Numerical Analysis Exercises

Abdelkader BENHARI
Numerical anlysis exercises
Numerical analysis is the area of mathematics and computer science that creates,
analyzes, and implements algorithms for solving numerically the problems of continuous
mathematics. Such problems originate generally from real-world applications of algebra,
geometry, and calculus, and they involve variables which vary continuously. These problems
occur throughout the natural sciences, social sciences, medicine, engineering, and business
Contents
Number Systems and Error
Solution of Nonlinear Equations
11
Systems of Linear Equations
23
Approximating Functions
84
Differentiation and Integration
96
Ordinary Differential Equations
130
Bibilography
154
Chapter 1
Number Systems and Error
1. Convert the following binary numbers to decimal form

(1010)2 , (100101)2 , (.1100011)2
Solution:
(1010)2 = (0 20 ) + (1 21 ) + (0 22 ) + (1 23 )
= 0 + 2 + 0 + 8 = (10)10
(100101)2 = (1 20 ) + (0 21 ) + (1 22 ) + (0 23 )
= (0 24 ) + (1 25 )
= 1 + 0 + 4 + 0 + 0 + 32 = (37)10
(.1100011)2 = (1 27 ) + (1 26 ) + (0 25 ) + (0 24 )
= (0 23 ) + (1 22 ) + (1 21 )
= 1/27 + 1/26 + 0 + 0 + 0 + 1/22 + 1/2 = (99/128)10
2. Convert the following binary numbers to decimal form.
(101101)2 , (10110)2 , (100111)2 , (10000001)2
Solution:
(101101)2 = (1 20 ) + (0 21 ) + (1 22 ) + (1 23 )
= (0 24 ) + (1 25 )
= 1 + 0 + 4 + 8 + 0 + 32 = (45)10
(101111)2 = (1 20 ) + (1 21 ) + (1 22 ) + (1 23 )
= (0 24 ) + (1 25 )
= 1 + 2 + 4 + 8 + 0 + 32 = (47)10
(100111)2 = (1 20 ) + (1 21 ) + (1 22 ) + (0 23 )
= (0 24 ) + (1 25 )
= 1 + 2 + 4 + 0 + 0 + 32 = (39)10
(10000001)2 = (1 20 ) + (0 21 ) + (0 22 ) + (0 23 )
= (0 24 ) + (0 25 ) + (0 26 ) + (1 27 )
= 1 + 0 + 0 + 0 + 0 + 0 + 0 + 128 = (129)
A.BENHARI
3. Find the first five binary digits of (0.1)10 . Obtain values for the absolute and
relative errors in yours results.
Solution: Apply the conversion procedure as follows:
0.1
2
0.2 + integer part 0 (d1 )
2
2
2
2
Thus
(0.1)10 = (.00011)2
4. Convert the following:
(a) decimal numbers to binary numbers form.
165, 3433, 111, 2345, 278.5, 347.45
(b) decimal numbers to hexadecimal decimal numbers.
1025, 278.5, 14.09375, 1445, 347.45
(c) hexadecimal numbers to both decimal and binary.
1F.C, F F F.118, 1A4.C, 1023, 11.1
Solution: (a)
165
82
41
20
10
5
2
1
A.BENHARI
=
=
=
=
=
=
=
=
2 82 + 1,
2 41 + 0,
2 20 + 1,
2 10 + 0,
2 5 + 0,
2 2 + 1,
2 1 + 0,
2 0 + 1,
b0 = 1
b1 = 0
b2 = 1
b3 = 0
b4 = 0
b5 = 1
b6 = 0
b7 = 1
Thus the binary representation for 165 is

165 = (b7 b6 b5 b4 b3 b2 b1 b0 )2 = (10100101)2
111
55
27
13
6
3
1
=
=
=
=
=
=
=
2 55 + 1,
2 27 + 1,
2 13 + 1,
2 6 + 1,
2 3 + 0,
2 1 + 1,
2 0 + 1,
b0 = 1
b1 = 1
b2 = 1
b3 = 1
b4 = 0
b5 = 1
b6 = 1
Thus the binary representation for 111 is

