Available online at http://scik.

org

J. Math. Comput. Sci. 2 (2012), No. 5, 1512-1521

ISSN: 1927-5307

EXTENDED AND MODIFIED HALLEY’S ITERATIVE METHOD FOR

SOLVING NONLINEAR SYSTEMS

M.A. HAFIZ1,* AND MOHAMED. S. M. BAHGAT1,2

1
Department of Mathematics, Faculty of Science and Arts, Najran University, 1988, Saudi Arabia
2
Department of Mathematics, Faculty of Science, Minia University, Minia 61111 Egypt

Abstract: In this paper, an efficient method is constructed and used for solving system of nonlinear

equations. The method based on a Halley’s iterative method (HIM). This technique is revised and

modified to solve a system of nonlinear equations of n-dimension with n-variables. One of the

obtained methods have order six of convergence. In addition, the proposed method has been tested on a

series of examples published in the literature and show good results when compared it with the

previous literature.

Keywords: Halley’s method; Nonlinear system; Iterative method; Taylor series; Newton’s method;

Householder’s method.

2000 AMS Subject Classification: 65H10; 65B99; 65K20

1. Introduction
Recently, several iterative methods have been made on the development for solving nonlinear
equations and system of nonlinear equations. These methods have been improved using several
different techniques including Taylor series, quadrature formulas, homotopy and decomposition
techniques, see [1–9] and references therein. He [10] suggested an iterative method for solving the
nonlinear equations by rewriting the given nonlinear equation as a system of coupled equations. This
technique has been used by Chun [11] and Noor [12,13] to suggest some higher order convergent

__________
*
Corresponding author
Received May 30, 2012
1512
1513 EXTENDED AND MODIFIED HALLEY’S ITERATIVE METHOD

iterative methods for solving nonlinear equations. Newton method is the well-known iterative method
for finding the solution of the nonlinear equations. There exist several classical multipoint methods
with fourth-order and sixth-order convergence for solving nonlinear equations. It is well known [9] that
the two-step Newton method has fourth-order convergence, which has been suggested by using the
technique of updating the solution. Hafiz and Bahgat [14] modified Householder iterative method for
solving system of nonlinear equations. He also show that this new method includes famous two step
Newton method as a special case. Noor [6] have suggested and analyzed a two-step Halley method
using the Newton method as a predictor and Halley’s method as corrector. It has been shown that this
two-step Halley method is of sixth-order convergence and is an efficient one. In this paper we consider
the case of nonlinear systems of equations and we extend and modify the Halley method which
constructed by Noor [6]. This technique, however, needs to be revised to solve the system of nonlinear
equations. Some illustrative examples have been presented, to demonstrate our method and the results
are compared with those derived from the previous methods. All test problems reveals the accuracy and
fast convergence of the new method. In the next sections, we use the same notation for the n-
dimensional case as for the one-dimensional case, interpreting the symbols appropriately.

2. Halley iterative method

Suppose we have system of nonlinear equations of the following form
f 1(x 1, x 2 , , x n )  0,
f 2 (x 1, x 2 , , x n )  0,

f n (x 1, x 2 , , x n )  0,

where f i :Rn R and the functions f i is differentiable up to any desired order [15],

can be thought of as mapping a vector X  (x 1, x 2 , , x n )T of the n-dimensional

space Rn , into the real line R . The system can alternatively be represented by
defining a functional F : Rn  Rn , by

F (x 1, x 2 , x n )  [f 1(x 1, x 2 , x n ), , f n (x 1, x 2 , x n )]T
Using vector notation to represent the variables x 1, x 2 , x n , the previous system
then assumes the form:

F (x)  0 (1)

For simplicity, we assume that X* is a simple root of Eq. (1) and X0 is an initial guess

sufficiently close to X*. Using the Taylor’s series expansion of the function f k (x) , we
have
 M.A. HAFIZ1,* AND M.S.M.BAHGAT1,2 1514

