Alliance School of Applied Mathematics

Semester I
Calculus and Linear Algebra

Module 1: Single-variable Calculus (Differentiation)

Syllabus
Functions of single variable; Limit, continuity and differentiability. Rolle’s
MODULE 1 theorem-geometrical interpretation, Lagrange’s and Cauchy’s mean value
Single-variable theorems and Application; Extreme values of functions; Linear
Calculus approximation; Indeterminate forms and L' Hospital's rule.
(Differentiation)

Motivation:

In Differential Calculus, differentiation is implemented in all fields of applied mathematics.

Mean Value Theorem is one of the important topics of calculus. This theorem relates the slope
of the tangent at one point of the arc of the curve to the slope of its secant. In other words, mean
value theorem relates the average value of the function with it derivative. Mean value theorem
can be applied in the comparison of instantaneous velocity with the average velocity in its
engineering applications. The mean value theorem is a very important result in Real Analysis
and is very useful for analyzing the behavior of functions in higher mathematics.

The operation of adding infinite terms is called infinite series. In mathematics, the enumerated
no of objects when added they give the structure of infinite series which are required for the
study of finite structures with help of generating functions. Infinite series emphasizes methods
for discussing convergence and divergence of series. It also helps in expanding different
functions which can be expressed into convergent series.

Limit
Meaning of 𝒙 → 𝒂:
Let ‘𝑥’ be a variable and ‘𝑎’ be constant. Since 𝑥 is a variable; we can change its value at our
pleasure. It can be changed in such a way that its value comes nearer and nearer to 𝑎. Then we
say that 𝑥 approaches 𝑎 and it is denoted by 𝑥 → 𝑎 .

Definition: Given a number 𝛿 > 0 however small, if 𝑥 takes up values, such that
0 < |𝑥 − 𝑎| < 𝛿. Then 𝑥 is said to tend to 𝑎, and is symbolically written as 𝑥 → 𝑎.

Limit of a function: Let 𝑓(𝑥) be a function defined on (𝑎, 𝑏) except possibly at 𝑐 ∈ (𝑎, 𝑏).
We say that lim 𝑓(𝑥) = 𝐿 if, for every real number 𝜖 > 0; there exists a real number 𝛿 > 0
𝑥→𝑐
such that
𝟎 < |𝒙 − 𝒄| < 𝜹 ⇒ |𝒇(𝒙) − 𝑳| < 𝝐

Important facts on limits:

a) The fact that the limit of 𝑓(𝑥) exists at 𝑥 = 𝑎 means that the graph of 𝑓(𝑥) approaches
the same value from both sides of 𝑥 = 𝑎 .
 Alliance School of Applied Mathematics
Semester I
Calculus and Linear Algebra

b) Right hand limit is the limit of the function as 𝑥 approaches 𝑎 from the positive side.
[RHL= lim+ 𝑓(𝑥) = lim 𝑓(𝑎 + ℎ)]
𝑥→𝑎 𝑥→0
c) Left hand limit is the limit of the function as 𝑥 approaches 𝑎 from the negative side.
[LHL= lim− 𝑓(𝑥) = lim 𝑓(𝑎 − ℎ)]
𝑥→𝑎 𝑥→0

Working Rule for the Evaluation of Limit:

Right hand limit of 𝒇(𝒙), when 𝒙 → 𝒂 = 𝐥𝐢𝐦+ 𝒇(𝒙)

𝒙→𝒂
Step I. Put 𝑥 = 𝑎 + ℎ and replace 𝑥 → 𝑎 + by ℎ → 0.
Step II. Simplify lim 𝑓(𝑎 + ℎ).
𝑥→0
Step III. The value obtained in step 2 is the right hand limit of 𝑓(𝑥) at 𝑥 = 𝑎.

Similarly for evaluating the left-hand limit put 𝒙 = 𝒂 − 𝒉.

Continuity

The word ‘continuous’ means without any break or gap. If the graph of a function has no break
or gap or jump, then it is said to be continuous. A function which is not continuous is called a
discontinuous function.

Definition: A real valued function 𝒇(𝒙) is said to be continuous at 𝒙 = 𝒄 if

(i) 𝒇(𝒙) is defined.
(ii) 𝐥𝐢𝐦 𝒇(𝒙) exists [i.e., 𝐥𝐢𝐦− 𝒇(𝒙) = 𝐥𝐢𝐦+ 𝒇(𝒙)].
𝒙→𝒄 𝒙→𝒄 𝒙→𝒄
(iii) 𝐥𝐢𝐦 𝒇(𝒙) = 𝒇(𝒄)
𝒙→𝒄

➢ A real valued function 𝒇(𝒙) is said to be continuous in [𝒂, 𝒃] if 𝒇(𝒙) is continuous

at every point 𝒄 ∈ [𝒂, 𝒃].

