See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/287209798

A Notebook on Numerical Methods

Preprint · December 2015

CITATIONS READS

0 32,815

1 author:

Shree Krishna Khadka

Nepal Telecom
50 PUBLICATIONS   25 CITATIONS   

SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Online Certification Courses View project

Research Proposal and Proceedings View project

All content following this page was uploaded by Shree Krishna Khadka on 24 January 2016.

The user has requested enhancement of the downloaded file.

 A Note Book On

NUMERICAL METHODS
For BE-V Semester (Computer Engineering)

Prepared By
Er. SriKisna

Inside
o Numerical Methods: Introduction & Scope
o Errors in Numerical Calculation
o Solution of Algebraic & Transcendental Equation
o Interpolation
o Curve Fitting, B Splines & Approximation
o Numerical Differentiation & Integration
o Matrices & Linear System of Equations
o Solution of Ordinary Differential Equations
o Solution of Partial Differential Equations
o Solution of Integral Equations
o Programming Examples
 HWHIC GSSE
BLEX IV

CONTENT
Introduction to Numerical Methods (1 – 6)
Course Plan & Strategy ----------------------------------- 2
Importance & Need of Numerical Methods 4
Scope & Application of Numerical Methods 6
Chapter: 1 Error in Numerical Calculations (8 – 14)

1.1 Numbers & their accuracy --------------------------- 8

 Significant Digits 8
 Accuracy & Precision 8
 Rules for Rounding off 8
1.2 Types of Errors & 9
1.3 Numerical Examples 10
1.4 Mathematical Preliminaries 11
1.5 General Error Formula 12
1.6 Convergence 13
1.7 Assignment 1
14
Chapter: 2 Solutions of Algebric & Transcendental Equations (15 – 35)

2.1 Introduction to types of equations ----------------- 16

 Algebric Functions 16
 Transcendental Equations 16
 Polynomial Equations 16
o Evaluation of Polynomial by Horner’s Method 16
 Linear & Non Linear Equations 17
2.2 Solving Non Linear Equations by Iterative Methods
18
 Bisection Method & Its Convergence
 Regular Falsi Method & Its Convergence
18
 Newton Raphson Method & Its Variance and 20
Convergence 22
 Secant Method & Its Converges 28
 Fixed Point Iteration Method & Its Convergence 31
 Horner’s Method of Finding Roots 33
2.3 Assignment 2 35

Chapter: 3 Interpolations (36 – 49)

3.1 Introduction & Forms of Polynomial Equations ---------- 37

3.2 Errors in Polynomial Interpolation 37
3.3 Finite Difference Methods 38
 Forward Differences 38
 Backward Differences 39
o Difference of a Polynomials 41
o Newton Interpolation Formula
42
 Central Differences
42
 Linear Interpolation
43
 Lagrange’s Interpolation
44
 Inverse Interpolation 46

1
 HWHIC GSSE
BLEX IV

3.4 Divided Differences 47

3.5 Assignment 3 49

Chapter: 4 Curve Fittings, B-Splines & Approximations (50 – 63)

4.1 Introduction --------------------------------------- 51
4.2 Least Square Regression 51
 Fitting Linear Equation 52
 Fitting Transcendental Equation 53
 Fitting Polynomial Equation
55
4.3 Multiple Linear Regression
4.4 Spline Interpolation 57
 Cubic B-Spline 58
4.5 Approximation of Functions 62
4.6 Assignment 4 63

Chapter: 5 Differentiation & Integration (64 – 88)

4.1 Numerical Differentiation: Introduction ------------ 65

4.2 Differentiating Continuous Functions 65
 Forward & Backward Difference Method 65
 Central Difference Method 66
 Error Analysis 67
 High Order Derivatives 69
4.3 Differentiating Tabulated Functions 70
4.4 Numerical Integration: Introduction 73
4.5 Newton Cote’s General Integration Formula 73
 Trapezoidal Rule 74
 Simpson’s Rule 77
4.6 Numerical Double Integration 80
4.7 Gaussian Integration 81
4.8 Romberg Integration
86
4.9 Assignment 5
88

Chapter: 6 Matrices & Linear System of Equations (89 – 104)

6.1 Introduction --------------------------------------- 90
6.2 Existence of Solution 90
6.3 Method of Solving Linear Equations 92
6.4 Solution by Elimination 93
 Basic Gauss Elimination & Its Limitation 93
 Gauss Elimination with Pivoting 95
 Gauss Jordan Method 96
 Triangular Factorization Method 97
 Singular Value Decompositions 99
6.5 Solution by Iteration 101
 Jacobi Method 101
 Gauss Seidal Method 102
6.6 Assignment 6 104

1
 HWHIC GSSE
BLEX IV

Chapter: 7 Numerical Solutions of Ordinary Differential Equations (105 – 120)

7.1 Introduction & Types of ODEs ----------------------- 106

 Order & Degree of Differential Equations 107
 Linear & Non Linear Differential Equations 107
 General & Particular Solution 107
 One Step & Multi Step Solution 107
7.2 Solution by Taylor’s Series Method
7.3 Euler’s Method
108
7.4 Heun’s Method 109
7.5 Fourth Order Runge Kutta Method 110
7.6 Simultaneous First Order Differential Equations 112
7.7 Solution of ODEs as Boundary Value Problem 114
 Finite Difference Method 116
 Shooting Method
7.8 Assignment 7
118
120

Chapter: 8 Numerical Solutions of Partial Differential Equations (121 – 132)

8.1 Introduction---------------------------------------- 122
8.2 Finite Difference Approximation 123
8.3 Solution of Elliptic Equations 124
 Laplace Equations 124
 Poisson’s Equations 126
8.4 Solution of Parabolic Equations 127
8.5 Solution of Hyperbolic Equations 129
8.6 Assignment 8
132

Chapter: 9 Numerical Solutions of Integral Equations (133 – 140)

9.1 Introduction --------------------------------------- 134
9.2 Method of Degenerated Kernels 135
9.3 Method of Generalized Quadrature 137
9.4 Chebyshev Series & Cubic Spline Method 138
9.5 Assignment 9 140

Chapter: 10 Programming Examples ( 141 – 156)

1
WhiteHouse Institute of Science & Technology
Khumaltar, Lalitpur-Satdobato

Bachelor in E & C/Computer Engineering

Bachelor in Information Technology

NUMRECIAL METHODS

Lecture Handouts

Prepared by: Er Shree Krishna Khadka

2012
1
Page

Prepared By
Er. Shree Krishna Khadka
1. Reference books:

a. Numerical Methods by E. Balagurusamy

b. Introductory Methods of Numerical Analysis by S.S. Sastry

2. Teaching Schedule:

a. Theory: 3hrs/week
b. Practical: 1.5hrs/week

3. Examination Scheme:

a. Internal Assessment
i. Theory: 20 marks
ii. Practical: 50 marks

b. Final Examination
i. Theory: 80 marks

c. Total: 150 marks

4. Internal marks evaluation:

Theory (20) Practical (50)

A &CP GD & P TP & MT PBE A & CP RS & P PE & MCQs
25% 25% 25% 25% 20% 40% 40%

5. Abbreviations:
a. A & CP: Attendance & Class Participation
b. GD & P: Group Discussion & Presentation
c. TP & MT: Test Papers & Mid Term
d. PBE : Pre Board Examination
e. RS & P: Report Submission & Presentation
f. PE & MCQs: Practical Exam & Multiple Choice Questions

6. Adjustments:

a. 80-100% accomplishment: Full marks (100%)

b. 60-80% accomplishment: Semi marks (80%)
c. 40-60% accomplishment: Fair marks (60%)

7. Ratings:

a. ***: Distinction with 100% marks.

2

b. ** : First Division with up to 80% marks.

Page

c. * : Fair with up to 60% marks.

Prepared By
Er. Shree Krishna Khadka
8. Final Examination:

a. Board Exam:

i. Full marks: 80
ii. Question Pattern: Generally consists of two groups

BIT BE C/C & E

Pattern
Group A Group B Group A Group B
Question type Long Short Long Long/Program
Total question 3 10 7 3
Attempt 2 8 6 2
Marks/each 12 7 10 10
Total 24 56 60 20

b. For BIT: Sometimes, one extra paper with 20 objectives questions can be asked
each carrying 1 marks for twenty minutes as a Group A. In that case, long answer
type questions weigh 10 marks each and short type questions weigh 5 marks for
each.
For BE: Choices may vary from more than one questions in each groups.

c. Lab Exam

i. Of about 12 exercises on lab

- As per the syllabus, selected topics will be exercised.
ii. Programming
- Programming is done using C/C++ language in Visual Basic platform.
iii. Submission of lab report is compulsory.
iv. Practical Exam
1. Programming test.
2. Viva/MCQs test.
3
Page

Prepared By
Er. Shree Krishna Khadka
Introduction

1. Why to study Numerical Methods?

a. Historical Review:
- In early age of about 3000 BC, computing starts with a device called ABACUS
by Chinese society, which was used for arithmetic operation only.
- Slide rule by John Napier to compute a logarithmic problems in early 16th
century.
- An accounting machine by Blaise Pascal called Pascal Calculator.
- Data storing facility at punched card by Loom Jacquard in 18th century.
- 18th Century: Revolutionary invention of Difference Engine & Analytical
Engine by Charles Babbage.
- Later on till the date, all advanced/most recent computers are based on that
principle.
b. Present Context:

Now-a-days, computers are great tools in numerical computation, which plays

an indispensable role in solving real-life mathematical, physical and engineering
problems.

However, without fundamental understanding of engineering problems, they

will be useless.

2. How is the engineering system understood?

- By observation and experiment.
- Theoretical analysis and generalization.

Problem Definition

Theory Mathematical Model Data

Problem solving tools:

Computers, Statistics, Numerical
Methods, Graphics etc.

Numerical/Graphical
Result

Validation by:
Societal interlaces, Scheduling,
Optimization, Communication &
Public Interaction

Implementation/
Application
4
Page

Fig: Computing Process

Prepared By
Er. Shree Krishna Khadka
3. Example:
Newton’s Second Law of Motion
1. Theory:
The time rate of change of momentum of a body is equal to the resulting force acting
on it.

2. Mathematical Model:

Where, F = Net force acting on the body.

m = Mass of the object
u = Initial velocity of the object Data
v = Final velocity of the object
t = time and a = acceleration

3. Analysis:
It characterizes the typical mathematical model of the physical world as:
a. Natural process or system in mathematical term.
b. Idealization and simplification of reality.
c. Yields reproducible results.

4. Solution:
Manipulation of mathematical model gives rise to the solution of the problem.

Analytical Vs. Numerical Solution

Analytical solution is like a ‘Math: a great example’, as we have been learning all the
way. Numerical solution is computed via computer with some approximation but
exact result.
For example, the quadratic equation ax2+bx+c = 0 has:
√
a) Analytic solution: ; which works for any set of values of a,
b & c. The real solution exists only if , i.e. the properties of the
solution are transparent.
b) Numerical solution: Can only deal with a given set (a, b, c) at a given time.
Solution is approximate. So, error estimation is needed.
We want the computer to do repeated search for the solution, but not blindly. A
clear set of rules (steps) has to be designed based on the sound mathematical
reasoning, which guarantee a solution at a desirable level of accuracy.

5. Validation & Implementation

Now, numerical method is a tool that deals with the mathematical model
(formulated to describe the theory) to give a valid numerical or graphical result.

This example is just a prototype; most of the real-world problems (99.9%) are
5

complex in nature but can be approximately solved with numerical computation.

Page

Prepared By
Er. Shree Krishna Khadka
4. Summary:

a. Science and engineering study convert physical phenomena into mathematical

model often with complex-analysis and calculation.
b. Numerical analysis is the study of procedure for solving the problem with
computer.
c. It is the study of algorithms and further problems of mathematics.
d. Numerical analysis is always numerical.
e. Goal of numerical analysis is the design and analysis of techniques to give
approximate but accurate solution to hard problems.

5. Applications:

a. Numerical weather forecasting.

b. Computing transactory of space-craft (path-of-projectile).
c. In super computers.
d. In improvement of car safety to avoid accidents.

6
Page

Prepared By
Er. Shree Krishna Khadka
Errors in Numerical Calculation

Contents:

o Numbers & Their Accuracy

o Different Types of Errors
o Mathematical Preliminaries
o General Error Formula
o Convergence
o Assignment 1
7
Page

Prepared By
Er. Shree Krishna Khadka
NUMBERS & THEIR ACCURACY
There are two kinds of numbers:
i) Exact numbers, e.g. 1, 2, 3 …1/2, 5/2 etc.
ii) Approximate numbers, e.g. PI, K … etc

o Significant Digit:
Digits that are used to express a number are called significant digits/figures. The
following statements describe the notion of significant digits.
- All non-zero digits.
- All zeroes occurring between non- zero digits.
- Trailing zeroes following a decimal point, e.g. 3.50, 65.0, 0.230 have three
significant digits.
- Zeroes between the decimal point are preceding a non-zero digits are not
significant, e.g. 0.01234, 0.001234, 0.0001234 all have four significant digits.
- When the decimal point is not written, trailing zeroes are not considered to be
significant, e.g. 4500 contain only two significant digits.

o Accuracy & precision

The concept of accuracy and precision are closely related to significant digits.
- Accuracy refers to the number of significant digits in a value. E.g. the number
57.345 is accurate to five significant digits.
- Precision refers to the number of decimal position, i.e. the order of magnitude of
last digit in a value. E.g. 57.396 have a precision of 0.001.
In numerical computation, we come across numbers which have large numbers of
digits and it will be necessary to cut them to usable numbers of figures. This
process is known as rounding off. The error caused due to cut-off a large number
into usable number of figure is called round-off error.

o Rules for rounding off:

To round off a number to n-significant digits, discard all digits to the right of the nth
digit if this discarded number is:
- Less than half a unit in the nth place, leave the nth digit unaltered.
- Greater than half a unit in nth place, increase the nth digit by unity.
- Exactly half a unit in the nth place, increase the nth digit by unity if it is odd;
otherwise leave it unchanged.
Examples: Following numbers are rounded off to four significant figures.

1.6583(1.658), 30.0567(30.06), 0.859378(0.8594), 3.14159(3.142)

8
Page

Prepared By
Er. Shree Krishna Khadka
DIFFERENT TYPES OF ERROR:

a) Absolute Error: Numerical difference between true value of a quantity and its
approximate value.

Mathematically;

b) Relative error: Ratio of absolute error to a true value of that quantity being
concerned.

Mathematically;

c) Percentage Error: The percentage value of relative error.

Mathematically;

d) Truncation Error: Occurs due to truncation or terminating an infinite sequence

of operation after a finite number have been performed.

Example:

e) Relative Accuracy: Ratio of change in true value to the modulus of true value.

Mathematically;
9
Page

Prepared By
Er. Shree Krishna Khadka
NUMERICAL EXAMPLES:

 An approximate value of PI is given as 3.1428571 and its true value is

3.1415926. Find EA and ER.

Solution:

EA =| X-X1| = |3.1415926-3.1428571| = 1.265*10-3

ER = EA/X = 1.265*10-3/3.1415926 = 4.025*10-4

 Three approximate values of the number 1/3 are given as 0.30, 0.33 and 0.34,
which of these is the best approximation?

Solution:

E1 = |(1/3)-0.30| = 1/30
E2 = |(1/3)-0.33| = 1/300
E3 = |(1/3)-0.34| = 1/150

Since, E2<E3<E1, 0.33 is the best approximation.

 Evaluate the sum S=_/3 + _/5 + _/7 to four significant digits and find its absolute
and relative error.

Solution:

_/3 = 1.73205 = 1.732

_/5 = 2.23606 = 2.236
_/7 = 2.64574 = 2.646

S = 1.732 + 2.236 + 2.646 = 6.614

Absolute Error (EA) = 0.0005 + 0.0005 + 0.0005 = 0.0015

Relative Error (ER) = EA/S = 0.0015/6.614 = 2.268*10-4 = 0.0002
10
Page

Prepared By
Er. Shree Krishna Khadka
MATHEMATICAL PRELIMINARIES

Theorem 1: If f(x) is continuous in a , and if f(a) and f(b) are of opposite

signs, then for at least one number , such that .

Theorem 2 (Rolle’s Theorem): If f(x) is continuous in a , exists in

and f(a)=f(b)=0, then there exist at least one value of x say , such that
.

Generalization: If f(x) be a function, which is n-times differentiable on [a, . If f(x)

vanishes at the (n+1) distinct points in (a,b), then there exist a
number in (a,b) such that , .

Theorem 3 (Intermediate Value Theorem): If f(x) be continuous in [a, b] and let K

be any number between f(a) and f(b), then there exists a number in (a, b) such
f( .

Theorem 4 (Mean Value Theorem for Derivative): If f(x) is continuous in [a, b]

and f’(x) exists in (a, b), there exists at least one value of x, say between a and b
such that:

Theorem 5 (Taylor’s Series for a function of one variable): If f(x) is continuous

and passes continuous derivates of order in an interval and includes x=a, then:

Where; .

Theorem 6 (McLaren’s Expansion):

Theorem 7 (Taylor’s Series for 2-variables):

( ) ( )
11
Page

Prepared By
Er. Shree Krishna Khadka
 Theorem 8 (Taylor’s Series for several variables):

( )

( )

{ }

GENERAL ERROR FORMULA

Formula for the error committed in using a certain formula or a functional relation.
Let; u=f(x1, x2, x3, x4, x5, … xn) be the function of several variable xi (i=1,2…n) and let the
error in each xi be xi , then the error u in u is given by:

….(i)

On expansion of RHS using Taylor’s Series, we get:

( )

( )

{ }

i.e. ( ) ∑

So, ignoring powers of errors, we get:

∑ … (ii)

Now, Relative Error: … (iii)

12
Page

Prepared By
Er. Shree Krishna Khadka
CONVERGENCE

- Numerical computing is based on the idea of iterative process.

- Iterative processes involve generation of sequence of approximation with the
hope that, the process will end of the required solution.
- Certain methods convert faster than others.
- It is necessary to know that convergence rate of any method to get the required
solution.
- Rapid convergent take less execution time.

13
Page

Prepared By
Er. Shree Krishna Khadka
ASSIGNMENT 1

Full Marks: 20
Pass Marks: 10

Que: 1 (3X2=6 Marks)

1.1) Round off the following numbers to two decimal places.

2.3742, 81.255, 52.275, 48.21416

1.2) Round off the following numbers to four significant digits.

0.70029, 0.00022218, 2.36425, 38. 46235

1.3) Calculate (_/102 - _/101) correct to four significant digits.

Que: 2 (2X2=4 Marks)

2.1) if u = 3v7-6v, find the percentage error in u at v=1, if the error in v is 0.05.

2.2) if y = (0.31x+2.73)/(x+0.35), where the coefficients are rounded off; find the
absolute and relative error in y when x= .

Que: 3 (3X2=6 Marks)

If
find the maximum absolute error in:
a) a+b+c+d
b) a+5c-d
c) c3

Que: 4 (1X4=4 Marks)

If u=5xy2/z3,
&
;

Find the maximum value of relative error.

14

(Hint: Use general error formula)

Page

Prepared By
Er. Shree Krishna Khadka
Solution of Algebric &
Transcendental Equations

Contents:

o Introduction
o Evaluation of Polynomials by Horner’s Method
o Methods of solving non linear equations
o Bracketing Methods
o Bisection Method
o Regular Falsi Method
o Open End Methods
o Newton Raphson Method
o Secant Method
o Fixed Point Iteration Method
o Horner’s Method of Finding Root
o Assignment 2
15
Page

Prepared By
Er. Shree Krishna Khadka
INTRODUCTION

Wide varieties of problems in science and engineering can be formulated into

equation of the form f(x) = 0. The solution process involves finding the value of x
that would satisfy the equation f(x) = 0. These values are called roots of the
equation.

o Algebric Function: Any equation of type y = f(x) is said to be algebraic if it

can be expressed in the form: f(x, y) = 0, i.e. the function is dependent
between the variables x and y.
e.g. 3x+5y-21 = 0; 2x+3xy-25=0; x3-xy-3y3 = 0

o Transcendental Equation: A non-algebraic equation is called a

transcendental equation, which includes trigonometric, exponential and
logarithmic functions.
e.g. 2sinx-x = 0; excosx-0.5x = 0; logx2-1 = 0
These may have a finite or infinite number of real roots or may not have real
root at all.

o Polynomial Equation: These are also a simple class of algebraic equation,

represented by:
an(xn)+ an-1(xn-1)+ an-2(xn-2)+ … + a1(x1)+ a0
This is called nth degree polynomial and has ‘n’ number of roots. The roots
may be:

i) Real and different; e.g. x2 + 4x -5 =0

ii) Real and repeated; e.g. x2 – 2x +1 = 0
iii) Complex numbers; e.g. x2 – x + 1 = 0

EVALUATION OF POLYNOMIALS BY HORNER’S METHOD

All iterative methods require the evaluation of function for which solution is sought.
Since, it is a recurring task; the design of an efficient algorithm for evaluating the
function assumes a greater importance.
As: ∑ ∑

Now, let us consider the evaluation of a polynomial using Horner’s Rule as:
16

( )
Page

Prepared By
Er. Shree Krishna Khadka
Where; innermost expression is evaluated first. Hence, the resulting value
constitutes a multiplicand for the expression at the next level. In this way, the number
level will be n for nth-order polynomial and consists of total n-addition and n-
multiplication.