1 n 1 n n

f k (X0 )  [ (x i  x i(0) )f k ,i (X0 )]  [
1! i 1  (x  x i(0) )(x j  x (0)
2! i 1 j 1 i j )f k ,ij (X0 )]  ...  0 (2)

f k 2f k
where k  1,2,..., n , f k ,i  , f k ,ij  and X0  [x 1(0) , x (0) ,..., x (0) ]T is
x i x i x j 2 n

the initial approximation of Eq. (1). Matrices of first and second partial derivatives
appearing in equation (2) are Jacobian J and Hessian matrix H respectively. In
matrix notation

1 n
F (X0 )  J (X0 )[X(1)  X0 ]   e [X(1)  X0 ]T H i (X0 )[X(1)  X0 ]  0
2! i 1 i
(3)

from which we have


 1 n 



J (X0 )  
2! i 1
ei [X(1)  X0 ]T H i (X0 )[X(1)  X0 ]  F (X0 )

 

1

 1 n 

X (1)
 X0  J (X0 ) 


2! i 1
ei [X(1)  X0 ]T H i (X0 ) F (X0 )

(4)
 
where H i is the Hessian matrix of the function fi ,  is the Kronecker product and ei

is a n 1 vector of zero except for a 1 in the position i. First two terms of the
equation (3) gives the first approximation, as

X(1)  X0  J 1(X0 )F (X0 ) (5)

Substitution again of (5) into the right hand side of (4) gives the second
approximation

1

 1 n 

X (1)
 X0  J (X0 ) 


2! i 1
ei [J 1(X0 )F (X0 )]T H i (X0 ) F (X0 )

(6)
 

This is extended Halley’s method and this formulation allows us to suggest the
following iterative methods for solving system of nonlinear equations (1).

3. Algorithms
1515 EXTENDED AND MODIFIED HALLEY’S ITERATIVE METHOD

In this section, we suggest four algorithms for solving system of nonlinear

equations:
(k ) (k ) (k )
Algorithm.1. For a given X(k )  [x 1 , x 2 ,..., x n ]T calculate the approximation
( k 1) ( k 1) ( k 1)
solution X(k 1)  [x 1 ,x2 ,..., x n ]T for k =0,1, 2, … by the iterative scheme

X(k 1)  X(k )  J 1(X(k ) )F (X(k ) ) (7)

which is the Newton–Raphson method for n dimension.

(k ) (k ) (k )
Algorithm.2. For a given X(k )  [x 1 , x 2 ,..., x n ]T calculate the approximation
( k 1) ( k 1) ( k 1)
solution X(k 1)  [x 1 ,x2 ,..., x n ]T for k =0, 1, 2, … by the iterative schemes

1


n 


ei
X(k 1)  X(k )  J (X(k ) )  [J 1(X(k ) )F (X(k ) )]T H i (X(k ) ) F (X(k ) ) (8)

 i 1
2! 

We also remark that, if f k ,ij  0, k , i , j  1,2,..., n , then Algorithm 2 reduces to the

Newton method. That is, Algorithm 2 which is the generalized of Halley’s method for
solving system of nonlinear equations. Several authors have already studied the
convergence of iteration (8) in Banach space setting (see [16] and etc.). Now using
Algorithm 1 as a predictor and Algorithm 2 as a corrector, we suggest and analyze a
new two-step iterative method for solving system of nonlinear equations.

