MTH647 Handout

Lecture NO. 01

Mathematical Modelling: It is the approximation of objects with geometrical objects.

Mathematical Formulation: In this step we drive the equation corresponds to the given phenomenon. It is
based on existing law i.e. Newton’s law of motion, laws of thermodynamics. If laws are not available for
given existing problem, then we go for experiments.
Example:

Lecture NO. 02

Example: Heat flow problem Formulation PDEs Solutions Expansion of function terms of ‘sine’
and ‘cosine’ (Fourier analysis).
dx
Physical Interpretation: E.g. x x  t   Aet A R
dt

Lecture NO. 03

Vibrating String Equation:

2 y 2  y
2
 a
t 2 x 2
Here we’ll try to derive the above given equation. So,
Let represents the elements of arc of string and is tension at and (constant).

For small angles;

So …… (1)
By newton second law of motion

( ) …… (2) Is linear mass density

Note: Here acceleration is in derivative form and while discussing the acceleration for a segment (large
number of particles), there will be some approximation/error. That’s why we are using as
approximation symbol.
By comparing equation 1 and 2, we have

( )

( )

As , so above equation become

[ ] ( )

Note: Here are respectively.

For small vibrations;
Page | 1
MTH647 Handout

By using this relation, we get

[ ] ( )

Dividing by on both sides

[ ] ( )

[ ] ( )

For and , it becomes

[ ( )]

( )

Here , so equation becomes

( )

Or it can be written as

( )

This relation is known as Vibrating string equation.

Lecture NO. 04

Boundary Conditions for Vibrating String Equation:

Statement: Write boundary conditions for a vibrating string of length for which i) Ends and
are fixed. ii) Initial shape of string is . iii) Initial velocity distribution is . iv) Displacement at any
instant is bounded.
Solution:
i) String is fixed at and , then its displacement at

Note: above relation is in form. At starting point and

at final point .
ii) Initial shape at is given as

Note: Just put in relation and here is just an expression.

iii) Initial velocity at time is given as

At
Note: Here velocity is in derivative form and is just an expression.
iv) Displacement is bounded: It implies that their exist some finite real number such that

These are some examples of boundary conditions applied on vibrating string equation.

Dedicated to unknown students who are our future heroes.

Page | 2
MTH647 Handout

Lecture NO. 05
Heat Flux across the Plane:
Suppose planes′𝐼′ and ′𝐼𝐼′are parallel and at distance ′∆𝑛′apart. Let
𝑢 = 𝑡𝑒𝑚𝑝 𝑜𝑓 𝑝𝑙𝑎𝑛𝑒 𝐼 And 𝑢 + ∆𝑢 = 𝑡𝑒𝑚𝑝 𝑜𝑓 𝑝𝑙𝑎𝑛𝑒 𝐼𝐼
Heat flows from higher temp plane to lower temp plane.
Heat flux is defined as; Heat per unit area per unit time. Mathematically
𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑐𝑒 𝑜𝑓 𝑡𝑒𝑚𝑝𝑒𝑟𝑎𝑡𝑢𝑟𝑒
𝐻𝑒𝑎𝑡 𝑓𝑙𝑢𝑥 ∝
𝑑𝑖𝑠𝑡𝑎𝑛𝑐𝑒 𝑏𝑒𝑡𝑤𝑒𝑒𝑛 𝑝𝑙𝑎𝑛𝑒𝑠
𝑢 − 𝑢 + ∆𝑢 ∆𝑢
𝐻𝑒𝑎𝑡 𝑓𝑙𝑢𝑥 𝑓𝑜𝑟𝑚 𝐼 𝑡𝑜 𝐼𝐼 = 𝑘 => −𝑘
∆𝑛 ∆𝑛
Where ′𝑘′ is constant of proportionality and is called thermal conductivity.
Thermal conductivity is defined as; Measure of how well a material can conduct or transfer heat.
Note: As in this case heat flows from plane ′𝐼 𝑡𝑜 𝐼𝐼′ (lower temp to higher temp), that’s why we have the
negative value of heat flux. ∴ ∆𝑢 > 0 (Heat value will be positive) if heat flows from plane ′𝐼𝐼 𝑡𝑜 𝐼′
(higher temp to low temp) and vice versa.
For limiting case: If ∆𝑛 → 0 𝑎𝑛𝑑 ∆𝑢 → 0 then
𝜕𝑢
𝐻𝑒𝑎𝑡 𝑓𝑙𝑜𝑤 𝑎𝑐𝑟𝑜𝑠𝑠 𝑝𝑙𝑎𝑛𝑒 𝐼 = −𝑘
𝜕𝑛
In vector form
∇𝑢
𝐻𝑒𝑎𝑡 𝑓𝑙𝑜𝑤 𝑎𝑐𝑟𝑜𝑠𝑠 𝑝𝑙𝑎𝑛𝑒 𝐼 = −𝑘
∇n
Note: As ∆𝑛 → 0 (distance between planes approaches to zero), so we can write it as 𝑝𝑙𝑎𝑛𝑒 ′𝐼′ only.

Lecture NO. 06

Partial Differential Equations (Definition and Related Terms):

Definition: It is an equation containing unknown functions of two or more variables and partial derivatives
w.r.to these variables. Example
𝜕2 𝑢
= 2𝑥 − 𝑦 𝑜𝑟𝑑𝑒𝑟 = 2
𝜕𝑥𝑑𝑦
Order: It is the order of highest derivative involved in PDE.
Solution: It is a function which satisfy the given DE. Example:
1
𝑢 = 𝑥 2 𝑦 − 𝑥𝑦 2
2
𝜕𝑢
= 𝑥 2 − 𝑥𝑦 1𝑠𝑡 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑤. 𝑟. 𝑡𝑜 ′𝑦′
𝜕𝑦
𝜕2 𝑢 𝜕 𝜕𝑢
= = 2𝑥 − 𝑦 2𝑛𝑑 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒 𝑤. 𝑟. 𝑡𝑜 ′𝑥′
𝜕𝑥𝑑𝑦 𝜕𝑥 𝜕𝑦
1
Similarly; 𝑢 = 𝑥 2 𝑦 − 𝑥𝑦 2 + 𝐹 𝑥 + 𝐺(𝑦) is also a solution. Here ′𝐹 𝑥 𝑎𝑛𝑑 𝐺(𝑥)′ are arbitrary functions.
2
Particular Solution: It is obtained from the general solution by particular choice of arbitrary functions.
Example:
1
𝑢 = 𝑥 2 𝑦 − 𝑥𝑦 2 + 2𝑠𝑖𝑛𝑥 + 3𝑦 4 − 5
2
Singular Solution: It cannot be obtained from the general solution by choosing arbitrary functions.

Lecture NO. 07
1
Example: 𝑥𝑦 ′ = 𝑦 => 𝑦 = 𝑙𝑛𝑐𝑥 2 (General Solution−Here ′𝑐′ is arbitrary constant).
4
But ′𝑦 = 0′is also a solution of given DE.As it is not obtained from general solution, that’s why it is called
singular solution. ∄ 𝑐 ∈ 𝑅 𝑠𝑢𝑐ℎ 𝑡ℎ𝑎𝑡 𝑦 = 0
Page | 3
MTH647 Handout

Lecture NO. 08

Linear PDEs and Their Classification:

General form of linear partial differential equation of order 2in two independent variables is of the form
𝜕2𝑢 𝜕2 𝑢 𝜕2 𝑢 𝜕𝑢
𝐴 2 +𝐵 +𝐶 2+𝐷 + 𝐹𝑢 = 𝐺
𝜕𝑥 𝜕𝑥𝑑𝑦 𝜕𝑦 𝜕𝑥
Here 𝐴, 𝐵, 𝐶, 𝐷, 𝐸, 𝐹 𝑎𝑛𝑑 𝐺 may depend on ′𝑥′ or ′𝑦′ only but not on both ′𝑥′and ′𝑦′.
If at least one of these(𝐴, 𝐵, 𝐶, 𝐷, 𝐸, 𝐹 𝑎𝑛𝑑 𝐺) is function of ′𝑥′ and ′𝑦′, then it is non-linear.
If 𝐺 = 0, then it is homogeneous.
Example:
𝜕3 𝑢 𝜕2 𝑢
𝑥2 3 = 𝑦3 2 (𝑜𝑟𝑑𝑒𝑟 = 3, 𝑙𝑖𝑛𝑒𝑎𝑟)
𝜕𝑦 𝜕𝑥
Example:
2 2
𝜕𝑢 𝜕𝑢
+ =1 (𝑜𝑟𝑑𝑒𝑟 = 1, 𝑛𝑜𝑛 − 𝑙𝑖𝑛𝑒𝑎𝑟)
𝜕𝑥 𝜕𝑦

Lecture NO. 09

Topic Continue…
𝜕2𝑢 𝜕2 𝑢 𝜕2 𝑢 𝜕𝑢
𝐴 2
+ 𝐵 + 𝐶 2
+𝐷 + 𝐹𝑢 = 𝐺
𝜕𝑥 𝜕𝑥𝑑𝑦 𝜕𝑦 𝜕𝑥
It is defined as
i) Elliptical if 𝐵2 − 4𝐴𝐶 < 0
ii) Hyperbolic if 𝐵2 − 4𝐴𝐶 > 0
iii) Parabolic if 𝐵2 − 4𝐴𝐶 = 0
Example:
𝜕2 𝑢 𝜕2 𝑢 𝜕𝑢
𝑥 2 + 𝑦 2 + 3𝑦 2 =0
𝜕𝑥 𝜕𝑦 𝜕𝑥
Here 𝐴 = 𝑥, 𝑏 = 0 𝑎𝑛𝑑 𝑐 = 𝑦. Now
𝑥𝑦 > 0, 𝑡ℎ𝑒𝑛 𝑖𝑡 𝑖𝑠 𝑒𝑙𝑙𝑖𝑝𝑡𝑖𝑐
2
𝐵 − 4𝐴𝐶 = −4𝑥𝑦 𝑥𝑦 < 0, 𝑡ℎ𝑒𝑛 𝑖𝑡 𝑖𝑠 ℎ𝑦𝑝𝑒𝑟𝑏𝑜𝑙𝑖𝑐
𝑥𝑦 = 0, 𝑡ℎ𝑒𝑛 𝑖𝑡 𝑖𝑠 𝑝𝑎𝑟𝑎𝑏𝑜𝑙𝑖𝑐

Lecture NO. 10

Show that 𝑢 𝑥, 𝑡 = 𝑒 −8𝑡 𝑠𝑖𝑛2𝑥, is a solution to BVP (Boundary Value Problem):

2
𝜕𝑢 𝜕 𝑢
𝐼) = 2 2 , 𝐼𝐼) 𝑢 0, 𝑡 = 𝑢 𝜋, 𝑡 = 0, 𝐼𝐼𝐼) 𝑢 𝑥, 0 = 𝑠𝑖𝑛2𝑥
𝜕𝑡 𝜕𝑥
Solution:
𝐼𝐼) 𝑢 𝑥, 𝑡 = 𝑒 −8𝑡 𝑠𝑖𝑛2𝑥 => 𝑢 0, 𝑡 = 𝑒 −8𝑡 sin 0 = 0
𝐼𝐼) 𝑢 𝑥, 𝑡 = 𝑒 −8𝑡 𝑠𝑖𝑛2𝑥 => 𝑢 𝜋, 𝑡 = 𝑒 −8𝑡 sin 2𝜋 = 0
𝐼𝐼𝐼) 𝑢 𝑥, 𝑡 = 𝑒 −8𝑡 𝑠𝑖𝑛2𝑥 => 𝑢 𝑥, 0 = 𝑒 −8(0) sin 2𝑥 = 𝑠𝑖𝑛2𝑥
Now, try to calculate the values for condition ′𝐼′
𝜕𝑢 𝜕2 𝑢
=2 2 … … (1)
𝜕𝑡 𝜕𝑥
Taking derivative of equation 1 w.r.to ′𝑡′
𝜕𝑢
= −8𝑒 −8𝑡 𝑠𝑖𝑛2𝑥
𝜕𝑡
Page | 4
MTH647 Handout

Taking derivative of equation 1 w.r.to ′𝑥′

𝜕𝑢
= 2𝑒 −8𝑡 𝑐𝑜𝑠2𝑥
𝜕𝑥
Taking 2nd derivative, we have
𝜕2 𝑢
2
= −4𝑒 −8𝑡 𝑠𝑖𝑛2𝑥
𝜕𝑥
Now, by putting the calculated values in equation 1 and checking, we get
𝜕𝑢 𝜕2 𝑢
=2 2
𝜕𝑡 𝜕𝑥
𝜕𝑢 −8𝑡 −8𝑡
𝜕2 𝑢
= −8𝑒 𝑠𝑖𝑛2𝑥 => 2 −4𝑒 𝑠𝑖𝑛2𝑥 = 2 2
𝜕𝑡 𝜕𝑥
Or simply
𝜕𝑢 𝜕2 𝑢
=2 2
𝜕𝑡 𝜕𝑥
Hence proved(′𝐼′) that the given function "𝑢 𝑥, 𝑡 = 𝑒 −8𝑡 𝑠𝑖𝑛2𝑥" is solution to given BVP.

Lecture NO. 11

Show that 𝑣 = 𝐹(𝑦 − 3𝑥), where ′𝐹′ is arbitrary differentiable function.

𝜕𝑣 𝜕𝑣
𝐼) The general solution of equation is +3 = 0.
𝜕𝑥 𝜕𝑦
𝐼𝐼) Also find its particular solution if 𝑣 0, 𝑦 = 4𝑠𝑖𝑛𝑦
Solution:
𝐼) Given that 𝑣 = 𝐹 𝑦 − 3𝑥 …… (1)
Let / say 𝑣 𝑥, 𝑦 = 𝐹 𝑢 𝑊ℎ𝑒𝑟𝑒 𝑢 = 𝑦 − 3𝑥
𝑣 = 𝐹 𝑢 𝑎𝑛𝑑 𝑢 = 𝑢(𝑥, 𝑦)
By applying chain rule w.r.to ′𝑥′
𝜕𝑣 𝜕𝑣 𝜕𝑢
= . = 𝐹 ′ 𝑢 −3 => −3𝐹 ′ 𝑢
𝜕𝑥 𝜕𝑢 𝜕𝑥
By applying chain rule w.r.to ′𝑦′
𝜕𝑣 𝜕𝑣 𝜕𝑢
= . = 𝐹 ′ 𝑢 1 => 𝐹 ′ 𝑢
𝜕𝑦 𝜕𝑢 𝜕𝑦
Now by putting value in equation
𝜕𝑣 𝜕𝑣
+3 = −3𝐹 ′ 𝑢 + 3𝐹 ′ 𝑢
𝜕𝑥 𝜕𝑦
By solving, we have
𝜕𝑣 𝜕𝑣
+3 =0
𝜕𝑥 𝜕𝑦
Hence, the general solution of equation proved.
𝐼𝐼) Now, also given that
𝑣 𝑥, 𝑦 = 𝐹(𝑦 − 3𝑥)
Given condition
𝑣 0, 𝑦 = 𝐹 𝑦 = 4𝑠𝑖𝑛𝑦
It implies
𝑣 𝑥, 𝑦 = 𝐹 𝑦 − 3𝑥 = 4sin⁡ (𝑦 − 3𝑥)
This is our required particular solution.
_____________

Page | 5
MTH647 Handout

Lecture NO. 12
Solving PDEs by the Method of ODEs:
𝜕2𝑢 𝜕𝑢
Question: 𝑡 +2 = 2𝑡𝑥
𝜕𝑥𝜕𝑡 𝜕𝑥
Solution: Given that
𝜕2𝑢 𝜕𝑢
𝑡 +2 = 2𝑡𝑥
𝜕𝑥𝜕𝑡 𝜕𝑥
𝜕 𝜕𝑢 𝜕
𝑡 + 2 𝑢 = 2𝑡𝑥
𝜕𝑥 𝜕𝑡 𝜕𝑥
𝜕 𝜕𝑢
𝑡 + 2𝑢 = 2𝑡𝑥
𝜕𝑥 𝜕𝑡
Integrating w.r.to ′𝑥′
𝜕 𝜕𝑢
𝑡 + 2𝑢 𝑑𝑥 = 2𝑡𝑥 𝑑𝑥
𝜕𝑥 𝜕𝑡
𝜕𝑢
𝑡 + 2𝑢 = 𝑥 2 𝑡 + 𝐺(𝑡)
𝜕𝑡
Dividing by ′𝑡′ on both sides
𝜕𝑢 2 𝐺(𝑡)
+ 𝑢 = 𝑥2 +
𝜕𝑡 𝑡 𝑡
2
𝑑𝑡
𝐼𝑛𝑡𝑒𝑔𝑟𝑎𝑡𝑖𝑛𝑔 𝐹𝑎𝑐𝑡𝑜𝑟 = 𝐼. 𝐹 = 𝑒 𝑡
2
𝐼. 𝐹 = 𝑒 2 𝑙𝑛 𝑡 = 𝑒 𝑙𝑛 𝑡 = 𝑡 2
Required solution is
𝐼. 𝐹 × 𝑢 = 𝑅. 𝐻. 𝑆 × 𝐼. 𝐹 𝑑𝑡 + 𝐹(𝑥)
By putting values
𝐺(𝑡)
𝑡2𝑢 = (𝑥 2 + ) × 𝑡 2 𝑑𝑡 + 𝐹(𝑥)
𝑡
𝐺(𝑡) 2
𝑡2𝑢 = 𝑥2𝑡2 + 𝑡 𝑑𝑡 + 𝐹(𝑥)
𝑡
𝑡3
𝑡 2 𝑢 = 𝑥 2 + 𝑡. 𝐺(𝑡)𝑑𝑡 + 𝐹(𝑥)
3
2 2
𝑡3
𝑡 𝑢=𝑥 + 𝐻 𝑡 + 𝐹(𝑥)
3

Lecture NO. 13
General Solution for Solving PDEs:
𝜕2𝑢 𝜕2𝑢 𝜕2𝑢
Question +3 +2 =0 …… (1)
𝜕𝑥 2 𝜕𝑥𝜕𝑡 𝜕𝑦 2
Solution: Let
𝑢 = 𝑒 𝑎𝑥 +𝑏𝑦
Taking 1st and 2nd derivatives w.r.to ′𝑥′
𝜕𝑢 𝜕
= 𝑒 𝑎𝑥 +𝑏𝑦 𝑎𝑥 + 𝑏𝑦 = 𝑎𝑢
𝜕𝑥 𝜕𝑥
𝜕2𝑢 𝜕 𝜕𝑢 𝜕 𝜕𝑢
2
= = 𝑎𝑢 = 𝑎 = 𝑎2 𝑢
𝜕𝑥 𝜕𝑥 𝜕𝑥 𝜕𝑥 𝜕𝑥
Similarly taking derivatives w.r.to ′𝑦′
𝜕2𝑢
= 𝑏2 𝑢
𝜕𝑦 2
And
𝜕2𝑢 𝜕 𝜕𝑢 𝜕 𝜕𝑢
= = 𝑏𝑢 = 𝑏 = 𝑎𝑏𝑢
𝜕𝑥𝜕𝑦 𝜕𝑥 𝜕𝑦 𝜕𝑥 𝜕𝑥
Putting values in equation 1, we have
Page | 6
MTH647 Handout

𝑎2 𝑢 + 3𝑎𝑏𝑢 + 2𝑏2 𝑢 = 0
𝑎2 + 3𝑎𝑏 + 2𝑏2 𝑢 = 0
𝑎 + 2𝑏 𝑎 + 𝑏 = 0
By solving, we get
𝑎 = −2𝑏 ; 𝑎 = −𝑏
𝐹𝑜𝑟 𝑎 = −2𝑏
𝑢1 = 𝑒 −2𝑏𝑥 +𝑏𝑦 = 𝑒 𝑏(𝑦 −2𝑥)
𝐹𝑜𝑟 𝑎 = −𝑏
𝑢2 = 𝑒 −𝑏𝑥 +𝑏𝑦 = 𝑒 𝑏 (𝑦−𝑥)
Given PDE is homogeneous, therefore by superposition principle
𝑢 = 𝛼𝑢1 + 𝛽𝑢2
By putting values
𝑢 = 𝛼𝑒 𝑏 (𝑦−2𝑥) + 𝛽𝑒 𝑏 (𝑦−𝑥)
Let
𝑒 𝑏 (𝑦−2𝑥) = 𝐺 𝑦 − 2𝑥 𝑎𝑛𝑑 𝑒 𝑏(𝑦−𝑥) = 𝐻(𝑦 − 𝑥)
So, above equation becomes
𝑢 = 𝐺 𝑦 − 2𝑥 + 𝐻(𝑦 − 𝑥)
Required general solution of given PDE.

Lecture NO. 14
Solving PDEs by Separation of Variables:
Question:
𝜕𝑢 𝜕𝑢
=4 𝑠𝑢𝑏𝑗𝑒𝑐𝑡 𝑡𝑜 𝑏𝑜𝑢𝑛𝑑𝑟𝑦 𝑐𝑜𝑛𝑑𝑖𝑡𝑖𝑜 𝑢 0, 𝑦 = 8𝑒 −3𝑦
𝜕𝑥 𝜕𝑦
Solution: Given equation
𝜕𝑢 𝜕𝑢
=4 … (1)
𝜕𝑥 𝜕𝑦
Let 𝑢 𝑥, 𝑦 = 𝑋 𝑥 𝑌 𝑦 = 𝑋𝑌 Be its solution… (i)
Taking derivative w.r.to 𝑥 & 𝑦
𝜕𝑢 𝜕𝑢
= 𝑋 ′ 𝑌 𝑎𝑛𝑑 = 𝑋𝑌′
𝜕𝑥 𝜕𝑦
Putting values in equation 1, we have
𝑋 ′ 𝑌 = 4 𝑋𝑌′
𝑋′ 𝑌′
= … (2)
4𝑋 𝑌
As
𝑋 𝑥 = 𝑋 ,𝑌 𝑦 = 𝑌
Here 𝑋 𝑎𝑛𝑑 𝑌 are independent variables. Each side of equation 2 must be a constant say ′𝑐′. So,
𝑋′ 𝑌′
=𝑐 =𝑐
4𝑋 𝑌
′
𝑋 = 4𝑐𝑋 𝑌 ′ = 𝑐𝑌
𝑑𝑋 𝑑𝑌
= 4𝑐𝑋 = 𝑐𝑦
𝑑𝑥 𝑑𝑦
𝑑𝑋 𝑑𝑌
= 4𝑐𝑑𝑥 = 𝑐𝑑𝑦
𝑋 𝑌
By integrating
𝑑𝑋 𝑑𝑌
= 4𝑐 𝑑𝑥 = 𝑐 𝑑𝑦
𝑋 𝑌
𝑙𝑛𝑋 = 4𝑐𝑥 + 𝐴 𝑙𝑛𝑌 = 𝑐𝑦 + 𝐵
𝑋 = 𝑒 4𝑐𝑥 +𝐴 𝑌 = 𝑒 𝑐𝑦 +𝐵
𝑋 = 𝑒 𝐴 𝑒 4𝑐𝑥 𝑌 = 𝑒 𝐵 𝑒 𝑐𝑦
𝑋 = 𝑘1 𝑒 4𝑐𝑥 𝑌 = 𝑘2 𝑒 𝑐𝑦
Page | 7
MTH647 Handout

Now by putting values in equation (i)

𝑢 𝑥, 𝑦 = 𝑋𝑌 = 𝑘1 𝑘2 𝑒 4𝑐𝑥 +𝑐𝑦 …… (3)
−3𝑦
By using boundary condition𝑢 0, 𝑦 = 8𝑒 in equation 3, we have
𝑢 0, 𝑦 = 𝑘 𝑒 𝑐𝑦 = 8𝑒 −3𝑦
Here 𝑘 = 8 𝑎𝑛𝑑 𝑐 = −3,
Hence
𝑢 𝑥, 𝑦 = 8𝑒 −3(4𝑥+𝑦) = 8𝑒 −12𝑥−3𝑦

Lecture NO. 15

Heat Conduction Equation and its Physical Interpretation:

Question:
𝜕𝑢 𝜕2 𝑢
= 2 2 … (1) ∴ 0 < 𝑥 < 3 ,𝑡 > 0
𝜕𝑡 𝜕𝑥
Given that (boundary values)
𝑢 0, 𝑡 = 𝑢 3, 𝑡 = 0 … 𝑖 ,
𝑢 𝑥, 0 = 5𝑠𝑖𝑛4𝜋𝑥 − 3𝑠𝑖𝑛8𝜋𝑥 + 2𝑠𝑖𝑛10𝜋𝑥 … (𝑖𝑖) ∴ 𝑢(𝑥, 𝑡) < 𝑀 < ∞
Solution: Let
𝑢 𝑥, 𝑡 = 𝑋 𝑥 . 𝑇 𝑡 = 𝑋𝑇 … (𝑧)
Taking derivative w.r.to ′𝑡′
𝜕𝑢
= 𝑋𝑇′
𝜕𝑡
Taking 1st derivative w.r.to ′𝑥′
𝜕𝑢
= 𝑋′𝑇
𝜕𝑥
Taking 2nd derivative w.r.to ′𝑥′
𝜕2 𝑢
= 𝑋′′𝑇
𝜕𝑥 2
By putting values equation 1 becomes
𝑋𝑇 ′ = 2𝑋′′𝑇
𝑋′′ 𝑇 ′
= = −𝜆2 (𝑠𝑎𝑦)
𝑋 2𝑇
So,
𝑋′′ 2
𝑇′
= −𝜆 = −𝜆2
𝑋 2𝑇
𝑋 ′′ + 𝜆2 𝑋 = 0 … 𝑎 𝑇𝑎𝑘𝑖𝑛𝑔 𝑑𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒
𝑑𝑇
𝐿𝑒𝑡 𝑋 = 𝑒 𝑚𝑥 = −2𝜆2 𝑇
𝑑𝑡
𝑑𝑇
𝑋 ′′ = 𝑚2 𝑒 𝑚𝑥 = 𝑚2 𝑋 = −2𝜆2 𝑑𝑡
𝑇
𝑑𝑇
𝑎 => 𝑚2 𝑋 + 𝜆2 𝑋 = 0 = −2𝜆2 𝑑𝑡
𝑇
𝑚 = ±𝜆𝑖 & 𝑋 ≠ 0 𝑙𝑛 𝑇 = −2𝜆2 𝑡 + 𝑐
2 𝑡+𝑐
∴ 𝑋 𝑥 = 𝛼1 𝑒 𝜆𝑖𝑥 + 𝛼2 𝑒 −𝜆𝑖𝑥 𝑇 = 𝑒 −2𝜆
2𝑡
𝑋 𝑥 = 𝛼1 𝑐𝑜𝑠𝜆𝑥 + 𝑖𝑠𝑖𝑛𝜆𝑥 + 𝛼2 𝑐𝑜𝑠𝜆𝑥 − 𝑖𝑠𝑖𝑛𝜆𝑥 𝑇 = 𝑒 𝑐 . 𝑒 −2𝜆
2𝑡
𝑋 𝑥 = 𝛼1 + 𝛼2 𝑐𝑜𝑠𝜆𝑥 + 𝑖 𝛼1 − 𝛼2 𝑠𝑖𝑛𝜆𝑥 𝑇 = 𝑐1 𝑒 −2𝜆
𝑋 𝑥 = 𝐴1 𝑐𝑜𝑠𝜆𝑥 + 𝐵1 𝑠𝑖𝑛𝜆𝑥 𝐻𝑒𝑟𝑒 𝑒 𝑐 = 𝑐1
Here we were solving both functions side by side. Now
Putting values in equation (𝑧), we get
2𝑡
𝑢 𝑥, 𝑡 = 𝑋𝑇 = 𝑒 −2𝜆 𝐴1 𝑐𝑜𝑠𝜆𝑥 + 𝐵1 𝑠𝑖𝑛𝜆𝑥 … (2)

