Mathematics-I

Tarakanta Nayak∗

1 Introduction
Mathematics is playing with ideas-creating them, understanding them, realizing them and
possibly making them work for a purpose, hopefully for a good one !
There are different levels of existence. Tables exist, so also ghosts! Colors exist though
not for a blind. Mathematics talks of things that exist independent of the observers. Ideas
exist. Plato’s world of ideas. Ideas are sometimes more important than objects. Is not the
idea of 5 more important than 5 mangos at a certain point of time?
Think, think and think. Meditate upon an idea you wish to grasp. In order to see the
ideas, your eyes should possibly be closed, the thought factory must stop working.
Ask, ask and ask. Ask anything that comes to your mind on the subject. Though each
of them may not be addressed upto your satisfaction, you should not stop asking. Let
us agree that silly questions donot exist. If asking a teacher is good, it is better to ask
yourself before coming to a teacher. That not only refines the question but also prepares
you to understand the answer effectively when it is presented to you. The ability to frame
a non-trivial question is not less than a talent by itself.
Questioning is the first step towards questionlessness.
Try to relate what you learn and what you have already learnt. This helps in effective
assimilation of new knowledge.
Mathematics is not to be read. Do it to understand it, to realize it, to enjoy it and to
love it. Be able to be delighted by doing mathematics. That is the key to excel.
Understand. Do not try to remember. Acquire the capacity to understand, it will be
remembered automatically.
∗
Delivered to B Tech students of IIT Bhubaneswar during Nov-Dec 2020, Email: [email protected]

1
T. Nayak IIT Bhubaneswar

2 Review of Basics
A set is a collection of well defined objects, not necessarily similar. Being similar depends
on the point of views and we decide to be independent of them to deal with greater truths.
A relation between two non-empty sets A and B is a subset of A × B. A function from A
to B is a specal type of relation. It is a way to associate each element of a non-empty set
A to that of a unique element of B. We denote it by f : A → B. The set A is called the
domain and B is called the range of f . It can be one-one or many-one.

2.1 Convergence of Sequences and Limits of functions

A sequence of real numbers is a function (not a relation) from N to R.

A sequence of real numbers xn converges to x ∈ R if for every positive real number

α there is a natural number K(α) such that |xn − x| < α for all n > K.

• The tail of the sequence is important!

• Sakha Chandra argument for existence of god !

• The K(α) game.

• The limit is unique.

• A sequence {xn } is said to be α−close to a real number x if |xn − x| < α for all n.

• It is called ultimately α−close to a real number x if |xn − x| < α for all n > n0 .

• xn → x if and only if xn is ultimately α−close to x ∀ α > 0.

Noting that |x − y| < α for all α > 0 means x = y, it is observed that xn is equal to
x ultimately. Converegence of sequences in any metric space (any non-empty set with a
notion of distance) is defined in the same way.
1 (−1)n
Example 1. The sequence n3 +n
is convergent (to 0) whereas n+2
is not.

To understand a sequence which is not convergent, two ideas are useful, namely lim
sup and lim inf.
Let {xn }n>0 be a bounded sequence and for each n, let sn = sup{xk : k ≥ n} and
tn = inf{xk : k ≥ n}. Note that sn is a monotonically decreasing and bounded below

2
T. Nayak IIT Bhubaneswar

sequence. Similarly, tn is monotonically increasing and bounded above sequence. By the

Monotone convergence theorem, sn and tn are convergent. The numbers limn→∞ sn and
limn→∞ sn are called limit superior and limit inferior respectively. Note that the lim sup
and lim inf are the supremum and infimum respectively, of all subsequential limits of the
sequences.

A function f : A ⊆ R and a ∈ A, we say the limit of f at a exist, and is l, written as

limx→a f (x) = l if for every  > 0 there is a δ > 0 such that |f (x) − l| <  whenever
|x − a| < δ.

• The right hand limit: limP x → a+ f (x) = l0 if for every  > 0 ∃δ > 0 such that
|f (x) − l0 | <  whenever 0 < x − a < δ.

• The left hand limit: limP x → a− f (x) = l0 if for every  > 0 ∃δ > 0 such that
|f (x) − l0 | <  whenever −δ < x − a < 0.

