University of Palestine

Gaza Strip

Civil Engineering College

Numerical analysis
CIVL 3309
Formerly CVL 3308

Dr. Suhail Lubbad

Numerical analysis
CVL 3308
Lecture Ch6.2

2
The Newton-Raphson open method

3
Start:
Initial guess at the root is xi
Draw the tangent at the point ( xi, f(xi) ),
The intersection of this tangent line with the x-axis is
the improved estimate of the root.

4
 Derivation and geometrical interpretation

f (xi )−f (x i '1 ) f (x i )−0

f '(x i )= =
*x x i −x i'1

f (x i )
x i'1=x i− Newton Raphson Formula
f ' (x i ) 5
What about error computation?

6
From
Taylor-expansion-based derivation of Newton-Raphson Formula
Box 6.2

f(xi+1 )= f(xi )+ f′(xi )(xi+1− xi ) + f′′(xi)/2! ( xi+1− xi )2+ ....

Shows a quadratic convergence for the true error

−f ' '(x r ) 2
Et , i'1= Et , i
2f '(x r )

Correct decimal places doubles within each iteration

Check Example 6.4

7
True root value: 0.56714329

8
Example 6.5:
Determine the positive root of f(x)= x10 – 1 using Newton-Raphson
method and an initial guess of x=0.5

This example illustrates troubled cases with Newton-Raphson

Formula

Troubled cases
A special case: multiple roots
A special case: When f'' =0 near the root
A special case: f' =0 near the root

9
Iteration 1
The new approximate Root 1=5.0000000e-01−(−9.9902344e-01)/(1.9531250e-02)=5.1650000e+01

∣
The True Error E t =
1
TrueRoot∣
TrueRoot−Root 1
∣
×100 %=
1.0000000e+00−(5.1650000e+01)
1.0000000e+00 ∣×100 %=5.0650000e+03 %

Iteration 2
The new approximate Root 2=5.1650000e+01−(1.3511490e+17)/(2.6159710e+16)=4.6485000e+01

∣
The True Error E t =
2
TrueRoot∣
TrueRoot−Root 2
∣
×100 %=
1.0000000e+00 ∣
1.0000000e+00−(4.6485000e+01)
×100 %=4.5485000e+03 %

The Approximate Error E a =

2 ∣ ∣
Root 2−Root 1
Root 2 ∣
×100 %=
4.6485000e+01 ∣
4.6485000e+01−(5.1650000e+01)
×100 %=1.1111111e+01%

Iteration 3
The new approximate Root 3=4.6485000e+01−(4.7111654e+16)/(1.0134808e+16)=4.1836500e+01

∣
The True Error E t =
3
TrueRoot∣
TrueRoot−Root 3
∣
×100 %=
1.0000000e+00 ∣
1.0000000e+00−(4.1836500e+01)
×100 %=4.0836500e+03 %

The Approximate Error E a =

3 ∣ ∣
Root 3 −Root 2
Root 3 ∣
×100 %=
4.1836500e+01 ∣
4.1836500e+01−(4.6485000e+01)
×100 %=1.1111111e+01%

Iteration 4
The new approximate Root 4 =4.1836500e+01−(1.6426818e+16 )/(3.9264322e+15)=3.7652850e+01

∣
The True Error E t =
4
TrueRoot∣
TrueRoot−Root 4
∣
×100 %=
1.0000000e+00 ∣
1.0000000e+00−(3.7652850e+01)
×100 %=3.6652850e+03 %

The Approximate Error E a =

4 ∣ ∣
Root 4 −Root 3
Root 4 ∣
×100 %=
3.7652850e+01 ∣
3.7652850e+01−(4.1836500e+01)
×100 %=1.1111111e+01%

Iteration 5
The new approximate Root 5=3.7652850e+01−(5.7276773e+15)/(1.5211803e+15)=3.3887565e+01

∣
The True Error E t =
5
TrueRoot∣
TrueRoot−Root 5
∣
×100 %=
1.0000000e+00 ∣
1.0000000e+00−(3.3887565e+01)
×100 %=3.2887565e+03 %

The Approximate Error E a =

5 ∣ ∣
Root 5 −Root 4
Root 5 ∣
×100 %=
3.3887565e+01 ∣
3.3887565e+01−(3.7652850e+01)
×100 %=1.1111111e+01% 10
Iteration 6
The new approximate Root 6=3.3887565e+01−(1.9971176e+15)/(5.8933641e+14 )=3.0498809e+01