111 = (b6 b5 b4 b3 b2 b1 b0 )2 = (1001111)2
278
139
69
34
17
8
4
2
1
Also
=
=
=
=
=
=
=
=
=
2 139 + 0,
2 69 + 1,
2 34 + 1,
2 17 + 0,
2 8 + 1,
2 4 + 0,
2 2 + 0,
2 1 + 0,
2 0 + 1,
b0 = 0
b1 = 1
b2 = 1
b3 = 0
b4 = 1
b5 = 0
b6 = 0
b7 = 0
b8 = 1
0.5
2
Thus the binary representation for 278.5 is

278.5 = (100010110.1)2
(c)
(1F.C)16 = C(16)1 + F (16)0 + 1(16)1
= 12/16 + 15 + 16 = (31.75)10
Thus the decimal and binary representation for (1F.C)16 is
(1F.C)16 = (31.75)10 = (11111.11)2
(1A4.C)16 = C(16)1 + 4(16)0 + A(16)1 + 1(16)2
= 12/16 + 4 + 160 + 256 = (420.75)10
Thus the decimal and binary representation for (1A4.C)16 is
(1A4.C)16 = (420.75)10 = (110100100.11)2
A.BENHARI
(11.1)16 = 1(16)1 + 1(16)0 + 1(16)1

= 1/16 + 1 + 16 = (17.0625)10
Thus the decimal and binary representation for (11.1)16 is
(11.1)16 = (17.0625)10 = (10001.0001)2
5. What is the absolute error in approximating 1/3 by 0.3333 ? What is the corresponding relative error ?
Solution: Let x = 1/3 = 0.33333 and x = 0.3333, then the absolute error is
Abs. Error = |0.33333 0.3333| = 3 105
and the relative error is
Rel. Error =
Since 9.0001 105 <
3 105
= 9.0001 105
0.33333
1
104 , therefore, x approximates x to 4 significant digits.
2
6. Evaluate the absolute error in each of the following calculations and hence give
the answer to a suitable degree of accuracy.
(a) 9.01 + 9.96, (b) 4.65 3.429, (c) 0.7425 0.7199, (d) 0.7078 0.87
Solution: (a) Let x1 = 9.01, x2 = 9.96, and S = x1 + x2 = 18.97. Also, let e1 and
e2 be the errors in x1 and x2 respectively. Then the respective absolute errors are
as follows:
1
1
|e1 | 102 , |e2 | 102
2
2
and
Abs. Error |e1 | + |e2 | = 1 102
The relative error is
Rel. Error =
Abs. Error
(1 102 )
=
= 5.2715 104
S
18.97
Thus, the result lies in the range S Abs. Error

18.97 1 102 ,
or 18.96 S 18.98
The answer may be rounded meaningfully to 18.9 which is correct to 3 significant

digits (sd)(1 decimal place (dp)).
(c) Let x1 = 0.7425, x2 = 0.7199, and P = x1 x2 = 0.5345. Let e1 and e2 be the
errors in x1 and x2 respectively and
|e1 |
A.BENHARI
1
104 ,
2
|e2 |
1
104
2
Then the relative error is

Rel. Error =
1
1
104
104
|e1 |
|e2 |
+
2
+ 2
= 0.1368 103
|x1 | |x2 |
0.7425
0.7425
and the absolute error is

Abs. Error = Rel. Error P = 7.3116 105
Thus the product lies in the range
P Abs. Error = 0.5345 7.3116 105 ,
or 0.5344 P 0.5346
which is correct to 3 sd (3 dp).

7. Find the absolute and relative errors in approximating by 3.1416. What are the
corresponding errors in the approximation 100 314.16
Solution: Let x = = 3.14159654 and x = 3.1416, then the absolute error is
Abs. Error = |3.14159654 3.1416| = 7.3464 106
Rel. Error =
7.3464 106
= 2.3384 106
3.141592654
1
105 , therefore, x approximates x to 5 significant digits.
2
Similarly, the absolute error in the approximation 100 314.16 is 7.3464 104
1
and the relative error is 2.3384 106 . Since 2.3384 106 < 105 , therefore,
2
x approximates x to 5 significant digits.
Since 2.3384 106 <
8. Calculate the absolute error, relative error, and number of significant digits in the
following approximations, with p x:
(a)
(b)
(c)
x = 25.234, p = 25.255
x =
e, p = 19/7
x = 2, p = 1.414
Solution: (a) Let x = 25.234 and p = 25.255, then the absolute error is define as
Abs. Error = |25.234 25.255| = 0.021
1
Since 0.021 < 101 , therefore, p approximates x to 1 significant digit.
2
Now we define the relative error as follows
Rel. Error =
A.BENHARI
0.021
= 8.3221 104
25.234
Since 8.3221 104 <
1
103 , therefore, p approximates x to 3 significant digits.
2
(b) Let x = e = 2.71828 and p = 19/7 = 2.71429, then the absolute error is define
as
Abs. Error = |2.71828 2.71429| = 3.9900 103
1
Since 3.9900 103 < 101 , therefore, p approximates x to 1 significant digit.
2
The relative error is define as
Rel. Error =
Since 1.4678 103 <
(c) Let x =
(3.9900 103 )
= 1.4678 103
2.71828
1
101 , therefore, p approximates x to 1 significant digit.
2
2 = 1.4142 and p = 1.414, then the absolute error is define as

