Ch2 Lecture3
Ch2 Lecture3
Chapter 2
Lecture #3
y=f(x)
0
x3 x2 x x
1
(x − xn )2 00
f (x) = f (xn ) + (x − xn )f 0 (xn ) + f (η(x)), (7)
2
where η(x) lies between x and xn . Since f (α) = 0, then (7), with x = α, gives
(α − xn )2 00
f (α) = 0 = f (xn ) + (α − xn )f 0 (xn ) + f (η(α)).
2
Since |α − xn | is small, then we neglect the term involving (α − xn )2 and so
0 ≈ f (xn ) + (α − xn )f 0 (xn ).
f (xn)
xn+1 = xn − ________ f 0(xn) 6= 0, for all n ≥ 0. (9)
,
f 0(xn)
The iterative method (9) is called the Newton’s method.
Example 0.1
Use Newton’s method to find the approximation x3 to the root of
cos x − x = 0,
n xn f (xn )
1 0.7853981635 -0.078291381
2 0.7395361337 -0.000754873
3 0.7390851781 -0.000000074
Therefore, x3 = 0.7390851781.
Example 0.1
Use Newton’s method to find the approximation x3 to the root of
cos x − x = 0,
n xn f (xn )
1 0.7853981635 -0.078291381
2 0.7395361337 -0.000754873
3 0.7390851781 -0.000000074
Therefore, x3 = 0.7390851781.
Example 0.2
Use the Newton’s method to find the root of x3 = 2x + 1 that is located in the
interval [1.5, 2.0] accurate to 10−2 , take an initial approximation x0 = 1.5.
Solution. Given f (x) = x3 − 2x − 1 and so f 0 (x) = 3x2 − 2. Now evaluating f (x)
and f 0 (x) at the give approximation x0 = 1.5, gives
f (x0 ) (−0.625)
x1 = x0 − = 1.5 − = 1.631579.
f 0 (x0 ) 4.75
Using the iterative formula (9) again to get other new approximation. The
successive iterates were shown in the Table 3.
f (x0 ) (−0.625)
x1 = x0 − = 1.5 − = 1.631579.
f 0 (x0 ) 4.75
Using the iterative formula (9) again to get other new approximation. The
successive iterates were shown in the Table 3.
To get the above results using MATLAB command, firstly the function
x3 − 2x − 1 and its derivative 3x2 − 2 were saved in m-files called fn.m and dfn.m,
respectively written
f unction y = fas follows:
n(x) f unction dy = df n(x)
y = x.ˆ 3 − 2 ∗ x − 1; dy = 3 ∗ x.ˆ 2 − 2;
after which we do the following:
>> x0 = 1.5; tol = 0.01; sol = newton(0 f n0 ,0 df n0 , x0, tol);
Example 0.3
If the difference of two numbers x and y is 6 and the square root of their product is 4,
then use Newton’s method to approximate, to within 10−4, Find the largest value of
the number x and the corresponding number y using initial approximation
x0 = 7.5.
Solution. Given
√
x − y = 6 and xy = 4.
Solving the above equations for x, we have
x(x − 6) = 16 or x2 − 6x − 16 = f (x) = 0.
x(x − 6) = 16 or x2 − 6x − 16 = f (x) = 0.
1
.
x
Thus
1
2 cos x = , gives, x cos x − 0.5 = 0,
x
and from this we have the function and its derivative as follows
xn cos xn − 0.5
xn+1 = xn − ,
cos xn − xn sin xn
and for finding the approximations, starting x0 = 8, we obtain, x1 = 7.7936 and
x2 = 7.7897. Taking x = 7.79, we have y = 2 sin 7.79 = 1.996. Therefore, the point
of contact is (7.79, 1.996). To find the value of k, we solve the equation,
1.996 = ln(7.79) + k, and it gives, k = −0.0568, the required value of k. •
Example 0.4
1
.
x
Thus
1
2 cos x = , gives, x cos x − 0.5 = 0,
x
and from this we have the function and its derivative as follows
xn cos xn − 0.5
xn+1 = xn − ,
cos xn − xn sin xn
and for finding the approximations, starting x0 = 8, we obtain, x1 = 7.7936 and
x2 = 7.7897. Taking x = 7.79, we have y = 2 sin 7.79 = 1.996. Therefore, the point
of contact is (7.79, 1.996). To find the value of k, we solve the equation,
1.996 = ln(7.79) + k, and it gives, k = −0.0568, the required value of k. •
Example 0.5
Develop an iterative procedure for evaluating the reciprocal of a positive number
N by using Newton’s method. Use the developed formula to find third
approximation to the reciprocal of 3, taking an initial approximation x0 = 0.4.
Compute absolute error.
Solution. Consider x = 1/N . This problem can be easily solved by noting that
we seek to find a root to the nonlinear equation
1/x − N = 0,
Hence, assuming an initial estimate to the root, say, x = x0 and by using iterative
formula (9), we get
(1/x0 − N )
x1 = x0 − = x0 + (1/x0 − N )x20 = x0 + x0 − N x20 = x0 (2 − N x0 ).
(−1/x20 )
In general, we have
xn+1 = xn (2 − N xn ), n = 0, 1, . . . , (10)
Example 0.5
Develop an iterative procedure for evaluating the reciprocal of a positive number
N by using Newton’s method. Use the developed formula to find third
approximation to the reciprocal of 3, taking an initial approximation x0 = 0.4.
Compute absolute error.
Solution. Consider x = 1/N . This problem can be easily solved by noting that
we seek to find a root to the nonlinear equation
1/x − N = 0,
(1/x0 − N )
x1 = x0 − = x0 + (1/x0 − N )x20 = x0 + x0 − N x20 = x0 (2 − N x0 ).
(−1/x20 )
In general, we have
xn+1 = xn (2 − N xn ), n = 0, 1, . . . , (10)
We have to find the approximation of the reciprocal of number N = 3. Given the
initial gauss of say x0 = 0.4, then by using the iterative formula (10), we get
x1 = 0.3200, x2 = 0.3328, x3 = 0.3333.
2 1
1.5
0.8 y=x
y = 1/x − 3
0.6
1 α
0.4
0.5
0.2
α
y
y
0 0
−0.2
−0.5 y = x(2 − 3x)
−0.4
−1
−0.6
−1.5
−0.8
−2 −1
0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
x x
After just three iterations the estimated value compares rather favorably with the
exact value of 1/3 ≈ 0.3333, (see Figure 7). Thus the absolute error is
1
|E| = − x3 = |0.3333 − 0.3333| = 0.0000.
3
We can calculate the other reciprocal of the number in the same way by using the
general iterative formula (10). •
Procedure
(Newton’s Method)
1. Find the initial approximation x0 for the root by sketching the graph of the
function.
2. Evaluate function f (x) and the derivative f 0 (x) at initial approximation.
Check: if f (x0 ) = 0 then x0 is the desire approximation to a root. But if
f 0 (x0 ) = 0, then go back to step 1 to choose new approximation.
3. Establish Tolerance ( > 0) value for the function.
4. Compute new approximation for the root by using the iterative formula (9).
5. Check Tolerance. If |f (xn )| ≤ , for n ≥ 0, then end; otherwise, go back to
step 4, and repeat the process.
Summary