Notes upon the subject of Numerical Analysis
𝕽udravisek
8 January 2026
Contents
Day 1 8 January 2026 Page 1
1.1 Introduction 1
Direct and Iterative Methods — 1
1.2 Types of Equations 1
1.3 General Iterative Scheme 2
1.4 Accuracy and Stopping Criteria 2
Error Tolerance — 3 • Stopping Conditions — 3
1.5 Initial Approximation 3
i
Day 1
1.1 Introduction
Numerical Analysis deals with the problem of obtaining approximate solutions of mathematical
problems which cannot be solved by analytical methods or for which exact solutions are already
known but are computationally inconvenient.
1.1.1 Direct and Iterative Methods
There are two broad classes of numerical methods:
• Direct methods
• Iterative methods
Direct methods yield a solution in a finite number of steps and do not involve successive
approximations.
Iterative methods start with an initial approximation and generate a sequence of improved
approximations, each obtained by modifying the previous one, until the required accuracy is
achieved.
Most methods studied in Numerical Analysis are iterative in nature. The successive steps involved
in such methods are called iterations.
1.2 Types of Equations
Definition 1.2.1: Polynomial Equations
An equation of the form
𝑓 (𝑥) = 0
is called a polynomial equation if 𝑓 (𝑥) is a polynomial.
Example:
𝑥 3 − 2𝑥 + 7 = 0.
1
Definition 1.2.2: Transcendental Equations
An equation is called a transcendental equation if it involves transcendental functions
such as trigonometric functions, inverse trigonometric functions, logarithmic functions,
exponential functions, et cetera.
Example:
sin 𝑥 − 𝑥 log 𝑥 = 0.
In this course, we shall consider methods for solving both polynomial and transcendental equa-
tions.
Note:-
The methods discussed are intended for finding simple roots only. If 𝛼 is a root of 𝑓 (𝑥) with
multiplicity 1, then 𝛼 is called a simple root. If 𝛼 has multiplicity 𝑚 ≥ 2, then 𝛼 is called a
multiple (or repeated) root.
1.3 General Iterative Scheme
To find a root of the equation
𝑓 (𝑥) = 0,
we require:
1. An initial approximation 𝑝0
2. An iteration formula of the form
𝑝 𝑘+1 = 𝑔(𝑝 𝑘 ), 𝑘 ≥ 0.
Starting from 𝑝0 , we generate a sequence
𝑝1 = 𝑔(𝑝 0 ), 𝑝 2 = 𝑔(𝑝1 ), ..., 𝑝 𝑛 = 𝑔(𝑝 𝑛−1 ).
The process is continued until the desired accuracy is obtained.
1.4 Accuracy and Stopping Criteria
Suppose the exact root is
𝑝 = 2.73498 . . .
and we want the answer correct up to three decimal places. Then the approximation should be
of the form
𝑝 𝑛 = 2.73423 . . .
so that the first three digits after the decimal point agree.
2
1.4.1 Error Tolerance
If
|𝑝 𝑛 − 𝑝| = 0.0008,
then rounding to three decimal places gives an error tolerance
𝜀 = 0.001.
In general, if the answer is required correct up to 𝑛 decimal places, then
𝜀 = 5 × 10−(𝑛+1) .
1.4.2 Stopping Conditions
We stop the iteration if either:
1. |𝑝 𝑛 − 𝑝 𝑛−1 | ≤ 𝜀, or
2. | 𝑓 (𝑝 𝑛 )| ≤ 𝜀.
1.5 Initial Approximation
To obtain an initial approximation, we use the Intermediate Value Theorem (IVT).
We first find an interval [𝑎, 𝑏] such that
𝑓 (𝑎) 𝑓 (𝑏) < 0.
This guarantees the existence of at least one root in (𝑎, 𝑏). A point in this interval, often the
midpoint, is chosen as the initial approximation.
Example 1.5.1
Solve
𝑓 (𝑥) = 𝑥 2 − 4𝑥 + 2 = 0.
Solution: Rewriting the equation in the form 𝑥 = 𝑔(𝑥), we obtain
𝑥2 + 2
𝑥= .
4
Thus, the iteration formula is
𝑥 2𝑘 + 2
𝑥 𝑘+1 = , 𝑘 ≥ 0.
4
Using IVT:
𝑓 (0) = 2, 𝑓 (1) = −1 ⇒ a root lies in (0, 1).
Taking 𝑥0 = 0.5, we compute:
𝑥1 = 0.5625, 𝑥 2 = 0.57910, 𝑥 3 = 0.583839.
The sequence {𝑥 𝑛 } converges to the exact root.
3
Note:-
Remark on Convergence. Not all iteration formulas converge. A sequence generated by
a particular formula may fail to converge to the exact root. Hence, convergence analysis is
essential for any iterative method.