Abs. Error = |1.4142 1.414| = 2.0 104
1
Since 2.0 104 < 103 , therefore, p approximates x to 3 significant digits.
2
Now we define the relative error as follows
Rel. Error =
Since 1.4142 104 <
(2.0 104 )
= 1.4142 104
1.4142
1
103 , therefore, p approximates x to 3 significant digits.
2
9. Write each of the following numbers in (decimal) floating point form, starting the
word length m and the exponent e
13.2,
12.532,
2/125
Solution: The floating decimal point representation of 13.2 is

13.2 has representation 0.132 102
(M = 0.132, e = 2)
Similarly, writing 12.532 in decimal floating point form

12.532 has representation 0.12532 102
(M = 0.12532, e = 2)
and the number 2/125 in decimal floating point form

2
125
has representation (
2
10) 101
125
(M =
20
, e = 1)
125
10. Find absolute error in each of the following calculations (all numbers are rounded):
(a)
(b)
(c)
A.BENHARI
187.2 + 93.5
0.281 3.7148
28.315
(d)
(6.2342 0.82137)/27.268
Solution: (a) Let x1 = 187.2, x2 = 93.5, and S = x1 + x2 = 280.7. Also, let e1

and e2 be the errors in x1 and x2 respectively, then the respective absolute errors
are as follows:
1
1
|e1 | 101 , |e2 | 101
2
2
and
Abs. Error |e1 | + |e2 | = 1 101
Thus, the result lies in the range S Abs. Error
280.7 1 101 ,
or 280.6 S 280.8
which is correct to 3 significant digits.

(b) Let x1 = 0.281, x2 = 3.7148, and P = x1 x2 = 1.0439. Let e1 and e2 be the
errors in x1 and x2 respectively, define as follows
|e1 |
1
103 ,
2
|e2 |
1
104
2

Rel. Error =
|e1| |e2 |
+
|x1 | |x2 |
1
2
1
103
104
+ 2
= 1.7929 103
0.281
3.7148

Abs. Error = Rel. Error P = 1.8716 103
Thus the product lies in the range
P Abs. Error = 1.0439 1.1.8716 103 ,
or 1.0420 P 1.0458
or 1.04 which is correct to 3 sd (2 dp).
(c) Let x = 28.315, n = 1/2, and R = x = 5.3212. Let e be the error in x and
define as follows
1
|e| 103
2
1 1
. 103
|e|
2
2
= 8.8292 106
Rel. Error =
|x|
28.315
Abs. Error = Rel. Error R = (8.8292 106 ) 5.3212 = 4.6982 105
The correct value of R lies in the range R Abs. Error, that is
5.3212 R 5.3213
A.BENHARI

s
x1 x2
=
x3
0.4333. Let e1 , e2 , and e3 be the errors in x1 , x2 , and x3 respectively, and are define
as follows
1
1
1
|e1 | 104 , |e2 | 105 , |e3 | 103
2
2
2
(d) Let x1 = 6.2342, x2 = 0.82137, x3 = 27.2680, n = 1/2, and D =
1
Rel. Error
2
12 104 12 105 12 103

