MA105: Calculus

Lecture 6 (D1)

Shripad M. Garge
IIT Bombay, Mumbai.

August 13, 2018.

Shripad M. Garge, IITB MA105: D1-06

Recall

Derivative of a function,

differentiability =⇒ continuity,

algebraic rules, increment function, chain rule,

maxima, minima,

Rolle’s theorem, Cauchy’s MVT,

August 15 and 22, August 18, 2-2.50 pm,

Shripad M. Garge, IITB MA105: D1-06

Concavity/convexity

A function f is said to be concave upward (or convex) on an

interval (a, b) if the graph of f lies above all its tangents on (a, b)
and f is concave downward (or concave) on (a, b) if graph of f lies
below all its tangents.

Shripad M. Garge, IITB MA105: D1-06

Second derivative test

Theorem
If f 00 (x) > 0 for all x ∈ (a, b) then f is convex and if f 00 (x) < 0 for
all x ∈ (a, b) then f is concave.

A point where the curve changes its direction of concavity is called

an inflection point. Thus, there is a point of inflection at any point
where f 00 changes sign.

Theorem
Suppose f 00 is continuous near c.
If f 0 (c) = 0 and f 00 (c) > 0 then f has a local minimum at c.
If f 0 (c) = 0 and f 00 (c) < 0 then f has a local maximum at c.

Shripad M. Garge, IITB MA105: D1-06

L’Hospital’s rules

Our limit theorems tell us that the limit of a quotient can be

computed provided the denominator limit is non-zero. Moreover, if
the denominator limit is zero and numerator limit is non-zero then
the limit does not exist.

What if both the numerator and denominator limits are zero?

Theorem (L’hospital’s rule)

Let f , g : (a, b) → R be differentiable functions and assume that
g 0 (x) 6= 0 on (a, b). Suppose that limx→c f (x) = 0 = limx→c g (x)
for some c ∈ (a, b). If

f 0 (x) f (x)
lim =L∈R then lim = L.
x→c g 0 (x) x→c g (x)

Shripad M. Garge, IITB MA105: D1-06

L’Hospital’s rules

Examples:
sin x cos x
lim = lim = 1,
x→0 x x→0 1

1 − cos x sin x cos x 1

lim = lim = lim = .
x→0 x2 x→0 2x x→0 2 2
ln x
lim = 1.
x→1 x − 1

Shripad M. Garge, IITB MA105: D1-06

Taylor’s theorem

A very useful technique in the study of differentiable functions is to

approximate them by polynomial functions. Recall the mean value
theorem:

Theorem
Let f : [a, b] → R be continuous and differentiable on (a, b) then
there exists c ∈ (a, b) such that

f (b) − f (a) = f 0 (c)(b − a).

By changing the point a to any x0 ∈ (a, b) and b to any

x ∈ (x0 , b), we can apply MVT to the interval [x0 , x] and get a
point c ∈ (x0 , x) such that

f (x) = f (x0 ) + f 0 (c)(x − x0 ).

Shripad M. Garge, IITB MA105: D1-06

Taylor’s theorem

Now, let us assume that our function f is twice differentiable on

(a, b). Fix x0 ∈ (a, b) and let x ∈ (x0 , b). Define

F (t) = f (x) − f (t) − (x − t)f 0 (t)

and  2
x −t
G (t) = F (t) − F (x0 ).
x − x0
Then G (x0 ) = 0 = G (x) and by Rolle’s theorem there exists
c ∈ (x0 , x) such that G 0 (c) = 0, and hence

f 00 (c)
f (x) = f (x0 ) + f 0 (x0 )(x − x0 ) + (x − x0 )2 .
2
Thus we have approximated f (x) by a linear polynomial in x0 and
the error term involves second derivative at some c ∈ (x0 , x).

Shripad M. Garge, IITB MA105: D1-06

Taylor’s theorem

Doing this inductively, we get following result:

Theorem (Taylor’s theorem)
Let f : [a, b] → R be such that f is (n + 1)-times differentiable on
(a, b). Fix x0 ∈ (a, b), then for any x ∈ (x0 , b) there exists
c ∈ (x0 , x) such that

f (n) (x0 )
f (x) = f (x0 ) + f 0 (x0 )(x − x0 ) + · · · + (x − x0 )n
n!

f (n+1) (c)
+ (x − x0 )n+1 .
(n + 1)!

Thus we have f (x) = Pn (x) + Rn (x) where Pn (x) is a polynomial

of degree ≤ n and Rn (x) is the remainder term.

Shripad M. Garge, IITB MA105: D1-06

Taylor’s theorem

As an application of Taylor’s theorem, let us approximate e up to

an error less than 10−5 .

We consider the function f (x) = e x and we want to find the value

f (1). Let x0 = 0, so that

f (n) (0) f n+1 (c)

f (1) = f (0) + f 0 (0) + · · · + +
n! (n + 1)!

for some c ∈ (0, 1). We need to find c ∈ (0, 1) and n such that
f n+1 (c) ec −5
(n+1)! = (n+1)! < 10 . One finds out that n = 8 works and so
we get
1 1
e ≈ 1 + 1 + + ··· + = 2.71827876984.
2 8!
The error is 0.00000305861.

Shripad M. Garge, IITB MA105: D1-06

Curve tracing

While drawing curves, we usualy draw a few points and join them
suitably.

The picture comes out nice if we

know the concavity/convexity
properties of the graph.

Given a function, we plot a few

points; maxima, minima, points on
the co-ordinate axes, and then join
them appropriately.