• f is −close to l near a for every .

• The limit is unique, whenever exist.

• limx→a f (x) = l if and only if f (xn ) → l for all sequence xn converging to a.

2.2 Continuity
Intuitively, a continuous function is one without abrupt change in its value. In Science
and Engineering, several quantities are continuous functions. The temperature at a par-
ticular point of a room varies with respect to time. Considered as a function of time, it is
continuous in a normal set of conditions.
A function f : A ⊆ R → R is continuous at x0 if for every  > 0, ∃ δ > 0 such that
|f (x) − f (x0 )| <  whenever |x − x0 | < δ.

• The limit of f at x0 is f (x0 ).

• f is -close to f (x0 ) near x0 for every .

• A smaller δ also works, but not the bigger ones!

3
T. Nayak IIT Bhubaneswar

A function can be discontinuous in a number of different ways. A comparatively well

behaved situation is when the left hand as well as the right hand limit exist at a point
but they are not equal. This is called jump discontinuity. The greatest integer function
f (x) = [x] has jump discontinuity at every integer.
When the left hand and right hand limit exist and are equal but is different from the
value of the function at the point, it is called removable discontinuity. The discontinuity
can be removed by redefining the function at the concerned point. The function defined
x2 −3
√ √
by f (x) = x− √ , x 6= 3 and f (x) = 1 has removable discontinuity at 3.
3

• (Sign preserving property) If f is continuous at c and f (c) 6= 0 then f (x) > 0 for all
x ∈ (c − δ, c + δ) for some δ > 0.

Proof. Assume f (c) > 0. For every  > 0, ∃δ > 0 such that |f (x)−f (c)| <  whenever
|x − c| < δ. In other words, f (c) −  < f (x) < f (c) +  whenever |x − c| < δ. Set  =
f (c)
2
> 0. Then ∃δ() such that f (c)
2
< f (x) < 3f2(c) whenever c − δ() < x < c + δ().
Now f (x) > f (c)
2
> 0. For f (c) < 0, use the same argument.

• (Bolzano Theorem)If f : [a, b] → R is continuous and f (a)f (b) < 0 then there is a
c ∈ (a, b) such that f (c) = 0.

Proof. Suppose that f (a) < 0. Then f (b) > 0. Consider S = {x ∈ [a, b] : f (x) ≤ 0}.
Clearly, S 6= ∅ and is bounded above (by b, for example). In fact, there are many
upper bounds for S. Let c = inf S, the least upper bound of S. It exists and is
finite. If f (c) > 0 then by the sign preserving property of f at c, there is a δ > 0
such that f (x) > 0 for all x ∈ (c − δ, c + δ). or (c − δ, b) whenever b = c. Now, c − δ
is an upperbound for S. But c is the least upper bound. This is a contradiction.
A contradiction can be arrived at similarly when f (c) < 0. The proof completes by
proceeding in a similar way for f (a) < 0.

• (Intermediate value theorem) Let f : [a, b] → R be continuous and x1 , x2 ∈ [a, b]. If

f (x1 ) 6= f (x2 ) then, for each α between f (x1 ) and f (x2 ) there is a c ∈ (x1 , x2 ) such
that f (c) = α.

Proof. Consider g(x) = f (x) − α and apply the Bolzano theorem. It is possible to
have more than one c as given above!

• If f : [a, b] → R is continuous then f attains its maximum and minimum in [a, b].

4
T. Nayak IIT Bhubaneswar

2.3 Differentiability
Continuous functions are not sufficiently suitable for the purposes of application. More
sophisticated functions are required.
f (x)−f (x0 )
A function f is differentiable at x0 if limx→x0 x−x0
exists and is a real number.

• f 0 (x0 ) is the slope of the tangent to the curve y = f (x) at (x0 , f (x0 )).

• f (x) = f (x0 ) + f 0 (x0 )(x − x0 ) + O(x − x0 ) where O(x − x0 ) is a function such that
O(x−x0 )
x−x0
→ 0 when x → x0 . We say, f (x) is infinitesimally equal to f (x0 ) + f 0 (x0 )(x −
x0 ) near x0 . Note that f (x0 ) + f 0 (x0 )(x − x0 ) = Tf (x0 ) gT−x0 (x) where Ta (x) = a + x
for every a ∈ R and g(x) = f 0 (x0 )x. Thus, f behaves like a linear map g near x0 .

