To

Dr. Babasaheb Ambedkar Marathwada University

Chhatrapati Sambhajinagar (Maharashtra)
A Study of Root-Finding Methods for Nonlinear Equations

Research Project
Submitted in partial fulfillment of the requirement for the degree
Master in Mathematics
By
KHAN MUSHARRAF SHOUKAT
([Link]. Mathematics, III Semester)

Under the Guidance of

Dr. P. N. Kamble
Associate Professor,
Department of Mathematics,
Dr. Babasaheb Ambedkar Marathwada University,
Chhatrapati Sambhajinagar – 431004, Maharashtra, India.
 Certificate

This is to certify that the dissertation entitled “A Study of Root-Finding Methods

for Nonlinear Equations” submitted by KHAN MUSHARRAF SHOUKAT in
partial fulfillment of the requirement for the award of the degree of [Link]. Math-
ematics under Dr. Babasaheb Ambedkar Marathwada University, Chhatrapati
Sambhajinagar (Maharashtra), embodies the work carried out by him under my
supervision and guidance.
This work has not previously been submitted for the award of any degree,
diploma, fellowship, or other similar title.

Place: Dept. of Mathematics, BAMU University, Chhatrapati Sambhajinagar

Date:

Dr. P. N. Kamble
Research Guide
Associate Professor,
Department of Mathematics,
Dr. Babasaheb Ambedkar Marathwada University,
Chhatrapati Sambhajinagar.

Prof. D. B. Pachpatte
Head, Department of Mathematics
Dr. Babasaheb Ambedkar Marathwada University,
Chhatrapati Sambhajinagar.

1
 Declaration

I hereby declare that the thesis entitled “A Study of Root-Finding Methods for
Nonlinear Equations” is an original research work carried out by me for the
award of the degree of [Link]. Mathematics under the guidance of Dr. P. N. Kam-
ble, Assistant Professor, Department of Mathematics, Dr. Babasaheb Ambedkar
Marathwada University.
I further declare that this work has not previously been used for any degree,
diploma, fellowship, or any similar title. The present work is completely original
to the best of my knowledge.

Place:
Date:

KHAN MUSHARRAF SHOUKAT

Research Student

2
 Acknowledgement

It is a genuine pleasure to express my deep sense of thanks and gratitude to my

mentor, philosopher and guide, Professor Dr. P. N. Kamble, Department of Math-
ematics, Dr. Babasaheb Ambedkar Marathwada University, [Link] guid-
ance, valuable discussions, constant motivation, and willingness to spare time at
any moment despite his busy schedule have been greatly instrumental in complet-
ing this project.
I am also thankful to the Head of Department, Prof. Deepak B. Pachpatte,
and Professors Dr. S. K. Panchal, Dr. Kirtiwant P. Ghadle, and Dr. Amol
Khandagale, for their support and encouragement during my project work.
I extend my heartfelt thanks to my friends and colleagues for their continuous
support. Above all, I am grateful to my parents and family members, who form
the foundation of my strength. This project is dedicated to them.

KHAN MUSHARRAF SHOUKAT

3
 ABSTRACT
Root-finding is one of the most fundamental problems in numerical analysis.
Many scientific and engineering problems lead to nonlinear equations whose exact
solutions are difficult or impossible to obtain analytically. In such cases, numerical
methods provide approximate solutions with desired accuracy. This dissertation
presents a detailed study of classical and modern root-finding methods for solv-
ing nonlinear equations. The methods discussed include the Bisection method,
Regula-Falsi method, Secant method, and Newton–Raphson method. Their con-
vergence behavior, advantages, and limitations are analyzed through theoretical
discussion and numerical examples. A comparative study is also carried out to
highlight the efficiency and applicability of each method.

4
Contents
Sr. No. Contents Page No.

1 Introduction 01

2 Nonlinear Equations 02-04

3 Errors and Convergence 05-08

4 Bisection Method 08-09

5 Rate of Convergence 09-13

6 Limitations of the Bisection Method 13

7 Method of False Position or Regula–Falsi Method or 14-19

Interpo- lation Method

8 Advantages and Disadvantages of the Regula–Falsi Method 19-20

9 Limitations of the Regula–Falsi Method 21-29

10 A Modification of Linear Interpolation Method for Comput- 29-30

ing the Roots of an Equation using Newton–Raphson
Method

11 THE MODIFIED LINEAR INTERPOLATION ALGO- 30-31

RITHM (MODIFIED LIM)

12 ERROR ESTIMATION 31-32

14 NUMERICAL AND GRAPHICAL ILLUSTRATIONS 32-34

15 Comparison Between Usual Methods and Modified Method 34-36

16 CONCLUSION 37

17 References 38

2
1. INTRODUCTION
Nonlinear equations arise naturally in various fields such as physics, engineering,
economics, and applied sciences. An equation of the form f(x) = 0 is called a
nonlinear equation if f(x) is a nonlinear function of x. In many cases, it is not
possible to obtain exact solutions using algebraic methods. Therefore, numerical
techniques are employed to approximate the roots of such equations.
Root-finding methods aim to find values of x for which f(x) = 0. These methods
are iterative in nature and generate a sequence of approximations that converge to
the actual root. The efficiency of a method depends on its rate of convergence,
computational cost, and robustness.

1
2. Nonlinear Equations
An equation of the form
f (x) = 0
is called a nonlinear equation if the function f (x) is not linear in the variable
x. In other words, an equation is said to be nonlinear if the unknown variable
appears with degree greater than one or occurs in transcendental functions such as
exponential, logarithmic, or trigonometric functions. Some common examples of
nonlinear equations are:

x2 − 4 = 0, ex − 3x = 0, sin x + x − 1 = 0.

Nonlinear equations arise frequently in various branches of science, engineer-

ing, and applied mathematics. Problems related to heat transfer, fluid mechan-
ics, electrical circuits, population growth models, chemical reaction rates, and
optimization often lead to nonlinear mathematical equations. In most practical
situations, such equations cannot be solved exactly using elementary algebraic
methods. Therefore, numerical methods are widely used to obtain approximate
solutions with acceptable accuracy.
The solution of a nonlinear equation is called its root or zero. Geometrically,
a root corresponds to the point where the graph of the function y = f (x) intersects
the x-axis. Depending on the nature of the function, a nonlinear equation may
possess one root, several roots, or no real root at all. Hence, identifying an interval
containing the root is an important step in numerical computation.
One of the most important theoretical tools for establishing the existence of a
root is the Intermediate Value Theorem. According to this theorem, if a func-
tion f (x) is continuous on a closed interval [a, b] and satisfies f (a) f (b) < 0, then
there exists at least one real root in the open interval (a, b). Many numerical meth-
ods such as the Bisection method and the Regula-Falsi method are based on this
principle.
Since exact solutions of nonlinear equations are often not available, numerical
techniques generate a sequence of approximations that converge to the actual root.
The efficiency of a root-finding method depends on factors such as the rate of
convergence, computational cost, and reliability. Therefore, a clear understanding
of nonlinear equations forms the foundation for studying and applying numerical
methods for solving root-finding problems.