(k ) (k ) (k )
Algorithm.3. For a given X(k )  [x 1 , x 2 ,..., x n ]T calculate the approximation
( k 1) ( k 1) ( k 1)
solution X(k 1)  [x 1 ,x2 ,..., x n ]T for k =0,1, 2, … by the iterative schemes

Y(k ) =X(k )  J 1(X(k ) )F (X(k ) ), (9)

1

 1 n 

X (k 1)
Y
(k )
 J (Y(k ) ) 
 2! ei [J 1(Y(k ) )F (Y(k ) )]T H i (Y(k ) ) F (Y(k ) )

(10)
 i 1 
The famous for Algorithm 3 states that the maximum order of convergence is six,
when the function is of one variable (see [6]). Algorithm 3 is extended of modified
 M.A. HAFIZ1,* AND M.S.M.BAHGAT1,2 1516

Halley method. It is clear that for Y(k )  X(k ) Algorithm 3 is exactly Algorithm 2. If,
f k ,ij  0, k , i , j  1,2,..., n then Algorithm 3 reduce to the following two-step

method.

(k ) (k ) (k )
Algorithm.4. For a given X(k )  [x 1 , x 2 ,..., x n ]T calculate the approximation
( k 1) ( k 1) ( k 1)
solution X(k 1)  [x 1 ,x2 ,..., x n ]T for k =0,1, 2, … by the iterative schemes

Y(k )  X(k )  J 1(X(k ) )F (X(k ) ), (11)

X(k 1)  Y(k )  J 1(Y(k ) )F (Y(k ) ). (12)

4. Numerical results
We present some examples to illustrate the efficiency of our proposed method.
Here, numerical results are performed by Maple 15 with 100 digits but only 14 digits are
displayed. In Tables 1, 2 we list the results obtained by generalized Halley method
(GHM), modified generalized Halley method, which we called, Hafiz and Bahgat
Method (HBM) and comparison them with Newton–Raphson method (NM) and
Householder method (MHIM) [14]. The following stopping criteria is used for
computer programs:

|| X (n 1)  X (n ) ||  || F (X (n ) ) || 1015

and the computational order of convergence (COC) can be approximated using the
formula,

COC 

ln || X (n 1)
X (n )
|| / || X (n )
X ( n 1)
|| 

ln || X (n )
X (n 1)
|| / || X (n 1)
X (n 2)
|| 
Table 2 shows the number of iterations, the computational order of convergence
(n 1)
(COC) , || X  X (n ) || and norm of the function F (X (n ) ) are also shown in
Table 2 for various methods.

4.1. Small systems of nonlinear equations.

1517 EXTENDED AND MODIFIED HALLEY’S ITERATIVE METHOD

Example 1. In a case of one dimension, consider the following nonlinear functions

2 2
7x 30
[14], f 1(x )  xe x  sin2 x  3cos x  5, with x 0  3 and f 2 (x )  e x 1

with x 0  4. .

Example 2. In a case two dimension, consider the following systems of nonlinear

functions [14],
 f 1(x , y )  x 2 10x  y 2  8  0
F3  x    , (x0, y0)=(2, 2).
f 2 (x , y )  xy  x 10 y  8  0
2

f (x , y )  x 4 y  xy  2x  y 1  0
F4  x    1 , (x0, y0)=(0.8, 0.8).
x 1
 f 2 (x , y )  ye  x  y  e  0
Example 3. In a case three dimension, consider the following systems of nonlinear
functions [18].
 f 1(x , y , z )  15x  y 2  4z 13  0


F5  x   f 2 (x , y , z )  x 2  10 y  e z 11  0 , X0=(10, 6, -5).

f 3 (x , y , z )  y  25z  22  0
3


f 1(x , y , z )  3x  cos( yz )  0.5  0



F6  x   f 2 (x , y , z )  x  81( y  0.1)  sin z  1.06  0 ,
2 2
X0=(1.1, 1.1, 1.1).
 xy 10  3
f 3 (x , y , z )  e  20z 
3
0

4.2 Large systems of nonlinear equations.

In this subsection, we test HPM with some sparse systems with m unknowns
variables. In examples 4 through 6, we compare the NR method with the proposed
method HPM focusing on iteration numbers [14].

Example 4. Consider the following system of nonlinear equations:

F7 : f i  e x i 1, i  1,2,..., n.
The exact solution of this system is X* = [0,0,...,0]T. To solve this system, we set

X0  [0.5,0.5,...,0.5]T as an initial value. Table 6 is shown the result.