Note: All elementary functions such as algebraic, exponential, trigonometric, logarithmic,

and hyperbolic functions are continuous functions. Also, the sum, difference, and product
of continuous functions is continuous. The quotient of continuous functions is continuous at
all those points at which the denominator does not become zero.
 Alliance School of Applied Mathematics
Semester I
Calculus and Linear Algebra

Continuity from Left and Right:

A function 𝑓(𝑥) is said to be

(i) Left continuous at 𝑥 = 𝑎 if, lim− 𝑓(𝑥) = 𝑓(𝑎).
𝑥→𝑎
(ii) Right continuous at 𝑥 = 𝑎 if, lim+ 𝑓(𝑥) = 𝑓(𝑎).
𝑥→𝑎
Thus, a function 𝑓(𝑥) is continuous at a point 𝑥 = 𝑎, if it is left continuous as well as right
continuous at 𝑥 = 𝑎.

Continuity of a Function in an Interval:

➢ A function f(x) is said to be continuous in an open interval (a, b), if it is continuous at every
point in (a, b). For example, the functions y = sin x, y = cos x, y = ex are continuous
in (−∞, ∞).

➢ A function f(x) is said to be continuous in the closed interval [a, b], if it is:
a) Continuous at every point of the open interval (a, b).
b) Right continuous at x = a, i.e. RHL|x=a = f(a).
c) Left continuous at x = b, i.e. LHL|x=b = f(b).

Reasons of Discontinuity
a) Limit does not exist i.e. 𝐥𝐢𝐦− 𝒇(𝒙) ≠ 𝐥𝐢𝐦+ 𝒇(𝒙)
𝒙→𝒂 𝒙→𝒂
b) 𝒇(𝒙) is not defined at 𝒙 = 𝒂.
c) 𝐥𝐢𝐦 𝒇(𝒙) ≠ 𝒇(𝒂).
𝒙→𝒂

Differentiability
Meaning of a Derivative: The instantaneous rate of change of a function with respect to the
independent variable is called the derivative. Let 𝑓(𝑥) be a given function of one variable and
let ∆𝑥 denote a number (positive or negative) to be added to the number 𝑥. Let ∆f denote the
corresponding change of ‘𝑓’, then
∆𝒇 𝒇(𝒙 + ∆𝒙) − 𝒇(𝒙)
∆𝒇 = 𝒇(𝒙 + ∆𝒙) − 𝒇(𝒙) ⇒ =
∆𝒙 ∆𝒙
∆𝑓
If ∆𝑥 approaches a finite value as ∆𝑥 approaches zero, this limit is the derivative of ‘𝑓’ at the
point 𝑥. The derivative of a function ‘𝑓’ is denoted by symbols such as
𝒅𝒇 𝒅
𝒇′ (𝒙); ; (𝒇(𝒙))
𝒅𝒙 𝒅𝒙

Note: If 𝑦 = 𝑓(𝑥)
𝑑𝑓 ∆𝑓 𝑓(𝑥+∆𝑥)−𝑓(𝑥)
Then 𝑦 ′ = 𝑑𝑥 = 𝑙𝑖𝑚 = 𝑙𝑖𝑚
∆𝑥→0 ∆𝑥 ∆𝑥→0 ∆𝑥

𝒇(𝒙)−𝒇(𝒙𝟎 )
Definition: A real-valued function 𝒇(𝒙) is said to be differentiable at point 𝒙𝟎 if 𝒍𝒊𝒎
𝒙→𝒙𝟎 𝒙−𝒙𝟎
′ (𝒙) 𝒅𝒇
exists uniquely and it is denoted by 𝒇 or 𝒅𝒙.
 Alliance School of Applied Mathematics
Semester I
Calculus and Linear Algebra

➢ The fact that 𝑓(𝑥) is differentiable at 𝑥 = 𝑎 means that the graph is smooth as 𝑥
moves from 𝑎 − to 𝑎 to 𝑎+ . Derivative of an even function is always an odd function.