+
+
6.2342
0.82137
27.2680
gives
Rel. Error 1.6222 105
Now the absolute error is define as
Abs. Error = Rel. Error D = (1.6222 105 ) 0.43333 = 7.0291 106
So the answer lies in the range D Abs. Error, that is
0.43329 D 0.4333
11. Express the base of natural logarithms e as a normalized floating point number,
using both chopping and symmetric rounding for each of the following system.
Solution: (a)
Chopping: e = 2.718 100
Rounding: e = 2.718 100
(b)
Chopping: e = 2.718281 100
Rounding: e = 2.718282 100
(c)
Chopping: e = 1.010110111 21
Rounding: e = 1.010111000 21
12. Write down the normalized binary floating point representations of 1/3, 1/5, 1/7, 1/9
and 1/10. Use enough bits in the mantissa to see the recurring patterns.
Solution:
written as
The normalized binary floating point representation of 1/5 can be
4
1
1
= ( 4)22 (M = , e = 2)
5
10
5
Similarly, the normalized binary floating point representation of 1/10 can be written
as
4
1
1
= ( 8)23 (M = , e = 3)
10
10
5
A.BENHARI
Chapter 2
Solution of Nonlinear Equations
1. Find the root of f (x) = ex 2 x in the interval [2.4, 1.6] accurate to 104
using the bisection method.
Solution: Using the bisection method gives a1 = 2.4 and b1 = 1.6, so f (2.4) =
0.4907 > 0 and f (1.6) = 0.1981 < 0. Thus [2.4, 1.6] we have x1 =
1
(a +b1 ) = 2.0 and f (x1 ) = 0.1353 > 0. Since f (x1 )f (b1 ) < 0 so [2.0, 1.6]
2 1
we have x2 = 21 (x1 + b1 ) = 1.8 and f (x2 ) = 0.0347 < 0. Continuing in this
manner, the bisection gives x10 = 1.84141, accurate to within 104 .
2. Use the bisection method to find solutions accurate to within 104 on the interval
[5, 5] of the following functions:
(a)
(b)
(c)
f (x) = x5 10x3 4
f (x) = 2x2 + ln(x + 6) 3
f (x) = ln(x + 1) + 30ex 3
Solution: (a) The bisection method gives x17 = 3.1818, accurate to within 104 .
(b) The bisection method fails because f (5) = 47 and f (5) = 49.3979, which
gives f (5)f (5) > 1.
(c) The bisection method gives x17 = 2.9084, accurate to within 104 .
3. The following equations have a root in the interval [0, 1.6]. Determine these with
an error less than 104 using the bisection method.
(a) 2x ex = 0
(b) e3x + 2x 2 = 0.
Solution: (a) The bisection method gives x16 = 0.35173, accurate to within 104 .
(b) The bisection method gives x15 = 0.9730, accurate to within 104 .
4. Estimate the number of iterations needed to achieve an approximation with accuracy 104 to the solution of f (x) = x3 + 4x2 + 4x 4 lying in the interval [0, 1]
using the bisection method.
Solution: To find an approximation of that is accurate to within 104 , we need to
determine the number of iterations n so that
| xn | <
ba
< 104
n
2
gives
2n > 104 ,
or n ln 2 > ln 104
Thus a bound for the number of iterations is

n
9.2103
= 13.2882,
0.96312
gives n = 14
So using the bisection method, we have x14 = 0.5943
A.BENHARI
5. Use the bisection method for f (x) = x3 3x + 1 in [1, 3] to find:

(a) The first eight approximation to the root of the given equation.
(b) Find an error estimate | x8 |.
Solution: (a) Using the bisection method, we have the first eight approximations
as
x1 = 1.5000, x2 = 1.7500, x3 = 1.6250, x4 = 1.5625
x5 = 1.5313,
x6 = 1.5469,
x7 = 1.5391,
x8 = 1.5352
(b) The error bound for the approximation in part (a), we have
| x8 |
31
= 0.00781
28
6. Solve the Problem 1 by the false position method.

Solution: Since f (x) = ex 2 x, using the false position method gives a1 = 2.4
and b1 = 1.6, so f (2.4) = 0.4907 > 0 and f (1.6) = 0.1981 < 0. Thus
[2.4, 1.6] we have
x1 =
(1.6)(0.4907) (2.4)(0.1981)
b1 f (a1 ) a1 f (b1 )
=
= 1.8301
f (a1 ) f (b1 )
(0.4907) (0.1981)
Since f (x1 ) = 0.0095 < 0 and f (a1 )f (x1 ) < 0 so [2.4, 1.8301] we have
x2 =
(1., 8301)(0.4907) (2.4)(0.0095)