Page | 8
MTH647 Handout

𝐶𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑖𝑛𝑔 𝑖 (𝑏𝑜𝑢𝑛𝑑𝑎𝑟𝑦 𝑣𝑎𝑙𝑢𝑒) 𝑏𝑦 𝑝𝑢𝑡𝑡𝑖𝑛𝑔 𝑣𝑎𝑙𝑢𝑒𝑠 𝑖𝑛 𝑒𝑞𝑢𝑎𝑖𝑡𝑖𝑜𝑛 2,

2𝑡 2 2𝑡
𝑢 0, 𝑡 = 0 = 𝑒 −2𝜆 𝐴𝑐𝑜𝑠𝜆(0) + 𝐵𝑠𝑖𝑛𝜆(0) = 𝑒 −2𝜆 𝑡 . 𝐴 => 𝐴 = 0 ∴ 𝑒 −2𝜆 ≠0
Equation 2 becomes
2
𝑢 𝑥, 𝑡 = 𝑋𝑇 = 𝐵. 𝑒 −2𝜆 𝑡 𝑠𝑖𝑛𝜆𝑥 ∴𝐴=0
𝐶𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑖𝑛𝑔 𝑠𝑒𝑐𝑜𝑛𝑑 𝑝𝑎𝑟𝑡 𝑜𝑓 𝑖 𝑏𝑦 𝑝𝑢𝑡𝑡𝑖𝑛𝑔 𝑣𝑎𝑙𝑢𝑒𝑠 𝑖𝑛 𝑎𝑏𝑜𝑣𝑒 𝑟𝑒𝑙𝑎𝑡𝑖𝑜𝑛
2
𝑢 3, 𝑡 = 0 = 𝑒 −2𝜆 𝑡 . 𝐵𝑠𝑖𝑛3𝜆
Two Cases
1; 𝑖𝑓 𝐵 = 0 => 𝑢 𝑥, 𝑡 = 0𝑎𝑠 𝐴 = 0 𝑎𝑙𝑟𝑒𝑎𝑑𝑦
𝑚𝜋
2; 𝑖𝑓 𝐵 ≠ 0 𝑏𝑢𝑡 𝑠𝑖𝑛3𝜆 = 0 => 3𝜆 = 𝑚𝜋 => 𝜆 = ∴𝑚∈𝑍
3
Hence
𝑚 2𝜋 2 𝑚𝜋
−2 𝑡
𝑢 𝑥, 𝑡 = 𝐵𝑒 9 sin 𝑥 … (3)
3
𝐶𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑖𝑛𝑔 𝑖𝑖 (𝑏𝑜𝑢𝑛𝑑𝑎𝑟𝑦 𝑣𝑎𝑙𝑢𝑒)
𝑢 𝑥, 0 = 5𝑠𝑖𝑛4𝜋𝑥 − 3𝑠𝑖𝑛8𝜋𝑥 + 2𝑠𝑖𝑛10𝜋𝑥 … (𝑏)
By applying superposition principle on equation 3,
𝑚 2𝜋 2 𝑚1 𝜋 𝑚 2𝜋 2 𝑚2 𝜋 𝑚 2𝜋 2 𝑚3 𝜋
−2 1 𝑡 −2 2 𝑡 −2 3 𝑡
𝑢 𝑥, 𝑡 = 𝐵1 𝑒 9 sin 𝑥 + 𝐵2 𝑒 9 sin 𝑥 + 𝐵3 𝑒 9 sin 𝑥 … (4)
3 3 3
For
𝑚1 𝜋 𝑚2 𝜋 𝑚3 𝜋
𝑢 𝑥, 0 = 𝐵1 sin 𝑥 + 𝐵2 sin 𝑥 + 𝐵3 sin 𝑥 … (𝑐)
3 3 3
By comparing 𝑏 & 𝐶, we have
𝐵1 = 5 , 𝐵2 = −3 , 𝐵3 = 2 , 𝑚1 = 12 , 𝑚2 = 24 & 𝑚3 = 30
Finally by putting calculated values in equation 4, we get
2𝑡 2𝑡 2𝑡
𝑢 𝑥, 𝑡 = 5𝑒 −32𝜋 𝑠𝑖𝑛4𝜋𝑥 − 3𝑒 −128𝜋 𝑠𝑖𝑛8𝜋𝑥 + 2𝑒 −200𝜋 𝑠𝑖𝑛10𝜋𝑥

Lecture NO. 16

Motivation Behind Fourier Series:

𝜕𝑢 𝜕2 𝑢
= 2 2 … (1) ∴ 0 < 𝑥 < 3 ,𝑡 > 0
𝜕𝑡 𝜕𝑥
Given that (boundary values)
𝑢 0, 𝑡 = 𝑢 3, 𝑡 = 0 … 𝑖 ,
𝑢 𝑥, 0 = 5𝑠𝑖𝑛4𝜋𝑥 − 3𝑠𝑖𝑛8𝜋𝑥 + 2𝑠𝑖𝑛10𝜋𝑥 … (𝑖𝑖) ∴ 𝑢(𝑥, 𝑡) < 𝑀 < ∞
Solution: If
𝑢 𝑥, 0 = 𝑓 𝑥 = ? … (1)
As we know from previous lecture
2 2
∞
𝑚 𝜋 𝑚1 𝜋 𝑚 2𝜋 2 𝑚2 𝜋 𝑚 2𝜋 2𝑡 𝑚𝜋𝑥
−2 1 𝑡 −2 2 𝑡 −2
𝑢 𝑥, 𝑡 = 𝐵1 𝑒 9 sin 𝑥 + 𝐵2 𝑒 9 sin 𝑥 +… = 𝐵𝑚 𝑒 9 sin
3 3 3
𝑚 =1
𝐶𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑖𝑛𝑔 (1)
∞ ∞
𝑚 2 𝜋 2 (0) 𝑚𝜋𝑥 𝑚𝜋𝑥
−2
𝑓 𝑥 = 𝑢 𝑥, 0 = 𝐵𝑚 𝑒 9 sin = 𝐵𝑚 𝑒 0 sin
3 3
𝑚 =1 𝑚 =1
As 𝑒 0 = 1
∞
𝑚𝜋𝑥
𝑓 𝑥 = 𝑢 𝑥, 0 = 𝐵𝑚 sin
3
𝑚 =1

____________________

Page | 9
MTH647 Handout

Lecture NO. 17

Periodic & Piecewise Continuous Functions:

Definition: A function 𝑓(𝑥) is said to have a period ′𝑝′ if ∀𝑥, we have 𝑓 𝑥 + 𝑝 = 𝑓(𝑥), and ′𝑝 > 0′.
Further if 𝑝 > 0 is least then it is said to be period of 𝑓(𝑥).
Example: 𝑠𝑖𝑛𝑥 = 𝑠𝑖𝑛 𝑥 + 2𝜋 = 𝑠𝑖𝑛 𝑥 + 4𝜋 = 𝑠𝑖𝑛 𝑥 + 6𝜋 = ⋯
But ′2𝜋′ is least => 𝑝 = 2𝜋 is a period of ′𝑠𝑖𝑛𝑥′.
In general; for
𝑠𝑖𝑛 𝑛𝑥 = 𝑠𝑖𝑛 𝑛𝑥 + 2𝜋
2𝜋
𝑠𝑖𝑛 𝑛 𝑥 + 0 = 𝑠𝑖𝑛 𝑛(𝑥 + )
𝑛
2𝜋
𝑃𝑒𝑟𝑖𝑜𝑑 = 𝑝 =
𝑛
Piecewise Continuous Functions:
𝑓: [𝑎, 𝑏] → 𝑅
It is a function that has at most a finite number of finite discontinuities.

Lecture NO. 18
Fourier Series:
Definition: let 𝑓(𝑥) be a periodic function of period ′2𝑙′ and defined on (−𝑙, 𝑙), then its Fourier series is
defined as
∞
𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
+ 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛
2 𝑙 𝑙
𝑛=1
Where
𝑙 𝑙
1 𝑛𝜋𝑥 1 𝑛𝜋𝑥
𝑎𝑛 = 𝑓 𝑥 𝑐𝑜𝑠 𝑑𝑥 , 𝑏𝑛 = 𝑓 𝑥 𝑠𝑖𝑛 𝑑𝑥 ∴ 𝑛 = 0,1,2,3 …
𝑙 𝑙 𝑙 𝑙
−𝑙 −𝑙
These are called Fourier coefficient.
By changing the period, we can also define coefficients as
𝑐+2𝑙 𝑐+2𝑙
1 𝑛𝜋𝑥 1 𝑛𝜋𝑥
𝑎𝑛 = 𝑓 𝑥 𝑐𝑜𝑠 𝑑𝑥 , 𝑏𝑛 = 𝑓 𝑥 𝑠𝑖𝑛 𝑑𝑥 ∴ 𝑛 = 0,1,2,3 …
𝑙 𝑙 𝑙 𝑙
𝑐 𝑐
𝑙
𝑎0 1
= 𝑓 𝑥 𝑑𝑥 → 𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 𝑓 𝑥 𝑖𝑛 𝑖𝑛𝑡𝑒𝑟𝑎𝑣𝑎𝑙 (−𝑙, 𝑙)
2 2𝑙
−𝑙
Example: Function
𝑓 𝑥 = 𝑥 2 & 𝑖𝑛𝑡𝑒𝑟𝑣𝑎𝑙 (0,2)
2 2
2
1 𝑥3 1 8 4
𝑥 𝑑𝑥 = = . = → 𝐴𝑣𝑒𝑟𝑎𝑔𝑒 𝑣𝑎𝑙𝑢𝑒 𝑖𝑛 𝑡𝑕𝑒 𝑖𝑛𝑡𝑒𝑟𝑣𝑎𝑙 (0,2)
2 3 0
2 3 3
0

Lecture NO. 19
Dirichlet Condition:
Fourier series corresponding to ′𝑓(𝑥)′ is
∞
𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
+ 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛
2 𝑙 𝑙
𝑛=1
Question is; Series converge or diverge? If converge then converges to 𝑓(𝑥) or not.
Suppose that:
Page | 10
MTH647 Handout

𝑖) 𝑓(𝑥) is defined and single valued on (−𝑙, 𝑙)

𝑖𝑖) 𝑓(𝑥) is periodic on (−𝑙, 𝑙) with period ′2𝑙′
𝑖𝑖𝑖) 𝑓 𝑥 & 𝑓 ′ (𝑥) are piecewise continuous, then the series
∞ 𝑓 𝑥 𝑖𝑓 ′𝑥′ 𝑖𝑠 𝑝𝑜𝑖𝑛𝑡 𝑜𝑓 𝑐𝑜𝑛𝑡𝑖𝑛𝑢𝑖𝑡𝑦
𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
+ 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛 = 𝑓 𝑥 + 0 + 𝑓(𝑥 − 0)
2 𝑙 𝑙 𝑖𝑓 ′𝑥′ 𝑖𝑠 𝑝𝑜𝑖𝑛𝑡 𝑜𝑓 𝑑𝑖𝑠𝑐𝑜𝑛𝑡𝑖𝑛𝑢𝑖𝑡𝑦
𝑛=1 2

Lecture NO. 20
Evaluation of Fourier Series:
Theorem: If the series;
∞
𝑛𝜋𝑥 𝑛𝜋𝑥
𝐴+ 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛
𝑙 𝑙
𝑛=1
Converges uniformly to 𝑓(𝑥) in (−𝑙, 𝑙), then for 𝑛 = 0,1,2,3 … show that
𝑙
1 𝑛𝜋𝑥
𝑎𝑛 = 𝑓 𝑥 𝑐𝑜𝑠 𝑑𝑥 … (𝑖)
𝑙 𝑙
−𝑙
𝑙
1 𝑛𝜋𝑥
𝑏𝑛 = 𝑓 𝑥 𝑠𝑖𝑛 𝑑𝑥 … (𝑖𝑖)
𝑙 𝑙
−𝑙
𝑎0
𝐴= … (𝑖𝑖𝑖)
2
Solution: Given that
∞
𝑛𝜋𝑥 𝑛𝜋𝑥
𝑓 𝑥 =𝐴+ 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛 … (1)
𝑙 𝑙
𝑛=1
𝐶𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑖𝑛𝑔 𝑖):
𝑚𝜋𝑥
Multiplying (equation 1) both sides by 𝑐𝑜𝑠 and integrating from – 𝑙 𝑡𝑜 𝑙, we have
𝑙
𝑙 𝑙 ∞ 𝑙 𝑙
𝑚𝜋𝑥 𝑚𝜋𝑥 𝑛𝜋𝑥 𝑚𝜋𝑥 𝑛𝜋𝑥 𝑚𝜋𝑥
𝑓 𝑥 𝑐𝑜𝑠 𝑑𝑥 = 𝐴 𝑐𝑜𝑠 𝑑𝑥 + 𝑎𝑛 𝑐𝑜𝑠 𝑐𝑜𝑠 𝑑𝑥 + 𝑏𝑛 𝑠𝑖𝑛 𝑐𝑜𝑠 𝑑𝑥 … (2)
𝑙 𝑙 𝑙 𝑙 𝑙 𝑙
−𝑙 −𝑙 𝑛=1 −𝑙 −𝑙

(a) (b) (c)

Now we’ll calculate the values for above functions.
𝑙 𝑙
𝑚𝜋𝑥 𝑙 𝑚𝜋 𝑚𝜋𝑥 𝐴𝑙 𝑚𝜋𝑥 𝑙
𝑎 ; 𝐴 𝑐𝑜𝑠 𝑑𝑥 = 𝐴 𝑐𝑜𝑠 𝑑𝑥 = 𝑠𝑖𝑛 =0
𝑙 𝑚𝜋 𝑙 𝑙 𝑚𝜋 𝑙 −𝑙
−𝑙 −𝑙
𝑙 𝑙
𝑛𝜋𝑥 𝑚𝜋𝑥 1 (𝑚 + 𝑛)𝜋𝑥 (𝑚 − 𝑛)𝜋𝑥
𝑏 ; 𝑐𝑜𝑠 𝑐𝑜𝑠 𝑑𝑥 = 𝑐𝑜𝑠 + 𝑐𝑜𝑠 𝑑𝑥 = 0 ; 𝑓𝑜𝑟 𝑚 ≠ 𝑛
𝑙 𝑙 2 𝑙 𝑙
−𝑙 −𝑙
1
∴ Here we use the trigonometric relation; 𝑐𝑜𝑠𝐴 𝑐𝑜𝑠𝐵 = cos 𝐴 + 𝐵 + cos⁡
(𝐴 − 𝐵)
2
For 𝑚 = 𝑛, we have
𝑙 𝑙 𝑙
𝑛𝜋𝑥 𝑚𝜋𝑥 1 𝑚𝜋𝑥 1 𝑚𝜋𝑥 1 𝑙
𝑐𝑜𝑠 𝑐𝑜𝑠 𝑑𝑥 = 𝑐𝑜𝑠 2 𝑑𝑥 = 1 + 𝑐𝑜𝑠2 𝑑𝑥 = 𝑥 −𝑙 =𝑙
𝑙 𝑙 2 𝑙 2 𝑙 2
−𝑙 −𝑙 −𝑙
𝑙 𝑙
𝑛𝜋𝑥 𝑚𝜋𝑥 1 (𝑚 + 𝑛)𝜋𝑥 (𝑚 − 𝑛)𝜋𝑥
𝑐 ; 𝑠𝑖𝑛 𝑐𝑜𝑠 𝑑𝑥 = 𝑠𝑖𝑛 + 𝑠𝑖𝑛 𝑑𝑥 = 0
𝑙 𝑙 2 𝑙 𝑙
−𝑙 −𝑙
1
∴ Here we use the trigonometric relation; 𝑠𝑖𝑛𝐴 𝑐𝑜𝑠𝐵 =
2
sin 𝐴 + 𝐵 + sin⁡
(𝐴 + 𝐵)
Page | 11
MTH647 Handout

Lecture NO. 21
By putting values in equation 2, we get
𝑙 𝑙
𝑚𝜋𝑥 1 𝑛𝜋𝑥
𝑓 𝑥 𝑐𝑜𝑠 𝑑𝑥 = 0 + 𝑎𝑚 𝑙 + 0 => 𝑎𝑛 = 𝑓 𝑥 𝑐𝑜𝑠 𝑑𝑥
𝑙 𝑙 𝑙
−𝑙 −𝑙
This is our required solution for (𝑖).
𝐶𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑖𝑛𝑔 𝑖𝑖 ;
𝑚𝜋𝑥
Multiplying (equation 1) both sides by 𝑠𝑖𝑛 and integrating from – 𝑙 𝑡𝑜 𝑙, we have
𝑙
𝑙 𝑙 ∞ 𝑙 𝑙
𝑚𝜋𝑥 𝑚𝜋𝑥 𝑚𝜋𝑥 𝑛𝜋𝑥 𝑚𝜋𝑥 𝑛𝜋𝑥
𝑓 𝑥 𝑠𝑖𝑛 𝑑𝑥 = 𝐴 𝑠𝑖𝑛 𝑑𝑥 + 𝑎𝑛 𝑠𝑖𝑛 𝑐𝑜𝑠 𝑑𝑥 + 𝑏𝑛 𝑠𝑖𝑛 𝑠𝑖𝑛 𝑑𝑥 … (3)
𝑙 𝑙 𝑙 𝑙 𝑙 𝑙
−𝑙 −𝑙 𝑛=1 −𝑙 −𝑙

(a) (b) (c)

Now we’ll calculate the values for above functions.
𝑙
𝑚𝜋𝑥
𝑎 ; 𝐴 𝑠𝑖𝑛 𝑑𝑥 = 0 ∴ 𝑠𝑖𝑛𝛼 𝑖𝑠 𝑜𝑑𝑑 ∀𝛼 ∈ 𝑅
𝑙
−𝑙
𝑙
𝑚𝜋𝑥 𝑛𝜋𝑥
𝑏 ; 𝑠𝑖𝑛 𝑐𝑜𝑠 𝑑𝑥 = 0 ∴ 𝑠𝑖𝑛𝛼 𝑐𝑜𝑠𝛽 𝑎𝑟𝑒 𝑜𝑑𝑑 𝑓𝑢𝑛𝑐𝑡𝑖𝑜𝑛𝑠 ∀𝛼, 𝛽 ∈ 𝑅
𝑙 𝑙
−𝑙
𝑙 𝑙
𝑚𝜋𝑥 𝑛𝜋𝑥 1 (𝑚 − 𝑛)𝜋𝑥 (𝑚 + 𝑛)𝜋𝑥
𝑐 ; 𝑠𝑖𝑛 𝑠𝑖𝑛 𝑑𝑥 = 𝑐𝑜𝑠 − 𝑐𝑜𝑠 𝑑𝑥 = 0 ; 𝑚 ≠ 𝑛
𝑙 𝑙 2 𝑙 𝑙
−𝑙 −𝑙
For 𝑚 = 𝑛, we have
𝑙 𝑙 𝑙
𝑚𝜋𝑥 𝑛𝜋𝑥 1 2
𝑚𝜋𝑥 1 𝑚𝜋𝑥 1 𝑙
𝑠𝑖𝑛 𝑐𝑜𝑠 𝑑𝑥 = 𝑠𝑖𝑛 𝑑𝑥 = 1 − 𝑐𝑜𝑠2 𝑑𝑥 = 𝑥 −𝑙 =𝑙
𝑙 𝑙 2 𝑙 2 𝑙 2
−𝑙 −𝑙 −𝑙
By putting values in equation 3, we get
𝑙 𝑙
𝑚𝜋𝑥 1 𝑛𝜋𝑥
𝑓 𝑥 𝑠𝑖𝑛 𝑑𝑥 = 0 + 0 + 𝑏𝑚 𝑙 => 𝑏𝑛 = 𝑓 𝑥 𝑠𝑖𝑛 𝑑𝑥
𝑙 𝑙 𝑙
−𝑙 −𝑙
This is our required solution for (𝑖𝑖).
𝐶𝑎𝑙𝑐𝑢𝑙𝑎𝑡𝑖𝑛𝑔 𝑖𝑖𝑖 :
Integrating equation 1 from – 𝑙 𝑡𝑜 𝑙 on both sides,
𝑙 𝑙 ∞ 𝑙 𝑙
𝑛𝜋𝑥 𝑛𝜋𝑥
𝑓 𝑥 = 𝐴 𝑑𝑥 + 𝑎𝑛 𝑐𝑜𝑠 𝑑𝑥 + 𝑏𝑛 𝑠𝑖𝑛 𝑑𝑥
𝑙 𝑙
−𝑙 −𝑙 𝑛=1 −𝑙 −𝑙
𝑙 𝑙
1
𝑓 𝑥 𝑑𝑥 = 2𝐴𝑙 => 2𝐴 = 𝑓 𝑥 𝑑𝑥 … (𝑎)
𝑙
−𝑙 −𝑙
Put 𝑛 = 0 𝑖𝑛 𝑎𝑛 ;
𝑙 𝑙
1 𝑛𝜋𝑥 1
𝑎𝑛 = 𝑓 𝑥 𝑐𝑜𝑠 𝑑𝑥 => 𝑎0 = 𝑓 𝑥 𝑑𝑥 … (𝑏)
𝑙 𝑙 𝑙
−𝑙 −𝑙
𝑎0
Comparing 𝑎 𝑎𝑛𝑑 𝑏, we have 𝑎0 = 2𝐴 => 𝐴 =
2
Hence equation 1 (Fourier Series) becomes
∞
𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
𝑓 𝑥 = + 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛
2 𝑙 𝑙
𝑛=1

Page | 12
MTH647 Handout

Lecture NO. 22
Problem:
Expand 𝑓 𝑥 = 𝑥 2 ; 0 < 𝑥 < 2𝜋 in a Fourier Series.
Solution: Fourier Series
∞
𝑛𝜋𝑥 𝑛𝜋𝑥
𝑓 𝑥 =𝐴+ 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛
𝑙 𝑙
𝑛=1
According to given function, (∴ 𝑙 = 𝜋)
∞

𝑓 𝑥 =𝐴+ 𝑎𝑛 𝑐𝑜𝑠 𝑛𝑥 + 𝑏𝑛 𝑠𝑖𝑛 𝑛𝑥 … 𝐴

𝑛=1
As we know from previous lectures,
𝑐+2𝑙
1 𝑛𝜋𝑥
𝑎𝑛 = 𝑓 𝑥 𝑐𝑜𝑠 𝑑𝑥
𝑙 𝑙
𝑐
According to given function,
2𝜋 2𝜋
1 𝑛𝜋𝑥 1
𝑎𝑛 = 𝑥 2 𝑐𝑜𝑠 𝑑𝑥 => 𝑥 2 𝑐𝑜𝑠 𝑛𝑥 𝑑𝑥 ∴𝑙=𝜋
𝜋 𝜋 𝜋
0 0
By integrating and applying limits, we have
2𝜋
1 𝑠𝑖𝑛(𝑛𝑥) 𝑐𝑜𝑠(𝑛𝑥) 𝑠𝑖𝑛(𝑛𝑥)
𝑎𝑛 = 𝑥 2 − 2𝑥 − + 2 −
𝜋 𝑛 𝑛2 𝑛3 0
1 4𝜋(1) 1 4𝜋 4
𝑎𝑛 = 4𝜋 2 0 + 2
+0 = 2
= 2 … (1)
𝜋 𝑛 𝜋 𝑛 𝑛
Now
𝑙
1 𝑛𝜋𝑥
𝑏𝑛 = 𝑓 𝑥 𝑠𝑖𝑛 𝑑𝑥
𝑙 𝑙
−𝑙
According to given function,
2𝜋
1
𝑏𝑛 = 𝑥 2 𝑠𝑖𝑛 𝑛𝑥 𝑑𝑥
𝜋
0
By integrating and applying limits, we have
2𝜋
1 𝑐𝑜𝑠(𝑛𝑥) 𝑠𝑖𝑛(𝑛𝑥) 𝑐𝑜𝑠(𝑛𝑥)
𝑏𝑛 = 𝑥 2 − − 2𝑥 − + 2
𝜋 𝑛 𝑛2 𝑛3 0
1 4𝜋 2 (−1) 4𝜋
𝑏𝑛 = 0 + +0 =− … (2)
𝜋 𝑛 𝑛
Now
2𝜋 2𝜋 2𝜋
𝑎0 1 1 2
1 𝑥3 8𝜋 2
𝐴= => 𝑓 𝑥 𝑑𝑥 => 𝑥 𝑑𝑥 => => … (3)
2 𝜋 𝜋 𝜋 3 0
3
0 0
Putting values in Fourier series,
∞
4𝜋 2 4 4𝜋
𝑓 𝑥 = + 2
𝑐𝑜𝑠 𝑛𝑥 − 𝑠𝑖𝑛 𝑛𝑥
3 𝑛 𝑛
𝑛=1

Lecture NO. 23
Fourier Expansion of Even Functions:
Theorem: Show that an even function does not have 𝑠𝑖𝑛𝑒 terms in its Fourier Series.
Proof: Since
∞
𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
𝑓 𝑥 = + 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛 … (1)
2 𝑙 𝑙
𝑛=1
Page | 13
MTH647 Handout

If 𝑓(𝑥) is even function, then we know

𝑓 𝑥 = 𝑓(−𝑥)
So, replacing 𝑥 𝑏𝑦 − 𝑥, we have
∞
𝑎0 −𝑛𝜋𝑥 −𝑛𝜋𝑥
𝑓 −𝑥 = + 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛
2 𝑙 𝑙
𝑛=1
As cos −𝑥 = cos & sin −𝑥 = −𝑠𝑖𝑛𝑥, hence
∞
𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
𝑓 −𝑥 = + 𝑎𝑛 𝑐𝑜𝑠 − 𝑏𝑛 𝑠𝑖𝑛 … (2)
2 𝑙 𝑙
𝑛=1
By comparing equation 1 & 2,
𝑓 𝑥 = 𝑓(−𝑥)
∞ ∞
𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥 𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
+ 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛 = + 𝑎𝑛 𝑐𝑜𝑠 − 𝑏𝑛 𝑠𝑖𝑛
2 𝑙 𝑙 2 𝑙 𝑙
𝑛=1 𝑛=1
∞ ∞ ∞ ∞
𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥 𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
+ 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛 = + 𝑎𝑛 𝑐𝑜𝑠 − 𝑏𝑛 𝑠𝑖𝑛
2 𝑙 𝑙 2 𝑙 𝑙
𝑛=1 𝑛=1 𝑛=1 𝑛=1
By cancelling the terms, we have
∞ ∞
𝑛𝜋𝑥 𝑛𝜋𝑥
𝑏𝑛 𝑠𝑖𝑛 =− 𝑏𝑛 𝑠𝑖𝑛
𝑙 𝑙
𝑛=1 𝑛=1
∞ ∞
𝑛𝜋𝑥 𝑛𝜋𝑥
2 𝑏𝑛 𝑠𝑖𝑛 = 0 => 𝑏𝑛 𝑠𝑖𝑛 =0
𝑙 𝑙
𝑛=1 𝑛=1
So equation (1) implies,
∞
𝑎0 𝑛𝜋𝑥
𝑓 𝑥 = + 𝑎𝑛 𝑐𝑜𝑠
2 𝑙
𝑛=1
Hence proved that an even function does not have 𝑠𝑖𝑛𝑒 terms in its Fourier Series.

