we must have

∂(S

c

) ⊂S

c

,

which means that S

c

contains all its boundary points. But this means that S

c

is

closed.

Thus, if a set is open then its complement is closed.

The converse is also true: if a set is closed then its complement is open. Thus,

Notes in Introductory Real Analysis 31

Theorem 8 A subset of R

∗

is open if and only if its complement is closed.

Exercise. Consider the open set (1, 5). Check that its complement if closed.

Exercise. Consider the closed set [4, ∞]. Show that its complement is open.

2.8 Closed sets in R and in R

∗

The set

[3, ∞)

is closed in R, but is not closed in R

∗

. This is because in Rit has only the boundary

point 3, which it contains; in contrast, in R

∗

the point ∞ is also a boundary point

and is not in the set. Thus, when working with closed sets it is important to bear

in mind the distinction between being closed in R and being closed in R

∗

. There

is no such distinction for open sets.

2.9 Closure of a set

The closure of a set S is obtained by throwing in all its boundary points:

S = S∪∂S (2.10)

Of course, if S is closed then its closure is itself.

For example, the closure of (3, 5) is [3, 5]. The closure of

(3, ∞)

is [3, ∞]. (But, in R the closure of (3, ∞) is [3, ∞).)

Let us see what the closure of Q is. Now every point in R

∗

is a boundary point

of Q:

∂Q =R

∗

,

because any neighborhood of any point contains both rationals (points in Q) and

irrationals (points outside Q).

It is useful to think of the closure

S = S∪∂S

in this way:

A point p is in S if and only if every neighborhood of p intersects S.

32 Ambar N. Sengupta

2.10 The closure of a set is closed

Consider the closure S of a set S. We will show that S is a closed set.

Take any boundary point p ∈ ∂S. We have to show that p is actually in S. Now

let N be any neighborhood of p. Then N contains a point q in S. Choose (as we

may) a neighborhood W of q lying entirely inside N. Since q ∈ S it follows that W

contains a point of S. Thus, N contains a point of S. So we have seen that every

neighborhood of p contains a point of S. This means p ∈ S. Thus, we have shown

that every boundary point of S is in S, and so S is closed.

2.11 S is the smallest closed set containing S

Consider any closed set K with

S ⊂K

Take any p ∈ S. Then every neighborhood of p intersects S, and hence also K.

Thus every neighborhood of p intersects K. So, either p is in the interior of K or

on its boundary. But K is closed, and so in either case p is in K. Thus,

S ⊂K

What we hae shown then is that S is the smallest closed set containing S as a

subset.

2.12 R

∗

is compact

There is a special property of R

∗

whose full signiﬁcance is best appreciated at a

later stage. However, we have the language and tools to state and prove it now

and shall do so.

An open cover of R

∗

is a collection U of open sets whose union covers all of

R

∗

. For example,

¦[−∞, 5), (−1, 8), (7, 9) ∪(11, 15), (8, ∞]¦

is an open cover of R

∗

. It is best to draw a picture for yourself to make this clear.

More formaly, an open cover of R

∗

is a set U of open subsets of R

∗

such that

every point x ∈ R

∗

falls inside some open set U in the collection U, i.e. for each

x ∈ R

∗

there exists U ∈ U such that

x ∈U ∈ U.

Notes in Introductory Real Analysis 33

Another example of an open cover of R

∗

is

¦[−∞, 50), (2, 100), (8, 200) ∪(701, 800), (150, 750), (760, ∞]¦.

The examples of open covers of R

∗

given above are both ﬁnite collections of

sets. Let us look at an example of a cover which uses inﬁnitely many sets.

We start with an open cover of R, as opposed to R

∗

: take all intervals of the

form (a, a+2), where a runs over all integers:

..., (−8, −6), (−7, −5), ...(−2, 0), (−1, 1), (0, 2), (1, 3), ...

In short we are looking at the collection

U

/

=¦(a, a+2) : a ∈ Z¦,

where, recall that Z is the set of all integers. This collection fails to include −∞

and ∞.

To obtain an open cover of R

∗

from U

/

we could, for example, just throw in

the open sets [−∞, −4) and (5, ∞]. Thus, this gives us an open cover of R

∗

:

U =¦[−∞, −4), (5∞]¦∪¦(a, a+2) : a ∈ Z¦.

If you look at this, however, you realize that even though this collection does

contain inﬁnitely many open sets, we don’t reall need all of them to cover R

∗

.

Indeed, we could just use the sub-collection

V =¦[−∞, −4), (−3, −1), (−2, 0), (−1, 1), (0, 2), (1, 3), (2, 4), (3, 5), (4, 6), (5, ∞]¦

and this would cover all of R

∗

.

This is not just a feature of one particular example. It turns out that

Every open cover of R

∗

has a ﬁnite subcover.

This property is called compactness, and so

Theorem 9 R

∗

is compact.

We can prove the compactness of R

∗

using the completeness of the real line.

Proof of compactness of R

∗

. Let U be an open cover of R

∗

. This is a collection

of open sets such that every point of R

∗

falls inside some set in the collection. In

particular, −∞ is in some open set W ∈ U. Thus W contains a neighborhood of

−∞, i.e.

[−∞, t) ⊂W

34 Ambar N. Sengupta

for some t ∈ R. So, for one thing, all the elements of R

∗

less than t are covered

by just the one set W in U; we would not need any other set from the cover if we

are to stick to points to the left of t. Now let

S =¦x ∈ R

∗

: ﬁnitelymanysetsinUcover [−∞, x]¦

This set is not empty because −∞∈ S. In fact, t −1 is also in S. By completeness,

the set S has a supremum supS. Let

m = supS.

We will argue in three steps:

(i) First we explain that m isn’t −∞.

(ii) Next we prove that m is in fact, ∞.

(iii) Finally we prove that [∞, m], i.e. all of R

∗

, can be covered by just ﬁnitely

many sets in U.

For (i), note, as before, that there is a real number t such that t −1 is in S, and

so m (being an upper bound of S) has to be ≥t −1. In particular, m is certainly

not −∞.

Now we show that m = ∞. Suppose to the contrary that m < ∞. Now there is

some open set U ∈ U such that m ∈U, because U covers all points of R

∗

. Since

m is not −∞, and is being assumed to be also not ∞, it is in R and so there is some

neighborhood

(m−ε, m+ε) ⊂U,

where ε is a positive real number. Now m−ε being less than the least upper bound

m of S, there has to be some x ∈ S which is greater than m−ε. Thus

m−ε < x ≤m for some x ∈ S.

So, x being in S, there are ﬁnitely many sets, say U

1

, ...,U

N

, in U, which cover the

ray segment [−∞, x]. The set U covers (m−ε, m+ε). Thus, the collection

¦U

1

, ...,U

N

,U¦

covers all of

[−∞, m+ε).

Notes in Introductory Real Analysis 35

But this means that, for example, m+

1

2

ε is in S, for the ray segment [−∞, m+1/2]

is covered by the ﬁnite collection U

1

, ...,U

N

,U. But now we have a contradiction,

because we have found a number, m+ε/2, greater than m, which is in S. Thus,

our original hypothesis concerning m must have been wrong. So m = ∞.

Finally we prove that R

∗

is covered by ﬁnitely many sets in U. The element

∞ ∈ R

∗

falls inside some open set V in the collection U. Therefore, there is some

‘ray-neighborhood’ of ∞

(r, ∞] ⊂V,

for some real number r. Now r being less than m = ∞, and the latter being the

least upper bound of S, there must be some y ∈ (r, ∞] which is in S. Thus,

[−∞, y]

is covered by ﬁnitely many open sets, say V

1

, ...,V

k

, in U. Note that

[−∞, y] ∪(r, ∞] =R

∗

.

But then

V

1

, ...,V

k

,V

cover all of [−∞, ∞], since V covers the segment (r, ∞]. This prove that ﬁnitely

many sets from U cover all of R

∗

.

2.13 Compactness of closed subsets of R

∗

.

Consider now any closed subset D of R

∗

. We will prove that it is compact, i.e.

that any open cover of D has a ﬁnite sub-cover.

Consider any open cover U of D. This is a collection of open sets such that

every point of D is covered by some set in the collection. Now the set D

c

, being

the complement of a closed set, is an open set. Throw this into the collection, and

consider

U

/

=U∪¦D

c

¦.

This covers all of R

∗

: any point of D would be covered by a set in U while any

point in D

c

is, of course, covered by D

c

. Then we know that there has to be a ﬁnite

sub-collection

V

/

⊂U

/

36 Ambar N. Sengupta

which covers R

∗

. Now throw out D

c

from V

/

in case it is there, and consider the

collection

V =V

/

`¦D

c

¦.

Of course, this is a ﬁnite subcollection of U. Moreover, it covers all points of D,

because no point in D could have been covered by the ‘rejected’ set D

c

. Thus, D

is covered by a ﬁnite sub-collection of sets from U.

Thus, every closed subset of R

∗

is compact. The converse is also true, and we

can state:

Theorem 10 A subset of R

∗

is compact if and only if it is closed.

We will leave out the proof of the converse part of this result.

2.14 The Heine-Borel Theorem: Compact subsets

of R

A subset B of R is said to be bounded if

B ⊂[−N, N]

for some real number N. Thus, for B to be bounded, there should exist a real

number N such that

[x[ < N for all x ∈ B.

Recall that a subset of Ris closed in Rif it contains no boundary point. Equiv-

alently, if a subset of R is closed if its complement in R is open.

Consider, for instance, the set

[4, ∞) ⊂R.

This is a closed subset of R because, in R, its only boundary point is 4, and this

point lies in the set. Note that [4, ∞) is not closed when considered as a subset of

R

∗

.

We can now state the Heine-Borel theorem:

Theorem 11 Every closed and bounded subset of R is compact. Conversely, ev-

ery compact subset of R is closed and bounded.

Notes in Introductory Real Analysis 37

We shall prove half of this. Suppose K ⊂R is closed and bounded as a subset

of R. Boundedness implies that

K ⊂[−N, N]

for some real number N. Then K is closed also as a subset of R

∗

. Let’s check

this. We will show that the complement U of K in R

∗

is open. Consider any

point p ∈ U. If p ∈ R then we already know, by closedness of K, that p has a

neighborhood lying inside U. If p = ∞ then the neighborhood of p given by

(N+1, ∞]

lies entirely outside K. If p =−∞ then the neighborhood

[−∞, −N−1)

is entirely outside K. Thus, in all cases, p has a neighborhood outside K in R

∗

. So

the complement of K in R

∗

is open, and hence K is closed in R

∗

. We know that

then K must be compact.

2.15 Sequences

A sequence of elements in a set A is a string

a

1

, a

2

, a

3

, ...

of elements of A.

More precisely, the sequence is actually a mapping

a : P →A : n →a

n

.

We will often be concerned with sequences in R

∗

.

Sometimes out sequence will be speciﬁed explicitly as a string of numbers;

for example,

−5, −3, 5, 6, 9, 12, ...

Sometimes we may have a formula for the n-th term:

a

n

=

(−1)

n

n+1

.

38 Ambar N. Sengupta

Sometimes we have a sequence speciﬁed recursively. For example, we might

know that

b

1

= 1, b

2

= 1

and then

b

n

= b

n−1

+b

n−2

,

for all n ≥3. This generates the Fibonacci numbers

1, 1, 2, 3, 5, 8, 13, 21, ...

Sometimes it is better to label a sequence starting with the index 0:

a

0

, a

1

, a

2

, ...

For example, for a ﬁxed real number r we can form the sequence

r

0

= 1, r

1

= r, r

2

, r

3

, ...

2.16 Limits points of a sequence

Consider a sequence

x

1

, x

2

, x

3

, ....

in R

∗

. A point p ∈ R

∗

is said to be a limit point of this sequence if every neigh-

borhood of p is visited inﬁnitely often by the sequence.

For example, for the sequence

1

2

, 4,

1

3

, 4+

1

3

,

1

4

, 4+

1

4

,

1

5

, 4+

1

5

,

1

6

, 4+

1

6

, ...

it is intuitively clear that both 0 and 4 are limit points.

Suppose a sequence (x

n

) lies entirely inside a set S ⊂R:

x

n

∈ S, for all n ∈ P

Consider then any limit point p of this sequence. Let U be any neighborhood of

p. We know that this neighborhood is visited inﬁnitely often by the sequence.

Therefore, at least one point of S must be in U. Thus, every neigborhood of p

contains a point of S, and so

p ∈ S.

Thus, any limit point of a sequence which lies always in a set S must be in the

closure of S.

Notes in Introductory Real Analysis 39

2.17 Bolzano-Weierstrass Theorem

Consider any sequence in R

∗

. We shall prove that it must have at least one limit

point.

Suppose to the contrary that no point is a limit of the given sequence. Then

each point p of R

∗

has a neighborhood U

p

which is visited only ﬁnitely often by

the sequence. The neighborhoods U

p

form an open cover of all of R

∗

. Then, by

compactness of R

∗

, there are ﬁnitely many of them, say

U

p

1

, ...,U

p

N

which cover all of R

∗

. But this is impossible: we have covered all of the extended

real line R

∗

with ﬁnitely many sets each of which is visited only ﬁnitely many

times by the sequence! Thus, we have a contradiction and so there exist a limit of

the sequence.

Notice that all we needed above was the compactness of R

∗

. Thus, we have

the more general Bolzano-Weierstrass theorem:

Theorem 12 Every sequence in a compact set has at least one limit point.

Of course, we have seen that a sequence may well have more than one limit

point. For example, the sequence,

1, 1, 2, 1, 2, 3, 1, 2, 3, 1, 2, 3, 4, 1, 2, 3, 4, 5, ...

visits every natural number inﬁnitely often and so every natural number is a limit

point of the sequence.

In the next subsection we will show how to actually obtain a limit point of a

sequence.

2.18 Limit points and suprema and inﬁma

Consider a sequence

x

1

, x

2

, x

3

, ...

Let p be any limit point of this sequence.

Now let S

1

be the least upper bound of the set ¦x

1

, x

2

, ...¦:

S

1

= sup¦x

1

, x

2

, ...¦

40 Ambar N. Sengupta

In particular,

S

1

≥x

n

for each n.

Then it is clear that no point ‘to the right’ of S

1

could possibly be a limit of the

sequence: indeed any point r > S

1

would have a neighborhood lying entirely to

the ‘right’ of S

1

and this would never be visited by the sequence. Thus, p cannot

be > S

1

. So

p ≤S

1

Now we can apply the same argument to

S

2

= sup¦x

2

, x

3

, x

4

, ...¦

Again, p could not be greater than S

2

for then it would have a neighborhood to the

right of S

2

and this would have to be visited inﬁnitely often by x

1

, x

2

, ... and so at

least once by x

2

, x

3

, .... Thus,

p ≤S

2

In this way, we can form

S

3

= sup¦x

3

, x

4

, ...¦

and have again

p ≤S

3

. Thus, p is a lower bound for all the S

k

’s:

p ≤S

k

for all k.

Now the inf of a set is the greatest lower bound of the set. Therefore,

p ≤inf¦S

1

, S

2

, S

3

, ...¦.

The infemum on the right is an important object and is called the limsup of the

given sequence:

limsup

n→∞

x

n

def

= inf¦sup

n≥k

x

n

¦ : k ∈ P¦ (2.11)

So we have proved that

Every limit point is ≤ the limsup

Notes in Introductory Real Analysis 41

for any sequence.

Similarly, we have the notion of liminf: it is the supermum of the sequence of

infemums:

liminf

n→∞

x

n

def

= sup¦inf

n≥k

x

n

¦ : k ∈ P¦ (2.12)

In just the same way as before, we can prove

Every limit point is ≥ the limsup

Thus, for any limit point p if any sequence (x

n

) we have

liminf

n→∞

x

n

≤anylimit point ≤limsup

n→∞

x

n

. (2.13)

In fact, the limsup and liminf of the sequence are also limit points, but we won’t

prove this here.

2.19 Limit of a sequence

A sequence (s

n

) is said to lie in a set S eventually if after a certain value of n, all

the s

n

lie in S. Put another way, we say that s

n

lies in S for large n, if s

n

∈ S for all

values of n beyond some value, say n

0

.

For example, the sequence

−5, −4, −3, −2, −1, 0, 1, 2, 3, ...

will lie in the set (20, ∞) eventually.

Let us note again: a sequence (s

n

) lies in a set S eventually if there is an n

0

∈P

such that

s

n

∈ S

for all n ∈ P with n ≥n

0

.

A sequence (x

n

) in R

∗

is said to have limit L ∈ R

∗

if for any neighborhood U

of L the sequence lies in this neighborhood eventually.

We denote this symbolically as

x

n

→L, as n →∞.

42 Ambar N. Sengupta

We shall see later that if a limit exists then it is unique; the limit L is denoted

lim

n→∞

x

n

.

The notion of limit is one of the central notions in mathematics.

For example, the sequence

1, 2, 3, 4, ....

has limit ∞. For, if we take any neighborhood of ∞, say

(t, ∞]

then eventually n will exceed t (by the archimidean property) and so the sequence

will stay in (t, ∞] from the n-th term onwards.

The sequence

1,

1

2

,

1

3

,

1

4

, ...

has limit 0. We will look at this more carefully soon.

2.20 Simple Examples of limits

Let us look at some examples of sequences and try to see what their limits are.

The simplest sequence is a constant sequence which keeps repeating the same

value. For example,

3, 3, 3, 3, ....

The limit of the this sequence is 3: clearly every neighborhood of 3 is hit even-

tually by the sequence (indeed it is hit every time, since the sequence is stuck at

3).

The sequence

−1, 4, 5, 7, 8, 8, 8, 8, 8, ....

which eventually stabilizes at the constant value 8 has limit 8. For, again, given

any neighborhood of 8 the sequence falls inside this neighborhood eventually and

stays there.

In contrast, the sequence

1, 3, 4, 1, 3, 4, 1, 3, 4, 1, 3, 4, ...

Notes in Introductory Real Analysis 43

does not have a limit. For example, the point 3 cannot be the limit of the sequence

because, for instance,

(2.5, 3.5)

is a neighborhood of 3, but the sequence keeps falling outside this neighborhood

(when it hits 1 or 4).

The sequence

1, 3, 5, 7, ...

has limit ∞. If you take any neighborhood of ∞, an interval of the form

(t, ∞]

then eventually the sequence falls inside the nighborhood and stays in there.

2.21 The sequence 1/n

Consider again the sequence

1, 1/2, 1/3, 1/4, ...

It is intuitively clear that this sequence has limit 0. But let us prove this.

Note ﬁrst that the n-th term of the sequence is

x

n

= 1/n.

We have to show that given any neighborhood of 0, our sequence will eventu-

ally lie inside this neighborhood. So consider any neighborhood of 0:

(−ε, ε),

where ε is a positive real number. We have to show that x

n

lies in (−ε, ε) for all n

beyond some value. Thus we should show that

1

n

< ε

for all n beyond some initial value. Now the condition 1/n < ε is equivalent to

nε > 1.

44 Ambar N. Sengupta

The archimedean property guarantees the existence of an n

0

∈ P for which

n

0

ε > 1.

Therefore, nε > 1 for all integers n ≥ n

0

. This proves that the sequence does

indeed tend to the limit 0:

1

n

→0, as n →∞

2.22 The sequence R

n

Consider the sequence

3

0

, 3

1

, 3

2

, 3

3

, ...

It is clear that in this sequence the terms get large very quickly, and intuitively it

is clear that the sequence has limit ∞.

Consider now the sequence of powers

R

n

where R > 1.

Let us see by how much each term exceeds the preceding:

R

n

−R

n−1

= R

n−1

(R−1)

But R is > 1, and so the multiplier R

n−1

is ≥1 (it is equal to 1 when n is actually

1). Thus,

R

n

−R

n−1

≥R−1.

Let us write x for R−1, and note that

x > 0

and we have

R

n

−R

n−1

> x.

Now it takes n ‘steps’ to climb from R

0

to R

n

, and so

R

n

≥1+nx,

Notes in Introductory Real Analysis 45

because each step is at least x. Now it is clear that R

n

→∞ as n →∞. We just have

to show that for any given t,

1+nx >t

when n is large enough. But this is just Archimedes again: some multiple n

0

x of

xexceeds t −1, and then

nx >t −1

for all n ≥ n

0

. Note again that it is important that x > 0, and this comes from the

fact that R > 1.

Thus,

lim

n→∞

R

n

= ∞ for all R > 1.

Now consider the case R = 1. In this case the sequence is just all powers of 1,

and so it is

1, 1, 1, 1, ....

Thus

lim

n→∞

R

n

= 1 if R = 1.

Next consider the case

0 < R < 1.

As an example, we have R = 1/3 and the sequence

1,

1

3

,

1

3

2

,

1

3

3

, ...

We have here a sequence with denomiantor going to ∞ and numerator ﬁxed at 1.

It seems then clear that the sequence goes to 0. Indeed, for any real ε > 0 we just

have to wait till

3

n

> 1/ε

and this would ensure that

1

3

n

< ε,

and so

1

3

n

∈ (−ε, ε) for large n.

Now consider the general case of

R

n

46 Ambar N. Sengupta

where R > 1. Observe that

R =

1

r

,

where r = R

−1

is > 1. Thus

R

n

=

1

r

n

,

where r > 1. Now, as we have seen above, the denominator r

n

goes to inﬁnity,

and so it seems clear that 1/r

n

should decrease to 0. So we will prove that

R

n

→0

in this case. Consider any neighborhood of 0:

(−ε, ε)

where ε > 0 is a positive real number. We want R

n

to be in this neighborhood, i.e.

R

n

should be < ε. But this means we should show that r

n

is > 1/ε for all n large

enough. But we know that

r

n

→∞

So, after some n

0

, we have

r

n

> ε

−1

for all n ≥n

0

. Consequently,

R

n

=

1

r

n

< ε

and so

R

n

∈ (ε, ε),

for all n ≥n

0

. Thus

lim

n→∞

R

n

= 0 if 0 < R < 1.

Now it is also clear that

lim

n→∞

R

n

= 0 if R = 0.

Next consider negative values. Suppose

−1 < R < 0.

For example, for R =−1/3 we have the sequence

1, −

1

3

,

1

3

2

, −

1

3

3

,

1

3

4

, ...

Notes in Introductory Real Analysis 47

It is clear that this is the same sequence as 1/3

n

except it swings back and forth

between negative and positive values. Thus, this sequence has limit 0.

More generally, if

−1 < R < 0

we can look at the distance between R

n

and 0:

[R

n

−0[ =[R

n

[ =[R[

n

and

[R[

n

→0 as n →∞,

because

0 <[R[ < 1.

Thus, the conclusion is that

lim

n→∞

R

n

= 0 if −1 < R < 0.

In short,

lim

n→∞

R

n

= 0 if [R[ < 1. (2.14)

Lastly, one should look at the case

R ≤−1.

You should examine a few examples and convince yourself that then there is no

limit, for the sequence swings back and forth between widely separated positive

and negative values.

Exerise Examine the sequence given by powers of −2:

1, −2, 4, −8, 16, ...

Does this sequence visit every neighborhood of ∞? Does it visit every

neighborhood of −∞?

48 Ambar N. Sengupta

2.23 Monotone Sequences

A sequence such as

1, 2, 5, 8, 9, 11, ...

where each term is ≥ the preceding term is said to be monotone increasing. An-

other example is the sequence

1, 1.1, 1.11, 1.111, 1.1111, ...

Thus, a sequence (x

n

) is

monotone increasing if x

n

≤x

n+1

for all n ∈ P (2.15)

We say that (x

n

) is

monotone decreasing if x

n

≥x

n+1

for all n ∈ P (2.16)

An example of a monotone decreasing sequence is

1, 1/2, 1/3, 1/4, 1/5, ...

It is intuitively clear that a monotone increasing sequence (x

n

) will tend to the

limit sup¦x

1

, x

2

, ...¦. This is indeed true:

If (x

n

) is a monotone increasing sequence then

lim

n→∞

x

n

= sup

n≥1

x

n

.

To prove this let

L = sup

n≥1

x

n

.

The most extreme case is when L is −∞; as will become clear shortly, we need to

take care of this case separately. If L =−∞ each x

n

must be −∞. But in this case

lim

n→∞

x

n

=−∞,

since x

n

is just stuck at −∞. Now consider the other situation: L is not −∞. We

will show again that x

n

→L, i.e. we will show that for any neigborhood U of L, the

sequence (x

n

) falls eventually in U and stays there. So consider any neighborhood

U of L. Pick a point t of U to the left of L, i.e.

t < L and t ∈U.

Notes in Introductory Real Analysis 49

(Note that there would be no such t if L is −∞.) Since L is the least upper bound

of ¦x

1

, x

2

, x

3

, ...¦, the number t can’t be an upper bound, and so some x

n

0

is >t:

x

n

0

>t.

But recall that we are dealing with a monotone increasing sequence. Conse-

quently,

x

n

>t for all n ≥n

0

.

But recall that L is an upper bound of ¦x

1

, x

2

, ...¦; therefore,

all the x

n

are ≤t.

Those x

n

which are >t but ≤L are, of course, in the neighborhood U of L. Thus,

x

n

∈U for all n ≥n

0

.

This proves that L is indeed the limit of the sequence (x

n

).

In a similar way, if x

n

is a monotone decreasing sequence then

lim

n→∞

x

n

= inf

n≥1

x

n

.

2.24 The limit of a sequence is unique

Now we shall prove that a sequence can have at most one limit.

Consider a sequence (x

n

) and suppose both L and L

/

are limits of the sequence,

and L =L

/

. We will arrive at a contradiction. Since L and L

/

are distinct, they have

disjoint neighborhoods U and U

/

respectively. Since x

n

→L we know that x

n

∈U

eventually. But x

n

→L

/

, and so x

n

∈U

/

eventually. But this is impossible since U

and U

/

are disjoint and thus have no element in common.

Thus, a sequence which has a limit must have a unique limit.

2.25 Convergent sequences and Cauchy sequences

A sequence (x

n

) in R is said to be convergent if it has a limit and the limit is a real

number. If x

n

→L, and L ∈ R, we also say that the sequence x

n

converges to L.