The True Error E t =

6 ∣ TrueRoot∣
TrueRoot−Root 6
×100 %= ∣ 1.0000000e+00 ∣
1.0000000e+00−(3.0498809e+01)
×100 %=2.9498809e+03 %

The Approximate Error E a =

6 ∣ Root 6 ∣
Root 6−Root 5
×100 %= ∣
3.0498809e+01−(3.3887565e+01)
3.0498809e+01 ∣
×100 %=1.1111111e+01%

Iteration 7
The new approximate Root 7=3.0498809e+01−(6.9635184e+14)/(2.2832100e+14)=2.7448928e+01

The True Error E t =

7 ∣ TrueRoot∣
TrueRoot−Root 7
×100 %= ∣ 1.0000000e+00 ∣
1.0000000e+00−(2.7448928e+01)
×100 %=2.6448928e+03 %

The Approximate Error E a =

7 ∣ Root 7 ∣
Root 7−Root 6
×100 %= ∣
2.7448928e+01−(3.0498809e+01)
2.7448928e+01 ∣
×100 %=1.1111111e+01%

Iteration 8
The new approximate Root 8=2.7448928e+01−(2.4280288e+14 )/(8.8456233e+13)=2.4704035e+01

The True Error E t =

8 ∣ TrueRoot∣
TrueRoot−Root 8
×100 %= ∣ 1.0000000e+00 ∣
1.0000000e+00−(2.4704035e+01)
×100 %=2.3704035e+03 %

The Approximate Error E a =

8 ∣ Root 8 ∣
Root 8 −Root 7
×100 %= ∣
2.4704035e+01−(2.7448928e+01)
2.4704035e+01 ∣×100 %=1.1111111e+01%

Iteration 9
The new approximate Root 9=2.4704035e+01−(8.4660128e+13)/(3.4269757e+13)=2.2233631e+01

The True Error E t =

9 ∣ TrueRoot∣
TrueRoot−Root 9
×100 %= ∣ 1.0000000e+00 ∣
1.0000000e+00−(2.2233631e+01)
×100 %=2.1233631e+03 %

The Approximate Error E a =

9 ∣ Root 9 ∣
Root 9−Root 8
×100 %= ∣
2.2233631e+01−(2.4704035e+01)
2.2233631e+01 ∣
×100 %=1.1111111e+01%

Iteration 10
The new approximate Root 10=2.2233631e+01−(2.9519161e+13)/(1.3276806e+13)=2.0010268e+01

The True Error E t =

10 ∣TrueRoot−Root 10
TrueRoot ∣ ×100 %= ∣
1.0000000e+00−(2.0010268e+01)
1.0000000e+00 ∣×100 %=1.9010268e+03 %
11
The Approximate Error E a =
10 ∣
Root 10 −Root 9
Root 10 ∣
×100 %= ∣
2.0010268e+01−(2.2233631e+01)
2.0010268e+01 ∣
×100 %=1.1111111e+01%
The Newton−Raphson Formula
f (x i )
xi'1=x i−
f ' ( xi )
The function in hand is: f (x )=x10 −1
9
The first derivative is: f ' ( x )=x
f ( xi ) X 10 −1
Calculating : x i'1=x i− =x i− 9
f ' ( xi) 10x
Iteration xr
0
1 0.5
2 51. ? ? ?
3 46. ? ? ?
4 .
. .
. .
. .
. .
∞ 1.00000 Clearly Slow convergence 12
Troubled cases
A special case: multiple roots
A special case: When f'' =0 near the root
A special case: f' =0 near the root

13
(a) Troubles with an inflection point [that is, f ′′(x) = 0] in the vicinity of a root. iterations
beginning at x0 progressively diverge from the root.
14

(b) Newton- Raphson technique oscillates around a local maximum or minimum

(c) Initial guess close to one root can jump to a location several roots away. This t is
because of the near-zero slopes are encountered
15
(d) The solution shoots off horizontally and never hits the x axis.
Algorithm for Newton-Raphson
Must be improved

16
BONUS CREDITS

Volunteer wanted for a little research

Check Page 164 in your textbook

The Modified Newton-Raphson Method mainly for multiple

roots function
'
f ( xi ) f ( xi )
x i '1 =x i− ' 2 ''
f ( x i ) −f ( x i ) f ( x i )

17