x1 f (a1 ) a1 f (x1 )
=
= 1.8409
f (a1 ) f (x1 )
(0.4907) (0.0095)
Since f (x2 ) = 0.0004 < 0 and f (a1 )f (x2 ) < 0 so [2.4, 1.8409], therefore,
the false position method gives x3 = 1.8414, accurate to within 104 .
7. Use the false position method to find the root of f (x) = x3 + 4x2 + 4x 4 on the
interval [0, 1] accurate to 104 .
Solution: The false position method gives x8 = 0.59431, accurate to within 104 .
8. Use the false position method to find solution accurate to within 104 on the interval [1, 1.5] of the equation 2x3 + 4x2 2x 5 = 0.
9. Use the false position method to find solution accurate to within 104 on the interval [3, 4] of the equation ex 3x2 = 0.
10. The cubic equation x3 3x 20 = 0 can be written as
A.BENHARI
10
(a)
x=
(b)
x=
(x3 20)
.
3
20
.
3)
(x2
(c) x = (3x + 20)1/3 .

Choose the form which satisfies the condition |g 0 (x)| < 1 on [1, 4] and then find
third approximation x3 when x0 = 1.5.
Solution: For (a) and (b), |g 0 (x)| > 0 on [1, 4], but for (c), we have g 0 (x) =
1
< 1 on [1, 4]. So by using (c), the third approximation by the fixed(3x + 20)2/3
point method is, x3 = 3.0789.
1
11. Consider the nonlinear equation g(x) = e0.5x defined on the interval [0, 1]. Then
2
(a) Show that there exists a unique fixed-point for g in [0, 1].
(b) Use the fixed-point iterative method to compute x3 , set x0 = 0.
(c) Compute an error bound for your approximation in part (b).
Solution: (a) Since given g is continuous in [0, 1] and g(0) = 0.5 [0, 1] and
g(1) = 0.8243 [0, 1]. Also, g 0 (x) = 41 e0.5x and g 0 (0) = 0.25, g 0 (1) = 0.41218, so
|g 0 (x)| < 1 for x [0, 1].
(b) The fixed-point iterative method using x0 = 0 gives
x1 = g(x0 ) = 0.5,
x2 = g(x1 ) = 0.64201,
x3 = g(x2 ) = 0.58705
(c) The error bound for the approximation is

| x3 |
k3
|x 1 x0 |
1k
where k = max |g 0 (x)| = g 0 (1) = 0.41218. Thus

0x1
(0.41218)3
|0.5 0| = 0.05957
| x3 |
1 0.41218
12. An equation x3 2 = 0 can be written in form x = g(x) in two ways:
(a) x = g1 (x) = x3 + x 2
(2 + 5x x3 )
(b) x = g2 (x) =
5
Generate first four approximations from xn+1 = gi (xn ), i = 1, 2 by using x0 = 1.2.
Show which sequence converges to 21/3 and why ?
Solution: (a) The fixed-point iterative method using x0 = 0 gives x4 = 16.3514.
(b) The fixed-point iterative method using x0 = 0 gives x4 = 1.2599.
The second sequence converges to 21/3 because |g20 (21/3 )| = 0.0476 < 1 whereas the
first sequence does not converge to 21/3 because |g10 (21/3 )| = 4.7622 > 1.
A.BENHARI
11
x2 4kxn + 7
, n 0 con13. Find value of k such that the iterative scheme xn+1 = n
4
verges to 1. Also, find the rate of convergence of the iterative scheme.
x2 4kx + 7
and = 1 is the fixed-point, then
4
1 4k + 7
, gives k = 1
g(1) = 1 =
4
Solution: Since g(x) =
Also
g 0 (x) =
2x 4k
,
4
1
gives g 0 (1) = =
6 0
2
a linear convergence.
14. Write the equation x2 6x + 5 = 0 in the form x = g(x), where x [0, 2], so that
the iteration xn+1 = g(xn ) will converge to the root of the given equation for any
initial approximation x0 [0, 2].
x2 + 5
, then
6
x
g 0 (x) = < 1 for all x [0, 2]
3
x2 + 5
will converge to the root of the given
Thus the iteration xn+1 = g(xn ) = n
6
equation for any initial approximation x0 [0, 2].
Solution: Taking g(x) =
15. Which of the following iterations
6
1 2
xn +
(a) xn+1 =
4
xn
!
6
(b) xn+1 = 4 2
xn
is suitable to find a root of the equation x3 = 4x2 6 in the interval [3, 4] ? Estimate
the number of iterations required to achieve 103 accuracy, starting from x0 = 3.
1 2 6
x +
which is continuous in [3, 4], but g10 (x) > 1
4
x
for all x (3,4). So g!1 (x) is not suitable.
6
(b) g2 (x) = 4 2 which is continuous in [3, 4] and g(x) [3, 4] for all x [3, 4].
xn
Also, |g20 (x)| = |12/x3 | < 1 for all x (3, 4). Then from the Theorem ?? implies
that a unique fixed-point exists in [3, 4]. To find an approximation of that is
accurate to within 103 , we need to determine the number of iterations n so that
Solution: (a) Let g1 (x) =
| xn |
kn
|x1 x0 | < 103
1k
With k = max |g 0 (x)| = 4/9 and using the fixed-point method by taking x0 = 3,
3x4
we have x1 = 10/3, we have and