2
Intermediate Value Theorem
The Intermediate Value Theorem (IVT) is one of the most fundamental results
in real analysis and forms the theoretical basis of several numerical methods used
for solving nonlinear equations. The theorem provides a sufficient condition for
the existence of at least one real root of an equation within a given interval and is
extensively used in root-finding algorithms.
Statement of the Theorem: Let f : [a, b] → R be a real-valued function that
is continuous on the closed interval [a, b]. If the function values at the endpoints
satisfy
f (a) f (b) < 0,
then there exists at least one real number c ∈ (a, b) such that
f (c) = 0.
The Intermediate Value Theorem guarantees the existence of a root but does
not provide information regarding its uniqueness or exact location. Moreover, the
theorem does not specify the number of roots that may lie within the interval.
This limitation makes numerical methods essential for approximating the root to
a desired level of accuracy.
From a geometrical point of view, the theorem states that if a continuous func-
tion takes values of opposite signs at the endpoints of an interval, then the graph
of the function must cross the x-axis at least once within that interval. The point
of intersection corresponds to a real root of the equation f (x) = 0. This geometric
interpretation provides a clear and intuitive understanding of the theorem.
In numerical analysis, the Intermediate Value Theorem plays a crucial role in
determining an initial interval that contains a root. Methods such as the Bisection
method and the Regula-Falsi method are based entirely on this theorem. These
methods repeatedly subdivide the interval [a, b] into smaller subintervals while
preserving the sign change condition f (a) f (b) < 0. As a result, the interval con-
taining the root becomes progressively smaller, leading to convergence toward the
exact solution.
One of the major advantages of methods based on the Intermediate Value The-
orem is their reliability. Since the existence of a root is guaranteed at each step,
these methods are robust and unlikely to fail. However, this reliability comes at
the cost of speed. Bracketing methods generally exhibit linear convergence, which
is slower when compared to open methods such as the Newton–Raphson or Secant
methods.

3
 Despite its limitations, the Intermediate Value Theorem remains an indispens-
able theoretical tool in numerical analysis. It ensures the existence of solutions,
provides a foundation for safe and reliable algorithms, and serves as a starting
point for more advanced and faster root-finding techniques. Therefore, a thorough
understanding of the Intermediate Value Theorem is essential for the effective ap-
plication of numerical methods to nonlinear equations.

4
3. Errors and Convergence
In numerical analysis, the concepts of error and convergence play a central role
in assessing the accuracy, reliability, and efficiency of numerical methods. Since
numerical techniques provide approximate solutions rather than exact ones, it is
essential to measure how close an approximation is to the true solution and how
rapidly a sequence of approximations approaches that solution. This chapter dis-
cusses different types of errors and the notion of convergence in the context of
root-finding methods for nonlinear equations.

Types of Errors
Let α be the exact root of a nonlinear equation f (x) = 0, and let xn be the approx-
imation obtained after the n-th iteration of a numerical method. The difference
between the exact value and its approximation is called the error. Errors may
arise due to several reasons such as truncation of infinite processes, rounding of
numbers, and limitations of digital computation.

Absolute Error

The absolute error is defined as

En = |xn − α|.

It measures the magnitude of the difference between the approximate solution and
the exact root. Absolute error provides a direct measure of accuracy; however, it
may not always be meaningful when the exact value α is very large or very small.

Relative Error

The relative error is given by

|xn − α|
Rn = , α ̸= 0.
|α|
Relative error expresses the error as a fraction of the exact value and is often
more informative than absolute error, especially when comparing the accuracy of
approximations of different magnitudes.

5
Percentage Error

The percentage error is defined as

Pn = Rn × 100%.

This form of error is commonly used in practical applications as it provides a

clearer understanding of accuracy in percentage terms.

Sources of Error
Errors in numerical computation mainly arise from two sources: truncation errors
and round-off errors.

Truncation Error

Truncation error occurs when an infinite process is approximated by a finite num-

ber of steps. For example, when a Taylor series is truncated after a finite number
of terms, the neglected higher-order terms introduce an error. In iterative meth-
ods, truncation error arises when the iterative process is terminated after a finite
number of iterations.

Round-Off Error

Round-off error occurs due to the finite precision of computer arithmetic. Since
computers can store only a limited number of digits, real numbers are approx-
imated, leading to small errors that may accumulate during successive compu-
tations. In some cases, round-off errors can significantly affect the final result,
especially in long iterative processes.

Convergence of Numerical Methods

A numerical method is said to be convergent if the sequence of approximations
{xn } generated by the method approaches the exact root α as the number of itera-
tions increases. Mathematically, convergence is defined as

lim xn = α.
n→∞

Convergence is a desirable property, as it ensures that repeated application of

the algorithm will eventually lead to an accurate solution.

6
Rate of Convergence
The rate of convergence describes how fast the sequence {xn } converges to the
root α. Let en = |xn − α| denote the error in the n-th iteration. A method is said to
converge with order p if there exists a constant C > 0 such that
en+1
lim = C.
n→∞ (en ) p

Based on the value of p, convergence can be classified as follows:

Linear Convergence

If p = 1, the method is said to have linear convergence. In this case, the error
decreases at a constant rate. The Bisection and Regula-Falsi methods are examples
of methods with linear convergence.

Superlinear Convergence

If 1 < p < 2, the method exhibits superlinear convergence. The Secant method
is a well-known example of a method with superlinear convergence.

Quadratic Convergence

If p = 2, the method is said to have quadratic convergence. This means that the
number of correct digits approximately doubles with each iteration. The Newton–
Raphson method is a classical example of a quadratically convergent method.

Stopping Criteria
In practical computations, an iterative method must be terminated after a finite
number of steps. Common stopping criteria include:
• The absolute error |xn+1 − xn | is less than a prescribed tolerance.
• The function value | f (xn )| is sufficiently close to zero.
• A maximum number of iterations has been reached.

Importance of Error and Convergence Analysis

Error and convergence analysis helps in selecting an appropriate numerical method
for a given problem. While methods with higher-order convergence are faster,

7
they may require more computational effort per iteration or good initial approxi-
mations. On the other hand, slower methods are often more reliable. Therefore, a
balance between accuracy, efficiency, and stability must be maintained.
In conclusion, understanding errors and convergence is essential for evaluat-
ing the performance of numerical root-finding methods. These concepts provide
insight into the accuracy of approximations, the speed of convergence, and the
overall reliability of numerical algorithms used to solve nonlinear equations.

