Journal of Physics: Conference Series

PAPER • OPEN ACCESS

A COMBINATORIAL APPROACH TO DECIDE INITIAL ROOT VALUE

FOR THE SOLUTION OF NON-LINEAR SYSTEM OF EQUATIONS
To cite this article: Gananathan Ramaiyan et al 2019 J. Phys.: Conf. Ser. 1228 012063

View the article online for updates and enhancements.

This content was downloaded from IP address 178.173.237.93 on 03/06/2019 at 13:36

International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012063 doi:10.1088/1742-6596/1228/1/012063

A COMBINATORIAL APPROACH TO DECIDE INITIAL

ROOT VALUE FOR THE SOLUTION OF NON-LINEAR
SYSTEM OF EQUATIONS

Gananathan Ramaiyan1, Sulaiman.S.M2 and Irene Hepzibah.R3

1
School of Engineering, Asia Pacific University of Technology and Innovation,
Bukit Jalil, Kuala Lumpur, Malaysia; E-mail: [email protected]
2
School of Computing, Asia Pacific University of Technology and Innovation,
Bukit Jalil, Kuala Lumpur, Malaysia; E-mail: [email protected]
3
Department of Mathematics, T. B. M. L. College, Porayar, Tamil Nadu State, India

E-mail: [email protected]

Abstract. This study aims at developing the real root of an algebraic or a transcendental
equation based on non-linear model and a combinatorial numerical approach for solving
nonlinear equations. The introduction of Bisection, False-Position and Newton–Raphson
methods to identify the real positive root of an equation which is provided here. The rate of
convergence is fastest in Newton–Raphson method due to the quadratic convergence compared
with the other two iterative methods. We have applied here the three iterative methods in solving
nonlinear equations in MATLAB environment and compared the outcomes of results as well as
geometrically also. Finally, an illustrative algebraic example is discussed, and the MATLAB
outputs are provided to check the validity of the new approach.

Keywords: Bisection method, False-Position method, Newton–Raphson method, Non-Linear

equations and Rate of Convergence.

1. Introduction
We can find the roots of an algebraic or transcendental equation by using 3 iterative methods in
Numerical methods namely Bisection, False position and Newton–Raphson method. Applications in
engineering is to determining the root value of t in Fourier analysis of the instantaneous value of a
waveform, the value of t in damped oscillation of a system and determining the value of t in motion of
a particle in an electrostatic field which is represented as y = f(t). This area of research has been studied
by many researchers. Claudio Gutierrez et al studied about Complexity of the bisection method in 2007.
Saba Akram et al [7] wrote a research paper about Newton Raphson Method which is all-inclusive to
solve the non-square and non-linear problems. The study also compared the rate of performance, rate of
convergence of Bisection method, root findings of the Newton method and False Position method. Later
a Comparative Study of Bisection, Newton-Raphson and False Position Methods of Root- Finding
Problems were done by Ehiwario, J.C [5]. In a more recent review, Ali Ebrahimnejad et al [2] proposed
a revisit of numerical approach for solving linear fractional programming problem in a fuzzy
environment. Extensive studies have been done and many methods of solving nonlinear equations are
also available in the literature [1,3,4,6,7]. In this paper we consider nonlinear equations and operational
methodologies for solving the problem are provided. We have applied here three iterative methods in
solving nonlinear equations and compared the outcomes with the help of results attained and their
MATLAB outputs.

Content from this work may be used under the terms of the Creative Commons Attribution 3.0 licence. Any further distribution
of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.
Published under licence by IOP Publishing Ltd 1
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012063 doi:10.1088/1742-6596/1228/1/012063

The paper is organized as follows. In section 2, the preliminaries are briefly introduced and the
perception of nonlinear equations with essential definitions and notions with iterative methods are
explained. An application of these results and their MATLAB outputs are specified by numerical
illustration in section 3 and some concluding remarks are given in section 4.