| xn |
A.BENHARI
(4/9)n
|10/3 3| < 103
(1 4/9)
12
Thus a bound for the number of iterations is

(4/9)n
|10/3 3| < 103
(1 4/9)
and solving for n, we get, n = 8.
16. An equation ex = 4x2 has a root in [4, 5]. Show that we cannot find that root
1
using x = g(x) = ex/2 for the fixed-point iteration method. Can you find another
2
iterative formula which will locate that root ? If yes, then find third iterations with
x0 = 4.5. Also find the error bound.
1 x/2
e
> 0 for all x (4, 5), therefore, the fixed-point
4
iteration fails to converge to the root in [4, 5]. Consider x = g(x) = ln(4x2 ) and
2
its derivative can be found as, g 0 (x) = . Note that |g 0 (x)| < 1 for all x (4, 5)
x
and the fixed-point iteration converges to the root in [4, 5]. Using the fixed-point
iteration method gives the third iteration as x3 = 4.3253. With k = max |g 0 (x)| =
Solution: Since g 0 (x) =
4x5
g (4) = 0.5 and taking x0 = 4.5, we have x1 = 4.3944. Thus the error bound for
the above approximation, gives
| x3 |
(0.5)3
|4.3944 4.5| = 0.0264
(1 0.5)
17. Solve the Problem 1 by the Newtons method by taking initial approximation
x0 = 2.
Solution: Since f (x) = ex 2 x and its derivative is f 0 (x) = ex 1. Using the
Newtons iterative formula, we get
x1 = x0
and
x 2 = x1
(0.1353)
f (x0 )
=
2
= 1.8435
f 0 (x0 )
(0.8647)
(0.0017)
f (x1 )
= 1.8435
= 1.8414
0
f (x1 )
(0.8417)
Thus, the Newtons method gives x2 = 1.8414, accurate to within 104 .

18. Let f (x) = ex + 3x2 . (a) Find the Newtons formula g(xk ).
(b) Start with x0 = 4 and compute x4 . (c) Start with x0 = 0.5 and compute x4 .
Solution: (a) Using the Newtons formula gives
xk+1 = g(xk ) = xk
(xk 1)exk + 3x2k

exk + 3x2k
=
exk + 6xk
exk + 6xk
(b) For x0 = 4, x4 = 0.1215

(c) For x0 = 0.5, x4 = 32.1077
A.BENHARI
13
19. Use the Newtons formula for the reciprocal of square root of a number 15 and
then find the 3rd approximation of number, with x0 = 0.05.
Solution: Let N be a positive number and x = 1/ N . If f (x) = 0, then

x = = 1/ N is the exact zero of the function
f (x) = 1/x2 N,
gives f 0 (x) = 2/x3
Hence, assuming an initial estimate to the root, say, x = x0 and by using the
Newtons iterative formula, we get
(1/x20 N )
= x0 (3 N x20 )/2.
x1 = x0
3
(2/x0 )
In general, we have
xn+1 = xn (3 N x2n )/2,
n = 0, 1, . . . ,
Now to find the reciprocal of square root of a number N = 15, using an initial
gauss of say x0 = 0.05, we have
n = 0,
n = 1,
n = 0,
x1 = x0 (3 N x20 )/2 = 0.05(3 (15)(0.05)2 )/2 = 0.0741

