0% found this document useful (0 votes)
15 views3 pages

Rate of Convergence: ST ND RD TH

Uploaded by

nazrin sultana
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
15 views3 pages

Rate of Convergence: ST ND RD TH

Uploaded by

nazrin sultana
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

L10: Rate of Convergence

Rate of Convergence

The general technique, to find the real root of a numerical equation, using
any iterative method, is that it begins with an initial approximate value and
then find the better approximations by repeating the same method.

If, at each step, the successive approximations approach the root more and
more closely, then the method is said to be convergent.

Let us suppose 𝑥1 , 𝑥2 , 𝑥3 ……. 𝑥𝑛 be the approximations of 1st, 2nd, 3rd, …….,


nth iterations, and 𝜀1 , 𝜀2 , 𝜀3 , ……, 𝜀𝑛 be their corresponding errors.

Therefore, for an iterative method to be convergent, it can be written as

|𝜀𝑛 | ≤ |𝜀𝑛−1 | … .. . (1)

Now, p will be the rate of convergence of the iterative method if p is the


largest positive real number for which
|𝜀𝑛 | = 𝐶|𝜀𝑛−1 |𝑝 … .. . (2)

where, C is a constant and 𝐶 ≠ 0. p is also known as the order of the


iterative method.

1
L10: Rate of Convergence

Rate of Convergence of Newton-Raphson Method

We have a general form of Newton-Raphson method as


f ( xn )
xn 1  xn  , n = 0, 1, 2, …… ……….. (3)
f / ( xn )
We now want to find the rate of convergence of this method. Suppose, a is
a root of an equation f ( x)  0 . Therefore, xn  a   n . Putting this relation in the
above equation, we obtain
f (a   n )
 n1   n  ………… (4)
f / (a   n )
Now, expanding f (a   n ) and f / (a   n ) by Taylor’s series about the point a, we
obtain after simplification as
1 f // (a) 2
 n1   n  ............
2 f / (a)

or,  n1  C n2 ………. (5)


1 f // (a)
Where, C  and higher powers of  n are neglected.
2 f / (a)

According to Eq. (5), Newton-Raphson method has second order


convergence.

Acceleration of Convergence: Aitkin’s ∆𝟐 Methodti

The linear convergence of an iterative method can be accelerated. For


example, equation (i) can also be written as

|𝑎 − 𝑥𝑛 | ≤ |𝑎 − 𝑥𝑛−1 |
or, |𝑎 − 𝑥𝑛 | ≅ 𝐾|𝑎 − 𝑥𝑛−1 |
Now, if xi 1 , x i and xi 1 be the three successive approximations. Then we can
write
a  xi 1  K (a  xi )
and a  xi  K (a  xi 1 )

2
L10: Rate of Convergence
Dividing, we get
a  xi a  xi 1

a  xi 1 a  xi

Simplifying the above relation it can be shown that-


( xi 1  xi ) 2
a  xi 1  ………….. (6)
xi 1  2 xi  xi 1

This gives a new approximate value that can be termed as the next
approximation by
(𝑥𝑖+1 −𝑥𝑖 )2
𝑥𝑖+2 = 𝑥𝑖+1 − … .. . (7)
𝑥𝑖+1 −2𝑥𝑖 +𝑥𝑖−2

Therefore, the rate of convergence can be accelerated by putting three


previously known successive approximations in Eq. (7).

Now, in the sequence of successive approximations, we have


𝑥𝑖+1 − 𝑥𝑖 = ∆𝑥𝑖
and 𝑥𝑖 − 𝑥𝑖−1 = ∆𝑥𝑖−1
 2 xi 1  (xi 1 )  ( xi  xi 1 )  xi  ( xi 1 )  ( xi 1  xi )  ( xi  xi 1 )  xi 1  2 xi  xi 1

Hence, equation (7) can be written as


(xi ) 2
xi  2  xi 1  … .. . (8)
2 ( xi 1 )
This modified iterative-scheme is known as Aitken’s 2 -method or Aitken’s
extrapolation method.

Note: In any numerical application, the values of xi and xi 1 can be
obtained by the following way

xi 1
xi 1
xi 2 xi 1
xi
xi 1

You might also like