Note that x

n

→L ∈ R if for any real r > 0 there is an n

0

∈ P such that

[x

n

−L[ < r

50 Ambar N. Sengupta

for all natural numbers n > n

0

.

This has a consequence: since the x

n

’s are all accumulating to L they are also

getting close to each other. More precisely, for any ε > 0 we can ﬁnd N ∈ P such

that

[x

n

−x

m

[ < ε for all natural numbers n, m > N.

To prove this, observe that there is some N ∈ P such that

[x

k

−L[ < ε/2

for all k ∈ P with k > N. But then for any n, m ∈ P with n, m > N we have

[x

n

−x

m

[ = [x

n

−L + L−x

m

[

≤ [x

n

−L[ + [L−x

m

[

= [x

n

−L[ + [x

m

−L[

< ε/2+ε/2

= ε.

A sequence (x

n

) in R which bunches up on itself in the sense above is said to

be a Cauchy sequence, i.e. if for any δ > 0 there is an N ∈ P such that

[x

n

−x

m

[ < ε for all natural numbers n, m > N.

2.26 Every Cauchy sequence is bounded

Consider a Cauchy sequence (x

n

) in R. Then, eventually the points of this se-

quence are at most distance 1 from each other; in fact, there is an N ∈ P such

that

[x

n

−x

m

[ < 1

for all natural numbers n, m > N. In particular, ﬁxing a particular j > N we have

x

n

∈ (x

j

−1, x

j

+1),

for all n ≥N. So

x

j

−1 < x

n

< x

j

+1

for all n ∈ ¦N +1, N +2, ...¦. Thus, at least from the (N +1)-th term on, the

sequence is bounded. But the terms not counted here,

x

1

, ..., x

N

Notes in Introductory Real Analysis 51

are just ﬁnitely many, and so they have a largest B and a smallest A among them.

Thus,

every x

n

is ≤max¦B, x

j

+1¦

and

every x

n

is ≥min¦B, x

j

−1¦

Thus the entire sequence is bounded.

2.27 Every Cauchy sequence is convergent

The completeness property of the real line R has an equivalent formulation:

Theorem 13 Every Cauchy sequence in R converges.

Let us prove this.

Consider a Cauchy sequence (x

n

) in R. We have seen that it is bounded. Thus,

a ≤x

n

≤b for all n ∈ P

for some real numbers a and b.

We know that a sequence always has a limit point in R

∗

. Let L be a limit point

of (x

n

). We will prove that L is a real number and x

n

→L.

First it is clear that L must also be in [a, b]. Therefore, L is not ∞ or −∞, and

is a real number.

Take any neighborhood

(L−δ, L+δ)

of L, where δ > 0 is a real number.

The sequence (x

n

) visits the neighborhood

(L−

δ

2

, L+

δ

2

)

inﬁnitely often. Now we also know that eventually, the terms of the sequence vary

from each other by < δ/2, i.e. there is some N ∈ P such that

[x

n

−x

m

[ < δ/2 for all n, m ∈ P with n, m > N.

52 Ambar N. Sengupta

We can choose some j > N such that

x

j

∈ (L−

δ

2

, L+

δ

2

)

because the sequence visits this neighborhood inﬁnitely often. Then

[x

n

−L[ = [x

n

−x

j

+ x

j

−L[

≤ [x

n

−x

j

[ + [x

j

−L[

<

δ

2

+

δ

2

= δ.

This means that

x

n

∈ (L−δ, L+δ)

for all n > N. Thus,

x

n

→L.

2.28 The rationals are countable

A set is set to be countable if it is ﬁnite or if its elements can be enumerated in a

sequence. Thus, S is countable if there is a sequence x

1

, x

2

, ... such that

S =¦x

1

, x

2

, x

3

, ...¦

For example, the even numbers form a countable set:

¦2, 4, 6, 8, ...¦

It may seem at ﬁrst that the set Z of integers is not countable, but we can certainly

lay out all the integers in a sequence:

0, 1, −1, 2, −2, 3, −3, ...

It is much harder to see that the rationals Q are also countable. This is what we

shall prove now.

We will constuct a sequence which enumerates all the rationals, i.e. a sequence

r

1

, r

2

, ... such that

¦r

1

, r

2

, r

3

, ...¦ = Q.

Notes in Introductory Real Analysis 53

Put another way, we will encode each rational by a unique natural number; thus

to each natural number n we would associate a rational r

n

, and in this way we will

cover all rationals.

There can be many such encoding schemes. Here is one: Take any rational

x ≥0 and write it as

p

q

where p and q are non-negative integers (of course q ≥1)

and q is the smallest possible denominator among all such ways of writing x. (For

example, 0.6 can be written as both 6/10 and 3/5, but we would take 3/5 as our

representation.) We know that every non-negative rational can be represented in

this way uniquely. Associate to this x the natural number

f (x) = 2

q

3

p

For example,

f (0) = f (0/1) = 2

1

3

0

= 2

f (1) = f (1/1) = 2

1

3

1

= 6, f (4/5) = 2

5

3

4

= 3281 = 2592.

This associates a unique natural number f (x) to each non-negative rational x. If

x ∈ Q is negative, x < 0, let us deﬁne f (x) to be just 5 times f (−x):

f (x) = f (−x) 5 if x < 0

Now we have labeled each rational by a unique natural number. To enumerate the

rationals in a sequence all we have to do then is to run this in reverse: let r

1

be

the rational number x with the smallest value for f (x) (so r

1

is in fact 0 because

f (0) = 2 is the lowest value possible for f ); let r

2

be the rational number with the

next lowest value for f (x), and so on. Thus, r

n

is the rational for which f (r

n

) is

the ﬁrst value of f (x) greater than f (r

1

), ..., f (r

n−1

). For example,

r

1

= 0

and

r

2

= 1 because f (1) = 2

1

3

1

= 6, the next value for f (x) after 2

Thus we have produced a sequence of distinct elements r

1

, r

2

, ... such that

¦r

1

, r

2

, ...¦ = Q

54 Ambar N. Sengupta

2.29 The real numbers are uncountable

We will prove that R is not countable, i.e. there is no sequence which touches on

all the real numbers.

In fact we will show that even the real numbers between 0 and 1 cannot be

enumerated in a sequence.

The argument used here is the celebrated diagonal method due to Georg Can-

tor (1845-1912). The strategy is to make a list of strings, then take the main

diagonal string, and then form a new string by altering each element of the diago-

nal.

Consider any sequence x

1

, x

2

, ... lying in (0, 1). Now each real number y in

(0, 1) can be expressed uniquely in the form

y = 0.y

1

y

2

y

3

... =

y

1

10

+

y

2

10

2

+

y

3

10

3

+ ,

where y

1

, y

2

, ... ∈ ¦0, 1, 2, 3, ..., 9¦ and we exclude all such representations which

use an inﬁnite string of 9’s at the end (for example, instead of 0.19999.... we would

use 0.20000....). Thus each of the numbers x

1

, x

2

, also has such an expansion:

x

1

= 0.x

11

x

12

x

13

. . .

x

2

= 0.x

21

x

22

x

23

. . .

x

3

= 0.x

31

x

32

x

33

. . .

.

.

. =

Now form a number

w = 0.w

1

w

2

w

3

...

as follows: take w

1

to be any number in ¦1, 2, ..., 8¦ other than x

11

; then choose

w

2

∈ ¦1, 2, ..., 8¦ other than x

22

, and so on. This way we make the n-th decimal

place of w different from the n-th decimal place of x

n

, for every n. Then w cannot

be equal to any of the x

n

, and w ∈ (0, 1). Thus the original sequence cannot

possibly have covered all of (0, 1). [The reason we excluded 0 and 9 from our

choices for w

n

was to avoid ending up at 0 or 1.]

Notes in Introductory Real Analysis 55

2.30 Connected Sets

Consider a subset S of R

∗

. If U is an open set then the part of U in S, i.e. U ∩S is

said to be open in S.

For example, in [2, 5] the set

[2, 4)

is open because, for example,

[2, 4) = (1, 4) ∩[2, 5]

Put another way, a subset J of S is said to be open in S if every p ∈ J has a

neighborhood N such that every point of N which is in S is in fact in J, i.e.

N∩J ⊂S

A set S is said to be connected if it cannot be split into two non-empty disjoint

pieces A and B each of which is open in S. The main result for connected sets is:

Theorem 14 A subset of R, or of R

∗

, is connected if and only if it is an interval.

56 Ambar N. Sengupta

Chapter 3

Integration

Modern integration theory was built from the works of Newton, Riemann, and

Lebesgue, but the origins of integration theory lie in the computation of areas

and volumes. The idea of measuring area of a curved region by lower and upper

bounds is present even Archimedes’ study of the area enclosed by a circle.

The area of a rectangle whose sides are 2 units and 3 units is 6 square units,

because if you tile the rectangle with unit squares you will need three rows of such

squares, each row having 2 tiles. By extension, the area of a rectangle of sides 1/4

unit and 1/5 unit should be 1/20 square units because we would need 20 of these

rectangles to cover a unit square. Thus, it is clear that the area enclosed by a

rectangle whose sides are a units and b units, where a and b are rational numbers,

is ab square units. Consider now a rectangle R whose sides a and b are possibly

not rational. Take any rationals a

/

, a

//

, and b

/

, b

//

with

a

/

< a < a

//

, and b

/

< b < b

//

.

Then a rectangle R

/

of sides a

/

by b

/

sits inside R, while a rectangle R

//

of sides a

//

by b

//

contains R. Thus, it makes sense to suppose that

area of R

/

≤area of R ≤area of R

//

which is to say:

a

/

b

/

≤area of R ≤a

//

b

//

It is intuitively clear, and not hard to prove, that ab is the unique real number that

lies between all the possible values of a

/

b

/

and a

//

b

//

. Thus,

area of R = ab

57

58 Ambar N. Sengupta

The rectangle R is made up of two congruent right angled triangles, and so each

of these would have area (ab)/2. This makes it possible to compute the areas of

all kinds of polygonal ﬁgures but cutting up these fgures into right angled trian-

gles. However, this strategy fails when we try to compute the area enclosed by a

circle. No amount of cutting would turn a disc into a ﬁnite number of right angled

triangles.

Archimedes computed area enclosed by a circle C by consider polygons P

/

and P

//

, where P lies inside the circle C and P

//

encloses the circle C; thus

area A

/

enclosed by P

/

≤area enclosed by C ≤area A

//

enclosed by P

//

Some computation shows that for any ε > 0 there are such polygons P

/

and P

//

such that

A

//

−A

/

< ε

This implies that there is a unique real number A which lies between the area of

all the polygons P

/

and the polygons P

//

. Clearly then this number should be the

area enclosed by the circle C.

Our development of the theory of the Riemann integrals is based on these

ideas.

3.1 Approaching the Riemann Integral

Consider a function

f : [a, b] →R

and think of its graph. Assume, for convenience of visualization, that f ≥0. Then

f speciﬁes a region which lies below its graph and above the x-axis. In general,

this is a region whose upper boundary, given by the graph of f , is curved. The

integral

Z

b

a

f

which is also, conveniently, written as

Z

b

a

f (x)dx

measures the area of this region.

The strategy used for computing the area is to cut up the region into vertical

slices. The area of each such slice is between the area of a larger ‘upper’ rectangle

Notes in Introductory Real Analysis 59

and a smaller ‘lower’ rectangle. Thus we should expect the actual area to be the

unique real number lying between these ‘upper rectangle’ areas and the ‘lower

rectangle’ areas.

3.2 Riemann Sums

We will work with functions on an interval

[a, b] ⊂R

] where < b.

A partition X of [a, b] is speciﬁed by a sequence

X = (x

0

, x

1

, ..., x

N

)

of points x

0

, x

1

, ..., x

N

∈ [a, b] with

a = x

0

< x

1

< < x

N

= b

Here N ∈ ¦1, 2, 3, ...¦.

We will often use the notation

∆x

j

def

= x

j

−x

j−1

(3.1)

to denote the length of the j-th interval

[x

j−1

, x

j

]

marked out by the partition X.

The width or norm of the partition X is the maximum size of these intervals:

[[X[[ = max

j∈¦1,...,N¦

∆x

j

(3.2)

Consider a function

f : [a, b] →R

on an interval [a, b] ⊂R, where a < b.

Let

M

j

( f ) = sup

x∈[x

j−1

,x

j

]

f (x) (3.3)

60 Ambar N. Sengupta

and

m

j

( f ) = inf

x∈[x

j−1

,x

j

]

f (x) (3.4)

Thus, if A

j

were the area of the region under the graph of f over [x

j−1

, x

j

] then

m

j

( f )∆x

j

≤A

j

≤M

j

( f )∆x

j

.

Our objective is to specify the actual area A under the graph of f , and this would

be the sum of the A

j

.

The upper Riemann sumU( f , X) is deﬁned to be

U( f , X)

def

=

N

∑

j=1

M

j

( f )∆x

j

(3.5)

and the lower Riemann sum L( f , X) is

L( f , X)

def

=

N

∑

j=1

m

j

( f )∆x

j

(3.6)

Note that

L( f , X) ≤U( f , X) (3.7)

We will show later that in fact every lower sum if less or equal to every upper sum,

i.e. L( f , X) is less or equal to U( f ,Y) for every partitions X and Y of [a, b].

Now consider a sequence

X

∗

= (x

∗

1

, ..., x

∗

N

)

obtained by picking a point from each interval [x

j−1

, x

j

]::

x

∗

j

∈ [x

j−1

, x

j

]

We shall indicate this by writing

X

∗

< X

The Riemann sum S( f , X, X

∗

) is

S( f , X, X

∗

) =

N

∑

j=1

f (x

∗

j

)∆x

j

(3.8)

Notes in Introductory Real Analysis 61

Note that the upper sum is ∞ if and only if one of the M

j

is ∞, and this occurs

if and only if the superemum of f is ∞:

U( f , X) = ∞ if and only if sup

x∈[a,b]

f (x) = ∞ (3.9)

However, U( f , X) cannot be −∞, because that would mean that at least one of

the M

j

is −∞ which can only be if f is equal to −∞ on that interval [x

j−1

, x

j

]

which contradicts the fact that f is real-valued. (Note that we are working with

non-empty intervals because x

j−1

< x

j

.)

Similarly,

L( f , X) =−∞ if and only if inf

x∈[a,b]

f (x) =−∞ (3.10)

and L( f , X) can never be ∞.

It is useful then to observe that the difference

U( f , X) −L( f , X)

is always deﬁned, i.e. we never have the ∞−∞ situation.

3.3 Deﬁnition of the Riemann Integral

Consider a function

f : [a, b] →R

on an interval [a, b] ⊂ R, where a < b. This function is said to be Riemann inte-

grable if there is a unique real number I lying between all the lower sums and all

the upper sums:

L( f , X) ≤I ≤U( f , X) (3.11)

for every partition X of [a, b]. This number I is the integral of f over [a, b] and

denoted

Z

b

a

f or

Z

b

a

f (x)dx (3.12)

The set of all Riemann integrable functions on [a, b] is denote

R [a, b] (3.13)

62 Ambar N. Sengupta

3.4 Reﬁning Partitions

We work with a function

f : [a, b] →R.

We use the notation

M( f , [s, t]) = sup

x∈[s,t]

f (x), and m( f , [s, t]) = inf

x∈[s,t]

f (x) (3.14)

Let

X = (x

0

, x

1

, ..., x

N

)

be a partition of the interval [a, b] and let X

/

be a partition obtained by adding one

more point x

/

to X. Let us say that x

/

lies in the j-th interval

x

/

∈ [x

j−1

, x

j

]

Thus, X

/

cuts up [x

j−1

, x − j] into two intervals

[x

j−1

, x

/

] and [x

/

, x

j

]

Let us compare the upper sums U( f , X) and U( f , X

/

). To this end, let us write

M

j

= M( f , [x

j−1

, x

j

]), M

/

j

= M( f , [x

j−1

, x

/

]), M

//

j

= M( f , [x

/

, x

j

])

It is important to observe that

M

j

≥M

/

j

, and M

j

≥M

//

j

(3.15)

These sums differ only in the contribution that comes from [x

j−1

, x

j

]:

U( f , X) −U( f , X

/

) = M

j

(x

j

−x

j−1

)

−

M

/

j

(x

/

−x

j−1

) +M

//

j

(x

j

−x

/

)

= (M

j

−M

/

j

)(x

/

−x

j−1

) +(M

j

−M

//

j

)(x

j

−x

/

)

≥ 0.

Thus

U( f , X

/

) ≤U( f , X), (3.16)

i.e. adding a point to a partition lowers upper sums.

If we keep adding points to the initial partition, we keep lower the upper sums.

Arguing similarly, it follows that adding points to a partition raises lower

sums.

This observation is important enough to state as a theorem:

Notes in Introductory Real Analysis 63

Theorem 15 Adding points to a partition lowers upper sums and raises lower

sums. Thus if X and X

/

are partitions of [a, b], with X

/

containing all the points of

X and some more, then

L( f , X) ≤L( f , X

/

) ≤U( f , X

/

) ≤U( f , X) (3.17)

for every function f : [a, b] →R.

Using this we can prove that upper sums always dominate lower sums:

Theorem 16 If f : [a, b] →R is a function on an interval [a, b] ⊂R, where a < b,

and if X and Y are any partitions of [a, b] then

L( f , X) ≤U( f ,Y) (3.18)

Thus, every lower sum is dominated above by every upper sum.

Proof. Here is a useful trick: we can combine the partitions X and Y to form a

partition Z which contains all points of X and all the points of Y as well. Therefore,

U( f , Z) ≤U( f ,Y)

and

L( f , X) ≤L( f , Z)

Stringing these together we have

L( f , X) ≤L( f , Z) ≤U( f , Z) ≤U( f ,Y)

and this gives the desired result. QED

We have used the fact that adding a point to a partition lowers upper sums and

raises lower sums. It will be useful to take a closer look at this and estimate by

how much the upper and lower sums move as points are added to the partition.

Recall the formula

U( f , X) −U( f , X

/

) = (M

j

−M

/

j

)(x

/

−x

j−1

) +(M

j

−M

//

j

)(x

j

−x

/

) (3.19)

where X

/

is the partition obtained from X by adding the single point x

/

to the j-th

interval [x

j−1

, x

j

] of the partition X = (x

0

, ..., x

T

). Now suppose instead of adding

just the one point x

/

, we add m distinct points y

1

, ..., y

m

to the interval [x

j−1

, x

j

] to

form the new partition X

/

. We label the new points y

i

in increasing order

y

1

< < y

m

.

64 Ambar N. Sengupta

For convenience of notation we write y

0

for x

j−1

and y

m+1

for x

j

. Thus

x

j−1

= y

0

< y

1

< < y

m

< y

m+1

= x

j

The intervals

[y

k−1

, y

k

]

make up the interval

[x

j−1

, x

j

]

Then:

U( f , X)−U( f , X

/

) =

m+1

∑

k=1

M( f , [x

j−1

, x

j

]) −M( f , [y

k−1

, y

k

])

(y

k

−y

k−1

) (3.20)

Now

0 ≤M( f , [x

j−1

, x

j

]) −M( f , [y

k−1

, y

k

]) ≤2[[ f [[

sup

,

where [[ f [[

sup

is supremum of [ f [ over the full original interval [a, b]. The interval

lengths y

k

−y

k−1

add up to ∆x

j

. Consequently,

U( f , X) −U( f , X

/

) ≤2[[ f [[

sup

∆x

j

≤2[[ f [[

sup

[[X[[ (3.21)

This provides an upper bound for how much the upper sum is decreased by addi-

tion of points all in one single interval of the original partition X.

If we add points to each of N

/

intervals to create a new partition X

/

then

U( f , X) −U( f , X

/

) ≤2N

/

[[ f [[

sup

[[X[[ (3.22)

Similarly,

L( f , X

/

) −L( f , X) ≤2N

/

[[ f [[

sup

[[X[[ (3.23)

Note that

N

/

≤N,

where N is the total number of new points added. Consequently, we have

Lemma 1 If

X = (x

0

, ..., x

T

)

is a partition of the interval [a, b] ⊂ R, where a < b, and X

/

is a partition of the

same interval containing all the points of X and possibly some more, then

U( f , X

/

) −L( f , X

/

) ≤U( f , X) −L( f , X) (3.24)

but the decrease in the value of U −L is at most 2N[[ f [[

sup

[[X[[:

[U( f , X) −L( f , X)] −

U( f , X

/

) −L( f , X

/

)

≤2N[[ f [[

sup

[[X[[ (3.25)

where N is the number of new points added.

Notes in Introductory Real Analysis 65

3.5 The Darboux Criterion

Now that we know that the upper sums dominate the lower sums, it is clear that

there will be a unique real number between them if and only if the sup of the lower

sums equals the inf of the upper sums:

Theorem 17 A function f : [a, b] →R is Riemann integrable if and only if

sup¦U( f , X) : all partitions X of [a, b]¦ = inf¦L( f , X) : all partitions X of [a, b]¦

(3.26)

The common value is

R

b

a

f .

Thus, the function f is Riemann integrable if and only if there is no ‘gap’

between the lower sums and upper sums. Thus, another equivalent formulation is

the Darboux criterion:

Theorem 18 T:Darbouxcrit A function f : [a, b] →Ris Riemann integrable if and

only if for any ε > 0 there is a partition X of [a, b] for which

U( f , X) −L( f , X) < ε (3.27)

This is an extremely useful result: virtually all of our results on integrability

will use the Darboux criterion.

Proof. Suppose f is integrable. Then there is a unique real number I which

lies between all upper sums and all lower sums. Take any ε > 0. Consider the

interval

[I, I +ε/2)

All the lower sums lie to the left, i.e. ≤I. All upper sums lie to the right (≥) of I.

So since I is the only real number lying between the upper sums and lower sums,

there must be at least one upper sum which lies in [I, I +ε/2). Thus, there is a

partition Y for which

I ≤U( f ,Y) < I +ε/2

Similarly, there is a partition Z for which

I −ε/2 < L( f , Z) ≤I.

Let X be the partition obtained by pooling together Y and Z. Then

I −ε/2 < L( f , Z) ≤L( f , X) ≤U( f , X) ≤U( f , Z) < I +ε/2 (3.28)

66 Ambar N. Sengupta

Since both L( f , X) and U( f , X) lie in

(I −ε/2, I +ε/2)

it follows that

U( f , X) −L( f , X) < ε. (3.29)

Conversely, suppose that for every ε >0 there is a partition for which (3.29) holds.

Suppose that I and I

/

are distinct real numbers which both lie between all upper

sums and all lower sums. Let

ε =[I

/

−I[

We know that there is a partition X satifying (3.29). Now I and I

/

both lie between

U( f , X) and L( f , X). So the difference between I and I

/

is < ε:

[I

/

−I[ < ε

But this contradicts the deifnition of ε taken above. Thus I and I

/

must be equal.

So there is a unique real number between all the upper sums and all the lower

sums. Thus, f is integrable. QED

3.6 Integrable functions are bounded

Suppose f : [a, b] →R is integrable. We will prove that f must be bounded.

Theorem 19 Every Riemann integrable function is bounded.

Proof. Consider

f : [a, b] →R

where [a, b] ⊂R and a < b. Assume that f ∈ R [a, b].

Let us suppose that f is unbounded. Then we will reach a contradiction.

Suppose, for example, that f is not bounded above, i.e.

sup

x∈[a,b]

f (x) = ∞.

Consider any partition

X = (x

0

, ..., x

N

)

Notes in Introductory Real Analysis 67

of [a, b]. Then there is some subinterval, say [x

j−1

, x

j

], on which f is not bounded

above, i.e.

M

j

= ∞.

But then the supper sum for f with this partition is also ∞:

U( f , X) =

N

∑

k=1

M

k

∆x

k

= ∞

Thus, every upper sum for f is ∞. Now let

I =

Z

b

a

f

and recall that this is the unique real number lying between all upper sums and all

lower sums. Then

L( f , X) ≤I < I +1 < ∞ =U( f , X) (3.30)

for every partition X. But then I +1 would also be a real number lying between

all upper sums and all lower sums. Thus, we have a contradiction. QED

3.7 Variation of a Function

For any non-empty subset S ⊂R the diameter of S is

diam(S) = sup¦a−b : a, b ∈ S¦ (3.31)

It is easy to believe and not hard to prove that the diameter of S is the difference

between supS and inf S:

diam(S) = supS−inf S (3.32)

Now consider a function f on an interval [s, t]. It will be useful to have a

measure of the ﬂuctuation of f over this interval.

The simplest such measure is given by the diameter of the range of f :

Var( f ) = diam(Rangeof f) (3.33)

More explicitly, the variation of f on the interval [s, t] is:

Var( f , [s, t]) = sup

x,y∈[s,t]

( f (x) − f (y)) (3.34)

68 Ambar N. Sengupta

It is also equal to

Var( f , [s, t]) = M( f , [s, t]) −m( f , [s, t]) (3.35)

Sometimes we may drop [s, t] and just write Var( f ):

Var( f ) = M( f ) −m( f ) (3.36)

We record the following algebraic facts about the variation of functions, which

will be very useful in proving corresponding facts about Riemann integration.

Lemma 2 For any functions f and g on an interval [a, b] ⊂ R with a < b, we

have:

(i) Var( f ) ≥0;

(ii) Var(k) = 0 if and only if k is constant;

(iii) the variation scales like length, i.e. the variation of a constant times a func-

tion is the absolute value of the function time the variation in the function:

Var(k f ) =[k[Var( f ) for any k ∈ R.