Lecture NO. 24
Half Range Expansion of Identity Function:
Expand 𝑓 𝑥 = 𝑥 ; 0 < 𝑥 < 2 in half-range 𝑠𝑖𝑛𝑒𝑠 & 𝑐𝑜𝑠𝑖𝑛𝑒 𝑠𝑒𝑟𝑖𝑒𝑠.
Solution:
(𝒊); For odd extension of 𝑓 𝑥 = 𝑥 in −2,2 ∴ −2 < 𝑥 < 2
This implies 2𝑙 = 4 𝑠𝑜 𝑙 = 2
For odd expansion; 𝑎𝑛 = 0
(∴ 𝑎0 = 0 𝑎𝑙𝑠𝑜) And
𝑙
2 𝑛𝜋𝑥
𝑏𝑛 = 𝑓(𝑥) 𝑠𝑖𝑛 𝑑𝑥
𝑙 𝑙
0
For given function,
2 2
2 𝑛𝜋𝑥 𝑛𝜋𝑥
𝑏𝑛 = 𝑥 𝑠𝑖𝑛 𝑑𝑥 => 𝑥 𝑠𝑖𝑛 𝑑𝑥
2 2 2
0 0
By integrating and applying limits, we have
2
−2 𝑛𝜋𝑥 4 𝑛𝜋𝑥
𝑏𝑛 = 𝑥 𝑐𝑜𝑠 − 1 𝑠𝑖𝑛
𝑛𝜋 2 𝑛2 𝜋 2 2 0
4
𝑏𝑛 = − 𝑐𝑜𝑠 (𝑛𝜋)
𝑛𝜋
Now, putting values in Fourier Series
Page | 14
MTH647 Handout

∞
𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
𝑓 𝑥 = + 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛
2 𝑙 𝑙
𝑛=1
∞ ∞
𝑛𝜋𝑥 4 𝑛𝜋𝑥
𝑓 𝑥 =0+0+ 𝑏𝑛 𝑠𝑖𝑛 => − 𝑐𝑜𝑠 (𝑛𝜋) 𝑠𝑖𝑛
2 𝑛𝜋 2
𝑛=1 𝑛=1
4 𝜋𝑥 1 2𝜋𝑥 1 3𝜋𝑥
𝑓 𝑥 = 𝑠𝑖𝑛 − 𝑠𝑖𝑛 + 𝑠𝑖𝑛 − …
𝜋 2 2 2 3 2
Required result.
Lecture NO. 25

(𝒊𝒊); For even extension of 𝑓 𝑥 = 𝑥 in −2,2

This implies 2𝑙 = 4 𝑠𝑜 𝑙 = 2
For odd expansion; 𝑏𝑛 = 0 and
𝑙
2 𝑛𝜋𝑥
𝑎𝑛 = 𝑓(𝑥) 𝑐𝑜𝑠 𝑑𝑥
𝑙 𝑙
0
For given function,
2 2
2 𝑛𝜋𝑥 𝑛𝜋𝑥
𝑎𝑛 = 𝑥 𝑐𝑜𝑠 𝑑𝑥 => 𝑥 𝑐𝑜𝑠 𝑑𝑥
2 2 2
0 0
By integrating and applying limits, we have
2 𝑛𝜋𝑥 −4 𝑛𝜋𝑥 2
𝑎𝑛 = 𝑥 𝑠𝑖𝑛 − 1 𝑐𝑜𝑠
𝑛𝜋 2 𝑛2 𝜋 2 2 0
4 4
𝑎𝑛 = 0 + 2 2 𝑐𝑜𝑠 𝑛𝜋 − 1 => 2 2 −1 𝑛 − 1
𝑛 𝜋 𝑛 𝜋
For 𝑛 = 0,
𝑙 2 2
2 2 𝑥2
𝑎0 = 𝑥𝑑𝑥 => 𝑥𝑑𝑥 => => 2
𝑙 2 2 0
0 0
𝑎0
=1
2
Now putting values in Fourier Series,
∞
𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
𝑓 𝑥 = + 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛
2 𝑙 𝑙
𝑛=1
∞ ∞
𝑛𝜋𝑥 4 𝑛
𝑛𝜋𝑥
𝑓 𝑥 =1+ 𝑎𝑛 𝑐𝑜𝑠 + 0 => 1 + −1 − 1 𝑐𝑜𝑠
2 𝑛2 𝜋 2 2
𝑛=1 𝑛=1
4 𝜋𝑥 2 3𝜋𝑥 2 5𝜋𝑥
𝑓 𝑥 = 1 + 2 −2𝑐𝑜𝑠 − 2 𝑐𝑜𝑠 − 2 𝑐𝑜𝑠 − …
𝜋 2 3 2 5 2
8 𝜋𝑥 1 3𝜋𝑥 1 5𝜋𝑥
𝑓 𝑥 = 1 − 2 𝑐𝑜𝑠 + 2 𝑐𝑜𝑠 + 2 𝑐𝑜𝑠 + …
𝜋 2 3 2 5 2
Required result.
Lecture NO. 26

Parseval’s Identity:
Theorem: If Fourier Series of 𝑓(𝑥) converges uniformly to 𝑓 𝑥 𝑖𝑛 (−𝑙, 𝑙), then prove that
𝑙 ∞
1 𝑎0 2
2
𝑓(𝑥) 𝑑𝑥 = + 𝑎𝑛 2 + 𝑏𝑛 2
𝑙 2
−𝑙 𝑛=1
Proof:
Page | 15
MTH647 Handout

∞
𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
𝑓 𝑥 = + 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛 … (1)
2 𝑙 𝑙
𝑛=1
Multiplying both sides by 𝑓(𝑥) and integrating form – 𝑙 𝑡𝑜 𝑙,
𝑙 𝑙 ∞ 𝑙 𝑙
𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
𝑓(𝑥) 2 𝑑𝑥 = 𝑓 𝑥 𝑑𝑥 + 𝑎𝑛 𝑓(𝑥) 𝑐𝑜𝑠 𝑑𝑥 + 𝑏𝑛 𝑓(𝑥) 𝑠𝑖𝑛 𝑑𝑥 … (𝐴)
2 𝑙 𝑙
−𝑙 −𝑙 𝑛=1 −𝑙 −𝑙
∴
𝑙 𝑙
1 𝑛𝜋𝑥 𝑛𝜋𝑥
𝑎𝑛 = 𝑓(𝑥) 𝑐𝑜𝑠 𝑑𝑥 => 𝑎𝑛 𝑙 = 𝑓(𝑥) 𝑐𝑜𝑠 𝑑𝑥
𝑙 𝑙 𝑙
−𝑙 −𝑙
Similarly for
𝑙
𝑛𝜋𝑥
𝑏𝑛 𝑙 = 𝑓(𝑥) 𝑠𝑖𝑛 𝑑𝑥
𝑙
−𝑙
𝑙

𝑎0 𝑙 = 𝑓 𝑥 𝑑𝑥
−𝑙
So, equation (A) can be written as
𝑙 ∞
𝑎0
𝑓(𝑥) 2 𝑑𝑥 = (𝑎0 𝑙) + 𝑎𝑛 𝑎𝑛 𝑙 + 𝑏𝑛 (𝑏𝑛 𝑙)
2
−𝑙 𝑛=1
𝑙 ∞
𝑎0 2
𝑓 𝑥 2 𝑑𝑥 = 𝑙 + (𝑙) 𝑎𝑛 2 + 𝑏𝑛 2
2
−𝑙 𝑛=1
𝑙 ∞
1 𝑎0 2
𝑓(𝑥) 𝑑𝑥 = 2
+ 𝑎𝑛 2 + 𝑏𝑛 2
𝑙 2
−𝑙 𝑛=1
Required result proved.
Lecture NO. 27

Problem:
Use Parseval’s identity for 𝑓 𝑥 = 𝑥 , −2 < 𝑥 < 2 to show that
1 1 1 𝜋4
+ + + …=
14 34 54 96
Where it is given
4
𝑎0 = 2 , 𝑎𝑛 = 2 2 −1 𝑛 − 1 , 𝑛 ≠ 0 & 𝑏𝑛 = 0
𝑛 𝜋
Also deduce that
1 1 1 𝜋4
+ + + …=
14 24 34 90
Solution: Parseval’s identity
𝑙 ∞
1 𝑎0 2
𝑓(𝑥) 𝑑𝑥 = 2
+ 𝑎𝑛 2 + 𝑏𝑛 2 … (1)
𝑙 2
−𝑙 𝑛=1
For 𝑓 𝑥 = 𝑥 ; −2,2 => 2𝑙 = 4 𝑠𝑜 𝑙 = 2
Equation (1) becomes
2 ∞
1 2
22 16 𝑛 2
𝑥 𝑑𝑥 = + −1 −1 +0
2 2 𝑛4 𝜋 4
−2 𝑛=1
By solving, we have

Page | 16
MTH647 Handout

8 64 1 1 1
=2+ 4 4+ 4+ 4+ …
3 𝜋 1 3 5
1 1 1 𝜋4
+ + + … =
14 34 54 96
Proved. Now let
1 1 1
+ 𝑆= + + …
14 24 34
1 1 1 1 1 1
𝑆= 4+ 4+ 4+ … + 4+ 4+ 4+ …
1 3 5 2 4 6
4
𝜋 1 1 1 1
𝑆= + 4 4+ 4+ 4+ …
96 2 1 2 3
𝜋4 1
𝑆= + 4 (𝑆)
96 2
4
𝜋
𝑆=
90
Deduced.
Lecture NO. 28

Finding a Fourier Series by Integration:

Theorem: Fourier series of 𝑓(𝑥) can be integrated by term form 𝑎 𝑡𝑜 𝑥 and the resulting series with
𝑥
converge uniformly to 0
𝑓 𝑥 provided that;
𝑖 : 𝑓(𝑥) is piecewise continued in (−𝑙, 𝑙) 𝑖𝑖 : 𝑎, 𝑥 ∈ (−𝑙, 𝑙)
2
Problem: Find the Fourier Series of 𝑓 𝑥 = 𝑥 , 0 < 𝑥 < 2, by integrating
4 𝜋𝑥 1 2𝜋𝑥 1 3𝜋𝑥
𝑥 = 𝑠𝑖𝑛 − 𝑠𝑖𝑛 + 𝑠𝑖𝑛 − …
𝜋 2 2 2 3 2
And further evaluate;
∞
−1 𝑛−1
=?
𝑛2
𝑛=1
Solution: Given that
4 𝜋𝑥 1 2𝜋𝑥 1 3𝜋𝑥
𝑥=
𝑠𝑖𝑛 − 𝑠𝑖𝑛 + 𝑠𝑖𝑛 − …
𝜋 2 2 2 3 2
Integrating from 0 𝑡𝑜 𝑥 term by term
𝑥 𝑥 𝑥 𝑥
4 𝜋𝑥 1 2𝜋𝑥 1 3𝜋𝑥
𝑥𝑑𝑥 = 𝑠𝑖𝑛 𝑑𝑥 − 𝑠𝑖𝑛 𝑑𝑥 + 𝑠𝑖𝑛 𝑑𝑥
𝜋 2 2 2 3 2
0 0 0 0
2 𝑥 𝑥 𝑥 𝑥
𝑥 4 2 𝜋𝑥 2 1 2𝜋𝑥 2 1 3𝜋𝑥
= −𝑐𝑜𝑠 + 𝑐𝑜𝑠 − 𝑐𝑜𝑠 − …
2 0
𝜋 𝜋 2 0 𝜋 22 2 0 𝜋 32 2 0
2
𝑥 4 2 𝜋𝑥 2 1 2𝜋𝑥 2 1 3𝜋𝑥
= −𝑐𝑜𝑠 +1 + 𝑐𝑜𝑠 − 1 − 𝑐𝑜𝑠 −1
2 𝜋 𝜋 2 𝜋 22 2 𝜋 32 2
16 1 1 1 16 𝜋𝑥 1 2𝜋𝑥 1 3𝜋𝑥
𝑥 2 = 2 1 − 2 + 2 − 2 + … − 2 𝑐𝑜𝑠 − 2 𝑐𝑜𝑠 + 2 𝑐𝑜𝑠 − …
𝜋 2 3 4 𝜋 2 2 2 3 2
Comparing above relation with general form;
∞
𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
𝑓(𝑥) = + 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛
2 𝑙 𝑙
𝑛=1
We have
𝑙
𝑎0 1 16 1 1 1
=> 𝑓 𝑥 𝑑𝑥 = 1 − + − + … … (1)
2 𝑙 𝜋2 22 32 42
0
Here
Page | 17
MTH647 Handout

𝑙 2 2
1 1 2
1 𝑥3 4
𝑓 𝑥 𝑑𝑥 = 𝑥 𝑑𝑥 = =
𝑙 2 2 3 0
3
0 0
So equation (1) becomes,
4 16 1 1 1
= 2 1− 2+ 2− 2+ …
3 𝜋 2 3 4
∞
𝑛−1
4 16 −1
= 2
3 𝜋 𝑛2
𝑛=1
∞
𝑛−1
−1 𝜋2
=
𝑛2 12
𝑛=1
Required result.
Lecture NO. 29

Problem:
Check the term by term differentiation of Fourier Series;
4 𝜋𝑥 1 2𝜋𝑥 1 3𝜋𝑥
𝑓(𝑥) = 𝑠𝑖𝑛 − 𝑠𝑖𝑛 + 𝑠𝑖𝑛 − …
𝜋 2 2 2 3 2
Solution: Taking derivative,
4 𝜋 𝜋𝑥 1 2𝜋 2𝜋𝑥 1 3𝜋 3𝜋𝑥
𝑓′(𝑥) = 𝑐𝑜𝑠 − 𝑐𝑜𝑠 + 𝑐𝑜𝑠 − …
𝜋 2 2 2 2 2 3 2 2
𝜋𝑥 2𝜋𝑥 3𝜋𝑥
𝑓′(𝑥) = 2 𝑐𝑜𝑠 − 𝑐𝑜𝑠 + 𝑐𝑜𝑠 − …
2 2 2
This implies
𝑛𝜋𝑥
𝑎𝑛 = −1 𝑛−1 𝑐𝑜𝑠 ⇏0
2
⇒ Series does not converge ⇒ It does not converges uniformly ⇒ Term by term differentiation is not
possible.
Lecture NO. 30

Heat Flow Problem:

A bar of length "𝑙" whose entire surface is insulated including its ends at 𝑥 = 0 and 𝑥 = 𝑙. Its initial
temperature is 𝑓 𝑥 , then determine the subsequent temperature of the bar.
Solution: It is a heat flow boundary value problem (B.V.P.). As we know heat equation
𝜕𝑢 𝜕2 𝑢
= 𝑘 2 … (1)
𝜕𝑡 𝜕𝑥
Where
𝑢(𝑥, 𝑡) < 𝑀 , 𝑢 𝑥, 0 = 𝑓 𝑥 , 𝑢𝑥 0, 𝑡 = 0 = 𝑢𝑥 (𝑙, 𝑡)
Let 𝑢 𝑥, 𝑡 = 𝑋 𝑥 𝑇 𝑡 = 𝑋𝑇 be its solution. Equation (1) implies,
𝑋𝑇 ′ = 𝑘𝑋′′𝑇
𝑇′ 𝑋′′
= = −𝜆2 (𝑠𝑎𝑦)
𝑘𝑇 𝑋
Here become two equations. We’ll solve them side by side.
𝑇 ′ + 𝜆2 𝑘𝑇 = 0 … (𝑖) 𝑋 ′′ + 𝜆2 𝑋 = 0 … (𝑖𝑖)
Let 𝑇 = 𝑒 𝑚𝑡 𝑋 = 𝑒 𝑚𝑥 => 𝑋 ′′ = 𝑚2 𝑒 𝑚𝑥 = 𝑚2 𝑋
𝑇 ′ = 𝑚𝑒 𝑚𝑡 => 𝑚𝑇 𝑚2 𝑋 + 𝜆2 𝑋 = 0 => 𝑚 = ±𝑖𝜆 (∴ 𝑥 ≠ 0)
So 𝑖 ; 𝑚𝑇 + 𝜆2 𝑘𝑇 = 0 𝑋 = 𝑎𝑒 𝜆𝑖𝑥 + 𝑏𝑒 −𝜆𝑖𝑥
𝑚 = −𝜆2 𝑘 𝑋 = 𝑎 𝑐𝑜𝑠𝜆𝑥 + 𝑖𝑠𝑖𝑛𝜆𝑥 + 𝑏 𝑐𝑜𝑠𝜆𝑥 − 𝑖𝑠𝑖𝑛𝜆𝑥
2 𝑘𝑡
∴ 𝑇 = 𝑐𝑒 −𝜆 𝑋 = 𝑎 + 𝑏 𝑐𝑜𝑠𝜆𝑥 + 𝑖 𝑎 − 𝑏 𝑠𝑖𝑛𝜆𝑥 = 𝐴𝑐𝑜𝑠𝜆𝑥 + 𝐵𝑠𝑖𝑛𝜆𝑥
By putting values,
Page | 18
MTH647 Handout

2 𝑘𝑡 2 𝑘𝑡
𝑢 𝑥, 𝑡 = 𝑋 𝑥 𝑇 𝑡 = 𝑋𝑇 = 𝑐𝑒 −𝜆 𝐴𝑐𝑜𝑠𝜆𝑥 + 𝐵𝑠𝑖𝑛𝜆𝑥 = 𝑒 −𝜆 𝛼𝑐𝑜𝑠𝜆𝑥 + 𝛽𝑠𝑖𝑛𝜆𝑥 … (2)
Now
2 𝑘𝑡
𝑢𝑥 = 𝑒 −𝜆 −𝛼𝜆 𝑐𝑜𝑠𝜆𝑥 + 𝛽𝜆 𝑠𝑖𝑛𝜆𝑥
2 𝑘𝑡
𝑢𝑥 0, 𝑡 = 𝛽 𝜆 𝑒 −𝜆 = 0 => 𝛽 = 0
Equation (2) implies,
2 𝑘𝑡
𝑢 𝑥, 𝑡 = 𝛼 𝑒 −𝜆 𝑐𝑜𝑠 𝜆𝑥 … (3)
−𝜆 2 𝑘𝑡
𝑢𝑥 = −𝜆 𝛼 𝑒 𝑠𝑖𝑛 𝜆𝑥
−𝜆 2 𝑘𝑡
𝑢𝑥 𝑙, 𝑡 = −𝜆 𝛼 𝑒 𝑠𝑖𝑛 𝜆𝑙 = 0
𝑚𝜋
=> 𝑠𝑖𝑛 𝜆𝑙 = 0 => 𝜆𝑙 = 𝑚𝜋 => 𝜆 =
𝑙
Equation (3) implies,
𝑚 2𝜋 2 𝑚𝜋
− 𝑘𝑡
𝑢 𝑥, 𝑡 = 𝛼 𝑒 𝑙2 𝑐𝑜𝑠 𝑥
𝑙
By using super-position principle;
∞
𝑎0 𝑚 2𝜋 2 𝑚𝜋
− 𝑘𝑡
𝑢 𝑥, 𝑡 = + 𝛼𝑚 𝑒 𝑙2 𝑐𝑜𝑠 𝑥 … (𝐴)
2 𝑙
𝑚 =1
Now; as given
∞
𝑎0 𝑚𝜋
𝑢 𝑥, 0 = 𝑓 𝑥 = + 𝛼𝑚 𝑐𝑜𝑠 𝑥
2 𝑙
𝑚 =1
Where
𝑙 𝑙
2 𝑚𝜋 𝑎0 1
𝛼𝑚 = 𝑓 𝑥 𝑐𝑜𝑠 𝑥 , = 𝑓(𝑥)𝑑𝑥
𝑙 𝑙 2 𝑙
0 0
So, equation (A) becomes (by putting values)
𝑙 ∞ 𝑙
1 2 𝑚𝜋 𝑚 2𝜋 2 𝑚𝜋
− 2 𝑘𝑡
𝑢 𝑥, 𝑡 = 𝑓 𝑥 𝑑𝑥 + 𝑓 𝑥 𝑐𝑜𝑠 𝑥 𝑒 𝑙 𝑐𝑜𝑠 𝑥
𝑙 𝑙 𝑙 𝑙
0 𝑚 =1 0
Required result.

Lecture NO. 31
Laplace Equation:
Problem: Suppose that the three sides of a square plate kept at zero temperature and fourth one at 𝑢1 .
Determine the steady state temperature at all parts on the plate.
For Solution; WATCH Lecture (Lengthy Calculations)

Lecture NO. 32

Orthogonal Functions And Orthogonal Sets:

𝑖 : Can we generalize the idea of vectors and orthogonality?
𝑓 𝑥 𝑑𝑒𝑓𝑖𝑛𝑒𝑑 𝑜𝑛 𝑎, 𝑏 ; 𝑣𝑎𝑙𝑢𝑒 𝑜𝑓 ′𝑓′ 𝑎𝑡 𝑒𝑎𝑐ℎ 𝑝𝑜𝑖𝑛𝑡 𝑜𝑛 𝑎, 𝑏 𝑟𝑒𝑝𝑟𝑒𝑠𝑒𝑛𝑡𝑠 𝑖𝑡𝑠 𝑐𝑜𝑚𝑜𝑝𝑜𝑛𝑒𝑛𝑡𝑠.
𝑏
𝑖𝑖 : Say 𝐴 𝑥 & 𝐵(𝑥) defined on (𝑎, 𝑏) if 𝑎
𝐴 𝑥 . 𝐵 𝑥 = 0 , then 𝐴 𝑥 & 𝐵(𝑥) are said to orthogonal.
𝑖𝑖𝑖 : A vector say 𝑟 = 𝑥𝑖 + 𝑦𝑗 + 𝑧𝑘 is said to normalized if its magnitude is unity i.e.
𝑟 𝑥𝑖 + 𝑦𝑗 + 𝑧𝑘
𝑟= =
𝑟 𝑥2 + 𝑦2 + 𝑧2
𝑖𝑣 : 𝐴(𝑥) normal or normalized in (𝑎, 𝑏) if
𝑏
2
𝐴 𝑥 𝑑𝑥 = 1
𝑎
Page | 19
MTH647 Handout

Orthogonal Sets:
A set of function say 𝛷𝑘 (𝑥) , 𝑘 = 1,2,3, … defined on (𝑎, 𝑏) such that
𝑏

𝑖 : 𝛷𝑚 𝑥 . 𝛷𝑛 𝑥 = 0 , 𝑚 ≠ 𝑛
𝑎
𝑏
2
𝑖𝑖 : 𝛷𝑚 𝑥 = 1 , 𝑚 = 1,2,3, …
𝑎
Then the set 𝛷𝑘 (𝑥) , 𝑘 = 1,2,3, … is orthogonal.
𝑖 & (𝑖𝑖) implies,
𝑏
0 𝑖𝑓 𝑚 ≠ 𝑛
𝛷𝑚 𝑥 . 𝛷𝑛 𝑥 𝑑𝑥 = 𝑆𝑚𝑛 =
1 𝑖𝑓 𝑚 = 𝑛
𝑎
Orthogonality w.r.to Weight Function:
If
𝑏

𝛷𝑚 𝑥 𝛷𝑛 𝑥 ⍵ 𝑥 𝑑𝑥 = 𝑆𝑚𝑛
𝑎
∞
Where ⍵ 𝑥 ≥ 0, then 𝛷𝑚 𝑥 𝑘=1 is orthogonal w.r.to weight function.

