This article was downloaded by: [New York University]

On: 03 October 2014, At: 04:30

Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer
House, 37-41 Mortimer Street, London W1T 3JH, UK

International Journal of Computer Mathematics

Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/gcom20

A sixth-order family of methods for nonlinear

equations
a
Beny Neta
a
Department of Mathematical Sciences , Northern llinois University , DeKalb,
Illinois, 60115, U.S.A
Published online: 20 Mar 2007.

To cite this article: Beny Neta (1979) A sixth-order family of methods for nonlinear equations, International Journal of
Computer Mathematics, 7:2, 157-161

To link to this article: http://dx.doi.org/10.1080/00207167908803166

PLEASE SCROLL DOWN FOR ARTICLE

Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”)
contained in the publications on our platform. However, Taylor & Francis, our agents, and our licensors
make no representations or warranties whatsoever as to the accuracy, completeness, or suitability
for any purpose of the Content. Any opinions and views expressed in this publication are the opinions
and views of the authors, and are not the views of or endorsed by Taylor & Francis. The accuracy of
the Content should not be relied upon and should be independently verified with primary sources of
information. Taylor and Francis shall not be liable for any losses, actions, claims, proceedings, demands,
costs, expenses, damages, and other liabilities whatsoever or howsoever caused arising directly or
indirectly in connection with, in relation to or arising out of the use of the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or
systematic reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in
any form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions
 Intern J Computer Math 1979, Section B, Vol 7, pp 157-161
0 Gordon and Bredih Sc~encePubl~rhersLtd, 1979
Prmted ~n Great Brltarn

A Sixth-Order Family of Methods

for Nonlinear Equations
BENY NETA
Northern Illinois University, Department of Mathematical Sciences,
DeKalb, Illinois 601 75, U.S.A.

A one-parameter family of slxth-order methods for finding slmple zeros of nonilnear

functions 1s developed. Each member of the famdy requires three evaluations of the glven
function and only one evaluation of the derlvatlve per step

1. INTRODUCTION
Newton's method for computing a simple zero 5 of a nonlinear equation
f(x)=O has been modified in a number of ways. For example, Ostrowski
[I] discusses a third-order method that evaluates the function f at every
substep but only requires the derivative f ' at every other substep. He also
Downloaded by [New York University] at 04:30 03 October 2014

introduced a fourth-order scheme that uses the same information. King

[2] has shown that there is a family of such methods. Traub [3]
introduced a third-order method which requires one function and two
derivative evaluation per step. Jarratt [4] developed fourth-order method
which uses the same information. King [5] developed a fifth-order scheme
that requires two evaluations o f f and f '.
Here we develop methods of order six. An iteration consists of one
Newton substep followed by two substeps of "modified Newton (i.e.,
using the derivative o f f at the first substep instead of the current one).
Let us recall the definition of order (see e.g. [3]).

DEFINITIONLet x,, x,, ...,xi, be a sequence converging to 5. Let

If there exists a real number p and a nonzero constant C such that

(

then p is called the'order of the sequence.

157
 158 B. N E T A

2. DEVELOPMENT O F THE SIXTH-ORDER FAMILY

Development of the sixth-order family.
Let

where A, B, C, D, G, H are arbitrary constants. This is a family of

methods which uses Newton's method in the first substep and two
Newton-like in the other two substeps. In each step we have to evaluate
the function f ( x ) at the three points x,, o n , z, and to evaluate the
derivative at one point x,.
In order to find the order of the method we used MACSYMA (Project
MAC'S SYmbolic MAnipulation system, which is a large computer
prbgramming system written in LISP and used for performing symbolic as
well as numerical mathematical manipulations [ 6 ] ) .
Downloaded by [New York University] at 04:30 03 October 2014

The error expression at w, is given by

where

and

For later use we also give the expression for

f (on)
= { 3 F 2 ~ ,+b(F3FI
2 - 3F:)c; +&3F4 +15Fi)~;
- 14F2F3

+&(4F5 - 25F2F4 - 20F; + 120F;F3 - 90F:)e;