Important facts on Differentiability:

➢ If a function is not differentiable but is continuous at a point, it geometrically implies

there is a sharp corner or a kink at that point.
➢ If a function is differentiable at a point, then it is also continuous at that point.
➢ If a function is continuous at point 𝑥 = 𝑎, then nothing can be guaranteed about the
differentiability of that function at that point.
➢ If a function 𝑓(𝑥) is not differentiable at 𝑥 = 𝑎, then it may or may not be continuous
at 𝑥 = 𝑎.
➢ If a function 𝑓(𝑥) is not continuous at 𝑥 = 𝑎, then it is not differentiable at 𝑥 = 𝑎.
➢ If the left hand derivative and the right-hand derivative of 𝑓(𝑥) at 𝑥 = 𝑎 are finite (they
may or may not be equal), then 𝑓(𝑥) is continuous at 𝑥 = 𝑎.
➢ Every differentiable function is necessarily continuous, but every continuous function
is not necessarily differentiable i.e. Differentiability ⇒ continuity but continuity ⇏
differentiability
➢ All polynomial, trigonometric, logarithmic and exponential functions are continuous
and differentiable in their domains.
➢ If 𝑓(𝑥) & 𝑔(𝑥) are differentiable at 𝑥 = 𝑎 then the functions 𝑓(𝑥) ± 𝑔(𝑥),
𝑓(𝑥). 𝑔(𝑥) will also be differentiable at 𝑥 = 𝑎 & if 𝑔(𝑎) ≠ 0, then the function
𝑓(𝑥)/𝑔(𝑥) will also be differentiable at 𝑥 = 𝑎.

Differentiability of a Function in an Interval:

➢ A function 𝑓(𝑥) is said to be differentiable over an open interval (𝑎, 𝑏), if it is
differentiable at every point in (𝑎, 𝑏).
➢ A function 𝑓(𝑥) is said to be differentiable over a closed interval [𝑎, 𝑏], if it is:
a) Differentiable at every point of the open interval (𝑎, 𝑏).
b) Right derivative exists at 𝑥 = 𝑎.
c) Left derivative exists at 𝑥 = 𝑏.

Rolle’s Mean value theorems

Introduction: The mean value theorem tells us (roughly) that if we know the slope of the
secant line of a function whose derivative is continuous, then there must be a tangent line
nearby with that same slope. This lets us draw conclusions about the behaviour of a function
based on knowledge of its derivative.

Statement: If a function 𝑓(𝑥) is

(i) 𝑓(𝑥) is continuous on the closed interval [𝑎, 𝑏];
(ii) 𝑓(𝑥) is differentiable on the open interval (𝑎, 𝑏);
(iii) 𝑓(𝑎) = 𝑓(𝑏),
then there exists at least one point 𝑐 in (𝑎, 𝑏) (i.e. 𝑎 < 𝑐 < 𝑏) such that
𝒇′ (𝒄) = 𝟎.
 Alliance School of Applied Mathematics
Semester I
Calculus and Linear Algebra

Alternative or Another Statement of Rolle’s Theorem: If a function 𝑓(𝑥) is

(i) continuous in [𝑎, 𝑎 + ℎ];

(ii) differentiable in (𝑎, 𝑎 + ℎ);
(iii) 𝑓(𝑎) = 𝑓(𝑎 + ℎ),
then there exists at least one real number 𝜃 such that
𝒇′ (𝒂 + 𝜽𝒉) = 𝟎 , 𝒇𝒐𝒓 𝟎 < 𝜽 < 𝟏
Geometrical Interpretation of Rolle’s Theorem: If graph of a function 𝑓(𝑥) is a continuous
curve between 𝑥 = 𝑎 and 𝑥 = 𝑏; having a unique tangent at all point in (𝑎, 𝑏) and 𝑓(𝑎) =
𝑓(𝑏), then there exists at least one point 𝑃 between 𝑥 = 𝑎 and 𝑥 = 𝑏 on the curve, such that
the tangent at this point is parallel to 𝑥 − axis i.e., 𝑓 ′ (𝑐) = 0.

Class Work [Session 2]:

1. Verify Rolle’s theorem for the function 𝑓(𝑥) = 𝑥 3 − 12𝑥 in [0,2√3].
2. In a network simulation, the normalized system response time (in seconds) of a distributed
server is modeled by
sin 𝑥
𝑅(𝑥) = 𝑥 , 𝑥 ∈ [0, 𝜋]
𝑒
where 𝑥 represents the normalized workload level.
Using Rolle’s Theorem, determine whether there exists a workload level at which the
instantaneous rate of change of the response time becomes zero.
3. In a network simulation, the data transfer rate (in normalized units) of a communication
system is modeled by
𝜋 5𝜋
𝑅(𝑥) = 𝑒 𝑥 (sin 𝑥 − cos 𝑥), 𝑥∈[ , ]
4 4
where 𝑥 represents the network load parameter.
Using Rolle’s Theorem, determine whether there exists a network load level within the
𝜋 5𝜋
interval [4 , 4 ] at which the instantaneous rate of change of the data transfer rate becomes
zero.