Lecture NO. 33

Obtaining Normalizing Constants From orthogonal Sets:

Show that the set
𝜋𝑥 𝜋𝑥 2𝜋𝑥 2𝜋𝑥
1, 𝑠𝑖𝑛 , 𝑐𝑜𝑠 , 𝑠𝑖𝑛 , 𝑐𝑜𝑠 ,…
𝑙 𝑙 𝑙 𝑙
Is an orthogonal set. Also find its corresponding normalizing constants, so that given set is orthogonal.
Solution: As we know condition for orthogonal set condition
𝑙
0 𝑖𝑓 𝑚 ≠ 𝑛
𝛷𝑚 𝑥 . 𝛷𝑛 𝑥 𝑑𝑥 =
1 𝑖𝑓 𝑚 = 𝑛
−𝑙
Possibilities for given set (using Fourier series results;
𝑙 𝑙
𝑘𝜋𝑥 𝑘𝜋𝑥
𝑖 : 1 𝑠𝑖𝑛 𝑑𝑥 = 0 = 1 𝑐𝑜𝑠 𝑑𝑥 ∀ 𝑘 = 1,2,3, …
𝑙 𝑙
−𝑙 −𝑙
𝑙
𝑘𝜋𝑥 𝑝𝜋𝑥
𝑖𝑖 : 𝑠𝑖𝑛 𝑐𝑜𝑠 𝑑𝑥 = 0 𝑘≠𝑝
𝑙 𝑙
−𝑙
𝑙 𝑙
𝑚𝜋𝑥 𝑛𝜋𝑥 𝑚𝜋𝑥 𝑛𝜋𝑥 0 𝑖𝑓 𝑚 ≠ 𝑛
𝑖𝑖𝑖 : 𝑠𝑖𝑛 𝑠𝑖𝑛 𝑑𝑥 = 𝑐𝑜𝑠 𝑐𝑜𝑠 𝑑𝑥 =
𝑙 𝑙 𝑙 𝑙 𝑙 𝑖𝑓 𝑚 = 𝑛
−𝑙 −𝑙
(iii) Implies
𝑙 𝑙
𝑚𝜋𝑥 𝑚𝜋𝑥
𝑠𝑖𝑛 2 𝑑𝑥 = 𝑙 & 𝑐𝑜𝑠 2 𝑑𝑥 = 𝑙
𝑙 𝑙
−𝑙 −𝑙
𝑙 𝑙
2 2
1 𝑚𝜋𝑥 1 𝑚𝜋𝑥
𝑠𝑖𝑛 𝑑𝑥 = 1 & 𝑐𝑜𝑠 𝑑𝑥 = 1
𝑙 𝑙 𝑙 𝑙
−𝑙 −𝑙
Similarly for (i):
𝑙 𝑙
2
2
1
1 𝑑𝑥 = 2𝑙 => 𝑑𝑥 = 1
2𝑙
−𝑙 −𝑙
Page | 20
MTH647 Handout

Hence, Orthonormalizing constants;

1 1
,
2𝑙 𝑙
Corresponding Orthonormal set;
1 1 𝜋𝑥 1 𝜋𝑥 1 2𝜋𝑥 1 2𝜋𝑥
, 𝑠𝑖𝑛 , 𝑐𝑜𝑠 , 𝑠𝑖𝑛 , 𝑐𝑜𝑠 ,…
2𝑙 𝑙 𝑙 𝑙 𝑙 𝑙 𝑙 𝑙 𝑙

Lecture NO. 34

Generalized Fourier Series:

Given that 𝛷𝑚 𝑥 ∞ 𝑚 =1 be an orthogonal set of functions and if possible to expand a function "𝑓(𝑥)" in a
set of orthonormal functions, i.e.
∞

𝑓 𝑥 = 𝐶𝑛 𝛷𝑛 𝑥 ; 𝑎≤𝑥≤𝑏
𝑛=1
Then such series is called Orthonormal Series or Generalized Fourier Series and "𝐶𝑛 " are called Generalized
Fourier Coefficients.
If 𝑓 𝑥 & 𝑓 ′ (𝑥) are piecewise continuous functions, then
∞
1
𝐶𝑛 𝛷𝑛 𝑥 = 𝑓 𝑥 + 0 + 𝑓(𝑥 − 0)
2
𝑛=1
“If a function and its derivative are piecewise continuous, then the series converge to their average”.

Lecture NO. 35
Theorem:
∞
Let 𝛷𝑛 𝑥 be a set of mutually orthonormal functions in (𝑎, 𝑏). Show that if 𝑛=1 𝐶𝑛 𝛷𝑛 𝑥 converges
𝑏
uniformally, then 𝐶𝑛 = 𝑎
𝑓 𝑥 𝛷𝑛 𝑥 𝑑𝑥.
Proof: As given that
∞
The series 𝑛=1 𝐶𝑛 𝛷𝑛 𝑥 converges uniformally to 𝑓(𝑥)
This implies,
∞

𝑓 𝑥 = 𝐶𝑛 𝛷𝑛 𝑥 … (1)
𝑛=1
Multiplying both sides by 𝛷𝑚 𝑥 and integrating form ′𝑎 𝑡𝑜 𝑏′
𝑏 ∞ 𝑏

𝑓 𝑥 𝛷𝑚 𝑥 𝑑𝑥 = 𝐶𝑛 𝛷𝑚 𝑥 𝛷𝑛 𝑥 𝑑𝑥 … (2)
𝑎 𝑛=1 𝑎
As
𝑏
0 𝑖𝑓 𝑚 ≠ 𝑛
𝛷𝑚 𝑥 𝛷𝑛 𝑥 𝑑𝑥 =
𝑙 𝑖𝑓 𝑚 = 𝑛
𝑎
Hence equation (2) impies
𝑏

𝐶𝑚 = 𝑓(𝑥) 𝛷𝑚 𝑥 𝑑𝑥
𝑎
Or it can be written as;
𝑏

𝐶𝑛 = 𝑓(𝑥) 𝛷𝑛 𝑥 𝑑𝑥
𝑎
Which are called the generalized fourier constants.

Page | 21
MTH647 Handout

Lecture NO. 36
MSE and RMSE:
MSE stands for “Mean Square Error” and RMSE stands for “Root Mean Square Error”.
Let 𝑓 𝑥 𝑎𝑛𝑑 𝑓 ′ (𝑥) be continuous piecewise functions in 𝑎, 𝑏 , 𝛷𝑚 (𝑥) ∞ 𝑚 =1 be an orthogonal set
in (𝑎, 𝑏).
Now suppose the sum 𝑆𝑚 𝑥 = ∞ 𝑛=1 𝛼𝑛 𝛷𝑛 (𝑥) be an approximation of 𝑓 𝑥 , where "𝛼𝑛 " are unknown
constants. Then the mean square error of this approximation is given by,
𝑏 2
𝑎
𝑓 𝑥 + 𝑆𝑚 𝑥 𝑑𝑥
𝑀𝑆𝐸 =
𝑏−𝑎
And
𝑏 2 𝑑𝑥
𝑎
𝑓 𝑥 + 𝑆𝑚 𝑥
𝑅𝑀𝑆𝐸 =
𝑏−𝑎
These are used to compare the accuracy of different mathematical methods of a physical system.
We are aimed to find "𝛼𝑛 " which with produce the least mean square error.
Use: (MSE; for large or greater errors/quantities & RMSE; for same units as that of given function)

Lecture NO. 37

Lease Square Approximation & Principle of Finality:

Theorem: The root mean square error (RMSE) is least when the coefficients "𝛼𝑛 " are equal to generalized
Fourier Coefficient; i.e.
𝑛

𝛼𝑛 = 𝑐𝑛 = 𝑓 𝑥 𝛷𝑛 (𝑥)𝑑𝑥
𝑎
Proof: As we known
𝑏 𝑏
𝑓 𝑥 + 𝑆𝑚 𝑥 2 𝑑𝑥
𝑎 1 2 𝑑𝑥
𝑅𝑀𝑆𝐸 = ⇒ 𝑓 𝑥 + 𝑆𝑚 𝑥
𝑏−𝑎 𝑏−𝑎
𝑎

Now
∞

𝑓 𝑥 − 𝑆𝑚 𝑥 = 𝑓 𝑥 − 𝛼𝑛 𝛷𝑛 (𝑥)
𝑛=1
∞
Where 𝛷𝑛 (𝑥) 𝑛=1 is an orthonormal set and "𝛼𝑛 " is unknown. Now,
∞ ∞ 2
2 2
𝑓 𝑥 − 𝑆𝑚 𝑥 = 𝑓(𝑥) −2 𝛼𝑛 𝛷𝑛 𝑥 𝑓 𝑥 + 𝛼𝑛 𝛷𝑛 (𝑥) … (1)
𝑛=1 𝑛=1
∞ 2 ∞ ∞ ∞ ∞

𝛼𝑛 𝛷𝑛 (𝑥) = 𝛼𝑛 𝛷𝑛 𝑥 × 𝛼𝑛 𝛷𝑛 𝑥 = 𝛼𝑛 𝛷𝑛 𝑥 × 𝛼𝑚 𝛷𝑚 𝑥
𝑛=1 𝑛=1 𝑛=1 𝑛=1 𝑚 =1
∞ 2 ∞ ∞

𝛼𝑛 𝛷𝑛 (𝑥) = 𝛼𝑛 𝛼𝑚 𝛷𝑛 𝑥 𝛷𝑚 (𝑥)
𝑛=1 𝑛=1 𝑚 =1
By putting equation (1) becomes,
∞ ∞ ∞
2 2
𝑓 𝑥 − 𝑆𝑚 𝑥 = 𝑓(𝑥) −2 𝛼𝑛 𝛷𝑛 𝑥 𝑓 𝑥 + 𝛼𝑛 𝛼𝑚 𝛷𝑛 𝑥 𝛷𝑚 (𝑥)
𝑛=1 𝑛=1 𝑚 =1
Integrating form "𝑎 𝑡𝑜 𝑏"
𝑏 𝑏 ∞ 𝑏 ∞ ∞ 𝑏
2 2
𝑓 𝑥 − 𝑆𝑚 𝑥 𝑑𝑥 = 𝑓(𝑥) 𝑑𝑥 − 2 𝛼𝑛 𝛷𝑛 𝑥 𝑓 𝑥 𝑑𝑥 + 𝛼𝑛 𝛼 𝑚 𝛷𝑚 𝛷𝑛 𝑑𝑥 … (2)
𝑎 𝑎 𝑛=1 𝑎 𝑛=1 𝑚 =1 𝑎
Page | 22
MTH647 Handout

∴ 𝛷𝑛 𝑥 𝑓 𝑥 𝑑𝑥 = 𝑐𝑛 = 𝐺𝑒𝑛𝑒𝑟𝑎𝑙𝑖𝑧𝑒𝑑 𝐹𝑜𝑢𝑟𝑖𝑒𝑟 𝐶𝑜𝑒𝑓𝑓𝑖𝑐𝑖𝑒𝑛𝑡𝑠

𝑎
So, equation (2) becomes,
𝑏 𝑏 ∞ ∞
2 2
𝑓 𝑥 − 𝑆𝑚 𝑥 𝑑𝑥 = 𝑓(𝑥) 𝑑𝑥 − 2 𝛼𝑛 𝑐𝑛 + (𝛼𝑛 )2
𝑎 𝑎 𝑛=1 𝑛=1
𝑏 𝑏 ∞
2 2
𝑓 𝑥 − 𝑆𝑚 𝑥 𝑑𝑥 = 𝑓(𝑥) 𝑑𝑥 + 𝛼 2 − 2 𝛼𝑛 𝑐𝑛 … (3)
𝑎 𝑎 𝑛=1
2 2 2
𝛼 − 2 𝛼𝑛 𝑐𝑛 =𝛼 − 2 𝛼𝑛 𝑐𝑛 + 𝑐𝑛 − 𝑐𝑛 2 = 𝛼𝑛 − 𝑐𝑛 2
− 𝑐𝑛 2

So, equation (3) becomes,

𝑏 𝑏 ∞
2 2 2 2
𝑓 𝑥 − 𝑆𝑚 𝑥 𝑑𝑥 = 𝑓(𝑥) 𝑑𝑥 + 𝛼𝑛 − 𝑐𝑛 − 𝑐𝑛
𝑎 𝑎 𝑛=1
𝑏 𝑏 ∞ ∞
2 2 2 2
𝑓 𝑥 − 𝑆𝑚 𝑥 𝑑𝑥 = 𝑓(𝑥) 𝑑𝑥 + 𝛼𝑛 − 𝑐𝑛 − 𝑐𝑛
𝑎 𝑎 𝑛=1 𝑛=1
Error will be minimum when
∞
2 2
𝛼𝑛 − 𝑐𝑛 =0 ⇒ 𝛼𝑛 − 𝑐𝑛 = 0 ⇒ 𝛼𝑛 − 𝑐𝑛 ⇒ 𝛼𝑛 = 𝑐𝑛
𝑛=1

Lecture NO. 38
Bissel’s Inequality:
Theorem: For generalized Fourier coefficients “𝑐𝑛 ” corresponding to 𝑓(𝑥), show that
∞ 𝑏

(𝑐𝑛 )2 ≤ {𝑓(𝑥)}2 𝑑𝑥
𝑛=1 𝑎
Proof: Since 𝑅𝑀𝑆𝐸 ≥ 0, so
𝑏 𝑏
𝑓 𝑥 + 𝑆𝑚 𝑥 2
𝑎 2 𝑑𝑥
𝑑𝑥 ≥ 0 ⇒ 𝑓 𝑥 + 𝑆𝑚 𝑥 ≥0
𝑏−𝑎
𝑎
𝑏
2
⇒ 𝑓 𝑥 + 𝑆𝑚 𝑥 𝑑𝑥 ≥ 0
𝑎
𝑏 ∞
2 2
⇒ 𝑓(𝑥) 𝑑𝑥 − 𝑐𝑛 ≥0
𝑎 𝑛=1
∞ 𝑏
2 2
⇒ 𝑐𝑛 ≤ 𝑓(𝑥) 𝑑𝑥
𝑛=1 𝑎
This is our required result. This inequality is known as Bissel’s inequality.

Lecture NO. 39
Limiting Value of Generalized Fourier Coefficients:
Theorem: Show that
𝑏

lim 𝑓 𝑥 𝛷𝑛 𝑥 𝑑𝑥 = 0
𝑛→∞
𝑎
Proof: As we know, Generlized Fourier Coefficients
Page | 23
MTH647 Handout

𝑐𝑛 = 𝑓 𝑥 𝛷𝑛 𝑥 𝑑𝑥
𝑎
And from Bissel’s inequality
∞ 𝑏
2
𝑐𝑛 ≤ 𝑓 𝑥 𝛷𝑛 𝑥 𝑑𝑥
𝑛=1 𝑎
Since
𝑏

𝑓 𝑥 𝛷𝑛 𝑥 𝑑𝑥 < 𝑘 ∀𝑘 ∈𝑅
𝑎
However large it may be.
∞
2
𝑐𝑛 𝑤𝑖𝑙𝑙 𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒 ⇒ 𝑐𝑛 → 0 𝑎𝑠 𝑛 → ∞
𝑛=1
This implies that
𝑏

lim 𝑐𝑛 = 0 ⇒ lim 𝑓 𝑥 𝛷𝑛 𝑥 𝑑𝑥 = 0
𝑛→∞ 𝑛→∞
𝑎
This is our required reslut.
Lecture NO. 40
Sturm-Liouville System:
A boundary value problem of the form:
𝑑 𝑑𝑦
𝑝(𝑥) + 𝑞 𝑥 + 𝜆 𝑟(𝑥) 𝑦 = 0 𝑎 ≤ 𝑥 ≤ 𝑏 … (1)
𝑑𝑥 𝑑𝑥
Subjected to
𝛼1 𝑦 𝑎 + 𝛼2 𝑦 ′ 𝑎 = 0 , 𝛽1 𝑦 𝑏 + 𝛽2 𝑦 ′ 𝑏 = 0 … (2)
Where 𝛼1 , 𝛼2 , 𝛽1 , 𝛽2 are given constants and 𝑝(𝑥), 𝑞(𝑥) 𝑎𝑛𝑑 𝑟(𝑥) are given differentiable functions. And λ
is unspecified parameter independent of 𝑥.
Here system has a squence of eigon value λ𝑛 and corresponding eigen functions satisfying boundary value
problem.
Example: 𝑦 ′′ + 𝜆𝑦 = 0 ; 𝐵. 𝑉. 𝑃: 𝑦 0 = 𝑦 1 = 0
This implies
𝑑 𝑑𝑦
1. + 0 + 𝜆. 1 𝑦 = 0 0≤𝑥≤1
𝑑𝑥 𝑑𝑥
Boundary values can be written as,
𝑦 0 = 1 𝑦 0 + 0 𝑦 ′ 0 = 0 = 𝛼1 𝑦 𝑎 + 𝛼2 𝑦 ′ 𝑎
𝑦 1 = 1 𝑦 1 + 0 𝑦 ′ 1 = 0 = 𝛽1 𝑦 𝑏 + 𝛽2 𝑦 ′ 𝑏
Comparing this with Sturm-Liouville System, we have
𝛼1 = 1 , 𝛼2 = 0 , 𝛽1 = 1 , 𝛽2 = 0 , 𝑝 𝑥 = 1 , 𝑞 𝑥 = 0 , 𝑟 𝑥 = 1

Lecture NO. 41

Heat Equation as Motivation for S-L System:

Consider a B.V.P,
𝜕𝑢 𝜕 𝜕𝑢
𝑔 𝑥 = 𝑘 𝑥 + 𝑕 𝑥 𝑢 … (1)
𝜕𝑡 𝜕𝑥 𝜕𝑥
Where
0 < 𝑥 < 𝑙 , 𝑡 > 0 , 𝑢 0, 𝑡 = 0 , 𝑢 𝑙, 𝑡 = 0 , 𝑢 𝑥, 0 = 𝑓 𝑥 & 𝑢(𝑥, 𝑡) < 𝑀
Let
𝑢 𝑥, 𝑡 = 𝑋 𝑥 . 𝑇 𝑡 = 𝑋𝑇
Page | 24
MTH647 Handout

So, equation (1) can be written as

𝜕
𝑔 𝑥 𝑋𝑇 ′ = 𝑘 𝑥 𝑋𝑇 ′ + 𝑕 𝑥 𝑋𝑇
𝜕𝑥
𝑑 𝑑𝑋
𝑔 𝑥 𝑋𝑇 ′ = 𝑇 𝑘 𝑥 + 𝑕 𝑥 𝑋𝑇
𝑑𝑥 𝑑𝑥
Dividing by 𝑔 𝑥 𝑋𝑇 , we get
𝑇′ 1 𝑑 𝑑𝑋 𝑕 𝑥
= 𝑘 𝑥 + = −𝜆 (𝑠𝑎𝑦)
𝑇 𝑔 𝑥 𝑋 𝑑𝑥 𝑑𝑥 𝑔(𝑥)
𝑇 ′ + 𝜆𝑇 = 0 … (2)
Boundary Values can be written as,
𝑢 0, 𝑡 = 0 ⇒ 𝑋 0 . 𝑇 𝑡 = 0 ⇒ 𝑋 0 = 0
𝑢 𝑙, 𝑡 = 0 ⇒ 𝑋 𝑙 . 𝑇 𝑡 = 0 ⇒ 𝑋 𝑙 = 0
On comparing;
𝑦 = 𝑥 , 𝑝 𝑥 = 𝑘 𝑥 , 𝑞 𝑥 = 𝑕 𝑥 , 𝑟 𝑥 = 𝑔(𝑥)
Solution: From equation (2)
𝑇 ′ + 𝜆𝑇 = 0 ⇒ 𝑇 = 𝐶 𝑒 −𝜆𝑡
By superposition principle
∞

𝑢 𝑥, 𝑡 = 𝐶𝑛 𝑒 −𝜆𝑛𝑡 𝑋𝑛 𝑥 … (3)
𝑛=1
As given
𝑢 𝑥, 0 = 𝑓(𝑥)
∞

𝑓 𝑥 = 𝑢 𝑥, 0 = 𝐶𝑛 𝑋𝑛 (𝑥)
𝑛=1
If 𝐶𝑛 is generalized coefficient, then
𝑙

𝐶𝑛 = 𝑓 𝑥 𝑋𝑛 𝑥 𝑑𝑥
0
Putting 𝐶𝑛 value in equation (3) we have
∞ 𝑙

𝑢 𝑥, 𝑡 = 𝑓 𝑥 𝑋𝑛 𝑥 𝑑𝑥 𝑒 −𝜆𝑛𝑡 𝑋𝑛 𝑥
𝑛=1 0
This is our required result.
Lecture NO. 42

Eigen Values and Eigen Functions of S-L System:

Given a S-L System;
𝑑 𝑑𝑦
𝑝 𝑥 + 𝑞 𝑥 + 𝜆 𝑟(𝑥) 𝑦 = 0
𝑑𝑥 𝑑𝑥
Subjected to;
𝛼1 𝑦 𝑎 + 𝛼2 𝑦 ′ 𝑎 = 0 , 𝛽1 𝑦 𝑏 + 𝛽2 𝑦 ′ 𝑏 = 0 𝑎 ≤ 𝑥 ≤ 𝑏
Here non-trivial solution of S-L System exists in generalized for a particular value of λ. These values are
called the Eigen values and the corresponding non-trivial solutions are called Eigen functions of S-L system.
Example:
𝑦 ′′ + 𝜆𝑦 = 0 ; 𝑦 0 = 𝑦 1 = 0 , 0 ≤ 𝑥 ≤ 1
Let
𝑦 = 𝑒 𝑚𝑥 ⇒ 𝑦 ′′ = 𝑚2 𝑦 ⇒ 𝑚 = ±𝑖 𝜆 ; 𝑦 ≠ 0
By superposition principle;
𝑦 = 𝐴 𝑐𝑜𝑠 𝜆𝑥 + 𝐵 𝑠𝑖𝑛 𝜆𝑥 … (1)
For 𝑦 0 = 𝐴 = 0 . So equation (1) becomes
𝑦 = 𝐵 𝑠𝑖𝑛 𝜆𝑥
Page | 25
MTH647 Handout

And for
𝑦 1 = 𝐵 𝑠𝑖𝑛 𝜆 = 0 ⇒ 𝐵 ≠ 0 & 𝑠𝑜 𝑠𝑖𝑛 𝜆 = 0
This implies,
𝜆 = 𝑚𝜋 ⇒ 𝜆𝑚 = 𝑚2 𝜋 2 ; 𝑚 ∈ 𝑍
These are called Eigen values of given S-L System. And
𝑦𝑚 = 𝐵𝑚 𝑠𝑖𝑛 𝑚𝜋𝑥
These are called the Corresponding Eigen Functions.

Lecture NO. 43

Orthogonality of Eigen Functions:

𝑦𝑚 = 𝐵𝑚 𝑠𝑖𝑛 𝑚𝜋𝑥 ; 0 ≤ 𝑥 ≤ 1
1

𝐵𝑚 𝑠𝑖𝑛 𝑚𝜋𝑥 𝐵𝑛 𝑠𝑖𝑛 𝑛𝜋𝑥 𝑑𝑥

0
1

= 𝐵𝑚 𝐵𝑛 𝑠𝑖𝑛 𝑚𝜋𝑥 𝑠𝑖𝑛 𝑛𝜋𝑥 𝑑𝑥

0
Using relation 2 𝑠𝑖𝑛𝐴 𝑠𝑖𝑛𝐵 = cos 𝐴 − 𝐵 − cos⁡
(𝐴 + 𝐵) , we have
1
𝐵𝑚 𝐵𝑛
= cos 𝑚 − 𝑛 𝜋𝑥 − cos 𝑚 + 𝑛 𝜋𝑥 𝑑𝑥
2
0
By integrating and applying limits, we get
1
𝐵𝑚 𝐵𝑛 sin 𝑚 − 𝑛 𝜋𝑥 sin 𝑚 + 𝑛 𝜋𝑥
= − =0
2 𝑚−𝑛 𝜋 𝑚+𝑛 𝜋 0
∞
Hence, 𝐵𝑚 𝑠𝑖𝑛 𝑚𝜋𝑥 𝑚 =1 is an orthogonal set.

Lecture NO. 44

Normalization of Eigen Functions:

Given set 𝛷𝑚 (𝑥) ∞ 𝑚 =1 is orthonormal if
𝑏
0; 𝑚≠𝑛
𝛷𝑚 𝛷𝑛 𝑑𝑥 = 𝑆𝑚𝑛 =
1; 𝑚=𝑛
𝑎
∞
Now for 𝐵𝑚 𝑠𝑖𝑛 𝑚𝜋𝑥 𝑚 =1 in (0,1), we apply the condition of orthonormality, if
1
2
𝐵𝑚 𝑠𝑖𝑛 𝑚𝜋𝑥 𝑑𝑥 = 1 ; 𝑖𝑓𝑓 𝑚 = 𝑛
0
1 1
1 − 𝑐𝑜𝑠2𝑚𝜋𝑥
(𝐵𝑚 )2 𝑠𝑖𝑛2 𝑚𝜋𝑥 𝑑𝑥 = 1 ⇒ (𝐵𝑚 )2 𝑑𝑥 = 1
2
0 0
1
1 𝑠𝑖𝑛2𝑚𝜋𝑥 1
(𝐵𝑚 )2 𝑥− = 1 ⇒ (𝐵𝑚 )2 1 −0−0+0 =1
2 42𝑚𝜋 0 2
(𝐵𝑚 )2 = 2 ⇒ 𝐵𝑚 = 2 ; ∀ 𝑚 ∈ 𝑁
∞
Hence, 2 𝑠𝑖𝑛 𝑚𝜋𝑥 𝑚 =1
is an orthonormal set.

