Unit 3.4 Numerical Solution of Equations
Unit 3.4 Numerical Solution of Equations
Specific Objectives:
1. To learn fixed point iteration method, Newton’s method, Secant method and method of false position.
2. To acquire the skill in using the relevant methods to find approximate roots of equations, and compute the errors of the roots.
In scientific and engineering work, a frequently occurring problem is to find the roots
of equations of the form f ( x ) = 0 . For quadratic, cubic or quartic equations, exact
solutions are readily obtained using algebraic methods. However, when f ( x ) is a
polynomial of higher degree than four or a transcendental function such as
e x − 4 cos x , algebraic methods are not easily available. It is natural to find the solution
by approximate methods.
17.1 Method of Fixed-point 7 Students should be taught that the method consists of several steps.
119
xi = g ( xi −1)
xi +1 = g ( xi )
convergence 5
teachers may ask students to find the roots using initial guesses x0 = 2 and 5
respectively. Students would find that for x0 = 2 , they get the root near 1 but for
x0 = 5 , the procedure diverges. The discussion of convergence then becomes natural.
It is worthwhile to geometrically discuss monotonic and oscillating convergence and
divergence. Mathematical treatment of monotonic and oscillating convergence as well
as divergence should be briefly mentioned.
To discuss convergence mathematically, the use of mean value theorem is essential. It
is advantageous for students to know that for the fixed-point iteration algorithm to be
useful the following are usually needed.
1. There is an interval I = [a, b] such that for all x in I, g ( x ) is defined and g ( x ) in I.
2. The iteration function g ( x ) is continuous on I.
3. The iteration function is differentiable on I and for all x in I, there exists a real number
K such that g ′( x ) ≤ K < 1 .
Exercises on proofs like the following can be given.
Detailed Content Time Ratio Notes on Teaching
Example
It is given that x = g ( x ) has exactly one root α in [a, b], and that g ′( x ) ≤ K for any x
in [a, b]. Suppose further that xn +1 = g ( xn ) where a ≤ xn ≤ b , n = 1, 2, 3, ....
1− K
can be derived and discussed. It may be seen that the smaller the value of K, the faster
the rate of convergence. For abler students, teachers may discuss the order of
convergence using Taylor's series expansion of the error about the fixed point α
εn 2 ε 3
εn +1 = g ′(α )εn + g ′′(α )
+ g ′′′(α ) n +
2! 3!
121
Alternatively, it may also be derived using Taylor’s series expansion of f ( xn +1) about
x0 as
( xn +1 − xn )2
f ( xn +1) = f ( xn ) + f ′( xn )( xn +1 − xn ) + f ′′( xn ) +
2!
(b) The condition of As such, the condition of convergence is the same as the general fixed-point
convergence and error iteration method. It is interesting to discuss the rate of convergence of Newton’s method
estimation as compared with that of the general method. Students could easily find that
f ( x )f ′′( x ) εn 2
g ′( x ) = and g′(α ) = 0 if α is a simple root, and that εn +1 ≈ g ′′(α ) .
[f ′( x )]2 2!
Examples in which Newton’s method excels others should be given. An example follows.
Example
122
(c) Application of Newton’s Problems on polynomial equations of degree higher than two and transcendental
method equations are relevant. For example, finding the root between x = 0 and x = 1 of the
equation 2 x 3 + x 2 − 20 x + 20 = 0 with an accuracy of 10− , and deriving Newton’s
6
xk − a
formula xn +1 = xn − for finding the kth root of a are common questions.
kxnk −1
Detailed Content Time Ratio Notes on Teaching
17.3 Secant Method
The secant method is another method for finding the roots of f ( x ) = 0 . It often
converges almost as fast as Newton’s method, but avoids the need for calculating the
derivative f ′( x ) . Instead of using a tangent line, a secant line is used.
(a) Derivation of the secant The derivation of the algorithm can be easily done by appealing to the geometry of
method the method in a similar way to that of Newton’s method.
123
xn − xn −1
The analogous formula is xn +1 = xn − f ( xn ) with the slope of the
f ( xn ) − f ( xn −1)
f ( xn ) − f ( xn −1)
secant replacing the slope of the tangent, f ′( x ) .
xn − xn −1
Students should be able to see that the secant method requires two initial guesses
at which the functional values need not be of different signs.
In-depth treatment of rate of convergence and error estimation is not needed, When
such is required for calculation, the relevant formulae will be given in the questions.
17.4 Method of False Position 2 This is still another method for finding a root of the equation f ( x ) = 0 lying in the
interval [a, b]. The method is similar to the Bisection method (which students have
learned in S.5) in that intervals [an, bn] are generated to bracket the root, and the method
is also similar to the Secant method in the manner of obtaining new approximate
iterates. The method is also named Regula Falsi.
(a) Derivation of the method Assuming that the interval [an, bn] contains a root of f(x) = 0 and with the help of a
of false position diagram like the following
124
teachers can guide students to compute the value of the x-intercept of the line joining the
points ( an , f (an ) ) and ( bn , f ( bn ) ) . This point labelled xn +1 will be found to be
f (an )(bn − an ) f (bn )(bn − an )
xn +1 = an − or xn +1 = bn − . Students should find it easy to
f (bn ) − f (an ) f (bn ) − f (an )
complete the algorithm by defining an +1 = an and bn +1 = xn +1 if f ( xn +1)f (an ) < 0 and
defining an +1 = xn +1 and bn +1 = bn if f ( xn +1)f (an ) > 0 .
Detailed Content Time Ratio Notes on Teaching
(b) Application of the Students should be reminded that two initial guesses are required by the method
method of false position and that the functional values at these initial guesses must be of different signs for the
method to be applicable. Exercises similar to those for Newton’s method and Secant
method are appropriate. Following are some examples.
Example 1
Show that the equation x 3 + 3 x − 12 = 0 has exactly one root in the interval [1, 2] and
find this root by the method of false position correct to 3 decimal places.
Example 2
The method of false position is used to find the root of x 3 = 2 x + 5 in the interval [2, 3].
Show that the sequence of iterates { xi } is given by
16 xi − 48
xi +1 = 3 − i = 1, 2, 3, ……
x i 3 − 2 x i − 21
125