2. Preliminaries

2.1. Algebraic Equation

An equation which involves only function of x is called an algebraic equation.

2.2 Transcendental Equation

A transcendental equation is an equation containing a transcendental function of the variable(s) being
solved for. It contains the exponential functions, the logarithmic functions and the trigonometric
functions.

2.3 Root

A real number which satisfies the real root of the equation f x( ) =0 if and only if f ( ) =0.

2.4 Intermediate value theorem

Let f(x) be a continuous function in [a, b]. Let f a( ) 0 and f b( ) 0then f x( ) =0 will have atleast one

root lies between a and b. This is depicted in figure1 shown below:

3. Iterative methods
There are three iterative methods (repetitive methods) are going to discuss and the corresponding figure
are explained as below: They are:
3.1. Bisection Method or Bolzano Iterative method
3.2. False Position Method or Regula-Falsi method
3.3. Newton-Raphson Method or Method of Tangents

2
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012063 doi:10.1088/1742-6596/1228/1/012063

3.1. The Bisection method (or) The Bolzano method (or) Interval Halving method

Consider the equation f x( ) =0 . Assume the value a and b so that f a( ) 0 and f b( ) 0or vice versa.
There is a change of sign between a and b so that the root lies between a and b. Let a < b then
+b
a find x1 = and f x( ) 01 = . Therefore x1 is the root of f x( )
=0 otherwise: 2

+b
(i) f x( ) 01 , the root lies between x1 and b. x2 = x1 and find f x( )2 and so on (or)
2

(ii) f x( ) 01 , the root lies between a and x1. x2 = a + x1 and find f x( )2 and so on. Proceeding in 2
this way until the last two consecutive roots are correct up to the desired accuracy equal, stop the
iterations. This method is known as the Bisection or Bolzano method.
For example, identify the real positive root of an algebraic equation f x( )= + − =x3x2 1 0 . The

MATLAB output results are depicted in table 3.1.1. and corresponding graph is shown in figure 3.1.2.

Table 3.1.1.

Bisection Method
a(ve) b Sign of
Iteration (+ve) c=(a+b)/2 f(c) f(c)
-
1 0 1 0.5 0.625 negative
-
2 0.5 1 0.75 0.016 negative
3 0.75 1 0.875 0.436 positive
4 0.75 0.875 0.813 0.197 positive
5 0.75 0.813 0.781 0.087 positive
6 0.75 0.781 0.766 0.035 positive
7 0.75 0.766 0.758 0.009 positive
8 0.75 0.758 0.754 0.003 negative
9 0.754 0.758 0.756 0.003 positive
10 0.754 0.756 0.755 0 positive

3
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012063 doi:10.1088/1742-6596/1228/1/012063

Figure 3.1.2

3.2. The Regula -Falsi method (or) The method of False-Position:

Consider the equation f x( ) =0 . Assume the value a and b so that f a( ) 0 and f b( ) 0or vice versa.
There is a change of sign between a and b so that the root lies between a and b. The graph of
y = f x( ) crosses x-axis and y = 0. The equation of the tangent line joining between two points A and

B is: =
y− =y0m x( − x0), where m is the slope of tangent line.m dydx (x y0 0, )

− f a(
y− f a( )= f b( )
)
(x−a) , b−a
x
The first approximation to the root is x = 1 and in x- axis y = 0

−0 f a( )= f
b( )− f a( )(x1 −a) b−a

−a)
= −x1 a (b f a( )
f b( )− f a( )

=x1 af b( )−af a( )−bf a( )+af a( )

f b( )− f a( )

xn+1 = af b( )−bf a( ) and

find f x( )1 f b( )− f a( )

4
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012063 doi:10.1088/1742-6596/1228/1/012063

If f x( ) 01 then x1is a and f x( )1 is f a( ) .

If f x( ) 01 then x1is b and f x( )1 is f b( ) . Proceeding like this we will get the root x2 , x3 and so on.