Lecture NO. 45
Expansion in Terms of Orthonormal Functions:
Let 𝑓 𝑥 = 1 (𝑠𝑎𝑦) and we want to express 𝑓(𝑥) as an expansion of orthonormal functions say;
Page | 26
MTH647 Handout

∞
∞
𝛷𝑛 (𝑥) 𝑛=1 = 2 𝑠𝑖𝑛 𝑛𝜋𝑥 𝑛=1
i.e.
∞

𝑓 𝑥 = 𝐶𝑛 𝛷𝑛 𝑥 … (1)
𝑛=1
Where
𝑙 𝑙 𝑙

𝐶𝑛 = 𝑓 𝑥 𝛷𝑛 𝑥 𝑑𝑥 ⇒ 1. 2 𝑠𝑖𝑛 𝑛𝜋𝑥 𝑑𝑥 ⇒ 2 𝑠𝑖𝑛 𝑛𝜋𝑥 𝑑𝑥

0 0 0
2 1
2 2 2
𝐶𝑛 = − 𝑐𝑜𝑠 𝑛𝜋𝑥 0 ⇒ − 𝑐𝑜𝑠 𝑛𝜋 − 1 ⇒ 1 − −1
𝑛𝜋 𝑛𝜋 𝑛𝜋
So, equation (1) can be written as
∞
2 2
𝑓 𝑥 = 1 − −1 𝑠𝑖𝑛 𝑛𝜋𝑥
𝑛𝜋
𝑛=1
This is our required result.

Lecture NO. 46
Characterizing the Eigen Values of S-L System:
Theorem: show that the Eigen values of a S-L System are real.
Proof: Given a S-L System in (𝑎, 𝑏);
𝑑 𝑑𝑦
𝑝 𝑥 + 𝑞 𝑥 + 𝜆 𝑟(𝑥) 𝑦 = 0
𝑑𝑥 𝑑𝑥
Or it can be written as,
𝑑
𝑦 𝑝 𝑦′ + 𝑞 𝑥 + 𝜆 𝑟(𝑥) 𝑦 = 0 … (1)
𝑑𝑥
Subjected to boundary condition;
𝛼1 𝑦 𝑎 + 𝛼2 𝑦 ′ 𝑎 = 0 … 𝑎 , 𝛽1 𝑦 𝑏 + 𝛽2 𝑦 ′ 𝑏 = 0 … (𝑏)
Where 𝛼1 , 𝛼2 , 𝛽1 , 𝛽2 are given real numbers and 𝑝(𝑥), 𝑞(𝑥) 𝑎𝑛𝑑 𝑟(𝑥) are real valued functions.
Taking conjugate of equation (1),
𝑑
𝑦 𝑝 𝑦 ′ + 𝑞 𝑥 + 𝜆 𝑟(𝑥) 𝑦 = 0 … (2)
𝑑𝑥
Subjected to;
𝛼1 𝑦 𝑎 + 𝛼2 𝑦 ′ 𝑎 = 0 … 𝑐 , 𝛽1 𝑦 𝑏 + 𝛽2 𝑦 ′ 𝑏 = 0 … (𝑑)
By 𝑦 2 − 𝑦 ′ (1);
𝑑 𝑑
𝑦 𝑝 𝑦′ − 𝑦 𝑝 𝑦′ + 𝑞 𝑥 + 𝜆 𝑟 𝑥 − 𝑞 𝑥 − 𝜆 𝑟(𝑥) 𝑦 𝑦 = 0
𝑑𝑥 𝑑𝑥
𝑦 𝑝 𝑦 ′′ + 𝑝′ 𝑦 ′ − 𝑦 𝑝 𝑦 ′′ + 𝑝′ 𝑦′ + 𝜆 − 𝜆 𝑟 𝑥 𝑦 𝑦 = 0
𝑝 𝑦 𝑦 ′′ − 𝑦 𝑦 ′′ + 𝑝′ 𝑦 𝑦 ′ − 𝑦 𝑦 ′ = 𝜆 − 𝜆 𝑟 𝑥 𝑦 𝑦
𝑑
𝑝 𝑦 𝑦 ′ − 𝑦 𝑦 ′ + 𝑝′ 𝑦 𝑦 ′ − 𝑦 𝑦 ′ = 𝜆 − 𝜆 𝑟 𝑥 𝑦 𝑦
𝑑𝑥
𝑑
𝑝 𝑥 𝑦 𝑦′ − 𝑦 𝑦′ = 𝜆 − 𝜆 𝑟 𝑥 𝑦 𝑦
𝑑𝑥
Integrating form "𝑎 𝑡𝑜 𝑏",
𝑏

𝜆−𝜆 𝑟 𝑥 𝑦 𝑦 𝑑𝑥 = 𝑝 𝑥 𝑦 𝑦′ − 𝑦 𝑦′ 𝑥=𝑏
𝑥=𝑎
𝑎
= 𝑝 𝑥 𝑦 𝑏 𝑦′ 𝑏 − 𝑦 𝑏 𝑦 ′ 𝑏 − 𝑦 𝑎 𝑦′ 𝑎 + 𝑦 𝑎 𝑦 ′ (𝑎) … (3)
Since,
𝛼1 𝑦 𝑎 + 𝛼2 𝑦 ′ 𝑎 = 0 , 𝛽1 𝑦 𝑏 + 𝛽2 𝑦 ′ 𝑏 = 0
This implies,
Page | 27
MTH647 Handout

𝑦(𝑎) 𝛼2 𝑦 𝑎 𝛼2
= − , = −
𝑦 ′ (𝑎) 𝛼1 𝑦′ 𝑎 𝛼1
𝑦(𝑎) 𝑦 𝑎
′
= ⇒ 𝑦 𝑎 𝑦′ 𝑎 − 𝑦 𝑎 𝑦′ 𝑎 = 0
𝑦 (𝑎) 𝑦′ 𝑎
Similarly,
𝑦 𝑏 𝑦′ 𝑏 − 𝑦 𝑏 𝑦′ 𝑏 = 0
2
So, equation (3) becomes, ∴ 𝑦𝑦 = 𝑦
𝑏
2
𝜆−𝜆 𝑟 𝑥 𝑦 𝑑𝑥 = 0
𝑎
As,
𝑏
2 2
𝑟 𝑥 ≥ 0 𝑎𝑛𝑑 𝑦 ≥0 ⇒ 𝑟 𝑥 𝑦 𝑑𝑥 ≥ 0
𝑎
So,
𝜆 − 𝜆 = 0 ⇒ 𝜆 = 𝜆 ⇒ 𝜆 𝑖𝑠 𝑟𝑒𝑎𝑙 𝑣𝑎𝑙𝑢𝑒𝑑.
Hence proved.
Lecture NO. 47

The Gamma Function (Definition and Recurrence Relation):

The Gamma function is defined and given by;
∞

𝛾 𝑛+1 = 𝑥 𝑛 𝑒 −𝑥 𝑑𝑥 𝑐𝑜𝑛𝑣𝑒𝑟𝑔𝑒𝑛𝑡 𝑓𝑜𝑟 𝑛 > 0.

0
∞ ∞
∞
= 𝑥 𝑛 . 𝑒 −𝑥 0 − −𝑒 −𝑥 . 𝑛𝑥 𝑛−1 𝑑𝑥 = 0 ⇒ 0 + 𝑛 𝑥 𝑛 −1 . 𝑒 −𝑥 𝑑𝑥
0 0
𝛾 𝑛 + 1 = 𝑛 𝛾 𝑛 ; ∀ 𝑛 > 0 … (1)
If 𝑛 ∈ 𝑁
𝐹𝑜𝑟 𝑛 = 1 ; 𝛾 2 = 1. 𝛾 1 = 1.1 = 1!
𝐹𝑜𝑟 𝑛 = 2 ; 𝛾 3 = 2. 𝛾 2 = 2.1! = 2!
𝐹𝑜𝑟 𝑛 = 3 ; 𝛾 4 = 3. 𝛾 3 = 3.2! = 3!
By induction; 𝛾 𝑛 + 1 = 𝑛!
Example: 𝛾 6 = 5! = 5.4.3.2.1 = 120

Lecture NO. 48

Analytic Continuation of Gamma Function:

The Gamma function is defined and given by;
∞

𝛾 𝑛+1 = 𝑥 𝑛 𝑒 −𝑥 𝑑𝑥 ; 𝑛 > 0.
0
It is technique which is used to extend the domain of a given function beyond its natural domain of
definition.
𝛾 𝛼+1
𝛾 𝛼+1 =𝛼𝛾 𝛼 ⇒ 𝛾 𝛼 = … (𝑎)
𝛼
Here
𝛾 𝛼+2
𝛾 𝛼+1 =
𝛼+1
Hence above (a) relation becomes,
1
𝛾 𝛼 = 𝛾 𝛼+2
𝛼 𝛼+1
Page | 28
MTH647 Handout

Similarly, we can write it as

1
𝛾 𝛼 = 𝛾 𝛼+3
𝛼 𝛼+1 𝛼+2
In general form it can be written as,
1
𝛾 𝛼 = 𝛾 𝛼+𝑘+1
𝛼 𝛼 + 1 𝛼 + 2 … (𝛼 + 𝑘)
𝛼 > 0 ⇒ 𝛾 𝛼 is not defined for 𝛼 = 0, −1, −2, …
Example:
𝛾(−0.5 + 1) 𝛾(0.5) 𝜋
𝛾 −0.5 = = = = −2 𝜋
−0.5 −0.5 −0.5

Lecture NO. 49
An Important Gamma Value:
1
Theorem: Prove that 𝛾 = 𝜋.
2
Proof: Since,
∞ ∞

𝛾 𝑛+1 = 𝑥 𝑛 −1 𝑒 −𝑥 𝑑𝑥 ⇒ 𝑥 −1/2 𝑒 −𝑥 𝑑𝑥
0 0
Let
𝑥 = 𝑢2 ⇒ 𝑑𝑥 = 2𝑢 𝑑𝑢
1 1
𝑥 −2 = 𝑢−1 ⇒ 𝑢−1 =
𝑢
By putting, we get
∞ ∞
1 1 −𝑢 2 2
𝛾 = 𝑒 2𝑢 𝑑𝑢 ⇒ 2 𝑒 −𝑢 𝑑𝑢
2 𝑢
0 0
Taking square on both sides, (∴ 𝑢 → 𝑣)
∞ 2 ∞ ∞
2
1 −𝑢 2 −𝑢 2 2
𝛾 =4 𝑒 𝑑𝑢 ⇒ 2 𝑒 𝑑𝑢 2 𝑒 −𝑣 𝑑𝑣
2
0 0 0
∞ ∞
2
1 𝑢 2 +𝑣 2
𝛾 =4 𝑒− 𝑑𝑢 𝑑𝑣
2
0 0
Let
𝑣 = 𝑟𝑠𝑖𝑛𝜃 ⇒ 𝑢2 + 𝑣 2 = 𝑟 2
𝑢 = 𝑟𝑐𝑜𝑠𝜃 ,
𝑣 𝜋
𝐴𝑠 𝑢, 𝑣 → ∞, 𝑠𝑜 𝜃 = tan−1 ⇒ 𝜃=
𝑢 2
𝐴𝑛𝑑 𝑑𝑢 𝑑𝑣 = 𝑟 𝑑𝑟 𝑑𝜃
By putting, we get
𝜋
𝜃=
2 𝑟→∞
2
1 2
𝛾 =4 𝑒 −𝑟 𝑟 𝑑𝑟 𝑑𝜃
2
𝜃 =0 𝑟=0
𝜋 𝜋
2 ∞ 2 2 ∞
−𝑟 2
𝑒 −𝑟
=4 𝑒 𝑟 𝑑𝑟 𝑑𝜃 ⇒ 4 𝑑𝜃
−2 0
0 0 0
𝜋 𝜋
2 2
𝜋
4 1 2
= − 1 𝑑𝜃 ⇒ 2 𝑑𝜃 ⇒ 2 𝜃 0 ⇒ 𝜋
−2 ∞
0 0
2
1 1
𝑠𝑜, 𝛾 =𝜋 ⇒ 𝛾 = 𝜋
2 2
This is required result.
Page | 29
MTH647 Handout

Lecture NO. 50

Evaluating Some Improper Integrals by Gamma Function:

Question: Evaluate
∞ ∞
2
𝑖 𝑥 𝑚 𝑒 −𝑎𝑥 𝑛 𝑑𝑥 𝑖𝑖 𝑒 −𝛼𝜆 𝑐𝑜𝑠𝛽𝜆 𝑑𝜆
0 0
Solution: (𝑖) let
∞

𝐼= 𝑥 𝑚 𝑒 −𝑎𝑥 𝑛 𝑑𝑥
0
Let
𝑢
𝑎𝑥 𝑛 = 𝑢 ⇒ 𝑥 𝑛 =
𝑎
1
𝑢 𝑛
𝑑𝑢 = 𝑛𝑎𝑥 𝑛−1 𝑑𝑥 ⇒ 𝑥 =
𝑎
𝐴𝑠 𝑥 → 0 ⇒ 𝑢 = 0 , 𝑥→∞ ⇒𝑢=∞
By putting, we get
∞ 1
𝑚 1
𝑢 𝑛 −𝑢
𝑢 𝑛
𝐼= 𝑒 .𝑑
𝑎 𝑎
0
∞ 1−𝑛
1 𝑚 /𝑛 −𝑢
1 𝑢 𝑛 1
𝐼= 𝑚 𝑢 𝑒 . 𝑑𝑢
𝑎 𝑛 𝑛 𝑎 𝑎
0
By solving and simplifying, we have
∞
1 1 𝑚 +1−1 −𝑢 1 𝑚+1
𝐼= 𝑚 1−𝑛 𝑢 𝑛 𝑒 𝑑𝑢 ⇒ 𝑚 +1 𝛾
𝑎𝑛 + 𝑛
+1 𝑛 𝑛𝑎 𝑛 𝑛
0
This is our required result.
𝑖𝑖 : 𝐷𝑜 𝑌𝑜𝑢𝑟 𝑆𝑒𝑙𝑓.

Lecture NO. 51

Inverse Linear Motion:

A particular is attracted toward a fixed point with a force inversely proportional to its instantaneous
distance from fixed point. If the particle is released form rest, then find the time for it to reach fixed point.
Proof: Let at 𝑡 = 0, the particle is a rest i.e. 𝑣 = 0 and be at position 𝑥 = 𝑎 and attracted towards 𝑥 = 0.
Then by given statement and by Newton’s law;
𝑑2 𝑥 𝑘
𝑚 2 = − … (1)
𝑑𝑡 𝑥
Here 𝑚 is the mass of particle and 𝑘 is constant of proportionality.
𝑑2 𝑥 𝑑 𝑑𝑣 𝑑𝑣 𝑑𝑣 𝑑𝑥 𝑑𝑣 𝑘
𝑚 2 =𝑚 ⇒ 𝑚 ⇒ 𝑚 ⇒ 𝑚𝑣 =−
𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑡 𝑑𝑥 𝑑𝑡 𝑑𝑥 2
Separating the variables;
𝑑𝑥
𝑚𝑣 𝑑𝑣 = −𝑘
𝑥
Integrating;
𝑑𝑥 𝑣2
𝑚 𝑣 𝑑𝑣 = −𝑘 ⇒ 𝑚 = −𝑘 𝑙𝑛𝑥 + 𝑐 … (2)
𝑥 2
∴ 𝑣 = 0 𝑎𝑡 𝑥 = 𝑎 ⇒ 0 = −𝑘 𝑙𝑛𝑎 + 𝑐 ⇒ 𝑐 = 𝑘 𝑙𝑛𝑎
By putting, equation (2) becomes,
Page | 30
MTH647 Handout

𝑣2 𝑎 2𝑘 𝑎
𝑚 = 𝑘 𝑙𝑛𝑎 − 𝑘 𝑙𝑛𝑥 = 𝑘 𝑙𝑛 ⇒ 𝑣2 = 𝑙𝑛
2 𝑥 𝑚 𝑥
2𝑘 𝑎 𝑑𝑥 2𝑘 𝑎
𝑣= 𝑙𝑛 ⇒ =− 𝑙𝑛 … (3)
𝑚 𝑥 𝑑𝑡 𝑚 𝑥
Here – 𝑣𝑒 sign is taken because ‘𝑥’ decreases as ‘𝑡’ increases. So,
𝑥=0 𝑡=𝑇 𝑥=0
𝑚 𝑑𝑥 𝑚 𝑑𝑥
=− 𝑑𝑡 ⇒ 𝑇 = … 4
2𝑘 𝑎 2𝑘 𝑎
𝑥=𝑎 𝑙𝑛 𝑡=0 𝑥=𝑎 𝑙𝑛
𝑥 𝑥

Let
𝑎 𝑎
𝑙𝑛 = 𝑢 ⇒ 𝑒 𝑢 = ⇒ 𝑥 = 𝑎𝑒 −𝑢 ⇒ 𝑑𝑥 = −𝑎𝑒 −𝑢 𝑑𝑢
𝑥 𝑥
𝑎
𝑊𝑕𝑒𝑛 𝑥 = 0 ⇒ 𝑢 → ∞ , 𝑥 → 𝑎 ⇒ 𝑙𝑛 = 𝑙𝑛 1 = 0 = 𝑢
𝑎
By putting, we get
∞ ∞
𝑚 −
1 𝑚 1
𝑇= 𝑢 2 – 𝑎𝑒 −𝑢 𝑑𝑢 ⇒ 𝑢−2 𝑒 −𝑢 𝑑𝑢
2𝑘 2𝑘
0 0
𝑚 1 𝑚 𝑚𝜋
𝑇= .𝛾 = 𝜋 ⇒
2𝑘 2 2𝑘 2𝑘
This is our required result.
Lecture NO. 52

Beta Function (Definition and Some Properties):

It is defined and given by;
1

𝐵 𝑚, 𝑛 = 𝑥 𝑚 −1 1 − 𝑥 𝑛−1
𝑑𝑥 ; 𝑚, 𝑛 > 0
0
Let
𝑥 =1−𝑦 ⇒ 𝑦 =1−𝑥
𝑥 =0 ⇒ 𝑦 =1, 𝑥=1 ⇒0
∴ 𝑑𝑥 = −𝑑𝑦
By putting, we get
1 1
𝑚 −1
𝐵 𝑚, 𝑛 = 1−𝑦 𝑦 𝑛−1 – 𝑑𝑦 ⇒ 𝑦 𝑛 −1 1 − 𝑦 𝑚 −1
𝑑𝑦
0 0
1

𝐵 𝑚, 𝑛 = 𝑥 𝑛 −1 1 − 𝑥 𝑚 −1
𝑑𝑥
0
Question: Prove that
𝜋/2

𝐵 𝑚, 𝑛 = 2 𝑠𝑖𝑛2𝑚 −1 𝜃. 𝑐𝑜𝑠 2𝑛−1 𝜃 𝑑𝜃

0
Proof: Since,
𝜋/2

𝐵 𝑚, 𝑛 = 2 𝑥 𝑚 −1 1 − 𝑥 𝑛−1
𝑑𝑥
0
Let
𝑥 = 𝑠𝑖𝑛2 𝜃 ⇒ 𝑑𝑥 = 2𝑠𝑖𝑛𝜃 𝑐𝑜𝑠𝜃
𝜋
𝑥 = 0 ⇒ 𝑠𝑖𝑛2 𝜃 = 0 , 𝑥 = 1 ⇒ 𝑠𝑖𝑛2 𝜃 = 1 ⇒ 𝜃 =
2
By putting, we get
Page | 31
MTH647 Handout

𝜋/2

𝐵 𝑚, 𝑛 = 𝑠𝑖𝑛2 𝜃 𝑚 −1
1 − 𝑠𝑖𝑛2 𝜃 𝑛−1
2𝑠𝑖𝑛𝜃𝑐𝑜𝑠𝜃 𝑑𝜃
0
𝜋/2

𝐵 𝑚, 𝑛 = 2 𝑠𝑖𝑛2 𝜃 𝑚 −1
𝑐𝑜𝑠 2 𝜃 𝑛−1
𝑠𝑖𝑛𝜃𝑐𝑜𝑠𝜃 𝑑𝜃
0
𝜋/2

𝐵 𝑚, 𝑛 = 2 𝑠𝑖𝑛2𝑚 −1 𝜃. 𝑐𝑜𝑠 2𝑛−1 𝜃 𝑑𝜃

0
Hence proved our require result.

Lecture NO. 53

Relation between Beta and Gamma Function:

Prove that
𝛾 𝑚 𝛾 𝑛
𝐵 𝑚, 𝑛 =
𝛾 𝑚+𝑛
Proof: Since
∞

𝛾 𝑚 = 𝑥 𝑚 −1 𝑒 −𝑢 𝑑𝑢 ; 𝑛 > 0.
0
Let 𝑥 2 = 𝑢 ⇒ 𝑑𝑢 = 2𝑥 𝑑𝑥
∞ ∞
−𝑥 2 2
𝛾 𝑚 = 𝑥2 𝑚 −1
𝑒 2𝑥𝑑𝑥 ⇒ 2 𝑥 2𝑚 −1 𝑒 −𝑥 𝑑𝑥
0 0
Similarly,
∞
2
𝛾 𝑛 =2 𝑦 2𝑛 −1 𝑒 −𝑦 𝑑𝑦
0
∞ ∞
𝑥 2 +𝑦 2
𝛾 𝑚 𝛾 𝑛 =4 𝑥 2𝑚 −1 𝑦 2𝑛 −1 𝑒 − 𝑑𝑥 𝑑𝑦
0 0
Let
𝑥 = 𝑟𝑐𝑜𝑠𝜃 , 𝑦 = 𝑟𝑠𝑖𝑛𝜃 , 𝑥 2 + 𝑦 2 = 𝑟 2
𝑦
𝜃 = 𝑡𝑎𝑛−1
𝑥
𝑑𝑥 𝑑𝑦 = 𝑟 𝑑𝑟 𝑑𝜃
By putting, we get
𝜋/2 ∞
2
𝛾 𝑚 𝛾 𝑛 =4 𝑟2 𝑚 +𝑛 −1
𝑒 −𝑟 𝑐𝑜𝑠 2𝑚 −1 𝜃 𝑠𝑖𝑛2𝑛−1 𝜃 𝑑𝜃 𝑑𝑟
0 0
𝜋/2 ∞
2
𝛾 𝑚 𝛾 𝑛 =2 𝑐𝑜𝑠 2𝑚 −1 𝜃 𝑠𝑖𝑛2𝑛−1 𝜃 𝑑𝜃 2 𝑟2 𝑚 +𝑛 −1
𝑒 −𝑟 𝑑𝑟
0 0
𝛾 𝑚 𝛾 𝑛 = 𝐵 𝑚, 𝑛 𝛾 𝑚 + 𝑛
𝛾 𝑚 𝛾 𝑛
𝐵 𝑚, 𝑛 =
𝛾 𝑚+𝑛
For better understanding, WATCH lecture.

Lecture NO. 54
Legendre’s Duplication Formula for Gamma Function:
By evaluating the integrals,
Page | 32
MTH647 Handout

𝜋/2 𝜋/2

𝐼= 𝑠𝑖𝑛2𝑝 𝑥 𝑑𝑥 𝑎𝑛𝑑 𝐽= 𝑠𝑖𝑛2𝑝 (2𝑥) 𝑑𝑥

0 0
Derive the Legendre’s duplication formula for Gamma function.
Proof: As given
𝜋/2

𝐼= 𝑠𝑖𝑛2𝑝 𝑥 𝑐𝑜𝑠𝑥 0
𝑑𝑥 … (1)
0
Comparing with,
𝜋/2
𝛾 𝑚 𝛾 𝑛
𝑠𝑖𝑛2𝑚 −1 𝜃. 𝑐𝑜𝑠 2𝑛−1 𝑥 𝑑𝑥 = … (2)
2𝛾 𝑚 + 𝑛
0
We have,
1 1
2𝑝 = 2𝑚 − 1 ⇒ 𝑚 = 𝑝 + & 2𝑛 − 1 = 0 ⇒ 𝑛 =
2 2
So, equation (2) can be written as,
1 1 1
𝛾 𝑝+ 𝛾 𝜋𝛾 𝑝+
2 2 2
= = 𝐼 … (𝐴)
2𝛾 𝑝 + 1 2𝑝 𝛾 𝑝

Lecture NO. 55
Now for
𝜋
2

𝐽= 𝑠𝑖𝑛2𝑝 2𝑥 𝑑𝑥
0
1
Let 2𝑥 = 𝑦 ⇒ 𝑑𝑥 = 𝑑𝑦 ; 𝑦 = 𝜋 by putting, we get
2
𝜋 𝜋/2 𝜋/2
1 2𝑝
2
= 𝑠𝑖𝑛2𝑦 𝑑𝑦 ⇒ 𝑠𝑖𝑛2𝑝 𝑥 𝑑𝑥 ⇒ 𝑠𝑖𝑛2𝑝 𝑥 𝑑𝑥 = 𝐼
2 2
0 0 0
Again
𝜋 𝜋 𝜋
2 2 2

𝐽= 𝑠𝑖𝑛2𝑝 2𝑥 𝑑𝑥 ⇒ 𝑠𝑖𝑛2𝑥 2𝑝
𝑑𝑥 ⇒ 2𝑠𝑖𝑛𝑥 𝑐𝑜𝑠𝑥 2𝑝
𝑑𝑥
0 0 0
𝜋
2 1 1
22𝑝 𝛾 𝑝 + 𝛾 𝑝+
2 2
𝐽 = 22𝑝 𝑠𝑖𝑛2𝑝 𝑥. 𝑐𝑜𝑠 2𝑝 𝑥 𝑑𝑥 ⇒ … (𝐵)
2𝛾 2𝑝 + 1
0
∴𝐼=𝐽
1 1 1
𝜋𝛾 𝑝+ 22𝑝 𝛾 𝑝 + 𝛾 𝑝+
2 2 2
=
2𝑝 𝛾 𝑝 2𝛾 2𝑝 + 1
By solving we get,
1
𝛾 2𝑝 = 22𝑝−1 𝜋 −1/2
𝛾 𝑝+ 𝛾 𝑝
2
This is our required result.