(iv) the variation satiﬁes the triangle inequality:

Var( f +g) ≤Var( f ) +Var(g) (3.37)

(v) the variation of the product of two functions is bounded above by the sum

of their variations, weighted by their sup-norms:

Var( f g) ≤[[ f [[

sup

Var(g) +[[g[[

sup

Var( f ) (3.38)

(vi) the variations in f and in g differ by at most the variation in [ f −g[:

Var( f ) −Var(g)

≤Var( f −g) (3.39)

(vii) if f is not equal to zero anywhere then

Var(1/ f ) ≤M([ f [

−1

)

2

Var( f ) (3.40)

Notes in Introductory Real Analysis 69

(viii) the variation of a function increases montonically with the interval of vari-

ation:

Var( f , [s, t]) ≤Var( f , [a, b]) if [s, t] ⊂[a, b] (3.41)

(ix) The variation in the absolute value of f is bounded by the variation in f :

Var([ f [) ≤Var( f ) (3.42)

(x) The variation of f is bounded above by twice [[ f [[

sup

:

Var( f ) ≤2[[ f [[

sup

(3.43)

(xi) If s, t, u ∈ [a, b] with s <t < u then

Var( f , [s, u]) ≤Var( f , [s, u]) +Var( f , [u, t]) (3.44)

(xii) If f is monotone on an interval [s, t] then

Var( f , [s, t]) =

f (t) − f (s) if f (t) ≥ f (s)

f (s) − f (t) if f (t) ≤ f (s)

(3.45)

We also record the following result on variations and partitions for continuous

functions:

Lemma 3 If f : [a, b] →R is continuous, where [a, b] ⊂R and a <b, then for any

ε > 0 there is a partition X = (x

0

, ..., x

N

) of [a, b] such that

Var( f , [x

j−1

, x

j

]) < ε

for every j ∈ ¦1, ..., N¦.

Proof. Since f is continuous on the compact set [a, b] it is uniformly continu-

ous. So for any ε > 0 there is a δ > 0 such that

[ f (x) − f (x

/

)[ < ε/2

whenever x, x

/

∈ [a, b] with [x −x

/

[ < δ. Take any partition X = (x

0

, ..., x

N

) with

all the intervals having length less than δ. Then, for each j ∈ ¦1, ..., N¦,

f (x) − f (x

/

) < ε/2

70 Ambar N. Sengupta

for every x, x

/

∈ [x

j−1

, x

j

] since these intervals all have length < δ. Consequently,

Var( f , [x

j−1

, x

j

]) ≤ε/2

and we are done. QED

The sup-norm bound on the variation has the following consequence:

Proposition 1 If f

1

, f

2

, ... : [a, b] →Rconverge uniformly to a function f : [a, b] →

R, then

lim

n→∞

Var( f

n

) = Var( f ) (3.46)

It will be convenient to use the notation

Var

j

( f )

def

= Var( f , [x

j−1

, x

j

]) (3.47)

The reason why we are interested in the variation is summarized by

Theorem 20 For any function f : [a, b] → R, where [a, b] ⊂ R and a < b, and

every partition X of [a, b] we have

U( f , X) −L( f , X) =

N

∑

j=1

Var

j

( f )∆x

j

(3.48)

Proof. Observe that

Var

j

( f ) = M

j

( f ) −m

j

( f ). (3.49)

Multiplying this by ∆x

j

and adding up over all j ∈ ¦1, ..., N¦ gives the result

(3.48). QED

3.8 The algebra R [a, b]

Recall that

R [a, b]

is the set of all Riemann integrable functions on [a, b].

Our main objective now is

Theorem 21 The set R [a, b] has the following properties:

(i) Every constant function belongs to R [a, b]

Notes in Introductory Real Analysis 71

(ii) If f , g ∈ R [a, b] then f +g ∈ R [a, b]

(iii) If f , g ∈ R [a, b] then f g ∈ R [a, b]

(iv) If f ∈ R [a, b], and f is never equal to zero and 1/ f is bounded, then 1/ f ∈

R [a, b].

(v) If f ∈ R [a, b] then [ f [ ∈ R [a, b] and

Z

b

a

f

≤

Z

b

a

[ f [ (3.50)

Properties (i)-(iii) say that R [a, b] is an algebra under pointwise addition and

multiplication of functions. It is important to note that the converse of (v) does not

hold, i.e. there are functions which are not Riemann integrable but whose absoulte

values are Riemann integrable.

Proof For a constant function k on [a, b] we have

L( f , X) = k(b−a) =U( f , X)

for every partition X and so k(b−a) is the unique real number lying between all

upper sums and all lower sums. Thus

Z

b

a

k = k(b−a)

Now supppose f , g ∈ R [a, b]. Let ε > 0. By the Darboux condition, there are

partitions Y and Z of [a, b] such that

U( f ,Y) −L( f ,Y) < ε/2 and U(g, Z) −L(g, Z) < ε/2

Let X be the partition obtained by combining Y and Z. Then, because upper sums

decrease and lower sums increase when points are added to a partition, we have

U( f , X) −L( f , X) < ε/2 and U(g, X) −L(g, X) < ε/2

Then, with X = (x

0

, ..., x

N

),

72 Ambar N. Sengupta

U( f +g, X) −L( f +g, X) =

N

∑

j=1

M

j

( f +g) −m

j

( f +g)

∆x

j

=

N

∑

j=1

Var( f +g, [x

j−1

, x

j

])∆x

j

≤

N

∑

j=1

Var( f , [x

j−1

, x

j

])∆x

j

+

N

∑

j=1

Var(g, [x

j−1

, x

j

])∆x

j

(by Lemma 2 (ii))

= U( f , X) −L( f , X) + U(g, X) −L(g, X) by (3.48)

< ε/2+ε/2 = ε.

Thus, by the Darbox criterion, f +g is Riemann integrable.

The other results follow in a similar way by applying other parts of Lemma

2. QED

Consider the function g on [0, 1] given by

g(x) =

1 if x is rational

−1 if x is irrational

Then g is not Riemann integrable, because no matter what partition X we take of

[0, 1] the upper sum is always

U(g, X) = 1

and the lower sum is always

L(g, X) =−1

But the absolute value of g is the constant function 1:

[g[ = 1

and so [g[ is Riemann integrable.

3.9 C[a, b] ⊂R [a, b]

Every continuous function on a compact interval is integrable. This is a central

result of integration theory.

Notes in Introductory Real Analysis 73

Theorem 22 For any interval [a, b] ⊂R, with a < b, the set C[a, b] of continuous

functions on [a, b] is contained in the set R [a, b] of Riemann integrable functions

on [a, b]:

C[a, b] ⊂R [a, b]

The proof has been presented in class. The key result used is Lemma 3 which says

that the variation of a continuous function can be controlled suitably to apply the

Darboux criterion for integrability.

It should be noted that discontinuous functions might also be integrable. In-

deed, any function which is dicontinuous at only ﬁnitely many points is integrable.

3.10 The Integral as a Non-negative Linear Func-

tional

We have seen that the set

R [a, b]

of all Riemann integrable functions on a compact interval [a, b] ⊂ R, with a < b,

is a linear space, i.e. that sums and constant multiples of Riemann integrable

functions are again Riemann integrable. Our next objective is to show that the

Riemann integral viewed as a function

R [a, b] →R : f →

Z

b

a

f (3.51)

is linear:

Theorem 23 For interval [a, b] ⊂R, where a <b, and any f , g ∈R [a, b], we have

Z

b

a

( f +g) =

Z

b

a

f +

Z

b

a

g (3.52)

and, for any k ∈ R,

Z

b

a

(k f ) = k

Z

b

a

f (3.53)

Thus, the Riemann integral

R [a, b] →R : f →

Z

b

a

f (3.54)

is a linear functional on the linear space R [a, b].

74 Ambar N. Sengupta

Proof Let f , g ∈ R [a, b], and k ∈ R. We have already seen, in Theorem 21,

that f +g and k f are also in R [a, b]. Now let

X = (x

0

, ..., x

N

)

be any partition of [a, b]. As usual, let Var

k

(h) denote the variation of a function h

over the interval [x

j−1

, x

j

]:

Var

k

(h) = sup

s,t∈[x

j−1

,x

j

]

(h(s) −h(t)) = M

j

(h) −m

j

(h) (3.55)

Then

U( f +g, X) =

N

∑

j=1

Var

k

( f +g)∆

j

x

≤

N

∑

j=1

Var

k

( f )∆

j

x +

N

∑

j=1

Var

k

(g)∆

j

x

= U( f , X) +U(g, X)

Similarly,

L( f +g, X) ≥L( f , X) +L(g, X) (3.56)

Thus, L( f +g, X) and U( f +g, X) are squeezed into the interval

[L( f , X) +L(g, X),U( f , X) +U(g, X)] (3.57)

Now let ε > 0. By the usual trick of combining partitions, there is a partition X of

[a, b] such that

U( f , X) −L( f , X) < ε/2

and

U(g, X) −L(g, X) < ε/2

So

U( f +g, X)−L( f +g, X) ≤U( f , X)+U(g, X) −[L( f , X) +L(g, X)] <ε (3.58)

which, by the Darboux criterion implies that f +g ∈ R [a, b]. (Of course, we have

just repeated the proof of Theorem 21 (ii).)

Notes in Introductory Real Analysis 75

Let

I( f ) =

Z

b

a

f , I(g) =

Z

b

a

g, (3.59)

I( f +g) =

Z

b

a

( f +g) (3.60)

Then I( f ) lies between U( f , X) and L( f , X), and I(g) lies between U(g, X) and

L(g, X). Consequently,

L( f , X) +L(g, X) ≤I( f ) +I(g) ≤U( f , X) +U(g, X)

Moreover,

L( f +g, X) ≤I( f +g) ≤U( f +g, X)

Putting all this together, we see that I( f +g) and the sum I( f ) +I(g) both lie in

the interval

[L( f , X) +L(g, X),U( f , X) +U(g, X)] (3.61)

and the width of this interval is < ε. Therefore, I( f ) +I(g) and I( f +g) differ by

less than ε. But ε is any positive real number. Therefore,

I( f +g) = I( f ) +I(g) (3.62)

Next, consider the function k f , where k ∈ R. Let

X = (x

0

, ..., x

N

)

be a partition of [a, b] such that

U( f , X) −L( f , X) <

ε

1+[k[

(3.63)

(The 1+ in the denominator is to avoid trouble if k happens to be 0.) Then

U(k f , X) −L(k f , X) =

N

∑

j=1

Var

j

(k f )∆x

j

=

N

∑

j=1

[k[Var

j

( f )∆x

j

(by Lemma 2 (iii))

= [k[ [U( f , X) −L( f , X)]

≤ [k[

ε

1+[k[

< ε.

76 Ambar N. Sengupta

Thus, the interval

[L(k f , X),U(k f , X)]

has width less than ε. The Darboux criterion implies that k f is integrable. Let

I(k f ) =

Z

b

a

(k f ) (3.64)

Now for k ≥0 we have

M(k f , [s, t]) = sup

x∈[s,t]

k f (x) = k sup

x∈[a,b]

f (x) = kM( f , [s, t]) (3.65)

and for k < 0 we have

M(k f , [s, t]) = sup

x∈[s,t]

k f (x) = k inf

x∈[a,b]

f (x) = km( f , [s, t]), (3.66)

because multiplying by a negative number reverses inequalities and transforms

sup into inf, and inf into sup. Thus, also

m(k f , [s, t]) = km( f , [s, t]) if k ≥0 (3.67)

m(k f , [s, t]) = kM( f , [s, t]) if k ≤0 (3.68)

(3.69)

Doing this for each interval [x

j−1

, x

j

], and mutiplying everything by ∆x

j

and

adding up, we see that I(k f ) lies in the interval

[kL( f , X), kU( f , X)] if k ≥0

and it lies in

[kU( f , X), kL( f , X)] if k ≤0

Now I( f ) lies between L( f , X) and U( f , X), and so kI( f ) lies in the same interval

mentioned above as I(k f ) does. Consequently,

I(k f ) −kI( f )

≤[k[ [U( f , X) −L( f , X)] < ε, (3.70)

as before. Now since ε > 0 is any positive real number we have

I(k f ) = kI( f ).

This completes the proof. QED

The Riemann integral is a non-negative linear functional in the sense that it

carried non-negative functions into non-negative numbers:

Notes in Introductory Real Analysis 77

Theorem 24 If [a, b] ⊂R with a < b, and f ∈ R [a, b] is non-negative, i.e. f (x) ≥

0 for all x ∈ [a, b] then

R

b

a

f ≥0:

f ∈ R [a, b] and f ≥0 imply that

R

b

a

f ≥0 (3.71)

Consequently, the integral is order-preserving

f , ∈ R [a, b] and f ≥g imply that

R

b

a

f ≥

R

b

a

g (3.72)

Proof. This is simply because if f ≥0 then all the lower sums are ≥0 and so the

integral

R

b

a

f , being ≥ all lower sums, is also ≥0.

Next, suppose f , g ∈ R [a, b] and f ≥g. Observe that

f −g = f +(−1)g

is also in R [a, b], and is, of course, ≥0. Thus,

f = ( f −g) +g

and so, by linearity,

Z

b

a

f =

Z

b

a

( f −g) +

Z

b

a

g

Now we have just shown that the ﬁrst term on the right is ≥0. Therefore,

Z

b

a

f ≥

Z

b

a

g. QED

The linear functional given by the Riemann integral is a bounded linear func-

tional on R [a, b] for the sup-norm in the following sense:

Theorem 25 For any compact interval [a, b] ⊂ R, with a < b, and for any f ∈

R [a, b] we have

Z

b

a

f

≤[[ f [[

sup

(b−a) (3.73)

Proof We know that if f ∈ R [a, b] then [ f [ ∈ R [a, b] and

Z

b

a

f

≤

Z

b

a

[ f [ (3.74)

78 Ambar N. Sengupta

Now the function [ f [ is bounded above by the constant [[ f [[

sup

(recall from Theo-

rem 19 that this is ﬁnite.) Therefore,

Z

b

a

[ f [ ≤

Z

b

a

[[ f [[

sup

=[[ f [[

sup

(b−a)

and we are done. QED

A useful but simple consequence of this result is:

Theorem 26 For any compact interval [a, b] ⊂R, with a <b, if f

1

, f

2

, ... ∈R [a, b]

converge uniformly to a function f on [a, b] then f ∈ R [a, b] and

lim

n→∞

Z

b

a

f

n

=

Z

b

a

f (3.75)

Proof First let us show that f ∈ R [a, b]. Let ε > 0. By uniform convergence, we

have an n ∈ P such that

[[ f

n

− f [[

sup

<

ε

4(b−a)

(3.76)

By integrability of f

n

we know that there is a partition X = (x

0

, ..., x

N

) of [a, b]

such that

U( f

n

, X) −L( f

n

, X) =

N

∑

j=1

Var

j

( f

n

)∆x

j

<

ε

2

(3.77)

Now

Var

j

( f

n

) −Var

j

( f )

≤ Var

j

( f

n

− f )

≤ 2[[ f

n

− f [[

sup

< 2

ε

4(b−a)

=

ε

2(b−a)

and so

Var

j

( f ) ≤Var

j

( f

n

) +

ε

2(b−a)

(3.78)

Therefore,

U( f , X) −L( f , X) =

N

∑

j=1

Var

j

( f )∆x

j

≤

N

∑

j=1

Var

j

( f

n

)∆x

j

+

ε

2(b−a)

N

∑

j=1

∆x

j

<

ε

2

+

ε

2

= ε.

Notes in Introductory Real Analysis 79

Thus, f ∈ R [a, b].

Next, we have

Z

f

n

−

Z

f

≤[[ f

n

− f [[

sup

(b−a) →0, as n →∞. QED

3.11 Additivity of the Integral

We will show that the integral of a function over an interval [a, b] is the sum of the

integrals of the function over [a, c] and [c, b] for any point c ∈ (a, b).

If

F : S →U

is a function, and T is a non-empty subset of S, then

F[T

denotes the restriction of F to the smaller domain T, i.e. F[T is the function

whose domain is T and whse values are given through F:

(F[T)(x) = F(x) for all x ∈ T (3.79)

Theorem 27 Let a, c, b be real numbers with a <c <b, and consider any function

f : [a, b] →R

which is intergable over [a, c] and over [c, b]. Then f ∈ R [a, b], and

Z

b

a

f =

Z

c

a

f +

Z

b

c

f (3.80)

Proof. Let ε > 0. By Darboux, there is a partition Y of [a, c] and a partition Z of

[c, b] such that

U( f [[a, c],Y) −L( f [[a, c],Y) < ε/2 (3.81)

and

U( f [[c, b], Z) −L( f [[c, b], Z) < ε/2 (3.82)

Now put together the points of Y and Z. This yields a partition X of the combined

interval [a, b]. Then,

U( f , X) =U( f [[a, c],Y) +U( f [[c, b], Z) ≥

Z

c

a

f +

Z

b

c

f (3.83)

80 Ambar N. Sengupta

and

L( f , X) = L( f [[a, c],Y) +L( f [[c, b], Z) ≤

Z

c

a

f +

Z

b

c

f (3.84)

Consequently,

U( f , X) −L( f , X) = U( f [[a, c],Y) +U( f [[c, b], Z) − [L( f [[a, c],Y) +L( f [[c, b], Z)]

= U( f [[a, c],Y) −L( f [[a, c],Y) + U( f [[c, b], Z) −L( f [[c, b], Z)

< ε/2+ε/2

= ε

Therefore, by Darboux, f ∈ R [a, b].

Now from (3.83) and (3.84) it follows that the sum

Z

c

a

f +

Z

b

c

f

lies between L( f , X) and U( f , X), and, of course, so does

R

b

a

f . Therefore,

R

c

a

f +

R

b

c

f and

R

b

a

f differ by less than ε.

Since ε > 0 is arbitrary, it follows that

Z

b

a

f =

Z

c

a

f +

Z

b

c

f . QED

Now we prove that if a function is ingtergable on an interval then it is inte-

grable on any sub-interval:

Theorem 28 If [a, b] ⊂ R, with a < b, and if s, t ∈ [a, b] with s < t then for any

function f ∈ R [a, b] we have f [[s, t] ∈ R [s, t].

Proof This is, as always, a matter of applying Darboux using one of the properties

of Var. Let ε > 0. Since f ∈ R [a, b] there is a partition Y of [a, b] such that

U( f ,Y) −L( f ,Y) < ε.

Add to Y the points s and t, in case they are not in Y, to obtain a partition

Z = (z

0

, ..., z

M

)

Notes in Introductory Real Analysis 81

of [a, b]. We know that this lowers the upper sum and raises the lower sum and so

U( f , Z) −L( f , Z) < ε.

Now let X be the partition of [s, t] obtained by taking the points of Z which are in

[s, t]. Then

U( f , Z) −L( f , Z) =U( f [[s, t], X) −L( f [[s, t], X) +

∑

j∈J

M

j

( f ) −m

j

( f )

∆x

j

(3.85)

where J consists of those j ∈ ¦1, ..., M¦ for which the interval [z

j−1

, z

j

] is not

contained in [s, t]. Therefore,

U( f , Z) −L( f , Z) ≤U( f [[s, t], X) −L( f [[s, t], X) < ε,

and we are done. QED

3.12 Monotone Functions are Riemann Integrable

We have made the remark that not every Riemann integrable function is continu-

ous. We will now prove that every monotone function is Riemann integrable on

any compact interval.

Let

f : [a, b] →R

be a monotone function, where [a, b] ⊂R and a < b. Let

X = (x

0

, x

1

, ..., x

N

)

be any partition of [a, b]. Then

U( f , X) −L( f , X) =

N

∑

j=1

Var

j

( f )∆x

j

≤

N

∑

j=1

Var

j

( f )

[[X[[

where

[[X[[ = max

1≤j≤N

∆x

j

,

82 Ambar N. Sengupta

is the maximum width of the intervals making up the partition.

Suppose for convenience that f is monotone non-decreasing, i.e.

f (x) ≤ f (y) for all x, y ∈ [a, b] with x ≤y

Then, by the property of Var for monotone functions given in (3.41) we have

Var

j

( f ) = f (x

j

) − f (x

j−1

) for all j ∈ ¦1, ..., N¦

Therefore, the sum of the variations over all the intervals is simply the variation

over the full interval:

N

∑

j=1

Var

j

( f ) = Var( f , [a, b]) (3.86)

The same conclusion holds even if f is monotone non-increasing, i.e. if

f (x) ≤ f (y) for all x, y ∈ [a, b] with x ≥y

Thus, in either case, we have

U( f , X) −L( f , X) = Var( f , [a, b])[[X[[ (3.87)

To make this less than any chosen ε > 0 all we have to do is take a partition X

with all the interval sizes less than

ε/[1+Var( f , [a, b])].

For example, we could divide [a, b] into N equal pieces, with N chosen large

enough that

b−a

N

<

ε

1+Var( f , [a, b])

.

Thus we have proved:

Theorem 29 If f is a monotone function on a compact interval [a, b] ⊂ R, with

a < b, then f ∈ R [a, b].

Notes in Introductory Real Analysis 83

3.13 Riemann Sums and the Riemann Integral

We have used the Archimedean strategy of capturing the value of the integral

between upper sums and lower sums. This approach led to a smooth development

of the central results of the theory. However, this method is not the most intuitive

in understanding concepts such as arc length. It is therefore useful to understand

the Riemann integral in terms of Riemann sums as well. This method is also

amenable to generalizations such as the notion of line integrals. Furthermore, the

Riemann sum approach motivates the construction of more advanced notions such

as the stochastic integral of Itô.

So consider a function

f : [a, b] →R

where [a, b] ⊂R with a < b. Let

X = (x

0

, ..., x

N

)

be any partition of [a, b]. Recall that the norm or width of X is the length of the

largest interval

[[X[[ = max

j

∆x

j

Now consider any sequence

X

∗

= (x

∗

1

, ..., x

∗

N

)

subordinate to X, i.e. with x

∗

j

∈ [x

j−1

, x

j

] for each j. We denote this by

X

∗

< X

Recall the Riemann sum

S( f , X, X

∗

) =

N

∑

j=1

f (x

∗

j

)∆x

j

(3.88)

Now

m

j

≤ f (x

∗

j

) ≤M

j

,

for each j, and so

L( f , X) ≤S( f , X, X

∗

) ≤U( f , X) (3.89)

If f is integrable, with I =

R

b

a

, then for any ε > 0 we can choose partition X such

that

U( f , X) −L( f , X) < ε.

84 Ambar N. Sengupta

Since both I and S( f , X, X

∗

) are squeezed in between the upper and lower sums,

it follows that

[S( f , X, X

∗

) −I[ < ε (3.90)

The following result is is often used to deﬁne the Riemann integral in alterna-

tive approaches to the theory.

Theorem 30 A function f : [a, b] →R is Riemann integrable if and only if there

is a real number I such that for any ε > 0 there is a δ > 0 such that

[S( f , X, X

∗

) −I[ < ε (3.91)

for every partition X of norm < δ and every X

∗

< X. In this case,

I =

Z

b

a

f

Proof. Suppose the given condition holds. Then there is a real number I such

that for any ε > 0 there is a δ > 0 for which the condition

[S( f , X, X

∗

) −I[ < ε/4 (3.92)

holds for all partitions X of [a, b] of width < δ and all X

∗

< X. Thus,

I −ε/4 <

N

∑

j=1

f (x

∗

j

)∆x

j

< I +ε/4

for every sequence X

∗

< X. Then, taking the supremum over all possible x

∗

1

in the

ﬁrst interval [x

0

, x

1

], we see that

M

1

∆x

1

+

N

∑

j=2

f (x

∗

j

)∆x

j

≤I +ε/4

and, taking the inﬁmum over all possible x

∗

1

in [x

0

, x

1

], we have

I −ε/4 ≤m

1

∆x

1

+

N

∑

j=2

f (x

∗

j

)∆x

j

Carrying this successively for j = 2, 3, ..., N, we conclude that

I −ε/4 ≤L( f , X) ≤U( f , X) ≤I +ε/4

Notes in Introductory Real Analysis 85

Consequently,

U( f , X) −L( f , X) ≤ε/2 < ε

and so, by Darboux, f ∈ R [a, b]. Moreover, since both I and the integral

R

b

a

f

are trapped in the interval [L( f , X),U( f , X)] whose width is ε it follows that I and

R

b

a

f differ by less than ε. But, ε is any positive real number. Thus,

I =

Z

b

a

f .

For the converse, suppose f ∈ R [a, b]. Let ε > 0. By Darboux, there is a

partition

Y = (y

0

, ..., y

N

)

of [a, b] such that U( f ,Y) and L( f ,Y) differ by less than ε:

U( f ,Y) −L( f ,Y) < ε/2 (3.93)

Now let

δ =

ε/2

1+2N[[ f [[

sup

. (3.94)

(Where we get this from will be clear later.) Consider any partition

X = (x

0

, ..., x

T

)

of [a, b] of norm less than δ:

[[X[[ < δ

We will compare U −L for X with that for Y and conclude that U −L for X is

indeed less than ε. Using our standard trick, let Z be the partition of [a, b] obtained

by combining X and Y. Then

U( f , Z) −L( f , Z) ≤U( f ,Y) −L( f ,Y) < ε/2 (3.95)

We also know by Lemma 1 that U −L for Z differs from that for X by at most

2N[[ f [[

sup

[[X[[, because at most N points were added to X to obtain Z. Thus, the

most U −L for X could be is

U( f , Z) −L( f , Z) +2N[[ f [[

sup

[[X[[ (3.96)

Thus,

U( f , X) −L( f , X) < ε/2+2N[[ f [[

sup

δ (3.97)

86 Ambar N. Sengupta

In (3.94) we chose δ just so this right side now works out to ε:

U( f , X) −L( f , X) < ε/2+ε/2 = ε (3.98)

Now take any X

∗

<X. Then the Riemann sum S( f , X, X

∗

) is sandwiched between

the lower sum L( f , X) and the upper sum U( f , X), and so it the integral

R

b

a

f .

Therefore, S( f , X, X

∗

) and

R

b

a

f both lie in the interval

[L( f , X),U( f , X)]

whose width is < ε. Thus,

S( f , X, X

∗

) −

Z

b

a

f

< ε (3.99)

We have shown that for any ε > 0 there is a δ > 0 such that (3.99) holds for any

partition X of width < δ and any X

∗

< X. QED

Bibliography

[1] Conway, John H. 2001. On Numbers and Games. (Second Edition) A.K.

Peters.

[2] Dedekind, Richard. 1901. Essays on the Theory of Numbers. The

Open Court Publishing Company. Online at http://www.gutenberg.org/

etext/21016

[3] Hilbert, David. 1956. Grundlagen der Geometrie. (Eight Edition, with

supplementary material and revisions by Paul Bernays.) H.G. Verlagsge-

sellschaft mbH, Stuttgart.