 A SIXTH-ORDER METHOD FOR NONLINEAR EQUATIONS 159
+&(SF, - 39F5F2- 85F4F3 + 2 5 5 ~ , F ; + 370F:F2
- 1095F3F: f630FZ)s: + . . .)f '(5) (7)
and

f ' ( x n ) [ f ( x n+) B f ( u n ) l = f f 2 ( 0f{3&

( B f, 3 ) F z ~ ;
4)F3 + 3~,2]s,3
+&[(2~+
+&[(3B+ 5)F4+ 10F2F3+ ~ B F : ] E , ~
+&[(4B+6)F5 + 15F2F4+ 1 0 ~ +: ~ o B F ; F ~
-15B~;ls:

+*[(5B + 7)F6+21F5F2 + 35F,F3 +~~BF,F;

+40BF2F: - 150BF3F: + ~ O B F ~ ] E , ...
~ + (8)

The error expression at the point z, is given by

+(3~-21)~+21A-27)F:]e,4+ , & [ ( l 8 ~ - 1 8 A + 3 0 ) F 2 ~ ,
Downloaded by [New York University] at 04:30 03 October 2014

+ (16B - 16A + 24)F: + ( - 36B2 + ( 3 6 A - 228)B + 228A

-264)F;F3 + ( 9 + 9~0 ~ ~' - 9 A B 2 - 9 0 A B +297B-297A

+ 2 7 0 ) F ~ ] $+&[(48B -48A + 78)F2F5+ (120B - 12A

+ 170)F3F4+ ( - 135B+ 135AB -825B +825A-930)FiF4

+ ( - 2408' + 240AB - 1360B + 1360A - 1 4 0 0 ) ~ +~ (~2: 4 0 ~ ~
-2 4 0 ~ + ~ 2' 2 2 0 ~ ' 2220AB + 6750B - 6750A + 5640)F:F3
-

+ ( - 4 5 B 4 + 4 5 ~ -B 585B3
~ + 585AB2 - 2 7 9 0 ~ ' + 2 7 9 0 A B
- 5805B + 5805A - 3960)FZ]s,6+ ...) (9)

In order to annihilate the coefficient of E: in the expression for ~ ( z , we

)
have chosen B = A - 2. We have also chosen c = - 1, G = - 3 and H = D so
that the low-order terms in the expression for e n + , will vanish. This choice
leads to the family
 B. NETA

with error term

en+,=&[2F:F, - 3(2A + l)F;F,]e,h + .. .. (11)

Note that the error term does not depend on D so we can let D = 0.
If the parameter A is chosen so that A = -4 then the F i F , term in error
is eliminated so that
Downloaded by [New York University] at 04:30 03 October 2014

and

3. N U M E R I C A L EXAMPLES
Computer tests for the functions f ' =x n - 1 with the root 4 = 1 for various
values of the integer n show that the method is indeed of order 6.

Kemnrk Our .method needs the same number of function evaluations

as a method constructed from one substep of Newton's method followed
by two substeps of the secant method. The order of the secant method is
approximately 1.62 (see e.g. [ 7 ] ) thus such a scheme will approximately be
of order 5.2 (see e.g. [ 4 ] ) .

References
[I] A M Ostrowskl, Solullon of Equutrons and Sqstems of Equuttorzs, Acadern~cPress, New
York, 1960
[2] R~chardF Kmg, A farn~lyof fourth-order methods for nonlmear equations, subrn~tted
to Slum J Numer Anul
 A SIXTH-ORDER METHOD FOR NONLINEAR EQUATIONS 161
[3] J F Traub, 1terutrt.e Method3 for the Solutlon of Equatlon~,Prentlce-Hall, Englewood
Cliffs, New Jersey, 1964
[4] P Jarratt, Some effic~entfourth-order multlpo~ntmethods for solvmg equations, BIT 9
(1969), 119-124
[ 5 ] R~chardF Klng, A fifth-order family of modlfied Newton methods, BIT 11 (1971), 409-
412
[6] R Bogen et a l , MACSYMA Rejerence Manual, MIT, Cambridge, Mass, 1975
[7] D M Young and R T Gregory, A Survej of Numerical Marhemat~cs,Addlson-Wesley,
Readmg, Mass , 1973
Downloaded by [New York University] at 04:30 03 October 2014