• (Darboux Theorem) If f : [a, b] → R is differentiable, f 0 (a) 6= f 0 (b) and λ is a number

between f 0 (a) and f 0 (b) then prove that there is a c ∈ (a, b) such that f 0 (c) = λ.

Proof. Assume f 0 (a) < λ < f 0 (b). Consider g(x) = λx − f (x). Since g 0 (a) > 0 g is
strictly increasing near a and a is not a local maximum of g. Similarly, g 0 (b) < 0 gives
that b is not a maximum of g. However, g attains its maximum in [a, b]. Therefore,
there is a c ∈ (a, b) such that g(c) is a local maximum which means that g 0 (c) = 0,
and consequently f 0 (c) = λ.
Note that continuity of f 0 is NOT required in the proof.

3 Rolle, Lagrange, Cauchy and Taylor

In a way, the most important point of a differentiable function is one where its derivative
vanishes, i.e. f 0 (c) = 0. Such a c is called a critical point of f .

If f has local maximum/minimum at c and f is differentiable at c then f 0 (c) = 0.

To see it when c is a local maximum, note that the right hand limit of f (x)−f
x−c
(c)
is non
positive and the left hand limit is non-negative. Since the two limits are equal we get
f 0 (c) = 0. The proof is similar when c is a local minimum. However, the converse is not
true. For example, x3 at 0.
Between two points with the same image, there is a critical point.

5
T. Nayak IIT Bhubaneswar

Rolle’s Theorem
Let f : [a, b] → R is continuous, f : (a, b) → R be differentiable and f (a) = f (b).
Then there is at least one c ∈ (a, b) such that f 0 (c) = 0.

Proof. Since f is continuous there are c, d ∈ [a, b] such that f (c) = supx∈[a,b] f (x) and
f (d) = inf x∈[a,b] f (x). If both c, d are end points then f is constant and f 0 (x) = 0 for all x.
Otherwise f assumes its maximum at c. Therefore f 0 (c) = 0.

Between two roots of f there can be more than one root of f 0 , e.g., sin x, x ∈ [0, 4π].
Lagrange’s Mean Value Theorem

Let f : [a, b] → R is continuous and f : (a, b) → R be differentiable. Then there is a

c ∈ (a, b) such that f 0 (c) = f (b)−f
b−a
(a)
.

Proof. The function h(x) = f (x)(b − a) − x(f (b) − f (a)) is continuous on [a, b] and differ-
entiable in (a, b). Note that h(a) = bf (a) − af (b) = h(b). Applying the Rolle’s theorem to
h, a point c ∈ (a, b) is found such that h0 (c) = 0. In other words f 0 (c) = f (b)−f
b−a
(a)
.

Average speed is attained. The speedometer must cross the the mark of average speed
at least once during the journey. The chord joining (a, f (a)) and (b, f (b)) is parallel to the
tangent to the curve y = f (x) at some point c. For a = b, Lagrange’s Mean value Theorem
gives Rolle’s Theorem and in this sense the LMT is a generalisation of Rolle’e Theorem.
Here is a generalisation further.
Cauchy’s Mean Value Theorem

Let f, g : [a, b] → R be continuous and f : (a, b) → R be differentiable. Then there

is a c ∈ (a, b) such that (g(b) − g(a))f 0 (c) = (f (b) − f (a))g 0 (c).

Proof. Apply Rolle’s theorem to h(x) = f (x)(g(b) − g(a)) − g(x)(f (b) − f (a)) on [a, b].

A function (a, b) → R is called increasing (or strictly increasing ) if for a < x1 < x2 < b,
f (x1 ) ≤ f (x2 ) (or f (x1 < f (x2 )) respectively). For x1 < x2 , there is a c ∈ (x1 , x2 ) such
that f 0 (c) = f (xx22)−f
−x1
(x1 )
. Following is the consequence.

6
T. Nayak IIT Bhubaneswar

Increasing and Decreasing Functions

Let f : (a, b) → R be differentiable. Then

• f 0 (x) ≥ 0 for all x =⇒ f is increasing.