Lecture NO. 56
The Walli’s Product Formula (By Beta Function):
Prove that

Page | 33
MTH647 Handout

𝜋 𝜋
1.3.5 … 𝑝 − 1 𝜋
2 2
. ; 𝑖𝑓 𝑝 𝑖𝑠 𝑒𝑣𝑒𝑛
2.4.6 … 𝑝 2
𝑠𝑖𝑛𝑝 𝜃 𝑑𝜃 = 𝑝
𝑐𝑜𝑠 𝜃 𝑑𝜃 =
2.4.6 … 𝑝 − 1
0 0 ; 𝑖𝑓 𝑝 𝑖𝑠 𝑜𝑑𝑑
1.3.5 … 𝑝
Solution: Since
𝜋/2
𝛾 𝑚 𝛾 𝑛
𝐵 𝑚, 𝑛 = = 2 𝑠𝑖𝑛2𝑚 −1 𝜃. 𝑐𝑜𝑠 2𝑛−1 𝜃 𝑑𝜃
𝛾 𝑚+𝑛
0
𝜋/2
𝛾 𝑚 𝛾 𝑛
= 𝑠𝑖𝑛2𝑚 −1 𝜃. 𝑐𝑜𝑠 2𝑛−1 𝜃 𝑑𝜃 … (1)
2𝛾 𝑚 + 𝑛
0
1 1
Let 2𝑚 − 1 = 𝑝 𝑎𝑛𝑑 2𝑛 − 1 = 0 ⇒ 𝑚 = 𝑃 + 1 , 𝑛= by putting, we get
2 2
𝜋
2 1 1
𝛾 𝑝+1 𝛾
2 2
𝑠𝑖𝑛𝑝 𝜃 𝑑𝜃 = … 2
1
0 2𝛾 𝑝+2
2
𝑖 : If 𝑝 𝑖𝑠 𝑒𝑣𝑒𝑛 ⇒ 𝑝 = 2𝑙 , 𝑙 ∈ 𝑁, then equation (2) becomes
𝜋
2 1 1
𝛾 𝑙+ 𝛾
2 2
𝑠𝑖𝑛𝑝 𝜃 𝑑𝜃 =
2𝛾 𝑙 + 1
0
1 3 1 1 1
𝑙− 𝑙− … 𝛾 𝛾
2 2 2 2 2
=
2. 𝑙 𝑙 − 1 … 1
1 2
2𝑙 − 1 2𝑙 − 3 … 1 𝛾
2
=
2𝑙 . 2𝑙 𝑙 − 1 … 3.2.1
2
2𝑙 − 1 2𝑙 − 3 … 5.3.1 𝜋
=
2𝑙 2𝑙 − 2 … 6.4.2
1.3.5 … 2𝑙 − 3 2𝑙 − 3 𝜋
=
2.4.6 … 2𝑙 − 2 2𝑙 2
𝑖𝑖 : If 𝑝 𝑖𝑠 𝑜𝑑𝑑 ⇒ 𝑝 = 2𝑘 + 1 , 𝑘 ∈ 𝑁
𝜋
2 1 1 1
𝛾 2𝑘 + 1 + 1 𝛾 𝛾 𝑘+1 𝛾
2 2 2
𝑠𝑖𝑛𝑝 𝜃 𝑑𝜃 = 1
= 2
0
2𝛾 2𝑘 + 3 2𝛾 𝑘+
2 3
𝑘+1
𝑘 𝑘 − 1 𝑘 − 2 … 3.2.1 𝜋 𝑘 𝑘 − 1 … 3.2.1 2
= 1 1 1 1
=
2 𝑘+ 𝑘− … 𝛾 2 2𝑘 + 1 2𝑘 − 1 … 5.3.1
2 2 2 2
= 2𝑘 2𝑘 − 2 … 6.4.2 ⇒ 2.4.6 … 2𝑘 − 2 2𝑘
This is our required result.
Lecture NO. 57
Fourier Integral (Definition and Related Theorem):
Since Fourier series of a function 𝑓(𝑥) in (−𝑙, 𝑙) is defined and given by;
∞
𝑎0 𝑛𝜋𝑥 𝑛𝜋𝑥
+ 𝑎𝑛 𝑐𝑜𝑠 + 𝑏𝑛 𝑠𝑖𝑛
2 𝑙 𝑙
𝑛=1
Where
𝑙 𝑙
1 𝑛𝜋𝑥 1 𝑛𝜋𝑥
𝑎𝑛 = 𝑓(𝑥) 𝑐𝑜𝑠 𝑑𝑥 & 𝑏𝑛 = 𝑓 𝑥 𝑠𝑖𝑛 𝑑𝑥 ; 𝑛 = 1,2,3, …
𝑙 𝜋 𝑙 𝑙
−𝑙 −𝑙
Page | 34
MTH647 Handout

What if 𝑙 → ∞, then in this case, the Fourier series becomes the Fourier Integral.
Fourier Integral Theorem:
Let 𝑓(𝑥) be a function define on (−∞, ∞) and satisfying;
(𝑖): 𝑓(𝑥) 𝑎𝑛𝑑 𝑓 ′ 𝑥 are piecewise continuous in every finite interval.
(𝑖𝑖): 𝑓(𝑥) is absolutely integerable in (−∞, ∞) then
2
𝑓 𝑥 𝑖𝑓 𝑥 𝑖𝑠 𝑎 𝑝𝑜𝑖𝑛𝑡 𝑜𝑓 𝑐𝑜𝑛𝑡𝑖𝑛𝑢𝑖𝑡𝑦
𝐴 𝛼 𝑐𝑜𝑠𝛼𝑥 + 𝐵 𝛼 𝑠𝑖𝑛𝛼𝑥 𝑑𝑥 = 𝑓 𝑥 + 0 + 𝑓 𝑥 − 0
𝑖𝑓 𝑑𝑖𝑠𝑐𝑜𝑛𝑡𝑖.
0 2
The L.H.S side is called Fourier Integral Expansion of 𝑓(𝑥).

Lecture NO. 58

Some Equivalent Forms of Fourier Integral:

∞

𝑓 𝑥 = 𝐴 𝛼 𝑐𝑜𝑠𝛼𝑥 + 𝐵 𝛼 𝑠𝑖𝑛𝛼𝑥 𝑑𝑥 … 𝑝
𝛼=0
∞ +∞
1
𝑓 𝑥 = 𝑓 𝑢 𝑐𝑜𝑠𝛼 𝑥 − 𝑢 𝑑𝑢 𝑑𝛼 … (𝑞)
𝜋
𝛼=0 𝑢=−∞
∞ ∞
1
𝑓 𝑥 = 𝑓 𝑢 𝑒 𝑖𝛼 𝑥−𝑢
𝑑𝑢 𝑑𝛼 … (𝑟)
2𝜋
−∞ −∞
∞ ∞
1
𝑓 𝑥 = 𝑒 𝑖𝛼𝑥 𝑑𝛼 𝑓 𝑢 𝑒 −𝑖𝛼𝑥 𝑑𝑢 … (𝑠)
2𝜋
−∞ −∞
All these forms imply each other.
Question: Prove that (𝑝 ⇄ 𝑞)
∞ ∞ +∞
1
𝐴 𝛼 𝑐𝑜𝑠𝛼𝑥 + 𝐵 𝛼 𝑠𝑖𝑛𝛼𝑥 𝑑𝑥 ⇄ 𝑓 𝑢 𝑐𝑜𝑠𝛼 𝑥 − 𝑢 𝑑𝑢 𝑑𝛼
𝜋
0 𝛼=0 𝑢=−∞
Where
∞ ∞
1 1
𝐴 𝛼 = 𝑓 𝑥 𝑐𝑜𝑠𝛼𝑥 𝑑𝑥 & 𝐵 𝛼 = 𝑓 𝑥 𝑠𝑖𝑛𝛼𝑥 𝑑𝑥
𝜋 𝜋
−∞ −∞
Proof: Since 𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (𝑞)
∞ +∞
1
𝑓 𝑥 = 𝑓 𝑢 𝑐𝑜𝑠𝛼 𝑥 − 𝑢 𝑑𝑢 𝑑𝛼
𝜋
𝛼=0 𝑢=−∞
∞ +∞
1
= 𝑓 𝑢 𝑐𝑜𝑠𝛼𝑥 𝑐𝑜𝑠𝛼𝑢 + 𝑠𝑖𝑛𝛼𝑥 𝑠𝑖𝑛𝛼𝑢 𝑑𝑢 𝑑𝑥
𝜋
𝛼=0 𝑢=−∞
∞ +∞ +∞
1 1
= 𝑓 𝑢 𝑐𝑜𝑠𝛼𝑢 𝑐𝑜𝑠𝛼𝑥 + 𝑓 𝑢 𝑠𝑖𝑛𝛼𝑢 𝑠𝑖𝑛𝛼𝑥
𝜋 𝜋
𝛼=0 𝑢=−∞ 𝑢=−∞
∞

= 𝐴 𝛼 𝑐𝑜𝑠𝛼𝑥 + 𝐵 𝛼 𝑠𝑖𝑛𝛼𝑥 𝑑𝑥
𝛼 =0
Where
∞ ∞
1 1
𝐴 𝛼 = 𝑓 𝑢 𝑐𝑜𝑠𝛼𝑢 𝑑𝑢 & 𝐵 𝛼 = 𝑓 𝑢 𝑠𝑖𝑛𝛼𝑢 𝑑𝑢
𝜋 𝜋
−∞ −∞
(𝑃𝑟𝑜𝑣𝑒𝑑 𝑞 ⇒ 𝑝)
Page | 35
MTH647 Handout

Now 𝐸𝑞𝑢𝑎𝑡𝑖𝑜𝑛 (𝑝)

∞

𝑓 𝑥 = 𝐴 𝛼 𝑐𝑜𝑠𝛼𝑥 + 𝐵 𝛼 𝑠𝑖𝑛𝛼𝑥 𝑑𝑥 … 1
𝛼=0
Where
∞ ∞
1 1
𝐴 𝛼 = 𝑓 𝑢 𝑐𝑜𝑠𝛼𝑢 𝑑𝑢 & 𝐵 𝛼 = 𝑓 𝑢 𝑠𝑖𝑛𝛼𝑢 𝑑𝑢
𝜋 𝜋
−∞ −∞
Putting 𝐴 𝛼 𝑎𝑛𝑑 𝐵 𝛼 values in equation (1),
∞ ∞ ∞
1 1
𝑓 𝑥 = 𝑓 𝑢 𝑐𝑜𝑠𝛼𝑢 𝑑𝑢 𝑐𝑜𝑠𝛼𝑥 + 𝑓 𝑢 𝑠𝑖𝑛𝛼𝑢 𝑑𝑢 𝑠𝑖𝑛𝛼𝑥 𝑑𝛼
𝜋 𝜋
𝛼=0 −∞ −∞
∞ ∞
1
= 𝑓 𝑢 𝑐𝑜𝑠𝛼𝑥 𝑐𝑜𝑠𝛼𝑢 + 𝑠𝑖𝑛𝛼𝑥 𝑠𝑖𝑛𝛼𝑢 𝑑𝑢 𝑑𝑥
𝜋
𝛼=0 −∞
∞ +∞
1
𝑓 𝑥 = 𝑓 𝑢 𝑐𝑜𝑠𝛼 𝑥 − 𝑢 𝑑𝑢 𝑑𝛼
𝜋
𝛼=0 𝑢=−∞
(𝑃𝑟𝑜𝑣𝑒𝑑 𝑝 ⇒ 𝑞)
Hence Proved. (𝑞 𝑖𝑚𝑝𝑙𝑖𝑒𝑠 𝑝 & 𝑝 𝑖𝑚𝑝𝑙𝑖𝑒𝑠 𝑞).

Lecture NO. 59
Fourier Transform and its Inverse (Definition):
For a function 𝑓(𝑥) defined on (−∞, ∞), Fourier Integral is defined and given by;
∞

𝑥 = 𝐴 𝛼 𝑐𝑜𝑠𝛼𝑥 + 𝐵 𝛼 𝑠𝑖𝑛𝛼𝑥 𝑑𝑥
0
Where
∞ ∞
1 1
𝐴 𝛼 = 𝑓 𝑥 𝑐𝑜𝑠𝛼𝑥 𝑑𝑥 & 𝐵 𝛼 = 𝑓 𝑥 𝑠𝑖𝑛𝛼𝑥 𝑑𝑥
𝜋 𝜋
−∞ −∞
And one of its equivalent form is;
∞ ∞
1
𝑓 𝑥 = 𝑒 𝑖𝛼𝑥 𝑑𝛼 𝑓 𝑢 𝑒 −𝑖𝛼𝑥 𝑑𝑢
2𝜋
−∞ −∞
Taking
∞

𝐹 𝛼 = 𝑓 𝑢 𝑒 −𝑖𝛼𝑥 𝑑𝑢 … 𝐾𝑒𝑟𝑛𝑎𝑙
−∞
Then above relation becomes,
∞
1
𝑓 𝑥 = 𝐹 𝛼 𝑒 𝑖𝛼𝑥 𝑑𝛼
2𝜋
−∞
Here 𝐹 𝛼 is called Fourier Transform of 𝑓(𝑥) and 𝑓(𝑥) is called Inverse Fourier Transform of 𝐹 𝛼 .
It can also be expressed as;
∞ ∞
1 1
𝐹 𝑘 = 𝑒 𝑖𝛼𝑥 𝑓 𝑥 𝑑𝛼 & 𝑓 −1 𝐹 𝑘 = 𝑒 −𝑖𝛼𝑥 𝐹 𝑥 𝑑𝛼 = 𝑓(𝑥)
2𝜋 2𝜋
−∞ −∞
1
Note that in both cases/forms/formulation𝑠 "𝑝𝑟𝑜𝑑𝑢𝑐𝑡 𝑜𝑓 𝑐𝑜𝑛𝑠𝑡𝑎𝑛𝑡𝑠 𝑠𝑕𝑜𝑢𝑙𝑑 𝑏𝑒 .”
2𝜋

____________________________________________
“Good Luck For The Mid-Term Exam”

Page | 36
MTH647 Handout

Lecture NO. 60
Fourier Transform of Unit Step Function:
Find the Fourier series of
1 𝑥 <𝑎
𝑓 𝑥 =
0 𝑥 >𝑎
Solution: Fourier Transform of 𝑓(𝑥)
∞

𝐹 𝛼 = 𝑓 𝑢 𝑒 −𝑖𝛼𝑢 𝑑𝑢
−∞
For given problem,
𝑎

𝐹 𝛼 = 1. 𝑒 −𝑖𝛼𝑢 𝑑𝑢 … (1)
−𝑎
−𝑖𝛼𝑢 𝑎
𝑒 𝑒 −𝑖𝛼𝑎 − 𝑒 −𝑖𝛼 −𝑎
𝑒 𝑖𝛼 𝑎 − 𝑒 −𝑖𝛼 𝑎
𝐹 𝛼 = ⇒ ⇒
−𝑖𝛼 −𝑎
−𝑖𝛼 𝑖𝛼
Divide and multiply by 2,
2 𝑒 𝑖𝛼 𝑎 − 𝑒 −𝑖𝛼 𝑎
𝛼 2𝑖
𝑒 𝑖 𝑥 −𝑒 −𝑖 𝑥
Using relation “𝑠𝑖𝑛 𝑥 = ” we have
2𝑖
2
𝐹 𝛼 = 𝑠𝑖𝑛𝛼𝑎
𝛼
It is the required function when 𝛼 ≠ 0.
If we put 𝛼 = 0 in equation (1), we get
𝑎

𝐹 𝛼 = 1 𝑑𝑢 ⇒ 2𝑎
−𝑎

Lecture NO. 61
Fourier Transform of Exponential Functions:
Fourier function of
2
𝑒 −𝑝𝑥 , 𝑝 > 0
Solution: As we know from previous lecture,
∞

𝐹 𝛼 = 𝑓 𝑢 𝑒 −𝑖𝛼𝑢 𝑑𝑢
−∞
For given problem,
∞
2
𝐹 𝛼 = 𝑒 −𝑝𝑢 . 𝑒 −𝑖𝛼𝑢 𝑑𝑢
−∞
∞ ∞
𝑖𝛼 𝑖𝛼 𝑖𝛼 2 𝑖𝛼 2
−𝑝 𝑢 2 − 𝑢 −𝑝 𝑢 2 − 𝑢+ −
𝐹 𝛼 = 𝑒 𝑝 𝑑𝑢 ⇒ 𝑒 𝑝 2𝑝 2𝑝 𝑑𝑢
−∞ −∞
∞ ∞
𝑖𝛼 2 𝛼 2 𝛼2 𝑖𝛼 2
−𝑝 𝑢− + 2 −𝑝 𝑢−
𝐹 𝛼 = 𝑒 2𝑝 4𝑝 𝑑𝑢 ⇒ 𝑒 4𝑝 2 𝑒 2𝑝 𝑑𝑢
−∞ −∞
Let
𝑖𝛼
𝑝 𝑢− =𝑦
2𝑝
1
𝑑𝑦 = 𝑝 𝑑𝑢 ⇒ 𝑑𝑢 = 𝑑𝑦
𝑝
By putting in above relation, we have
Page | 37
MTH647 Handout

𝛼2 ∞ 𝛼2
𝑒 4𝑝 2
−𝑦 2
𝑒 4𝑝 2 𝜋 𝛼 22
𝐹 𝛼 = 𝑒 𝑑𝑦 ⇒ 𝜋 ⇒ 𝑒 4𝑝
𝑝 𝑝 𝑝
−∞
This is our required result.

Lecture NO. 62
Conjugate of a Fourier Series:
Theorem: For a real valued function 𝑓(𝑥)on (−∞, ∞), show that 𝐹 𝛼 = 𝐹(−𝛼)
Proof: As we know
∞

𝐹 𝛼 = 𝑓 𝑢 𝑒 −𝑖𝛼𝑢 𝑑𝑢
−∞
Taking conjugate, we have
∞ ∞ ∞

𝐹 𝛼 = 𝑓 𝑢 𝑒 −𝑖𝛼𝑢 𝑑𝑢 ⇒ 𝑓 𝑢 (𝑒 −𝑖𝛼𝑢 ) 𝑑𝑢 ⇒ 𝑓 𝑢 𝑒 𝑖𝛼𝑢 𝑑𝑢

−∞ −∞ −∞
∞ ∞

𝐹 𝛼 = 𝑓 𝑢 𝑒 −(−𝑖𝛼𝑢 ) 𝑑𝑢 ⇒ 𝑓 𝑢 𝑒 −𝑖(−𝛼)𝑢 𝑑𝑢 ⇒ 𝐹(−𝛼)

−∞ −∞
Hence proved that
𝐹 𝛼 = 𝐹(−𝛼)

Lecture NO. 63
Fourier Transforms of Even and Odd Functions:
Let 𝑓(𝑥) be an odd function
∞

𝐹 𝑓 𝑥 = 𝑓 𝑢 𝑒 −𝑖𝛼𝑢 𝑑𝑢 = 𝐹(𝛼)
−∞
Here put 𝑢 = −𝑡 ⇒ 𝑑𝑢 = −𝑑𝑡
As 𝑢 → ∞ , 𝑡 → ∞ 𝑎𝑛𝑑 𝑢 → −∞ ⇒ 𝑡 → ∞
∞ ∞ ∞

𝐹 𝛼 =− 𝑓 −𝑡 𝑒 −𝑖𝛼 (−𝑡) 𝑑𝑡 ⇒ − 𝑓 𝑡 𝑒 −𝑖𝛼 −𝑡

𝑑𝑡 = − 𝑓 𝑡 𝑒 −𝑖 −𝛼 𝑡
𝑑𝑡
−∞ −∞ −∞
Hence proved that
𝐹 𝛼 = −𝐹 −𝛼 = −𝐹(𝛼)

Lecture NO. 64
Attenuation Property of Fourier Transforms:
Theorem:
𝐹 𝑓 𝑥 𝑒 −𝑝𝑥 = 𝐹 𝛼 − 𝑝𝑖 𝑤𝑕𝑒𝑟𝑒 𝐹 𝛼 = 𝐹{𝑓(𝑥)}
Proof: Aswe know
∞

𝐹 𝑓 𝑥 =𝐹 𝛼 = 𝑓 𝑢 𝑒 −𝑖𝛼𝑢 𝑑𝑢
−∞
For given function,
∞

𝐹 𝑓 𝑥 𝑒 −𝑝𝑥 = 𝑓 𝑢 𝑒 −𝑝𝑢 𝑒 −𝑖𝛼𝑢 𝑑𝑢

−∞
∞ ∞
2 𝑝𝑢 2 𝑝𝑢 −𝑖𝛼𝑢
= 𝑓 𝑢 𝑒𝑖 𝑒 −𝑖𝛼𝑢 𝑑𝑢 ⇒ 𝑓 𝑢 𝑒𝑖 𝑑𝑢
−∞ −∞

Page | 38
MTH647 Handout

𝐹 𝑓 𝑥 𝑒 −𝑝𝑥 = 𝑓 𝑢 𝑒 −𝑖 𝛼 −𝑖𝑝 𝑢
𝑑𝑢 = 𝐹 𝛼 − 𝑖𝑝
−∞
Hence proved that
𝐹 𝑓 𝑥 𝑒 −𝑝𝑥 = 𝐹 𝛼 − 𝑝𝑖

Lecture NO. 65
Shifting Property of Fourier Transform:
Theorem:
𝐼𝑓 𝐹 𝑓 𝑥 = 𝐹 𝛼 𝑡𝑕𝑒𝑛 𝐹 𝑓 𝑥−𝑝 = 𝑒 −𝑖𝛼𝑝 𝐹 𝛼
Proof: As we know
∞

𝐹 𝛼 =𝐹 𝑓 𝑥 = 𝑓 𝑢 𝑒 −𝑖𝛼𝑢 𝑑𝑢
−∞
For given function,
∞

𝐹 𝑓 𝑥−𝑝 = 𝑓 𝑢 − 𝑝 𝑒 −𝑖𝛼𝑢 𝑑𝑢
−∞
Put 𝑢 − 𝑝 = 𝑡 ⇒ 𝑢 = 𝑝 + 𝑡 𝑎𝑛𝑑 𝑑𝑡 = 𝑑𝑢
∞ ∞

𝐹 𝑓 𝑥−𝑝 = 𝑓 𝑡 𝑒 −𝑖𝛼 (𝑝+𝑡) 𝑑𝑡 ⇒ 𝑓 𝑡 𝑒 −𝑖𝛼𝑝 𝑒 −𝑖𝛼𝑡 𝑑𝑢

−∞ −∞
∞

= 𝑒 −𝑖𝛼𝑝 𝑓 𝑡 𝑒 −𝑖𝛼𝑡 𝑑𝑢 ⇒ 𝑒 −𝑖𝛼𝑝 𝐹 𝑓 𝑥 = 𝑒 −𝑖𝛼𝑝 𝐹 𝛼

−∞
This is our required result.

Lecture NO. 66
Fourier Transform of Derivatives:
Theorem: For a function 𝑓(𝑥)on −∞, ∞ , show that if 𝑓(𝑥) is n-times differentiable and 𝑓 𝑛 −1 𝑥 → 0
as 𝑥 → ±∞, then
𝐹 𝑓 𝑛 𝑥 = 𝑖𝛼 𝑛 𝐹 𝑢
Proof:
∞
𝑑𝑛 𝑦 𝑑 𝑛 𝑦 −𝑖𝛼𝑢
𝐹 = 𝑒 𝑑𝑢
𝑑𝑢𝑛 𝑑𝑢𝑛
−∞
By integrating,
∞ ∞
−𝑖𝛼𝑢
𝑑 𝑛 −1 𝑦 𝑑 𝑛−1 𝑦 −𝑖𝛼𝑢
=𝑒 + 𝑖𝛼 𝑒 𝑑𝑢
𝑑𝑢𝑛−1 −∞
𝑑𝑢𝑛−1
−∞
∞ ∞
1 −𝑖𝛼𝑢
𝑑 𝑛−2 𝑦 𝑑 𝑛−2 𝑦 −𝑖𝛼𝑢
= 0 + 𝑖𝛼 𝑒 + 𝑖𝛼 𝑒 𝑑𝑢
𝑑𝑢𝑛−2 −∞
𝑑𝑢𝑛−2
−∞
𝑛 −2
1
𝑑 𝑦 2
𝑑 𝑛−2 𝑦
= 𝑖𝛼 0 + 𝑖𝛼 𝐹 ⇒ 𝑖𝛼 𝐹
𝑑𝑢𝑛−2 𝑑𝑢𝑛−2
Now by induction,
𝑑𝑛 𝑦
𝐹 = 𝑖𝛼 𝑛 𝐹 𝑓 𝑥 = 𝑖𝛼 𝑛
𝐹 𝑢
𝑑𝑢𝑛
Hence proved that
𝑛 𝑛
𝐹 𝑓 𝑥 = 𝑖𝛼 𝐹 𝑢

“Dedicated To My Unknown Students Which Are Our Future Heroes.”

Regards: Virtual Alerts

Page | 39
MTH647 Handout

Lecture NO. 67
Parseval’s Theorems:
Theorem: If 𝑓(𝑥) and 𝑔(𝑥) are real valued on (−∞, ∞), then show
∞ ∞

𝐼) 𝐹 𝛼 𝐺 −𝛼 𝑑𝛼 = 𝑓 𝑢 𝑔 𝑢 𝑑𝑢
−∞ −∞
∞ ∞
2 2
𝐼𝐼) 𝑓 𝑢 𝑑𝑢 = 𝐹 𝛼 𝑑𝛼
−∞ −∞
Proof: I)
∞ ∞ ∞

𝐹 𝛼 𝐺 −𝛼 𝑑𝛼 = 𝐹 𝛼 𝑔 𝑢 𝑒 −𝑖(−𝛼)𝑢 𝑑𝑢
−∞ −∞ −∞
∞ ∞ ∞

= 𝐹 𝛼 𝑒 𝑖𝛼𝑢 𝑔 𝑢 𝑑𝑢 = 𝑓 𝑢 𝑔 𝑢 𝑑𝑢
−∞ −∞ −∞
II): Putting 𝑓 = 𝑔
∞ ∞ ∞ ∞

𝐹 𝛼 𝐺 −𝛼 𝑑𝛼 = 𝐹 𝛼 𝐺 𝛼 𝑑𝛼 = 𝐹 𝛼 𝐺 𝛼 𝑑𝛼 = 𝑓 𝑢 𝑔 𝑢 𝑑𝑢
−∞ −∞ −∞ −∞
∞ ∞ ∞ ∞
2 2 2 2
⇒ 𝐺 𝛼 𝑑𝛼 = 𝑔 𝑢 𝑑𝑢 𝑂𝑟 𝑓 𝑢 𝑑𝑢 = 𝐹 𝛼 𝑑𝛼
−∞ −∞ −∞ −∞
Required result proved.