4. Bisection Method
The Bisection Method is based on the repeated application of the Intermediate
Value Theorem. Let the function f (x) be continuous in the interval [a, b]. For def-
initeness, let f (a) be negative and f (b) be positive. Then the first approximation
to the root is given by
1
x1 = (a + b).
2

If f (x1 ) = 0, then x1 is a root of the equation f (x) = 0. Otherwise, the root lies
between a and x1 or between x1 and b, depending on whether f (x1 ) is positive or

8
negative. The interval is then bisected again, and the process is continued until the
root is obtained with the desired accuracy.
If f (x1 ) is positive, then the root lies between a and x1 , and the second approx-
imation to the root is
1
x2 = (a + x1 ).
2
If f (x2 ) is negative, then the root lies between x1 and x2 . The third approximation
to the root is then
1
x3 = (x1 + x2 ),
2
and so on.

Note
Since the new interval containing the root is exactly half the length of the previous
one, the interval width is reduced by a factor of 12 at each step. At the end of the
n-th step, the length of the interval is
b−a
.
2n
If this process is repeated n times, and the final interval is required to be as
small as a given tolerance ε, then
b−a
≤ ε.
2n
Taking logarithms, we obtain
log(b − a) − log ε
n≥ .
log 2
This gives the number of iterations required for achieving an accuracy ε.
In particular, the minimum number of iterations required for converging to a
root in the interval (0, 1) for different values of ε are given below:

ε 10−2 10−3 10−4

n 7 10 14

5. Rate of Convergence
1
As the error decreases with each step by a factor of 2 (i.e., en+1 /en = 12 ), the
convergence in the bisection method is linear.

9
Example 1
(a) Find a root of the equation

x3 − 4x − 9 = 0,

using the bisection method correct to three decimal places.

(b) Using the bisection method, find the negative root of the equation

x3 − 4x + 9 = 0.

Solution
(a) Let
f (x) = x3 − 4x − 9
Since f (2) is negative and f (3) is positive, a root lies between 2 and 3.
1
First approximation to the root: x1 = (2 + 3) = 2.5
2

f (x1 ) = (2.5)3 − 4(2.5) − 9 = −3.375 i.e., negative

Hence, the root lies between x1 and 3. The second approximation to the root is:
1
x2 = (x1 + 3) = 2.75
2

f (x2 ) = (2.75)3 − 4(2.75) − 9 = 0.7969 i.e., positive

Hence, the root lies between x1 and x2 . The third approximation is:
1
x3 = (x1 + x2 ) = 2.625
2

f (x3 ) = (2.625)3 − 4(2.625) − 9 = −1.4121 i.e., negative

The root lies between x2 and x3 . The fourth approximation is:
1
x4 = (x2 + x3 ) = 2.6875
2

10
 Repeating this process, the successive approximations are:
x5 = 2.71875, x6 = 2.70313, x7 = 2.71094,
x8 = 2.70703, x9 = 2.70508, x10 = 2.70605,
x11 = 2.70654, x12 = 2.70642
Hence, the root is
x ≈ 2.7064
(b) If α, β , γ are the roots of the given equation, then −α, −β , −γ are the roots
of
(−x)3 − 4(−x) + 9 = 0
Hence, the negative root of the given equation is the positive root of

x3 − 4x − 9 = 0

which we have found above to be 2.7064.

Advantages and Disadvantages of the Bisection Method

The Bisection Method is one of the oldest and most widely used numerical tech-
niques for solving nonlinear equations. Its popularity is mainly due to its sim-
plicity and guaranteed convergence. However, like all numerical methods, it pos-
sesses both advantages and limitations. A detailed discussion of these aspects is
presented below.

Advantages of the Bisection Method

One of the most important advantages of the Bisection Method is its guaranteed
convergence. As long as the function f (x) is continuous on the interval [a, b] and
satisfies the condition f (a) f (b) < 0, the method is certain to converge to at least
one real root. This reliability makes the method highly robust and suitable for
practical applications where stability is more important than speed.
Another major advantage is the simplicity of implementation. The algorithm
involves only basic arithmetic operations such as addition, division by two, and
function evaluation. No derivatives are required, which makes the method easy to
apply even when the derivative of the function is difficult or impossible to com-
pute.

11
 The Bisection Method also provides a clear error bound. At each iteration,
the length of the interval containing the root is reduced by half. After n iterations,
the error is bounded by b−a2n . This feature allows the user to estimate the accuracy
of the approximation and determine the number of iterations required to achieve a
prescribed tolerance.
Another advantage is its numerical stability. Since the method always works
within a fixed interval and does not rely on extrapolation, it is less affected by
round-off errors compared to some open methods. This makes it particularly suit-
able for computer implementation.
Finally, the Bisection Method is insensitive to the shape of the function. Un-
like methods such as Newton–Raphson, it does not require the function to be dif-
ferentiable or well-behaved near the root. As a result, it can be applied to a wide
variety of nonlinear equations.

Disadvantages of the Bisection Method

Despite its reliability, the Bisection Method has several limitations. The most
significant disadvantage is its slow rate of convergence. The method converges
linearly, which means that the number of correct digits increases slowly with each
iteration. As a result, a large number of iterations may be required to achieve high
accuracy.
Another drawback is the requirement of an initial bracketing interval. The
method can only be applied if an interval [a, b] is known such that f (a) f (b) < 0.
Finding such an interval may not always be straightforward, especially for com-
plex functions or when little information about the location of the root is available.
The Bisection Method is also inefficient for problems requiring high preci-
sion. Compared to faster methods such as the Secant or Newton–Raphson meth-
ods, the Bisection Method takes considerably more time to reach the same level of
accuracy. This makes it less suitable for problems where computational efficiency
is critical.
Another limitation is that the method does not exploit the functional behav-
ior. It does not use information about the slope or curvature of the function, which
could otherwise accelerate convergence. As a result, the method may waste com-
putational effort by repeatedly bisecting the interval even when the function be-
haves smoothly.
In conclusion, the Bisection Method is a simple, reliable, and stable numer-
ical technique for solving nonlinear equations. However, its slow convergence

12
and dependence on an initial bracketing interval limit its efficiency. Therefore, in
practice, the Bisection Method is often used as a preliminary step to locate a root
accurately, after which faster methods are employed to refine the solution.