• f 0 (x) > 0 for all x =⇒ f is strictly increasing.

• f 0 (x) ≤ 0 for all x =⇒ f is decreasing.

• f 0 (x) < 0 for all x =⇒ f is strictly decreasing.

3.1 Taylor’s series

If a function f has n-th order derivative at a point α, we can construct a polynomial Pn of
degree at most n such that P (α) = f (α), P 0 (α) = f 0 (α), P 00 (α) = f 00 (α), ...P n (α) = f n (α).
00 n
Indeed, Pn (x) can be taken to be f (α) + f 0 (α)(x − α) + f 2!(α) (x − α)2 + ... + f n!(α) (x − α)n .
This Pn is usually referred as the Taylor polynomial of f at α with degree n. But this does
not say when Pn can be equal (at least approximately) to f at other points of its domain.
Taylor’s Theorem does exactly that.
Taylor’s Theorem

Let f : [a, b] → R and n ∈ N. If f n−1 is continuous on [a, b] and f n exists in (a, b)

00
then for every α, β ∈ [a, b], f (β) = f (α) + f 0 (α)(β − α) + f 2!(α) (β − α)2 + ... +
f n−1 (α) n

(n−1)!
(β − α)n−1 + f n!(c) (β − α)n for some c between α and β.

Here unlike some text books, α < β is not assumed!

f (k) (α)
Proof. Note that P (t) = n−1 (t − α)k is a polynomial of degree n − 1 and, f (k) (α) =
P
k=0 k!
P (k) (α) for k = 0, 1, 2, · · · , n − 1. Let M be such that f (β) = P (β) + M (β − α)n . The
claim that
M n! = f (n) (c)
for some c between α and β will complete the proof. Define g(t) = f (t) − P (t) − M (t −
α)n , a ≤ t ≤ b. Observe that g(α) = g 0 (α) = g 00 (α) = · · · = g (n−1) (α) = 0. Since g(α) = 0
and g(β) = 0 by definition, by Rolle’s Theorem, there is a c1 between α and β such that
g 0 (c1 ) = 0. Again by Rolle’s Theorem applied to g 0 on [α, c1 ] (if α < c1 ) or on [c1 , α] (if
c1 < α) there is a c2 between α and c1 such that g 00 (c2 ) = 0. Continuing this we get a cn

7
T. Nayak IIT Bhubaneswar

between α and cn−1 , hence between α and β such that g n (cn ) = 0. Take cn = c. Since P n
vanishes identically, f n (c) = M n!, as desired.

Few important remarks follow.

Remark 1. 1. For α = 0, the series is called the Maclaurin’s series.

2. For n = 1, the Taylor’s theorem is nothing but the Lagrange’s Mean Value Theorem.

3. The function f , considered as a function of β can be written as a polynomial in

(β − α) plus a remainder term for every α ∈ [a, b].

4. If |f n (x)| < K for all x ∈ (a, b), then the error term can be estimated to be bounded
above by K n!
|β − α|n .

If a function is differentiable for any given number of times then it can be written as
an infinite series, called the Taylor series.
Taylor’s Series

Let f (n) exist at all points of an interval I = (a−δ, a+δ). Then the Taylor series of f
f n (a) (x−a)n n
at a is given by ∞ n
P
n=0 n! (x−a) . Here the remainder term is Rn (x, a) = n!
f (c)
where c is a point between a and x. This series converges to f (x) for all x ∈ I
whenever Rn (x, a) → 0 as n → ∞ for all x, c ∈ I.

1
Example 2. Use Taylor’s theorem with n = 2 to approximate (1 + x) 3 , x > −1.
1
Hints: Take f (x) = (1 + x) 3 , x0 = 0 and n = 2. Ans: 1.09.

4 Indeterminate forms
Let f, g be differentiable in (a, b) and g 0 (x) 6= 0 for any x. If limx→s f (x) = 0 = limx→s g(x)
or limx→s f (x) = ∞ = limx→s g(x) then sufficient information is not available toevalu-
ate the limit and limx→s fg(x)(x)
is said to assume an indeterminate form. In this case if
f 0 (x) f (x)
limx→s g 0 (x)
= L then limx→s g(x)
= L. Here s, L can be ±∞.