Lecture NO. 68
Convolution (Definition and Related Theorem):
𝐹 𝑓 =𝐹 𝛼 , 𝐺 𝑔 =𝐺 𝛼
𝐹 𝑓. 𝑔 ≠𝐹 𝛼 𝐺 𝛼
𝐹 𝑓∗𝑔 =𝐹 𝛼 𝐺 𝛼
Convolution: of function 𝑓 𝑥 𝑎𝑛𝑑 𝑔(𝑥) is defined and given,
∞

𝑓∗𝑔= 𝑓 𝑢 𝑔 𝑥 − 𝑢 𝑑𝑢
−∞
Now we have to prove some theorems.
1) Commutative 𝑓 ∗ 𝑔 = 𝑔 ∗ 𝑓 : By definition
∞

𝑓∗𝑔= 𝑓 𝑢 𝑔 𝑥 − 𝑢 𝑑𝑢
−∞
𝑢 → ∞ ⇒ 𝑡 → ±∞
Put 𝑥 − 𝑢 = 𝑡 , 𝑢 = 𝑥 − 𝑡 ⇒ 𝑑𝑢 = −𝑑𝑡 𝑎𝑛𝑑 𝑎𝑠
𝑢 → −∞ ⇒ 𝑡 → ∞
So, we get
−∞ ∞

𝑓∗𝑔 = 𝑓 𝑥 − 𝑡 𝑔 𝑡 −𝑑𝑡 ⇒ 𝑔 𝑡 𝑓 𝑥 − 𝑡 𝑑𝑡 = 𝑔 ∗ 𝑓
+∞ −∞
Hence proved
𝑓∗𝑔 =𝑔∗𝑓
2) 𝐹 𝑓 ∗ 𝑔 = 𝐹 𝛼 𝐺 𝛼 𝑂𝑟 𝐹 −1 𝐹 𝛼 𝐺 𝛼 =𝑓∗𝑔
By definition,
∞ ∞
1 1
𝐹 −1 𝐹 𝛼 𝐺 𝛼 = 𝐹 𝛼 𝐺 𝛼 𝑒 𝑖𝛼𝑥 𝑑𝑥 ⇒ 𝐹 𝛼 𝑒 𝑖𝛼𝑥 𝑑𝑥 . 𝐺 𝛼
2𝜋 2𝜋
−∞ −∞
Page | 40
MTH647 Handout

As we know (from previous lectures)

∞

𝐺 𝛼 = 𝑔 𝑡 𝑒 −𝑖𝛼𝑡 𝑑𝑡
−∞
By putting, we have
∞ ∞
1
𝐹 −1 𝐹 𝛼 𝐺 𝛼 = 𝐹 𝛼 𝑒 𝑖𝛼𝑥 𝑑𝑥 𝑔 𝑡 𝑒 −𝑖𝛼𝑡 𝑑𝑡
2𝜋
−∞ −∞
∞ ∞
1
𝐹 −1 𝐹 𝛼 𝐺 𝛼 = 𝑔 𝑡 𝑑𝑡 𝐹 𝛼 𝑒 𝑖𝛼𝑥 𝑒 −𝑖𝛼𝑡 𝑑𝑡
2𝜋
−∞ −∞
∞ ∞
1
𝐹 −1 𝐹 𝛼 𝐺 𝛼 = 𝑔 𝑡 𝑑𝑡 𝐹 𝛼 𝑒 𝑖𝛼 𝑥−𝑡
𝑑𝑡
2𝜋
−∞ −∞
By using inverse Fourier transformation,
∞
1
𝐹 −1 𝐹 𝛼 𝐺 𝛼 = 𝑔 𝑡 𝑓 𝑥 − 𝑡 𝑑𝑡 ⇒ 𝑔 ∗ 𝑓
2𝜋
−∞
As "𝑔 ∗ 𝑓 = 𝑓 ∗ 𝑔" so
𝐹 −1 𝐹 𝛼 𝐺 𝛼 =𝑓∗𝑔
Applying "𝐹" on both sides, we get
𝐹 𝑓∗𝑔 =𝐹 𝛼 𝐺 𝛼
Examples:
2 1 𝑥 ≤𝑝
𝑖): 𝑓 𝑥 = 𝑒 −𝑥 ; 𝑔 𝑥 =
0 𝑥 ≥𝑝
2 2
𝑖𝑖) 𝑓 𝑥 = 𝑒 −𝛼𝑥 ; 𝑔 𝑥 = 𝑒 −𝛽𝑥 ; 𝛼, 𝛽 > 0
Do Your Self.

Lecture NO. 69
Solving Integral Equation by Convolution:
Solve the integral equation,
∞

𝑦 𝑥 =𝑔 𝑥 + 𝑦 𝑢 𝑟 𝑥 − 𝑢 𝑑𝑢
−∞
Solution: For the given integral equation, taking fourier transformation on both sides,
∞

𝐹 𝑦 𝑥 =𝐹 𝑔 𝑥 +𝐹 𝑦 𝑢 𝑟 𝑥 − 𝑢 𝑑𝑢
−∞
Convolution of last function,
∞

𝑦 𝑢 𝑟 𝑥 − 𝑢 𝑑𝑢 = 𝑦(𝑥) ∗ 𝑟 𝑥
−∞
So, above relation can be writtena as
𝐹 𝑦 𝑥 =𝐹 𝑔 𝑥 + 𝐹 𝑦 𝑥 ∗ 𝑟(𝑥) … (1)
𝐹 𝑦 𝑥 = 𝑌(𝛼)
Let 𝐹 𝑔 𝑥 =𝐺 𝛼 By putting equation (1) becomes
𝐹 𝑟 𝑥 =𝑅 𝛼
𝑌 𝛼 = 𝐺 𝛼 + 𝐹 𝑦 𝑥 ∗ 𝑟(𝑥) ⇒ 𝐺 𝛼 + 𝑌(𝛼)𝑅 𝛼
By solving we have
𝐺 𝛼
𝑌 𝛼 =
1−𝑅 𝛼
Taking inverse fourier transformation,
Page | 41
MTH647 Handout

∞
𝐺 𝛼 𝐺 𝛼
𝐹 −1 𝑌 𝛼 = 𝐹 −1 ⇒ 𝑒 𝑖𝛼𝑥 𝑑𝑥
1−𝑅 𝛼 1−𝑅 𝛼
−∞
1
𝑆𝑖𝑛𝑐𝑒 𝐹 → 1 𝑡𝑕𝑒𝑛 𝐹 −1 → 𝑎𝑛𝑑 𝑕𝑒𝑟𝑒 𝐹 −1 𝑌 𝛼 =𝑦 𝑥
2𝜋
So the above equation becomes,
∞
1 𝐺 𝛼
𝑦 𝑥 = 𝑒 𝑖𝛼𝑥 𝑑𝑥
2𝜋 1−𝑅 𝛼
−∞
This is our required result.

Lecture NO. 70
Fourier Sine and Cosing Tranforms:
𝑰): If 𝑓(𝑥) is an odd function, then Fourier 𝑠𝑖𝑛𝑒 transform is defined and given by
∞

𝐹𝑠 𝛼 = 𝑓 𝑢 𝑠𝑖𝑛𝛼𝑢 𝑑𝑢
0
And its inverse Fourier 𝑠𝑖𝑛𝑒 transfrom is given by
∞
2
𝑓 𝑥 = 𝐹𝑠 𝛼 𝑠𝑖𝑛𝛼𝑢 𝑑𝛼
𝜋
0
𝑰𝑰): If 𝑓(𝑥) is an even function, then fourier 𝑐𝑜𝑠𝑖𝑛𝑒 transform of 𝑓(𝑥) is defined and given by
∞

𝐹𝑐 𝛼 = 𝑓 𝑢 𝑐𝑜𝑠𝛼𝑢 𝑑𝑢
0
And its inverse Fourier 𝑐𝑜𝑠𝑖𝑛𝑒 transfrom is given by
∞
2
𝑓 𝑥 = 𝐹𝑐 𝛼 𝑐𝑜𝑠𝛼𝑢 𝑑𝛼
𝜋
0
Example:
𝐼𝑓; 𝑓 𝑥 = 𝑒 −𝑚𝑥 , 𝑚 > 0 𝑡𝑕𝑒𝑛 𝐹𝑐 𝛼 =?
As we know
∞

𝐹𝑐 𝛼 = 𝑓 𝑢 𝑐𝑜𝑠𝛼𝑢 𝑑𝑢
0
For given function, it can be written as
∞

𝐹𝑐 𝛼 = 𝑒 −𝑚𝑢 𝑐𝑜𝑠𝛼𝑢 𝑑𝑢 … (1)

0
Using integral formula, which is given as
𝑝𝑥
𝑒 𝑝𝑥
𝑒 𝑐𝑜𝑠𝑞𝑥 𝑑𝑥 = 2 𝑝 𝑐𝑜𝑠𝑞𝑥 + 𝑞 𝑠𝑖𝑛𝑞𝑥
𝑝 + 𝑞2
Here 𝑝 → −𝑚 , 𝑞 → 𝛼 𝑎𝑛𝑑 𝑥 → 𝑢
So, by applying integral formula, equation (1) becomes
∞
𝑒 −𝑚𝑢
𝐹𝑐 𝛼 = −𝑚 𝑐𝑜𝑠𝛼𝑢 + 𝛼 𝑠𝑖𝑛𝛼𝑢
𝑚2 + 𝛼 2 0
By applying limits, we get
𝑚
𝐹𝑐 𝛼 = 2 2
= 𝐹𝑐 𝑒 −𝑚𝑥
𝑚 +𝛼
Now its inverse, 𝑐𝑜𝑠𝑖𝑛𝑒 transform;
∞
2
𝑓 𝑥 = 𝐹𝑐 𝛼 𝑐𝑜𝑠𝛼𝑢 𝑑𝛼
𝜋
0
Page | 42
MTH647 Handout

For given function, it can be written as (Putting 𝐹𝑐 𝛼 calculated value)

∞
2 𝑚
𝑓 𝑥 = 𝑐𝑜𝑠𝛼𝑢 𝑑𝛼
𝜋 𝑚2 + 𝛼 2
0
As 𝑓 𝑥 = 𝑒 −𝑚𝑥 ,
∞
2𝑚 𝑐𝑜𝑠𝛼𝑢
𝑒 −𝑚𝑥 = 𝑑𝛼
𝜋 𝑚2 + 𝛼 2
0
∞
−𝑚𝑥
𝜋𝑒 𝑐𝑜𝑠𝛼𝑢
= 𝑑𝛼
2𝑚 𝑚2 + 𝛼 2
0
This is our required result.

Lecture NO. 71
Integral Equations Solution by Fourier Sine Transform:
Solve the integral equation
∞
1−𝛼 0≤𝛼≤1
𝑓 𝑥 𝑠𝑖𝑛𝛼𝑥 𝑑𝑥 =
0 𝛼>1
0
Solution: The given function can be writen as
∞
1−𝛼 0≤𝛼≤1
𝐹𝑠 𝑓 𝛼 = 𝐹𝑠 𝛼 = 𝑓 𝑥 𝑠𝑖𝑛𝛼𝑥 𝑑𝑥 =
0 𝛼>1
0
Its 𝐹 −1 for given function,
∞
2
𝐹𝑠−1 𝑓 𝑥 =𝑓 𝑥 = 𝐹𝑠 𝛼 𝑠𝑖𝑛𝛼𝑥 𝑑𝛼
𝜋
0
1
2
𝑓 𝑥 = 1 − 𝛼 𝑠𝑖𝑛𝛼𝑥 𝑑𝛼
𝜋
0
Integrating and applying limits, we have
1
2 𝑐𝑜𝑠𝛼𝑥 1 𝑐𝑜𝑠𝛼𝑥
= 1−𝛼 − + −1 𝑑𝛼
𝜋 𝑥 0 𝑥
0
1
2 1 1 𝑠𝑖𝑛𝛼𝑥 2 1 𝑠𝑖𝑛𝑥
= 0− 1 − − ⇒ − 2
𝜋 𝑥 𝑥 𝑥 0 𝜋 𝑥 𝑥
2 𝑥 − 𝑠𝑖𝑛𝑥
𝑓 𝑥 =
𝜋 𝑥2
This is our required result.

Lecture NO. 72
Laplace Transforms:
Objectives/Aim/Steps:

Page | 43
MTH647 Handout

Usefulness:
Deals with discontinuos functions (like electrical signals, eechanical forces etc) and sine, cosines forms (like
Fourier analysis).
It is Direct method to solve differential equations.
Applications:
Physics, Engineering Problems → Transform into → Differential Equations (Like ODEs, PDEs, BVPs , etc.)

Lecture NO. 73
Integral Transform:
For given function 𝑓(𝑡) defined on 0, ∞ 𝑜𝑟 𝑡 ≥ 0, if improper function
∞

𝐾 𝑠, 𝑡 𝑓 𝑡 𝑑𝑡
0
Is convergent, then it is called an integral transformation of 𝑓 𝑡 where 𝑠 ∈ 𝐶. And if choosen
∞

𝐾 𝑠, 𝑡 = 𝑒 −𝑠𝑡 𝑡𝑕𝑒𝑛 𝐹 𝑠 =𝐿 𝑓 𝑡 = 𝑒 −𝑠𝑡 𝑓 𝑡 𝑑𝑡 , 𝑠 ∈ 𝐶.

0
Is said to be Laplace transform of 𝑓 𝑡 provided that it is convergent. Here 𝐾 𝑠, 𝑡 = 𝑒 −𝑠𝑡 is known as
Kernal.
Q: Why 𝐾 𝑠, 𝑡 = 𝑒 −𝑠𝑡 ? Why we choose/use Kernal?
Ans: 1) Kernal smooth out the given functions (Like sharp egdges, corners etc.)
2) Easy to integrate or differetinate.
3) By kernal, differntial equation transform into algebric equation. So that it can be easily solved.

Lecture NO. 74
Geometical Interpretition of Laplace Transform:
Since we know
∞

𝐿 𝑓 𝑡 =𝐹 𝑠 = 𝑒 −𝑠𝑡 𝑓 𝑡 𝑑𝑡 , 𝑡 ≥ 0 , 𝑠 ∈ 𝐶, 𝑅.
0
Trnasformation of 𝑓 𝑡 𝑖𝑛𝑡𝑜 𝑓(𝑠):

𝒇 𝒕 Transform Throug Laplace 𝑭 𝒔

Time Frequency
Domain Domain

𝐹(𝑠) gives us information about the different frequency components that makes the 𝑓(𝑡).
Laplace relations:
𝐿 𝑓 𝑡 =𝐹 𝑠
𝑓 𝑡 = 𝐿−1 𝐹 𝑠
𝐿−1 𝐿 𝑓 𝑡 =𝑓 𝑡
𝐿 𝐿−1 𝑓 𝑡 = 𝐹(𝑠)

Lecture NO. 75
Laplace Transform of “1”:
Since we know
∞

𝐿 𝑓 𝑡 =𝐹 𝑠 = 𝑒 −𝑠𝑡 𝑓 𝑡 𝑑𝑡
0
For given function,
Page | 44
MTH647 Handout

∞ 𝑇

𝐿 1 = 1. 𝑒 −𝑠𝑡 𝑑𝑡 ⇒ lim 𝑒 −𝑠𝑡 𝑑𝑡

𝑇→∞
0 0
Integrating and applying limits, we have
𝑇 𝑇
𝑒 −𝑠𝑡 1 1 1 𝑇
𝐿 1 = lim lim 𝑒 −𝑠𝑡 ⇒ − lim⁡ 𝑠𝑡
⇒ − … 1
𝑇→∞ −𝑠 𝑠 𝑇→∞ 0 𝑠 𝑇→∞ 𝑒 0
0
1 1 1 1 1 1
𝐿 1 = − lim 𝑠𝑇 − 𝑠 0 ⇒ − −
𝑠 𝑇→∞ 𝑒 𝑒 𝑠 𝑒∞ 1
1
𝐿 1 =
𝑠
This is our required result for 𝑠 > 0. What if 𝑠 < 0?
𝐼𝑓; 𝑠 < 0 ⇒ −𝑠 > 0 ⇒ −𝑠 = 𝑚 (𝑠𝑎𝑦) ∈ 𝑅+
If so, then equation (1) becomes
1
lim⁡ 𝑠𝑡 = lim 𝑒 −𝑠𝑡 = lim 𝑒 𝑚𝑡 ⇒ 𝑒 ∞ = ∞
𝑇→∞ 𝑒 𝑇→∞ 𝑇→∞
So, whenever 𝑠 < 0, we cannot calculate the transform of “1”. Hence, we can take only 𝑠 > 0.

Lecture NO. 76
Laplace Transform of Exponential Function:
Let 𝑓 𝑡 = 𝑒 𝑎𝑡 . Here “a” is a constant, then evaluate 𝐿 𝑒 𝑎𝑡
Solution: Laplace transform for given function,
∞ ∞

𝐿 𝑒 𝑎𝑡 = 𝑒 −𝑠𝑡 𝑒 𝑎𝑡 𝑑𝑡 ⇒ 𝑒 −(𝑠−𝑎)𝑡 𝑑𝑡
0 0
Integrating and applying limits, we have
∞
𝑎𝑡
𝑒 − 𝑠−𝑎 𝑡 −1 𝑡 ∞
𝐿 𝑒 = ⇒ 𝑒 − 𝑠−𝑎
− 𝑠−𝑎 0
𝑠−𝑎 0
∞
𝑎𝑡
−1 1 −1 1 1
𝐿 𝑒 = 𝑠−𝑎 𝑡
= ∞
− 0
𝑠−𝑎 𝑒 0 𝑠−𝑎 𝑒 𝑒
1
𝐿 𝑒 𝑎𝑡 =
𝑠−𝑎
This is our required result. (𝐹𝑜𝑟 𝑠 − 𝑎 > 0 ⇒ 𝑠 > 𝑎)

Lecture NO. 77
Linearity of Laplace Transform:
Theorem: If 𝐿 𝑓 𝑡 and 𝐿 𝑔 𝑡 exist, then for any constants "𝛼" and "𝛽", transform of
𝛼𝐿 𝑓 𝑡 +𝛽𝐿 𝑔 𝑡
Also exist and further
𝐿𝛼 𝑓 𝑡 +𝛽 𝑔 𝑡 =𝛼𝐿 𝑓 𝑡 +𝛽𝐿 𝑔 𝑡
Also holds.
Proof: Applying Laplace transform on L.H.S.
=𝐿 𝛼 𝑓 𝑡 +𝛽 𝑔 𝑡
∞ ∞

= 𝑒 −𝑠𝑡 𝛼 𝑓 𝑡 +𝛽 𝑔 𝑡 𝑑𝑡 ⇒ 𝛼 𝑒 −𝑠𝑡 𝑓 𝑡 + +𝛽 𝑒 −𝑠𝑡 𝑔 𝑡 𝑑𝑡

0 0
∴ Integration is a linear transformation on 𝑅.
∞ ∞ ∞ ∞

= 𝛼 𝑒 −𝑠𝑡 𝑓 𝑡 𝑑𝑡 + 𝛽 𝑒 −𝑠𝑡 𝑔 𝑡 𝑑𝑡 ⇒ 𝛼 𝑒 −𝑠𝑡 𝑓 𝑡 𝑑𝑡 + 𝛽 𝑒 −𝑠𝑡 𝑔 𝑡 𝑑𝑡

0 0 0 0
Page | 45
MTH647 Handout

=𝛼𝐿 𝑓 𝑡 +𝛽𝐿 𝑔 𝑡 = 𝑅. 𝐻. 𝑆
Required result proved.
∴ 𝐿 𝑓 𝑡 𝑎𝑛𝑑 𝐿 𝑔 𝑡 𝑒𝑥𝑖𝑠𝑡.
⇒ 𝛼𝐿 𝑓 𝑡 +𝛽𝐿 𝑔 𝑡 <∞ 𝑎𝑙𝑠𝑜 𝑒𝑥𝑖𝑠𝑡. (𝐼𝑚𝑝𝑟𝑜𝑝𝑒𝑟 𝐼𝑛𝑡𝑒𝑔𝑟𝑎𝑙)

Lecture NO. 78
Corollary:
𝐿𝛼 𝑓 𝑡 +𝛽 𝑔 𝑡 =𝛼𝐿 𝑓 𝑡 + 𝛽 𝐿{𝑔 𝑡 }
If 𝛽 = 0;
𝐿 𝛼 𝑓 𝑡 =𝛼𝐿 𝑓 𝑡
Here we use “Scalar Composition” property.
This property also used in calculus. For example:
𝐿𝑖𝑛𝑒𝑎𝑟 𝑇𝑟𝑎𝑛𝑠𝑓𝑜𝑟𝑚𝑎𝑡𝑖𝑜𝑛: 𝑇 𝛼𝑣 = 𝛼 𝑇𝑣
𝑑 𝑑
𝐷𝑒𝑟𝑖𝑣𝑎𝑡𝑖𝑣𝑒: 𝛼 𝑠𝑖𝑛𝑥 = 𝛼 𝑠𝑖𝑛𝑠
𝑑𝑥 𝑑𝑥
𝐼𝑛𝑡𝑒𝑔𝑟𝑎𝑡𝑖𝑜𝑛: 𝑐 𝑓 𝑥 𝑑𝑥 = 𝑐 𝑓(𝑐)𝑑𝑥

Lecture NO. 79
Laplace Transform of Hyperbolic Function:
As we know
𝑒 𝑥 + 𝑒 −𝑥
𝑐𝑜𝑠𝑕 𝑥 =
2
Calculate: 𝐿(𝑐𝑜𝑠𝑕 𝑎𝑡), where “a” is a constant
Solution: As given
1 𝑎𝑡 1
𝐿 𝑐𝑜𝑠𝑕 𝑎𝑡 = 𝐿 𝑒 + 𝑒 −𝑎𝑡 ⇒ 𝐿 𝑒 𝑎𝑡 + 𝑒 −𝑎𝑡
2 2
1 1 1 1
𝐿 𝑐𝑜𝑠𝑕 𝑎𝑡 = 𝐿 𝑒 𝑎𝑡 ) + 𝐿(𝑒 −𝑎𝑡 ⇒ +
2 2 𝑠−𝑎 𝑠+𝑎
𝑎𝑡 1 −𝑎𝑡 1
∴𝐿 𝑒 = 𝑎𝑛𝑑 𝐿 𝑒 = , so
𝑠−𝑎 𝑠+𝑎
1 𝑠+𝑎+𝑠−𝑎 𝑠
𝐿 𝑐𝑜𝑠𝑕 𝑎𝑡 = 2 2
⇒ 2
2 𝑠 −𝑎 𝑠 − 𝑎2
This is our required result.

Lecture NO. 80
Laplace Transform of Cosine and Sine:
Show that
𝑠 𝑎
𝐿 𝑐𝑜𝑠 𝑎𝑡 = 𝑎𝑛𝑑 𝐿 𝑠𝑖𝑛 𝑎𝑡 =
𝑠 2 + 𝑎2 𝑠 2 + 𝑎2
Proof: Suppose that
𝐿 𝑐𝑜𝑠 𝑎𝑡 = 𝐹𝑐 𝑎𝑛𝑑 𝐿 𝑠𝑖𝑛 𝑎𝑡 = 𝐹𝑠
𝑳 𝒄𝒐𝒔 𝒂𝒕 : Applying Laplace Transform
∞

𝐹𝑐 = 𝑒 −𝑠𝑡 (𝑐𝑜𝑠 𝑎𝑡) 𝑑𝑡

0
Integrating and applying limits, we have
∞ ∞
𝑒 −𝑠𝑡 𝑒 −𝑠𝑡
𝐹𝑐 = 𝑐𝑜𝑠 𝑎𝑡 − −𝑎𝑠𝑖𝑛 𝑎𝑡 𝑑𝑡
−𝑠 0
−𝑠
0
∞
∞
1 𝑐𝑜𝑠 𝑎𝑡 𝑎
𝐹𝑐 = − − 𝑒 −𝑠𝑡 𝑠𝑖𝑛 𝑎𝑡 𝑑𝑡
𝑠 𝑒 𝑠𝑡 0 𝑠
0
Page | 46
MTH647 Handout

1 𝑐𝑜𝑠 ∞ 𝑐𝑜𝑠(0) 𝑎
𝐹𝑐 = − ∞
− 0
− 𝐿 𝑠𝑖𝑛 𝑎𝑡
𝑠 𝑒 𝑒 𝑠
1 𝑎
𝐹𝑐 = − 0 − 1 − 𝐹𝑠
𝑠 𝑠
1 𝑎
𝐹𝑐 = − 𝐹𝑠 … 1
𝑠 𝑠
𝑳 𝒔𝒊𝒏 𝒂𝒕 :
∞ ∞ ∞
−𝑠𝑡
𝑒 −𝑠𝑡 𝑒 −𝑠𝑡
𝐹𝑠 = 𝑒 (𝑠𝑖𝑛 𝑎𝑡) 𝑑𝑡 ⇒ 𝑠𝑖𝑛 𝑎𝑡 − 𝑎𝑐𝑜𝑠 𝑎𝑡 𝑑𝑡
−𝑠 0
−𝑠
0 0
∞
∞
1 𝑠𝑖𝑛 𝑎𝑡 𝑎
𝐹𝑠 = − + 𝑒 −𝑠𝑡 𝑐𝑜𝑠 𝑎𝑡 𝑑𝑡
𝑠 𝑒 𝑠𝑡 0 𝑠
0
1 𝑠𝑖𝑛 ∞ 𝑠𝑖𝑛(0) 𝑎
𝐹𝑠 = − − + 𝐿 𝑐𝑜𝑠 𝑎𝑡
𝑠 𝑒∞ 𝑒0 𝑠
1 𝑎
𝐹𝑠 = − 0 − 0 + 𝐹𝑐
𝑠 𝑠
𝑎
𝐹𝑠 = 𝐹𝑐 … (2)
𝑠
By putting Fs value in equation (1), we get
1 𝑎 𝑎 1 𝑎2
𝐹𝑐 = − 𝐹 ⇒ − 𝐹
𝑠 𝑠 𝑠 𝑐 𝑠 𝑠2 𝑐
𝑎2 1 𝑠 2 + 𝑎2 1
𝐹𝑐 + 2 𝐹𝑐 = ⇒ 𝐹𝑐 =
𝑠 𝑠 𝑠2 𝑠
1 𝑠2 𝑠
𝐹𝑐 = 2 2
⇒ 2
𝑠 𝑠 +𝑎 𝑠 + 𝑎2
As 𝐿 𝑐𝑜𝑠 𝑎𝑡 = 𝐹𝑐 , so
𝑠
𝐿 𝑐𝑜𝑠 𝑎𝑡 = 2
𝑠 + 𝑎2
This is our required result.
Now, from equation (2)
𝑠
𝐹𝑐 = 𝐹𝑠
𝑎
By putting in equation (1) we get
𝑠 1 𝑎 𝑠 𝑎 𝑠 1
𝐹𝑠 = − 𝐹𝑠 ⇒ + 𝐹𝑠 =
𝑎 𝑠 𝑠 𝑎 𝑠 𝑎 𝑠
2 2
𝑠 +𝑎 1 𝑎
𝐹𝑠 = ⇒ 𝐹𝑠 = 2
𝑎𝑠 𝑠 𝑠 + 𝑎2
As 𝐿 𝑠𝑖𝑛 𝑎𝑡 = 𝐹𝑠 , so
𝑎
𝐿 𝑠𝑖𝑛 𝑎𝑡 = 2
𝑠 + 𝑎2
This is our required result.

