Newton-Raphson Method
Newton-Raphson Method
Newton-Raphson Method
Mircea Cirnu1
Irina Badralexi 2
Abstract
First of the authors of present paper obtained higher order polynomial equations, which
for orders 2 and 3 are reduced to equations given by He and respectively by Wei-Wu-Mei,
with much improved form.
1. Newton-Raphson method
f (x) = 0 ,
the approximations x n of an exact real root x of the equation has the following from:
f ( xn )
x n +1 = x n − , n = 0,1,2,....
f ' ( xn )
2. He’s method
Using second order Taylor’s expansion, He [3] developed a faster convergent iteration
method, obtaining for the variation t n = x n +1 − x n , the second order polynomial equation
1
Mircea Cirnu, Phd, professor, Faculty Of Applied Sciences,
University “Politehnica” Of Bucharest
2
Irina Badralexi, Student, Faculty Of Applied Sciences,
University “Politehnica” Of Bucharest
lkjlklk
where x 0 and x1 , hence t 0 = x1 − x0 are given and
1
g ( x n ) = f ( x n ) − f ( x n −1 ) − f ' ( x n −1 ) t n −1 − f ' ' ( x n −1 ) t n2−1 .
2
He’s method is indeed faster convergent than Newton’s method, but it does not have
solutions for all initial values, for the following condition must be fulfilled at every step:
B 2 − 4 A(C + g ( x n )) ≥ 0 ,
1
A = 2 f ' ' ( xn )
where B = f ' ( x n ) − f ' ' ( x n ) x n
1
C = f ( x n ) − f ' ( x n ) x n + f ' ' ( x n ) x n 2
2
Following He’s example, Wei, Wu and Mei, [4], proposed an even more quickly
convergent method under the form of a third order polynomial equation:
where
1 1
g ( x n ) = f ( x n ) − f ( x n −1 ) − f ' ( x n −1 ) t n −1 − f ' ' ( x n −1 ) t n2−1 + f ' ' ' ( x n −1 ) t n3−1 .
2 3!
Being a cubic equation it will have at least one real solution for any initial values, thus
being more convenient than He’s method.
f (k ) ( x1 ) k
∑
m
k =0
t1 = 0 ,
k!
f (k ) (x 2 ) k
∑ t 2 + 2 f (x2 ) = 0 ,
m
k =1
k!
where x 2 = x1 + t1 ,
………………………………………………………………………………
f (k ) (x n ) k
∑ t n + 2 f ( x n ) + ∑ j = 2 f (x j ) = 0 , n = 3,4, ,
m n −1
k =1
k!
where x j = x j −1 + t j −1 , 3 ≤ j ≤ n , n = 3,4, .
f ′′( x1 ) 2
t1 + f ′( x1 )t1 + f ( x1 ) = 0 ,
2
f ′′( x 2 ) 2
t 2 + f ′( x 2 )t 2 + 2 f ( x 2 ) = 0 ,
2
where x 2 = x1 + t1 ,
………………………………………………………………………….
f ′′( x n ) 2
t n + f ′( x n )t n + 2 f (x n ) + ∑ j = 2 f (x j ) = 0 ,
n −1
where x j = x j −1 + t j −1 , 3 ≤ j ≤ n , n = 3,4, .
For m = 3 , are obtained the improved Wei-Wu-Mei equations
f ′′′( x1 ) 3 f ′′( x1 ) 2
t1 + t1 + f ′( x1 )t1 + f ( x1 ) = 0 ,
6 2
f ′′′( x 2 ) 3 f ′′( x 2 ) 2
t2 + t 2 + f ′(x 2 )t 2 + 2 f (x 2 ) = 0 ,
6 2
where x 2 = x1 + t1 ,
………………………………………………………………………….
f ′′′( x n ) 3 f ′′( x n ) 2
t n + f ′( x n )t n + 2 f (x n ) + ∑ j = 2 f (x j ) = 0 ,
n −1
tn +
6! 2
where x j = x j −1 + t j −1 , 3 ≤ j ≤ n , n = 3,4, .
The improvement of these equations consists in replacing of g ( x n ) from constant
term with simpler expressions.
5. Numerical example
We give an example, taken from [6], in which He’s method does not apply.
Consider equation f ( x ) = x 3 − e − x = 0 . Newton’s formula (11) gives recurrence
x n3 − e − xn
relation x n +1 = x n − 2 . Taking x0 = 0 , we obtain x1 = 1 , x 2 = 0.8123 ,
3 x n + e − xn
x3 = 0.7743 and x 4 = 0.7729 .
For m = 2 , taking x0 = x1 = 0 , He’s method give quadratic equation
t − 2 t1 + 2 = 0 , which has complex roots, therefore this method is not applicable.
2
1
References.