[4] Knuth, Donald.1974. Surreal Numbers. Addison-Wesley Professional.

[5] Rudin, Walter. 1976. Principles of Mathematical Analysis. (Third Edition.)

McGraw-Hill.

87

88 Ambar N. Sengupta

Appendix A

Question Bank

Set 1

1. Prove that there is no rational number whose square is 7.

89

90 Ambar N. Sengupta

2. Let ε be any positive real number.

(i) Show that there is an n

0

∈ N such that

1

n

2

< ε

for all n ∈ N with n ≥n

0

.

(ii) Let a be any real number. Show that there is some n

1

∈ N such that

2a

1

n

< ε

for all n ∈ N with n ≥n

1

.

Notes in Introductory Real Analysis 91

3. Suppose a is a positive real number with a

2

< 7. Show that

a+

1

n

2

< ε

for all n ∈ N large enough (i.e. there is some n

/

∈ N such that the preceding

inequality holds for all natural numbers n ≥n

/

).

92 Ambar N. Sengupta

4. Let

S =¦x ∈ R : x > 0andx

2

< 7¦.

(i) Show that S = / 0.

(ii) Show that S is bounded above in R.

Notes in Introductory Real Analysis 93

(iii) Let a = supS. Prove that a

2

≥7. (Hint: Suppose a

2

were less than 7.

Then use the result of Problem 3.)

(iv) Prove that a

2

is, in fact, equal to 7. Thus, there is a real number whose

square is 7.

94 Ambar N. Sengupta

Set 2

1. For the set B =¦−4, 8¦∪[1, 7) ∪[9, ∞), viewed as a subset of R

∗

:

(i) B

0

=

(ii) ∂B =

(iii) B

c

=

(iv) the interior of the complement B

c

is

(B

c

)

0

=

(v) (B

0

)

c

=

(vi) B =

(vii) ∂B =

(viii) ∂B

0

=

(ix) The interior of (B

0

)

c

is:

(x) The closure of B

c

is:

Notes in Introductory Real Analysis 95

2. Provide brief explanations/answers for the following:

(i) If U is an open set then U

0

=

(ii) For any set A, the interior of the interior of A is the interior of A, i.e.

(A

0

)

0

= A

0

.

(iii) The set [0, ∞) is closed as a subset of R.

(iv) If S is closed then its closure S is S itself, i.e. S = S.

(v) Is there a subset of R whose boundary in R is all of R?

96 Ambar N. Sengupta

3. Give an example of an open cover of (−1, 1) which does not have a ﬁnite

subcover.

Notes in Introductory Real Analysis 97

Set 5

1. Consider the partition

X = (.25, .5, .75, 1)

of [0, 1]. For the function f on [0, 1] given by

f (x) = x

2

for all x ∈ [0, 1]

(i) Work out the upper sumU( f , X)

(ii) Work out the Riemann sum S( f , X, X

∗

), where

X

∗

= (0.1, 0.3, 0.6, 0.8)

98 Ambar N. Sengupta

2. Consider the partition

X =

1

N

,

2

N

, ...,

N

N

**of [0, 1]. For the function f on [0, 1] given by
**

f (x) = x

2

for all x ∈ [0, 1]

(i) Show that

U( f , X) =

1

6

1+

1

N

2+

1

N

**[Hint: Use the sum formula 1
**

2

+2

2

+ +k

2

= k(k +1)(2k +1)/6.]

(ii) Show that

L( f , X) =

1

6

1−

1

N

2−

1

N

**Notes in Introductory Real Analysis 99
**

(iii) Assuming that f is integrable, prove that

Z

1

0

x

2

dx =

1

3

using the deﬁnition of the Riemann integral and the results of (i) and

(ii).

100 Ambar N. Sengupta

3. Consider the function g on [0, 1] given by

g(x) =

1 if x is rational

−1 if x is irrational

Prove that g is not Riemann integrable.

Notes in Introductory Real Analysis 101

4. Prove that, for a, b ∈ R with a < b, if f , g ∈ R [a, b], then f +g ∈ R [a, b].

102 Ambar N. Sengupta

Set 6

1. Write out in a complete and neat way the proof that every continuous func-

tion on a compact interval is Riemann integrable.

Notes in Introductory Real Analysis 103

2. For a function

f : [a, b] →R

we deﬁne the variation Var( f ) by

Var( f ) = sup

x,y∈[a,b]

( f (x) − f (y)) = M( f ) −m( f ),

where

M( f ) = sup

x∈[a,b]

f (x), andm( f ) = inf

x∈[a,b]

f (x).

Prove that for functions f and g on [a, b],

Var( f g) ≤M([ f [)Var(g) +m([g[)Var( f )

104 Ambar N. Sengupta

3. Explain the notion of the Riemann integral by clearly stating the deﬁnitions

of upper sums and lower sums, all properly explained in your own terms,

and then stating and explaining the deﬁnition of the Riemann integral. Work

out a simple integral in such a way as to illustrate the deﬁnition of the Rie-

mann integral. Clearly explain, in your own words, the Darboux criterion.

(Please note that every piece of notation you bring in must explained.)

Notes in Introductory Real Analysis 105

Test 1

1. Mark True or False:

(i) If S is a bounded subset of R then it contains a largest element

(ii) If S is a bounded, non-empty subset of R then it has a least upper

bound in R

(iii) The point ∞ is a boundary point of [1, ∞]

(iv) The point 1 is an interior point of [0, 1]

(v) The point 4 is a boundary point of [0, 4)

(vi) The set [0, ∞) is a closed subset of R

∗

(vii) The set [0, ∞) is a closed subset of R

(viii) Every real number is a boundary point of Q

(ix) The set ¦1, 2, 3¦ is compact

(x) The set (1, 5] is compact

106 Ambar N. Sengupta

2. Prove one of the following statements:

(i) The union of any collection of open sets is open.

(ii) The complement of a closed set is open.

(ii) If S is a non-empty subset of R

∗

then supS is in the closure S.

Notes in Introductory Real Analysis 107

3. State

(i) the Heine-Borel Theorem

(ii) the Bolzano-Weierstrass theorem.

108 Ambar N. Sengupta

4. State the deﬁnitions of the following:

(i) Limit point of a sequence (x

n

) in R

∗

.

(ii) Limit of a sequence (x

n

)in R

∗

.

(iii) A Cauchy sequence in R.

Notes in Introductory Real Analysis 109

Test 2

1. Mark True or False:

(i) A continuous function on any subset of R is bounded.

(ii) A continuous function on any closed subset of R is bounded.

(iii) A continuous function on a compact set is bounded.

(iv) A continuous function on a compact set attains a mimimum value at

some point in the set.

(v) If a sequence of functions converges to a function pointwise then it

converges uniformly.

(vi) If a sequence of functions converges to a function unformly then it

converges pointwise.

(vii) If f

n

: [0, 1] →R is a sequence of functions, and f

n

→ f , as n →∞,

pointwise, and if each f

n

is continuous at a point p ∈ [0, 1], then f is

continuous at p.

(viii) Every uniformly Cauchy sequence of functions converges uniformly.

(ix) Every uniformly convergent sequence of functions is uniformly Cauchy.

(x) The uniform limit of a sequence of continuous functions is continuous.

110 Ambar N. Sengupta

2. Suppose f : [0, 1] →Ris continuous, and f (p) >0 for some p ∈[0, 1]. Show

that there is a neighborhood U of p such that f (x) > 0 for all x ∈U ∩[0, 1].

Notes in Introductory Real Analysis 111

3. State

(i) the Intermediate Value Theorem

(ii) the Extreme Value theorem.

112 Ambar N. Sengupta

4. Show that the equation

x

3

+5x −2 = 0

has a solution in the interval (0, 1).

Notes in Introductory Real Analysis 113

Test 3

1. Consider a function f on an interval [a, b] ⊂R, where a < b,:

f : [a, b] →R.

(i) Let X = (a, b) be the simplest partition of [a, b]. Write down an ex-

pression for L( f , X), explaining your notation clearly. (5pts)

(ii) Now consider a partition X

/

which contains one addtional point t, i.e.

X

/

= (a, t, b),

where a < t < b. Write down an expression for L( f , X

/

), again ex-

plaining the notation clearly. (5pts)

114 Ambar N. Sengupta

(iii) Show that (5pts)

L( f , X

/

) ≥L( f , X).

Notes in Introductory Real Analysis 115

2. Consider a function f : [a, b] →R, where [a, b] ⊂R and a < b. Let X and Y

be partitions of [a, b] with Y containing all the points of X and some more.

Prove that (5pts)

U( f ,Y) −L( f ,Y) ≤U( f , X) −L( f , X)

116 Ambar N. Sengupta

3. If f ∈ R [a, b] and [c, d] ⊂[a, b] prove that f ∈ R [c, d]. (5pts)

Notes in Introductory Real Analysis 117

4. Mark TRUE or FALSE: (10pts)

a. For any function f : [a, b] → R and any partition X = (x

0

, ..., x

N

) of

[a, b],

U( f , X) −L( f , X) =

N

∑

j=1

Var

j

( f )∆x

j

,

where Var

j

( f ) = sup

x∈[x

j−1

,x

j

]

f (x) −inf

x∈[x

j−1

,x

j

]

f (x), and ∆x

j

= x

j

−

x

j−1

.

b. Every Riemann integrable function is bounded.

c. Every bounded function is Riemann integrable.

d. Every monotone function is Riemann integrable.

e. Every continuous function is Riemann integrable.

f. Every Riemann integrable function is continuous.

g. If f ∈ R [a, b] and f is never zero then 1/ f ∈ R [a, b].

h. For a bounded function on an interval [a, b] ⊂R, there is always a real

number which is ≥ all the lower sums and ≤ all the upper sums.

i. If [ f [ ∈ R [a, b] then f ∈ R [a, b].

j. If f ∈ R [a, b] then [ f [ ∈ R [a, b].

118 Ambar N. Sengupta

Cumulative Question Set

1. Consider the set

S = [2, 4) ∪(5, 6) ∪¦1, 8¦.

Mark True or False:

(i) 2 is an interior point of S.

(ii) 4 is an interior point of S.

(iii) 4 is a boundary point of S.

(iv) 9 is an isolated point of S.

(v) 1 is an isolated point of S.

(vi) S is an open set.

(vii) S is a closed set.

(viii) S is compact.

(ix) Every sequence in S has a subsequence which is convergent.

(x) Every sequence in S has a subsequence converging to a point in S.

Notes in Introductory Real Analysis 119

2. Let S be a non-empty subset of R.

(i) If t a real number with t < supS. Explain why there exists x ∈ S with

x >t.

(ii) Show that there is a sequence of elements s

n

∈ S such that s

n

→supS.

[Hint: Let U = supS. First assume that S is bounded above; in this

case U ∈ R. Now apply (i) to produce an s

n

using 1/n for t. If you

have time, try also the case U = ∞.]

120 Ambar N. Sengupta

3. Suppose f is continuous at a point p, and f (p) > 2. Prove that there is

a δ > 0 such that f (x) > 0 for all x in the domain of f which lie in the

neighborhood (p−δ, p+δ).

Notes in Introductory Real Analysis 121

4. Let f be a continuous function on a closed and bounded set S ⊂ R. Prove

that f reaches its maximum value on S, i.e. there is a point p ∈ S such that

f (p) = sup

x∈S

f (x). [Hint: Let U = sup

x∈S

f (x). Then choose a sequence

of points s

n

∈ S such that f (s

n

) →U. Apply Bolzano-Weierstrass.]

122 Ambar N. Sengupta

5. Let a, b ∈ R, with a < b. Let ( f

n

) be a sequence of functions on [a, b]. Mark

True or False:

(i) If there is a function f on [a, b] such that f

n

(x) → f (x) for all x ∈ [a, b]

then f

n

→ f uniformly.

(ii) If f

n

→ f uniformly then f

n

(x) → f (x) for all x ∈ [a, b].

(iii) If ( f

n

) is Cauchy in sup-norm then f

n

→ f uniformly, for some func-

tion f on [a, b].

(iv) If f

n

→ f uniformly and each f

n

is continuous then f is continuous.

(v) If f

n

(x) → f (x) for every x ∈ [a, b], and if each f

n

is continuous, then

f is continuous.

Notes in Introductory Real Analysis 123

6. Consider an interval [a, b] ⊂R. Mark true or false.

(i) C[a, b] ⊂R[a, b].

(ii) If f ∈ R[a, b] then

Z

b

a

f

≤

Z

b

a

[ f [

(iii) If f

n

∈ R[a, b], for n ∈ N, and f

n

(x) → f (x) for every x ∈ [a, b] then

R

b

a

f

n

→

R

b

a

f .

(iv) If X is a partition of [a, b], and f ∈ R[a, b], then

L( f , X) ≤

Z

b

a

f

(v) If f is a function on [a, b], and X and Y are partitions of [a, b] with

X ⊂Y then

U( f , X) <U( f ,Y)

2

Ambar N. Sengupta

Contents

Introductory Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . 1 Ordered Fields and The Real Number System 1.1 Ordered Fields . . . . . . . . . . . . . . . . . . . . . . . 1.1.1 Fields . . . . . . . . . . . . . . . . . . . . . . . . 1.1.2 Order Relations . . . . . . . . . . . . . . . . . . . 1.1.3 Ordered Fields . . . . . . . . . . . . . . . . . . . 1.1.4 The absolute value function . . . . . . . . . . . . 1.1.5 The Archimedean Property . . . . . . . . . . . . . 1.2 The Real Number System R . . . . . . . . . . . . . . . . 1.2.1 Hilbert Maximality and the Completeness Property 1.2.2 Completeness of R and measurement . . . . . . . Problem Set 1 . . . . . . . . . . . . . . . . . . . . . . . . The Extended Real Line and Its Topology 2.1 The extended real line . . . . . . . . . . . . . . . 2.2 Neighborhoods . . . . . . . . . . . . . . . . . . 2.3 Types of points for a set . . . . . . . . . . . . . . 2.4 Interior, Exterior, and Boundary of a Set . . . . . 2.5 Open Sets and Topology . . . . . . . . . . . . . 2.6 Closed Sets . . . . . . . . . . . . . . . . . . . . 2.7 Open Sets and Closed Sets . . . . . . . . . . . . 2.8 Closed sets in R and in R∗ . . . . . . . . . . . . 2.9 Closure of a set . . . . . . . . . . . . . . . . . . 2.10 The closure of a set is closed . . . . . . . . . . . 2.11 S is the smallest closed set containing S . . . . . 2.12 R∗ is compact . . . . . . . . . . . . . . . . . . . 2.13 Compactness of closed subsets of R∗ . . . . . . . 2.14 The Heine-Borel Theorem: Compact subsets of R 3 5 7 8 9 13 14 16 18 18 19 20 20 23 23 24 25 27 28 30 30 31 31 32 32 32 35 36

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

. . . . . . . . . .

2

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

.26 2. . . . . . . 3. . . . .8 The algebra R [a. . . .21 2. . . . . . . . . . . . .9 C[a. . . . . . . . . . . . .15 2. . . . . . . . . . . . . . . . . .4 2. . Connected Sets . . . . . . .16 2. .28 2. . . . . . . . . . . . . . . . . . . . . . . . .25 2. . . .18 2.17 2. . . . . . . . . . . . . 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Simple Examples of limits . . .23 2. . . . . . . .3 Deﬁnition of the Riemann Integral . .27 2. . .12 Monotone Functions are Riemann Integrable . .24 2. . . . . . . . . . . . . . . . . . . . .5 The Darboux Criterion .11 Additivity of the Integral . . . . . . . . . . 3. . . . . . . The sequence 1/n . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. . Limits points of a sequence . . . . .13 Riemann Sums and the Riemann Integral .6 Integrable functions are bounded . . . 3. . . . . . . .29 2. . . 37 38 39 39 41 42 43 44 48 49 49 50 51 52 54 55 57 58 59 61 62 65 66 67 70 72 73 79 81 83 86 89 3 Integration 3. . . . . . . . . . . The real numbers are uncountable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Sengupta . . . . . . . . . . . . . . . . . . . . . . . . . . 3. . . . . . .30 Sequences . . 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . 3. . . . . b] ⊂ R [a. . . . . . . . . .22 2. . . . . . . . . . . . . . . . . . . . . . . . . . . . . b] . . . . . . . . b] . 3. . .20 2. . . . . . . Limit points and suprema and inﬁma . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Every Cauchy sequence is convergent . . . . . . .7 Variation of a Function . .10 The Integral as a Non-negative Linear Functional 3. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . The sequence Rn . . Bibliography . . . . . . . . . . Convergent sequences and Cauchy sequences Every Cauchy sequence is bounded . . . . . . . . . . . . . . . .2 Riemann Sums . . . . . . . . . . The limit of a sequence is unique . . . . . . .19 2. . . . . . . . . . . . . . . . . . 3.1 Approaching the Riemann Integral . Limit of a sequence . . . . . . . . . . . . . . 3. Monotone Sequences . . . . . . . . . A Question Bank . . . . . . . . . . . Bolzano-Weierstrass Theorem . . . . . . .4 Reﬁning Partitions . . . Ambar N. . . . . . . . . . . . . . . . . The rationals are countable . .

it conforms to the classical geometric notion of a positive real number being speciﬁed by quantities greater than it and those less than it. I have chosen the latter. explaining the insights from several different approaches The reality of constraints of time makes (iii) the most convenient approach. at LSU in the Fall of 2006. Richardson were used. Math 4031. a choice has to be made whether to treat the Cauchy sequence point of view or the existence of suprema as fundamental. the Riemann integral of a function is the unique real number lying between the upper Riemann sums and lower Riemann sums. In studying the notion of completeness. The notes here do not include a chapter on continuous functions. I will update them time to time. In addition to these notes. There are several different ideologies that would guide the presentation of concepts and proofs in any course in real analysis: (i) the historical way (ii) the most natural way (iii) the most efﬁcient way (iv) a comprehensive way. The downside is that there is little possibility of conveying any insights or intuition. for which we followed the Richardson notes. This point of view also guides the choice of approach in the treatment of the Riemann integral. and perhaps the best example of this approach is Rudin’s Principles of Mathematical Analysis [5].Notes in Introductory Real Analysis 5 Introductory Remarks These notes were written for an introductory real analysis class. a set of notes by Professor L. These notes have not been proof read carefully. .

Sengupta .6 Ambar N.

the segment nAB (which is n copies of AB laid contiguously) exceeds the segment mCD if and only if the segment nPQ exceeds the segment mRS.Chapter 1 Ordered Fields and The Real Number System In this chapter we go over the essential. foundational. 7 . such as segments or planar regions or even angles. In Euclid’s Elements. In the discussion below we focus on segments. some multiple of any one of them exceeds the other. Then aAa pair PQ and RS if for any positive numbers n and m. The ratio AB CD is then essentially the equivalence class of all segment pairs which are in the same ratio as AB is to CD. symbolically EF>GH if EF is congruent to a segment GK. facts about the real number system. where K is some point between E and F. An important feature of this order relation is encapsulated in the archimedean axiom: given any two segments. as ratios of magnitudes. a segment EF is taken to exceed a segment GH. Euclid also deﬁnes the ratio XY/ZW to be greater than the ratio PQ/RS : XY PQ > ZW RS if they are unequal and if whenever mZW>nXY then also mRS>nPQ. Positive real numbers arose from geometry in Greek mathematics.

and absolute values. Dedekind’s method has beed generalized in a striking. and proceed to subsection 1. Such ratios are the rational numbers. the real number system may be constructed by starting with the empty set. ordered ﬁelds. and is denoted by m . and vastly more powerful way. Other ratios are irrational. .8 Ambar N. show that ratios of segments can be added and multiplied and. and the Euclidean construction procedures. say nAB. say mCD. constitutes the real number system R. The maximal such ﬁeld. Leaving aside the historical background. for example the set of all those rationals which exceed the given ratio. who shows how the Dedekind cut method can be applied to abstract sets leading to the construction of all real numbers as well as transcendentals and inﬁnitesimals. Euclid’s considerations suggest that a ratio of segments may be understood in terms of a set of rational numbers. is congruent to a whole multiple of CD. n It is readily checked that m p = n q if and only if qm = pn. The axioms of geometry. Sengupta A special case is that of commensurate segments: if a whole multiple of AB. an algebraic structure called a ﬁeld emerges (this is discussed at length by Hilbert [3]). constructing the natural numbers. 1. by Conway [1]. and then the real numbers by Dedekind’s method of identifying a real number with a splitting of the rationals into two disjoint classes with members of one class exceeding those of the other. when 0 and negatives are included.4.1 Ordered Fields In this section we deﬁne and prove simple properties of ﬁelds. Knuth’s novel [4] is an unusual and entertaining presentation of this construction. The reader wishing to move on to properties of the real numbers may skim the contents of the ﬁrst few subsections. then the ratio AB is CD rational. respecting the axioms of geometry pertaining to the order relation and congruence.1. In either case. then the rationals.

1 Fields A ﬁeld F is a set along with two binary operations Addition : F × F → F : (a. The commutative law holds for addition: a+b = b+a for all a.9) .1) 2. There is an element 0 ∈ F. b ∈ F (1. Every element a ∈ F has an additive inverse −a. called the reciprocal of a: aa−1 = 1 = a−1 a for all non-zero a ∈ F (1. b. b) → ab satisfying the following axioms: 1. The associative law holds for addition a(bc) = (ab)c for all a. c ∈ F (1. c ∈ F (1. for which a1 = a = 1a for all a ∈ F (1.2) (1.6) (1.Notes in Introductory Real Analysis 9 1. The associative law holds for addition a + (b + c) = (a + b) + c for all a.8) 8. for which a+0 = a = 0+a for all a ∈ F (1.7) 7. the unit or multiplicative identity element.4) 3. called the negative of a: a + (−a) = 0 = (−a) + a 4. Every non-zero element a ∈ F has an multiplicative inverse a−1 . the zero or additive identity element.3) (1. b. There is an element 1 ∈ F. b) → a + b and Multiplication : F × F → F : (a.5) 5. The associative law holds for multiplication 6.1.

Let us note a few simple consequences: Theorem 1 If x ∈ F is such that a + x = a for some a ∈ F and y ∈ F is such that by = b for some non-zero b ∈ F then x=0 and y = 1. The commutative law holds for multiplication: ab = ba 10. and some of the axioms may be deduced from others. (b + c)a = ba + ca for all a.11) 11. −0 = 0 and 1−1 = 1 . the additive identity and the multiplicative identity are unique. Moreover.10) for all a. we will just write a+b+c instead of a + (b + c). b. The element 1 is not equal to the element 0: 1=0 We have not attempted to provide a minimal axiom set. and abc instead of abc. For instance. In particular. Sengupta (1. b ∈ F Ambar N. the commutativity of addition can be deduced from the other axioms. The distributive law holds: a(b + c) = ab + ac. c ∈ F (1.10 9. Because of the associative laws.

First deﬁne 1a = a. Next. b and the product ab−1 as a = ab−1 b There are many other easy consequences of the axioms which we will use without comment. The multiplicative inverse a−1 is best written as the reciprocal: 1 = b−1 . where now 1 is the number one in Z. and b = 0. 3. 3..13) (1. def . The other claims follow from 0+0 = 0 and 1 · 1 = 1. −2. and (−a)(−b) = ab. −3. QED 11 Theorem 2 If a.Notes in Introductory Real Analysis Proof.12) and (b−1 )−1 = b. (−a)b = −ab.. 2. Multiplying by = b by b−1 shows that y is 1.}. Moreover. We can multiply any element a ∈ F by an integer as follows. 2.. −1. Adding −a to a+x = a shows that x is 0. (1. then −(−a) = a. 2a = a + a.. and the set of integers by Z = {0. . b ∈ F. 1. . We denote the set of natural numbers by P: P = {1.}.

y ∈ Z and all a ∈ F (1. Finally. Sengupta (1.14) (1.19) There is a multiplicative analog of this given by powers of elements. n ∈ Z. y ∈ Z and all a ∈ F x(a + b) = xa + xb for all x ∈ Z and all a.12 and. for example. inductively. The following facts can be readily veriﬁed: x(ya) = (xy)a for all x. The simplest example of a ﬁeld is the two-element ﬁeld Z2 = {0. however. 1}. b ∈ F (x + y)a = xa + ya for all x. with addition and multiplication deﬁned modulo 2.17) (1. valid for suitable a ∈ F and m. .15) (1. (n + 1)a = na + a for all a ∈ F and all n ∈ P. We also deﬁne a0 = 1 for non-zero a ∈ F and. If m is a positive integer and a ∈ F we deﬁne a1 = a and am+1 = am a. be concerned with ﬁelds which permit a consistent ordering of their elements. for any positive integer m and non-zero a ∈ F.18) (1. 1 am We will use without comment simple facts such as a−m = (a−1 )n = (am )n = amn . 1+1 = 0 in Z2 We will. Next deﬁne negative multiples by (−n)a = −(na) for all n ∈ P and all a ∈ F. 0a = a def Ambar N.16) where 0 is the integer 0.

x ≥ y means x = y or x > y If T is a subset of an ordered set S then an element u ∈ S is said to be an upper bound of T if t ≤ u for all t ∈ T (1.23) (1. For any x.22) (1.Notes in Introductory Real Analysis 13 1. (1. y) ∈ O We also write y>x to mean x < y. or x < y.2 Order Relations An order relation on a set S is a set O of ordered pairs (x.20) If there is a least such upper bound then that element is called the supremum of T : sup T = the least upper bound of T We deﬁne similarly lower bounds and inﬁmums: if l ≤ t for every t ∈ T then l is called a lower bound of T and inf T = the greatest lower bound of T Of course.21) . If x < y and y < z then x < z. It is convenient to adopt the convention that x < y means (x. the sup or the inf might not exist. or y < x. similarly. y) of elements of S satisfying the conditions O1 and O2 below. O2. It is also convenient to use the notation: x ≤ y means x = y or x < y and. y ∈ F exactly one of the following hold: x = y.1. The axioms of order are: O1.

y. Then (i) x > 0 if and only if −x < 0 (ii) For any non-zero x ∈ F we have x2 > 0 (iii) 1 > 0 (iv) For any r ∈ Z the element r1 ∈ F is > 0 if r is a positive integer and is < 0 if r is a negative integer (v) x > y holds if and only if x − y > 0 (vi) If x ∈ F and x > 0 then 1/x > 0 (vii) The product of two positive elements is positive (viii) The product of a positive and negative is negative (ix) The product of two negative elements is negative (x) If x > y and z < 0 then xz < yz. z ∈ F. the following hold: OF1. then xz < yz: x < y and z > 0 imply xz < yz for all x. y. If x.14 Ambar N. We have the following simple observations for an ordered ﬁeld: Theorem 3 Let F be an ordered ﬁeld.25) (1. z ∈ F and x < y then x + z < y + z: x < y implies x + z < y + z for all x. z ∈ F and x < y. Sengupta 1. (xi) If x > y then −x < −y (xii) If x > y > 0 then 1/x < 1/y (1. and if also z > 0.24) . y. If x < 0 we say that x is negative.1. z ∈ F OF2.3 Ordered Fields An ordered ﬁeld is a ﬁeld F with an order relation in which. in addition to the ﬁeld and order axioms stated above. y. If x. If x > 0 we say that x is positive.

however. and 1 = 0. This then leads to a copy of the integers inside F. where 1F is the unit element in F. it follows that 1 > 0. if −x < 0 then adding x to both sides shows that x > 0. if x < 0 then we know that −x > 0 and so x2 = (−x)(−x) > (−x)0 = 0.26) In particular. QED Observe that if x > 0 then 2x = x + x > x + 0 = x. and so 0 > −x. but we know that 1 > 0. on identifying n ∈ P with n1F . (ii) If x > 0 then. and 3x = 2x + x > 2x + 0 = 2x. x(1/x) would have to be < 0. and. proceeding inductively. Multiplying x > y by the positive element 1/(xy) gives 1/y > 1/x. it follows that 1/x cannot be zero. inside the ordered ﬁeld F we have a copy of the natural numbers 1. If 1/x < 0 then. by OF2.27) 0 < x < 2x < 3x < · · · < nx < (n + 1)x < · · · . 1/x > 0. Conversely. 3. (xi) If x > y then. (i) Suppose x > 0. we have x2 = xx > x0 = 0. and we can assume that Z⊂F (1. 2. We prove some of the results. 15 (iii) Since 1 is 12 . (xii) If x > y > 0 then xy > 0 and hence 1/(xy) > 0. On the other hand. adding −x − y shows that −y < −x. Thus.. (vi) Suppose x > 0. we have for all n ∈ P and x > 0 in F (1.Notes in Introductory Real Analysis Proof. . Then we have x + (−x) > 0 + (−x).. Since x(1/x) = 1.