6. Limitations of the Bisection Method

Although the Bisection Method is simple and reliable, it suffers from several lim-
itations that restrict its efficiency and applicability in practical problems.
The most significant limitation of the Bisection Method is its slow rate of con-
vergence. The method converges linearly, which means that the error decreases
by a constant factor at each iteration. As a result, a large number of iterations may
be required to achieve high accuracy, especially when compared to faster methods
such as the Newton–Raphson or Secant methods.
Another major limitation is the requirement of an initial bracketing interval.
The method can only be applied if two initial points a and b can be found such
that f (a) f (b) < 0. Identifying such an interval is not always straightforward, par-
ticularly when the behavior of the function is complex or when little information
about the location of the root is available.
The Bisection Method is also inefficient for problems requiring high pre-
cision. Since the interval width is reduced by only one half at each iteration,
obtaining solutions with very small tolerances demands a considerable number of
iterations, leading to increased computational time.
Another drawback is that the method does not utilize information about the
function’s behavior. It does not take advantage of derivative information, slope,
or curvature of the function, which could otherwise accelerate convergence. Con-
sequently, the method may perform unnecessary computations even when the
function behaves smoothly.
Furthermore, the Bisection Method is restricted to continuous functions. If
the function is discontinuous in the interval [a, b], the method may fail, even if a
sign change occurs. This restricts its applicability to a specific class of functions.
In summary, while the Bisection Method is robust and easy to implement, its
slow convergence, dependence on an initial bracketing interval, and inability to
exploit additional information about the function limit its effectiveness. Therefore,
it is often used as a preliminary method to locate a root before applying more
efficient numerical techniques.

13
7. Method of False Position or Regula–Falsi Method or Interpo-
lation Method
The Method of False Position, also known as the Regula–Falsi Method or the
Interpolation Method, is one of the oldest numerical techniques for finding the
real root of an equation
f (x) = 0.
This method closely resembles the Bisection Method and is based on the idea of
linear interpolation.
Let us choose two points x0 and x1 such that f (x0 ) and f (x1 ) are of opposite
signs, that is,
f (x0 ) f (x1 ) < 0.
This implies that the graph of y = f (x) crosses the x-axis between the points x0
and x1 , and hence there exists at least one real root in the interval (x0 , x1 ).
The equation of the chord joining the points
A(x0 , f (x0 )) and B(x1 , f (x1 ))
is given by
f (x1 ) − f (x0 )
y − f (x0 ) = (x − x0 ).
x1 − x 0
The basic idea of the method is to replace the curve AB by the straight line
(chord) AB and take the point where this chord intersects the x-axis as an approx-
imation to the root. Setting y = 0, the abscissa of the point where the chord cuts
the x-axis is obtained as
x1 − x 0
x2 − x 0 = f (x0 ),
f (x1 ) − f (x0 )
or equivalently,
f (x0 )(x1 − x0 )
x2 = x 0 − .
f (x1 ) − f (x0 )
The value x2 serves as an approximation to the root of the equation f (x) = 0.
If f (x0 ) and f (x2 ) are of opposite signs, then the root lies between x0 and x2 ,
and x1 is replaced by x2 . Otherwise, if f (x2 ) and f (x1 ) are of opposite signs, then
the root lies between x2 and x1 , and x0 is replaced by x2 . This procedure is repeated
until the root is obtained with the desired accuracy.
The iterative process based on this procedure is known as the Method of False
Position or Regula–Falsi Method.

14
Rate of Convergence
The Regula–Falsi Method has a linear rate of convergence. However, it generally
converges faster than the Bisection Method because it uses linear interpolation
instead of simple interval halving.

Example 2
Find a real root of the equation

x3 − 2x − 5 = 0

by the method of false position correct to three decimal places.

Solution:
Let
f (x) = x3 − 2x − 5.
Then
f (2) = −1 and f (3) = 16.
Hence, a root lies between 2 and 3.

15
 Taking x0 = 2, x1 = 3, f (x0 ) = −1, and f (x1 ) = 16, by the method of false
position, we get
x1 − x 0 1
x2 = x 0 − f (x0 ) = 2 + = 2.0588. (i)
f (x1 ) − f (x0 ) 17
Now,
f (2.0588) = −0.3908.
Hence, the root lies between 2.0588 and 3.
Taking x0 = 2.0588, x1 = 3, f (x0 ) = −0.3908, f (x1 ) = 16, in (i), we get
0.9412
x3 = 2.0588 − (−0.3908) = 2.0813.
19.3908
Repeating the process, the successive approximations are:
x4 = 2.0862, x5 = 2.0915, x6 = 2.0934,
x7 = 2.0941, x8 = 2.0943, etc.
Hence, the root is
x = 2.094
correct to three decimal places.
—

Example 3
Find the root of the equation
cos x = xex
using the regula–falsi method correct to four decimal places.
Solution:
Let
f (x) = cos x − xex .
Then
f (0) = 1, f (1) = cos 1 − e = −2.17798.
Hence, the root lies between 0 and 1.
Taking x0 = 0, x1 = 1, f (x0 ) = 1, and f (x1 ) = −2.17798, by the regula–falsi
method, we get
x1 − x0 1
x2 = x 0 − f (x0 ) = 0 + = 0.31467. (i)
f (x1 ) − f (x0 ) 3.17798
16
 Now,
f (0.31467) = 0.51987,
i.e., the root lies between 0.31467 and 1.
Taking x0 = 0.31467, x1 = 1, f (x0 ) = 0.51987, f (x1 ) = −2.17798, in (i), we
get
0.68533
x3 = 0.31467 + × 0.51987 = 0.44673.
2.69785
Now,
f (0.44673) = 0.20356,
hence the root lies between 0.44673 and 1.
Taking x0 = 0.44673, x1 = 1, f (x0 ) = 0.20356, f (x1 ) = −2.17798, in (i), we
get
0.55327
x4 = 0.44673 + × 0.20356 = 0.49402.
2.38154
Repeating the process, the successive approximations are:

x5 = 0.50995, x6 = 0.51520, x7 = 0.51692,

x8 = 0.51748, x9 = 0.51767, x10 = 0.51775, etc.

Hence, the root is
x = 0.5177
correct to four decimal places.

Secant Method
The Secant Method is an improvement over the method of false position as it does
not require the condition
f (x0 ) f (x1 ) < 0
of that method.
Here also, the graph of the function y = f (x) is approximated by a secant line.
At each iteration, the two most recent approximations to the root are used to find
the next approximation. It is not necessary that the interval must contain the root.
Taking x0 and x1 as the initial approximations, the equation of the chord joining
these points is
f (x1 ) − f (x0 )
y − f (x1 ) = (x − x1 ).
x1 − x 0

17
 The abscissa of the point where it crosses the x-axis (y = 0) is given by
x1 − x 0
x2 = x 1 − f (x1 ),
f (x1 ) − f (x0 )
which is an approximation to the root.
The general formula for successive approximations is
xn − xn−1
xn+1 = xn − f (xn ), n ≥ 1.
f (xn ) − f (xn−1 )

Rate of Convergence

If at any iteration f (xn ) = f (xn−1 ), the method fails and does not converge neces-
sarily. This is a drawback of the secant method over the method of false position,
which always converges.
However, when the secant method does converge, its rate of convergence is
approximately 1.6, which is faster than that of the method of false position.
—