Finally we will see that the two successive roots are equal to up to the desired accuracy then stop the
iterations. This method is called the method of false position. The procedure is geometrically depicted
in figure

B[b,f(b
x=d

x=c
x=a

x=b

A[a,f(a) C[c,f(c) Root

Figure 3.2.1.
For example, identify the real positive root of an algebraic equation f x( )= + − =x3x2 1 0 . The

MATLAB output results are depicted in table 3.2.1. and corresponding graph are as shown in figure
3.2.1

Regula-Falsi Method
a(ve)
Iteration b(+ve) f(a) f(b) c f(c)
-
1 0 1 -1 1 0.5 0.625
- -
2 0.5 1 0.625 1 0.692 0.189
- -
3 0.692 1 0.189 1 0.741 0.043
- -
4 0.741 1 0.043 1 0.752 0.009
- -
5 0.752 1 0.009 1 0.754 0.002
-
6 0.754 1 0.002 1 0.755 0

5
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012063 doi:10.1088/1742-6596/1228/1/012063

Figure 3.2.1
3.3 The Newton-Raphson method (or) The method of Tangents:

Consider the equation f x( ) =0 . Draw the tangent to the curve and it meets at point P to the curve and

meets the x-axis is ( (x n+1,

0)
).
The equation of tangent at a point
P
((x f x
n, ( )n ) is y−

=yn m x( − xn), where m is the slope of tangent line.

m= dydx (x yn n,) y− f x( )n = dydx (x yn n, ) (x− xn) ,

y− f x( )n = f1( )(xn x− −−−−xn) (1)

(x
Since ( n+1,
0)
)lies in equation (1), we get
−0 f x( )n = f1( )(xn xn+1−xn) + f x( n) ,n =

0,1,2,3...... xn 1 = −xn f 1(xn)

xn+1 = x fn 1(fx1n()x−n)f x( n) ,n = 0,1,2,3......

6
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012063 doi:10.1088/1742-6596/1228/1/012063

This is known as the Newton-Raphson method. Taking the initial root value choose from the root lies
between a and b in [a, b]. Substituting the root value in the formula consecutively we will get the root
which is equal to the desired accuracy.
The rate of convergence is faster in Newton-Raphson method compare with the other two methods.
The condition for convergence of Newton-Raphson method is similar to the iteration method

y y

A[x0,f(x0)]

B[x1,f(x1)]

X
O x

X Roo

'
( x) 1
fx( )
xn +1 = xn − ' n = ( xn )
f ( xn )

' (x) = dxd x− ff x'((x))

f ' (x f). ' (x) − f x f( ). ''(x)

= −1 [ f '(x)]2

f x f( ). ''(x)
1
= − +1 [ f '(x)]2

' (x) = f x f[(f ).'(x)]''(2x)

' (x) 1f x f[(f ).'(x)]''(2x) 1

7
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012063 doi:10.1088/1742-6596/1228/1/012063

ie. ., f x f( ). ''(x) [ f '(x)] 2

Newton Raphson method has quadratic convergence.

Suppose we decide the initial root value as

x0 = 0.500 f x f( ) ''
( )x f ' ( )x 2

f x f( ) '' ( )x f ' ( )x 2

We are deciding the initial root value from

Regula-Falsi method as x0 = 0.500and check

with condition that satisfy the relation

2 f x f( ) '' ( )x f ' ( )x

f (0.5) f '' (0.5) f' (0.5)2

Table 3.3.1

Newton-Raphson Method
Iteration x x(n+1)
1 0.5 0.857
2 0.857 0.764
3 0.764 0.755
4 0.755 0.755

For example, identify the real positive root of an algebraic equation f x( )= + − =x3x2 1 0 . The

MATLAB output results are depicted in table 3.3.1. and the corresponding graph are shown in figure
3.3.2.
f x( )= + − =x3 x2 1

0 f ' ( ) 3x = x2 +2x f'

( ) 6x = +x 2 The rate of

8
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012063 doi:10.1088/1742-6596/1228/1/012063

convergence is: f x f( )
'' ( )x f ' ( )x 2
We can decide any value from [0, 1] must satisfy the relation.