The deﬁnition of |x| shows directly that | − x| = |x| ≥ 0 It is also useful to observe that |x| is the larger of x and −x We think of |a − b| as measuring the difference between a and b.29) .32) (1.30) x −x if x ≥ 0 if x ≤ 0 (1. Sengupta (1. n The set of all such ratios m/n is the set of rationals Q⊂F and is an ordered subﬁeld of the ordered ﬁeld F. b ∈ F we have: (i) the triangle inequality |a + b| ≤ |a| + |b| Equality holds if and only if a and b are both ≥ 0 or both ≤ 0.16 Going further.4 The absolute value function The absolute value | · | function in an ordered ﬁeld F is deﬁned by |x| = For instance. and n = 0. (1. |1| = 1.1.28) 1.31) for all x ∈ F (1. we have then the ratio m ∈ F. if m. and | − 1| = 1. n ∈ Z. We have then Theorem 4 For any a. Ambar N.

Therefore. |a||b| is ab or −ab. (−a)b. QED . Thus. Thus. equality holds if and only if a and b are either both ≥ 0 or both ≤ 0. |a| + |b| is greater or equal to a + b and −(a + b). whichever is ≥ 0. Consequently. Next. |a| + |b| is greater or equal to ±a + (±b). it is greater or equal to a + b and (−a) + (−b). Consequently. and so it is |ab|. Thus. and so |a| + |b| ≥ |a + b|. Equality holds if and only if |a| = a and |b| = b or |a| = −a and |b| = −b. We know that |x| is x or −x. (−a)(−b). 17 (1.33) Proof. and so |a − b| ≥ |a| − |b|.e. Recall that |x| is the larger of x and −x. Observe that |b − a| = |a − b|.33). i. one of the elements ab and −ab. the product |a||b| is one of the elements ab. using the triangle inequality we have |a − b| + |b| ≥ |a + b − b| = |a|. |a − b| is ≥ to both |a| − |b| and −(|a| − |b|). in particular. Since the larger of the latter is |a| − |b| . Interchanging a and b yields: |b − a| ≥ |b| − |a|. whichever is ≥ 0.Notes in Introductory Real Analysis (ii) the absolute values differ by at most the difference in a and b: |a| − |b| ≤ |a − b| (iii) |ab| = |a||b|. a(−b). we have proved (1.

s (1. 3. . We will assume that it is an ordered ﬁeld in which the completeness axiom of completeness holds. 2.18 Ambar N.34) and we are done..} The ﬁeld Q of rationals is clearly archimedean: Theorem 5 The ordered ﬁeld Q of rationals is acrhimedean. y > 0.. Sengupta 1. Proof. Simply choose n for which nx is > w. . Then nx = qr(p/q) + p/q > pr ≥ r = y. Now suppose x. with x > 0. it is essential to actually construct such a system so as to be sure that there is no hidden contradiction between the axioms. there exists a multiple of x which exceeds y: nx > y for some n ∈ {1. n Proof. y ∈ Q. but we shall not describe a construction in these notes.1. If y ≥ 0 then we have 1x > y and we are done.5 The Archimedean Property An ordered ﬁeld F is said to have the archimedean property if for any x. Needless to say. r. y ∈ F. s p q r y= . q. Then x= where p. s ∈ P. Take n = qr + 1.2 The Real Number System R We shall work with the real number system in an axiomatic way. QED In an archimedean ordered ﬁeld there are no inﬁnities. Let x. QED 1. and there are also no inﬁnitesimals: Theorem 6 If F is an archimedean ﬁeld then for any w > 0 in F there is an n ∈ P such that 1 w < x. with x > 0.

The completeness property can be formulated equivalently as: Every non-empty subset of R which has an upper bound has a least upper bound.Notes in Introductory Real Analysis 19 1. and if every element of L is ≤ every element of U then there is a unique element x ∈ R which lies between L and U: l ≤ x ≤ u for all l ∈ L and all u ∈ U. A crucial fact about R is the completeness property: If L and U are non-empty subsets of R such that every element of L is ≤ every element of U.36) (1.2. one can add to these axioms the further requirement that this ﬁeld is maximal in the sense that it cannot be embedded inside any larger archimedean ordered ﬁeld.1 Hilbert Maximality and the Completeness Property As we have mentioned before. It turns out then that any such ordered ﬁeld is isomorphic to any other. and thus there is essentially one such ordered ﬁeld. (1. . (1.37) The completeness property implies the archimedean peoperty: Theorem 7 If in an ordered ﬁeld the property (1. as formalized through the axioms of Hilbert. with two special points O and I marked on it.35) It is useful to view the real numbers geometrically. Consider a line. produces an archimedean ordered ﬁeld. This property is also often expressed as: If R is partitioned into two disjoint subsets L and U whose union is R. and the point O itself should be thought of as 0. For any point P on the line on the same side of O as P we think of the ratio OP/OI as a positive real number. Points on the other side from I correspond to negative real numbers. The completebess property says that there are no ‘gaps’ in the line. To complete the story. This ordered ﬁeld is the real number system R. the structure of Euclidean geometry.37) holds then the archimedean property also holds. then there is a real number m which lies between L and U: l ≤ m ≤ u for all l ∈ L and all u ∈ U.

that there are polyedra in three dimensions which have equal volumes (as deﬁned by requirements of ‘upper’ and ‘lower’ approximations) which cannot be decomposed into congruent pieces. However. For a truly useful theory of measure. The simplest measurement problem is to devise a measure of sets of points in the plane which are made up of a ﬁnite collection of segments constructed by Euclidean geometry. Moving up a dimension. a real number was. However. as we shall see later. of course. Two such sets should have the same measure if they can be decomposed into a ﬁnite collection of congruent segments. This. can work with a ﬁeld which is larger than the rationals but smaller than R. 2. For this we would not need the full complete system R of real numbers. between any two distinct elements there is at least one other element. implicitly. this method lives on in the Riemann integral. Prove that in any ordered ﬁeld.2. along with. and completeness is not essential. with requirements on intersections of lines and circles. it was shown by Max Dehn in 1900. understood in terms of all rationals which exceeded it and all rationals below it. The modern understanding and point of view on completeness grew out of the work of Richard Dedekind [2]. Archimedes and others computed areas of curved regions by such upper and lower approximations. Prove that in any ordered ﬁeld. the traditional axioms of Euclidean geometry.2 Completeness of R and measurement Even in Euclid’s geometry. the completeness of the number system is essential. In modern calculus. Sengupta A proof of this is outlined in an exercise below. with the task of measuring areas of polygonal regions constructed by Euclidean geometry. . in resolving Hilbert’s Third Problem. between any two distinct element lie inﬁnitely many elements.20 Ambar N. Capturing a real number between upper approximations and lower approximations proves to be very useful. Problem Set 1 1. one could still get away with a less-than-complete system of numbers. makes it absolutely essential to work with a notion of measure that goes beyond simply decomposing into geometrically congruent ﬁgures. the utility of measuring areas of curved regions even in two dimensions. 1.

show that between any two distinct elements there is a rational element. Use this to produce an element of S greater than u. Prove that the completeness property (1. Suppose x ∈ R is ≥ all elements of a non-empty subset S ⊂ R. all real numbers which are < some element of S. and let S = {nx : n ∈ P}. Prove that the completeness property implies the archimdean propery. y ∈ F. Then y is. Show that there exists y > x which is also not an upper bound of S. Let x.36). 9. the least upper bound of S. 7. with x > 0. an upper bound for S. such as R. If z ∈ F show that there is a rational multiple qz of z for which x < qz < y.e. 5. Consider the element u − 2 x. Show that x is the least upper bound of S. 4. then we can take U to be the set of all upper bounds of S and L to be the complement.] 8. y ∈ F with x < y.36). Let 1 u = sup S.37) implies the property (1. This. [Hint: Suppose (1. Let x ∈ F be provided by (1.36) implies the completeness property (1. Suppose L and U are non-empty subsets of R such that (i) every element of L is ≤ every element of R.37) holds. Prove that (1. is not an upper bound. being < u. Prove that there is a unique real number which is ≤ all elements of U and ≥ all elements of L. and (ii) for any ε > 0 there is an element l ∈ L and an element u ∈ U with u − l < ε. Suppose that no positive integral multiple of x exceeds y. Let F be an archimedean ordered ﬁeld.37) holds. and let x. i. thus reaching a contradiction. If S ⊂ R is a non-empty set which is bounded above. [Hint: Suppose F is an ordered ﬁeld in which (1. by assumption. .Notes in Introductory Real Analysis 21 3.37). If F is an archimedean ordered ﬁeld. 6. Suppose S is a non-empty subset of R and x ∈ R is not an upper bound of S.] 10. Explain why this implies x ≥ sup S.

Sengupta .22 Ambar N.

The primary reason for introducing them is to simplify the statements of several theorems. and a smallest element −∞. For technical purposes. No metaphysical meaning need be attached to these inﬁnities. instead of just R.1 The extended real line The extended real line is obtained by a largest element ∞. We deﬁne addition on R∗ as follows: x + ∞ = ∞ = ∞ + x for all x ∈ R∗ with x > −∞ y + (−∞) = −∞ = (−∞) + y for all y ∈ R∗ with y < ∞.1) Here ∞ and −∞ are abstract elements. 23 (2.4) . to the real line R: R∗ = R ∪ {−∞. it will be convenient to extend the real line R by adjoining to it a largest element ∞ and a smallest element −∞. 2.3) (2. ∞} (2.Chapter 2 The Extended Real Line and Its Topology In this chapter we study topological concepts in the context of the real line. Note that ∞ + (−∞) is not deﬁned. We extend the order relation to R by declaring that −∞ < x < ∞ for all x ∈ R (2.2) Much of our work will be on R∗ .

A typical neighborhood of 0 is of the form (−ε. ∞] is a neigborhood of ∞. ε) for any positive real number ε.e. s) = {x ∈ R∗ : x < s} (2. there is no useful or consistent deﬁnition for it. the neighborhood consists of all points distance less than δ from p: (p − δ. (5. 1. Thus. ∞} is a ray of the form (t. p + δ) where δ > 0 is any positive real number.5) whenever either side of these equations holds (i.2. (2. if one side is deﬁned then so is the other and the equality 2.2 Neighborhoods A neighborhood of a point p ∈ R is an interval of the form (p − δ.24 i. For example. p + δ) = {x ∈ R : |x − p| < δ}. Sengupta x + (y + z) = (x + y) + z. Ambar N. For example. ∞] = {x ∈ R∗ : x > t} with t any real number. A neighborhood of ∞ in R∗ = R ∪ {−∞.e.6) . (1.8) is a neighborhood of 2. The following algebraic facts continue to hold in R∗ : x + y = y + x. A neighborhood of −∞ in R∗ is a ray of the form [−∞.

∞] Exerise Give examples of disjoint neighborhoods of (i) 2 and −4 (ii) −∞ and 5 (iii) ∞ and −∞ (iv) 1 and −1 2.7) The points 2 and ∞ have disjoint neighborhoods. 3) of 2. and we can form the neighborhood (2. Consider a set A point p ∈ R∗ is said to be an interior point of S if it has a neighborhood U lying entirely inside S.e. For example. with U ⊂ S. and so U ∩V is just the smaller of the two neighborhoods. Here is a simple but fundamental observation: Distinct points of R∗ have disjoint neighborhoods.3 Types of points for a set S ⊂ R∗ .5 lying entirely inside (2. 3 and 5 have the neighborhoods (2.Notes in Introductory Real Analysis where s ∈ R.5. 4) 25 Observe that if U and V are neighborhoods of p then U ∩V is also a neighborhood of p. 4) of 3 contains 2. and if q ∈ N then q had a neighborhood lying entirely inside N. For example. the neighborhood (2. Observe also that if N is a neighborhood of a point p. i. 4). 5) and (12. such as (−1. .5. This is called the Hausdorff property of R∗ . 5.5) (2. An example is [−∞. 4) and (4. In fact. either U contains V as a subset of vice versa.

Next consider the set {3} ∪ (5. 7). because it doesn’t exist as far as R is concerned. 8} ∪ [9. In the example above. points −5. . −5) ∪ {2. i. Thus p is a boundary point of S if every neighborhood of p intersects both S and Sc . For example. U ⊂ Sc . 9} ∪ [6. 4) ∪ {5. ∞]. ∞. decide which of the following are true and which false: (i) −6 is an interior point (T) (ii) 6 is an interior point (F) (iii) 9 is a boundary point (T) (iv) 5 is an interior point (F) Exerise For the set B = [−∞. 5] ∪ {6. 5. 8} ∪ [4. and −∞ are exterior to E. the boundary points of E are −4. decide which of the following are true and which false: (i) −6 is an interior point (ii) −5 is an interior point (iii) 5 is a boundary point (iv) 4 is an interior point (v) 7 is a boundary point.26 For example. Example For the set A = [−∞. Sengupta the points −2. 5. 6. 5. and ∞. The point ∞ is also an interior point of E. 11 are interior points. 7). 3. 8. A point p is an exterior point if it has a neighborhood U lying entirely outside S. for the set E above. for the set E = (−4. It is important to observe that if we work with the real line R instead of the extended line R∗ then we must exclude ∞ as a boundary point. A point which is neither interior to S nor exterior to S is a boundary point of S. 9. 7.e. Ambar N. ∞) The boundary points are 3.

∞] (2. Thus ∂S = ∂Sc .Notes in Introductory Real Analysis 27 2. 5) ∪ (5.4 Interior. Example For the set A = [−∞. The set of all boundary points of S is denoted ∂S and is called the boundary of S. 9} ∪ [6. 7). 5.9) . 6) ∪ (7. Exterior. The interior of the entire extended line R∗ is all of R∗ . 9. (i) A0 = [−∞. Thus. and Boundary of a Set S0 The set of all interior points of a set S is denoted and is called the interior of S. and we shall denote it Sext . ∞] (iv) the interior of the complement Ac is (Ac )0 = (4. 4) ∪ (6. 4) ∪ {5. 7. 5) ∪ (5. So / ∂R∗ = 0. The set of all points exterior to S is the exterior of S.8) Recall that a point p is on the boundary of S if every neighborhood of the point intersects both S and Sc . But this is exactly the condition for p to be on the boundary of Sc . 6} (iii) Ac = [4. 9) ∪ (9. 7) (ii) ∂A = {4. the whole extended line R∗ is split up into three disjoint pieces: R∗ = S0 ∪ ∂S ∪ Sext (2. 6) ∪ [7. 9) ∪ (9.

∞] is open. . a set S is open means that S0 = S. it doesn’t have any boundary points. (i) B0 = (ii) ∂B = (iii) Bc = (iv) the interior of the complement Bc is (Bc )0 = 2. Exerise For the set B = {−4. For example. / Moreove. 9) is open. the empty set 0 is open. ∞}. The entire extended line R∗ is open. 3) ∪ (5. 8} ∪ [1. when viewed as a subset of R∗ . 7) ∪ [9. ∞) the boundary of G. (2. 4] is not open. Thus.5 Open Sets and Topology We say that a set is open if it does not contain any of its boundary points. Sengupta But if we decide to work only inside R then the boundary of G is just {3}. because it has no boundary points. because. again. because the point 4 is a boundary point. is ∂G = {3. ∞). Notice then that every point of an open set is an interior point. Ambar N.28 For the set G = (3. But (3. Also (4.

Then p is in both A and B. Take any point p ∈ A ∩ B. Since p ∈ A and since A is open. This type of argument works for any ﬁnite number of open sets. Thus: The intersection of a ﬁnite number of open sets is open. Now consider two open sets A and B. In other words. Exerise Check that the intersection of the sets (4. 6) is open. it becomes clear that the union of open sets is an open set. S is made up of a union of neighborhoods. B. The collection of all open subsets of R is called the topology of R. The set of all open subsets of R∗ is called the topology of R∗ . Thus. We will show that A ∩ B is open. Viewed in this way. there is a neighborhood U of p which is a subset of A: U ⊂ A.C. . A ∩ B is open.Notes in Introductory Real Analysis 29 Thus for an open set S each point has a neighborhood contained entirely inside S. and hence so is their intersection. Consequently. A ∩ B ∩C is open. ∞) and (−3 5) and (2. Thus every point in A ∩ B has a neighborhood lying inside A ∩ B. Similarly there is a neighborhood V of p which is a subset of B: V ⊂B But then U ∩V is a neighborhood of p which is a subset of both A and B: U ∩V ⊂ A ∩ B. The intersection A ∩ B ∩C can be viewed as (A ∩ B) ∩C But here both A ∩ B and C are open. Now consider three open sets A.

But [4. viewed as a subset of R it is closed. The set [3. Thus. Thus. So we need to be careful in deciding what is close and what isn’t: a set may be closed viewed as a subset of R but not as a subset of R∗ . But this means that Sc is closed. ∞] is closed. 2. which means that Sc contains all its boundary points.6 Closed Sets A set S is said to be closed if it contains all its boundary points. if a set is open then its complement is closed. / The empty set 0 is also closed. 8] ∪ [9. In other words.30 Ambar N. Sengupta 2. for S to be open we must have ∂(Sc ) ⊂ Sc . The full extended line R∗ is closed. The converse is also true: if a set is closed then its complement is open. 5) is not closed because the boundary point 5 is not in this set. . But. ∞) is not closed (as a subset of R∗ ) because the boundary point ∞ is not inside the set. [4. If S is open then its boundary points are all outside S: ∂S ⊂ Sc . But recall that the boundary of S is the same as the boundary of the complement Sc . Thus. / Note that the sets R∗ and 0 are both open and closed.7 Open Sets and Closed Sets Consider a set S ⊂ R∗ . S is closed if ∂S ⊂ S Thus.

in contrast. For example. 31 Exercise. Check that its complement if closed. Now every point in R∗ is a boundary point of Q: ∂Q = R∗ .10) .) Let us see what the closure of Q is. which it contains. It is useful to think of the closure S = S ∪ ∂S in this way: A point p is in S if and only if every neighborhood of p intersects S. ∞]. This is because in R it has only the boundary point 3. The closure of (3. in R∗ the point ∞ is also a boundary point and is not in the set. 5) is [3.8 The set Closed sets in R and in R∗ [3.9 Closure of a set The closure of a set S is obtained by throwing in all its boundary points: S = S ∪ ∂S Of course. ∞) is [3. ∞]. Show that its complement is open. Consider the closed set [4. the closure of (3. Consider the open set (1. ∞) is [3. Exercise. ∞). (2. 2. if S is closed then its closure is itself. 2. but is not closed in R∗ . Thus. in R the closure of (3. ∞) is closed in R. 5]. 5).Notes in Introductory Real Analysis Theorem 8 A subset of R∗ is open if and only if its complement is closed. There is no such distinction for open sets. (But. because any neighborhood of any point contains both rationals (points in Q) and irrationals (points outside Q). when working with closed sets it is important to bear in mind the distinction between being closed in R and being closed in R∗ .

But K is closed. For example. Take any boundary point p ∈ ∂S. 2. 8). (8. Since q ∈ S it follows that W contains a point of S.11 S is the smallest closed set containing S S⊂K Consider any closed set K with Take any p ∈ S.e. An open cover of R∗ is a collection U of open sets whose union covers all of R∗ . Thus every neighborhood of p intersects K. ∞]} is an open cover of R∗ . an open cover of R∗ is a set U of open subsets of R∗ such that every point x ∈ R∗ falls inside some open set U in the collection U . for each x ∈ R∗ there exists U ∈ U such that x ∈ U ∈ U. 15). {[−∞. N contains a point of S. (7. This means p ∈ S. Thus. (−1. we have the language and tools to state and prove it now and shall do so. Thus. Choose (as we may) a neighborhood W of q lying entirely inside N. . We have to show that p is actually in S. and so in either case p is in K. and so S is closed. We will show that S is a closed set. So. S⊂K What we hae shown then is that S is the smallest closed set containing S as a subset. More formaly. 2. So we have seen that every neighborhood of p contains a point of S. Then N contains a point q in S.12 R∗ is compact There is a special property of R∗ whose full signiﬁcance is best appreciated at a later stage. 9) ∪ (11. we have shown that every boundary point of S is in S. However. Sengupta 2. i.10 The closure of a set is closed Consider the closure S of a set S.32 Ambar N. and hence also K. Now let N be any neighborhood of p. either p is in the interior of K or on its boundary. 5). Thus. Then every neighborhood of p intersects S. It is best to draw a picture for yourself to make this clear.

This is a collection of open sets such that every point of R∗ falls inside some set in the collection. −∞ is in some open set W ∈ U . and so Theorem 9 R∗ is compact. −6). 0). 1). It turns out that Every open cover of R∗ has a ﬁnite subcover. 100). (−1. (0. (150. Thus. This is not just a feature of one particular example. ∞]}. 50). If you look at this. . 33 The examples of open covers of R∗ given above are both ﬁnite collections of sets.. you realize that even though this collection does contain inﬁnitely many open sets. 6). 3). This collection fails to include −∞ and ∞. We can prove the compactness of R∗ using the completeness of the real line. Let us look at an example of a cover which uses inﬁnitely many sets. (−7.. ∞]} and this would cover all of R∗ . To obtain an open cover of R∗ from U we could. (−2. (760. 0). This property is called compactness. 1). just throw in the open sets [−∞. (0. −1). for example. [−∞. 800). recall that Z is the set of all integers. In short we are looking at the collection U = {(a. 5).(−2. where a runs over all integers: . however. −4). 3). (1. 200) ∪ (701. . (3. ∞]. 4). 2).. (2. as opposed to R∗ : take all intervals of the form (a.e. (1. In particular. a + 2) : a ∈ Z}. i. 750).. a + 2) : a ∈ Z}. (−1. we could just use the sub-collection V = {[−∞.. a + 2). (5∞]} ∪ {(a. (−8. we don’t reall need all of them to cover R∗ . (4. 2). (2. this gives us an open cover of R∗ : U = {[−∞. Thus W contains a neighborhood of −∞. Let U be an open cover of R∗ ..t) ⊂ W . (5. (8. Proof of compactness of R∗ . We start with an open cover of R. −4) and (5.. −4). −5). where. (−3. Indeed.Notes in Introductory Real Analysis Another example of an open cover of R∗ is {[−∞.

say U1 . and is being assumed to be also not ∞.UN . x]} This set is not empty because −∞ ∈ S. all of R∗ .. m]. the set S has a supremum sup S. which cover the ray segment [−∞. i. Suppose to the contrary that m < ∞. m + ε) ⊂ U. in U . In particular. We will argue in three steps: (i) First we explain that m isn’t −∞. By completeness.. . m is certainly not −∞. there has to be some x ∈ S which is greater than m − ε.U} covers all of [−∞. In fact. it is in R and so there is some neighborhood (m − ε.. So. Thus m − ε < x ≤ m for some x ∈ S. where ε is a positive real number. (ii) Next we prove that m is in fact. all the elements of R∗ less than t are covered by just the one set W in U . Thus. Sengupta for some t ∈ R.UN . x]. that there is a real number t such that t − 1 is in S.. m + ε). x being in S. Now we show that m = ∞. The set U covers (m − ε. there are ﬁnitely many sets.. Since m is not −∞. Now there is some open set U ∈ U such that m ∈ U. the collection {U1 . t − 1 is also in S.e. ∞.34 Ambar N. because U covers all points of R∗ . Now let S = {x ∈ R∗ : ﬁnitely many sets in U cover [−∞. Let m = sup S. as before. (iii) Finally we prove that [∞. and so m (being an upper bound of S) has to be ≥ t − 1. . for one thing. m + ε). note. . can be covered by just ﬁnitely many sets in U . Now m − ε being less than the least upper bound m of S. For (i). we would not need any other set from the cover if we are to stick to points to the left of t. So..

there is some ‘ray-neighborhood’ of ∞ (r.Vk .Vk . Note that [−∞. Therefore. Throw this into the collection.V cover all of [−∞. there must be some y ∈ (r.. and the latter being the least upper bound of S... This prove that ﬁnitely many sets from U cover all of R∗ . Consider now any closed subset D of R∗ . which is in S. But now we have a contradiction. Now the set Dc . being the complement of a closed set. of course. But then V1 . because we have found a number. in U . Thus. Finally we prove that R∗ is covered by ﬁnitely many sets in U .U. . ∞] = R∗ . is an open set. ∞]. 2. for example. since V covers the segment (r. Consider any open cover U of D.e. Thus.Notes in Introductory Real Analysis 35 1 But this means that. We will prove that it is compact. i. say V1 . and consider U = U ∪ {Dc }.. for the ray segment [−∞. . ∞] ⊂ V. The element ∞ ∈ R∗ falls inside some open set V in the collection U . This covers all of R∗ : any point of D would be covered by a set in U while any point in Dc is. . y] ∪ (r. our original hypothesis concerning m must have been wrong. m + 1/2] is covered by the ﬁnite collection U1 . for some real number r. greater than m.UN . ∞] which is in S. So m = ∞. Now r being less than m = ∞.. m + ε/2. This is a collection of open sets such that every point of D is covered by some set in the collection. that any open cover of D has a ﬁnite sub-cover. covered by Dc .. ∞].. y] is covered by ﬁnitely many open sets.13 Compactness of closed subsets of R∗.. Then we know that there has to be a ﬁnite sub-collection V ⊂U . m + 2 ε is in S.. [−∞.