Self-Practice Problems:
1. Examine whether Rolle's Theorem is applicable for:
a) 𝑓(𝑥) = |𝑥| 𝑖𝑛 [−1,1].
b) 𝑓(𝑥) = tan 𝑥 , 0 ≤ 𝑥 ≤ 𝜋.
𝑥 2 +𝑎𝑏
c) 𝑓(𝑥) = log [(𝑎+𝑏)𝑥] 𝑖𝑛[𝑎, 𝑏], 𝑎 > 0, 𝑏 > 0
 Alliance School of Applied Mathematics
Semester I
Calculus and Linear Algebra

2. The average transmission delay in a hybrid routing network is given by

𝑥 2 + 1, 0 ≤ 𝑥 ≤ 1,
𝑇(𝑥) = {
3 − 𝑥, 1 ≤ 𝑥 ≤ 2.
where 𝑥 represents the routing load parameter.
Using Rolle’s Theorem, determine whether there exists a load level at which the
instantaneous rate of change of the transmission delay has a zero rate of change.

3. The efficiency of data transfer in a cloud network is represented by

𝐸(𝑥) = 1 − 3(𝑥 − 1)2/3 , 0 ≤ 𝑥 ≤ 2
where 𝑥 denotes the normalized network traffic intensity.
Using Rolle’s Theorem, determine whether there exists a traffic intensity at which the rate
of change of transfer efficiency is zero.

Lagrange’s mean value theorem

Statement: Suppose 𝑓(𝑥) be a function that satisfies following conditions:
(i) 𝑓(𝑥) is continuous on the closed interval [𝑎, 𝑏];
(ii) 𝑓(𝑥) is differentiable on the open interval (𝑎, 𝑏).
Then there exists a point 𝑐 ∈ (𝑎, 𝑏) (i.e. 𝑎 < 𝑐 < 𝑏) such that
𝒇(𝒃)−𝒇(𝒂)
𝒇′ (𝒄) = 𝒃−𝒂 .
Note: Rolle’s Theorem is a particular case of Lagrange’s MVT when the value of function is
same at the end points of [𝑎, 𝑏] i.e., 𝑓(𝑎) = 𝑓(𝑏).

Geometrical Interpretation of L.M.V.T.: If

graph of a function 𝒇(𝒙) is a continuous curve
between 𝒙 = 𝒂 and 𝒙 = 𝒃; having a unique
tangent at all point in (𝒂, 𝒃) and 𝒇(𝒂) = 𝒇(𝒃),
then there exists at least one point 𝑪 between
𝒙 = 𝒂 and 𝒙 = 𝒂 on the curve, such that the
tangent at this point is parallel to 𝒙 − axis i.e.,
𝑓(𝑏)−𝑓(𝑎)
𝒇′ (𝒄) = 𝑏−𝑎 .

𝒇(𝒃)−𝒇(𝒂)
Slope of the chord; 𝐭𝐚𝐧 𝜶 = 𝒃−𝒂

Cauchy’s mean value theorem

Statement: Suppose 𝑓(𝑥) and 𝑔(𝑥) be functions such that:

(i) Both are continuous on the closed interval [𝑎, 𝑏];
(ii) Both are differentiable on the open interval (𝑎, 𝑏);
(iii) 𝑔′ (𝑥) ≠ 0, ∀𝑥 ∈ (𝑎, 𝑏),
then there exists a point 𝑐 ∈ (𝑎, 𝑏) (i.e. 𝑎 < 𝑐 < 𝑏) such that
𝒇′ (𝒄) 𝒇(𝒃)−𝒇(𝒂)
= 𝒈(𝒃)−𝒈(𝒂).
𝒈′ (𝒄)
 Alliance School of Applied Mathematics
Semester I
Calculus and Linear Algebra

➢ Cauchy's Mean Value Theorem generalizes Lagrange’s Mean Value Theorem. This theorem
is also called the Extended or Second Mean Value Theorem. It establishes the relationship
between the derivatives of two functions and changes in these functions on a finite interval.

Geometrical Interpretation of C.M.V.T.:

Suppose that a curve 𝛾 is described by the
parametric equations 𝑥 = 𝑓(𝑡), 𝑦 =
𝑔(𝑡), where the parameter t ranges in the
interval [𝑎, 𝑏]. When changing the
parameter t, the point of the curve runs from
𝐴(𝑓(𝑎), 𝑔(𝑎)) to 𝐵(𝑓(𝑏), 𝑔(𝑏)). According
to the theorem, there is a point (𝑓(𝑐), 𝑔(𝑐))
on the curve 𝛾 where the tangent is parallel
to the chord joining the ends A and B of the
curve.