Example 5. Consider the following system of nonlinear equations:

 M.A. HAFIZ1,* AND M.S.M.BAHGAT1,2 1518

F8 : f i  x i2  cos(x i 1), i  1,2,..., n.

One of the exact solutions of this system is X* = [1,1,...,1]T. To solve this system, we
set X0 = [2,2.,...,2]T as an initial value. The results are presented in Table 6.

Example 6. Consider the following system of nonlinear equations [17]:

F9 : f i  cos x i 1, i  1,2,..., n.
One of the exact solutions of this system is X* = [0,0,...,0]T. To solve this system, we
set X0 = [2,2.,...,2]T as an initial guess. The results are presented in Table 6.

Table 1. Number of iterations for Examples 4-6.

Number of iterations F7 F8 F9 F7 F8 F9 F7 F8 F9
  1013 n =50 n =75 n =100
NM 6 7 26 6 7 26 6 7 26
GHM 4 5 18 4 5 18 4 5 18
MHIM 2 3 11 2 3 11 2 3 11
HBM 2 3 10 2 3 10 2 3 10

Fig. 1. Approximate solution against the number of iterations for f2

Number of iterations
1519 EXTENDED AND MODIFIED HALLEY’S ITERATIVE METHOD

Table 2.Numerical results for Examples 1-6

Methods & functions IT COC || X (n 1)  X (n ) || || F (X (n ) ) ||

F1, x0 =-3.
NM 15 2.00000001842997 4.62461580585008E-28 6.52399684550115E-54
GHM 8 3.00210029304863 2.57758679135367E-23 8.83803794298868E-68
MHIM 7 5.99944705395686 6.2904349110937E-65 5E-99
HBM 6 5.99637282489181 1.64283557214949E-42 5E-99
F2, x0 =4
NM 20 2.00000177738031 7.04037622176027E-21 4.23796972290579E-39
GHM 11 3.00002034672401 1.70247250684898E-23 9.03798392491937E-67
MHIM 9 5.99117673752232 7.41858978732436E-45 0
HBM 8 5.99110889033126 2.1845216055587E-47 0
F3, X0=(2, 2)
NM 9 2.00000140914961 4.37934644418127E-21 8.47815841458272E-41
GHM 6 3.00024293865263 3.62804255716776E-26 9.11319560570549E-77
MHIM 5 5.99934528347617 1.70074129087303E-85 1.01980390271855E-98
HBM 4 5.71112646345308 2.83986519147165E-19 1.0770329614269E-98
F4, X0=(2, 2)
NM 9 2.0003080790074 7.53061926402066E-19 2.30110082351703E-36
GHM 6 3.0051732467309 1.87851882722682E-26 1.85601502813661E-77
MHIM 5 6.00352249872292 1.94315799290308E-78 1.22065556157337E-99
HBM 4 5.86845135076057 5.61471494398326E-20 1E-99
F5, X0=(10, 6, -5)
NM 8 2.26089072336615 1.30827439612494E-16 5.41580264395725E-32
GHM 6 2.91590769468174 1.13291396861354E-42 1.41421356237309E-98
MHIM 5 3.9794335938325 1.28885942108228E-31 1E-98
HBM 4 3.84649791029683 1.54387130390901E-17 6.1593046396652E-71
F6, X0=(1.1, 1.1, 1.1)
NM 9 2.0000000032895 6.45188221156175E-20 3.36918186538804E-37
GHM 7 2.44656798277244 6.89405862265531E-17 1.4311394529322E-34
MHIM 5 4.19520706694058 2.82751409494534E-57 3E-99
HBM 4 5.55967632924813 2.10299366302151E-17 4.07677528685015E-70

In Tables 1-2, we list the results obtained by modified Halley iteration method. As
we see from this Tables, it is clear that, in most cases, the result obtained by HBM is
very superior to that obtained by MHIM, GHM and NM. Fig. 1 confirm this result.