Algorithm:

Pn = an
Pn-1 = Pn(x) + an-1
Pn-2 = Pn-1(x) + an-2
………………………………………………
………………………………………………
Pj = Pj+1(x) + aj
………………………………………………
………………………………………………
P1 = P2(x) + a1
P0 = P1(x) + a0 = f(x)

Example:

Evaluate the polynomial: f(x) = x3 - 4x2 + x + 6 using Horner’s Rule at x = 2.

Solution:

n = 3, a3 = 1, a2 = -4, a1 = 1, and a0 = 6

Using algorithm:

P3 = a3 = 1
P2 = P3(x) + a2 = 1(2) + (-4) = -2
P1 = P2(x) + a1 = (-2)(2) +1 = -3
P0 = P1(x) + a0 = (-3)(2) +6 = 0 = f(x=0)

o Linear and Non Linear Equations: If the dependent variable changes in

exact proportion to the changes in independent variables, then the function is
linear.
e.g. y = f(x) = x + 5;
Where, y – dependent parameter and x- independent parameter

If the response of the dependent variable is not in direct or exact proportion

to the changes in independent variable, then the function is non linear.
e.g. y = f(x) = x2 + 1;
17
Page

Prepared By
Er. Shree Krishna Khadka
METHOD OF SOLVING NON-LINEAR EQUATIONS

There are different methods to deal with non-linear equations:

i) Direct analytical method
ii) Graphical method
iii) Trial and error method
iv) Iterative method

We only deal with iterative method.

1. Iterative Methods

Iterative methods begin with one or more guesses at the solution being sought. It
generates a sequence of estimate of the solution which is expected to converge at
the solution. Based on the number of guess, we have two different methods.

A) Bracketing Method: It starts with two initial guesses that brackets the root
and then, systematically reduced the width of bracket until the solution is
reached. Example: Bisection Method, False Position Method etc.

B) Open End Method: It uses single starting values or two values that do not
necessarily bracket the root. Example: Newton Raphson Method with one
initial guess, Secant Method with two starting guesses, Fixed Point Iteration
Method etc.

BISECTION METHOD

This is one of the bracketing methods and is also known as Bolzano Method, Binary
Chopping or Half Interval Method.

F(x)

xn
x
xm xp

Fig: Bisection Method

Let f(x) be the function defined in the interval [xn, xp], where f(xn) and f(xp) are of
opposite sign i.e.[ f(xn)*f(xp)] < 0, then the root of the function lies in between xn and
18

xp.
Page

Prepared By
Er. Shree Krishna Khadka
 Algorithm:
1. Read xn and xp , define stopping criteria: EPS
2. Calculate, xm = (xn+xp)/2
3. If |(xp – xn)/xn| < EPS, go to step 6 (Optional)
4. Check for f(xm)
a. If f(xm) = 0, goto step 6
b. F(xm) < 0, xn = xm
c. F(xm) > 0, xp = xm
5. Repeat steps 2, 3 and 4, until EPS is achieved.
6. Print xm
7. Stop

Example:
Find the root of the equation x3-2x-5, using bisection method.

Solution:
x -5 -4 -3 -2 -1 0 1 2 3 4 5
f(x) -120 -61 -26 -9 -6 -5 -6 -1 16 51 110

Here, the sign changes occur, when the value of x shifts from +2 to +3. So, xn = 2
and xp = 3

n xn xp xmF(xm)
1 2 3 2.5 Positive
2 2 2.5 2.25 Positive
3 2 2.25 2.125 Positive
4 2 2.125 2.0625 Negative
5 2.0625 2.125 2.09375 Negative
6 2.09375 2.125 2.109375 Positive
7 2.09375 2.109375 2.1015625 Positive
8 2.09375 2.1015625 2.09765625 Positive

Since, it has not mentioned the value of stopping criteria, so for our convenient
stopping criteria, EPS = 0.005.
Since, xp = 2.1015625, xn = 2.09375, i.e. EPS = 0.003
Hence, the root of the equation is xm = 2.09765625 and with rounded off to four
significant digits, xm = 2.0976

/Home Work/
Find the root of the equation x(ex ) = 1, correct to three decimal places using Bisection
Method. Ans: 0.567
19
Page

Prepared By
Er. Shree Krishna Khadka
 Convergence of Bisection Method:
In Bisection Method, we choose a mid point xm in the interval between xp and xn.
Depending upon the sign of f(xm), xp or xn is set equal to xm such that the root lies In
the interval. In other case the interval containing the root is reduced by a function of 2.
The same process is repeated for the new interval. If the procedure is repeated ‘n’
times, then the interval containing the root is reduced to the size:

After iteration the root must lies within of our estimate, this mean the error
bound at nth iteration is:
.
Similarly,

Thus, the error decreases linearly with each step by a factor of 0.5.This method is therefore
linearly convergent.

REGULAR FALSI METHOD/FALSE POSITION METHOD

(b, f(b))
f(x)

a
x
(c,0) b

(a, f(a))

Fig: Regular Falsi Method

Suppose, we are given f(x) = 0. By the method of ‘Falsi Position’, the geometrical
interpretation of is given by two point formula as follows:

i.e.

Or,

Or,

Or,
20
Page

Prepared By
Er. Shree Krishna Khadka
 Algorithm:
1. Find a and b such that f(a)*f(b) < 0.

2. Compute,

3. If f(c) = 0; c is the required root, stop. If f(c) , then go to next step.

4. If f(c) < 0; then root lies between (c, b) otherwise root lies between (a, c)
5. Repeat steps 2, 3 and 4 until the root is found to the desired degree of
accuracy.

Example:
Find the root of the equation (x)log(x)=1; which lies in between 2 and 3 correct to
3 decimal places using Regular Falsi Method.
Solution:
We have: f(x) = (x)log(x)-1

As given: a = 2 and b = 3, i.e. the starting values. Now tabulating the data as below:

N a f(a) b f(b) c f(c)

1 2 -0.39794 3 0.43136 2.47985 -0.02188
2 2.47985 -0.02188 3 0.43136 2.50496 -0.00102
3 2.50496 -0.00102 3 0.43136 2.50612 -0.000053
4 2.50612 0.000053 3 0.43136 2.50618 -0.000003

Hence, the root of the equation is: 2.506 (correct to 3-decimal places).

Where; and when; f(c) < 0, a=c otherwise, b = c;

Convergence of Regular Falsi Method:

In the false position iteration, one of the starting points is fixed while the other moves
towards the solution. Assume that the initial points bracketing the solution are ‘a’ and
‘b’, where a moves towards the solution and ‘b’ is fixed as illustrated in following
figure.
21
Page

Prepared By
Er. Shree Krishna Khadka
 (b, f(b))
f(x)

x1 x2 x3
xr x
b
(d, f(d))

(c, f(c))
(a, f(a))

Fig: Convergence of Regular Falsi

Method

Let x1 = ‘a’ and ‘xr‘ be the solution. Then:

e1 = xr – x1
e2 = xr – x2
………………………………………….

………………………………………….
That is:
ei = xr – xi

In general:
ei+1 = (Ref: Go online via. Google Search)

NEWTON RAPHSON METHOD

Consider a graph f(x) as shown in the above figure. Let us assume that x1 is an
approximate root of f(x) = 0. Draw a tangent at the curve f(x) at x=x1 as shown in the
figure.
(x1, f(x1))

(x2, f(x2))
f(x)

xr
x
x3 x2 x1
22

Fig: Newton Raphson Method

Page

Prepared By
Er. Shree Krishna Khadka
The point of intersection of this tangent with the x-axis gives the second
approximation to the root. Let, the point of intersection be x2. The slope of the tangent
is given by:

Where; f’(x1) is the slope of f(x) at x=x1.

Solving for x2, we get:

This is called Newton Raphson Formula for first iteration. Then the next
approximation will be:

In general;

Algorithm:
1. Assign an initial value to x, say x0 and read the stopping criteria (EPS).
2. Evaluate f(x0) and f’(x0). If f(x0) = 0; then the root will be x0.
3. Find the improved estimate of x0 a:

4. Check for accuracy of estimate: if | | ; then stop and go to step 6

otherwise continue.
5. Replace x0 by x1 and repeat steps 3 & 4.
6. Print x1.
7. Stop.

Example:
Find the root of the equation f(x) = x2 – 3x + 2 in the vicinity of x = 0 using
Newton Raphson Method correct to 4-decimal places.
Solution:
23
Page

Prepared By
Er. Shree Krishna Khadka
Here, initial estimate of the root (x0) = 0;
We have, f(x) = x2 – 3x + 2, So, f’(x) = 2x – 3

At x0 = 0; f(x0 = 0) = 2 and f’(x0 = 0) = -3

Now, for the better approximation, we have the tabular data as follow:

n x0 f(x0) f'(x0) x1
0 0 2 -3 0.6667
1 0.6667 0.4444 -1.6667 0.9333
2 0.9333 0.0711 -1.1333 0.9961
3 0.9961 0.0032 -1.0078 0.9961

Hence, the root is: 0.9961 (correct to 4-decimal places)

Where: and in each iteration, x0 = x1.

Derivation of Newton Raphson Formula by Taylor’s Series Expansion:

Assume that: xn is an estimate root of the function f(x) = 0 and consider a small
interval ‘h’ such that:

h = xn+1 - xn …… (i)

Using Taylor’s Series Expansion for f(xn+1), we have:

Neglecting the terms containing higher derivatives, we get:

…… (ii)

If xn+1 is the root of f(x), then: f(xn+1) = 0. Putting this value in equation (ii) it gives:

i.e.

i.e.
24

i.e.
Page

Prepared By
Er. Shree Krishna Khadka
Special cases of Newton Raphson Method:
Case 1: Find the solution to compute the sequence root of any positive integer ‘N’.
Solution:
As per the case: √
i.e. √
i.e.
i.e. f(x) = and f’(x) =

If x0 be the initial estimation of root, then:

i.e. { }

In general:
{ }

Case 2: Show that the Pth root of any given positive integer ‘N’ is:

{ }

Solution:

As per the case: √

i.e.
i.e. f(x) = and f’(x) =

If x0 be the initial estimation of root, then:

i.e. { }

In general:
{ }
25
Page

Prepared By
Er. Shree Krishna Khadka
Case 3: Show that the reciprocal of the Pth root of any positive number ‘N’ is:

Solution:
As per the case:
√
i.e.
i.e. f(x) = and f’(x) =

If x0 be the initial estimation of root, then:

i.e.

In general:

Case 4: Derive the general equation to find the value of reciprocal of any positive
integer ‘N’.

Solution:

As per the case:

i.e.
i.e. f(x) = and f’(x) =

If x0 be the initial estimation of root, then:

i.e.

In general:
26
Page

Prepared By
Er. Shree Krishna Khadka
Convergence of Newton Raphson Method:
Let xn be the estimate to the root of the function f(x)=0. If xn and xn+1 are close to each
other, then using Taylor’s Series Expansion:

(i)

Let us assume that, the exact root of f(x) be xr, then: xn+1 = xr, then: f(xn+1)=0. So,
equation (i) becomes:

… (ii)

As we have:

Putting the value of f(xn) into equation (ii), we get:

i.e. … (iii)

We know that, the error in the estimate xn+1 is:

Similarly,

Now, neglecting the higher power terms and expressing equation (iii) in terms of error:

Rearranging the term, we get:

This shows that error is roughly proportional to the square of the error at previous
iteration. Therefore, Newton Raphson method has quadratic convergence.

Limitation of Newton Raphson Method:

1. Division of zero may occur when f’(x) = 0.

2. If the initial guess is too far away from the required root, the process may
converge to some other root.
27
Page

Prepared By
Er. Shree Krishna Khadka
SECANT METHOD
Secant method uses two initial estimates but does not required that they must bracket
the root. It can use two points x1 and x2 as shown in the figure below as starting values
although they do not bracket the root.

X1, f(x1)
f(x)
X2, f(x2)

x3
x
x2 x1

Fig: Secant Method

Slope of secant line passing through x1 and x2 is:

On rearranging the terms, we get:

…. (i)

On adding and subtracting f(x2)x2 at the numerator, equation (i) can be

represented in the form:

… (ii)

Hence, the approximate value of the root can be refined by repeating the process
to the desired level of accuracy. So, in general:

…. (iii)

Algorithm:
1. Decide the initial points: x1 and x2 and EPS.
2. Compute: f1 = f(x1) and f2 = f(x2)

3. Compute:

4. Test for accuracy of x3: if | | then:

Set (x1 = x2, f1 = f2) and set (x2 = x3, f2 = f(x3)); go to step 3, otherwise set root =
x3.
28

5. Print root
6. Stop
Page

Prepared By
Er. Shree Krishna Khadka
Example
Use Secant Method to estimate the root of the equation x2 – 4x -10 = 0, with
initial estimate x1 = 4 and x2 = 2.

Solution:
x1 = 4, f(x1) = f1 = 42 – 4*4 – 10 = -10
x2 = 2, f(x1) = f1 = 22 – 4*2 – 10 = -14

Let EPS = 0.05

For the better estimation, we have the tabular data as below:

n x1 f1 x2 f2 x3 f3 EPS
0 4 -10 2 -14 9 35 0.7778
1 2 -14 9 35 4 -10 1.2500
2 9 35 4 -10 5.1111 -4.321 0.2174
3 4 -10 5.1111 -4.321 5.9565 1.6539 0.1419
4 5.1111 -4.321 5.9565 1.6593 5.7225 -0.1429 0.0409

Hence, the root is: 5.7225, which is less than EPS = 0.05
Where EPS =| | and in each iteration: x1 = x2, f1 = f2 and x2 = x3, f2 = f(x3).

Comparison of Secant Iterative Formula with Newton Raphson Formula:

We have, Newton Raphson Formula:

…. (i)

And Secant Iterative Formula:

… (ii)

Comparing equation (i) and (ii):

Hence, the major advantage of Secant Method of iteration is no need to evaluate

29

the derivative.
Page

Prepared By
Er. Shree Krishna Khadka
Convergence of Secant Method:
We have, Secant Iterative Formula:

Or … (i)

If xr be the actual root fo f(x) and ei be the error in estimate of xi, then:
xn+1 = en+1 + xr
xn = en + xr
xn-1 = en-1 + xr
Now, rewriting equation (i) in terms of error with these values, we get:

… (ii)

According to Mean Value Theorem, if x = Rn lie in the interval xn and xr, then:

…. (iii)

As obvious: f(xr) = 0 and xn – xr = en. Therefore, equation (ii) becomes:

Similarly:
Substituting these values in the numerator of equation (ii), we get:

.. (iv)

i.e. … (v)
As we also know that, if the order of convergence of an iterative process is P, then:
i.e. … (vi)
Now, from equation (iv) and (v):
… (vii)
Finally, from equation (vi) and (vii):

So, the order of convergence of Secant Method is 1.618, which is a slow rate (<2). It
also requires previous two iterative to estimate the new one. These are recognized
30

as drawbacks of Secant Method.

Page

Prepared By
Er. Shree Krishna Khadka
ITERATION METHOD:
It is also known as, direct substitution method or method of fixed point iteration or
method of successive approximation. It is applicable if the equation f(x) = 0 can be
expressed in terms of x = g(x). If x0 is the initial approximation to the root, then the
next approximation will be: x1 = g(x0) and the next again will be: x2 = g(x1) and so on.
In general: xn+1 = g(xn)
Where, x = g(x) is known as fixed point equation.

Algorithm:
1. Read an initial guess: x0 and EPS
2. Evaluate: x1 = g(x0)
3. Check for error: if | | EPS, x1 is root, else continue.
4. Assign the value of x1 to x0, i.e. x0 = x1 and repeat steps 2 & 3.
5. Print x1.
6. Stop

Example:
Evaluate the square root of 5 using Fixed Point Iterating method.
Solution:
As, √
Let, x0 = 1, then:
g(x0) = 5/1 = 5 = x1
g(x1) = 5/5 = 1 = x2
g(x2) = 5/1 = 5 = x3
g(x3) = 5/5 = 1 = x4 and so on.
Here, the solution does not converge to the solution, this type of divergence is
known as oscillatory divergence.

Let’s try another form:

Adding x on both sides of , we get: =g(x)
Let x0 = 0, then:
g(x0) = -5 = x1
g(x1) = (-5)2+(-5)-5=15 = x2
g(x3) = 152+15-5=235 = x4
g(x4) = 2352+235-5=55455= x5 and so on.
This is also divergent i.e. does not converge to the solution. This type of divergence
is known as monotonic divergence.

Again Let’s try another form.

31
Page

Prepared By
Er. Shree Krishna Khadka
As x = 5/x can also be written in the form: ( )
For x0 = 1,
g(x0) = 3 = x1
g(x1) = 2.3333 = x2
g(x2) = 2.2381 = x3
g(x3) = 2.2361 = x4
g(x4) = 2.2361 = x5
Now, the process converges rapidly to the solution. Hence, the root of 5 is 2.2361.

Convergence of Iteration Method:

The iteration formula is:

….. (i)

If xf be the real root of the equation, then:

….. (ii)

Substituting equation (i) from equation (ii), we get:

…. (iii)

According to mean value theorem: if x = R lies in the interval between xf and xi ,

then:

… (iv)

Now, from equation (iii) and (iv)

…. (v)

If ei is the error in ith iteration then equation (v) in terms of error will be:

…. (vi)

This shows that, the error will decrease with each iteration iff g’(R) < 1. Hence, the
Iteration Method converges only when |g’(x)|<1.
32
Page

Prepared By
Er. Shree Krishna Khadka
HORNER’S METHOD FOR FINDING ROOTS
- Historic name for synthetic division is Horner’s Method.
- Let positive root of f(x) = 0 lies in between and , where is an integer.
Then the value of the root will be , where d1, d2 … are digits in their
decimal parts.
- To find the value of d1, first diminish the roots of f(x) = 0 by . For example, if
root lies in between 0 and 1 then root will be 0. . And make the
transformed equation.
- Multiply the root of the equation by the order of 10 and diminish other decimal
parts and so on.
- Continue the process to obtain the root of the equation at any desired degree of
accuracy by digit.

Example:
Find the root of the equation x3 + 9x2 – 18 = 0 to two decimal digit using
Horner’s Method.
Solution:
As: x3 + 9x2 – 18 = 0 =f1 (x) … (i)
So, f(x1)*f(x2) < 0 is analysed using following table.

x -5 -4 -3 -2 -1 0 1 2 3 4 5
f(x) 82 62 36 10 -10 -18 -8 26 90 190 332

So, root lies in between 1 and 2 as f(x) shifted from negative to positive at these
values. Hence, the integer part of the root is: 1. d1d2 up to two decimal places.

Now, we are going to diminish the root of the equation by 1 using synthetic
division as follow.

Coeff. x3 x2 x1 x0
1) 1 9 0 -18
0 1 10 10
1) 1 10 10 (-8)
0 1 11
1) 1 11 (21)
0 1
(1) (12)

Now, the transformed equation is: x3 + 12x2 + 21x – 8 = 0. Multiplying the

coefficients by the order of 10, we get: x3 + 120x2 + 2100x – 8000 = 0 = f2(x)... (ii)
33

Again finding d1 by following table:

Page

Prepared By
Er. Shree Krishna Khadka
 x -4 -3 -2 -1 0 1 2 3 4
f(x) -14544 -13247 -11728 -9981 -8000 -5779 -3312 -593 2384

Here, f2(x=3) * f2 (x=4) < 0. So, d1 must be 3. Hence, the integer part updated to
1.3d2. Now again, diminishing the root by 3 using synthetic division.

Coeff. x3 x2 x1 x0
3) 1 120 2100 -8000
0 3 369 10
3) 1 123 2469 (-593)
0 3 378
3) 1 126 (2847)
0 3
(1) (129)

Again the transformed equation is: x3 + 129x2 + 2847x – 593 = 0. Again,

multiplying the coefficients by the order of 10, we get: x3 + 1290x2 + 284700x –
593000 = 0 = f3(x)… (iii)

Finally, for d2, we have the following table:

X -3 -2 -1 0 1 2 3
f(x) -1435517 -1157248 -876411 -593000 -307009 -18432 272737

Here f3(x=2) * f3(x=3) < 0. So, d3 must be 2.

Finally, the root of the equation f(x) = 0 is: 1.32

34
Page

Prepared By
Er. Shree Krishna Khadka
Assignment: 2
Full Marks: 50
Pass Marks: 25

1. (a) Evaluate the polynomial by

Horner’s Method at x=1.5 [5]

(b) Use Fixed Point Iteration Method to evaluate: , correct

to four significant figures. [5]

2. Use Bisection Method to evaluate for :

(a) EPS = 0.05, [5]

(b) EPS = 0.001 and compare the result. [5]

3. Use False Position Method to evaluate correct to:

(a) 2 decimal places [5]

(b) 3 decimal places and compare the result. [5]

4. Use Newton Raphson Method to evaluate following function correct to 3

decimal places:

(a) [5]

(b) & find the value of [5]

√ √

5. Use Secant Method to evaluate correct to 3 decimal places:

(a) – & compare the output with the true value of x as 0.567143. [5]

(b) [5]
35
Page

Prepared By
Er. Shree Krishna Khadka
Interpolation

Contents:

o Introduction
o Errors in Polynomial Interpolation
o Interpolation with Equally Spaced Values
o Finite Difference Method
 Forward Difference Method
 Backward Difference Method
 Central Difference Method
o Difference of Polynomial
o Interpolation with Unequally Spaced Values
o Linear Interpolation Formula
o Lagrange Interpolation Formula
o Newton Interpolation Polynomial
o Inverse Interpolation
o Divided Differences
o Assignment 3
36
Page

Prepared By
Er. Shree Krishna Khadka
INTERPOLATION

If a function, say f(x) is constructed, such that it passes through all the set of data
points and then evaluating f(x) for the specified value of x is known as interpolation.
This method of constructing a function gives the estimation of values at non tabular
print. The functions are known as interpolation polynomials.