Class Work [Session 3]:

1. Verify Lagrange’s mean value theorem for the function 𝑓(𝑥) = 2𝑥 2 − 7𝑥 − 10 in [2,5].
2. Verify Cauchy’s mean value theorem for the function 𝑓(𝑥) = 𝑠𝑖𝑛 𝑥 ; 𝑔(𝑥) = 𝑐𝑜𝑠 𝑥 in
𝜋
[0, 2 ].
3. The reliability index of a communication network is represented by
𝑅(𝑥) = sin−1 𝑥 , 0 ≤ 𝑥 ≤ 1
where 𝑥 denotes the probability of successful packet transmission.
Using Lagrange’s Mean Value Theorem, determine whether there exists a transmission
success probability at which the instantaneous rate of change of the reliability index equals
its average rate of change over the interval.
4. In a secure communication network, the data transmission rate and encryption overhead
are represented by
𝑅(𝑥) = 𝑒 𝑥 , 𝐻(𝑥) = 𝑒 −𝑥 , 𝑎 ≤ 𝑥 ≤ 𝑏
where 𝑥 denotes the network operating parameter.
Using Cauchy’s Mean Value Theorem, determine whether there exists an operating
condition within the interval [𝑎, 𝑏] at which the ratio of the instantaneous rates of change
of transmission rate and security overhead equals the ratio of their overall changes across
the interval.

Self-Practice Problems:
1. Examine whether Lagrange’s mean value theorem is applicable for the functions:
a) 𝑓(𝑥) = 𝑥 2/3 in [−8,8].
1
b) 𝑓(𝑥) = 𝑥(𝑥 − 1)(𝑥 − 2) in [0, 2] .
c) 𝑓(𝑥) = 𝑙𝑛 𝑥 in [1, 𝑒].
2. The normalized signal propagation delay is modeled by
1 3
𝐷(𝑥) = cot(𝜋𝑥) , ≤𝑥≤
4 4
where 𝑥 represents a transmission parameter.
 Alliance School of Applied Mathematics
Semester I
Calculus and Linear Algebra

Using Lagrange’s Mean Value Theorem, determine whether there exists a transmission
parameter for which the instantaneous rate of change of delay equals the average rate of
change over the specified interval.
3. The performance index of a packet routing algorithm is given by
1
𝑃(𝑥) = 𝑥(𝑥 − 1)(𝑥 − 2), 0≤𝑥≤
2
where 𝑥 denotes the normalized routing load.
Using Lagrange’s Mean Value Theorem, determine a routing load at which the
instantaneous change in performance equals the average change in performance over the
interval.
4. Verify Cauchy’s mean value theorem for:
1
a) 𝑓(𝑥) = 𝑙𝑜𝑔 𝑥 ; 𝑔(𝑥) = 𝑥 in [1, 𝑒].
b) 𝑓(𝑥) = 𝑒 𝑥 ; 𝑔(𝑥) = 𝑥 + 1 in [0,1].
5. In a distributed computing environment, the computational workload and data storage
requirement are represented by
𝑊(𝑥) = 𝑥 2 , 𝑆(𝑥) = 𝑥 3 , 1≤𝑥≤2
where 𝑥 denotes the normalized processing load.
Using Cauchy’s Mean Value Theorem, determine whether there exists a processing load
level at which the ratio of the instantaneous growth rates of workload and storage
requirement equals the ratio of their average changes over the interval.
𝑠𝑖𝑛 𝑏−𝑠𝑖𝑛 𝑎
6. Prove that 𝑐𝑜𝑡 𝑐 = 𝑐𝑜𝑠 𝑎−𝑐𝑜𝑠 𝑏 where 𝑐 ∈ (𝑎, 𝑏) using Cauchy’s mean value theorem.

Extreme values of functions – Maxima and Minima

Definition: A function 𝑓(𝑥) is said to have a maximum at x = a, if there exists a small number
′ℎ′, however small, such that f(a) > f(a − h) and f(a) > f(a + h), both.

A function 𝑓(𝑥) is said to have a minimum at x = a, if there exists a small number ′ℎ′,
however small, such that f(a) < f(a − h) and f(a) < f(a + h), both.

Note: The maximum and minimum values of a function taken together are called its extreme
values and the points at which the function attains the extreme values are called the turning
points of the function.
 Alliance School of Applied Mathematics
Semester I
Calculus and Linear Algebra

Necessary Condition for an Extremum

Definition: The points at which the derivative of the function 𝒇(𝒙) is equal to zero are called
the stationary points.
The points at which the derivative of the function 𝒇(𝒙) is equal to zero or does not
exist are called the critical points of the function. Consequently, the stationary points are a
subset of the set of critical points.