Example 4
Find a root of the equation
x3 − 2x − 5 = 0
using the secant method correct to three decimal places.
Solution:
Let
f (x) = x3 − 2x − 5,
so that
f (2) = −1 and f (3) = 16.
Taking initial approximations x0 = 2 and x1 = 3, by the secant method, we have
x1 − x 0 3−2
x2 = x 1 − f (x1 ) = 3 − × 16 = 2.058823.
f (x1 ) − f (x0 ) 16 + 1
Now,
f (x2 ) = −0.390799.

x2 − x 1
x3 = x 2 − f (x2 ) = 2.081263,
f (x2 ) − f (x1 )
18
and
f (x3 ) = −0.147204.

x3 − x 2
x4 = x 3 − f (x3 ) = 2.094824,
f (x3 ) − f (x2 )
and
f (x4 ) = 0.003042.

x4 − x 3
x5 = x 4 − f (x4 ) = 2.094549.
f (x4 ) − f (x3 )
Hence, the root is
x = 2.094
correct to three decimal places.

8. Advantages and Disadvantages of the Regula–Falsi Method

The Regula–Falsi method, also known as the Method of False Position, is a numer-
ical technique for finding roots of nonlinear equations. It combines the reliability
of bracketing methods with the idea of linear interpolation. Like all numerical
methods, it has certain advantages as well as limitations, which are discussed be-
low.

Advantages of the Regula–Falsi Method

One of the major advantages of the Regula–Falsi method is its guaranteed con-
vergence. As long as the function f (x) is continuous on the interval [a, b] and
satisfies the condition f (a) f (b) < 0, the method ensures convergence to at least
one real root. This property makes the method reliable and robust for a wide class
of problems.
Another important advantage is that the Regula–Falsi method generally con-
verges faster than the Bisection Method. Instead of simply halving the interval,
it uses linear interpolation to estimate the root. This often leads to better approx-
imations in fewer iterations, especially when the function behaves nearly linearly
near the root.
The method does not require the computation of derivatives, which is a signif-
icant advantage over methods such as the Newton–Raphson method. As a result,

19
it can be applied to problems where the derivative of the function is difficult or
expensive to compute.
The Regula–Falsi method also provides a simple and intuitive geometric in-
terpretation. By replacing the curve with a straight line joining two points on
the graph, the method offers a clear visualization of how the approximations are
generated. This makes it easier to understand and implement.
Furthermore, the method maintains a bracketing interval throughout the iter-
ation process. This ensures numerical stability and prevents divergence, making
the method suitable for computer implementation.

Disadvantages of the Regula–Falsi Method

Despite its advantages, the Regula–Falsi method has several limitations. The most
significant drawback is its slow convergence in certain cases. Although it is
generally faster than the Bisection Method, the convergence is still linear. In some
situations, one endpoint of the interval remains fixed for many iterations, leading
to very slow progress toward the root.
Another limitation is the dependence on an initial bracketing interval. The
method requires two initial points where the function values have opposite signs.
Finding such an interval may not be easy, especially when the function is compli-
cated or when little information about the location of the root is available.
The Regula–Falsi method is also inefficient for problems requiring very high
precision. Due to its linear convergence, a large number of iterations may be
needed to achieve small tolerances, increasing computational time.
In addition, the method does not exploit higher-order information such as deriva-
tives or curvature of the function. This prevents the method from achieving faster
convergence rates seen in open methods like the Newton–Raphson or Secant meth-
ods.
In conclusion, the Regula–Falsi method is a reliable and moderately efficient
root-finding technique that improves upon the Bisection Method by using linear
interpolation. However, its slow convergence in certain cases and dependence on
an initial bracketing interval limit its effectiveness. Therefore, the method is often
used as an intermediate approach or combined with other techniques to improve
performance.

20
9. Limitations of the Regula–Falsi Method
Although the Regula–Falsi method is more efficient than the Bisection Method
and guarantees convergence under suitable conditions, it suffers from several im-
portant limitations.
The most significant limitation of the Regula–Falsi method is its slow con-
vergence in certain situations. In many practical problems, one endpoint of the
bracketing interval remains fixed for a large number of iterations, while the other
endpoint changes slowly. This results in poor progress toward the root and causes
the method to converge very slowly.
Another major drawback is its linear rate of convergence. Although the
method often converges faster than the Bisection Method, it is still much slower
than higher-order methods such as the Secant or Newton–Raphson methods. Con-
sequently, a large number of iterations may be required to obtain highly accurate
results.
The Regula–Falsi method also requires an initial bracketing interval [a, b]
such that f (a) f (b) < 0. Identifying such an interval is not always straightforward,
especially when the function is complicated or when prior information about the
location of the root is limited.
Another limitation is that the method does not utilize derivative information.
As a result, it fails to take advantage of the function’s slope or curvature, which
could otherwise significantly improve the rate of convergence.
Furthermore, the method is restricted to continuous functions. If the func-
tion is discontinuous within the chosen interval, the sign-change condition may be
misleading, and the method may fail to locate the root.
In conclusion, while the Regula–Falsi method is reliable and easy to imple-
ment, its slow convergence in certain cases and dependence on a suitable initial
interval limit its efficiency. Therefore, in practice, it is often modified or combined
with other numerical methods to improve its performance.

21
Newton–Raphson Method
Let x0 be an approximate root of the equation f (x) = 0. If x = x0 + h is the exact
root, then
f (x0 + h) = 0.
Expanding f (x0 + h) by Taylor’s series, we get

′ h2 ′′
f (x0 + h) = f (x0 ) + h f (x0 ) + f (x0 ) + · · · = 0.
2!
Since h is small, neglecting h2 and higher powers of h, we obtain

f (x0 ) + h f ′ (x0 ) = 0,

or
f (x0 )
h=− . (1)
f ′ (x0 )
Therefore, a closer approximation to the root is given by
f (x0 )
x1 = x 0 − .
f ′ (x0 )
Similarly, starting with x1 , a still better approximation x2 is given by
f (x1 )
x2 = x 1 − .
f ′ (x1 )
In general,
f (xn )
xn+1 = xn − , (n = 0, 1, 2, . . .), (2)
f ′ (xn )
which is known as the Newton–Raphson formula or Newton’s iteration for-
mula.

Note

Obs. 1. Newton’s method is useful in cases of large values of f ′ (x), i.e., when the
graph of f (x) while crossing the x-axis is nearly vertical.
If f ′ (x) is small in the vicinity of the root, then by equation (1), h will be large
and the computation of the root is slow or may not be possible. Thus, this method
is not suitable in those cases where the graph of f (x) is nearly horizontal while
crossing the x-axis.