Most common form of nth order polynomials are:

1. … (i), this is known as nth order

polynomial in Power Form.

2. …(ii), this
is known as nth order polynomial in Shifted Power Form, and

3. ...(iii),
this is known as nth
order polynomial in Newton Form, which reduces to
shifted power form, when c1 = c2 = c3 = … = cn and reduces to simple power form,
when ci = 0 for all i.

ERRORS IN POLYNOMIAL INTERPOLATION

Let the function y(x), defined by the (n+1) points (xi, yi), i = 0, 1, 2 … n, be the
continuous and differentiable (n+1) times and let y(x) be approximated by a
polynomial of degree not exceeding ‘n’ such that:

i = 0, 1 … n

Then error is given by: .

Where, and if then:

… (i)

Since, y(x) is generally unknown and hence, we do not have any information
concerning y(n+1)(x). It is almost useless in practical computation. But, on the other
37

hand, it is extremely useful in theoretical work in different branches of numerical

analysis. It is useful in determining errors in Newton Interpolating Formula.
Page

Prepared By
Er. Shree Krishna Khadka
FINITE DIFFERENCES: INTERPOLATION WITH EQUALLY SPACED VALUES
Assume that, we have a table of values: (xi, yi), (i = 0, 1, 2, … n) of any function y =f(x),
the values of x being equally spaced, i.e. xi = x0 + ih, (i = 0, 1, 2, … n). The method of
finding the solution of f(x) for some ‘x’ in the range is based on the
concept of such finite differences.

Forward Difference:
If y0, y1, …. Yn denote a set of values of y at x0, x1, … xn respectively, then (y1 – y0),
(y2 – y1), …. (yn – yn-1) are called difference of y; i.e. , ,
….. , where is the forward difference operator. Usually, the
functions are tabulated at equal intervals, i.e. x1 – x0 = x2 – x1 = x3 – x2 …. xn – xn-1 =
na. Here, with tabulation at equal intervals a difference table for ‘n’ points may
be expressed as shown below:
x f(x) ….

….

…. ….
…. …. ….
…. …. …. ….
…. …. …. …. …. ….

Here ‘h’ is the uniform difference in the value of ‘x’.

, , ….
In general: ( , which is called first forward difference.
Similarly:

In general: ( , which is called second forward difference

and so on.
For the polynomial that passes through a group of equidistant point i.e. Newton
Gregory Forward Polynomial, we write in terms of the index ‘s’, such that s = (x-
x0)/h.

This equation for polynomial is called Newton Gregory Forward Interpolation

38

Formula. It is applied when the required point is close to the start of the table.
Page

Prepared By
Er. Shree Krishna Khadka
Example:
For the given table, find the value of f(0.16).
x 0.1 0.2 0.3 0.4
f(x) 1.005 1.020 1.045 1.081

Solution:
x0 = 0.1, x1 = 0.2 … i.e. h = x1 – x0 = 0.2 – 0.1 = 0.1
We have given, x = 0.16. So: s = (x – x0)/h = (0.16 – 0.1)/0.1 = 0.6

Now, generating the Forward Difference Table:

X f(x)

Now:

Backward Difference:
If the required point is close to the end of the table, we can use another formula
known as Newton Gregory Backward Difference Formula. Here, the reference
point is: ‘xn’ instead of x0. Therefore, s = (x – xn)/h.
So, the Newton Gregory Formula is given by:

The table for the backward difference will be identical to the forward difference
table. However, the reference point will be below the point for which the
estimate is required. So, the value of ‘s’ will be negative for backward
interpolation. The backward difference table is shown below.
39
Page

Prepared By
Er. Shree Krishna Khadka
 X f(x) ….

….

…. ….
…. …. ….
…. …. ….
…. …. ….
…. ….
…. ….

Where, is a backward difference operator, and

Here ‘h’ is the uniform difference in the value of ‘x’.

, , ….
In general: ( , which is called first backward difference.
Similarly:

In general: ( , which is called second backward difference

and so on.

Example:
Estimate the value of sin(45o) using backward difference method with the
following set of data.
x 10 20 30 40 50
f(x)=sin(x) 0.1736 0.342 0.5 0.6428 0.768

Solution:
x0 = 10, x1 = 20 … i.e. h = x1 – x0 = 20 – 10 = 10
We have given, x = 45. So: s = (x – xn)/h = (45 – 50)/10 = -0.5

Where, the necessary data are fetched from the following Backward Difference
Table.
40
Page

Prepared By
Er. Shree Krishna Khadka
Backward Difference Table:

x f(x)

0.0024

0.1252
50 0.768

Difference of a polynomial:
As, we have discussed difference table by forward and backward difference
method. Let us analyse, what actually differs in the first, second and … difference
in their degrees.

If y(x) be a polynomial of nth degree:

i.e; ,

Then we obtain:

[ [
[ ]
[

Where, a1’, a2’ … are new coefficients.

Then:
[ [
41

This shows that the first difference of polynomial of the nth degree is a
polynomial of degree (n-1) and so on.
Page

Prepared By
Er. Shree Krishna Khadka
Where;
Coefficient of xn-1 is a0(nh) in first difference.
Coefficient of xn-2 is a0n(n-1)h2 in second difference.
……………….. ….. … …………….. ………………………….
……………….. ….. …. …………….. …………………………..
Coefficient of x0 is a0n!hn in nth difference.

Newton’s Formula for Interpolation

We had used Newton Gregory Formula for two different methods:

a) Newton’s Gregory Formula for Forward Difference, and

Where, s = (x-x0)/h; ‘x’ is referred near to the start of the table and ‘h’ is a
interval constant

b) Newton’s Gregory Formula for Backward Difference

Where, s = (x-xn)/h; ‘x’ is referred near to the end of the table and ‘h’ is a
interval constant

Central Difference Method

If interpolation is desired near the beginning or end of the table, there is no
alternative to Newton’s Forward and Backward Difference Formula respectively,
simply because higher order differences do not exist at the beginning or end of
the table values.

For interpolation near the middle of the table, Stirling’s Formula gives the most
accurate result for and Bessel’s Formula is most efficient near ,
say . But in the case, where a series of calculations have to made, it
would be inconvenient to use both of these formulae and a choice must be made
between them. Everett’s Formula may be gainfully employed for the aforesaid
reason.
42
Page

Prepared By
Er. Shree Krishna Khadka
LINEAR INTERPOLATION
The simplex form of interpolation is to approximate two data points by straight line.
Suppose, we are given two points, [x1, f(x1)] and [x2, f(x2)], these two points can be
connected linearly as shown in the figure below.

f(x)

f(x2)

f(x)

f(x1)

x
x1 x x2

Fig: Representation of Linear

Interpolation Method

Using the concept of similar triangle, we have:

Solving for f(x), we get:

{ } … (i)

This equation is known as Linear Interpolation Formula, where,

represents the slope of the line. Comparing equation (i) with Newton form of
polynomial of first order given by: Pn(x) = a0 + a1(x-c1), we get:

C1 = x 1, a0 = f(x1) and a1 =

Hence, the co-efficient ‘a1’ represent the first derivative of the given function.
43
Page

Prepared By
Er. Shree Krishna Khadka
 Example:
Given the table of data:

x 1 2 3 4 5
f(x) 1 1.4142 1.7321 2 2.2361

Determine the square root of 2.5.

Solution:
Since, the given value 2.5 lie in between the point 2 and 3. So, x1 = 2 and x2 = 3
whereas x=2.5.
Now:

{ }

i.e. { }

i.e. { }

Again, if we choose, x1 = 2 and x2 = 4, then we get: f(2.5) = 1.5607

Since, the true value of square root of 2.5 from calculator is 1.5811 so, error
produced on each cases will be:

Error_1 = |1.5811-1.5732| = 0.0079

Error_2 = |1.5811-1.5607| = 0.0204

Here, Error_2 > Error_1. Hence, it can be concluded that: smaller the interval
between the interpolation data points, the better will be the approximation.

LAGRANGE’S INTERPOLATION POLYNOMIAL FORMULA:

Let (x0, f0), (x1, f1), (x2, f2) … (xn, fn) be the (n+1) points. To derive the formula for the
polynomial of degree ‘n’, let us consider a second order polynomial of the form:
… (i)
If (x0, f0), (x1, f1) and (x2, f2) are three interpolations then:

….. (ii)
44

Substituting the values of b1, b2 and b3 from equation (iii) to equation (i), we get:
Page

Prepared By
Er. Shree Krishna Khadka
i.e. … (iii)

Now, equation (iii) may be represented as:

… (iv)

Where:

Equation (iv) second degree Lagrange’s Polynomial. For Lagrange’s Polynomial of

nth degree, we should have given (n+1) interpolating points. Then the polynomial
will be:

… (iv)

This can be generally expressed as:

∑ … (v)

Where;

∏ ( ) … (vi)

Here, equation (v) is called Lagrange’s Interpolation Polynomial of degree ‘n’ and
then equation (vi) is called Lagrange’s Basic Polynomial.
45
Page

Prepared By
Er. Shree Krishna Khadka
Example:
Find the Lagrange’s Interpolation Polynomial to fit the following data. Use
polynomial to estimate the value of e1.5.

0 1 2 3
0 1 2 3
0 1.7183 6.3891 19.0855

Solution:

= 0 + 1.7183{ }+6.3891{ } { }

; This is the required polynomial.

Now:

e1.5 = P(1.5) + 1 = 3.3677 + 1 = 4.3677

Concept of Inverse Interpolation

For a given set of values of x and y, the process of finding the value of x for a certain
value of y is called inverse interpolation. When the values of x are at unequal
intervals, the way of performing inverse interpolation is by interchanging x and y in
Lagrange’s method.
Inverse interpolation is meaningful only if the function is single valued in the
interval. When the values of x are equally spaced, the method of successive
46

approximation is used.
Page

Prepared By
Er. Shree Krishna Khadka
Divided Differences
The Lagrange’s Interpolation Formula has the disadvantage that if another
interpolation point were added, then the interpolation coefficients li(x) will have to
be recomputed; we therefore seek an interpolation polynomial which has the
property that a polynomial of higher degree may be derived from it by simply
adding new terms.

Newton general interpolation formula is one such formula, which employs what are
called divided differences. It is our principle purpose to define sch differences and
discuss certain of their properties to obtain the basic formula.

We have Newton’s form of polynomial as:

In this polynomial, coefficients can be calculated using divided difference table as

given below.

i xi f(xi) IDD IIDD IIIDD IVDD

0 x0 f(x0)
f(x0, x1)
1 x1 f(x1) f(x0, x1, x2)
f(x1, x2) f(x0, x1, x2,x3)
2 x2 f(x2) f(x1, x2, x3) f(x0, x1, x2,x3,x4)…
f(x2, x3) f(x1, x2, x3, x4)
3 x3 f(x3) f(x2, x3, x4)
f(x3, x4)
4 x4 f(x4)
…

If (x0,y0), (x1,y1)…. (xn,yn) be the given (n+1) points, then the divided difference of
order 1, 2, … n are defined by the relation>

f(x0) = a0

f(x0, x1) = f(x1, x2) = …..

f(x0, x1, x2) = f(x0, x1, x2) = ….

47
Page

Prepared By
Er. Shree Krishna Khadka
 Example:
A table of polynomial is given below. Fit the polynomial and find the value of
f(x) at x = 2.5. Use divided difference table.

x -3 -1 0 3 5
f(x) -30 -22 -12 330 3458

Solution:

Let’s build up the divided difference table as follow:

i xi f(xi) IDD IIDD IIIDD IVDD

0 -3 -30
4
1 -1 -22 2
10 4
2 0 -12 26 5
114 44
3 3 330 290
1564
4 5 3458
…

From table, a0 = -30, a1 = 4, a2 = 2, a3 = 4 and a4 = 5

Therefore,
f(x) = -30+4(x+3)+2(x+3)(x+1)+4(x+3)(x+1)(x-0)+5(x+3)(x+1)(x-0)(x-3)
i.e. f(2.5)= -30+4(2.5+3)+2(2.5+3)(2.5+1)+4(2.5+3)(2.5+1)(2.5-0)+5(2.5+3)(2.5+1)(
2.5-0)( 2.5-3)

Example:2 (Home Work)

Given the following set of data points, obtain the table of divided differences.
Use the table to estimate the value of f(1.5)

i 0 1 2 3 4
xi 1 2 3 4 5
f(xi) 0 7 26 63 124

Solution:
a0 = 0, a1 = 7, a2 = 6, a3 = 1 and a4 = 0
48

f(1.5) = 2.375
Page

Prepared By
Er. Shree Krishna Khadka
Assignment: 3
Full Marks: 40
Pass Marks: 20

1. From the following table of values of x and f(x), determine

(a) f(0.23) [5]

(b) f(0.29) [5]

x 0.20 0.22 0.24 0.26 0.28 0.30

f(x) 1.6596 1.6698 1.6804 1.6912 1.7024 1.7139

2. Use linear interpolation method to calculate the square root of 4.5 from
following table.

(a) Take two initial values as 4 and 5 [5]

(b) Take two initial values as 3 and 6
Compare the result of (a) and (b).

x 1 2 3 4 5 6
f(x) 1 1.4142 1.7321 2 2.2361 2.4494

3. Applying Lagrange’s Interpolation Formula, find a cubic polynomial which

approximates the following data. [10]

x -2 -1 2 3
f(x) -12 -8 3 5

Also find the polynomial values at -2.5 and 2.5.

4. Given the table of values

x 50 52 54 56
f(x)= √ 3.684 3.732 3.779 3.825

Use Lagrange Interpolation to find x when √ [10]

(Example: Inverse Interpolation)

5. Given the following set of data points. Obtain the table of divided
difference and use that table to estimate the value of f(1.5), f(3.45) & [5]
(4.2)3.

x 1 2 3 4 5
f(x)=x3-1 0 7 26 63 124
49
Page

Prepared By
Er. Shree Krishna Khadka
`

Curve Fittings, B-Splines

& Approximations

Contents:

o Curve Fitting & Regression: Introduction

o Least Square Regression Technique
o Fitting Linear Equations
o Fitting Transcendental Equations
o Fitting Polynomial Equations
o Multiple Linear Equations
o Spline Interpolation: Introduction
o Cubic B-Splines
o Approximation of Functions
o Assignment 4
50
Page

Prepared By
Er. Shree Krishna Khadka
CURVE FITTING & REGRESSION: INTRODUCTION
Previously, we have discussed the methods of curve fitting for the data points of well
defined functions. In this topic, we will discuss on methods of curve fitting for
experimental data. In many applications it often becomes necessary to establish a
mathematical relationship between experimental values. The mathematical equation
can be used to predict values of the dependent variables. For example: we would like
to know the maintenance cost of equipment as a function of age or mileage.

The process of establishing such a relationship in the form of mathematical model is

known as regression analysis or curve fitting. In this topic, we shall discuss a
technique known as least square regression to fit the data under the following
situation.

a) Relation is linear
b) Relation is transcendental
c) Relation is polynomial

LEAST SQUARE REGRESSION

Let us consider the mathematical equation for a straight line: y = f(x) = a + bx, to
describe the data. We know that ‘a’ is the intercept of the line and ‘b’ is its slope.

f(x)
f(x) = a + bx
ai
(Xi, yi)

a
x

Fig: Least Square Regression

Consider a point (xi, yi) as shown in figure. The vertical distance of this point from the
line f(x) = a + bx is the error: ai. Then ai = yi – f(xi) = yi – a – bxi ….. (i)

The best approach that could be tried for fitting a best line through the data is to
minimize the sum of squares of errors.
i.e. ∑ ∑
The technique of minimizing the sum of square of errors is known as least square
regression.
51
Page

Prepared By
Er. Shree Krishna Khadka
 FITTING LINEAR EQUATIONS

Let, the sum of square of individual errors be as:

∑ ∑ … (i)

In this method of least square, we choose ‘a’ and ‘b’ such that Q is minimum. Since, Q
depends on ‘a’ and ‘b’, a necessary condition for Q to be minimum is given by:

&

Then:
∑ & ∑

i.e. ∑ ∑ …. (ii) & ∑ ∑ ∑ … (iii)

Equations, (ii) and (iii) are called normal equations, solving for ‘a’ and ‘b’, we get:

∑ ∑ ∑ ∑ ∑
∑ ∑
…. (iv) & ̅ ̅ … (v)

Example
Fit a straight line to the following set of data.

x 1 2 3 4 5
y 3 4 5 6 8

Solution:
The various summations are calculated as below:

xi yi
1 3 1 3
2 4 4 8
3 5 9 15
4 6 16 24
5 8 25 40
∑ = 15 ∑ = 26 ∑ = 55 ∑ = 90

Here, n = 5
52

On calculation using above formula, we get: b = 1.2 and a = 1.6

Therefore, the required linear equation is: y = a + bx = 1.6 + 1.2x
Page

Prepared By
Er. Shree Krishna Khadka
 FITTING TRANSCENDENTAL EQUATIONS
The relationship between dependent and independent variables is not always linear.
The non-linear relationship may exist in the form of transcendental equation. For
example, the familiar equation of population growth is given by:

… (i)

Where, ‘P0’ is the initial population, ‘K’ is the rate of growth and ‘t’ is time.

Another example is the non-linear model in which relation is established between

pressure and volume.

… (ii)

Where, ‘P’ is pressure and ‘v’ is volume. If we observe the values of ‘P’ for the various
values of ‘v’, we can then determine the parameter ‘a’ and ‘b’.

The problem of finding the values of ‘a’ and ‘b’ can be solved by using the algorithm
given for linear equation. Let us rewrite the equation using the conversion of variable
as ‘x’ and ‘y’. So, the equation becomes:

… (iii)

Taking log on both sides, we get:

… (iv)

Now, comparing equation (iv) with standard linear equation, y = a + bx, we found:

…. (v)

Then, obviously we will have the following results:

∑ ∑ ∑
∑ ∑
…. (vi) &

∑ ∑

∑ ∑
53

i.e. { } …. (vi)
Page

Prepared By
Er. Shree Krishna Khadka
Example:
Given the data table:

x 1 2 3 4 5
y 0.5 2 4.5 8 12.5

Fit a power function model of the form y = axb.

Solution:

xi yi ln(xi) ln(yi) ln(xi).ln(yi)

1 0.5 0 -0.6931 0 0
2 2 0.6931 0.6931 0.4804 0.4804
3 4.5 1.0986 1.5641 1.2069 1.6524
4 8 1.3863 2.0794 1.9218 2.8827
5 12.5 1.6094 2.5257 2.5901 4.0649
∑ ∑ ∑ 4.7869 ∑ ∑ ∑

Now:
∑ ∑ ∑
∑ ∑

∑ ∑
ln(a) =

i.e. a = e-0.9641 = 0.381

Finally, the power form equation is: y = 0.381.x2.2832

Home Work
The temperature of a metal strip was measured at various time intervals during
heating and the values are given in the table below:

Time ‘t(min)’ 12 2 3 4
Temperature ‘T( C) 70 83 100 124
o

If the relation between the time ‘t’ and temperature ‘T’ is of the form: T = b.et/4 +
a, estimate the temperature at t = 6 minute.

Solution:
Hint: As T = b.et/4 +a … (i), we can write this equation in the form: y = b.f(x) +a …. (ii),
54

where, y = T and f(x) = et/4. Now, use linear formula and adjust the values at final. Ans:
37.56et/4+21.31
Page

Prepared By
Er. Shree Krishna Khadka
FITTING A POLYNOMIAL EQUATION
When a given series of data does not appear to satisfy a linear equation, we can try a
suitable polynomial as a regression to fit the data. The least square technique can be
used to fit the data to a polynomial
Consider the polynomial of degree (m-1) as below:

… (i)

If the data contains ‘n’ sets of ‘x’ and ‘y’, then the sum of squares of the error is given
by:
∑ …. (ii)

Hence; f(x) is a polynomial and contain coefficients: a1, a2,….,am. We have to estimate
all ‘m’ coefficients. As before, we have the following ‘m’ equations that can be solved
for these coefficients.

i.e.

Consider a general term:

∑ ; Where:

Thus we have:

∑ ; Where:

Or ∑ { } ; …. (iii)

Now, Substituting for f(xi):

∑ ∑

So, there are ‘m’ equations: (j=1, 2, 3, …. ,n) and each summation is for i =1 to i =n.

∑ ∑ ∑ ∑
∑ ∑ ∑ ∑ ∑
……………………………… …….. ……..
……………………………… …….. ……..
55

∑ ∑ ∑ ∑ ∑
Page

Prepared By
Er. Shree Krishna Khadka
#Note: For a second order polynomial, we will have the following simultaneous
equations:
∑ ∑ ∑
∑ ∑ ∑ ∑
∑ ∑ ∑ ∑

Example:
Fit a second order polynomial to the data in the table below:
x 1 2 3 4
y 6 11 18 27

Solution:
Below is the tabulation the values as second order polynomial components.
x y x2 x3 x4 xy x2y
1 6 1 1 1 6 6
2 11 4 8 16 22 44
3 18 9 27 81 54 162
4 27 16 64 256 108 432
∑ ∑ ∑ ∑ ∑ ∑ ∑

So, we have the following simultaneous equations:

4a1 + 10a2 + 30a3 = 62 … (i)

10a1 + 30a2 + 100a3 = 190 … (ii)
30a1 + 100a2 + 226a3 = 644 … (iii)

On calculation: a1 = 3, a2 = 2 and a3 = 1
Hence, the least square quadratic polynomial is: y = 3 + 2x + x2.

