Professional Documents
Culture Documents
1 Introduction
In this paper, we consider iterative methods to find a simple root x* of a nonlinear
⊆
equation f(x) = 0, where f: D R→R for an open interval D is a scalar function and it
is sufficiently smooth in a neighborhood of x*.
It is well known that classical Newton's method (NM) is a basic and important
method for solving non-linear equation [1] by the iterative scheme
f(xn)
xn+1=xn- , (1)
f'(xn)
which is quadratically convergent in the neighborhood of x*. In recent years, much
attention has been given to develop iterative methods for solving nonlinear equations
and many iterative methods have been developed [2-14].
Motivated and inspired by the on going activities in this direction, in this paper, we
present a sixth-order convergent method. The method is free from second derivatives
and permits f'(x)=0 in some points. It requires three evaluations of the given function
and two evaluations of its derivative in each step. Several examples are given to
illustrate the efficiency and advantage of the method.
S. Lin and X. Huang (Eds.): CSEE 2011, Part IV, CCIS 217, pp. 470–473, 2011.
© Springer-Verlag Berlin Heidelberg 2011
A Modified Newton-Type Method with Sixth-Order Convergence 471
f(xn)
yn=xn- , (2)
λnf(xn)+f'(xn)
f(yn)
zn=yn- , (3)
μnf(yn)+f'(xn)
f(zn)
xn+1=zn- , (4)
νnf(zn)+f'(zn)
where λn are real numbers chosen such that 0≤|λn|,|μn|,|νn|≤1, n=1,2,… and
sign(λnf(xn))=sign(f'(xn)), sign(μnf(yn))=sign(f'(xn)), sign(νnf(zn))=sign(f'(zn)), where
sign(x) is the sign function.
For Algorithm 1, we have the following convergence result.
⊆
Theorem 1. Assume that the function f: D R→R has a single root x* D, where D ⊆
is an open interval. If f(x) has first, second and third derivatives in the interval D, then
Algorithm 1 is fifth-order convergent in a neighborhood of x* and it satisfies error
equation
en+1=4c22(λn+c2)2(νn+c2)en6+O(en7), (5)
(k)
f (x*)
where en=xn-x*, ck= ,k=2,3,....
k!f'(x*)
f(k)(x*)
Proof. Let x* be the simple root of f(x), ck= ,k=2,3,… and en=xn-x*. Consider
k!f'(x*)
the iteration function F(x) defined by
f(z(x))
F(x) =z(x)- , (6)
νnf(z(x))+f'(z(x))
f(y(x)) f(x)
where z(x)=y(x)- , y(x)=x- .
μnf(y(x))+f'(x) λnf(x)+f'(x)
By some computations using Maple we can obtain F(x*)=x*, F(i)(x*)=0, i=1,2,3,4,5,
and
90f''(x*)2[f''(x*)+2λnf'(x*)]2[f''(x*)+2νnf'(x*)]
F(6)(x*)= . (7)
f'(x*)5
Furthermore, from the Taylor expansion of F(xn) around x*, we get
F''(x*) F'''(x*)
xn+1=F(xn) =F(x*)+F’(x*)(xn-x*)+ (xn-x*)2 + (xn-x*)3
2! 3!
(8)
F(4)(x*) F(5)(x*) F(6)(x*)
+ (xn-x*)4 + (xn-x*)5 + (xn-x*)6 +O((xn-x*)7).
4! 5! 6!
Substituting (7) into (8) yields
xn+1=x*+ en+1=4c22(λn+c2)2(νn+c2)en6+O(en7). (9)
Therefore, we have
472 L. Sun and L. Fang
en+1=4c22(λn+c2)2(νn+c2)en6+O(en7), (10)
which means the order of convergence of the Algorithm 1 is six. The proof is
completed.
Now, we consider efficiency index defined as p1/w, where p is the order of the
method and w is the number of function evaluations per iteration required by the
method. It is not hard to see that the efficiency index of the Algorithm 1 is 1.431
which is better than that of Newton's method (NM) 1.414.
3 Numerical Results
Now, we employ Algorithm 1 presented in the paper to solve some nonlinear
equations and compare it with NM. Displayed in Table 1 are the number of iterations
(ITs) required such that |f(xn)|<1.E-14.
Functions xo NM Algorithm 1
f1 -1 Failure 6
f1 0 72 4
f2 0 Failure 5
f2 -1 8 3
f3 1 Failure 8
f3 1.01 28 11
The computational results in Table 1 show that Algorithm 1 require less ITs than
NM. Therefore, the present sixth-order convergent method is of practical interest and
can compete with NM.
4 Conclusion
In this paper, we present and analyze a sixth-order convergent method for solving
nonlinear equations under mild conditions. The method is free from second
derivatives and permits f'(x)=0 in some points. It requires three evaluations of the
given function and two evaluations of its derivative in each step. Several numerical
tests demonstrate that the sixth-order method given in this paper is more efficient and
performs better than classical Newton's method.
References
1. Traub, J.F.: Iterative Methods for Solution of Equations. Prentice-Hall, Englewood Cliffs
(1964)
2. Noor, M.A., Noor, K.I.: Fifth-order iterative methods for solving nonlinear equations.
Appl. Math. Comput. 188, 406–410 (2007)
3. Kou, J.: The improvements of modified Newton’s method. Appl. Math. Comput. 189,
602–609 (2007)
4. Fang, L., He, G., Hu, Z.: A cubically convergent Newton-type method under weak
conditions. J. Comput. Appl. Math. 220, 409–412 (2008)
5. Fang, L., He, G.: Some modifications of Newton’s method with higher-order convergence
for solving nonlinear equations. J. Comput. Appl. Math. 228, 296–303 (2009)
6. Fang, L., Sun, L., He, G.: An efficient Newton-type method with fifth-order for solving
nonlinear equations. Comput. Appl. Math. 27, 269–274 (2008)
7. Muhammad, A.N., Faizan, A.: Fourth-order convergent iterative method for nonlinear
equation. Appl. Math. Comput. 182, 1149–1153 (2006)
8. Chun, C.: A simply constructed third-order modifications of Newton’s method. J. Comput.
Appl. Math. 219, 81–89 (2008)
9. Gautschi, W.: Numerical Analysis: An Introduction. Birkhauser Boston Inc., Cambridge
(1997)
10. Potra, F.A., Potra-Pták: Nondiscrete induction and iterative processes. Research Notes in
Mathematics 103 (1984)
11. Muhammad, A.N., Khalida, I.N.: Modified iterative methods with cubic convergence for
solving nonlinear equations. Appl. Math. Comput. 184, 322–325 (2007)
12. Fang, L., He, G., Hu, Y., Sun, L.: Some modified Newton-type methods with order of
convergence varied from two to six under weak conditions, pp. 637–640. IEEE CSO, Los
Alamitos (2009)
13. Grau, M., Diaz-Barrero, J.L.: An improvement to Ostrowski root-finding method. Appl.
Math. Comput. 173, 450–456 (2006)
14. Chun, C.: Some second-derivative-free variants of super-Halley method with fourth-order
convergence. Appl. Math. Comput. 192, 537–541 (2007)
15. Grau, M., Noguera, M.: A Variant of Cauchy’s Method with Accelerated Fifth-Order
Convergence. Applied Mathematics Letters 17, 509–517 (2004)
16. Sharma, J.R.: A composite third-order Newton-Steffensen method for solving nonlinear
equations. Appl. Math. Comput. 169, 242–246 (2005)
17. Homeier, H.H.H.: A modified Newton method for rootfinding with cubic convergence. J.
Comput. Appl. Math. 157, 227–230 (2003)