22
 Obs. 2. Geometrical Interpretation. Let x0 be a point near the root α of the
equation f (x) = 0. Then the equation of the tangent at A(x0 , f (x0 )) is

y − f (x0 ) = f ′ (x0 )(x − x0 ).

It cuts the x-axis at

f (x0 )
x1 = x 0 − ,
f ′ (x0 )
which is a first approximation to the root α.

If A1 is the point corresponding to x1 on the curve, the tangent at A1 will cut

the x-axis at x2 , which is nearer to α and is therefore a second approximation to
the root.
Repeating this process, we approach the root quite rapidly. Hence, the method
consists in replacing the part of the curve between the point A and the x-axis by
means of the tangent to the curve at A.
Obs. 3. Newton’s method is generally used to improve the results obtained by
other methods. It is applicable to the solution of both algebraic and transcendental
equations.

23
Convergence of Newton–Raphson Method
Newton’s formula converges provided the initial approximation x0 is chosen suf-
ficiently close to the root. If it is not near the root, the procedure may lead to an
endless cycle. A bad initial choice may lead one astray. Thus, a proper choice of
the initial guess is very important for the success of Newton’s method.
Comparing equation (2) with the relation xn+1 = φ (xn ) of the iteration method,
we get
f (xn )
φ (xn ) = xn+1 = xn − ′ .
f (xn )
In general,
f (x)
φ (x) = x − ′ ,
f (x)
which gives
f (x) f ′′ (x)
φ ′ (x) = .
[ f ′ (x)]2
Since the iteration method converges if

|φ ′ (x)| < 1,

Newton’s formula will converge if

| f (x) f ′′ (x)| < [ f ′ (x)]2

in the interval considered.

Assuming f (x), f ′ (x), and f ′′ (x) to be continuous, we can select a small inter-
val in the vicinity of the root α, in which the above condition is satisfied. Hence,
the result follows.
Newton’s method converges conditionally, while the regula–falsi method al-
ways converges. However, when the Newton–Raphson method converges, it does
so faster and is therefore preferred.

Quadratic Convergence of Newton–Raphson Method

Newton’s method has a quadratic convergence.

Suppose xn differs from the root α by a small quantity εn , so that

xn = α + εn and xn+1 = α + εn+1 .

24
 Then equation (2) becomes
f (α + εn )
α + εn+1 = α + εn − ,
f ′ (α + εn )
i.e.,
f (α + εn )
εn+1 = εn − .
f ′ (α + εn )
Using Taylor’s expansion, we have

′ εn2 ′′
f (α + εn ) = εn f (α) + f (α) + · · · ,
2!
and
f ′ (α + εn ) = f ′ (α) + εn f ′′ (α) + · · · .
Substituting these, we obtain
ε2
εn f ′ (α) + 2n f ′′ (α) + · · ·
εn+1 = εn − ′ .
f (α) + εn f ′′ (α) + · · ·
Neglecting higher-order terms and using f (α) = 0, we get

εn2 f ′′ (α)
εn+1 = .
2 f ′ (α)
This shows that the subsequent error at each step is proportional to the square
of the previous error, and as such the convergence is quadratic. Thus, the Newton–
Raphson method has second-order convergence.

Example 5
Find the positive root of
x4 − x − 10 = 0
using the Newton–Raphson method, correct to three decimal places.

Solution
Let
f (x) = x4 − x − 10

25
 Then,
f (1) = −10 < 0, f (2) = 16 − 2 − 10 = 4 > 0
Hence, a root of f (x) = 0 lies between 1 and 2. Let us take the initial approxi-
mation x0 = 2.

f ′ (x) = 4x3 − 1
Newton–Raphson formula is
f (xn )
xn+1 = xn −
f ′ (xn )
First approximation:
f (x0 ) 4 4
x1 = x0 − = 2 − = 2 − = 1.871
f ′ (x0 ) 4(2)3 − 1 31
Second approximation:
f (x1 ) (1.871)4 − (1.871) − 10
x2 = x1 − ′ = 1.871 −
f (x1 ) 4(1.871)3 − 1
0.3855
x2 = 1.871 − = 1.856
25.199
Third approximation:
f (x2 ) (1.856)4 − (1.856) − 10
x3 = x2 − = 1.856 −
f ′ (x2 ) 4(1.856)3 − 1
0.0010
x3 = 1.856 − = 1.856
24.574
Since x2 = x3 , the required root is

x = 1.856

correct to three decimal places.

26
Example 6
Find by Newton’s method, the real root of the equation
3x = cos x + 1
correct to four decimal places.

Solution
Let
f (x) = 3x − cos x − 1
Then,
f (0) = −2 < 0, f (1) = 3 − 0.5403 − 1 = 1.4597 > 0
Hence, a root of f (x) = 0 lies between 0 and 1. It is nearer to 1; let us take
x0 = 0.6.

f ′ (x) = 3 + sin x
Newton’s iteration formula is
f (xn ) xn sin xn + cos xn + 1
xn+1 = xn − =
f ′ (xn ) 3 + sin xn
First approximation:
(0.6) sin(0.6) + cos(0.6) + 1
x1 =
3 + sin(0.6)
0.6(0.5729) + 0.8253 + 1
x1 = = 0.6071
3 + 0.5729
Second approximation:
(0.6071) sin(0.6071) + cos(0.6071) + 1
x2 =
3 + sin(0.6071)
0.6071(0.57049) + 0.8213 + 1
x2 = = 0.6071
3 + 0.57049
Since x1 = x2 , the required root is
x = 0.6071
correct to four decimal places.

27
Advantages, Disadvantages and Limitations of the Newton–Raphson Method
The Newton–Raphson method is one of the most powerful and widely used nu-
merical techniques for solving nonlinear equations. It is based on the idea of suc-
cessive linear approximation using tangents and is known for its rapid convergence
when suitable conditions are satisfied.

Advantages

One of the most important advantages of the Newton–Raphson method is its quadratic
convergence. Once the iteration reaches sufficiently close to the root, the number
of correct digits approximately doubles at each step, making the method extremely
efficient compared to linear convergence methods such as the bisection or regula–
falsi methods.
Another major advantage is its high computational efficiency. Because fewer
iterations are required to achieve a desired level of accuracy, the method is well
suited for problems where high precision is needed.
The Newton–Raphson method is also widely applicable to both algebraic and
transcendental equations, provided the function is differentiable in the neighbor-
hood of the root. It is frequently used to refine approximate solutions obtained
from other numerical methods.
Furthermore, the method has a clear geometrical interpretation. Each iter-
ation corresponds to the intersection of the tangent to the curve y = f (x) at the
current approximation with the x-axis, which helps in understanding the conver-
gence behavior.