Home Work:
From the given data table, fit a second order polynomial.

x 1 2.1 3.2 4
y 2 2.5 3.0 4

Equations:
4a1 + 10.3a2 + 31.65a3 = 11.5 … (i)
10.3a1 + 31.65a2 + 107.029a3 = 32.85 … (ii)
31.65a1 + 107.029a2 + 381.3057a3 = 107.745 … (iii)
56

The polynomial is: y = 2.074 -0.205x + 0.168x2

Page

Prepared By
Er. Shree Krishna Khadka
MULTIPLE LINEAR REGRESSIONS:
There are number of situations, where the dependent variable is a function of two or
more variables. For example, the salary of a sales person may be expressed as: y = 500
+ 5x1 + 8x2, where x1 and x2 are the number of unit sold of product 1 and 2
respectively. We shall discuss an approach to fit the experimental data, where the
variable under consideration is a linear function of two independent variables.

Let us consider a two variables linear function as follows:

… (i)

Then, the sum of square of errors is given by:

∑ ∑ … (ii)

Differentiation equation … (ii) w. r. t. a1, a2, a3 and equating them to zero, we will get
the condition for minimum error.

With these conditions, we will get the following simultaneous equation:

∑ ∑ ∑
∑ ∑ ∑ ∑
∑ ∑ ∑ ∑

Example:
Given the table of data:

x 1 2 3 4
z 0 1 2 3
y 12 18 24 30

Solution:
57

Constructing the table as a component of multiple linear regression equation.

Page

Prepared By
Er. Shree Krishna Khadka
 x z Y x2 z2 z.x x.y y.z
1 0 12 1 0 0 12 0
2 1 18 4 1 2 36 18
3 2 24 9 4 6 72 48
4 3 30 16 9 12 120 90
∑ ∑ ∑ ∑ ∑ ∑ ∑ ∑

So, three simultaneous equations will be:

4a1 + 10a2 + 6a3 = 84 … (i)

10a1 + 30a2 + 20a3 = 240 … (ii)
6a1 + 20a2 + 14a3 = 156 … (iii)

On solving we can get the values of unknowns.

Home Work
Given the data points

X 5 4 3 2 1
Z 3 -2 -1 4 0
Y 15 8 -1 26 8

Fit the multiple linear regressions.

Solution:

Ans: y = 10 – 2x + 5z

SPLINE INTERPOLATION:
Data fitting by means of polynomials has been considered in the previous exercise.
From the viewpoint of cubic splines, the resulting approximation called the cubic
spline approximation, suffers from the disadvantage of being a global approximation,
which means that a change in one point affects the entire approximation curve. So, a
method based on the basis splines, which posses a local character, viz; a change in one
point introduces a change only in the immediate neighbourhood of that point, called
B-spline method that finds important application in computer graphics and smoothing
of data.

Cubic B-Splines:
58

It resembles the ordinary cubic spline in which a separate cubic is derived for each
interval. Specially, a cubic B-spline (or a B-spline of order-4) denoted by B4i(x), is a
Page

cubic spline with knots Ki-4, Ki-3, Ki-2, Ki-1 and Ki, which is zero everywhere except in the
Prepared By
Er. Shree Krishna Khadka
 range Ki-4 < x < Ki. In such case, B4i(x) is said to have a support [Ki-1, Ki]. It may be
noted that a B-spline need not necessarily pass through any or all of the data points.
Cubic splines are popular because of their ability to interpolate data with smooth
curves. If we consider the construction of a cubic spline function which would
interpolate the points (x0, f0), (x1, f1) … (xn, fn). The cubic spline S(x) consists of (n-1)
cubic corresponding to (n-1) sub intervals. If we denote such cubic by Si(x), then:
S(x) = Si(x); for I = 1, 2, … n.

These cubic must satisfy the following conditions:

a) S(x) must interpolate the function at all the points: x0 , x1, …. xn; i.e. for each ‘i',
S(xi) = fi.
b) The function values must be equal at all the interior knots, i.e. Si(xi) = Si+1(xi)
c) The first derivative at the interior knots must be equal; Si’(x) = Si+1’(x)
d) The second derivative at the interior knots must be equal; Si’’(xi) = Si+1’’(xi)
e) The second derivative at the end points are zero; S’’(x0) = S’’(xn) = 0

Let us consider (n+1) sample points with ‘n’ number of different section as shown in
the figure below. Suppose, gi(x) represents a cubic polynomial between [xi, f(xi)] and
[xi+1, f(xi+1)].

Then: gi(x) = ai(x-xi)3 + bi(x-xi)2 + ci(x-xi) + di …. (i)

g(xi)
F(x) g(xi+1)

……… ………

x
xi-2 xi-1 xi xi+1… x

Fig: Spline Interpolation

1.) Since, function must pass through all the points; i.e; gi(xi) = di = f(xi) ….(i)

2.) Differentiating equation (i) w. r. t. ‘x’ we get:

gi’(x) = 3ai(x-xi)2 + 2bi(x-xi) + Ci … (ii)

59

gi’’(x) = 6ai(x-xi) + 2bi … (iii)

Page

Prepared By
Er. Shree Krishna Khadka
 at x = xi , equation (iii) becomes: gi’’(x) = 2bi = Si say; i.e. bi = Si/2 …(iv)

Similarly:
gi’’(xi+1) = Si+1 = 6ai(xi+1 - xi) + 2bi = 6aihi + 2bi

Where; xi+1 - xi = hi
Then; ai = (Si+1 – Si)/6hi …. (v)
3.) Again, gi(xi+1) = f(xi+1);
But, f(xi+1) = aihi3 + bihi2 + Cihi + di = aihi3 + bihi2 + Cihi + f(xi)
i.e. Ci = [{f(xi+1) – f(xi)}/h] - aihi2 - bihi ….(vi)

4.) Also from equation (ii) we have: gi-1’(x) = 3ai-1(x-xi-1)2 + 2bi-1(x-xi-1) + Ci-1 … (vii)
And we also have the condition that, at x = xi, gi’(xi) = gi-1’(xi), then:
Ci = 3ai-1(hi-1)2 + 2bi-1(hi-1) + Ci-1 …. (viii)
5.) Finally on manipulating all these values, we get a final equation as:

{ }
6.) Expressing above equation in matrix form:

[ ][ ]=6[ ]

Where, …. are first divided difference values calculated from the

tabulated data and h0, h1 … are the successive intervals between two data points
that may or may not be equal. For natural spline, we have: S0 =Sn = 0 (Second order
derivative at end points is zero)

Example:
Fit the data of table with a natural cubic spline curve and evaluate the spline
values, g(0.66) and g(1.75), with h0 = 1, h1 = 0.5 and h2 = 0.75
X 0 1 1.5 2.25
f(x) 2 4.4366 6.7134 13.9130
(Hint: The relation is described as: f(x) = 2ex – x2)
Solution:

1.) From the given table, we have:

h0 = x1 – x0 = 1 – 0 = 1
60

h1 = x2 – x1 = 1.5 – 1 = 0.5
h2 = x3 – x2 = 2.25 – 1.5 = 0.75
Page

Prepared By
Er. Shree Krishna Khadka
2.) For natural spline, S0 = 0 and S3 = 0

3.) Now, Finding Si = S1 and S2. Generating Matrix Model as follow:

[ ][ ] [ ]

i.e. [ ][ ] [ ]

i.e. 3S1 + 0.5S2 = 12.702 … (i) and 0.5S1 + 2.5S2 = 30.2742 …(ii)

On solving: S1 = 2.292 and S2 = 11.65

4.) Now, finding ai, bi, ci and di

i 0 1 2
ai = (Si+1 – Si)/6hi 0.381 3.12 -2.598
bi = Si/2 0 1.146 5.825
Ci = [{f(xi+1) – f(xi)}/h] - aihi2 - bihi 2.0556 3.200 6.680
di = f(xi) 2 4.4366 6.7134

5.) Finally, g(0.66) = a0(x-x0)3 + b0(x-x0)2 + c0(x-x0) + d0 = 3.466

And, g(1.75) = a2(x-x2)3 + b2(x-x2)2 + c2(x-x2) + d0 = 8.70

Home Work:
Given the data points:

i 0 1 2
xi 4 9 16
f(xi) 2 3 4

Estimate the value of f(7)

Solution:
h0 = 5 and h1 = 7
S0 = 0 = S2, but S1 = -0.01428
a0 = -0.000476, a1 = 0.00034
b0 = 0, b1 = -0.00714
c0 = 0.2119, c1 = 0.6259
61

d0 = 2, d1 = 3
Page

Finally, f(7) = g0(x) = 2.6228

Prepared By
Er. Shree Krishna Khadka
APPROXIMATION OF FUNCTIONS:
The problem of approximating a function is a central problem in numerical analysis
due to its importance in the development of software for digital computers. Functions
evaluation through interpolation techniques over stored table of values has been
found to be quite costlier, when compared to the use of efficient function
approximations.

If f1, f2 … fn be the values of the given function and g1, g2, … gn be the corresponding
values of the approximating function; then the error vector ‘e’, where the components
of ‘e’ is given by: ei = fi – gi …..(i).

The approximation may be chosen in a number of ways. For example, we may find the
approximation such that the quantity √ is minimum. This leads us
to the least square approximation, which we have already studied. On the other hand,
we may choose the approximation such that the maximum component of ‘e’ is
minimized. This leads us to the ‘Celebrated Chebyshev Polynomials’, which have found
important application in the approximation of function in digital computers.

62
Page

Prepared By
Er. Shree Krishna Khadka
Assignment 4

Full Marks: 25
Pass Marks: 15

6. The table below gives the temperatures T(OC) and lengths(mm) of a heated
rod. If l = a0 + a1T, find the values of a0 and a1 using linear least squares.
Find the length of the rod when T = 72.5oC. [5]

T 40 50 60 70 80
l 600.5 600.6 600.8 600.9 601.0

7. The temperature of a metal strip was measured at various time intervals

during heating and the values are given in the table below.

Time(‘t’ min) 1 2 3 4
Temp(‘T’ OC) 70 83 100 124

If the relation between the time ‘t’ and temperature ‘T’ is of the form: T =
bet/4 + a. Estimate the temperature at t = 6 minute. [5]

8. Fit a second order polynomial to the data given in the table below. [5]

x 1 2.1 3.2 4
f(x) 2 2.5 3.0 4

Find the polynomial value at 1.5 and 3.8.

9. Given the table of values

x 5 4 3 2 1
z 3 -2 -1 4 0
y 15 8 -1 26 8

Use multiple linear regression formula to fit the data of the form:
y = a1 + a2(x) +a3(z) & Find f(5.3) and f(3.7). [5]

10. Fit a natural cubic B-Spline, S to the data given below in the table. [5]

x 1.2 2.8 3.5 4.1 5.9

f(x)=x3-1 -2 -1 0 1 2

Find the functional value at 1.5 and 6

63
Page

Prepared By
Er. Shree Krishna Khadka
`

Numerical Differentiation
& Integration

Contents:

o Numerical Differentiation: Introduction

o Differentiating Continuous Functions
o Differentiating Tabulated Functions
o Higher Order Derivatives
o Numerical Integration: Introduction
o Newton Cotes General Formula
o Trapezoidal & Composite Trapezoidal Rule
o Simpson’s & Composite Simpson’s Rule
o Gaussian Integration & Changing the limit of integration
o Romberg Integration
o Numerical Double Integration
64
Page

Prepared By
Er. Shree Krishna Khadka
NUMERICAL DIFFERENTIATION: INTRODUCTION
The method of obtaining the derivative of a function using a numerical technique is
known as numerical differentiation. The general method for deriving the numerical
differentiation formulae is to differentiate the interpolating polynomial. There are
essentially two situations where numerical differentiation is required. They are:

1. The function values are known but the function is unknown. Such functions are
called tabulated function.

2. The function to be differentiated is continuous and therefore complicated and

difficult to differentiate.

Since, analytical methods give exact answers; the numerical techniques provide only
approximations to derivatives. Numerical differentiation methods are very sensitive
to round off errors, in addition to the truncation error introduced by the methods of
themselves. So, it is necessary to discuss the errors and ways to minimize them.

DIFFERENTIATING CONTINUOUS FUNCTION

Here, the numerical process of approximating the derivative f’(x) of a function f(x) is
carried out when the function is continuous and itself available.

Forward Difference & Backward Difference Quotient

Consider a small increment, . According to Taylor’s Theorem, we have

Rearranging the terms, we get: …. (a) with the truncation

error: . Equation (a) is called the first order forward different

quotient. This is also called as two point formula. The truncation error is in the order
of ‘h’ and can be decreased by decreasing ‘h’.

Similarly, we can also show that the first order backward difference quotient is:

…. (a)
65
Page

Prepared By
Er. Shree Krishna Khadka
Example:
Estimate approximate derivative of f(x) = x2 at x = 1, for h = 0.2, 0.2, 0.05, 0.01
using first order forward difference formula.

Solution:
We have: f(x) = x2
a) Analytical Method: f’(x) = 2x, i.e. f’(x=1) = 2 X 1 = 2 (true value)

b) Numerical Method:
, which will give the approximation of derivative.

Now, derivative approximations are tabulated below:

h f'(1) Error
0.2 2.2 0.2
0.1 2.1 0.1
0.05 2.05 0.05
0.01 2.01 0.01

Conclusively, it is found that the derivative approximation approaches the exact value
as ‘h’ decreases. The truncation error decreases proportionally with decrease in ‘h’.
There is no round off error.

Central Difference Quotient

Equation (a) was obtained using the linear approximation to f(x). This would give
large truncation error if the functions were of higher order. In such cases, we can
reduce truncation errors for a given ‘h’ by using a quadratic approximation rather
than a linear one. This can be achieved by the following form of Taylor’s Expansion.
i.e. … (a) and similarly,

… (b)

Subtracting equation (b) from equation (a), we get:

… (c)

Thus we have: … (d) with truncation error of:

{ } { } ; which is order of ‘h2’. Equation © is called
second order central difference quotient. This is also known as three point formula.
66
Page

Prepared By
Er. Shree Krishna Khadka
Example:
Repeat the previous example for three point formula.

Solution:
We have: f(x) = x2 , i.e. f(x=1) = 2 (true value)
Now, for different values of ‘h’ the derivative approximations are tabulated below:

h f'(1) Error
0.2 2 0
0.1 2 0
0.05 2 0
0.01 2 0

Where:

Error Analysis
Numerical differentiation is very sensitive to round off errors. If Er(h) is the round off
error introduced in an approximation of derivative, then the total error is given by:
E(h) = Er(h) + Et(h) …. (i)

Let us consider two point formulae for the purpose of analysis.

i.e.

If we assume the round off error in f1 and f0 as e1 and e0 respectively, then:

( ) ( )

If the errors e1 and e0 are of the magnitude ‘e’ and of opposite signs, the we get bound
for round off error as:

We know, the truncation error for two point formula is:

67

Where, M2 is the bound given by M2 = max|f’’( |

Page

Thus the boundary for total error in the derivative is:

Prepared By
Er. Shree Krishna Khadka
 …(a)

We can obtain a rough estimation of ‘h’ that gives the minimum error. By
differentiating equation (a) with respect to ‘h’, we get:

Now, we know that E(h) is minimum when E’(h) = 0. So, solving for ‘h’, we obtain

Where, hopt is optimum step size. Substituting this value in equation (a), we get:

( ) √

Example:
Compute the approximate derivative of f(x) = sin(x) at x = 0.45 radian at
increasing value of ‘h’ from 0.01 to 0.04 with a step size of 0.005. Analyse the
total error. What is the optimum step size?

Solution:
We have given: y = f(x) = sin(x) and f’(x) = cos(x)
So, f(x=0.45rad) = sin(0.45rad) = 0.4350(rounded to four digits)
And f’(x=0.45rad) = cos(0.45rad) = 0.9004

Using two point formula, table below gives the approximate derivatives of sin(x) at x =
0.45 radian using different values of ‘h’.

h f(x+h) f'(x) Error

0.010 0.4439 0.8900 0.0104
0.015 0.4484 0.8933 0.0071
0.020 0.4529 0.8950 0.0054
0.025 0.4573 0.8935 0.0069
0.030 0.4618 0.8933 0.0071
0.035 0.4662 0.8914 0.0090
0.040 0.4706 0.8900 0.0104
68
Page

Prepared By
Er. Shree Krishna Khadka
The total error decreases from 0.0104 (at h = 0.01) till h = 0.02 and again increases
when ‘h’ is increased.

Since, we have used four significant digits; the bound for round off error ‘e’ is 0.5x10 -4.
For the two point formula, the bound M2 is given by:

M2 = max|f’’( )| for:

i.e. M2 = |-sin(0.49)| = 0.4706

Therefore, the optimum step size is:

√ √

Finally, the error at optimum step size is:

( ) √

Higher Order Derivatives:

We can also obtain approximations to higher order derivatives using Taylor’s Series
Expansion.

As we know that:
… (a) and similarly,

… (b)

Where: R1 = and R2 = =

So, the addition of two expansions gives:

…. (c)

Therefore:

with truncation error of: Et(h) = -(R1 + R2)/h2

i.e. Et(h) =
69
Page

So, the error is of the order h2 .

Prepared By
Er. Shree Krishna Khadka
 Example:
Find the approximation to second derivative of cos(x) at x = 0.75 with = 0.01.
Compare with the true value.

Solution:

We have given: y = f(x) = cos(x)

Then: y’ = f’(x) = -sin(x) and y’’ = f’’(x) = -cos(x)
Therefore: f’’(x=0.75) = -cos(0.75) = -0.7316888

Using the formula,

i.e. (at h = 0.01)

Therefore,
f’’(0.75) =

Finally, Error = |-0.7316888-(-0.7300000)| = 0.0016888

DIFFERENTIATING TABULATED FUNCTION

If we are given a set of data points (xi, fi), for i = 0, 1, 2 …. n, which corresponds to the
values of an unknown function f(x) and we wish to estimate the derivative at these
points. Assume that the points are equally spaced with a step size of ‘h’.

When function values are available in tabulated from, we may approximate this
function by an interpolation polynomial P(x) as discussed in earlier chapter and then
differentiate P(x).

If we consider the quadratic approximation, we need to use three point as referenced

by Newton Polynomial Form:

i.e. …. (i)
Then: …. (ii)

If x0 = x, x1 = x + h, x2 = x + 2h, then:

[ … (a), which is well known as two point formula for the

approximation fo first derivative, and on calculation:
70

[ … (b)
Page

Prepared By
Er. Shree Krishna Khadka
Now, substituting the values of a1 and a2 in equation (ii), we will get the three-point
forward difference formula for the approximation of first derivative as below:

… (iii)

Hence, the backward form of equation (iii) will be:

… (iv)

Similarly, we can obtain the three-point central difference formula by letting x0 = x, x1

= x - h and x2 = x + h in equation (ii), Then:

[ … (c) and

Finally, by substituting these values in equation (ii), we get:

…(v) which is the same result as derived in the earlier section

for the approximation of derivative of continuous function.

Example:

The table below gives the values of distance travelled by a car at various time
intervals during the initial running.

Time, t(s) 5 6 7 8 9
Distance Travelled, s(t) (km) 10.0 14.5 19.5 25.5 32.0

Estimate the velocity at time t = 5s, t = 7s and t = 9s. Also find the acceleration at
t = 7s.

Solution:

The velocity is given by the first derivative of s(t).

71
Page

Prepared By
Er. Shree Krishna Khadka
Case I:
At time, t = 5, we use the three-point forward difference formula.

i.e. v(t=5)

4.25km/s

Case II:
At time, t = 7, we use the central difference formula.

i.e. v(t=7)

5.5 km/s

Case III:
At time, t = 9, we use the three-point backward difference formula.

i.e. v(t=9)

= 6.75 km/s

Case IV:
Acceleration is given by the second derivative of s(t). Therefore:
72

i.e. = 1 km/s2
Page

Prepared By
Er. Shree Krishna Khadka
NUMERICAL INTEGRATION: INTRODUCTION
The general objective of numerical integration is to compute the value of the definite
integral given below for a given set of data points (x0, y0), (x1, y1), (x2, y2), (x3, y3), …
,(xn, yn) of a function y = f(x), where f(x) is not know explicitly. So, f(x) is determined
using the interpolation technique and then integration is carried out.

i.e. ∫ …(i)

Like in numerical differentiation, we need to evaluate the numerical integration in the

following cases.

1. Functions do not possess closed form solutions.

Example: ∫
2. Closed form solutions exist but these solutions are complex and difficult to use
for calculations.
3. Data for variables are available in the form of table, but no mathematical
relationship between them is known, as is often the case with the experimental
data.

A definite integral of the form: ∫ can be treated as the area under curve y =
f(x), enclosed between the limits x = a and x = b. Then the problem of integration is
then simply reduced to the problem of finding the shaded area.

NEWTON COTES GENERAL INTEGRATION FORMULA:

It is the most popular and widely used numerical integration formula, which provides
the basis for a number of numerical integration methods known as Newton Cotes
Method.

It is based on the polynomial interpolation. An nth degree polynomial Pn(x) that

interpolates the values of f(x) at (n+1) evenly spaced points can be used to replace the
integrand f(x) of the integral: ∫ and the resultant formula is called (n+1)
Newton Cotes Formula.

If the limits of integration ‘a’ and ‘b’ are in the set of interpolating points xi = 0, 1, 2, …,
n, then the formula is referred to as closed form. Whereas if the points ‘a’ and ‘b’ lie
beyond the set of interpolating points, then the formula is termed as open form. Since,
open form formula is not used for definite integration; we consider here only the
73

closed form methods.