Lecture NO. 81
𝒏
Laplace Transform of 𝒕 (𝒏 ∈ 𝑵):
Theorem: Prove that
𝑛!
𝐿 𝑡𝑛 = ; 𝑛∈𝑁 … 1
𝑠 𝑛+1
Proof: We’ll prove it by mathematical induction.
For 𝑛 = 0;
0! 1
𝐿 𝑡 0 = 0+1 ⇒ 𝐿 1 =
𝑠 𝑠
Page | 47
MTH647 Handout

Suppose equation (1) is true for fixed "𝑛" and we will prove it for "𝑛 + 1".
For 𝑛 + 1;
(𝑛 + 1)!
𝐿 𝑡 𝑛+1 =
𝑠 𝑛+2
Applying Laplace Transform,
∞

𝐿 𝑡 𝑛+1 = 𝑒 −𝑠𝑡 𝑡 𝑛+1 𝑑𝑡

0
Integrating and applying limits, we have
∞ ∞
𝑛+1 𝑛+1
𝑒 −𝑠𝑡 𝑒 −𝑠𝑡
𝐿 𝑡 = 𝑡 − 𝑛 + 1 𝑡 𝑛 𝑑𝑡
−𝑠 0
−𝑠
0
∞
𝑛+1 ∞
1 𝑡 𝑛+1
𝐿 𝑡 𝑛+1 =− + 𝑒 −𝑠𝑡 𝑡 𝑛 𝑑𝑡
𝑠 𝑒 𝑠𝑡 0
𝑠
0
𝑛+1
1 𝑡 𝑛+1
𝐿 𝑡 𝑛+1 = − lim 𝑠𝑡 − 0 + 𝐿 𝑡𝑛
𝑠 𝑡→∞ 𝑒 𝑠
𝑛!
As given 𝐿 𝑡 𝑛 = , so
𝑠 𝑛 +1
1 𝑛 + 1 𝑛!
𝐿 𝑡 𝑛+1 = −
0−0 +
𝑠 𝑠 𝑠 𝑛+1
(𝑛 + 1)!
𝐿 𝑡 𝑛+1 =
𝑠 𝑛+2
i.e. it is true for "𝑛 + 1". Hence, proved our required result.

Lecture NO. 82
𝜶
Laplace Transform of 𝒕 (𝜶 > −1):
Aaplying Laplace Transorm,
∞

𝐿 𝑡𝛼 = 𝑒 −𝑠𝑡 𝑡 𝛼 𝑑𝑡
0
1 𝑠
𝑣 = 𝑠𝑡 ⇒ 𝑑𝑣 = 𝑠𝑑𝑡 ⇒ 𝑑𝑡 = 𝑑𝑣 𝑎𝑛𝑑 𝑡 =
Put 𝑠 𝑣
𝐴𝑠 𝑡 → 0 ⇒ 𝑣 → 0 𝑎𝑛𝑑 𝑎𝑠 𝑡 → ∞ ⇒ 𝑣 → ∞
∞ ∞
𝛼 −𝑣
𝑣 𝛼 1 −𝑣
𝑣𝛼
𝐿 𝑡 = 𝑒 𝑑𝑣 ⇒ 𝑒 𝑑𝑣
𝑠 𝑠 𝑠 𝛼+1
0 0
∞
1
𝐿 𝑡𝛼 = 𝑒 −𝑣 𝑣 𝛼 𝑑𝑣
𝑠 𝛼+1
0
Here
∞

𝑒 −𝑣 𝑣 𝛼 𝑑𝑣 = 𝐺𝑎𝑚𝑚𝑎 𝐹𝑢𝑛𝑐𝑡𝑖𝑜𝑛 = 𝛾 𝛼 + 1
0
So,
𝛾 𝛼+1
𝐿 𝑡𝛼 =
𝑠 𝛼 +1
This is our required result.

Lecture NO. 83
S-Shifting Theorem:
If 𝑓 𝑡 has the transform 𝐹 𝑠 where 𝑠 > 𝑘, then "𝑒 𝛼𝑡 𝑓 𝑡 " has the transform 𝐹 𝑠 − 𝛼 where 𝑠 − 𝛼 > 𝑘
i.e.
𝐿 𝑒 𝛼𝑡 𝑓 𝑡 = 𝐹 𝑠 − 𝛼
Page | 48
MTH647 Handout

In term of inverse Laplace transform,

𝐿−1 𝐹 𝑠 − 𝛼 = 𝑒 𝛼𝑡 𝑓 𝑡
Proof: As we know Laplace Transform,
∞

𝐿 𝑓 𝑡 =𝐹 𝑠 = 𝑒 −𝑠𝑡 𝑓 𝑡 𝑑𝑡
0
For given function,
∞ ∞ ∞

𝐹 𝑠−𝛼 = 𝑒 −(𝑠−𝛼)𝑡 𝑓 𝑡 𝑑𝑡 ⇒ 𝑒 −𝑠𝑡 𝑒 𝛼𝑡 𝑓 𝑡 𝑑𝑡 ⇒ 𝑒 −𝑠𝑡 𝑒 𝛼𝑡 𝑓 𝑡 𝑑𝑡

0 0 0
𝛼𝑡
𝐹 𝑠−𝛼 =𝐿 𝑒 𝑓 𝑡
It can also be written as,
𝐿−1 𝐹 𝑠 − 𝛼 = 𝑒 𝛼𝑡 𝑓 𝑡
This is our required result.

Lecture NO. 84
Application of S-Shifting Theorem:
Evaluate
2𝑠 − 27
𝑖): 𝐿 𝑒 𝛼𝑡 𝑐𝑜𝑠𝜔𝑡 𝑖𝑖): 𝐿 𝑒 𝛼𝑡 𝑠𝑖𝑛𝜔𝑡 𝑖𝑖𝑖) 𝐿−1
𝑠 2 + 2𝑠 + 401
Solution: (Here we are going to use previous results)
𝒊): 𝑳 𝒆𝜶𝒕 𝒄𝒐𝒔𝝎𝒕
𝑠 𝑠−𝑜
𝐿 𝑐𝑜𝑠𝜔𝑡 = 2 2
⇒
𝑠 +𝜔 𝑠 − 𝑜 2 + 𝜔2
𝑠−𝛼
𝐿 𝑒 𝛼𝑡 𝑐𝑜𝑠𝜔𝑡 =
𝑠 − 𝛼 2 + 𝜔2
𝒊𝒊): 𝑳 𝒆𝜶𝒕 𝒔𝒊𝒏𝝎𝒕
𝜔
𝐿 𝑒 𝛼𝑡 𝑠𝑖𝑛𝜔𝑡 =
𝑠 − 𝛼 2 + 𝜔2
𝟐𝒔−𝟐𝟕
𝒊𝒊𝒊) 𝑳−𝟏 𝟐
𝒔 +𝟐𝒔+𝟒𝟎𝟏
2𝑠 − 27 2𝑠 + 2 − 2 − 27 2 𝑠 + 1 − 29
𝐿−1 = 𝐿−1
⇒ 𝐿−1
𝑠 2 + 2𝑠 + 401 𝑠 2 + 2𝑠 + 1 − 1 + 401 𝑠 + 1 2 + 400
2 𝑠+1 29
= 𝐿−1 2 2
− 𝐿−1
𝑠 + 1 + 20 𝑠 + 1 2 + 20 2
2 𝑠+1 29 20
= 𝐿−1 − 𝐿−1
×
𝑠 + 1 2 + 20 2 𝑠 + 1 2 + 20 2 20
𝑠+1 20
= 2𝐿−1 − 29𝐿 −1
𝑠 + 1 2 + 20 2 20 𝑠 + 1 2 + 20 2
𝑠 − −1 29 −1 20
= 2𝐿−1 2 − 𝐿 2
𝑠 − (−1 + 20 2 20 𝑠 − (−1 + 20 2

Using method of calculations i) and ii)

29 −𝑡
= 2𝑒 −𝑡 𝑐𝑜𝑠20𝑡 − 𝑒 𝑠𝑖𝑛20𝑡
20
29
= 𝑒 −𝑡 2 𝑐𝑜𝑠20𝑡 − 𝑠𝑖𝑛20𝑡
20
This is our required result.

Lecture NO. 85
Piecewise Continuous Function:
Infromlly, a function is a piecwise continuous on an interval if it has finite jumps (discontinuities) on that
interval.
Page | 49
MTH647 Handout

A function 𝑓 𝑡 is a piecewise continuous on a finite interval [𝑎, 𝑏] if satisfy the following conditions.
1) Function 𝑓 𝑡 defined on an/that interval. (i.e. [𝑎, 𝑏])
2) Interval can be sub-divided into finitely many sub-intervals in each of which function 𝑓 𝑡 is
continuous.
3) Function 𝑓 𝑡 has finite limits as "𝑡" approaches either end points of sub-intervals from interior
of [𝑎, 𝑏]. 𝑓 𝑎0 + 0 < ∞ , 𝑓 𝑎0 − 0 < ∞ ; 𝐹𝑖𝑛𝑖𝑡𝑒}
Example:
𝑠𝑖𝑛𝑥 𝑖𝑓 0 ≤ 𝑥 ≤ 𝜋/2
𝑓 𝑡 = 𝑒𝑥 𝑖𝑓 𝜋/2 ≤ 𝑥 ≤ 𝜋 𝑂𝑣𝑒𝑟𝑎𝑙𝑙 𝑖𝑛𝑡𝑒𝑟𝑣𝑎𝑙 [0,2𝜋]
−2 𝑖𝑓 𝜋 ≤ 𝑥 ≤ 2𝜋

Lecture NO. 86
Existence Theorem for Laplace Transform:
Let 𝑓 𝑡 be a piecewise continuous function on every finite interval in [0, ∞) and satisfies |𝑓 𝑡 𝑒 −𝑘𝑡 | ≤ 𝑚,
∀ 𝑡 ≥ 0 and for some "𝑘 𝑎𝑛𝑑 𝑚" then Laplace Transform of 𝑓(𝑡) exists for all 𝑠 > 𝑘.
Answer: As given
𝑓 𝑡 𝑒 −𝑘𝑡 ≤ 𝑚 … 1 ∀ 𝑡 ≥ 0 & 𝑓𝑜𝑟 𝑠𝑜𝑚𝑒 "𝑘 𝑎𝑛𝑑 𝑚"
Here equation (1) is known as “Growth Restriction”. Equation (1) implies,
𝑓 𝑡 ≤ 𝑚 𝑒 𝑘𝑡
𝑓(𝑡)
𝑓 𝑡 𝑒 −𝑘𝑡 → 0 𝑎𝑠 𝑡 → ∞ ⇒ lim 𝑘𝑡 → 0
𝑡→∞ 𝑒
Proof: As given that 𝑓 𝑡 be a piecewise continuous function. 𝑓 𝑡 𝑒 −𝑘𝑡 is integerable over any finite
interval on [0, ∞). Laplace Transform,
∞ ∞

𝐿 𝑓 𝑡 = 𝑒 −𝑠𝑡 𝑓 𝑡 𝑑𝑡 ≤ 𝑓 𝑡 𝑒 −𝑠𝑡 𝑑𝑡
0 0
For given function,
∞ ∞

𝐿 𝑓 𝑡 ≤ 𝑚 𝑒 𝑘𝑡 𝑒 −𝑠𝑡 𝑑𝑡 ⇒ 𝑚 𝑒 −(𝑠−𝑘)𝑡 𝑑𝑡 … (2)

0 0
Integrating and applying limits, (without “m”) , we have
∞
∞
1 ∞ 1 1
𝑒 −(𝑠−𝑘)𝑡 𝑑𝑡 = − 𝑒 −(𝑠−𝑘)𝑡 ⇒ −
𝑠−𝑘 0 𝑠 − 𝑘 𝑒 𝑠−𝑘 𝑡
0
0
1 1 1 1
=− − ⇒ ; 𝑓𝑜𝑟 𝑠 > 𝑘
𝑠 − 𝑘 𝑒∞ 𝑒0 𝑠−𝑘
So, equation (2) becomes,
𝑚
𝐿 𝑓 𝑡 ≤ ⇒ 𝐿 𝑓 𝑒𝑥𝑖𝑠𝑡 𝑖𝑓 𝑔𝑟𝑜𝑤𝑡𝑕 𝑟𝑒𝑠𝑡𝑟𝑖𝑐𝑡𝑖𝑜𝑛 𝑖𝑠 𝑠𝑎𝑡𝑖𝑠𝑓𝑖𝑒𝑑.
𝑠−𝑘
Hence prove our required result.

Lecture NO. 87
Counter Example of Existence Theorem:
Example:
1
𝑓 𝑡 = 𝐴𝑠 𝑡 → 0 𝑡𝑕𝑒𝑛 𝑓 𝑡 → ∞
𝑡
Here 𝑓(𝑡) is not a piecewise continuous function on [0, ∞). Also
𝑓 𝑡 ≤ 𝑚 𝑒 𝑘𝑡 𝐼𝑓 𝑚 = 1, 𝑘 = 0
Applying Laplace transform on given function,
Page | 50
MTH647 Handout

∞
1
𝐿 𝑓 𝑡 = 𝑒 −𝑠𝑡 𝑑𝑡
𝑡
0
2
𝐿𝑒𝑡 𝑠𝑡 = 𝑥 ⇒ 𝑠𝑡 = 𝑥
Put 1 𝑑𝑡 2
𝑥= 𝑠 𝑑𝑡 ⇒ = 𝑑𝑥
2 𝑡 𝑡 𝑠
∞ ∞
−𝑥 2
2 2 2
𝐿 𝑓 𝑡 = 𝑒 𝑑𝑥 ⇒ 𝑒 −𝑥 𝑑𝑥 … (1)
𝑠 𝑠
0 0
Here
∞
2 𝜋
𝑒 −𝑥 𝑑𝑥 = (𝐺𝑎𝑢𝑠𝑠𝑖𝑎𝑛 𝐼𝑛𝑡𝑒𝑔𝑟𝑎𝑙)
2
0
Hence relation (1) becomes,
2 𝜋 𝜋
𝐿 𝑓 𝑡 = × = ; 𝑠>0
𝑠 2 𝑠
1
Hence is not a piecewise continuous function but its Laplace Transform exists. This is our required result.
𝑡

Lecture NO. 88
Laplace Transform of Derivatives:
Prove that
𝑖) 𝐿 𝑓 ′ = 𝑠𝐿 𝑓 − 𝑓 0 𝑖𝑖) 𝐿 𝑓 ′ ′ = 𝑠 2 𝐿 𝑓 − 𝑠𝑓 0 − 𝑓′ 0
Proof: since we know
𝐼𝑓; 𝑓 𝑖𝑠 𝑑𝑖𝑓𝑓𝑒𝑟𝑒𝑛𝑡𝑖𝑎𝑏𝑙𝑒 ⇒ 𝑓 𝑖𝑠 𝑐𝑜𝑛𝑡𝑖𝑛𝑢𝑜𝑢𝑠.
I):
Laplace transform fro given function
∞

𝐿 𝑓′ = 𝐿 𝑓′ 𝑡 = 𝑒 −𝑠𝑡 𝑓′ 𝑡 𝑑𝑡
0
Here 𝑓(𝑡) is continuous. Integrating and applying limits, we have
∞
∞
𝐿 𝑓 ′ = 𝑒 −𝑠𝑡 𝑓(𝑡) 0 − 𝑓 𝑡 – 𝑠 𝑒 −𝑠𝑡 𝑑𝑡
0
∞
𝑓(𝑡) 𝑓(0)
𝐿 𝑓′ = lim 𝑠𝑡 − 0 + 𝑠 𝑒 −𝑠𝑡 𝑓′ 𝑡 𝑑𝑡
𝑡→∞ 𝑒 𝑒
0
′
𝐿 𝑓 = 0 − 𝑓 0 + 𝑠𝐿(𝑓)
𝐿 𝑓 ′ = 𝑠𝐿 𝑓 − 𝑓 0 … (1)
II):
Taking derivative of equation (1),
𝐿 𝑓 ′′ = 𝑠𝐿 𝑓′ − 𝑓′ 0
Putting value from equation (1),
𝐿 𝑓 ′′ = 𝑠 𝑠𝐿 𝑓 − 𝑓 0 − 𝑓′ 0
𝐿 𝑓 ′ ′ = 𝑠 2 𝐿 𝑓 − 𝑠𝑓 0 − 𝑓′ 0
Hence prove.

Lecture NO. 89
Laplace Transform of nth Derivatives:
𝐿 𝑓 𝑛 = 𝑠 𝑛 𝐿 𝑓 − 𝑠 𝑛 −1 𝑓 0 − 𝑠 𝑛−2 𝑓 ′ 0 … − 𝑠 0 𝑓 𝑛 −1 (0)
Watch Lecture(s).
Page | 51
MTH647 Handout

Lecture NO. 90
Applications of Laplace of Derivatives:
Find 𝐿(𝑐𝑜𝑠𝜔𝑡) y using derivative expression of Laplace.
Solution: since we know from previous lectures
𝐿 𝑓 ′ = 𝑠𝐿 𝑓 − 𝑓 0 … (1)
′ 2 ′
𝐿 𝑓 ′ = 𝑠 𝐿 𝑓 − 𝑠𝑓 0 − 𝑓 (0) … (2)
Here
𝑓 𝑡 = 𝑐𝑜𝑠𝜔𝑡 ⇒ 𝑓 ′ 𝑡 = −𝜔𝑠𝑖𝑛𝜔𝑡 & 𝑓 ′′ 𝑡 = −𝜔2 𝑐𝑜𝑠𝜔𝑡
𝑓 0 = 1 ⇒ 𝑓′ 0 = 0
By putting values in equation 2,
𝐿 𝑓 ′ ′ = 𝑠 2 𝐿 𝑓 − 𝑠𝑓 0 − 𝑓 ′ 0
𝐿 −𝜔2 𝑐𝑜𝑠𝜔𝑡 = 𝑠 2 𝐿 𝑓 − 𝑠
−𝜔2 𝐿 𝑐𝑜𝑠𝜔𝑡 = 𝑠 2 𝐿 𝑓 − 𝑠
−𝜔2 𝐿 𝑓 = 𝑠 2 𝐿 𝑓 − 𝑠
𝑠 = 𝐿 𝑓 𝑠 2 + 𝜔2
𝑠
𝐿 𝑓 = 2
𝑠 + 𝜔2

Lecture NO. 91
Evaluate:
𝐿(𝑡 𝑠𝑖𝑛𝜔𝑡)
Solution: Here
𝑓 𝑡 = 𝑡 𝑠𝑖𝑛𝜔𝑡 ⇒ 𝑓(0) = 0
𝑓 𝑡 = 𝑡 𝜔 𝑐𝑜𝑠𝜔𝑡 + 1 𝑠𝑖𝑛𝜔𝑡 ⇒ 𝑓 ′ 0 = 0
′

𝑓 ′′ 𝑡 = 𝜔 𝑡 −𝜔 𝑠𝑖𝑛𝜔𝑡 + 1 𝑐𝑜𝑠𝜔𝑡 + 𝜔 𝑐𝑜𝑠𝜔𝑡

𝑓 ′′ 𝑡 = −𝜔2 𝑡 𝑠𝑖𝑛𝜔𝑡 + 2𝜔 𝑐𝑜𝑠𝜔𝑡
By putting values in equation 2,
𝐿 𝑓 ′ ′ = 𝑠 2 𝐿 𝑓 − 𝑠𝑓 0 − 𝑓 ′ 0
𝐿 −𝜔2 𝑡 𝑠𝑖𝑛𝜔𝑡 + 2𝜔 𝑐𝑜𝑠𝜔𝑡 = 𝑠 2 𝐿 𝑓
−𝜔2 𝐿 𝑡 𝑠𝑖𝑛𝜔𝑡 + 2𝜔𝐿 𝑐𝑜𝑠𝜔𝑡 = 𝑠 2 𝐿 𝑓
𝑠
−𝜔2 𝐿 𝑓 + 2𝜔𝐿 2 = 𝑠2𝐿 𝑓
𝑠 + 𝜔2
2𝜔𝑠
= 𝐿 𝑓 𝑠 2 + 𝜔2
𝑠 + 𝜔2
2
2𝜔𝑠
𝐿 𝑓 = 2
𝑠 + 𝜔2 2
This is our required result.

Lecture NO. 92
Laplace Transform of Integral:
Theorem: let 𝑓 𝑡 is a piecewise continuous function for 𝑡 ≥ 0, & 𝑑 𝑓 𝑡 = 𝐹(𝑠). Further |𝑓 𝑡 | ≤ 𝑚𝑒 𝑘𝑡
for some𝑚 > 0 𝑎𝑛𝑑 𝑘 > 0, then
𝑡 ∞
1 𝐹(𝑠)
𝐿 𝑓 𝜏 𝑑𝜏 = 𝐹 𝑠 𝑎𝑛𝑑 𝐿−1 = 𝑓 𝜏 𝑑𝜏
𝑠 𝑠
0 0
Proof: Let
𝑡

𝑔 𝑡 = 𝑓 𝜏 𝑑𝜏
0
∴ 𝑓 𝑡 𝑖𝑠 𝑝𝑖𝑒𝑐𝑒𝑤𝑖𝑠𝑒 𝑐𝑜𝑛𝑡𝑖𝑛𝑢𝑜𝑢𝑠 ⇒ 𝑔 𝑡 𝑖𝑠 𝑐𝑜𝑛𝑡𝑖𝑛𝑢𝑜𝑢𝑠

Page | 52
MTH647 Handout

𝑡 𝑡 𝑡

𝑔 𝑡 = 𝑓 𝜏 𝑑𝜏 ≤ 𝑓 𝜏 𝑑𝜏 ≤ 𝑚𝑒 𝑘𝑡 𝑑𝜏
0 0 0
𝑘𝑡 𝑡
𝑒 𝑚 𝑘𝑡 𝑚
𝑔 𝑡 =𝑚 ⇒ 𝑒 − 1 ≤ 𝑒 𝑘𝑡
𝑘 0
𝑘 𝑘
⇒ 𝑔(𝑡) satisfies growth restriction. Now
𝑡
𝑑
𝑔′ 𝑡 = 𝑓 𝜏 𝑑𝜏 = 𝑓(𝑡)
𝑑𝑡
0
Except at point of discontinuities. This implies 𝑔 𝑡 𝑖𝑠 𝑐𝑜𝑛𝑡𝑖𝑛𝑢𝑜𝑢𝑠.
𝐿 𝑔′ 𝑡 = 𝐿 𝑓 𝑡 𝑎𝑛𝑑 𝑠𝐿 𝑔 𝑡 − 𝑔 0 = 𝐹 𝑠
𝑡 𝑡
𝐹 𝑠
𝑠𝐿 𝑓 𝜏 𝑑𝜏 = 𝐹 𝑠 ⇒ 𝐿 𝑓 𝜏 𝑑𝜏 =
𝑠
0 0

Lecture NO. 93
Evaluate:
1 1
𝑖) 𝐿−1 𝑎𝑛𝑑 𝑖𝑖) 𝐿−1
𝑠 𝑠2 + 𝜔2 𝑠2 𝑠2 + 𝜔2
I):
1 1 𝜔 1 −1 𝜔
𝐿−1 = 𝐿−1 . 2 ⇒ 𝐿
𝑠2 + 𝜔2 𝜔 𝑠 + 𝜔2 𝜔 𝑠 + 𝜔2
2

1 𝜔
𝐿−1 𝐹 𝑠 𝑠𝑖𝑛𝜔𝑡 = ; ∴ 𝐿−1 = 𝑠𝑖𝑛𝜔𝑡
𝜔 𝑠 2 + 𝜔2
Now, (Going to calculate required function)
𝑡
1
𝐿−1 𝐹 𝑠 = 𝑓 𝜏 𝑑𝜏
𝑠
0
𝑡
1 1
𝐿−1 = 𝑠𝑖𝑛𝜔𝜏 𝑑𝜏
𝑠 𝑠 2 + 𝜔2 𝜔
0
Integrating and applying limits, we have
1 1
𝐿−1 2 2
= − 2
𝑐𝑜𝑠𝜔𝜏 𝑡0
𝑠 𝑠 +𝜔 𝜔
1 1
𝐿−1 = 2 1 − 𝑐𝑜𝑠𝜔𝑡
𝑠 𝑠 2 + 𝜔2 𝜔
This is our required result.
II):
1 1 1 1
𝐿−1 = 𝐿−1 ⇒ 𝐿−1 𝐹(𝑠)
𝑠2 𝑠2 + 𝜔2 𝑠 𝑠 𝑠 + 𝜔2
2 𝑠
𝑡 𝑡
1 1
𝐿−1 = 𝑓 𝜏 𝑑𝜏 ⇒ 1 − 𝑐𝑜𝑠𝜔𝑡 𝑑𝜏
𝑠 2 𝑠 2 + 𝜔2 𝜔2
0 0
1 1 𝑠𝑖𝑛𝜔𝜏
𝐿−1 = 2 𝜏−
𝑠 2 𝑠 2 + 𝜔2 𝜔 𝜔
1 1 𝑠𝑖𝑛𝜔𝑡
𝐿−1 = 2 𝑡−
𝑠 2 𝑠 2 + 𝜔2 𝜔 𝜔
This is our required result.
____________________

Page | 53