∞) ⊂ R. Note that [4. Recall that a subset of R is closed in R if it contains no boundary point. ∞) is not closed when considered as a subset of R∗ . Thus. We can now state the Heine-Borel theorem: Theorem 11 Every closed and bounded subset of R is compact.14 The Heine-Borel Theorem: Compact subsets of R A subset B of R is said to be bounded if B ⊂ [−N. Consider. it covers all points of D. D is covered by a ﬁnite sub-collection of sets from U . Thus. Sengupta which covers R∗ . Of course. every closed subset of R∗ is compact. and consider the collection V = V \ {Dc }. there should exist a real number N such that |x| < N for all x ∈ B. This is a closed subset of R because. for instance. N] for some real number N. and we can state: Theorem 10 A subset of R∗ is compact if and only if it is closed. the set [4. because no point in D could have been covered by the ‘rejected’ set Dc . for B to be bounded. . We will leave out the proof of the converse part of this result. if a subset of R is closed if its complement in R is open. this is a ﬁnite subcollection of U . Moreover. Thus. every compact subset of R is closed and bounded.36 Ambar N. and this point lies in the set. its only boundary point is 4. in R. The converse is also true. 2. Conversely. Now throw out Dc from V in case it is there. Equivalently.

n+1 . Consider any point p ∈ U. Sometimes we may have a formula for the n-th term: an = (−1)n . If p = ∞ then the neighborhood of p given by (N + 1. a3 . −3. and hence K is closed in R∗ . in all cases.15 Sequences a1 . If p ∈ R then we already know. −5. 5. a2 . for example. We know that then K must be compact. . Boundedness implies that K ⊂ [−N.. A sequence of elements in a set A is a string of elements of A. 2. 12.. Thus. . p has a neighborhood outside K in R∗ .Notes in Introductory Real Analysis 37 We shall prove half of this. Let’s check this. More precisely. by closedness of K.. 9. N] for some real number N. −N − 1) is entirely outside K. Sometimes out sequence will be speciﬁed explicitly as a string of numbers. ∞] lies entirely outside K. If p = −∞ then the neighborhood [−∞. We will show that the complement U of K in R∗ is open. 6.. Then K is closed also as a subset of R∗ . Suppose K ⊂ R is closed and bounded as a subset of R. We will often be concerned with sequences in R∗ . the sequence is actually a mapping a : P → A : n → an . that p has a neighborhood lying inside U. So the complement of K in R∗ is open.

. . Thus.. Therefore.. .. Let U be any neighborhood of p.. Sengupta Sometimes we have a sequence speciﬁed recursively. r2 . Sometimes it is better to label a sequence starting with the index 0: a0 .. . b2 = 1 and then bn = bn−1 + bn−2 . For example. We know that this neighborhood is visited inﬁnitely often by the sequence. 13. 5. For example. any limit point of a sequence which lies always in a set S must be in the closure of S. for the sequence 1 1 1 1 1 1 1 1 1 .. r3 . 2. in R∗ .16 Limits points of a sequence Consider a sequence x1 . . .. 1. 2 3 3 4 4 5 5 6 6 it is intuitively clear that both 0 and 4 are limit points. for all n ∈ P Consider then any limit point p of this sequence. x2 . r1 = r.. Thus. and so p ∈ S. . we might know that b1 = 1. 8. a1 . 2. . . 4+ . Suppose a sequence (xn ) lies entirely inside a set S ⊂ R: xn ∈ S.. for a ﬁxed real number r we can form the sequence r0 = 1. 21. 4. x3 . at least one point of S must be in U. 3. For example.. 4 + . A point p ∈ R∗ is said to be a limit point of this sequence if every neighborhood of p is visited inﬁnitely often by the sequence.. This generates the Fibonacci numbers 1. . 4+ . every neigborhood of p contains a point of S. 4+ . for all n ≥ 3. a2 .38 Ambar N.

. Of course. 2. Thus. there are ﬁnitely many of them.. visits every natural number inﬁnitely often and so every natural number is a limit point of the sequence. Then... x3 . say U p1 . 3. 3. by compactness of R∗ .Notes in Introductory Real Analysis 39 2.. x2 . 1.. we have the more general Bolzano-Weierstrass theorem: Theorem 12 Every sequence in a compact set has at least one limit point. For example.17 Bolzano-Weierstrass Theorem Consider any sequence in R∗ . . x2 . 2. 1. we have seen that a sequence may well have more than one limit point.. The neighborhoods U p form an open cover of all of R∗ . . We shall prove that it must have at least one limit point. 5. . 3... 2. 1. x2 . Now let S1 be the least upper bound of the set {x1 .U pN which cover all of R∗ . 2. . 2. In the next subsection we will show how to actually obtain a limit point of a sequence. But this is impossible: we have covered all of the extended real line R∗ with ﬁnitely many sets each of which is visited only ﬁnitely many times by the sequence! Thus. 2. Notice that all we needed above was the compactness of R∗ . the sequence. Then each point p of R∗ has a neighborhood U p which is visited only ﬁnitely often by the sequence. Suppose to the contrary that no point is a limit of the given sequence..}: S1 = sup{x1 . 1. Consider a sequence Let p be any limit point of this sequence. 4. 1. 1. 3.. we have a contradiction and so there exist a limit of the sequence. 4..18 Limit points and suprema and inﬁma x1 .} .

Thus.} and have again p ≤ S3 .11) . S2 . and so at least once by x2 . .. Thus. . .. .... So p ≤ S1 Now we can apply the same argument to S2 = sup{x2 . Sengupta Then it is clear that no point ‘to the right’ of S1 could possibly be a limit of the sequence: indeed any point r > S1 would have a neighborhood lying entirely to the ‘right’ of S1 and this would never be visited by the sequence.}. Ambar N. p ≤ inf{S1 .. x4 . S3 . .. x2 . p is a lower bound for all the Sk ’s: p ≤ Sk for all k. p cannot be > S1 . p could not be greater than S2 for then it would have a neighborhood to the right of S2 and this would have to be visited inﬁnitely often by x1 . S1 ≥ xn for each n.40 In particular.. The infemum on the right is an important object and is called the limsup of the given sequence: lim supn→∞ xn = inf{supn≥k xn } : k ∈ P} So we have proved that Every limit point is ≤ the limsup def (2. p ≤ S2 In this way.} Again... Therefore. Thus. we can form S3 = sup{x3 .. x3 . Now the inf of a set is the greatest lower bound of the set. x3 . x4 .

say n0 . will lie in the set (20. if sn ∈ S for all values of n beyond some value.19 Limit of a sequence A sequence (sn ) is said to lie in a set S eventually if after a certain value of n. 1.13) In fact. −2. For example. ∞) eventually. we can prove Every limit point is ≥ the limsup Thus. for any limit point p if any sequence (xn ) we have def (2. We denote this symbolically as xn → L. the limsup and liminf of the sequence are also limit points. . 0. A sequence (xn ) in R∗ is said to have limit L ∈ R∗ if for any neighborhood U of L the sequence lies in this neighborhood eventually. but we won’t prove this here. all the sn lie in S. −3. 3.. .. −4. −1. we have the notion of liminf: it is the supermum of the sequence of infemums: lim infn→∞ xn = sup{infn≥k xn } : k ∈ P} In just the same way as before. we say that sn lies in S for large n. Similarly. 2. Let us note again: a sequence (sn ) lies in a set S eventually if there is an n0 ∈ P such that sn ∈ S for all n ∈ P with n ≥ n0 . Put another way. (2.12) lim infn→∞ xn ≤ any limit point ≤ lim supn→∞ xn .Notes in Introductory Real Analysis 41 for any sequence. the sequence −5. as n → ∞. 2.

7. .. The notion of limit is one of the central notions in mathematics. ∞] from the n-th term onwards. . 3.. . 3.. . 3.. . 4... the sequence 1. 3. if we take any neighborhood of ∞. 8. The sequence −1. . 1. the sequence 1. 1. 4. The sequence 1 1 1 1. . ... For example. 3. 3. 4. The simplest sequence is a constant sequence which keeps repeating the same value. 5. For example.20 Simple Examples of limits Let us look at some examples of sequences and try to see what their limits are.. 2. say (t. For. 3. 4.. again. 8. 2 3 4 has limit 0.. 4. 3. 1.42 Ambar N. 3. The limit of the this sequence is 3: clearly every neighborhood of 3 is hit eventually by the sequence (indeed it is hit every time. 8. In contrast. since the sequence is stuck at 3). ∞] then eventually n will exceed t (by the archimidean property) and so the sequence will stay in (t. 4. has limit ∞. Sengupta We shall see later that if a limit exists then it is unique. 8.. For. .. given any neighborhood of 8 the sequence falls inside this neighborhood eventually and stays there. 8. We will look at this more carefully soon. 2. the limit L is denoted n→∞ lim xn . which eventually stabilizes at the constant value 8 has limit 8.

Thus we should show that 1 <ε n for all n beyond some initial value. We have to show that given any neighborhood of 0. 7.. If you take any neighborhood of ∞. 3. where ε is a positive real number.Notes in Introductory Real Analysis 43 does not have a limit.5) is a neighborhood of 3. . ε).. (2. We have to show that xn lies in (−ε. 1/2.5. 1/4.. the point 3 cannot be the limit of the sequence because. 2. The sequence 1. But let us prove this. ∞] then eventually the sequence falls inside the nighborhood and stays in there. 5. Note ﬁrst that the n-th term of the sequence is xn = 1/n. For example. Consider again the sequence It is intuitively clear that this sequence has limit 0. has limit ∞.. an interval of the form (t. our sequence will eventually lie inside this neighborhood.21 The sequence 1/n 1. . 3. but the sequence keeps falling outside this neighborhood (when it hits 1 or 4). So consider any neighborhood of 0: (−ε. 1/3. . Now the condition 1/n < ε is equivalent to nε > 1. for instance. ε) for all n beyond some value.

32 . and so Rn ≥ 1 + nx. 33 . . . n as n → ∞ 2.. 31 . and intuitively it is clear that the sequence has limit ∞. Let us see by how much each term exceeds the preceding: Rn − Rn−1 = Rn−1 (R − 1) But R is > 1. This proves that the sequence does indeed tend to the limit 0: 1 → 0. and so the multiplier Rn−1 is ≥ 1 (it is equal to 1 when n is actually 1). Now it takes n ‘steps’ to climb from R0 to Rn .44 Ambar N. Sengupta The archimedean property guarantees the existence of an n0 ∈ P for which n0 ε > 1. Let us write x for R − 1. Rn − Rn−1 ≥ R − 1. and note that x>0 and we have Rn − Rn−1 > x. Consider now the sequence of powers Rn where R > 1. It is clear that in this sequence the terms get large very quickly.22 The sequence Rn Consider the sequence 30 . Therefore. Thus. nε > 1 for all integers n ≥ n0 ..

But this is just Archimedes again: some multiple n0 x of xexceeds t − 1. It seems then clear that the sequence goes to 0. 1 + nx > t when n is large enough. Indeed. and then nx > t − 1 for all n ≥ n0 . Now it is clear that Rn → ∞ as n → ∞. 1.. 2 . 1. Thus.Notes in Introductory Real Analysis 45 because each step is at least x. lim Rn = ∞ for all R > 1. .. and this comes from the fact that R > 1. As an example. We just have to show that for any given t. ε) 3n Now consider the general case of . 1. Rn and so 1 ∈ (−ε. we have R = 1/3 and the sequence 1 1 1 1. Thus n→∞ lim Rn = 1 if R = 1. n→∞ Now consider the case R = 1. Note again that it is important that x > 0. 3 3 3 We have here a sequence with denomiantor going to ∞ and numerator ﬁxed at 1. for any real ε > 0 we just have to wait till 3n > 1/ε and this would ensure that 1 < ε. In this case the sequence is just all powers of 1. . 3 .... 3n for large n. and so it is 1. . Next consider the case 0 < R < 1.

For example. Consequently. Suppose −1 < R < 0. as we have seen above. for all n ≥ n0 . 2 . lim Rn = 0 if R = 0. i. for R = −1/3 we have the sequence 1 1 1 1 1. − . But this means we should show that rn is > 1/ε for all n large enough. We want Rn to be in this neighborhood. Consider any neighborhood of 0: (−ε. Now. . the denominator rn goes to inﬁnity. So we will prove that Rn → 0 in this case. we have for all n ≥ n0 .. r Rn = 1 . Sengupta 1 R= . 4 . Observe that where r = R−1 is > 1. and so it seems clear that 1/rn should decrease to 0. Next consider negative values. Thus Ambar N. − 3 . Rn should be < ε. ε). Thus Now it is also clear that n→∞ rn > ε−1 1 <ε rn n→∞ lim Rn = 0 if 0 < R < 1..46 where R > 1. rn where r > 1.e. after some n0 . 3 3 3 3 . But we know that rn → ∞ So. ε) where ε > 0 is a positive real number. Rn = and so Rn ∈ (ε.

Does this sequence visit every neighborhood of ∞? Does it visit every neighborhood of −∞? if |R| < 1. −2. 4. if −1 < R < 0 we can look at the distance between Rn and 0: |Rn − 0| = |Rn | = |R|n and |R|n → 0 because 0 < |R| < 1. this sequence has limit 0. . Thus. lim Rn = 0 if −1 < R < 0.. Exerise Examine the sequence given by powers of −2: 1. −8..Notes in Introductory Real Analysis 47 It is clear that this is the same sequence as 1/3n except it swings back and forth between negative and positive values. 16.14) . one should look at the case R ≤ −1. for the sequence swings back and forth between widely separated positive and negative values. the conclusion is that n→∞ as n → ∞. limn→∞ Rn = 0 Lastly. (2. In short. You should examine a few examples and convince yourself that then there is no limit. More generally. Thus.

15) (2..111. the sequence (xn ) falls eventually in U and stays there. 2. 1. Pick a point t of U to the left of L. .23 Monotone Sequences A sequence such as 1. So consider any neighborhood U of L. 1. where each term is ≥ the preceding term is said to be monotone increasing. This is indeed true: If (xn ) is a monotone increasing sequence then n→∞ (2. since xn is just stuck at −∞. Now consider the other situation: L is not −∞. 1/2.1.48 Ambar N. . t < L and t ∈ U.. Another example is the sequence 1. 11. we need to take care of this case separately. as will become clear shortly.e... 8. i. 1.}. 9. 1. 1/3. n≥1 The most extreme case is when L is −∞. n≥1 To prove this let L = sup xn .16) lim xn = sup xn . It is intuitively clear that a monotone increasing sequence (xn ) will tend to the limit sup{x1 . 1/5.1111. .e. ..11. Thus. Sengupta 2. 1/4.. 5. a sequence (xn ) is monotone increasing if xn ≤ xn+1 for all n ∈ P We say that (xn ) is monotone decreasing if xn ≥ xn+1 for all n ∈ P An example of a monotone decreasing sequence is 1. i.. If L = −∞ each xn must be −∞. But in this case n→∞ lim xn = −∞. We will show again that xn → L. .. x2 . we will show that for any neigborhood U of L.

2.24 The limit of a sequence is unique Now we shall prove that a sequence can have at most one limit. and L = L . of course. In a similar way. in the neighborhood U of L. the number t can’t be an upper bound. . Thus. But recall that L is an upper bound of {x1 .}. therefore. x3 . Since L and L are distinct. if xn is a monotone decreasing sequence then n→∞ lim xn = inf xn . x2 .. and L ∈ R. and so some xn0 is > t: xn0 > t. Consider a sequence (xn ) and suppose both L and L are limits of the sequence. But this is impossible since U and U are disjoint and thus have no element in common. a sequence which has a limit must have a unique limit. But recall that we are dealing with a monotone increasing sequence. This proves that L is indeed the limit of the sequence (xn ). . x2 . all the xn are ≤ t. they have disjoint neighborhoods U and U respectively. We will arrive at a contradiction. Consequently.25 Convergent sequences and Cauchy sequences A sequence (xn ) in R is said to be convergent if it has a limit and the limit is a real number. and so xn ∈ U eventually. we also say that the sequence xn converges to L. Thus.) Since L is the least upper bound of {x1 . But xn → L . If xn → L. Note that xn → L ∈ R if for any real r > 0 there is an n0 ∈ P such that |xn − L| < r . xn > t for all n ≥ n0 .Notes in Introductory Real Analysis 49 (Note that there would be no such t if L is −∞. Since xn → L we know that xn ∈ U eventually. xn ∈ U for all n ≥ n0 ... n≥1 2.}.. Those xn which are > t but ≤ L are.

in fact. More precisely. So x j − 1 < xn < x j + 1 for all n ∈ {N + 1. at least from the (N + 1)-th term on. m > N. i. Then.e. Sengupta for all natural numbers n > n0 . In particular. A sequence (xn ) in R which bunches up on itself in the sense above is said to be a Cauchy sequence. m > N. m > N.}... for any ε > 0 we can ﬁnd N ∈ P such that |xn − xm | < ε for all natural numbers n. Thus. x j + 1).26 Every Cauchy sequence is bounded Consider a Cauchy sequence (xn ) in R. if for any δ > 0 there is an N ∈ P such that |xn − xm | < ε for all natural numbers n.50 Ambar N. there is an N ∈ P such that |xn − xm | < 1 for all natural numbers n.. x1 . eventually the points of this sequence are at most distance 1 from each other. the sequence is bounded. N + 2. To prove this. observe that there is some N ∈ P such that |xk − L| < ε/2 for all k ∈ P with k > N.. . for all n ≥ N. m > N we have |xn − xm | = ≤ = < = |xn − L + L − xm | |xn − L| + |L − xm | |xn − L| + |xm − L| ε/2 + ε/2 ε. But the terms not counted here. ﬁxing a particular j > N we have xn ∈ (x j − 1.. . But then for any n. m ∈ P with n. 2. This has a consequence: since the xn ’s are all accumulating to L they are also getting close to each other. xN .

x j − 1} Thus the entire sequence is bounded. The sequence (xn ) visits the neighborhood δ δ (L − . L + δ) of L. L is not ∞ or −∞. Let L be a limit point of (xn ). We have seen that it is bounded. Now we also know that eventually. the terms of the sequence vary from each other by < δ/2. Take any neighborhood (L − δ. m ∈ P with n. where δ > 0 is a real number. L + ) 2 2 inﬁnitely often. and is a real number. Thus. b]. We will prove that L is a real number and xn → L. there is some N ∈ P such that |xn − xm | < δ/2 for all n. x j + 1} and every xn is ≥ min{B.Notes in Introductory Real Analysis 51 are just ﬁnitely many.e. Let us prove this. First it is clear that L must also be in [a. every xn is ≤ max{B. Consider a Cauchy sequence (xn ) in R. Thus. a ≤ xn ≤ b for all n ∈ P for some real numbers a and b. 2. m > N. Therefore. i. .27 Every Cauchy sequence is convergent The completeness property of the real line R has an equivalent formulation: Theorem 13 Every Cauchy sequence in R converges. and so they have a largest B and a smallest A among them. We know that a sequence always has a limit point in R∗ .

e. the even numbers form a countable set: {2. This means that xn ∈ (L − δ.. but we can certainly lay out all the integers in a sequence: 0. x3 . a sequence r1 ...52 We can choose some j > N such that δ δ x j ∈ (L − . such that {r1 .} It may seem at ﬁrst that the set Z of integers is not countable.. Then |xn − L| = |xn − x j + x j − L| ≤ |xn − x j | + |x j − L| δ δ < + 2 2 = δ. . −2. 2. xn → L.. 8. . This is what we shall prove now. Thus. We will constuct a sequence which enumerates all the rationals. It is much harder to see that the rationals Q are also countable. . −1. .. ... 4. x2 .. Sengupta because the sequence visits this neighborhood inﬁnitely often. 3. −3. L + δ) for all n > N. 1. i.} = Q. r3 . . . r2 . 6.. L + ) 2 2 Ambar N. S is countable if there is a sequence x1 .} For example. Thus... 2. r2 . such that S = {x1 .28 The rationals are countable A set is set to be countable if it is ﬁnite or if its elements can be enumerated in a sequence. x2 .

such that {r1 . 0.. x < 0. and so on. If x ∈ Q is negative.. Thus. There can be many such encoding schemes. (For example.6 can be written as both 6/10 and 3/5. r2 . . r2 . let us deﬁne f (x) to be just 5 times f (−x): f (x) = f (−x) × 5 if x < 0 Now we have labeled each rational by a unique natural number.. thus to each natural number n we would associate a rational rn . . This associates a unique natural number f (x) to each non-negative rational x. f (4/5) = 25 × 34 = 32 × 81 = 2592. f (rn−1 ).. Here is one: Take any rational p x ≥ 0 and write it as q where p and q are non-negative integers (of course q ≥ 1) and q is the smallest possible denominator among all such ways of writing x. we will encode each rational by a unique natural number.. rn is the rational for which f (rn ) is the ﬁrst value of f (x) greater than f (r1 ).) We know that every non-negative rational can be represented in this way uniquely. f (0) = f (0/1) = 21 × 30 = 2 f (1) = f (1/1) = 21 31 = 6.. the next value for f (x) after 2 Thus we have produced a sequence of distinct elements r1 .Notes in Introductory Real Analysis 53 Put another way. let r2 be the rational number with the next lowest value for f (x). and in this way we will cover all rationals. For example. Associate to this x the natural number f (x) = 2q × 3 p For example.} = Q . but we would take 3/5 as our representation.. To enumerate the rationals in a sequence all we have to do then is to run this in reverse: let r1 be the rational number x with the smallest value for f (x) (so r1 is in fact 0 because f (0) = 2 is the lowest value possible for f ). . r1 = 0 and r2 = 1 because f (1) = 21 31 = 6.

1). 1)..19999.. = y2 y3 y1 + 2 + 3 +··· .. [The reason we excluded 0 and 9 from our choices for wn was to avoid ending up at 0 or 1.. for every n.20000. 2. . lying in (0. 10 10 10 where y1 .e.... 8} other than x11 . Now form a number w = 0. there is no sequence which touches on all the real numbers. Thus each of the numbers x1 . then take the main diagonal string. . x2 = 0. x2 .x11 x12 x13 .y1 y2 y3 .. i. . Now each real number y in (0.. ∈ {0.. Sengupta 2. as follows: take w1 to be any number in {1. 1. . 8} other than x22 ..] . . and then form a new string by altering each element of the diagonal. and w ∈ (0. . 9} and we exclude all such representations which use an inﬁnite string of 9’s at the end (for example. also has such an expansion: x1 = 0. 1).x31 x32 x33 . The strategy is to make a list of strings. we would use 0. . 2. In fact we will show that even the real numbers between 0 and 1 cannot be enumerated in a sequence. . This way we make the n-th decimal place of w different from the n-th decimal place of xn . .. = ··· . Consider any sequence x1 .. . Thus the original sequence cannot possibly have covered all of (0. Then w cannot be equal to any of the xn ..29 The real numbers are uncountable We will prove that R is not countable... 3.x21 x22 x23 .... then choose w2 ∈ {1..). and so on. 2. The argument used here is the celebrated diagonal method due to Georg Cantor (1845-1912). x3 = 0. 1) can be expressed uniquely in the form y = 0.. instead of 0.w1 w2 w3 .. x2 .. y2 .54 Ambar N. . . .

or of R∗ . 5] the set [2. U ∩ S is said to be open in S.30 Connected Sets Consider a subset S of R∗ . [2. For example. i. If U is an open set then the part of U in S. a subset J of S is said to be open in S if every p ∈ J has a neighborhood N such that every point of N which is in S is in fact in J. The main result for connected sets is: Theorem 14 A subset of R.e. in [2. . 4) is open because. 4) ∩ [2. is connected if and only if it is an interval. 4) = (1.e.Notes in Introductory Real Analysis 55 2. 5] Put another way. N ∩J ⊂ S A set S is said to be connected if it cannot be split into two non-empty disjoint pieces A and B each of which is open in S. for example. i.