Page

Prepared By
Er. Shree Krishna Khadka
Let us consider a General Newton’s Interpolation Formula, given by:

(ii)

Where, , then
When, and When

Now, substituting all these values in equation (i), we get:

∫ ∫

{ ( ) ( ) }

{ ( ) ( ) } …. (iii)

Where: h = (xn – x0)/n and equation (iii) is the general equation for numerical
integration and is called General Newton Cotes Formula.

TRAPEZOIDAL RULE (TWO POINT FORMULA)

Setting, n = 1 in General Newton Cotes Formula and neglecting second and higher
order differences gives a two point formula by which numerical integration is carried
out known as Trapezoidal Rule.

F(x) F(b)

F(a)

x
x0=a x1=b
Area(a) ={F(b)+F(a)/2}(b-a)

Fig: Trapezoidal Rule

i.e. ∫ Area of trapezoid.

74
Page

Prepared By
Er. Shree Krishna Khadka
Example:
Evaluate the integral: I = ∫ for (i) (1, 2) and (ii) (1, 1.5)

Solution:

(a,b) H f(a)=a3 + 1 f(b)= b3 + 1 I=(f(a)+f(b))(h/2)

(1, 2) 1 2 9 5.5
(1, 1.5) 0.5 2 4.375 1.59375

COMPOSITE TRAPEZOIDAL RULE:

If the range is to be integrated and the function is non linear, the trapezoidal rule can
be improved by dividing the range (a, b) into number of small intervals and applying
the rule discussed above for all sub intervals. The sum of areas of all the sub intervals
is the integral of the interval (a, b) as shown in the figure below.

F(x)

x
x0=a x1 x2 x3 xn-1 xn=b

Fig: Composite Trapezoidal Rule

If the data range (a, b) is divided into ‘n’ stripes of equal width ‘h’, then:

∫ ∫ ∫ ∫

( ) ( ) ( )

i.e. +…+ … (i)

Here, equation (i) is the general formula for composite trapezoidal rule.
75
Page

Prepared By
Er. Shree Krishna Khadka
Example:
Evaluate: ∫ by dividing into four stripes.

Solution:
We have given: f(x) = sin(x)
Number of stripes (n) = 4
Limits (a, b) = (0, )
Interval (h) = (b-a)/n = ( /4 = /4

Now the approximation of the integration is tabulated below.

x 0 /4 /2 3 /4
f(x) 0 0.7071 1 0.7071 0

As we have the formula:

Compute the integral, : ∫ using composite trapezoidal rule for (a) n =

2 and (b) n = 4.

Solution:

I Case:
For n = 2, we have: h = (1-(-1))/2=1

x -1 0 1
f(x) 0.368 1 2.718

II case:
For n = 4, we have: h = (1-(-1))/4 = 0.5

x -1 -0.5 0 0.5 1
f(x) 0.368 0.607 1 1.649 2.718

} = 2.3995
76
Page

Prepared By
Er. Shree Krishna Khadka
SIMPSON (1/3) RULE: (THREE POINT FORMULA)
Here, the function f(x) is approximated by a second order polynomial, which passes
through three sampling points. So, taking n = 2 in General Newton Cotes Formula and
neglecting third and higher order differences, we get:

I ∫

Since, it uses three points to evaluate the integral; it is also known as three-point
formula. The three points are: x0 = a, x2 = b and x1 = (a + b)/2. Where, the interval ‘h’ is
given by: h = (b – a)/2.

Example:
Evaluate the integral, ∫ using Simpson’s 1/3 Rule.

Solution:

We have given: f(x) = ex

Limit Interval (a, b) = (-1, 1) = (x0, x2)
Then, x1 = (-1+1)/2 = 0 and h = (1+1)/2 = 1

Now, for h = 1, the range (a, b), the function will generate the following table:

x -1 0 1
f(x) = e 0.3698 1 2.7182
x

Hence, by formula we have:

77
Page

Prepared By
Er. Shree Krishna Khadka
COMPOSITE SIMPSON’S 1/3 RULE:
As in the case of composite trapezoidal rule, the range over which the integral have to
be carried out is divided into ‘n’ number of stripes in order to improve the accuracy of
the estimate of the area. Each stripe is then integrate by Simpson’s 1/3 Rule and
finally tied together to give the Composite Simpson’s Output. Hence, the integration
will be:

∫ ∫

∫ ∫ ∫ ∫

{ } { } { }

∑ ∑

Example:
Compute the integral: ∫ √ applying Composite Simpson’s 1/3 Rule
for n = 4 and n = 6 with accuracy to five decimal places.

Solution:

I Case: For n = 4
We have given: f(x) = √ , Limit Range (a, b) = (0, )
Interval (h) = (b – a)/n = ( )/4 =
So, the functional table will be:

x 0
f(x)= √ 0 0.61861 0.84090 0.96119 1

Now, by formula we have:

78

II Case: For n = 6, h = , and the integration will be: I = 1.18728 (same way)
Page

Prepared By
Er. Shree Krishna Khadka
SIMPSON’S 3/8 RULE
Simpson’s 1/3 rule was derived using three sampling points that fit a quadratic
equation. We can extent this approach to incorporate four sampling points so that the
rule can be exact for f(x) of degree 3.
By using the first four terms from General Newton Cotes Formula and applying the
same procedure followed in the previous case, we will get the following result:

Where, h = (b – a)/3

Example:
Use Simpson’s 3/8 Rule to evaluate, ∫ for n = 3.

Solution:
Here, a = 1, b = 2; So, h = (b – a)/3 = 1/3
Now, for different values of ‘x’ f(x) will generate the following values.

x 1 1.33 1.66 2
f(x) = x3 + 3 4 5.353 7.574 11

Finally, using Simpson’s 3.8 Rule, we have:

COMPOSITE SIMPSON’S 3/8 RULE:

In this case, the total range of the integral is divided into ‘n’ number of stripes and the
integral is estimated taking three stripes at a time (4-points) using Simpson’s 3/8
Rule. Finally, the total integral is obtained by adding all the sub integrals.

So, following the same procedure as we did in earlier section, we will get the formulae
for the Composite Simpson’s 3/8 Rule as below:

Where, h = (b – a)/n

Example:
Evaluate ∫ for h = 0.1 using Composite Simpson’s 3/8 Rule.
79
Page

/Home Work/

Prepared By
Er. Shree Krishna Khadka
NUMERICAL DOUBLE INTEGRATION

Formula for the evaluation of a double integral can be obtained by repeatedly

applying the trapezoidal and Simpson’s rule derived in earlier section. Let us consider,
as an example, the double integral defined by:

∫ ∫ … (i)

Dividing the range [a, b] into ‘n’ equal sub-intervals of length ‘h’ and dividing the range
[c, d] into ‘m’ equal sub intervals of equal length ‘k’.

Here we are using composite trapezoidal rule repeatedly.

∫ ∫

[∫ ∫ ∫ ∫
∫ ]

I [

Example:
Evaluate the double integral, ∫ ∫ using composite trapezoidal
rule. Take h = k = 0.5.
80
Page

Prepared By
Er. Shree Krishna Khadka
Solution:

We have given that: f(x) = ex + y

Range for integration over ‘x’ is: [a, b] = [0, 1]
Range for integration over ‘y’ is: [c, d] = [0, 1]
Interval in ‘x’ is: (h) = 0.5
Interval in ‘y’ is: (k) = 0.5

Now, for different combination of (x, y) characterized by the interval (h, k), we can
generate the following table:

x x0 x1 x2
y 0 0.5 1
y0 0 1 1.648 2.718
y1 0.5 1.648 2.718 4.482
y2 1 2.718 4.482 7.389

Now, the double integral is evaluated using the composite trapezoidal formula given
by:

I [

= 0.0625(7.014+23.132+19.067)

= 3.0758

GAUSSIAN INTEGRATION (GAUSS LEGENDRE’S INTEGRATION)

We have discussed so far a set of rules based on the Newton Cotes Formula, which was
derived by integrating Newton Gregory Forward Difference Interpolating Polynomial.
Consequently, all the rules were based on evenly faced sampling points (functional
values) within the range of integral.

Gauss Integration is based on the concept that the accuracy of numerical integration
can be improved by choosing the sampling points wisely, rather than on the basis of
81

equal spacing.
Page

Prepared By
Er. Shree Krishna Khadka
So, Gauss Integration assumes an approximation of the form:

∫ ∑ …. (i)

Here, the method of implementing the strategy of finding appropriate values of xi and
wi and obtaining the integral f(x) over the range [-1, 1] is called Gaussian Integration
or Quadrature. Where wi are known as weights of that function associated with and xi
are the nodes (both are unknowns).

Therefore, equation (i) contains 2n unknowns to be determined, which can be

determined using the condition given in the integration formula. So, it should give the
exact value of the integral for polynomials of as high a degree as possible.

Rewriting the equation (i) in the following form of representation:

∫ … (ii)

Let us find the Gaussian Quadrature Formula for n = 2. In this case, we need to find out
the values of four unknowns, namely: and x1 and x2. Let us assume that the
integral will be exact up to cubic polynomials. This implies that the function 1, x, x2, x3
can be numerically integrated to obtain the exact results. Then from equation (i) we
will get:

∫ … (a) (for f(x) = 1)

∫ … (b) (for f(x) = x)

∫ … (c) (for f(x) = x2)

∫ … (d) (for f(x) = x3)

Solving these four simultaneous equations we will get:

and ( , )
√ √

Thus we have the Gaussian Quadrature Formula for n = 2 as:

82

∫ ∑ ( ) ( )
√ √
Page

Prepared By
Er. Shree Krishna Khadka
Example:
Compute the integral, ∫ using two-point Gauss Legendre Formula.

Solution:

∫ ( ) ( ) √ √
√ √

CHANGING LIMITS OF INTEGRATION

Since, the Gaussian Formula imposes a restriction on the limits of integration to be the
form -1 to 1. This restriction can be overcome by using the technique of “interval
transformation”. Let:

∫ ∫ …. (i)

Assume the following transformation between ‘x’ and the new variable ‘z’.
x = Az + B …. (a)
This must satisfy the following conditions:
At x = a, z = -1 and at x = b, z = 1, which gives: B – A = a & A + B = b.
Solving for A and B, we get: A = (b – a)/2 and B = (a + b)/2
Therefore, Equation (a) becomes:

… (b)

Then: … (c)

Now, the equation (i) will be:

∫ ∫ ∫ … (ii)

Hence, the Gaussian formula for this integration is:

∫ ∑ … (iii)

Where and zi are the weights and Quadrature points for the integration domain (-1,
1).
83
Page

Prepared By
Er. Shree Krishna Khadka
Example:
Compute the integral, ∫ using Gaussian two point formula.

Solution:

For n = 2 (Gaussian two point formula)

Where, and therefore,

Also, for two point formula:

√ √

Now, upon substitution of these values, we get:

[ ( ) ( )] [ √ √ ]
√ √

/Home Work/
Compute the integral, ∫ using GL-2-Point Method. [Ans: 0.9984]

GAUSS LEGENDRE THREE-POINT FORMULA:

By using a procedure similar to one applied in deriving two-point formula, we can
obtain the parameters and zi for higher order versions of Gaussian Quadrature.

So, for Gauss Legendre Three-Point Formula, we will have:

And

The integration is again given by:

∫ ∑
84
Page

Prepared By
Er. Shree Krishna Khadka
Example:
Use Gauss Legendre three-point formula to evaluate the following integral:

∫ .

Solution:

For 3-point formula, n= 3

We have given, [a, b] = [2, 4] and

The function: f(x) = (x4 +1)

Then:

∫ ∑

, then:

For Gauss Legendre 3-point formula, we have:

and

So,

Ig = 0.55556[(-0.77460+3)4+1] + 0.88889[(0+3)4+1] + 0.55556[(0.77460+3)4+1]

= 14.18140 + 72.88898 + 113.33105

= 200.40143
85
Page

Prepared By
Er. Shree Krishna Khadka
ROMBERG INTEGRATION
In conclusion, what we have found that, the accuracy of a numerical integration
process can be improved in two ways:

1. By increasing the number of subintervals (i.e. by decreasing ‘h’): this decreases the
magnitude of error terms. Here, the order of the method is fixed.
2. By using higher order methods: this eliminates the lower order error terms. Here,
the order of the method is varied and, therefore, this method is known as variable
order approach.

The variable order method involves combining two estimates of a given order to
obtain a third estimate of higher order. The method that incorporates this process (i.e.
Richardson’s Extrapolation) to the trapezoidal rule is called Romberg Integration.

In order to achieve the Romberg Integration, we have the following steps:

1. Compute the integration with given value of ‘h’ by trapezoidal rule.

2. Each time, halve the value of ‘h’ and again compute the integral using composite
trapezoidal rule.
3. Refine the above computed values using the relation given below:

Example:
Compute the integral: ∫ using Romberg Integration, take h = 0.5, 0.25,
and 0.125

Solution:
Case I: For h = 0. 5
86
Page

Prepared By
Er. Shree Krishna Khadka
Case II: For h = 0.25

Case III: For h = 0.125

Now, from I1 and I2, we have to implement the Romberg Integration.

Let,

Again for I2 and I3, we have:

Finally, we have to refine the integration using Romberg Integration for I’ and I’’ as:
87
Page

Prepared By
Er. Shree Krishna Khadka
Assignment 5

Full Marks: 45
Pass Marks: 25
Grace Mark: 5

11. Estimate the approximate derivative of f(x) = sin2(x) at x = 0.45 radian,

for h = 0.05 and 0.1 using:

a) First order Forward Difference Quotient Formula and [5]

b) First order Central Difference Quotient Formula [5]

Compare the result of a) and b).

(Ref: Differentiation of continuous function)

12. The table below gives the values of distance travelled by a car at
various time intervals during the initial running.

Time(‘t’ sec) 5 6 7 8 9
Temp(‘T’ OC) 10.0 14.5 19.5 25.5 32.0

a) Estimate the velocity at time t = 5sec, 7sec and 9sec. [5]

b) Also estimate the acceleration at t = 7sec. [5]

(Ref: Differentiation of tabulated function)

13. Use trapezoidal rule with n =4 to estimate, I = ∫ correct to five

decimal places. Repeat the same question using composite trapezoidal
rule and analyse the results obtained by two methods. [5]

14. Estimate the integral, I = ∫ with n = 4, using Simpson’s (3/8)

Rule. Repeat the same question using Composite Simpson’s (3/8) Rule
and analyse the two results. [5]

15. Evaluate the double integral, ∫ ∫ for h =

and k = . [5]

( )
16. Evaluate by Gauss Integration, the integral, ∫ for n = 2. [5]

√
17. Use Romberg Integration to evaluate the integral, ∫ for h =
[5]
e/2 and e/4.
88
Page

Prepared By
Er. Shree Krishna Khadka
`

Matrices & Linear

System of Equations

Contents:

o Introduction
o Method of Solving System of Linear Equations
o Elimination/Direct Method
 Gauss Elimination Method
 Gauss Elimination with Pivoting
 Gauss Jordan Method
 Triangular Factorization Method
 Singular Value Decomposition
o Iterative Method
 Jacobi Iteration Method
 Gauss Seidel Method
89
Page

Prepared By
Er. Shree Krishna Khadka
INTRODUCTION: MATRICES & SYSTEM OF LINEAR EQUATIONS
Matrices occur in a variety of problems of interest; e.g. in the solution of linear
algebraic system, solution of partial and ordinary differential equations and Eigen
value problems. In this chapter, we introduce the matrices, through the theory o linear
transformations.

A linear equation involving two variables ‘x’ and ‘y’ has the standard form, ax + by = c,
where a, b and c are real numbers such that (a and b) 0. This equation becomes non
linear if any of the variables ( x or y) have the exponent other than one.
e.g. 4x + 4y = 15 and 3u – 2v = -0.5 are the examples of linear equations, whereas 2x –
xy + y = 2 and x + √ = 6 are the examples of non linear equations.

A linear equation with ‘n’ variables has the form:

… (i)

If we need a unique solution of an equation with ‘n’ variables or unknown then we

need a set of ‘n’ such independent equations known as system of simultaneous
equations. Such as:

-------------------------------------------------
-------------------------------------------------

The above equation can be expressed in matrix form as: AX = B

Where; A is ‘nxn’ matrix, b is an ‘n’ vector and ‘x’ is a vector of ‘n’ unknowns.

Existence of Solution
In solving systems of equations, we are interested in identifying values of the variables
that satisfy all equations in the system simultaneously. If we have given an arbitrary
system of equations, it is difficult to say whether the system has solution or not. There
may be four possibilities that the system has:

i) Unique Solution ii) No Solution

iii) Solution but not unique iv) Ill Conditioned
90
Page

Prepared By
Er. Shree Krishna Khadka
i) Systems with Unique Solution
For example: x + 2y = 9 and 2x – 3y = 4 has a solution (x, y) = (5, 2). Since no other
pair of values of x and y would satisfy the equation, the solution is said to be
unique.

ii) Systems with No Solution

For example: 2x – y = 5 and 3x – 3/2y = 4 has no solution. These two lines are
parallel and, therefore, they never meet. Such equations are called inconsistent
equations.

iii) System has Solutions but not Unique (i.e. infinite solution)
For example: -2x + 3y = 6 and 4x – 6y = -12 has many different solutions. We can
see that these are two different forms of the same equation and, therefore, they
represent the same line. Such equations are called dependent equations.

iv) System is Ill-Conditioned

Ill conditioned systems are very sensitive to round off errors. These errors during
computing process may induce small changes in the coefficients which, in turn,
may result in a large error in the solution.

Graphically, if two lines appear almost parallel, then we can say the system is ill-
conditioned, since it is hard to decide just at which point they intersect.

The problem of ill-condition can be mathematically described for following two

equation systems.

If these two lines are almost parallel, their slopes must be nearly equal.

i.e.

Alternatively: i.e. … (i)

Where, equation (i) is the determinant of the coefficient matrix A. This shows that
the determinant of all ill-conditioned system is very small ore nearly equal to zero.
91
Page

Prepared By
Er. Shree Krishna Khadka
Example:
Solve the following equations:
2x + y = 25
2.001x + y = 25.01
And thereby discuss the effect of ill conditioning.

Solution:

Solving above equation, we will get: x = 10 and y = 5.

If we change the coefficient of x in the second equation to 2.0005 then values of x and
y will be 20 and -15 respectively.

On conclusion, a small change in one of the coefficients has resulted in a large change
in the result. If we substituted these values back into the equations, we get the
residuals as: R1 = 0 and R2 = 0.01. This illustrates the effect of round off errors on ill
conditioned systems.

METHODS OF SOLVING SYSTEM OF LINEAR EQUATIONS

The techniques and methods for solving system of linear algebraic equations belong to
two fundamentally different approaches.

1. Elimination/Direct Method
It reduces the given system of equations to a form from which the solution can be
obtained by simple substitution. We will discuss the following elimination
method.

a) Basic Gauss Elimination Method

b) Gauss Elimination with Pivoting
c) Gauss Jordan Method
d) Triangular Factorization Method
e) Singular Value Decomposition

2. Iterative Method
Iterative approach, as usual, involves assumption of some initial values which
are then refined repeatedly till they reach some accepted level of accuracy. We
will discuss following iterative methods in this chapter.

a) Jacobi Iterative Method

b) Gauss Seidel Iterative Method
92
Page

Prepared By
Er. Shree Krishna Khadka
SOLUTION BY ELIMINATION
Elimination is a method of solving simultaneous linear equations. It involves
elimination of a term containing one of the unknowns in all but one equation. One
such step reduces the order of equation by one. Repeated elimination leads finally to
one equation with one unknown.
Rule:
a) An equation can be multiplied or divided by a constant.
b) One equation can be added or subtracted from another equation.
c) Equations can be written in any order.

Basic Gauss Elimination Method

Gauss Elimination Method is a computer base technique for solving large systems. It
proposes a systematic strategy for reducing the system of equations to the upper
triangular form using the forward elimination approach and then for obtaining values
of unknowns using the back substitution process.

Example:
Solve the following 3X3 system using Gauss Elimination Method.
3x + 2y + z = 10
2x +3y + 2z = 11
x + 2y + 3z = 14

Solution:
Representing the above system of three simultaneous equation in Standard Matrix
Form: i.e AX = B

[ ][ ] [ ] [ ]

[ ] [ ]

[ ]

The job up to this level is a forward elimination process. Finally, on backward

substitution, we will get: z = 3, y = 2 and x = 1.
93
Page

Prepared By
Er. Shree Krishna Khadka
Limitation of Basic GEM and its Overcome:
Let us try another example, to solve the following system of 3-simultaneous equations,
given by:
3x + 6y + z = 16
2x + 4y + 3z = 13
x + 3y + 2z = 9

i.e. [ ][ ] [ ] [ ]

[ ]

Here, the elimination procedure breaks down since, A22 = 0. So, second row cannot be
normalized. Therefore, the procedure fails. One way to overcome this problem is to
interchange this row with another row below it which does not have a zero element in
that position. Then, we will have:

[ ]

Now, on backward substitution: z = 1, y = 2 and x = 1.

Gauss Elimination with Pivoting:

Here, aij, when i = j, is known as a pivot element. Each row is normalized by dividing
the coefficient of that row by its pivot element.

i.e. , for j = 1, 2, … , n

If akk = 0; kth row cannot be normalized. Therefore, the procedure fails as in the above
case. And of course, to overcome this problem is to interchange this row with another
row below it which does not have a zero element in that position. But, there may be
more than one non-zero values in the kth column below the element akk. So, the
question is: which one of them is to be selected? It can be proved that round off error
would be reduced if the absolute value of the pivot element is large. Therefore, it is
suggested that the row with zero pivot element should be interchanged with the row
having the largest (absolute) coefficient in that position.
94

In general, the reordering of equations is done to improve accuracy, even if the pivot
Page

element is not zero.