Suppose we decide the initial root value as x0 = 0.500 f (0.5) f '' (0.5) f ' (0.5) 2

( 0.625)(5)− 1.75 2
−3.125 3.063

We are deciding the initial root value from Regula-Falsi method as x0 = 0.500

3.4 Numerical Example:

The critical speed of oscillation of a loaded beam is denoted by x are given by the equation f
x( )= + − =x3 x2 1 0 .
Identify the real root of critical speed of oscillation of a loaded beam given in an algebraic equation
f x( )= + − =x3 x2 1 0 , correct to 3 decimal places writing MATLAB code:
(a) the Bisection method,
(b) the Regula-Falsi method, and
(c) the Newton-Raphson method, taking the initial root value from Regula-Falsi method,

9
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012063 doi:10.1088/1742-6596/1228/1/012063

(d) Identify with reason which method converges fastest. Compare the results with other two methods in table
and project these values in graph.

3.5 MATLAB Programming code:

% Finding the root of the following f
= @(x) x^3 + x^2 - 1;
% f = @(x) 4 * x^2 - 6 * x - 7;
% Finding the initial guess for
i = -5:5 low = i;
high = i+1; if
f(low) * f(high) < 0
break; end end
% Bisection Method fprintf('\nBisection
Method');
fprintf('\nIteration\ta\t\t\tb\t\t\tc\t\t\tf(c)\t\tSign of f(c)\n');
i = 1; a = low; b = high; c = (a + b)/2; x(i) = c; prevc = c
+ 10; while abs(c - prevc) >= 1e-3 if ( i ~= 1)
prevc =c; end c = (a + b)/2; x(i) = c; sgn = '-ve';
if f(c) > 0 sgn = '+ve';
end fprintf('%d\t\t\t%.3f\t\t%.3f\t\t%.3f\t\t%.3f\t\t\t%s\n',i,a,b,c,f(c),sgn);
if f(a)*f(c)<0 b = c; else a = c; end i = i + 1;

end
% Plot the results figure('Color','White')
plot(1:i-1,x(1:i-1),'*-b') hold on

title('Root of: $f(x) = x^3 + x^2 - 1 $','FontSize', 20,'Interpreter','latex')

xlabel('Iteration','FontSize',16)
ylabel('$x$','FontSize',16,'Interpreter','latex') bs =x;
% Regula-Falsi Method fprintf('\nRegula-Falsi
Method');
fprintf('\nIteration\ta\t\t\tb\t\t\tf(a)\t\tf(b)\t\tc\t\t\tf(c)\n');
i = 1; a = low; b = high;
c = (a * f(b) - b * f(a)) / (f(b) - f(a));
x(i) = c; prevc = c + 10; while abs(c
- prevc) > 1e-3 if ( i ~= 1)
prevc =c; end

10
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012063 doi:10.1088/1742-6596/1228/1/012063

c = (a * f(b) - b * f(a)) / (f(b) - f(a));