A necessary condition for an extremum is formulated as follows:

If the point 𝒙 = 𝒂 is an extremum point of the function 𝒚 = 𝒇(𝒙); then
𝒅𝒚
𝒇′ (𝒂) = | =𝟎
𝒅𝒙 𝒙=𝒂

Note: The necessary condition does not guarantee the existence of an extremum. A classic
illustration here is the cubic function 𝑓(𝑥) = 𝑥 3 . Despite the fact that the derivative of the
function at the point 𝑥 = 0 is zero i.e., 𝑓 ′ (0) = 0, this point is not an extremum at 𝑥 = 𝑎.

Conditions for Maxima and Minima:

1) 𝑓(𝑥) is maximum at x = a if 𝑓 ′ (𝑎) = 0 and 𝑓 ′′ (𝑎) is ‘- ve’ [i.e., 𝑓 ′ (𝑎) changes sign from
‘+ ve’ to ‘– ve’].
2) 𝑓(𝑥) is minimum at x = a if 𝑓 ′ (𝑎) = 0 and 𝑓 ′′ (𝑎) is ‘+ ve’ [i.e., 𝑓 ′ (𝑎) changes sign from
‘- ve’ to ‘+ ve’].
 Alliance School of Applied Mathematics
Semester I
Calculus and Linear Algebra

Procedure for finding maxima and minima:

I. Consider the given function as 𝑓(𝑥).

II. Find 𝑓 ′ (𝑥) and equate it to zero. Solve 𝑓 ′ (𝑥) = 0 and let its roots be a,b,c,….

III. Find 𝑓 ′′ (𝑥) and then calculate the value of 𝑓 ′′ (𝑥) at x = a, b, c, …

➢ If 𝑓 ′′ (𝑎) < 0, 𝑓(𝑥) is maximum at x = a.
➢ If 𝑓 ′′ (𝑎) > 0, 𝑓(𝑥) is minimum at x = a.

IV. Sometimes 𝑓 ′′ (𝑥) maybe difficult to find out or 𝑓 ′′ (𝑥) = 0 at x = a. In such cases,
a) If 𝑓 ′ (𝑥) changes its sign from ‘+ ve’ to ‘- ve’ as x passes through a, then 𝑓(𝑥) is
maximum at x = a.
b) If 𝑓 ′ (𝑥) changes its sign from ‘- ve’ to ‘+ ve’ as x passes through a, then 𝑓(𝑥) is
minimum at x = a.
c) If 𝑓 ′ (𝑥) does not change its sign while passing through x = a, then 𝑓(𝑥) is neither
maximum nor minimum at x = a.

Class Work [Session 4]:

1. Find the maximum and minimum values of 3𝑥 4 − 2𝑥 3 − 6𝑥 2 + 6𝑥 + 1 in the interval
(0,2).
2. Find the maxima and minima for the following functions:
a) x 3 − 3𝑥 + 2;
b) 2x 3 − 3𝑥 2 + 6;
3. The performance gain of a search algorithm is modeled by
𝑃(𝑥) = (𝑥 − 1)(𝑥 − 2)2 ,
where 𝑥 denotes the optimization parameter.
Determine the local extrema of the performance function and identify the parameter value
corresponding to the maximum performance using derivative tests.

Class Work [Session 5]:

1. Show that sin 𝑥 (1 + cos 𝑥) is a maximum when 𝑥 = π/3.

2. The cost function of an optimization algorithm is represented by
𝑎
𝑓(𝑥) = + 𝑏𝑥.
𝑥
It is known that 𝑓(2) = 1 and the function has an extremum at 𝑥 = 2.
Determine the values of 𝑎 and 𝑏, and use the second derivative test to classify the
extremum as a maximum or minimum.
3. The accuracy score of a machine learning model over time is represented by
𝐴(𝑡) = 3 + 14𝑡 − 𝑡 2 ,
where 𝑡 denotes the training iteration count.
Find the maximum accuracy score attained by the model and determine the iteration at
which it occurs.
4. The utility function of a resource allocation algorithm is given by
𝑈(𝑥) = 𝑥 3 − 12𝑥 2 + 𝑘𝑥 − 8.
It is known that the utility attains its maximum value at 𝑥 = 2 over the interval [0,3].
Determine the value of 𝑘 and verify the nature of the extremum using derivative tests.
 Alliance School of Applied Mathematics
Semester I
Calculus and Linear Algebra