Disadvantages

Despite its many advantages, the Newton–Raphson method has several drawbacks.
The most significant disadvantage is that it requires the computation of the
derivative f ′ (x). For complicated functions, evaluating the derivative may be
difficult, time-consuming, or computationally expensive.
Another disadvantage is that the method does not guarantee convergence. If
the initial approximation is not chosen sufficiently close to the actual root, the
iterations may diverge or converge to a different root.
The method may also fail when the derivative becomes zero or very small
near the root. In such cases, the correction term becomes large, causing the itera-
tion to jump far away from the desired root.

28
 In addition, the Newton–Raphson method may exhibit oscillatory behavior or
enter into an endless cycle for poorly chosen initial approximations.

Limitations

One of the major limitations of the Newton–Raphson method is its sensitivity

to the initial guess. A poor initial approximation can lead to divergence, slow
convergence, or convergence to an unintended root.
The method is also not suitable for functions with multiple roots unless spe-
cial modifications are applied. In such cases, the convergence rate may reduce
from quadratic to linear.
Another limitation is that the method cannot be applied directly to non-
differentiable functions, as the derivative is essential for its formulation.
Moreover, the Newton–Raphson method is locally convergent, meaning that
convergence is assured only when the initial guess lies within a sufficiently small
neighborhood of the root.
In conclusion, while the Newton–Raphson method is extremely fast and effi-
cient under ideal conditions, its dependence on a good initial approximation and
derivative information limits its reliability. Therefore, in practical applications, it
is often combined with bracketing methods to ensure both reliability and speed.

10. A Modification of Linear Interpolation Method for Comput-

ing the Roots of an Equation using Newton–Raphson Method
Let xi−1 , xi be two initial approximations to the root suggested by the Intermediate
Value Theorem. By joining the points P(xi−1 , fi−1 ) and Q(xi , fi ), a new approxi-
mation xi+1 to the root is obtained which is the point of intersection of the chord
PQ passing through the points on the curve and the x-axis. The new value xi+1 is
now computed from the rule
(xi − xi−1 )
xi+1 = xi − fi , (1.2)
( fi − fi−1 )
which is obtained by setting y = 0 in the equation of the chord PQ.
In the next iteration, xi+1 replaces one among xi−1 or xi , the sign of the function
value of which is same as the sign of fi+1 . The process continues until a suitable
halting point is obtained. For example, one can use any one of the following
criteria to terminate the process.

29
 • f (xk ) is bounded by some positive real number µ, known as tolerance error,
i.e.
| f (xk )| ≤ µ.
• The difference between two successive iterations is less than a prescribed
value ε, known as accuracy, i.e.
|xk+1 − xk | ≤ ε.

Although the convergence is guaranteed, the method sometimes retains one

of the end points resulting in a slow rate of convergence. This kind of problem
arises when the given function becomes convex or concave during the process in
an interval that contains a root. To overcome this, a number of methods have been
developed in recent times. Our method mainly focuses on functions which become
convex or concave in some interval that contains a root of the equation.
We are now all set to describe our method and our main motivation of the work
is the Illinois method. We analysed the behaviour of the method with the help of
asymptotic convergence and compared the efficiency and rate of convergence of
our method with some well-known iteration methods.

11. THE MODIFIED LINEAR INTERPOLATION ALGORITHM

(MODIFIED LIM)
We developed our new method by modifying the linear interpolation method. So
our method behaves similar to that of the unmodified method except the case when
it retains one of the end points. We use the following rule to overcome this kind
of problem.
Let xi−1 and xi be the initial approximations to the root of (1.1) and xi+1 be
obtained using (1.2). Our method suggests that
1. If fi+1 · fi < 0, then (xi , fi ) replaces (xi−1 , fi−1 ) and
2. If fi+1 · fi > 0, then we provide a better approximation using Newton–Raphson
method.
We proceed in a similar manner for the next iterations. We apply the Interme-
diate Value Theorem each time to guarantee the confinement of the root within the
interval and the introduction of Newton–Raphson method will remove the retained
end point thereby enhances the rate of convergence.

30
 In the next section, we will analyse the rate of convergence and efficiency of the
Modified LIM with the assumption that the given function in (1.1) is continuously
differentiable in the interval (xi−1 , xi ), where xi−1 , xi are the initial approximations.

12. ERROR ESTIMATION

Let x0 and x1 be two sufficiently close initial approximations such that the interval
(x0 , x1 ) contains a root of (1.1). If εi is the error in the ith iteration, then we have
εi = xi − α.
Now using the Taylor series expansion of f (xi + α) we have
∞
f (k) (α)
fi = f (xi ) = ∑ ak εik , where ak = , k = 1, 2, 3, . . .
k=1 k!
By substituting this in (1.2) and neglecting the third and higher order deriva-
tives, it is easy to show that
a2
εi+1 ∼ − εi−1 εi , (3.1)
a1
which is the error estimation for the unmodified iteration. If fi+1 fi < 0 then the
error estimation in the next iteration is
a2
εi+2 ∼ − εi εi+1 , (3.2)
a1
otherwise, we have
a2 2
εi+2 ∼ −
ε , (3.3)
a1 i+1
which is the error estimation for the modified iteration. Now using (3.1) and (3.3)
we can examine the asymptotic behaviour of our modified LIM. Throughout the
analysis we assume that εi−1 is negative and εi positive. Now, if we consider the
c2
case where is positive, we find a sequence of values, calculated from modified
c1
(M) and unmodified (U) iterations in the order UUM, UUM, UUM, . . . and the se-
c2
quence is UM, UUM, UUM, . . . for the case where is negative. So in each step
c1
we have two unmodified iterations followed by a modified iteration. Therefore
from (3.1) and (3.3) we have
a32 5
εi+2 ∼ − 3 εi−1 , (3.4)
a1
31
and if we consider each sequence ‘UUM’ as a single iteration, then we have
a32 5
ζi+1 ∼ − 3 ζi−1 , (3.5)
a1

where ζi is the error in ith iteration, when each sequence is considered as a single
iteration.
Now our method requires 5 computations per iteration, so the computational ef-
1
ficiency of our method is 5 5 = 1.37972966 . . ., which is better than the unmodified
method, which has both rate of convergence and efficiency 1.

14. NUMERICAL AND GRAPHICAL ILLUSTRATIONS

We show first in Table 1, the behaviour of the method when applied to the equation

sin(x) + xex + 1 = 0.

The actual root is 0.709, which lies in the interval [−1, 0]. We begin with the
approximations
x0 = −1 and x1 = 0.

A Modification of Linear Interpolation Method for Computing the Roots of

an Equation using Newton–Raphson Method

Table 1

i εi iterations
2 −0.11789 U
3 −0.44419 × 10−1 U
4 0.209 M
5 −0.6017 × 10−2 U
6 −0.799 × 10−3 U
7 0.24972 × 10−1 M
8 −0.106 × 10−3 U
9 −0.2 × 10−5 U
10 −0.1569 × 10−2 M
11 −0.1 × 10−6 U

From Table 1 above we can see the fast convergence after the modified (M)
iterations.