x(i) = c;
fprintf('%d\t\t\t%.3f\t\t%.3f\t\t%.3f\t\t%.3f\t\t%.3f\t\t%.3f\n',i,a,b,f(a),f(b),c,f(c));
if f(a)*f(c)<0 b = c; else a = c; end i = i + 1; end
% Plot the results
plot(1:i,x(1:i),'*-g')
hold on rf = x;
Newton-Raphson Method fprintf('\nNewton-Raphson
Method'); fprintf('\nIteration\tx\t\t\tx(n+1)\n'); %
First-Order derivative of f fp = @(x) 3*x^2 + 2*x; x0
= 1; % Initial guess
N = 10; % Maximum number of iterations tol
= 1e-3; % Convergence tolerance
x = zeros(N + 1,1); % Pre allocate solution vector where row => iteration
x(1) = x0; % Set initial guess n = 2;
nfinal = N + 1; % Store final iteration if tol is reached before N iterations
while (n <= N + 1) fe = f(x(n - 1)); fpe = fp(x(n - 1)); x(n) = x(n -
1) - fe/fpe;
fprintf('%d\t\t\t%.3f\t\t%.3f\n',n-1,x(n-1),x(n));
if (abs(fe) <= tol)
nfinal = n; % Store final iteration
break; end n = n + 1; end
% Plot the results plot(1:nfinal,x(1:nfinal),'*-r')
legend('show','Bisection','Regula-Falsi','Newton-Raphson') nw
= x;

4. Results and Discussion

Comparison the results with three iterative methods:

Iterations N-R R-F Bisection method

method method
1 0.857 0.5 0.5
2 0.764 0.692 0.75
3 0.755 0.741 0.875
4 0.755 0.752 0.813
5 0.754 0.781
6 0.755 0.766
7 0.755 0.758

11
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012063 doi:10.1088/1742-6596/1228/1/012063

8 0.754
9 0.756
10 0.755
11 0.755

The following MATLAB Figure shows the convergence of the three methods and the number of
iterations to be taken to obtain the root. It can be seen that the error of the Bisection method are
frustrated along the iteration. While the errors of the Regula-Falsi and Newton Raphson method
decrease when the number of iteration increases. The result shows that the Bisection method presents
the biggest number of iterations and time in all cases. However, the Regula Falsi method have fairly
similar results whereas the Newton Raphson method have the smallest number of iterations and have
the smallest time in most cases.

Figure 4.2.1

From the figure 4.2.1, we see that the Newton-Raphson method is the fastest in the rate of
convergence. Comparing Newton Raphson with Bisection method 2.5 times faster and with
RegulaFalsi method 1.5 times faster.
Conclusion

In this work, we have presented the three iterative methods in solving non-linear equations in
MATLAB environment. The MATLAB outputs of the optimal solutions are also provided to check the
effectiveness of the proposed method and compared the outcomes. The solution methodology offered
here will be useful for many real life problems in engineering. This research work can be extended to
solve non-linear equations in fuzzy environment in the future.

12
International conference on computer vision and machine learning IOP Publishing
IOP Conf. Series: Journal of Physics: Conf. Series 1228 (2019) 012063 doi:10.1088/1742-6596/1228/1/012063

References

[1] Abdulaziz G. Ahmad, Comparative Study of Bisection and Newton-Rhapson Methods of RootFinding
Problems, international Journal of Mathematics Trends and Technology,19 (2015) 121-129.
[2] Ali Ebrahim nejad et al, A revisit of numerical approach for solving linear fractional programming
problem in a fuzzy environment, Applied Mathematical Modelling 57 (2018) 459–47.
[3] Ching-Long Su et al, A two-stage algorithm integrating genetic algorithm and modified Newton method
for neural network training in engineering systems, Expert Systems with Applications 38 (2011) 12189–
12194.
[4] Claudio Gutierrez et al, Complexity of the bisection method, Theoretical Computer Science 382 (2007)
131–138.
[5] Ehiwario, J.C., Aghamie, S.O, Comparative Study of Bisection, Newton-Raphson and Secant Methods
of Root- Finding Problems, IOSR Journal of Engineering 04 (2014) 01-07.
[6] Khalid Ali Hussein et al., Parallel Hybrid Algorithm of Bisection and Newton-Raphson Methods to Find
Non-Linear Equations Roots, IOSR Journal of Mathematics 11(2015) 32-36.
[7] Saba Akram et al, Newton Raphson Method, International Journal of Scientific & Engineering Research,
6(2015) 1748-1752.

13