Self-Practice Problems:
1. Find the maxima and minima for the following functions:
a) −3x 2 + 4𝑥 + 7;
b) 5x 3 + 2𝑥 2 − 3𝑥.
log 𝑥
2. What is the maximum value of the function 𝑥e ?
3. Find the minimum value for the function 𝑦 = 𝑥𝑒 𝑥 .
4. The performance score of a signal-processing algorithm is modeled by
sin 3𝑥
𝑓(𝑥) = 𝑝sin 𝑥 + .
3
Determine the value of 𝑝for which the performance function has an extremum at 𝑥 = 𝜋/3.
5. The computational cost of a data-processing algorithm depends on two design parameters
whose product is fixed. Show that among all rectangular data structures having a fixed
storage capacity (area), the square configuration minimizes the communication overhead
(perimeter).
6. A cylindrical data-storage unit is designed with a fixed storage volume.
Show that the storage unit with minimum material usage (surface area) has its height equal
to its diameter. Use derivative-based optimization methods.

Linear approximation
Consider a function 𝒚 = 𝒇(𝒙) is differentiable at a point 𝒙 = 𝒂. Then, the tangent line
to the graph of 𝒇(𝒙) at 𝒙 = 𝒂 is given by the equation:

𝒚 = 𝒇(𝒂) + 𝒇′ (𝒂)(𝒙 − 𝒂)

In general, for a differentiable function 𝒚 = 𝒇(𝒙), the equation of tangent line to 𝒇(𝒙) at 𝒙 =
𝒂 can be used to approximate 𝒇(𝒙) for 𝒙 near 𝒂. Therefore, we can write:

𝒇(𝒙) ≈ 𝒇(𝒂) + 𝒇′ (𝒂)(𝒙 − 𝒂) for 𝒙 near 𝒂.

We call the linear function,

𝑳(𝒙) = 𝒇(𝒂) + 𝒇′ (𝒂)(𝒙 − 𝒂)

The linear approximation, or tangent line approximation of the function 𝒇(𝒙) at 𝒙 = 𝒂. This
function L is known as the linearization of 𝒇(𝒙) at 𝒙 = 𝒂.
1
Example: Find the linear approximation of 𝑓(𝑥) = at 𝑥 = 2 and use the approximation to
𝑥
1
estimate 2.1.
Solution: Since, we are looking for the linear approximation at 𝑥 = 2, then using the
linear approximation formula, we have:
𝐿(𝑥) = 𝑓(2) + 𝑓 ′ (2)(𝑥 − 2)
We need to find 𝑓(2) .and 𝑓 ′ (2)
1 1 1 1
𝑓(𝑥) = ⇒ 𝑓(2) = ; 𝑓 , (𝑥) = − 2 ⇒ 𝑓 , (2) = −
𝑥 2 𝑥 4
 Alliance School of Applied Mathematics
Semester I
Calculus and Linear Algebra

Therefore, the linear approximation is given by:

𝟏 𝟏
𝑳(𝒙) = − (𝒙 − 𝟐)
𝟐 𝟒
𝟏
Using the linear approximation, we can estimate 𝟐.𝟏 by writing:
𝟏 𝟏 𝟏
= 𝒇(𝟐. 𝟏) ≈ 𝟐 − 𝟒 (𝟐. 𝟏 − 𝟐) ≈ 𝟎. 𝟒𝟕𝟓.
𝟐.𝟏

The actual value of 𝒇(𝟐. 𝟏) is given by:

𝟏
𝒇(𝟐. 𝟏) = 𝟐.𝟏 = 𝟎. 𝟒𝟕𝟔𝟏𝟗.
Therefore, the tangent line gives us a fairly good approximation of 𝒇(𝟐. 𝟏).

Note: However, for the value of 𝒙 far from 2, the equation of tangent line does not give us
good approximation. For example, if 𝒙 = 𝟏𝟎, the y-value of the corresponding point on the
tangent line is:
𝟏 𝟏 𝟏
𝐲 = − (𝟏𝟎 − 𝟐) = − 𝟐 = −𝟏. 𝟓
𝟐 𝟒 𝟐
𝟏
where as, the value of the function at 𝒙 = 𝟏𝟎 is 𝒇(𝟏𝟎) = 𝟏𝟎 = 𝟎. 𝟏.