Prepared By
Er. Shree Krishna Khadka
Example:
Solve the following system of linear equation by Gauss Elimination with
Pivoting: 2x + 2y + z = 6
4x + 2y + 3z = 4
x – y +z = 0
Solution:

i.e. [ ][ ] [ ] [ ]

1. Original System, 4 in the first column below 2 is the largest absolute value among
all. So, by interchanging first and second row, we will get the following modified
original system with first row as a PIVOT after interchanged.

[ ]

2. Now, applying the Basic GEM, we will get the first derived system as below:

[ ]

3. Here, the largest absolute value is: 3/2 below 1. So, interchanging second and third
row to have a modified first derived system as below with second row as a PIVOT
after interchanged.

[ ]

4. Again applying Basic GEM, we will get the second and final derived system as:

[ ]

5. Finally, on backward substitution: z = -10, y = -1 and x = 9.

95
Page

Prepared By
Er. Shree Krishna Khadka
Gauss Jordan Method:
This method also uses the process of elimination of variables. But, the variable is
eliminated from all other rows (both below and above). This process thus eliminates
all the off-diagonal terms producing a diagonal matrix rather than a triangular matrix.
Further, all rows are normalized by dividing them by their pivot elements.

Example:
Solve by Gauss Jordan Method:
2x + 4y -6z = -8
x + 3y + z = 10
2x - 4y – 2z = -12

Solution:

i.e. [ ][ ] [ ] [ ]

[ ]

[ ]

[ ]

[ ]

[ ]
96
Page

Prepared By
Er. Shree Krishna Khadka
Triangular Factorization Method
Since, the system of linear equation can be expressed in the matrix form as: AX = B. So,
here in the triangular factorization method, the coefficient matrix A of a system of
linear equations can be factorized or decomposed into two triangular matrices L and
U such that: A = LU …. (i).

Where, , known as lower triangular matrix.

[ ]

& , known as upper triangular matrix.

[ ]

Once, A is factorized into L and U, the system of equations AX = B can be expressed by:
(LU)X = B, i.e. L(UX) = B … (ii)

If we assume, UX = Y, where Y is an unknown vector. Then:

LY = B …. (iii)

Now, we can solve AX = B in two stages:

a) Solving the equation: LY = B for Y by forward substitution and
b) Solving the equation UX = Y for X using Y by backward substitution.

The elements of L and U can be determined by comparing the elements of the product
o L and U with those of A. This is done by assuming the diagonal elements of L or U to
be unity.

o The decomposition with L having unit diagonal values is called the Dolittle LU
Decomposition.

o The decomposition with U having unit diagonal elements is called the Crout LU
97

Decomposition.
Page

Prepared By
Er. Shree Krishna Khadka
Example:
Solve the system of three simultaneous linear equations by using Dolittle LU
Decomposition Method.
3x + 2y + z = 10
2x + 3y + 2z = 14
x + 2y + 3z = 14

Solution:
Using Dolittle LU Decomposition, we have:

[ [ ] [ ][ ]

[ ] [ ]

On comparisons, we will have the following relations:

u11 = 3, u12 = 2, u13 = 1, l21u11 = 2 : l21 = 2/3,
l21u12 + u22 = 3: u22 = 3 – (2/3)2 = 5/3,
l21u12 +u23 = 2: u23 = 2 – (2/3)1 = 4/3,
l31u11 = 1: l31 = 1/3, l31u12 + l32u22 = 2: l32 = (2 – (1/3)2)/(5/3) = 4/5
l31u13 + l32u23 + u33 = 3: u33 = (3 – (4/5)(4/3) – (1/3)(1) = 24/15

Thus we have: [ ] [ ]

(i) Forward Substitution: Solving: [ ][ ] [ ] , we will get:

y1 = 10, y2 = 22/3 and y3 = 72/15.

(ii) Backward Substitution: Solving:[ ][ ] [ ] , we will get:

98

z = 3, y = 2 and x = 1.
Page

Prepared By
Er. Shree Krishna Khadka
SINGULAR VALUE DECOMPOSITION
Since, we used LU Decomposition only for a square matrix A. But a similar
decomposition is also possible for a rectangular matrix, which is given by Singular
Value Decomposition. This method is of great importance in matrix theory since it is
useful in finding the generalized inverse of a singular matrix and has several image
processing applications.

Let A be an (mxn) real matrix with m n. Then, the matrices ATA and AAT are non-
negative, symmetric and have identical Eigen values, say . We can then obtain the ‘n’
ortho-normalized Eigen vectors, say Xn of ATA such that:

(ATA)Xn = Xn … (i)

If we assume Yn to be the ‘n’ ortho-normalized Eigen vectors of AAT, then we have:

(AAT)Yn = Yn … (ii)

Then, A can be decomposed into the form: A = UDVT … (iii)

Where: [⃑⃑⃑ ⃑⃑⃑⃑⃑ ⃑⃑⃑⃑⃑ ⃑⃑⃑⃑ [⃑⃑⃑ ⃑⃑⃑ ⃑⃑⃑ ⃑⃑⃑⃑ T

And D is a diagonal matrix, given by: Diag (√ , √ , √ …. √ )

It follows that the Eigen vectors of AAT, i.e. Yn are given by:

… (iv)
√

Example:

Obtain the Singular Value Decomposition of: [ ].

Solution:
We have:

1. [ ] So, [ ], Then: ATA = [ ]

2. Now, to find Eigen Vectors of ATA, we have the characteristics equation:

(Eigen Value Problem: An Example)
99
Page

|ATA – |=0
Prepared By
Er. Shree Krishna Khadka
 |[ ] [ ]|

|[ ]|

On solving for , we will get: = 16.64 and = 0.36

Then: D = Diag.(√ ,√ ) = Diag.( , 0.60)

3. Now, for the corresponding Eigen Vectors of ATA, we have:

4. [ ][ ] [ ] [ ][ ] [ ]

This gives the systems:

The solution is then given by:

[ ] [ ]

Then: Xn = [⃑⃑⃑ ⃑⃑⃑ T =[ ]

5. Finally, the Eigen Vectors of AAT, we have:

√ √

[ ][ ] [ ] [ ][ ] [ ]

So, the singular value decomposition of A is:

100

[ [ ] [ ][ ][ ]
Page

Prepared By
Er. Shree Krishna Khadka
SOLUTION BY ITERATION
Solution by direct method poses some problems when the systems grow larger or
when most of the coefficients are zero. They require prohibitively large number of
floating point operation and therefore, not only time consuming but also severely
affect the accuracy of the solution due to round off errors. Iterative methods provide
an alternative for this and also ill conditioned systems can be solved by this method
without facing the problem of round off errors.

Jacobi Iteration Method

Jacobi method is base on the idea as that for the fixed point iteration method
discussed in chapter two. Recalling that equation of the form: f(x) = 0 can be
rearranged into a form: x = g(x). The function g(x) can be evaluated iteratively using
an initial approximation x as: xi+1 = g(xi) … (i)

Jacobi method extends this idea to a system of equations. It is direct substitution

method where the values of unknowns are improved by substituting directly the
previous values.

Let us consider a system of ‘n’ equations in ‘n’ unknowns:

-------------------------------------------------
-------------------------------------------------

We can rewrite these original systems as:

-------------------------------------------------
-------------------------------------------------

Now, we can compute x1, x2… xn by using initial guesses for these values. These new
values are again used to compute the next set of x values. The process can continue till
we obtain a desired level of accuracy in the x values.
101
Page

Prepared By
Er. Shree Krishna Khadka
Example:
Obtain the solution of the following system using the Jacobi Iteration Method.
2x + y + z = 5
3x + 5y + 2z = 15
2x + y + 4z = 8

Solution:
Rewriting the above equation in the form:

x = (5 - y -z)/2
y = (15 – 3x – 2z)/5
z = (8 – 2x – y)/4

If we assume initial values of x, y and z to be zero, then going by iteration we will get
the following results:

n x y Z
0 0 0 0
1 2.5 3 2
2 0 0.7 0
3 2.15 3 1.825
4 0.0875 1.225 0.175
5 … … …

The process can be continued till the value of x reach a desired level of accuracy.

Gauss Seidel Iteration Method

It is an improved version of Jacobi Iteration Method. In Jacobi method, we begin with
the initial values: and obtain the next
approximation . Note that, in computing , we used
and not which has just been computed. Since, at this point, both and
are available, we can use which is a better approximation for computing .
Similarly, for computing , we can use and along with ,…, .
This idea can be extended to all subsequent computations. This approach is called the
Gauss Seidel Method.
102
Page

Prepared By
Er. Shree Krishna Khadka
Example:
Obtain the solution of the following system using the Jacobi Iteration Method.
2x + y + z = 5
3x + 5y + 2z = 15
2x + y + 4z = 8

Solution:
Rewriting the above equation in the form:

x = (5 - y -z)/2
y = (15 – 3x – 2z)/5
z = (8 – 2x – y)/4

If we assume initial values of x, y and z to be zero, then going by iteration we will get
the following results:

n x y Z
0 0 0 0
1 2.5 1.5 0.4
2 1.6 1.9 0.7
3 1.2 2 0.9
4 1.1 2 1
5 1 2 1

Hence, x = 1, y = 2 and z = 1.

103
Page

Prepared By
Er. Shree Krishna Khadka
Assignment 6

Full Marks: 50
Pass Marks: 25

18. Use Gauss Elimination Method to solve:

2w +2x + y + 2z = 7
w - 2x - z = 2
3w - x -2y – z = 3
w - 2z = 0 [7]

19. A system of simultaneous linear equations are given below:

2x + y – z = -1
x -2y + 3z = 9
3x – y + 5z = 14

a) Solve by Gauss Elimination with pivoting and [5]

b) Solve by Gauss Jordan Method, Compare two methods. [5]

20. Decompose the matrix [ ] into the form DoLittle/Crout

LU and hence solve the system AX = B, where B = [4 8 10]T [8]

21. Determine the Eigen Values and corresponding Eigen Vectors for the

matrix: A = [ ]. [5]

22. Compute the singular value decomposition of the matrix: [ ]. [10]

23. Find the solution to three decimals, of the system:

83x + 11y - 4z = 95
7x + 52y + 13z = 104
3x + 8y +29z = 71
a) Using Jacobi Iteration Method [4]
b) Using Gauss Seidel Iteration Method and [4]
[2]
104

c) Compare the results and execution time.

Page

Prepared By
Er. Shree Krishna Khadka
`

Numerical Solution of
Ordinary Differential Equations

Contents:

o Introduction
o Initial Value Problems
o Solution by Taylor’s Series Method
o Euler’s Method
o Heun’s Method
o Runge Kutta Method
o Predictor Corrector Method
o Solution of Simultaneous & Higher Order Equations
o Boundary Value Problem
o Finite Difference Method
o Shooting Method
105
Page

Prepared By
Er. Shree Krishna Khadka
INTRODUCTION: DIFFERENTIAL EQUATIONS
Most of the real time physical systems are expressed in terms of rate of change. The
Mathematical models that describe the state of such systems are often expressed in
terms of not only certain system parameters but also their derivatives. Such model
which used differential calculus to express relationship between variables is known as
differential equation.

Types of Differential Equations:

1.) Ordinary Differential Equations

The equation which involves an independent variable and dependent variable with
its ordinary derivatives of any order is called ordinary differential equation.
Examples of such equations are:

a) Law of Cooling: : The rate of loss of heat from a liquid is

proportional to the difference of temperature between the liquid and the
surroundings.

b) Law of Motion: : The time rate change of velocity of a moving

body is proportion to the force exerted by the body.

c) Kirchhoff’s Law for and Electric Circuit: : The voltage across

an electric circuit containing an inductance L and a resistance R.

The general form of the ordinary differential equation is: F(x, y, y’, y’’, … yn) = 0.
Where, ‘x’ is independent variable, y is dependent and y’, y’’ … are derivatives of ‘y’
with respect to ‘x’.

2.) Partial Differential Equations

The equation which involves two or more independent variables, dependent
variables and its partial derivative is called Partial Differential Equation. Examples
are:
a) Heat Flow in a Rectangular Plate: etc.

More about partial differential equations, we will discuss in next chapter.

106
Page

Prepared By
Er. Shree Krishna Khadka
Order and Degree of Ordinary Differential Equation
The order of a differential equation is the highest derivative that appears in the
equation. When the equation contains only a first derivative, it is called a first
order differential equation. On the other hand, the degree of a differential equation
is the power of the highest order derivative.

For example:

xy’’ + y2y’ = 2y + 3 is a 2nd order 1st degree equation, while:

(y’’’)2 + 5y’ = 0 is a 3rd order, 2nd degree equation.

Linear and Non Linear Differential Equation

A differential equation is said to be linear when it does not contain terms involving
the products of the dependent variable or its derivative.

On the other side, equations avoiding the above rules are known to be non linear
equations. For example: y’’ + 3y’ = 2y + x2 is a 2nd order 1st degree linear equation.
But the equation y’’ + (y’)2 = 1 and y’ = -ay2 are non linear equations.

General and Particular Solution

A solution to a differential equation is a relationship between the dependent and
independent variables that satisfy the differential equation. The solution that
contains arbitrary constants and is not unique is known as the general solution.
For example: y’ = 6x +1 has number of solutions as, y = 3x2 + x +2, y = 3x2 + x – 10…
Therefore, y = 3x2 + x + C is considered as general solution.

On the other part, if the values of the constants are known, then on substitution of
these values in the general solution, a unique solution is obtained which is known
as particular solution.

One Step and Multi Step Method of Solution

In one step methods, we use information from only one preceding point, i.e. to
estimate the value of yi, we need the conditions at the previous point yi-1 only.
Multistep methods use information at two or more previous steps to estimate a
value.
107
Page

Prepared By
Er. Shree Krishna Khadka
SOLUTION BY TAYLOR’S SERIES METHOD
Let us consider a differential equation in the form: y’ = f(x, y) with the initial condition
y(0) = y(x = 0) = y(x0) = y0. If y(x) is the exact solution, then we can expand a function
y(x) about a point x = x0 using Taylor’s Theorem of Expansion.

... (i)

Now, we must repeatedly differentiate f(x, y) implicitly with respect to x and evaluate
them at x0. Then equation (i) gives a power series of y.

Example:
From the Taylor’s Series for y(x), find y(0.1) correct to four decimal places if
y(x) satisfies: y’ = x – y2 and y(0) = 1.

Solution:
The Taylor’s Series for y(x) is given by:

We have:
y’(x) = x – y2 y’(0) = -1
y’’(x) = 1 - 2yy’ y’’(0) = 3
y’’’(x) = -2yy’’-2y’2 y’’’(0) = -8
yiv(x) = -2yy’’’ – 6y’y’’ yiv(0) = 34
yv(x) = -2yyiv – 8y’y’’’ – 6y’’2 yv(0) = -186 and so on…

Using these values, equation (i) gives a power series as follows:

Now: for the value of y(0.1) correct up to four decimal place, it is found that the terms
up to x4 should be considered, then we will have: y(0.1) = 0.9138 with the truncation
of (31/20)x5 0.00005.

/Home Work/
Given the differential equation: y’’ – xy’ – y = 0 with the conditions y(0) = 1 and
y’(0), use Taylor’s Series Method to determine the value of y(0.1).
108
Page

Prepared By
Er. Shree Krishna Khadka
EULER’S METHOD: (1ST ORDER RUNGE KUTTA METHOD)
The solution of differential equation by Taylor’s Series Method gives the equation in
the form of power series. We will now discuss the methods which give the solution in
the form of a set of tabulated values. Euler’s Method is one of the simplest one-step
methods and it has limited application because of its low accuracy.

True Values
y(x) Actual Solution

F(x)
y2
Errors
y0
y1
x
x0 x1 x2
h h

Fig: Illustration of Euler’s Method

Let us consider the first two terms of the Taylor’s Series Expansion as:

y(x) = y(x0) + y’(x0)(x – x0) … (i)

If we have given the differential equation y’(x) = f(x, y) with the initial condition by
y(x0) = y0. Then: y’(x0) = f(x0, y0).
Hence, equation (i) becomes: y(x) = y(x0) + f(x0, y0).(x – x0)
And also : y(x1) = y(x0) + f(x0, y0).( x1 – x0)
i.e. : y1 = y0 + f(x0, y0).h
Similarly, : y2 = y1 + f(x1, y1).h
In general, : yi+1 = yi + f(xi, yi).h … (ii)

Which is the main formula given by the Euler’s Method.

Example:
Given the equation y’ = 3x2 + 1 with y(1) = 2, estimate y(2) bye Euler’s Method
using (i) h = 0.5 and (ii) h = 0.25.
109
Page

Prepared By
Er. Shree Krishna Khadka
Solution:
Case I: For h = 0.5
y(1) = 2
y(1.5) = 2 + 0.5[3(1.0)2 + 1] = 4.0
y(2.0) = 4.0 + 0.5[3(1.5)2 + 1] = 7.875

Case II: For h = 0.25

y(1) = 2
y(1.25) = 2 + 0.25[3(1.0)2 + 1] = 3.0
y(1.50) = 3.0 + 0.25[3(1.25)2 + 1] = 5.42188
y(1.75) = 5.42188 + 0.25[3(1.50)2 + 1] = 7.35938
y(2.0) = 7.35938 + 0.25[3(1.75)2 + 1] = 9.90626

/Howe Work/For y’ = 2y/x, find y(2) at h = 0.25/

HEUN’S METHOD: (MODIFIED EULER’S METHOD/2ND ORDER RUNGE KUTTA METHOD)

Since, Euler’s Method does not require any differentiation and is easy to implement on
computers. However, it major weakness is large truncation errors. This is due to its
linear characteristic because it uses only the first two terms of the Taylor’s Series. So,
Heun’s Method is considered to be an improvement to Euler’s Method.
In Euler’s Method, the slope at the beginning of the interval is used to extrapolate yi to
yi+1 over the entire interval by the formula:
yi+1 = yi + m1h … (i)
An alternative is use the line which is parallel to the tangent at the point (xi+1, y(xi+1))
to extrapolate from yi, to yi+1 as shown in the figure by the formula:
yi+1 = yi + m2h … (ii)
y(x)

F(x)
y2
y1
y0
y1'
x
x0 x1 x2
h h

Fig: Illustration of Heun’s Method

110
Page

Prepared By
Er. Shree Krishna Khadka
The estimate given by equation (ii) appears to be overestimated.

Now, a third approach is to use a line whose slope is the average of the slope at the
end points of the interval. Then:

yi+1 = yi + … (iii)

As shown in figure, this gives a better approximation to yi+1. This approach is known
as Heun’s Method. Now from equations (i) and (ii), equation (iii) becomes:

yi+1 = yi + [ [ …(iv)

Here, the term yi+1 is appearing on both sides of the equation (iv) and therefore, yi+1
cannot be evaluated until the value of yi+1 inside the function is available.
This value can be predicted using the Euler’s Formula given in equation (i). So, finally,
equation (iv) becomes:

yi+1 [ …(v)

Hence, equation (v) is called improved version of Euler’s Method. It is also classified as
a one – step Predictor Corrector Method. Equation (i) is known as the predictor
equation whereas equation (v) is known as corrector equation.

Example:
Given the equation: y’ = 2y/x. Estimate y(2) using Heun’s Method at h = 0.25.
Solution:
y' = f(x, y) = 2y/x : For: x0 = 0, y0 = 2 : h= 0.25
Iteration: 1
m1 = (2X2)/1 = 4.0 ye(1.25) = 2 + 0.25(4.0) = 3.0
m2 = (2X3.0)/1.25 = 4.8 y(1.25) = 2 + (0.25)/2(4.0 + 4.8) = 3.1
Iteration: 2
m1 = (2X3.1)/1.25 = 4.96 ye(1.5) = 3.1 + 0.25(4.96) = 4.34
m2 = (2X4.34)/1.5 = 5.79 y(1.5) = 3.1 + (0.25)/2(4.96 + 5.79) = 4.44
Iteration: 3
m1 = (2X4.4)/1.5 = 5.92 ye(1.75) = 4.44 + 0.25(5.92) = 5.92
m2 = (2X5.92)/1.75 = 6.77 y(1.75) = 4.44 + (0.25)/2(5.92 + 6.77) = 6.03
Iteration: 4
m1 = (2X6.03)/1.75 = 6.89 ye(2) = 6.03 + 0.25(6.89) = 7.75
m2 = (2X7.75)/2 = 7.75 y(2) = 6.03 + (0.25)/2(6.89 + 7.75) = 7.86
111

Therefore: y(2) = 7.86

Page

Prepared By
Er. Shree Krishna Khadka
FOURTH ORDER RUNGE KUTTA METHOD
It is also a family of one step methods used for numerical solution of initial value
problems. It is also based on the general form of the extrapolation equation:

y i+1 = yi + mh …(i)

Where, ‘m’ represents the slope that is weighted averages of the slopes at various
points in the interval ‘h’. Runge Kutta (RK) methods are known by their order. For
example, an RK method is called rth order RK method when slopes at ‘r’ points are
used to construct the weighted average slope ‘m’. In Euler’s method we used only one
slope at (xi, yi) and Heun’s method employs slopes at two end points of the interval so,
they are called first-order and second-order RK method.

Higher the order better would be the accuracy of estimates. Therefore, selection of
order for using RK methods depends on the problem under consideration.