56 Ambar N. Sengupta .

but the origins of integration theory lie in the computation of areas and volumes. it is clear that the area enclosed by a rectangle whose sides are a units and b units. Thus. it makes sense to suppose that area of R ≤ area of R ≤ area of R which is to say: a b ≤ area of R ≤ a b It is intuitively clear. each row having 2 tiles. Take any rationals a . area of R = ab 57 . and not hard to prove. Consider now a rectangle R whose sides a and b are possibly not rational. a . and b <b<b .Chapter 3 Integration Modern integration theory was built from the works of Newton. is ab square units. Riemann. b with a <a<a . Then a rectangle R of sides a by b sits inside R. and Lebesgue. the area of a rectangle of sides 1/4 unit and 1/5 unit should be 1/20 square units because we would need 20 of these rectangles to cover a unit square. Thus. By extension. and b . The area of a rectangle whose sides are 2 units and 3 units is 6 square units. while a rectangle R of sides a by b contains R. that ab is the unique real number that lies between all the possible values of a b and a b . where a and b are rational numbers. because if you tile the rectangle with unit squares you will need three rows of such squares. The idea of measuring area of a curved region by lower and upper bounds is present even Archimedes’ study of the area enclosed by a circle. Thus.

1 Approaching the Riemann Integral Consider a function f : [a. Sengupta The rectangle R is made up of two congruent right angled triangles. The integral Z b f a which is also. The area of each such slice is between the area of a larger ‘upper’ rectangle . is curved. This makes it possible to compute the areas of all kinds of polygonal ﬁgures but cutting up these fgures into right angled triangles. Archimedes computed area enclosed by a circle C by consider polygons P and P . However. this strategy fails when we try to compute the area enclosed by a circle. this is a region whose upper boundary. The strategy used for computing the area is to cut up the region into vertical slices. written as Z b f (x) dx a measures the area of this region.58 Ambar N. for convenience of visualization. Then f speciﬁes a region which lies below its graph and above the x-axis. conveniently. In general. thus area A enclosed by P ≤ area enclosed by C ≤ area A enclosed by P Some computation shows that for any ε > 0 there are such polygons P and P such that A −A < ε This implies that there is a unique real number A which lies between the area of all the polygons P and the polygons P . Clearly then this number should be the area enclosed by the circle C. Our development of the theory of the Riemann integrals is based on these ideas. Assume. and so each of these would have area (ab)/2. that f ≥ 0. where P lies inside the circle C and P encloses the circle C. given by the graph of f . No amount of cutting would turn a disc into a ﬁnite number of right angled triangles. b] → R and think of its graph. 3.

Thus we should expect the actual area to be the unique real number lying between these ‘upper rectangle’ areas and the ‘lower rectangle’ areas...3) .}. b] is speciﬁed by a sequence X = (x0 . 3.... b] ⊂ R We will work with functions on an interval ] where < b.. b] ⊂ R. A partition X of [a. where a < b.... b] → R on an interval [a. b] with a = x0 < x1 < · · · < xN = b Here N ∈ {1. ...1) ∆x j (3. We will often use the notation ∆x j = x j − x j−1 to denote the length of the j-th interval [x j−1 . . The width or norm of the partition X is the maximum size of these intervals: ||X|| = Consider a function f : [a.Notes in Introductory Real Analysis 59 and a smaller ‘lower’ rectangle. . Let M j( f ) = max j∈{1. 3. x j ] marked out by the partition X.N} def (3.2) sup x∈[x j−1 .2 Riemann Sums [a. x1 .. x1 . xN ) of points x0 .x j ] f (x) (3. xN ∈ [a. 2.

. L( f . Our objective is to specify the actual area A under the graph of f .. if A j were the area of the region under the graph of f over [x j−1 . x j ] j We shall indicate this by writing X∗ < X The Riemann sum S( f . x j ] then m j ( f )∆x j ≤ A j ≤ M j ( f )∆x j . The upper Riemann sum U( f .. X. x j ]:: x∗ ∈ [x j−1 . X) (3. Sengupta inf f (x) (3.4) Thus.Y ) for every partitions X and Y of [a. and this would be the sum of the A j .5) ∑ m j ( f )∆x j (3. i.60 and m j( f ) = x∈[x j−1 . X) is less or equal to U( f . X. X) = Note that L( f . X) ≤ U( f .x j ] Ambar N. . X) is L( f . X) = and the lower Riemann sum L( f . xN ) def N j=1 def N j=1 ∑ M j ( f )∆x j (3.6) obtained by picking a point from each interval [x j−1 .e. Now consider a sequence ∗ ∗ X ∗ = (x1 . X ) = ∗ j=1 ∑ f (x∗j )∆x j N (3.8) . X ∗ ) is S( f . X) is deﬁned to be U( f .7) We will show later that in fact every lower sum if less or equal to every upper sum. b].

X) can never be ∞.10) and L( f . b] → R Consider a function on an interval [a. b] is denote R [a. we never have the ∞ − ∞ situation. X) = −∞ if and only if inf f (x) = −∞ x∈[a. X) cannot be −∞. It is useful then to observe that the difference U( f .e. b] ⊂ R. x j ] which contradicts the fact that f is real-valued.b] (3. X) ≤ I ≤ U( f . 3. (Note that we are working with non-empty intervals because x j−1 < x j . U( f . X) (3. b] and denoted Z Z b b f a or a f (x) dx (3. X) = ∞ if and only if sup f (x) = ∞ x∈[a.b] (3. This function is said to be Riemann integrable if there is a unique real number I lying between all the lower sums and all the upper sums: L( f . This number I is the integral of f over [a.) Similarly.Notes in Introductory Real Analysis 61 Note that the upper sum is ∞ if and only if one of the M j is ∞. where a < b. L( f .11) for every partition X of [a.12) The set of all Riemann integrable functions on [a.3 Deﬁnition of the Riemann Integral f : [a. b] (3.9) However. i. X) is always deﬁned. b]. because that would mean that at least one of the M j is −∞ which can only be if f is equal to −∞ on that interval [x j−1 . X) − L( f . and this occurs if and only if the superemum of f is ∞: U( f .13) .

Sengupta 3. It is important to observe that M j ≥ M j. x j ]: U( f .62 Ambar N.e. X) and U( f . If we keep adding points to the initial partition.14) Let X = (x0 . X cuts up [x j−1 . This observation is important enough to state as a theorem: . M j = M( f . b] and let X be a partition obtained by adding one more point x to X..t] (3. x j ] Let us compare the upper sums U( f .t]) = sup f (x). Arguing similarly. b] → R. it follows that adding points to a partition raises lower sums. [s. X ).t] and m( f . . and M j ≥ M j (3. x∈[s. let us write M j = M( f . X). adding a point to a partition lowers upper sums. xN ) be a partition of the interval [a. [x .15) M j = M( f . Thus U( f . [x j−1 .16) i. X ) ≤ U( f . [x j−1 . To this end. x j ]). (3. x j ] Thus. x − j] into two intervals [x j−1 . we keep lower the upper sums. Let us say that x lies in the j-th interval x ∈ [x j−1 .t]) = inf f (x) x∈[s. x ] and [x . X) −U( f . X ) = M j (x j − x j−1 ) − M j (x − x j−1 ) + M j (x j − x ) = (M j − M j )(x − x j−1 ) + (M j − M j )(x j − x ) ≥ 0.. x j ]) These sums differ only in the contribution that comes from [x j−1 . [s. We use the notation M( f . x1 . x ]).4 Reﬁning Partitions We work with a function f : [a..

X ) = (M j − M j )(x − x j−1 ) + (M j − M j )(x j − x ) (3. with X containing all the points of X and some more. Z) ≤ U( f .Y ) Thus.Y ) and L( f . X ) ≤ U( f .. U( f . x j ] of the partition X = (x0 . Z) ≤ U( f . then L( f . It will be useful to take a closer look at this and estimate by how much the upper and lower sums move as points are added to the partition. Now suppose instead of adding just the one point x . b] → R is a function on an interval [a. b] ⊂ R. . x j ] to form the new partition X .. .. Therefore. b] then L( f . we add m distinct points y1 . X) ≤ L( f . X) −U( f . X ) ≤ U( f . every lower sum is dominated above by every upper sum.. Using this we can prove that upper sums always dominate lower sums: Theorem 16 If f : [a..19) (3. xT ).18) (3. b]. Z) ≤ U( f . b] → R. Thus if X and X are partitions of [a..Y ) and this gives the desired result. where a < b. . X) ≤ L( f . Here is a useful trick: we can combine the partitions X and Y to form a partition Z which contains all points of X and all the points of Y as well. and if X and Y are any partitions of [a.Notes in Introductory Real Analysis 63 Theorem 15 Adding points to a partition lowers upper sums and raises lower sums. We label the new points yi in increasing order y1 < · · · < ym . ym to the interval [x j−1 . X) ≤ L( f . X) ≤ U( f . X) for every function f : [a. Z) Stringing these together we have L( f .17) where X is the partition obtained from X by adding the single point x to the j-th interval [x j−1 . QED We have used the fact that adding a point to a partition lowers upper sums and raises lower sums. Recall the formula U( f . Proof.

Sengupta For convenience of notation we write y0 for x j−1 and ym+1 for x j . U( f . X) ≤ 2N || f ||sup ||X|| Note that N ≤ N. X) −U( f . X ) − L( f . X) −U( f .23) (3. If we add points to each of N intervals to create a new partition X then U( f . X) − L( f . X) but the decrease in the value of U − L is at most 2N|| f ||sup ||X||: [U( f .25) (3. yk ]) ≤ 2|| f ||sup .22) . Consequently. b] ⊂ R.20) 0 ≤ M( f . we have Lemma 1 If X = (x0 . (3. X) −U( f . X ) − L( f .. X)] − U( f . X ) ≤ 2N || f ||sup ||X|| Similarly. [x j−1 .. . X ) − L( f . [yk−1 . [x j−1 . x j ] Then: m+1 U( f . X) − L( f . X ) ≤ 2N|| f ||sup ||X|| where N is the number of new points added. and X is a partition of the same interval containing all the points of X and possibly some more. yk ]) (yk − yk−1 ) (3.24) (3. Thus x j−1 = y0 < y1 < · · · < ym < ym+1 = x j The intervals [yk−1 . x j ]) − M( f . X ) ≤ U( f . then U( f . yk ] make up the interval [x j−1 . [yk−1 . where a < b. X ) ≤ 2|| f ||sup ∆x j ≤ 2|| f ||sup ||X|| (3. X ) = Now ∑ k=1 M( f . x j ]) − M( f .. The interval lengths yk − yk−1 add up to ∆x j . where N is the total number of new points added. b]. L( f .64 Ambar N. xT ) is a partition of the interval [a. Consequently.21) This provides an upper bound for how much the upper sum is decreased by addition of points all in one single interval of the original partition X. where || f ||sup is supremum of | f | over the full original interval [a.

the function f is Riemann integrable if and only if there is no ‘gap’ between the lower sums and upper sums. there is a partition Y for which I ≤ U( f .26) R The common value is ab f . there is a partition Z for which I − ε/2 < L( f . Take any ε > 0. I + ε/2). Consider the interval [I. Thus. Proof.5 The Darboux Criterion Now that we know that the upper sums dominate the lower sums. there must be at least one upper sum which lies in [I. Then I − ε/2 < L( f .Y ) < I + ε/2 Similarly. Then there is a unique real number I which lies between all upper sums and all lower sums. another equivalent formulation is the Darboux criterion: Theorem 18 T:Darbouxcrit A function f : [a. Z) ≤ L( f . it is clear that there will be a unique real number between them if and only if the sup of the lower sums equals the inf of the upper sums: Theorem 17 A function f : [a. b] → R is Riemann integrable if and only if sup{U( f .28) .27) This is an extremely useful result: virtually all of our results on integrability will use the Darboux criterion. X) < ε (3. Thus. So since I is the only real number lying between the upper sums and lower sums. b]} = inf{L( f . I + ε/2) All the lower sums lie to the left. Z) ≤ I. X) : all partitions X of [a. X) ≤ U( f . b] → R is Riemann integrable if and only if for any ε > 0 there is a partition X of [a.Notes in Introductory Real Analysis 65 3. X) ≤ U( f . Thus. b] for which U( f . Z) < I + ε/2 (3. All upper sums lie to the right (≥) of I. i. Suppose f is integrable. ≤ I.e. X) : all partitions X of [a. b]} (3. Let X be the partition obtained by pooling together Y and Z. X) − L( f .

X) lie in (I − ε/2. X) and U( f . b] → R is integrable. Thus. Theorem 19 Every Riemann integrable function is bounded. X) < ε. for example. sup f (x) = ∞. X)...29) Conversely. Now I and I both lie between U( f . So the difference between I and I is < ε: |I − I| < ε But this contradicts the deifnition of ε taken above. Let ε = |I − I| We know that there is a partition X satifying (3.29). Ambar N. Then we will reach a contradiction. QED 3. We will prove that f must be bounded.66 Since both L( f . Assume that f ∈ R [a. that f is not bounded above. Let us suppose that f is unbounded.e. Suppose. x∈[a.. Sengupta (3. Consider f : [a. suppose that for every ε > 0 there is a partition for which (3. So there is a unique real number between all the upper sums and all the lower sums. I + ε/2) it follows that U( f .29) holds. Thus I and I must be equal. xN ) . X) − L( f . Suppose that I and I are distinct real numbers which both lie between all upper sums and all lower sums. .6 Integrable functions are bounded Suppose f : [a. Proof. f is integrable. X) and L( f . b]. i. b] → R where [a. b] ⊂ R and a < b.b] Consider any partition X = (x0 .

Now let Z b I= a f and recall that this is the unique real number lying between all upper sums and all lower sums.e. we have a contradiction. say [x j−1 .7 Variation of a Function diam(S) = sup{a − b : a.t]) = sup ( f (x) − f (y)) x. M j = ∞.32) Now consider a function f on an interval [s. QED 3. X) (3. Then L( f . But then the supper sum for f with this partition is also ∞: N U( f .t] is: Var( f . Then there is some subinterval. i. b].30) for every partition X. every upper sum for f is ∞. [s. It will be useful to have a measure of the ﬂuctuation of f over this interval.y∈[s. But then I + 1 would also be a real number lying between all upper sums and all lower sums. b ∈ S} (3. X) = ∑ Mk ∆xk = ∞ k=1 Thus.31) For any non-empty subset S ⊂ R the diameter of S is It is easy to believe and not hard to prove that the diameter of S is the difference between sup S and inf S: diam(S) = sup S − inf S (3. The simplest such measure is given by the diameter of the range of f : Var( f ) = diam (Range of f) More explicitly.t] (3.33) (3.Notes in Introductory Real Analysis 67 of [a. X) ≤ I < I + 1 < ∞ = U( f . on which f is not bounded above. Thus.t]. x j ]. the variation of f on the interval [s.34) .

which will be very useful in proving corresponding facts about Riemann integration.t] and just write Var( f ): Var( f ) = M( f ) − m( f ) Ambar N.38) . [s. weighted by their sup-norms: Var( f g) ≤ || f ||sup Var(g) + ||g||sup Var( f ) (vi) the variations in f and in g differ by at most the variation in | f − g|: Var( f ) − Var(g) ≤ Var( f − g) (vii) if f is not equal to zero anywhere then Var(1/ f ) ≤ M(| f |−1 )2 Var( f ) (3.t]) − m( f . the variation of a constant times a function is the absolute value of the function time the variation in the function: Var(k f ) = |k|Var( f ) for any k ∈ R. b] ⊂ R with a < b. (iii) the variation scales like length. (ii) Var(k) = 0 if and only if k is constant. i.35) (3. Lemma 2 For any functions f and g on an interval [a.39) (3. (iv) the variation satiﬁes the triangle inequality: Var( f + g) ≤ Var( f ) + Var(g) (3. we have: (i) Var( f ) ≥ 0.t]) = M( f .e.37) (v) the variation of the product of two functions is bounded above by the sum of their variations.t]) Sometimes we may drop [s. [s.36) We record the following algebraic facts about the variation of functions. Sengupta (3.40) (3.68 It is also equal to Var( f . [s.

44) (3. where [a. b] (3. xN ) with all the intervals having length less than δ..41) (ix) The variation in the absolute value of f is bounded by the variation in f : Var(| f |) ≤ Var( f ) (x) The variation of f is bounded above by twice || f ||sup : Var( f ) ≤ 2|| f ||sup (xi) If s.. . b] such that Var( f . Since f is continuous on the compact set [a.t]) ≤ Var( f .43) (3.t. N}.. Proof. [s. [s. u]) ≤ Var( f . [s. b] with s < t < u then Var( f . then for any ε > 0 there is a partition X = (x0 . u]) + Var( f . b] → R is continuous. x j ]) < ε for every j ∈ {1.. x ∈ [a..Notes in Introductory Real Analysis 69 (viii) the variation of a function increases montonically with the interval of variation: Var( f .42) Var( f .. So for any ε > 0 there is a δ > 0 such that | f (x) − f (x )| < ε/2 whenever x.t]) = (3. N}.. b] it is uniformly continuous. [s. b] ⊂ R and a < b..t] then f (t) − f (s) if f (t) ≥ f (s) f (s) − f (t) if f (t) ≤ f (s) (3. u ∈ [a. Take any partition X = (x0 . f (x) − f (x ) < ε/2 . .. .. b] with |x − x | < δ.t]) (xii) If f is monotone on an interval [s. .. Then. [u.. b]) if [s.45) We also record the following result on variations and partitions for continuous functions: Lemma 3 If f : [a. xN ) of [a. [a.t] ⊂ [a. for each j ∈ {1. [x j−1 .

70

Ambar N. Sengupta

for every x, x ∈ [x j−1 , x j ] since these intervals all have length < δ. Consequently, Var( f , [x j−1 , x j ]) ≤ ε/2 and we are done. QED The sup-norm bound on the variation has the following consequence: Proposition 1 If f1 , f2 , ... : [a, b] → R converge uniformly to a function f : [a, b] → R, then lim Var( fn ) = Var( f ) (3.46)

n→∞

It will be convenient to use the notation Var j ( f ) = Var( f , [x j−1 , x j ]) The reason why we are interested in the variation is summarized by Theorem 20 For any function f : [a, b] → R, where [a, b] ⊂ R and a < b, and every partition X of [a, b] we have

N def

(3.47)

U( f , X) − L( f , X) = Proof. Observe that

j=1

∑ Var j ( f )∆x j

(3.48)

Var j ( f ) = M j ( f ) − m j ( f ).

(3.49)

Multiplying this by ∆x j and adding up over all j ∈ {1, ..., N} gives the result (3.48). QED

3.8 The algebra R [a, b]

Recall that

R [a, b]

is the set of all Riemann integrable functions on [a, b]. Our main objective now is Theorem 21 The set R [a, b] has the following properties: (i) Every constant function belongs to R [a, b]

Notes in Introductory Real Analysis (ii) If f , g ∈ R [a, b] then f + g ∈ R [a, b] (iii) If f , g ∈ R [a, b] then f g ∈ R [a, b]

71

**(iv) If f ∈ R [a, b], and f is never equal to zero and 1/ f is bounded, then 1/ f ∈ R [a, b]. (v) If f ∈ R [a, b] then | f | ∈ R [a, b] and
**

Z b

a

f ≤

Z b

a

|f|

(3.50)

Properties (i)-(iii) say that R [a, b] is an algebra under pointwise addition and multiplication of functions. It is important to note that the converse of (v) does not hold, i.e. there are functions which are not Riemann integrable but whose absoulte values are Riemann integrable. Proof For a constant function k on [a, b] we have L( f , X) = k(b − a) = U( f , X) for every partition X and so k(b − a) is the unique real number lying between all upper sums and all lower sums. Thus

Z b

a

k = k(b − a)

Now supppose f , g ∈ R [a, b]. Let ε > 0. By the Darboux condition, there are partitions Y and Z of [a, b] such that U( f ,Y ) − L( f ,Y ) < ε/2 and U(g, Z) − L(g, Z) < ε/2

Let X be the partition obtained by combining Y and Z. Then, because upper sums decrease and lower sums increase when points are added to a partition, we have U( f , X) − L( f , X) < ε/2 Then, with X = (x0 , ..., xN ), and U(g, X) − L(g, X) < ε/2

72

Ambar N. Sengupta

N

U( f + g, X) − L( f + g, X) = = ≤

j=1 N j=1 N j=1

∑

M j ( f + g) − m j ( f + g) ∆x j

∑ Var( f + g, [x j−1, x j ])∆x j

N j=1

**∑ Var( f , [x j−1, x j ])∆x j + ∑ Var(g, [x j−1, x j ])∆x j
**

by (3.48)

(by Lemma 2 (ii)) = U( f , X) − L( f , X) + U(g, X) − L(g, X) < ε/2 + ε/2 = ε.

Thus, by the Darbox criterion, f + g is Riemann integrable. The other results follow in a similar way by applying other parts of Lemma 2. QED Consider the function g on [0, 1] given by g(x) = 1 if x is rational −1 if x is irrational

Then g is not Riemann integrable, because no matter what partition X we take of [0, 1] the upper sum is always U(g, X) = 1 and the lower sum is always L(g, X) = −1 But the absolute value of g is the constant function 1: |g| = 1 and so |g| is Riemann integrable.

3.9 C[a, b] ⊂ R [a, b]

Every continuous function on a compact interval is integrable. This is a central result of integration theory.

b] ⊂ R.10 The Integral as a Non-negative Linear Functional R [a. b] ⊂ R [a. b]. b] → R : f → Z b f a (3.52) and.53) Thus.54) is a linear functional on the linear space R [a. we have Z b Z b Z b ( f + g) = a a f+ a g (3. is a linear space. with a < b. with a < b. . that sums and constant multiples of Riemann integrable functions are again Riemann integrable. the set C[a. i. and any f .e. b] is contained in the set R [a. b] of continuous functions on [a. any function which is dicontinuous at only ﬁnitely many points is integrable. b] ⊂ R.51) Theorem 23 For interval [a.Notes in Introductory Real Analysis 73 Theorem 22 For any interval [a. It should be noted that discontinuous functions might also be integrable. Z b Z b (k f ) = k a a f (3. where a < b. b] → R : f → is linear: Z b f a (3. b]: C[a. b] We have seen that the set of all Riemann integrable functions on a compact interval [a. Our next objective is to show that the Riemann integral viewed as a function R [a. Indeed. b]. b] ⊂ R. for any k ∈ R. The key result used is Lemma 3 which says that the variation of a continuous function can be controlled suitably to apply the Darboux criterion for integrability. b] The proof has been presented in class. 3. the Riemann integral R [a. b] of Riemann integrable functions on [a. g ∈ R [a.

we have just repeated the proof of Theorem 21 (ii). there is a partition X of [a. As usual. g ∈ R [a.74 Ambar N. (Of course. X) + L(g.t∈[x j−1 . in Theorem 21. X)] (3.58) which. by the Darboux criterion implies that f + g ∈ R [a. b]. X)] < ε (3. X) + L(g. X) = ≤ j=1 N j=1 ∑ Vark ( f + g)∆ j x N j=1 ∑ Vark ( f )∆ j x + ∑ Vark (g)∆ j x = U( f . let Vark (h) denote the variation of a function h over the interval [x j−1 . X) +U(g. x j ]: Vark (h) = Then N sup s. X) Thus. X) + L(g. b]. We have already seen. Sengupta Proof Let f .55) U( f + g. b].x j ] (h(s) − h(t)) = M j (h) − m j (h) (3. X) +U(g. b]. X) are squeezed into the interval [L( f . xN ) be any partition of [a.57) (3. X) − L(g.) . X) − L( f + g. X). X) +U(g. b] such that U( f .U( f . X) − [L( f .. By the usual trick of combining partitions. and k ∈ R. .. Now let X = (x0 . X) and U( f + g.. X) − L( f . X) < ε/2 and U(g. X) < ε/2 So U( f + g. that f + g and k f are also in R [a. X) ≥ L( f . X) ≤ U( f .56) Now let ε > 0. X) Similarly. L( f + g. L( f + g.

X).U( f .Notes in Introductory Real Analysis Let Z b Z b 75 I( f ) = a f.62) (The 1+ in the denominator is to avoid trouble if k happens to be 0.63) (3. consider the function k f . I( f ) + I(g) and I( f + g) differ by less than ε. X) Putting all this together. . X)] ε ≤ |k| 1 + |k| < ε. X) +U(g. we see that I( f + g) and the sum I( f ) + I(g) both lie in the interval [L( f . X).60) Z b I( f + g) = a ( f + g) Then I( f ) lies between U( f . b] such that U( f .59) (3.. X) − L( f . Consequently. X) + L(g. X) +U(g. But ε is any positive real number. X) + L(g. L( f .. I(g) = a g. and I(g) lies between U(g.) Then N U(k f . X) Moreover. X) ≤ I( f + g) ≤ U( f + g.61) and the width of this interval is < ε. (3. X) and L(g.. X) ≤ I( f ) + I(g) ≤ U( f . X) = = j=1 N j=1 ∑ Var j (k f )∆x j (by Lemma 2 (iii)) ∑ |k|Var j ( f )∆x j = |k| [U( f . I( f + g) = I( f ) + I(g) Next. L( f + g. X) − L(k f . X) < ε 1 + |k| (3. . Therefore. X)] (3. where k ∈ R. X) − L( f . X) and L( f . xN ) be a partition of [a. Let X = (x0 . Therefore. X).