Class Work [Session 6]:

1. The memory utilization index of a software application is modeled by

𝑈(𝑥) = √𝑥,
where 𝑥 represents the allocated memory units.
a) Find the linear approximation of the memory utilization function at x = 9.
b) Use the linear approximation to estimate the memory utilization index when the
allocated memory becomes 9.1 units.
 Alliance School of Applied Mathematics
Semester I
Calculus and Linear Algebra

2. The utility score of a software application is modeled by

𝑈(𝑥) = (1 + 𝑥)𝑛 ,
where 𝑥 represents the fractional increase in allocated resources.
a) Find the linear approximation of the utility function at 𝑥 = 0.
b) Using the linear approximation with 𝑛 = 3, estimate the utility score after a 1%
increase in resources, i.e., estimate (1.01)3 .

Self-Practice Problems:

1. The resource efficiency of a cloud-based software service is represented by

3
𝐸(𝑥) = √𝑥,
where 𝑥 denotes the normalized amount of computational resources allocated.
a) Find the linear approximation of the resource efficiency function at 𝑥 = 8.
b) Use the linear approximation to estimate the resource efficiency when the allocated
resources become 8.1 units. Give the estimate correct to five decimal places.

2. The utility score of a machine learning model is represented by

𝑀(𝑥) = 𝑒 𝑥 ,
where 𝑥 is the normalized resource allocation factor.
a) Find the linear approximation of 𝑀(𝑥) at 𝑥 = 0.
b) Use the approximation to estimate 𝑒 0.05 .

3. The reliability score of a software application is represented by

𝑅(𝑥) = sin 𝑥,
where 𝑥 is the normalized reliability parameter.
𝜋
a) Find the linear approximation of 𝑅(𝑥) at 𝑥 = 6 .
b) Use the approximation to estimate sin(0.55).

INDETERMINATE FORMS
L’Hospital Rule: If f(x) and g(x) are two functions of x which can be expanded by Taylor’s
series in the neighborhood of x=a and f(a)=g(a)=0, then
𝒇(𝒙) 𝒇′ (𝒙)
𝐥𝐢𝐦 = 𝐥𝐢𝐦 ′
𝒙→𝒂 𝒈(𝒙) 𝒙→𝒂 𝒈 (𝒙)

Indeterminate Forms: There are seven types of indeterminate forms given as follows:
0
• 0
∞
• ∞
• 0.∞
• ∞− ∞
• 1∞
• 00
• ∞0
where the limits for these can be calculated using L’Hospital Rule.
 Alliance School of Applied Mathematics
Semester I
Calculus and Linear Algebra

0 ∞
NOTE: To apply the L’Hospital rule, we need the given function in 0 or ∞ form.
Four standard limits:
𝑆𝑖𝑛 𝑥
(i) lim ( 𝑥 ) = 1
𝑥→0
𝑥
(ii) lim (𝑆𝑖𝑛 𝑥) = 1
𝑥→0
𝑡𝑎𝑛 𝑥
(iii) lim ( )=1
𝑥→0 𝑥
𝑥
(iv) lim (𝑡𝑎𝑛 𝑥) = 1
𝑥→0

Self-Practice Problems:
𝑥− 𝑥 𝑥
1. Find lim (𝑥−1−𝑙𝑜𝑔 𝑥) Ans 2
𝑥→1
𝑎 𝑥
2. Find lim (𝑥 − Cot 𝑎) Ans 0
𝑥→0
𝜋𝑥
𝑥 tan
2𝑎
3. Find lim (2 − 𝑎)
𝑥→𝑎
𝑒 2𝑥 − (1+𝑥)2
4. Find lim ( 𝑥 log(1+𝑥) ) Ans 1
𝑥→0
𝑆𝑖𝑛−1 𝑥−𝑥
5. Find lim ( ) Ans 1/6
𝑥→0 𝑥3
6. Find lim log tan 𝑥 tan 2𝑥
𝑥→0
1
(1+𝑥)𝑥 −𝑒 𝑒
7. Prove that lim ( )=-2
𝑥→0 𝑥

8. Show that lim log 𝑥 Sin 𝑥=1

𝑥→0
9. Evaluate lim tan 𝑥. log 𝑥 Ans 0
𝑥→0
1 2
10. Evaluate lim - 𝐶𝑜𝑡 2 x Ans
𝑥→0 𝑥 2 3

Note: For solutions, refer to the class notes.

Reference Books:

1. B.S. Grewal, Higher Engineering Mathematics, Latest edition, Khanna Publishers.

2. Peter V. O’Neil, Engineering Mathematics CENGAGE Learning India Pvt Ltd.

Publishers.

3. B.V. Ramana, Higher Engineering Mathematics, Latest Edition, Tata Mc. Graw Hill
Publications

4. Erwin Kreyszig, Advanced Engineering Mathematics, Latest edition, Wiley

Publications.