5. Conclusions
 M.A. HAFIZ1,* AND M.S.M.BAHGAT1,2 1520

The Halley’s method continues to be an important subject of investigation. In our

study we extend the standard iteration in order to obtain robust algorithms based on
Halley iteration method. This method is generalized and modified and applied to solve
nonlinear system of equations. The numerical examples show that our method is very
effective and efficient. Moreover, our proposed method provides highly accurate
results in a less number of iterations as compared with Newton–Raphson method and
generalized Householder method.

REFERENCES
[1] A.R.Vahidi, Sh. Javadi and S.M. Khorasani s Solving System of Nonlinear Equations by restarted
Adomain,s method. Appl. Math. Comput. 6 (2012), 509 -516.
[2] JoséL. Hueso, Eulalia Martí
nez, Juan R. Torregrosa, Third and fourth order iterative methods free
from second derivative for nonlinear systems, App. Math. Comput. 211 (2009) 190–197.
[3] M. Grau, J.L. Diaz-Barrero, An improvement to Ostrowski root-finding method, Appl. Math.
Comput. 173 (2006) 450–456.
[4] A.S. Householder, The Numerical Treatment of a Single Nonlinear Equation, McGraw-Hill, New
York, 1970.
[5] M. Aslam Noor, Numerical Analysis and Optimization, Lecture Notes, Mathematics Department,
COMSATS Institute of Information Technology, Islamabad, Pakistan, 2006.
[6] Khalida Inayat Noor, M. Aslam Noor, Predicot-corrector Halley method for nonlinear equations,
Appl. Math. Comput. (2006), doi:10.1016/j.amc.2006.11.023.
[7] M. Aslam Noor, K. Inayat Noor, Iterative schemes for solving nonlinear equations, Appl. Math.
Comput. 183 (2006) 774–779.
[8] M. Aslam Noor, K. Inayat Noor, Three-step iterative methods for nonlinear equations, Appl. Math.
Comput. 183 (2006) 322–327.
[9] J.F. Traub, Iterative Methods for Solution of Equations, Prentice-Hall, Englewood Cliffs, NJ, 1964.
[10] J.H. He, A new iterative method for solving algebric equations, Appl. Math. Comput. 135 (2003),
81–84.
[11] C. Chun, Iterative methods improving Newton’s method by the decomposition method, Comput.
Math. Appl. 50 (2005) 1559–1568.
[12] M. Aslam Noor, New family of iterative methods for nonlinear equations, Appl. Math. Comput.
(2007), doi:10.1016/j.amc.2007.01.045.
[13] M. Aslam Noor, K. Inayat Noor, Modified iterative methods with cubic convergence for solving
nonlinear equations, Appl. Math. Comput. (2006), doi:10.1016/j.amc:2006.05.155.
[14] M.A.Hafiz, M.S.M. Bahgat, Modified of Householder iterative method for solving nonlinear
system of equations. J. Mathematical and Computational Science (2012) (Under Publication) .
[15] R.Burden,J.D.Faires, Numerical Analysis , 9th Edition ,Brooks/cole ,Boston,USA , 2011.
1521 EXTENDED AND MODIFIED HALLEY’S ITERATIVE METHOD

[16] S. Kanno, Convergence for the method of tangent hyperbolas, Math. Japonica 37 (4) (1992) 711-
722
[17] M.T. Darvishi and Byeng-Chun Shin, High-Order Newton-Krylov Methods to Solve Systems of
Nonlinear Equations. J. KSIAM, 15 (2011), 19–30.
[18] M.A.Hafiz, M.S.M. Bahgat, An Efficient Two-step Iterative Method for Solving System of
Nonlinear Equations, Journal of Mathematics Research; 4,(4), (2012) (Under Publication) .