So, in 4th order RK method, we use four slopes and they are given by:

( )
( )

Then, the general form of extrapolation equation (i) will be:

… (ii)

Example

Given y’(x) = x2 + y2, y(0) =0 using 4th order RK method, estimate y(0.4), take h =
0.2

Solution:

We have given: f(x, y) = x2 + y2 and y0 = 0 at x0 =0

To find: yn = ? at xn = 0.4 for h = 0.2
112

Total number of iteration (n) = (xn – x0)/h = (0.4 - 0)/0.2 = 2

Page

Prepared By
Er. Shree Krishna Khadka
For iteration 1, (i =0)

( ) ( ) ( )

( ) ( ) ( )

For Iteration 2, (i =1 x1 = 0.2, y1 = 0.002667)

( ) ( )
( )

( ) ( )
( )
113
Page

Prepared By
Er. Shree Krishna Khadka
SIMULTANEOUS FIRST ORDER DIFFERENTIAL EQUATIONS
Mathematical models of many applications involve a system of several first order
differential equations. They may be represented as follows:

…………………………………………………………

…………………………………………………………

These equations should be solved for y1(x), y2(x) … yn(x) over interval (a, b).

Let us consider two such simultaneous differential equations, given by:

&

Given that: y(x0) = y0 and z(x0) = z0

Then we follow the solution approach given by the following algorithm.

( )
( )

Then:
… (i)

( )
( )
114

Then:
… (ii)
Page

Prepared By
Er. Shree Krishna Khadka
Example:

Find y(0.1) and z(0.1) from the system of equations: y’ =x +z and z’ = x – y2 with
y(0) = 2, z(0) = 1 and h = 0.1

Solution:
We have, x0 = 0, y0 = 2, z0 = 1 and h = 0.05

m1 = f(x0, y0, z0) = f(0, 2, 1) = (x0 + z0) = 0 + 1 = 1

l1 = g(x0, y0, z0) = g(0, 2, 1) = (x0 – y02) = 0 – 22 = -4

m2 = f(0.05, 2.05, 0.8) = 0.85

l2 = g(0.05, 2.05, 0.8) = -4.1525

Similarly,

m3 = 0.84 l3 = -4.22
m4 = 0.687 l4 = -4.244

Finally,

And

115
Page

Prepared By
Er. Shree Krishna Khadka
SOLUTION OF ORDINARY DIFFERENTIAL EQUATION AS A BOUNDARY VALUE PROBLEM:

We have seen that we require ‘m’ conditions to be specified in order to solve ‘m’ order
differential equation. And all the ‘m’ conditions were specified at one point, x = x0 and,
therefore we call this problem as an initial value problem. It is not always necessary to
specify the condition at one point of the independent variable. They can be specified at
different points in the interval (a, b) and, therefore, such problems are called the
boundary value problems. A large number of problems fall into this category.

In this topic, we will only discuss two special methods to solve such ordinary
differential equations they are:
a) Finite Difference Method
b) Shooting Method

FINITE DIFFERENCE METHOD

In finite difference method, the derivatives are replaced by their finite difference
equivalents, thus converting the differential equation into a system of algebraic
equations. Here, we will use the following “Central Difference Approximations”.

… (i)

… (ii)

These are the second order equations and the accuracy of estimates can be improved
by using higher order equations. The interval (a, b) is divided into ‘n’ sub-intervals,
each of width ‘h’. Then:

xi = x0 + ih = a + ih
yi = y(xi) = y(a + ih)
y0 = y(a)
yn = y(a + nh) = y(b)

Now, the difference equation is written for each of the interval points i = 1 to (n-1). If
the differential equation is linear, it has (n-1) unknowns y1, y2… yn-1, which can be
solved by any of the elimination methods. Smaller the size of ‘h’, more the sub-
intervals and, therefore, more are the equations to be solved. However, a smaller ‘h’
116

yields better estimates.

Page

Prepared By
Er. Shree Krishna Khadka
Example:
Given the equation: y’’ = with y(0) = 0 and y(1) = 0. Estimate the values of
y(x) at x = 0.25, 0.5 and 0.75

Solution:
We have: h = 0.25 and
y0 = y(0) = 0
y1 = y(0.25)
y2 = y(0.5)
y3 = y(0.75)
y4 = y(1) = 0

We also have:

For, i =1, x = 0.25:

Then: y2 – 2y1 = 0.0665 … (i)

For, i =2, x = 0.5:

Then: y3 – 2y2 + y1 = 0.0803 … (ii)

For, i =3, x = 0.75:

Then: – 2y3 + y2 = 0.0665 … (iii)

Solving, these three simultaneous algebraic equations, we will get:

y1 = y(0.25) = -0.1175
y2 = y(0.5) = -0.1684
y3 = y(0.75) = -0.1391
117
Page

Prepared By
Er. Shree Krishna Khadka
SHOOTING METHOD:
In shooting method, the given boundary value problem is first converted into an
equivalent initial value problem and then solved using any of the methods discussed
in the previous chapter.

Let us consider the equation:

y’’ = f(x, y, y’) with y(x0) = y(a) = A and y(xn) = y(b) = B

By letting y’ = z, we obtain the following set of two equations:

y' = z = f1(x, y, z) and z’ = y’’ = f2(x, y, z)

In order to solve this set as an initial value problem, we need two conditions at x = a.
We have one condition y(a) = A and, therefore, require another condition for z at x = a.

Let us assume: z(x0) = z(a) = M1(Just a guess but represent the slope y’(x) at x =a)
Thus, the problem is reduced to a system of first order equation with the initial
conditions as:

y' = z = f1(x, y, z) with y(a) = A

z’ = y’’ = f(x, y, z) with z(a) = M1 (= y’(a))

Now, using z(a) = z(x0) = M1, estimate y(b) = y(xn) = B1.

If B1 = B (Exact Solution), else

Suppose z(a) = z(x0) = M2(Another Guess) and estimate y(b) = y(xn) = B2

If B2 = B (Exact Solution), else

Estimate ‘M’ using the following relation:

Now, with z(x0) = z(a) = M3, we can again obtain the solution of y(x).
118
Page

Prepared By
Er. Shree Krishna Khadka
Example:
Using shooting method, solve the equation: y’’ = 6x2, with y(0) = 1 and y(1) = 2 in
the interval (0, 1) for y(0.5) taking h = 0.5

Solution:

We have given:
For x0(=x(a)) = 0, y0(=y(0)) = 1,
For xn(=x(b)) = 1, y0(=y(1)) = 2 = B and h = 0.5

Assume: y’ = z = f1(x, y, z)
Then: y’’ = z’ = 6x2 = f2(x, y, z)

Let: z(0) = 1.5 (=M1)

So, using Euler’s Method:
y1 = y(0.5) = y0 + hf1(x0, y0, z0) = 1 + 0.5X1.5 = 1.75 and
z1 = z(0.5) = z0 + hf2(x0, y0, z0) = 1.5 + 0.5X6X02 = 1.5

Therefore:
y2 = y(1) = y1 + hf1(x1, y1, z1) = 1.75 + 0.5X1.5 = 2.50 = B1

Since, B1 > B, assume another guess: z(0) = 0.5 (=M2)

So, again using Euler’s Method:
y1 = y(0.5) = y0 + hf1(x0, y0, z0) = 1 + 0.5X0.5 = 1.25 and
z1 = z(0.5) = z0 + hf2(x0, y0, z0) = 0.5 + 0.5X6X02 = 0.5

Therefore:
y2 = y(1) = y1 + hf1(x1, y1, z1) = 1.25 + 0.5X0.5 = 1.50 = B2

Again, B2 < B, so finding M using the following relation:

M = 1 = z(0)

Finally, y1 = y(0.5) = y0 + hf1(x0, y0, z0) = 1 + 0.5X1 = 1.5

119
Page

Prepared By
Er. Shree Krishna Khadka
Assignment 7

Full Marks: 40
Pass Marks: 20

24. Given the differential equation y’’ – xy’ – y = 0 with initial conditions
y(0) = 1 and y’(0) = 0, use Taylor’s series method to determine the
value of y(0.1). [5]

25. Use Euler’s method to solve the differential equation y’ = - y' with the
initial condition y(0) = 1 with step size 0.01 and find the value of
y(0.04). [5]

26. Solve the differential equation 10y’’ + (y’)2 + 6x = 0 with y(0) = 1 and
y’(0) = 0 by Heun’s method to estimate y(0.2) using h = 0.1
[7]

27. Use 4th order Runge Kutta method to estimate y(0.5) of the differential
equation y’ = y + _/y with y(0) = 1. Take h = 0.25. [5]

28. Solve the following simultaneous first order differential equations to

estimate y(0.2) and z(0.2) using any method of your choice.

y' = z with y(0) = 0 and z’ = yz + x2 + 1 with z(0) = 0 [8]

29. A body of mass 2 Kg is attached to a spring with a spring constant of 10.

The differential equation governing the displacement of the body ‘y’
and time ‘t’ is given by: y’’ + 2y’ + 5y = 0. Find the displacement ‘y’ at
time t = 0.5, 1, 1.5 using finite difference method. Given that y(0) = 2,
y’(0) = -4 and y(2) = 4

(Note: Make Correction if needed) [10]

120
Page

Prepared By
Er. Shree Krishna Khadka
`

Numerical Solution of
Partial Differential Equations

Contents:

o Introduction
o Finite Difference Approximation
o Solution of Elliptic Equations
o Laplace’s Equation
o Poisson’s Equation
o Solution Parabolic Equation
o Solution of Hyperbolic Equation
o Finite Difference Method
o Shooting Method
121
Page

Prepared By
Er. Shree Krishna Khadka
INTRODUCTION: PARTIAL DIFFERENTIAL EQUATION

Physical phenomena in applied science and engineering when formulated into

mathematical models fall into a category of systems known as partial differential
equations that involves more than one independent variable which determine the
behaviour of the dependent variable as described by their partial derivative contained
in the equation.

Examples: Heat flow in a rectangular plate

The model for heat flow in a rectangular plate that is heated is given by:

Where u(x, y) denotes the temperature at point (x, y) and f(x, y) is the heat source.
Here, the rate of change of a variable is expressed as a function of variables and
parameters. Although most of the differential equations may be solved analytically in
their simplest form, analytical techniques fail when the models are modified to take
into account the effect of other conditions of real life situations. In all such cases,
numerical approximation of the solution may be considered as a possible approach.

If we represent the dependent variable as ‘f’ and the two independent variables as ‘x’
and ‘y’, then we will have three possible second order partial derivatives:

Then we can write a second order equation involving two independent variables in
general form as:

… (i)

Where the coefficients a, b and c may be constants or function of ‘x’ and ‘y’. Depending
on the values of these coefficients equation (i) can be classified into one of the three
types of equations, namely:

a) Elliptic Equation: If b2 – 4ac < 0

b) Parabolic Equation: If b2 – 4ac = 0
c) Hyperbolic Equation: If b2 – 4ac > 0
122
Page

Prepared By
Er. Shree Krishna Khadka
FINITE DIFFERENCE APPROXIMATION
In finite difference method, we replace derivatives that occur in the partial differential
equation by their finite difference equivalents. We then write the difference equation
corresponding to each ‘grid point’ (where derivative is required) using functional
values at the surrounding grid points. Solving these equations simultaneously gives
the values for the function at grid point.

Consider a two-dimensional solution domain as shown in the figure below. The

domain is split into regular rectangular grids of width ‘h’ and height ‘k’. The pivotal
values at the points of intersection (grid points or node) are denoted by fij which is a
function of the two-space variable ‘x’ and ‘y’.

xi+1 = xi + h
y
yi+1 = yi + k

yj+1
k
yj
yj-1
yj-2
x
xi-2 xi-1 xi xi+1 xi+2
h
Fig: Two Dimensional Finite Difference Grid

If the function f(x) has continuous fourth derivative, then its first and second
derivatives are given by the following central difference approximations.

When ‘f’ is a function of two variables ‘x’ and ‘y’, the partial derivatives of ‘f’ with
respect to ‘x’ (or ‘y’) are the ordinary derivatives of ‘f’ with respect to ‘x’ (or ‘y’) when
‘y’ (or ‘x’) does not change. So, we can use above equation to determine derivatives
with respect to ‘x’ and in the ‘y’ direction. Thus, we have:
123

( ) ( )
Page

( )
Prepared By
Er. Shree Krishna Khadka
 ( ) ( )
( )

( ) ( )
( )

( ) ( )
( )

( ) ( ) ( ) ( )
( )

SOLUTION OF ELLIPTIC EQUATIONS

Elliptic equations are governed by conditions on the boundary of closed domain. We
consider here the two most commonly encountered elliptic equations, namely:
Laplace’s Equation and Poisson’s Equation.

Laplace Equation:
The general second order partial differential equation (i), when a = 1, b = 0, c = 1 and
F(x, y, fx, fy) = 0 becomes:

… (ii) { ( )}

The operator is called the Laplacian Operator and equation (ii) is called Laplace’s
Equation. To solve the Laplace’s Equation on a region in the xy-plane, we subdivide
the region as shown in the figure below. Consider the portion of the region near (xi, yi).
We have to approximate: .

y
h f1
yj+1

f4 f5 f2
yj
k
f3
yj-1

x
xi-1 xi xi+1
Fig: Solution of Laplace’s Equation
124

Replacing the second order derivatives by their finite difference equivalents at point
Page

(xi, yi), we obtain.

Prepared By
Er. Shree Krishna Khadka
i.e. f2 + f4 + f1 + f3 – 4f5 = 0 (for h = k)

Example:
Consider a steel plate of size 15X15 cm. If two of the sides are held at 100oC and
the other two sides are held at 0oC. What is the steady state value of
temperature at interior knots, assuming a grid size of 5X5 cm.

5
100 100

f1 f2
100

0
5
f3 f4
100 0

0 0

Solution:

At point 1:
f2 + f3 + 100 + 100 – 4f1 = 0 f2 + f3 - 4f1 + 200 = 0 … (i)

At point 2:
f1 + f4 + 100 + 0 – 4f2 = 0 f1 - 4f2 + f4 + 100 = 0 … (ii)

At point 3:
f1 + f4 + 100 + 0 – 4f3 = 0 f1 - 4f3 + f4 + 100 = 0 … (iii)

At point 4:
f3 + 0 + 0 + f2 – 4f4 = 0 f2 + f3 - 4f4 = 0 f2 = 4f4 - f3 … (iv)

Putting the values of of f2 from equation (iv) into equations (i) and (ii), we get:

f2 + f3 - 4f1 + 200 = 0 4f4 - f3 + f3 - 4f1 + 200 = 0 f1 – f4 = 50 … (vi)

f1 - 4f2 + f4 + 100 = 0 f1 – 4(4f4 - f3 ) + f4 + 100 = 0 f1 – 4f3 – 15f4 = -100 … (vii)
f1 - 4f3 + f4 + 100 = 0 … (viii)
125

Now, solving equations (vi), (vii) and (viii) we will get: f1 = 75, f3 = 50 and f4 = 25
Putting f3 and f4 in equation (iv), we get: f2 = 50.
Page

Prepared By
Er. Shree Krishna Khadka
Poisson’s Equation
The general second order partial differential equation, when a = 1, b = 0, c = 1 and F(x,
y, fx, fy) = g(x, y) becomes:

… (iii) { ( )}

Equation (ii) is called Poisson’s Equation. Now, Laplace’s equation may be modified to
solve Poisson’s Equation. Then the finite difference formula for soving Poisson’s
Equation takes the form:

i.e. f2 + f4 + f1 + f3 – 4f5 = h2gij (for h = k)

Example:
Solve the Poisson’s Equation over the square domain and
with f = 0 on the boundary and h = 1.
y

f1 f2

0
0
f3 f4

x
0

Solution:
At point 1: 0 + 0 + f2 + f3 – 4f1 = 2(1)2(2)2 f2 + f3 - 4f1 = 8 … (i)
At point 2: 0 + 0 + f1 + f4 – 4f2 = 2(2)2(2)2 f1 - 4f2 + f4 = 32 … (ii)
At point 3: 0 + 0 + f1 + f4 – 4f3 = 2(1)2(1)2 f1 - 4f3 + f4 = 2 … (iii)
At point 4: 0 + 0 + f2 + f3 – 4f4 = 2(2)2(1)2 f2 + f3 - 4f4 = 8 … (iv)

On solving these simultaneous equations by elimination method, we will get the

answers.

f1 = -22/4 f2 = -43/4
f3 = -13/4 f4 = -22/4
126
Page

Prepared By
Er. Shree Krishna Khadka
SOLUTION OF PARABOLIC EQUATIONS
Elliptic equations describe the problems that are time independent. Such problems
are known as steady state problems. But we come across problems that are not steady
state. This means that the function is dependent on both space and time. Parabolic
equations, for which b2 – 4ac = 0 describe the problems that depend on space and time
variables. The popular case for parabolic type equation is the study of heat flow in one
dimensional direction in an insulated rod. Such problems are governed by both
boundary and initial conditions.

f(x, t)

Insulation
Rod
f(0, t) f(L, t)

x=0 x x=L

Fig: Heat flow in a rod

If ‘f’ represent the temperature at any point in rod whose distance from the left end is
‘x = 0’ to ‘x = L’. Heat is flowing from left to right under the influence of temperature
gradient. The temperature f(x, t) in the rod at the position ‘x’ and time ‘t’, is governed
by the heat equation:

… (i)

Where, K1 = Coefficient of thermal conductivity

K2 = Specific Heat
K3 = Density of the material

Equation (i) can be written as: K.fxx(x, t) = ft(x, t) … (ii), where K = K1/(K2K3)

The initial condition will be the initial temperatures at all points along the rod i.e. f(x,
0) = f(x) for 0 x L.

The boundary conditions f(0, t) and f(L, t) describe the temperature at each end of the
rod as functions of time. If they are held constant, then:

f(0, t) = C1, for 0 t and f(L, t) = C2, for 0 t

127

We can solve the heat equation given by equation (ii) using finite difference formula
as below:
Page

Prepared By
Er. Shree Krishna Khadka
 y
h

f1
yj+1

f4 f2
τ
f3
yj

yj-1
x
xi-1 xi xi+1

Fig: Solution of Parabolic Equation

As: K.fxx(x, t) = ft(x, t)

i.e.

So, at point f2:

i.e. ( )

i.e.

i.e. [ ]

For, ,

i.e.

Example:
Solve the equation: 2fxx(x, t) = ft(x, t) for 0 < t < 1.5 and 0 < x < 4.
Given, f(x, 0) = 50(4 – x) for 0 < x < 4;
f(0, t) = 0 for 0 < t < 1.5;
f(4, t) = 0 for 0 < t < 1.5.

Take h = 1.

Solution:

Here, we have: K = 2 and h = 1

So, for: , = 1/2K = 1/(2X2) = 0.25

128
Page

Prepared By
Er. Shree Krishna Khadka
Now constructing the grid with h = 1 and = 0.25

t h=1
12.5 12.5 12.5
1.50

12.5 25 12.5
1.25
t\x 0.0 1.0 2.0 3.0 4.0
25 25 25 0.00 0 150 100 50 0
0.10
0.25 0 50 100 50 0
τ= 0.25 0.50 0 50 50 50 0
25 50 25
0 0.75 0
0.75 0 25 50 25 0
50 50 50 1.00 0 25 25 25 0
0.50
1.25 0 12.5 25 12.5 0
50 100 50 1.50 0 12.5 12.5 12.5 0
0.25

0 150 100 50 0
x
1 2 3 4

SOLUTION OF HYPERBOLIC EQUATIONS

Hyperbolic equations model the vibration of structures such as buildings, beams and
machines. We consider here the case of a vibrating string that is fixed at both the ends
as shown in figure below:

f =f(x, t)

x
x=0 x=L

Fig: Displacement of vibrating spring

The lateral displacement of string ‘f’ varies with time ‘t’ and distance ‘x’ along the
string. The displacement f(x, t) is governed by the wave equation:

… (i)

Where, T is the tension in the string and is the mass per unit length. Hyperbolic
problems are also governed by both boundary and initial conditions, if time is one of
129

the independent variables. Two boundary conditions for the vibrating string problem
under consideration are:
Page

Prepared By
Er. Shree Krishna Khadka
f(0, t) = 0 for 0 t b and f(L, t) = 0 for 0 t b

Two initial conditions are:

f(x, 0) = f(x) for 0 x a and ft(x, 0) = g(x) for 0 x a

Now, we can solve the heat equation given by equation (ii) using finite difference
formula as below:

y
h

f1
yj+1

f4 f0
τ
yj f2

yj-1 f3
x
xi-1 xi xi+1

Fig: Solution of Hyperbolic Equation

As the wave equation is given by:

So, at point f0

i.e. ( ) ( )

i.e.

i.e. [ ]

For, ,

i.e.
This can be also written in coordinate form as:
130

( ) …(ii)
Page

Prepared By
Er. Shree Krishna Khadka
Starting Values
Here, we need two row of starting values, corresponding to j = 1 and j = 2 in order to
compute the values at the third row. First row is obtained using the condition: f(x, 0 )
= f(x). The second row can be obtained using the second initial condition as given by:
ft(x, 0) = g(x)
As we know: (for t = 0 only). By this, substituting the value
of fi, -1 in equation (ii) for j = 0, we will get at t = t1:
( )
As for many cases: gi = 0, therefore: ( )

Example:
Solve the wave equation: 4fxx = ftt for 0 , take h = 1.
Given: f(0, t) = 0 and f(5, t) = 0
f(x, 0) = x(5 – x)
ft(x, 0) = 0 = gi

Solution:

Here we have, h = 1, T = 4 and = 1

So, for: ,

Hence, constructing the table, we will have:

t
h=1

-3 -5 -5 -3
2.00
τ= 0.5
-1 -2 -2 -1 t\x 0.0 1.0 2.0 3.0 4.0 5.0
1.50
0.00 0 4 6 6 4 0
1 2 2 1 0 0.50 0 3 5 5 3 0
0
1.00
1.00 0 1 2 2 1 0
3 5 5 3 1.50 0 -1 -2 -2 -1 0
0.50
2.00 0 -3 -5 -5 -3 0
0 4 6 6 4 0
x
1 2 3 4 5
131
Page

Prepared By
Er. Shree Krishna Khadka
Assignment 8

Full Marks: 30
Pass Marks: 15
Grace Mark: 4

30. Solve the steady-state temperature in a rectangular plate of 8cm x 10

cm, if one 10 cm side is held at 50oC, and the other 10 cm side is held at
30oC and the other two sides are held at 10oC. Assume square grids of
size 2cm x 2 cm. [5]

(Ref: Elliptic Equation – Solution of Laplace’s Equation)

31. Solve the equation: with F(x, y) = xy and f = 0 on

boundary. The domain is a square with corners at (0, 0) and (4, 4). Use
h = 1. [5]

(Ref: Elliptic Equation – Solution of Poisson’s Equation)

32. Estimate the values at grid points of the following equation using
Bender Schmidt Recurrence Equation. Assume h = 1.

a) 0.5fxx – ft = 0
Given: f(0, t) = -5, f(5, t) = 5 [8]
f(x, 0) = -5, for x: [0, 2.5]
= 5, for x: [2.5, 5]

(Ref: Parabolic Equation)

33. Solve the following wave equation using finite difference method.

4fxx = ftt
Given:
f(0, t) = 0 and f(1, t) = 0
f(x, 0) = sin( ) + sin(2 ) [8]
ft(x, 0) = 0

(Ref: Hyperbolic Equation)

132
Page

Prepared By
Er. Shree Krishna Khadka
`

Numerical Solution of
Integral Equations

Contents:

o Introduction
o Method of Degenerate Kernels
o Method of Generalized Quadrature
o Chebyshev Series Method
o The Cubic Spline Method
o Assignment 9
133
Page

Prepared By
Er. Shree Krishna Khadka
INTRODUCTION:

Any equation in which the unknown function appears under the integral sign is known
as an integral equation. General integral equation of first and second kind can be
represented by:

∫ … (i)

∫ …. (ii)

Here, the limits of integrals are constants; therefore it is called Fredholm Integral
Equation of first and second kinds respectively. In each case, f(x) is unknown and
occurs to the first degree but and the Kernel K(x, t) are known functions.