This is where calculus comes handy,

to tell us the shape of the graph
between these points.

Shripad M. Garge, IITB MA105: D1-06

Asymptotes

There are three types of asymptotes; vertical, horizontal and slant;

they basically measure limiting linear behavior of the function.

A line x = a is called a vertical asymptote to f if one of the

following happens

lim f (x) = ±∞, lim f (x) = ±∞.

x→a+ x→a−

Shripad M. Garge, IITB MA105: D1-06

Asymptotes

A line y = L is called a horizontal asymptote to f if either

lim f (x) = L or lim f (x) = L.
x→∞ x→−∞

A line y = mx + k is called a
slant asymptote to f ifeither
lim f (x) − (mx + k) = 0
x→∞

or
 
lim f (x) − (mx + k) = 0.
x→−∞

Shripad M. Garge, IITB MA105: D1-06

Symmetries of curves

A function f is called even (respectively, odd) if f (−x) = f (x)

(respectively, if f (−x) = −f (x)).

If f (x) = f (x + p) for a p > 0 and for all x, then f is called a

periodic function with period being the smallest such p.

Shripad M. Garge, IITB MA105: D1-06

Curve tracing

The key to curve tracing is

DISA&
% ext 

where

D Domain of the function,

I Intercepts, these are the intersections of the curve with
co-ordinate axes,
S Symmetries of the curve, f being even/odd/periodic,
A Asymptotes, vertical/horizontal/slant,
&
% intervals where f is increasing/decreasing,
ext relative/global extrema, and
 concavity/convexity behaviour of f .

Shripad M. Garge, IITB MA105: D1-06

Integration: history

Newton and Leibnitz studied integration as a method of finding

areas and volumes. Soon they realised that this process, which
involves summing, is inverse to the process of differentiation, which
is the process of taking differences.
In the 1850s, Bernhard Riemann (1826–1866) adopted a new and
different viewpoint. He separated the integration from its
companion, differentiation, and examined the summation process
itself. He also broadened the scope by considering all functions on
an interval which can be integrated.
Riemann’s viewpoint led other mathematicians to invent other
integration theories, most significant being the theory of Lebesgue.
But, to learn that you will have to take MA 403 (real analysis) and
MA 408 (Measure theory).
Walter Rudin: Analysis books.

Shripad M. Garge, IITB MA105: D1-06

Integration: various gadgets

Consider a closed and bounded interval I = [a, b].

A partition of I is a finite, ordered set of points
a = x0 < x1 < x2 < · · · < xn = b, and is usually denoted by P.
The points in a partition are used to divide the interval I into
non-overlapping subintervals, [x0 , x1 ], . . . , [xn−1 , xn ].

If a point ti is selected from each subinterval [xi , xi+1 ] then we call

the partition to be tagged. It will be denoted by Ṗ. Tags can be
selected in any way.


P = (x0 < x1 < · · · < xn ) and Ṗ = [xi , xi+1 ], ti : i = 0, . . . , n−1 .

We define the norm of a partition P to be the number

kPk := max{x1 − x0 , x2 − x1 , . . . , xn − xn−1 }.

For a tagged partition Ṗ we define kṖk = kPk.

Shripad M. Garge, IITB MA105: D1-06

Riemann sum

If f : [a, b] → R is a function, the Riemann sum of f corresponding

to a tagged partition Ṗ of [a, b] is the number
n−1
X
S(f , Ṗ) := f (ti )(xi+1 − xi ).
i=0

Shripad M. Garge, IITB MA105: D1-06

Integration: Definition

A function f : [a, b] → R is said to be Riemann integrable over

[a, b] if there exists a number L ∈ R such that for every  > 0
there exists δ > 0 such that if Ṗ is any tagged partition of [a, b]
with kṖk < δ, then
|S(f , Ṗ) − L| < .

The number L, in this case, will be called the Riemann integral of

f over [a, b]. We denote it as
Z b Z b Z b
f or f (x)dx or f (t)dt.
a a a

R[a, b] = the set of all Riemann integrable functions over [a, b].

Shripad M. Garge, IITB MA105: D1-06

Integration: Examples

• If f : [a, b] → R is the constant function f (x) = c then for any

tagged partition Ṗ the Riemann sum is c(b − a). Hence constant
functions are always Riemann integrable with
Z b
c dx = c(b − a).
a

• Let us now consider a step function f : [a, b] → R defined by


m a≤x <c
f (x) =
n c ≤ x ≤ b.

We assume m, n > 0. Let Ṗ be a tagged partition of [a, b] with

kṖk < δ. Let P˙1 be the subset of Ṗ consisting of tags in [a, c)
and P˙2 be the remaining subset consisting of tags in [c, b].

Shripad M. Garge, IITB MA105: D1-06

Integration: Examples

The union of all subintervals in P˙1 contains [a, c − δ) and is

contained in [a, c + δ). So the Riemann sum is bounded by
m(c − a − δ) ≤ S(f , P˙1 ) ≤ m(c − a + δ).
Similarly, n(b − c − δ) ≤ S(f , P˙2 ) ≤ n(b − c + δ). Thus we get

m(c−a)+n(b−c)−(m+n)δ ≤ S(f , Ṗ) ≤ m(c−a)+n(b−c)+(m+n)δ

and hence
 
S(f , Ṗ) − m(c − a) + n(b − c) < 

whenever Ṗ is a tagged partition of [a, b] with kṖk < m+n .

Thus, a step function is integrable over [a, b] and its integral is

given by summing the integrals of its steps.

Shripad M. Garge, IITB MA105: D1-06