32
 We now check the performance of our method by comparing the results and
number of iterations with those already available methods. The algorithms chosen
for comparison were the Regula–Falsi algorithm, Bisection algorithm and a hybrid
method of Regula–Falsi and the Bisection algorithm.
We chose 0.5 × 10−19 as the tolerance error, i.e. we will terminate the process
when
| f (xi )| ≤ 0.5 × 10−19 .
Also in each case there exists a root in the interval [0, 1].
Table 2
Table 1: f (x) = x2 − (1 − x)n . One inflexion on [0, 1].
No of iterations
N Bisection Regular–Falsi Illinois Modified LIM
2 1 1 1 1
5 64 54 8 8
15 61 179 11 9
20 62 245 12 9

Table 3
Table 2: f (x) = 2xe−n + 1 − 2e−nx . No inflexions on [0, 1].
No of iterations
N Bisection Regular–Falsi Illinois Modified LIM
1 64 23 19 21
5 64 40 21 21
15 67 41 23 17
20 67 42 23 17

Table 4
Table 3: f (x) = 1 + (1 − n)2 x − (1 − nx)2 . One turning point on [0, 1].

No of iterations
N Bisection Regular–Falsi Illinois Modified LIM
2 64 25 9 8
5 62 16 9 7
15 71 11 7 7
20 71 10 7 6

33
 Table 5
Table 4: f (x) = 1 + (1 − n)4 x − (1 − nx)4 . One turning point and one inflexion on [0, 1].

No of iterations
N Bisection Regular–Falsi Illinois Modified LIM
2 64 40 10 10
5 71 9 7 7
15 76 6 6 6
20 77 5 6 5

table 6
Table 5: f (x) = e−nx (x − 1) + xn . A family of curves which lie increasingly close to the x-axis for
large n.
No of iterations
N Bisection Regular–Falsi Illinois Modified LIM
1 64 26 9 11
5 63 114 9 8
15 57 1286 13 6
20 55 > 10 4 16 10

table 7
nx − 1
Table 6: f (x) = . A family of curves with y-axis asymptotic.
(n − 1)x
No of iterations
N Bisection Regular–Falsi Illinois Modified LIM
2 59 2008 14 15
5 56 809 14 14
15 61 262 14 13
20 58 192 15 13

15. Comparison Between Usual Methods and Modified Method

Numerical methods for solving nonlinear equations can broadly be classified into
usual (classical) methods and modified (hybrid) methods. Each method differs
in terms of convergence, efficiency, reliability, and computational cost. In this
section, a detailed comparison between these methods is presented.

0.0.1 Classification of Methods

• Usual (Classical) Methods:

34
 – Bisection Method
– Regula–Falsi Method
– Secant Method
– Newton–Raphson Method
• Modified Method:
– Modified Linear Interpolation Method (Modified LIM using Newton–
Raphson correction)

0.0.2 Detailed Comparison of Methods

Feature Bisection Regula–Falsi Secant

Nature of method Bracketing Bracketing + Interpolation Open method
Uses IVT Yes Yes No
Derivative required No No No
Initial requirement f (a) f (b) < 0 f (a) f (b) < 0 Two guesses
Convergence guarantee Guaranteed Guaranteed Not guaranteed
Rate of convergence Linear Linear Superlinear (≈ 1.618
Speed of convergence Very slow Moderate Fast
Endpoint stagnation No Yes No
Sensitivity to initial guess No No Yes
Computational cost Low Low Moderate
Numerical stability Very high High Moderate

0.0.3 Conceptual Comparison

The Bisection method is the most reliable but converges very slowly. The Regula–
Falsi method improves convergence by interpolation but often suffers from end-
point retention. The Secant method converges faster but lacks guaranteed conver-
gence. The Newton–Raphson method exhibits quadratic convergence but is highly
sensitive to the initial guess and requires derivative evaluation.

35
 The Modified Linear Interpolation Method combines the reliability of brack-
eting methods with the rapid convergence of the Newton–Raphson method. By
applying Newton–Raphson correction whenever endpoint retention occurs, the
modified method significantly improves convergence speed while maintaining sta-
bility.

0.0.4 Error and Convergence Comparison

Method Error Relation Order

Bisection Method en+1 = 21 en 1
Regula–Falsi Method Linear decrease of error 1
Secant Method en+1 ≈ e1.618
n 1.618
Newton–Raphson en+1 ≈ e2n 2
Method
Modified LIM en+1 ≈ e5n (asymptotic) >1

0.0.5 Computational Efficiency

Method Efficiency Index
Bisection Method 1
Regula–Falsi Method 1
Secant Method ≈ 1.47
Newton–Raphson Method ≈ 1.41
Modified LIM ≈ 1.38

0.0.6 Conclusion
The usual root-finding methods are either reliable but slow (Bisection and Regula–
Falsi) or fast but conditionally convergent (Secant and Newton–Raphson). The
Modified Linear Interpolation Method successfully overcomes these limitations
by combining bracketing reliability with Newton–Raphson acceleration. As a re-
sult, the modified method exhibits improved convergence, enhanced efficiency,
and greater practical applicability for solving nonlinear equations.

36
16. CONCLUSION
This dissertation presents a comprehensive study of root-finding methods for non-
linear equations. Each method has its own strengths and limitations. The choice
of method depends on the nature of the problem, required accuracy, and com-
putational efficiency. Newton–Raphson method is the fastest among the studied
methods, while the Bisection method is the most reliable.
The modified LIM has the rate of convergence 5, which is relatively fast com-
pared to other known methods. The efficiency of the method is something that re-
quires some improvement. As our method is a modification by Newton–Raphson
method, the modified LIM sometimes diverges when the initial approximations
are taken far away from the exact root or the given function is not differentiable at
some point in the interval. Despite this, the method provides results of high accu-
racy when the initial approximations are taken with special care. Similar to other
iterative methods, there is a scope of modification of our method and we wish to
have one in future. With this we conclude our work.

37
17. REFERENCES
[1] Dowell, M. and Jarratt, P., A modified regula falsi method for computing the
root of an equation. BIT Numerical Mathematics, 11:168–174, 1971.
[2] Isaacson, E. and Keller, H., Analysis of Numerical Methods. Wiley, 1966
(or Dover 1994).
[3] Stoer, J. and Bulirsch, R., Introduction to Numerical Analysis. Springer, 3rd
edition, 2002.
[4] Traub, J. F., Iterative Methods for the Solution of Equations. Prentice-Hall,
Englewood Cliffs, N. J., 1964.
[5] Quarteroni, A. and Saleri, F., Scientific Computing with Matlab. Springer,
2003.

38