If the constant ‘b’ in above equations is replaced by ‘x’, the variable of integration, the
equations are called Volterra Integral Equation given by:

∫ …. (iii)

If in equation (ii), then the equation is called homogeneous, otherwise non-

homogeneous. For non-homogeneous equations ‘ is a numerical parameter whereas
for homogeneous equations, it is an eigen-value parameter.

If the kernel: K(x, t) is bounded and continuous, then the integral equation is said to be
non-singular. If the range of integration is infinite, or if the kernel violates the above
conditions, then the equation is said to be singular.

To solve an integral equation of any type is to find the unknown function satisfying that
equation. In this chapter we deal with the Fredholm Integral Equation, particularly
those of the second kind since, it occurs quite frequently in practice.
134
Page

Prepared By
Er. Shree Krishna Khadka
METHOD OF DEGENERATE KERNELS
This method is important in the theory of integral equations but does not seem to be
much useful in the numerical work, since the kernel is unlikely to have the simple
form in practical problems. In general, however, it is possible to take a partial sum of
Taylor’s series for the kernel.

Let us consider the integral equation of type:

∫ …(i)

A kernel K(x, t) is said to be degenerate if it can be expressed in the form:

∑ …(ii)

Substituting equation (ii) in equation (i), we obtain:

∫ ∑

∑ ∫ … (iii)

Let us suppose: ∫

Then equation (iii) becomes:

∑ ∫ … (iv)

Hence, the coefficients Ai are determined from equations (iv) and (iii), which will give
a system of ‘n’ equations with ‘n’ unknowns A1, A2 … An. When Ai are determined,
equation (iv) gives f(x).
135
Page

Prepared By
Er. Shree Krishna Khadka
Example:

Solve the integral equation: ∫ ( ) .

Solution:
Comparing the given integral equation with the original equation, we get:

K(x, t) = ( ) ( ) +(neglecting the

other terms of the Taylor’s Series)

i.e. K(x, t) … (i)

Hence, the given integral equation becomes:

∫ ( )

∫ ∫

… (ii)

Now, generating simultaneous equations:

∫ ∫ { }

… (iii) ∫

Again:

∫ ∫ { }

… (iv)

Solving equations (iii) and (iv) we get: K1 = 0.2522 and K2 = 0.1685

Hence, the solution of the given integral equation is:

136
Page

Prepared By
Er. Shree Krishna Khadka
METHOD OF USING GENERALIZED QUADRATURE

Let us consider the integral equation of type:

∫ … (i)

Since, a definite integral can be closely approximated by a Quadrature Formula and

obviously, different types of Quadrature Formulas can be employed (e.g. Trapezoidal,
Simpson etc). We approximate the integral term of equation (i) by a formula of the
form:

∫ ∑ … (ii)

Where Am and xm are the weights and abscissa, respectively. Consequently eqn (ii) can
be written as:

∑ … (iii)

Where t1, t2 … , tn are the points in which the interval (a, b) is subdivided. Further
equation (iii) must hold for all values of ‘x’ in the interval (a, b); in particular, it must
hold for x = t1, x = t2, … , x = tn. Hence we obtain:

∑ for i = 1, 2, … , n … (iv)

Which is a system of ‘n’ linear equations in the ‘n’ unknowns f(t1), f(t2) … f(tn). When
the f(ti) are determined, equation (iii) gives an approximation for f(x).
137
Page

Prepared By
Er. Shree Krishna Khadka
Example:
Solve: ∫

Solution:
Here, K(x, t) = (x + t)
The integral term can be directly approximated by the Quadrature Formula. For the
numerical solution, we divide the range [0, 1] into two equal subintervals so that h =
0.5. Let us apply composite trapezoidal rule to approximate the integral term, we
obtain:

{ ( ) } … (i)

Setting x = ti (t0 = 0, t1 = 0.5, t1 = 1) and f(ti) = fi, we will have:

a) For x = t0 = 0

{ ( ) }

{ ( ) }

{ ( ) }

… (a)

b) Similarly, for x = t1 = 0.5

… (b)

c) And for x = t2 = 1
… (c)

On solving, we get: f0 = -1/2, f1 = -5/6 and f2 = ½

Now, using these values in equation (i), we get the solution as:

{ ( ) ( )( ) ( )}

i.e. { ( ) ( )( ) ( )}

Note: Since, the integrand is a second degree polynomial in‘t’ so, if we implement the
138

composite Simpson’s 1/3 rule, it will give the exact solution i.e. f(x) = x - 1.
Page

Prepared By
Er. Shree Krishna Khadka
CHEBYSHEV SERIES METHOD & CUBIC SPLINE METHOD

Chebyshev Series Method: The Fredholm Integral Equation can also be manipulated
by Chebyshev Series Method, which is somewhat laborious. For example the integral
equation of type:

∫ ; where

- can be solved. However, this method can give better accuracy than the trapezoidal
method but for the smaller values of d, this method is unsuitable. Because, for example
with 32 subdivisions and d = 0.0001, the value obtained for x = 0 is 0. 04782
compared to the true value of 0.50015 that means: somewhat large fluctuation in the
results.

Cubic Spline Method: In contrast with the previous methods, cubic spline method
can be applied when the values of d are small. However, for large value of d, the
convergence would rather slowly to give the exact solution.

The spline method for the numerical solution of Fredholm Integral Equations is
potentially useful. Its application to more complicated problems will have to be
examined together with estimation to error in the method. It seems probable that the
condition of continuity of the kernel may be relaxed, and the advantage to be achieved
by using unequal intervals may also be explored. Finally the solution obtained by the
spline method can be improved upon by regarding it as the initial iterate in an
iterative method of higher order convergence.
139
Page

Prepared By
Er. Shree Krishna Khadka
Assignment 9

Full Marks: 30
Pass Marks: 15
Grace Marks: 4

1. Solve the following integral equations with degenerate kernels:

a) ∫
[10]

2. Solve the integral equation given in problem 1 by using general

Quadrature method for using:

a) Composite Trapezoidal Rule [8]

b) Composite Simpson’s Rule [8]

(Note: Make necessary assumptions)

140
Page

Prepared By
Er. Shree Krishna Khadka
Programming
Exercises

Contents:

o Horner’s Method
o Bisection Method
o Regular Falsi Method
o Newton Raphson Method
o Secant Method
o Fixed Point Iteration
o Linear Interpolation
o Lagrange Interpolation
141
Page

Prepared By
Er. Shree Krishna Khadka
LAB 1
------------------------------------------------------------------------------------
HORNER’S METHOD OF FINDING POLYNOMIAL AT GIVE VALUE
------------------------------------------------------------------------------------

#include<stdio.h>
#include<conio.h>
#include<math.h>

#define G(x) (x)*(x)*(x)-4*(x)*(x)+(x)+6

void main()
{
int n,i;
float x, a[10],p;
printf("--------------------------------------------------------------\n");
printf("HORNER'S METHOD OF FINDING THE POLYNOMIAL VALUE AT GIVEN POINT\n");
printf("--------------------------------------------------------------\n\n");
printf("\nINPUT DEGREE OF POLYNOMIAL:\t");
scanf("%d", &n);
printf("\nINPUT POLYNOMIAL COEFFICENTS WITH ORDER a[%d] to a[0]:\t", n);
for(i=n; i>=0; i--)
scanf("%f", &a[i]);
printf("\nTHE POLYNOMIAL IS:\t");
for(i=n; i>=0; i--)
{
printf("%f(x^%d) ",a[i],i);
}
printf("\n\nINPUT VALUE OF 'X' (EVALUATION POINT):\t");
scanf("%f", &x);
p=a[n];
for(i=n-1; i>=0; i--)
{
p=p*x+a[i];
}
printf("\nTHE POLYNOMIAL VALUE: f(x) = %f at x = %f\t", p,x);
}

OUTPUT
142
Page

Prepared By
Er. Shree Krishna Khadka
LAB 2

-------------------------------------------------------------------------------
SOLUTION OF A GIVEN EQUATION BY BISECTION METHOD
-------------------------------------------------------------------------------

#include<stdio.h>
#include<conio.h>
#include<math.h>

#define EPS 0.000001

#define F(x) log(x)-1

void main()
{
float xn,xp,xm,a,b,c,d;
int count;

begin:
printf("Enter Bracketing Values:\t");
scanf("%f%f",&xn,&xp);

a=F(xn);
b=F(xp);

if((a*b)>0)
{
printf("Entered Values Didnot Bracket The ROOT\n");
goto begin;
}
else
{
count=1;
printf("\ncount\t xn\t\t xp\t\t xm\t\t F(xm)\n");

iteration:

xm=((xn+xp)/2);

c=F(xm);

printf("\n%d\t %f\t %f\t %f\t %f\t", count, xn, xp, xm, F(xm));

if(c==0)
{
printf("\n\nROOT: %f\t", xm);
printf("\nITERATION: %d\t", count);
}
else
{
if(c<0)
143

xn=xm;
else
Page

xp=xm;

Prepared By
Er. Shree Krishna Khadka
 d=fabs((xn-xp)/xn);
if(d<EPS)
{
xm=((xn+xp)/2);
printf("\n\nROOT: %f\t", xm);
printf("\nITERATION: %d\t", count);
}
else
{
count=count++;
goto iteration;
}
}
}
}

OUTPUT

144
Page

Prepared By
Er. Shree Krishna Khadka
LAB 3

-----------------------------------------------------------------------------------
SOLUTION OF GIVEN EQUATION BY REGULAR FALSI METHOD
-----------------------------------------------------------------------------------

#include<stdio.h>
#include<conio.h>
#include<math.h>
#define EPS 0.000001
#define F(x) x*log(x)-1
void main()
{
float a,b,c,d,e,fa,fb;
int count;
begin:
printf("Enter Bracketing Values:\t");
scanf("%f%f",&a,&b);
fa=F(a);
fb=F(b);
if((fa*fb)>0)
{
printf("Entered Values Didnot Bracket The ROOT\n");
goto begin;
}
else
{
count=1;
printf("\ncount\t a\t\t b\t\t c\t\t F(c)\n");
iteration:
c=((a*fb)-(b*fa))/(fb-fa);
e=F(c);
printf("\n%d\t %f\t %f\t %f\t %f\t", count, a, b, c, e);
if(e==0)
{
printf("\n\nROOT: %f\t",c);
printf("\nITERATION: %d\t", count);
printf("\nFunction Value: %f\t", F(c));
}
else
{
if(e<0)
145

a=c;
else
Page

b=c;
Prepared By
Er. Shree Krishna Khadka
 d=fabs((b-a)/b);

if(d<EPS)
{
c=((a*fb)-(b*fa))/(fb-fa);

printf("\n\nROOT: %f\t", c);

printf("\nITERATION: %d\t", count);
printf("\nFunction Value: %f\t", F(c));
}
else
{
count=count++;
goto iteration;
}
}
}
}

OUTPUT

146
Page

Prepared By
Er. Shree Krishna Khadka
LAB 4

----------------------------------------------------------------------------------------
SOLUTION OF GIVEN EQUATION BY NEWTON RAPHSON METHOD
----------------------------------------------------------------------------------------

#include<stdio.h>
#include<conio.h>
#include<math.h>

#define EPS 0.000001

#define F(x) (x)*(x)-3*(x)+2
#define FD(x) 2*(x)-3

void main()
{
int count;
float x0, xn, fx,fn,fdx,e;

printf("\n");
printf("----------------------------\n");
printf("TITLE: NEWTON RAPHSON METHOD\n");
printf("----------------------------\n\n");

printf("--------------------------------------------------------------------------\n");
printf("OBJECTIVE: TO FIND THE SOLUTION OF GIVEN EQUATION BY NEWTON RAPHSON METHOD\n");
printf("--------------------------------------------------------------------------\n\n");

printf("\nINPUT INITIAL GUESS (X0):\t");

scanf("%f", &x0);

printf("\n\n");

count=1;
printf("Count\tx0\t\tx1\t\tError\t\tFunction Value\N");

begin:
fx=F(x0);
fdx=FD(x0);

xn=x0-(fx/fdx);
fn=F(xn);

e=fabs((xn-x0)/xn);
147

printf("\n%d\t%f\t%f\t%f\t%f",count,x0,xn,e,fn);
Page

Prepared By
Er. Shree Krishna Khadka
 if(e<EPS)
{
printf("\n\n");
printf("\nRoot = %f\t", xn);
printf("\nFunction Value = %f\t", F(xn));
printf("\nIteration = %d\t", count);
}
else
{
x0=xn;
count++;
goto begin;
}
}

OUTPUT

148
Page

Prepared By
Er. Shree Krishna Khadka
LAB 5
------------------------------------------------------------------------
SOLUTION OF GIVEN EQUATION BY SECANT METHOD
------------------------------------------------------------------------
#include<stdio.h>
#include<math.h>
#include<conio.h>
#define EPS 0.000001
#define F(x) (x)*(x)+4*(x)-10
void main()
{
int count;
float x1,x2,x3,f1,f2,f3;
printf("\n");
printf("-----------------------\n");
printf("TITLE: SECANT METHOD\n");
printf("-----------------------\n\n");
printf("---------------------------------------------------------------------\n");
printf("OBJECTIVE: TO FIND THE SOLUTION OF GIVEN EQUATION BY SECANT METHOD\n");
printf("---------------------------------------------------------------------\n\n");
printf("\nInput two starting values (x1 & x2):\t");
scanf("%f%f", &x1, &x2);
f1=F(x1);
f2=F(x2);
count=1;
printf("\nCount\t\tx1\t\tx2\t\tx3\t\tf(x3)\n");
begin:
x3=(f2*x1-f1*x2)/(f2-f1);
f3=F(x3);
printf("\n%d\t\t%f\t%f\t%f\t%f",count,x1,x2,x3,f3);
if(fabs((x3-x2)/x3)<EPS)
{
printf("\n\nRoot = %f\t", x3);
printf("\nFunction Value at %f\t = %f\t", x3, F(x3));
printf("\nNo of Iteration = %d\t", count);
}
else
{
x1=x2;
149

f1=f2;
x2=x3;
Page

f2=F(x3);

Prepared By
Er. Shree Krishna Khadka
 count++;
goto begin;
}
}

OUTPUT

150
Page

Prepared By
Er. Shree Krishna Khadka
LAB 6
-----------------------------------------------------------------------------------------------
SOLUTION OF GIVEN EQUATION BY FIXED POINT ITERATION METHOD
-----------------------------------------------------------------------------------------------

#include<stdio.h>
#include<math.h>
#include<conio.h>
#define EPS 0.000001
//#define G(x) pow(x,2)-5 //#define G(x) 5/x //#define G(x) pow(x,2)+x-5
#define G(x) (pow(x,2)+5)/(2*(x))

void main()
{
int count;
float x0, x, e, f;
printf("\n");
printf("-----------------------------------\n");
printf("TITLE: FIXED POINT ITERATION METHOD\n");
printf("-----------------------------------\n\n");
printf("-------------------------------------------------------------------------------\n");
printf("OBJECTIVE: TO FIND THE SOLUTION OF GIVEN EQUATION BY FIXED PNT ITERATION METHOD\n");
printf("-------------------------------------------------------------------------------\n\n");
printf("\nInput Initial Estimate(x0):\t");
scanf("%f", &x0);
count=1;
printf("\nitn\tx0\t\tx1\t\tError\t\tF(x)");
printf("\n");
iteration:
x=G(x0);
e=fabs((x-x0)/x);
f=G(x);
printf("\n%d\t%f\t%f\t%f\t%f",count,x0,x,e,f);
if(e>EPS)
{
x0=x;
count = count++;
goto iteration;
}
151

else
{
Page

printf("\n\nRoot is: %f\t", x);

Prepared By
Er. Shree Krishna Khadka
 printf("\nFunction value: %f\t", G(x));
printf("\nIteration: %d\t", count);
}
}

OUTPUT

152
Page

Prepared By
Er. Shree Krishna Khadka
LAB 7
-----------------------------------------------------
LINEAR INTERPOLATION TECHNIQUES
-----------------------------------------------------
#include<stdio.h>
#include<math.h>
#include<conio.h>
#define F(x) sqrt(x)
void main()
{
float x1, x2, fx1, fx2, fx, f, x, e;
printf("\n-------------------------------------\n");
printf("TITLE: LINEAR INTERPOLATION TECHNIQUE\n");
printf("-------------------------------------\n\n");
printf("-------------------------------------------------------------------------------\n");
printf("OBJECTIVE: TO FIND THE FUNCTIONAL VALUE AT GIVEN PNT USING LINEAR INTERPOLATION\n");
printf("-------------------------------------------------------------------------------\n\n");
printf("\nFunction F(x): sqrt(x)\n");
printf("\nAt Which Point You Want To Evaluate Function (x):\t");
scanf("%f", &x);
x1 = x-1; x2 = x+1;
fx1 = F(x1); fx2 = F(x2);
f = fx1 + (x-x1)*(fx2-fx1)/(x2-x1);
fx = F(x); e = fabs(f-fx);
printf("\nInterpolating Points (x1, x2):(%f, %f)", x1, x2);
printf("\nEvaluation Point (x):%f", x);
printf("\nFunctional Value F(x):%f", f);
printf("\nError From Exact value (Error):%f", e);
}

OUTPUT
153
Page

Prepared By
Er. Shree Krishna Khadka
LAB 8
--------------------------------------------------------
LAGRANGE INTERPOLATION TECHNIQUE
--------------------------------------------------------
#include<stdio.h>
#include<math.h>
#include<conio.h>
#define F(x) sqrt(x)
void main()
{
int n, i, j;
float x[10], y[10], lbp, lp, xp, sum;

printf("\n-----------------------------------------\n");
printf("TITLE: LAGRANGE'S INTERPOLATION TECHNIQUE\n");
printf("-----------------------------------------\n\n");
printf("-------------------------------------------------------------------------------\n");
printf("OBJECTIVE: TO FIND THE FUNCNL VALUE AT GIVEN PNT USING LAGRANGE INTERPOLATION\n");
printf("-------------------------------------------------------------------------------\n\n");
printf("\nEnter The Number Of Data Points (n):\t");
scanf("%d", &n);
printf("\nEnter Data Points & Values (x[i], y[i]):\t");
for(i=1; i<=n; i++)
scanf("%f%f", &x[i], &y[i]);
printf("\nAt Which Point You Want To Evaluate Function (xp):\t");
scanf("%f", &xp);
sum = 0;
for(i=1; i<=n; i++)
{
lbp = 1;
for(j=1; j<=n; j++)
{
if(i!=j)
lbp = lbp*(xp-x[j])/(x[i]-x[j]);
}
sum = sum + lbp*y[i];
}
lp = sum;
printf("\n--------------------");
printf("\nInterpolating Points\n");
printf("--------------------\n");
printf("\nx[i]\t\ty[i]\n");
for(i=1; i<=n; i++)
154

printf("\n%f\t%f", x[i], y[i]);

printf("\n\nEvaluation Point (xp):%f", xp);
printf("\nFunctional Value F(x):%f", lp);
Page

}
Prepared By
Er. Shree Krishna Khadka
 OUTPUT

155
Page

Prepared By
Er. Shree Krishna Khadka

View publication stats