Sengupta has width less than ε.t]) = sup k f (x) = k sup f (x) = kM( f . [s.76 Thus.t]) = kM( f .t]) = sup k f (x) = k inf f (x) = km( f . x∈[s. kU( f .b] (3.64) Now for k ≥ 0 we have M(k f .t]) = km( f . [s.t] x∈[a. X)] Ambar N. also m(k f . Let Z b I(k f ) = a (k f ) (3. [s. [s. Consequently.70) if k ≥ 0 . Thus. and mutiplying everything by ∆x j and adding up.t]) if k ≤ 0 (3.U(k f . [s.t]) if k ≥ 0 m(k f . I(k f ) − kI( f ) ≤ |k| [U( f . QED The Riemann integral is a non-negative linear functional in the sense that it carried non-negative functions into non-negative numbers: (3. The Darboux criterion implies that k f is integrable. and so kI( f ) lies in the same interval mentioned above as I(k f ) does. Now since ε > 0 is any positive real number we have I(k f ) = kI( f ). X)] if k ≤ 0 Now I( f ) lies between L( f . x j ].69) Doing this for each interval [x j−1 . the interval [L(k f . X) − L( f . X)] < ε. as before.t]) x∈[s. X). kL( f . we see that I(k f ) lies in the interval [kL( f . X)] and it lies in [kU( f .65) and for k < 0 we have M(k f . X). [s.t] x∈[a.68) (3.66) because multiplying by a negative number reverses inequalities and transforms sup into inf. X). and inf into sup. [s.b] (3. This completes the proof. X).t]).67) (3. X) and U( f . [s.

b] and Z b a f ≤ Z b a |f| (3. b] is non-negative. b] we have Z b a f ≤ || f ||sup (b − a) (3. and for any f ∈ R [a. Z b a f≥ Z b g. b] ⊂ R. g ∈ R [a. b] and f ≥ g imply that Rb a Rb a f ≥0 (3. b] ⊂ R with a < b. Therefore. This is simply because if f ≥ 0 then all the lower sums are ≥ 0 and so the R integral ab f . with a < b.Notes in Introductory Real Analysis 77 Theorem 24 If [a. Z b Z b Z b f= a a ( f − g) + g a Now we have just shown that the ﬁrst term on the right is ≥ 0. b] for the sup-norm in the following sense: Theorem 25 For any compact interval [a.e. b] then | f | ∈ R [a. of course. b] and f ≥ 0 imply that Consequently. and f ∈ R [a. b]. b] then ab f ≥ 0: f ∈ R [a. b] and f ≥ g.71) f≥ Rb a g (3.73) Proof We know that if f ∈ R [a. is also ≥ 0. Thus. Next. f = ( f − g) + g and so.74) . Observe that f − g = f + (−1)g is also in R [a. ≥ 0. ∈ R [a. by linearity. suppose f . i. a QED The linear functional given by the Riemann integral is a bounded linear functional on R [a. the integral is order-preserving f .72) Proof. f (x) ≥ R 0 for all x ∈ [a. and is. being ≥ all lower sums.

X) − L( fn .77) 2 j=1 Now Var j ( fn ) − Var j ( f ) ≤ Var j ( fn − f ) ≤ 2|| fn − f ||sup ε ε < 2 = 4(b − a) 2(b − a) and so Var j ( f ) ≤ Var j ( fn ) + Therefore.. b] ⊂ R. b] converge uniformly to a function f on [a. with a < b..78 Ambar N.78) U( f . QED A useful but simple consequence of this result is: Theorem 26 For any compact interval [a. xN ) of [a. b] and Z b n→∞ a Z b lim fn = f a (3.76) 4(b − a) By integrability of fn we know that there is a partition X = (x0 .. . b] then f ∈ R [a.75) Proof First let us show that f ∈ R [a. 2 2 . b]. Let ε > 0.. b] such that N ε U( fn . X) = ∑ Var j ( fn )∆x j < (3. Z b a |f| ≤ Z b a || f ||sup = || f ||sup (b − a) and we are done. we have an n ∈ P such that ε || fn − f ||sup < (3. if f1 . f2 . N ε 2(b − a) (3. Sengupta Now the function | f | is bounded above by the constant || f ||sup (recall from Theorem 19 that this is ﬁnite. X) − L( f . X) = ≤ < j=1 N j=1 ∑ Var j ( f )∆x j ε 2(b − a) N j=1 ∑ Var j ( fn)∆x j + ∑ ∆x j ε ε + = ε.. ∈ R [a. By uniform convergence. .) Therefore.

i. Z) < ε/2 (3. then F|T denotes the restriction of F to the smaller domain T . and consider any function f : [a. b]. Then f ∈ R [a. Next. and Z b Z c Z b f= a a f+ c f (3. c].Y ) − L( f |[a. c] and a partition Z of [c. c. F|T is the function whose domain is T and whse values are given through F: (F|T )(x) = F(x) for all x ∈ T (3. c]. This yields a partition X of the combined interval [a. we have Z 79 fn − Z f ≤ || fn − f ||sup (b − a) → 0. b]. By Darboux. b].80) Proof. b). c].e. b]. there is a partition Y of [a. b]. f ∈ R [a.82) Now put together the points of Y and Z. b] such that U( f |[a. U( f . If F :S →U is a function. c] and [c. b] is the sum of the integrals of the function over [a.Y ) +U( f |[c.Y ) < ε/2 (3. as n → ∞. b]. b be real numbers with a < c < b.11 Additivity of the Integral We will show that the integral of a function over an interval [a. b] → R which is intergable over [a.81) and U( f |[c. Z) ≥ Z c Z b f+ a c f (3. Z) − L( f |[c.83) . b]. Then. b] for any point c ∈ (a.Notes in Introductory Real Analysis Thus. and T is a non-empty subset of S.79) Theorem 27 Let a. c] and over [c. Let ε > 0. QED 3. X) = U( f |[a.

X) and U( f . Therefore. b]. b]. and. Z)] U( f |[a. by Darboux. Z) ≤ Consequently. . b] such that U( f .Y ) +U( f |[c. Z) ε/2 + ε/2 ε Therefore. X) = = < = Ambar N. Let ε > 0. Add to Y the points s and t. of course. it follows that Z b Z c Z b f= a a f+ c f. Since f ∈ R [a. QED Now we prove that if a function is ingtergable on an interval then it is integrable on any sub-interval: Theorem 28 If [a..Y ) − L( f .. c]. c]. and if s. Sengupta Z c Z b f+ a c f (3. b].83) and (3.t] ∈ R [s. X) − L( f . Z) − [L( f |[a. b] ⊂ R. f+ Rb c f and Rb a f differ by less than ε.Y ) + U( f |[c.84) U( f |[a. c].. in case they are not in Y .80 and L( f . c]. with a < b. b] we have f |[s. Proof This is. b]. b] with s < t then for any function f ∈ R [a.Y ) − L( f |[a. b]. Since ε > 0 is arbitrary. to obtain a partition Z = (z0 . zM ) .Y ) + L( f |[c.t ∈ [a.Y ) < ε. as always. X).84) it follows that the sum Z c Z b f+ a c f Rb a lies between L( f . c]. f ∈ R [a. a matter of applying Darboux using one of the properties of Var. Z) − L( f |[c. X) = L( f |[a. b].t]. Now from (3. b] there is a partition Y of [a. U( f . so does Rc a f .Y ) + L( f |[c.

X) − L( f . Z) − L( f . We will now prove that every monotone function is Riemann integrable on any compact interval. Z) − L( f . Then N U( f .t] obtained by taking the points of Z which are in [s.t].t]. .12 Monotone Functions are Riemann Integrable We have made the remark that not every Riemann integrable function is continuous. X) < ε. Therefore.Notes in Introductory Real Analysis 81 of [a.85) where J consists of those j ∈ {1.. X) − L( f |[s.. Then U( f . 1≤ j≤N . We know that this lowers the upper sum and raises the lower sum and so U( f . z j ] is not contained in [s. where [a.. QED 3. Now let X be the partition of [s. Z) < ε. U( f .t].t]. b]... X) = ≤ where j=1 ∑ Var j ( f )∆x j N j=1 ∑ Var j ( f ) ||X|| ||X|| = max ∆x j .t]. b] ⊂ R and a < b. Let X = (x0 .. Z) ≤ U( f |[s. M} for which the interval [z j−1 . x1 . b]. b] → R be a monotone function. xN ) be any partition of [a. Z) = U( f |[s.t]. Let f : [a. X) − L( f |[s. and we are done. . Z) − L( f . X) + ∑ M j ( f ) − m j ( f ) ∆x j j∈J (3.

X) = Var( f . Sengupta is the maximum width of the intervals making up the partition.. y ∈ [a.41) we have Var j ( f ) = f (x j ) − f (x j−1 ) for all j ∈ {1. [a. For example. N 1 + Var( f . y ∈ [a. X) − L( f .87) To make this less than any chosen ε > 0 all we have to do is take a partition X with all the interval sizes less than ε/[1 + Var( f . b] ⊂ R.. i.e. b] with x ≤ y Then. . Suppose for convenience that f is monotone non-decreasing. b] with x ≥ y Thus. b])||X|| (3. [a. .86) The same conclusion holds even if f is monotone non-increasing. we have U( f . b] into N equal pieces. the sum of the variations over all the intervals is simply the variation over the full interval: N j=1 ∑ Var j ( f ) = Var( f . in either case. b].. b])]. [a. we could divide [a. i. by the property of Var for monotone functions given in (3.e. then f ∈ R [a. f (x) ≤ f (y) for all x. [a. with N chosen large enough that b−a ε < .82 Ambar N. with a < b. if f (x) ≤ f (y) for all x. b]) (3. b]) Thus we have proved: Theorem 29 If f is a monotone function on a compact interval [a. N} Therefore.

X) If f is integrable. This method is also amenable to generalizations such as the notion of line integrals.. However. X) ≤ S( f . X. xN ) be any partition of [a. X) − L( f . and so L( f . i. . . X ∗ ) = Now m j ≤ f (x∗ ) ≤ M j . X) < ε. . b] → R where [a. with x∗ ∈ [x j−1 . We denote this by j X∗ < X Recall the Riemann sum S( f . this method is not the most intuitive in understanding concepts such as arc length.. b] ⊂ R with a < b. j for each j..88) (3.e.. X.. the Riemann sum approach motivates the construction of more advanced notions such as the stochastic integral of Itô. x j ] for each j. b]. Recall that the norm or width of X is the length of the largest interval ||X|| = max ∆x j j Now consider any sequence ∗ ∗ X ∗ = (x1 . xN ) subordinate to X. Furthermore.Notes in Introductory Real Analysis 83 3. X ∗ ) ≤ U( f . This approach led to a smooth development of the central results of the theory. with I = that Rb a.. Let X = (x0 . It is therefore useful to understand the Riemann integral in terms of Riemann sums as well. So consider a function f : [a.89) then for any ε > 0 we can choose partition X such U( f .13 Riemann Sums and the Riemann Integral We have used the Archimedean strategy of capturing the value of the integral between upper sums and lower sums. j=1 ∑ f (x∗j )∆x j N (3.

taking the inﬁmum over all possible x1 in [x0 . taking the supremum over all possible x1 in the ﬁrst interval [x0 . it follows that |S( f . Z b (3. X ∗ ) − I| < ε/4 holds for all partitions X of [a. x1 ]. we have N I − ε/4 ≤ m1 ∆x1 + ∑ f (x∗ )∆x j j j=2 N Carrying this successively for j = 2... X ∗ ) − I| < ε for every partition X of norm < δ and every X ∗ < X. Then there is a real number I such that for any ε > 0 there is a δ > 0 for which the condition |S( f . X. b] → R is Riemann integrable if and only if there is a real number I such that for any ε > 0 there is a δ > 0 such that |S( f . In this case. I − ε/4 < (3. we see that M1 ∆x1 + ∑ f (x∗ )∆x j ≤ I + ε/4 j j=2 ∗ and..84 Ambar N. X) ≤ U( f . X. X) ≤ I + ε/4 . N.92) j=1 ∑ f (x∗j )∆x j < I + ε/4 N ∗ for every sequence X ∗ < X. X ∗ ) are squeezed in between the upper and lower sums. . b] of width < δ and all X ∗ < X. x1 ]. Suppose the given condition holds. X. Then. Theorem 30 A function f : [a.91) I= a f Proof. X. Sengupta Since both I and S( f . 3. we conclude that I − ε/4 ≤ L( f .90) The following result is is often used to deﬁne the Riemann integral in alternative approaches to the theory. X ∗ ) − I| < ε (3. Thus.

. because at most N points were added to X to obtain Z. U( f .U( f .Y ) and L( f ..Y ) − L( f . . X) − L( f . Thus.Y ) differ by less than ε: U( f . Z) − L( f ...Y ) − L( f . since both I and the integral ab f are trapped in the interval [L( f . Moreover. there is a partition Y = (y0 .95) We also know by Lemma 1 that U − L for Z differs from that for X by at most 2N|| f ||sup ||X||. Thus. the most U − L for X could be is U( f .Notes in Introductory Real Analysis Consequently. Using our standard trick. For the converse. U( f . Z) ≤ U( f .97) (3. suppose f ∈ R [a. X) ≤ ε/2 < ε 85 and so. X)] whose width is ε it follows that I and Rb a f differ by less than ε. yN ) of [a. By Darboux. xT ) of [a. by Darboux. X).93) (Where we get this from will be clear later.. b] such that U( f . Z) − L( f . b] obtained by combining X and Y .. let Z be the partition of [a. Then U( f . b]. 1 + 2N|| f ||sup (3. X) − L( f . R Z b I= a f.) Consider any partition X = (x0 . But.94) (3. Z) + 2N|| f ||sup ||X|| Thus. X) < ε/2 + 2N|| f ||sup δ (3.96) .Y ) < ε/2 (3. b]. f ∈ R [a.Y ) < ε/2 Now let δ= ε/2 . . b] of norm less than δ: ||X|| < δ We will compare U − L for X with that for Y and conclude that U − L for X is indeed less than ε. Let ε > 0. ε is any positive real number.

86 Ambar N. X) < ε/2 + ε/2 = ε (3. X) − L( f . R Therefore. QED . and so it the integral ab f . X ∗ ) is sandwiched between R the lower sum L( f . X). X. X. X). S( f . X ∗ ) and ab f both lie in the interval [L( f .99) holds for any partition X of width < δ and any X ∗ < X. X. Then the Riemann sum S( f .U( f . Thus. X ) − ∗ Z b f <ε a (3.94) we chose δ just so this right side now works out to ε: U( f .98) Now take any X ∗ < X. X)] whose width is < ε.99) We have shown that for any ε > 0 there is a δ > 0 such that (3. S( f . X) and the upper sum U( f . Sengupta In (3.

Walter. 2001. (Eight Edition. [5] Rudin.) McGraw-Hill.G. Principles of Mathematical Analysis. 1956. Addison-Wesley Professional.) H. The Open Court Publishing Company. On Numbers and Games. Essays on the Theory of Numbers.gutenberg. Surreal Numbers.org/ etext/21016 [3] Hilbert. John H. 1901. Stuttgart. 87 . Richard. David. [2] Dedekind. Peters. (Second Edition) A. 1976.Bibliography [1] Conway. Donald.K. with supplementary material and revisions by Paul Bernays. [4] Knuth. Online at http://www. Grundlagen der Geometrie. Verlagsgesellschaft mbH.1974. (Third Edition.

Sengupta .88 Ambar N.

Appendix A Question Bank Set 1 1. Prove that there is no rational number whose square is 7. 89 .

(i) Show that there is an n0 ∈ N such that 1 <ε n2 for all n ∈ N with n ≥ n0 . Let ε be any positive real number. Show that there is some n1 ∈ N such that 1 2a < ε n for all n ∈ N with n ≥ n1 .90 2. Sengupta (ii) Let a be any real number. . Ambar N.

there is some n ∈ N such that the preceding inequality holds for all natural numbers n ≥ n ).e. Suppose a is a positive real number with a2 < 7. . Show that a+ 1 n 2 91 <ε for all n ∈ N large enough (i.Notes in Introductory Real Analysis 3.

Ambar N. Let S = {x ∈ R : x > 0 and x2 < 7}.92 4. Sengupta (ii) Show that S is bounded above in R. . / (i) Show that S = 0.

Thus.Notes in Introductory Real Analysis 93 (iii) Let a = sup S. equal to 7. . in fact. (Hint: Suppose a2 were less than 7.) (iv) Prove that a2 is. Prove that a2 ≥ 7. there is a real number whose square is 7. Then use the result of Problem 3.

7) ∪ [9. For the set B = {−4. ∞). viewed as a subset of R∗ : (i) B0 = (ii) ∂B = (iii) Bc = (iv) the interior of the complement Bc is (Bc )0 = (v) (B0 )c = (vi) B = (vii) ∂B = (viii) ∂B0 = (ix) The interior of (B0 )c is: (x) The closure of Bc is: .94 Set 2 Ambar N. 8} ∪ [1. Sengupta 1.

Provide brief explanations/answers for the following: (i) If U is an open set then U 0 = 95 (ii) For any set A. ∞) is closed as a subset of R. (iv) If S is closed then its closure S is S itself. (iii) The set [0. i. (A0 )0 = A0 .Notes in Introductory Real Analysis 2. S = S. (v) Is there a subset of R whose boundary in R is all of R? . the interior of the interior of A is the interior of A.e.e. i.

96

Ambar N. Sengupta 3. Give an example of an open cover of (−1, 1) which does not have a ﬁnite subcover.

Notes in Introductory Real Analysis Set 5 1. Consider the partition X = (.25, .5, .75, 1) of [0, 1]. For the function f on [0, 1] given by f (x) = x2 for all x ∈ [0, 1]

97

(i) Work out the upper sum U( f , X)

(ii) Work out the Riemann sum S( f , X, X ∗ ), where X ∗ = (0.1, 0.3, 0.6, 0.8)

98 2. Consider the partition X= 1 2 N , , ..., N N N

Ambar N. Sengupta

of [0, 1]. For the function f on [0, 1] given by f (x) = x2 for all x ∈ [0, 1]

(i) Show that U( f , X) = 1 1 1+ 6 N 2+ 1 N

[Hint: Use the sum formula 12 + 22 + · · · + k2 = k(k + 1)(2k + 1)/6.]

(ii) Show that L( f , X) = 1 1 1− 6 N 2− 1 N

prove that Z 1 0 99 x2 dx = 1 3 using the deﬁnition of the Riemann integral and the results of (i) and (ii).Notes in Introductory Real Analysis (iii) Assuming that f is integrable. .

. Consider the function g on [0. Sengupta Prove that g is not Riemann integrable. 1] given by g(x) = 1 if x is rational −1 if x is irrational Ambar N.100 3.

if f . g ∈ R [a. b].Notes in Introductory Real Analysis 101 4. for a. then f + g ∈ R [a. . Prove that. b]. b ∈ R with a < b.

. Sengupta 1. Write out in a complete and neat way the proof that every continuous function on a compact interval is Riemann integrable.102 Set 6 Ambar N.

y∈[a. Var( f g) ≤ M(| f |)Var(g) + m(|g|)Var( f ) .Notes in Introductory Real Analysis 2. x∈[a. x. b] → R we deﬁne the variation Var( f ) by Var( f ) = sup ( f (x) − f (y)) = M( f ) − m( f ).b] Prove that for functions f and g on [a. For a function f : [a.b] 103 where M( f ) = sup f (x). b].b] andm( f ) = inf f (x). x∈[a.

the Darboux criterion. Work out a simple integral in such a way as to illustrate the deﬁnition of the Riemann integral.) . and then stating and explaining the deﬁnition of the Riemann integral. in your own words. (Please note that every piece of notation you bring in must explained. Explain the notion of the Riemann integral by clearly stating the deﬁnitions of upper sums and lower sums. all properly explained in your own terms. Sengupta 3. Clearly explain.104 Ambar N.

Notes in Introductory Real Analysis Test 1 1. 3} is compact (x) The set (1. ∞) is a closed subset of R∗ (vii) The set [0. 2. Mark True or False: (i) If S is a bounded subset of R then it contains a largest element 105 (ii) If S is a bounded. ∞] (iv) The point 1 is an interior point of [0. non-empty subset of R then it has a least upper bound in R (iii) The point ∞ is a boundary point of [1. 1] (v) The point 4 is a boundary point of [0. 5] is compact . 4) (vi) The set [0. ∞) is a closed subset of R (viii) Every real number is a boundary point of Q (ix) The set {1.

. Prove one of the following statements: (i) The union of any collection of open sets is open. Sengupta (ii) If S is a non-empty subset of R∗ then sup S is in the closure S.106 2. Ambar N. (ii) The complement of a closed set is open.

Notes in Introductory Real Analysis 3. . State (i) the Heine-Borel Theorem 107 (ii) the Bolzano-Weierstrass theorem.

108 4. Sengupta (ii) Limit of a sequence (xn )in R∗ . . Ambar N. (iii) A Cauchy sequence in R. State the deﬁnitions of the following: (i) Limit point of a sequence (xn ) in R∗ .

as n → ∞. pointwise. Mark True or False: (i) A continuous function on any subset of R is bounded. (viii) Every uniformly Cauchy sequence of functions converges uniformly. (ii) A continuous function on any closed subset of R is bounded. (iii) A continuous function on a compact set is bounded. (x) The uniform limit of a sequence of continuous functions is continuous. and fn → f . (vi) If a sequence of functions converges to a function unformly then it converges pointwise. (vii) If fn : [0. (ix) Every uniformly convergent sequence of functions is uniformly Cauchy. (v) If a sequence of functions converges to a function pointwise then it converges uniformly. 1]. 1] → R is a sequence of functions.Notes in Introductory Real Analysis Test 2 1. and if each fn is continuous at a point p ∈ [0. . 109 (iv) A continuous function on a compact set attains a mimimum value at some point in the set. then f is continuous at p.

Show that there is a neighborhood U of p such that f (x) > 0 for all x ∈ U ∩ [0. . Sengupta 2. Suppose f : [0.110 Ambar N. 1]. and f (p) > 0 for some p ∈ [0. 1] → R is continuous. 1].

State (i) the Intermediate Value Theorem 111 (ii) the Extreme Value theorem.Notes in Introductory Real Analysis 3. .

Sengupta . 1). Ambar N. Show that the equation x3 + 5x − 2 = 0 has a solution in the interval (0.112 4.

Notes in Introductory Real Analysis Test 3 1. Consider a function f on an interval [a, b] ⊂ R, where a < b,: f : [a, b] → R.

113

(i) Let X = (a, b) be the simplest partition of [a, b]. Write down an expression for L( f , X), explaining your notation clearly. (5pts)

(ii) Now consider a partition X which contains one addtional point t, i.e. X = (a,t, b), where a < t < b. Write down an expression for L( f , X ), again explaining the notation clearly. (5pts)

114

Ambar N. Sengupta

(iii) Show that L( f , X ) ≥ L( f , X).

(5pts)

Notes in Introductory Real Analysis

115

2. Consider a function f : [a, b] → R, where [a, b] ⊂ R and a < b. Let X and Y be partitions of [a, b] with Y containing all the points of X and some more. Prove that (5pts) U( f ,Y ) − L( f ,Y ) ≤ U( f , X) − L( f , X)

116 3. Sengupta (5pts) . If f ∈ R [a. d] ⊂ [a. b] and [c. d]. b] prove that f ∈ R [c. Ambar N.

Mark TRUE or FALSE: 117 (10pts) a. b] → R and any partition X = (x0 . Every Riemann integrable function is bounded. j. X) = j=1 ∑ Var j ( f )∆x j . Every monotone function is Riemann integrable. b. . If | f | ∈ R [a. Every bounded function is Riemann integrable.. b] then f ∈ R [a. . b]. For a bounded function on an interval [a. b] and f is never zero then 1/ f ∈ R [a. If f ∈ R [a. If f ∈ R [a. b]. d.Notes in Introductory Real Analysis 4. xN ) of [a. b]. where Var j ( f ) = supx∈[x j−1 . e. there is always a real number which is ≥ all the lower sums and ≤ all the upper sums.. g. f. i. c. For any function f : [a.. h. b] ⊂ R. Every Riemann integrable function is continuous.x j ] f (x). and ∆x j = x j − x j−1 . b] then | f | ∈ R [a.x j ] f (x) − infx∈[x j−1 . b]. Every continuous function is Riemann integrable. X) − L( f . N U( f .

8}. (x) Every sequence in S has a subsequence converging to a point in S. 4) ∪ (5. Consider the set S = [2. (viii) S is compact. (iv) 9 is an isolated point of S. (ii) 4 is an interior point of S. (vii) S is a closed set. 6) ∪ {1. (vi) S is an open set. Mark True or False: (i) 2 is an interior point of S. . Sengupta (ix) Every sequence in S has a subsequence which is convergent. (iii) 4 is a boundary point of S. (v) 1 is an isolated point of S. Ambar N.118 Cumulative Question Set 1.

119 (i) If t a real number with t < sup S.] . First assume that S is bounded above. in this case U ∈ R.Notes in Introductory Real Analysis 2. Let S be a non-empty subset of R. (ii) Show that there is a sequence of elements sn ∈ S such that sn → sup S. [Hint: Let U = sup S. If you have time. Explain why there exists x ∈ S with x > t. Now apply (i) to produce an sn using 1/n for t. try also the case U = ∞.

and f (p) > 2. p + δ). Prove that there is a δ > 0 such that f (x) > 0 for all x in the domain of f which lie in the neighborhood (p − δ. Suppose f is continuous at a point p. . Sengupta 3.120 Ambar N.

e. there is a point p ∈ S such that f (p) = supx∈S f (x). i.] .Notes in Introductory Real Analysis 121 4. Apply Bolzano-Weierstrass. [Hint: Let U = supx∈S f (x). Then choose a sequence of points sn ∈ S such that f (sn ) → U. Prove that f reaches its maximum value on S. Let f be a continuous function on a closed and bounded set S ⊂ R.

b ∈ R. Sengupta 5.122 Ambar N. b]. b] such that fn (x) → f (x) for all x ∈ [a. b] then fn → f uniformly. Let ( fn ) be a sequence of functions on [a. and if each fn is continuous. b]. (ii) If fn → f uniformly then fn (x) → f (x) for all x ∈ [a. Mark True or False: (i) If there is a function f on [a. Let a. (iv) If fn → f uniformly and each fn is continuous then f is continuous. b]. (v) If fn (x) → f (x) for every x ∈ [a. (iii) If ( fn ) is Cauchy in sup-norm then fn → f uniformly. for some function f on [a. then f is continuous. b]. with a < b. .

b] then Z b a 123 f ≤ Z b a |f| (iii) If fn ∈ R[a. b] ⊂ R. X) ≤ Z b f a (v) If f is a function on [a. (ii) If f ∈ R[a. b]. Consider an interval [a. and f ∈ R[a. and fn (x) → f (x) for every x ∈ [a. (iv) If X is a partition of [a. b]. then L( f . b] then Rb Rb a fn → a f . b] with X ⊂ Y then U( f . b].Notes in Introductory Real Analysis 6. and X and Y are partitions of [a. Mark true or false. for n ∈ N. X) < U( f . b]. (i) C[a.Y ) . b] ⊂ R[a. b].

Are you sure?

This action might not be possible to undo. Are you sure you want to continue?

We've moved you to where you read on your other device.

Get the full title to continue

Get the full title to continue reading from where you left off, or restart the preview.

scribd