x2 = x1 (3 N x21 )/2 = 0.0741(3 (15)(0.0741)2 )/2 = 0.1081
x3 = x2 (3 N x22 )/2 = 0.1081(3 (15)(0.1081)2 )/2 = 0.1527
20. Use the Newtons method to find solution accurate to within 104 of the equation
tan(x) 7x = 0, with initial approximation x0 = 4.
Solution: The Newtons method gives x6 = 4.1231e 014, accurate to within
104 .
21. Find the Newtons formula for f (x) = x3 3x + 1 in [1, 3] to calculate x3 , if
x0 = 1.5. Also, find the rate of convergence of the method.
Solution: The Newtons method gives x3 = 1.5321. To find the order of convergence, we do the following:
2x3 1
x3 3x + 1
= 2
g(x) = x
3x2 3
3x 3
and its derivative is
g 0 (x) =
6x(x3 3x + 1)
6x4 18x2 + 6x
=
=0
(3x2 3)2
(3x2 3)2
Thus g 0 () = 0, gives at least quadratic convergence.

1
22. Rewrite the nonlinear equation g(x) = e0.5x which defined in the interval [0, 1] in
2
the equivalent form f (x) = 0 and then use the Newtons method with x0 = 0.5 to
find third approximation x3 .
Solution: Writing x = g(x) or x g(x) = 0 gives f (x) = 2x e0.5x . Using the
Newtons method gives x3 = 0.7148.
A.BENHARI
14
f (xn )
, n 0 with f () = f 0 () = 0 and
0
f (xn )
f 00 () 6= 0. Find the rate of convergence for this scheme.
23. Given the iterative scheme xn+1 = xn
Solution: Since the Newtons iteration function is

f (x)
g(x) = x 0
f (x)
g 0 (x) = 1
f (x)f 00 (x)
[f 0 (x)]2 f (x)f 00 (x)
=
[f 0 (x)]2
[f 0 (x)]2
Since f 0 () = 0, so by using the Hopital rule, it gives us

g 0 (x) =
f 0 (x)f 00 (x) + f (x)f 000 (x)

2f 0 (x)f 00 (x)
[f 00 (x)]2 + f (x)f 000 (x) + f 0 (x)f 000 (x) + f (x)f (4)(x)

2[f 0 (x)]2 + 2f 0 (x)f 000 (x)
1
6= 0.
g 0 () =
2
Thus the rate of convergence for the given scheme is linear.
g 0 (x) =
24. The Halleys iterative method

"
xn+1
f (xn )f 00 (xn )
f (x)
1
= xn 0
f (x)
2[f 0 (xn )]2
#1
n0
has a order of convergence 3 for a simple root of f (x) = 0. Use it to find a iterative
procedure
to approximate the square root of a number N . Then use it to approximate 8, using x0 = 2.5 and compute first three iterations.
Solution: We shall compute x = N 1/2 by finding a positive root for the nonlinear
equation
x2 N = 0
where N > 0 is the number whose square root is to be found. Therefore, if f (x) = 0,
then x = N 1/2 is the exact root. Let
f (x) = x2 N
then f 0 (x) = 2x,
and f 00 (x) = 2
Hence, taking N = 8 and the initial estimate to the root, say, x = x0 = 2.5 and by
using the Halleys iterative formula, we get first approximation
x1 =
(2x30 + 2x0 N )
= 2.6636
(3x20 + N )
x2 =
(2x31 + 2x1 N )
= 2.7459
(3x21 + N )
second approximation
and third approximation is

x3 =
A.BENHARI
(2x32 + 2x2 N )
= 2.7872
(3x22 + N )
15
25. Find the positive root of f (x) = x10 1 by the secant method by using starting
values x0 = 1.2 and x1 = 1.1 accurate to within 104 .
Solution: Using the secant method gives x0 = 1.2 and x1 = 1.1, so f (1.2) = 5.1917
and f (1.1) = 1.5937, and the new approximation
x2 =
(1.2)(1.5937) (1.1)(5.1917)
x0 f (x1 ) x1 f (x0 )
=
= 1.0557
f (x1 ) f (x0 )
1.5937 5.1917
Similarly, the secant method gives other approximations

x3 = 1.0192,
x4 = 1.0042,
x5 = 1.0004,
x6 = 1.0000,
x7 = 1.0000
accurate to within 104 .

26. Find the first three estimates for the equation x3 2x5 = 0 by the secant method
using x0 = 2 and x1 = 3.
Solution: Using the secant method gives x0 = 2 and x1 = 3, so f (2) = 1 and
f (3) = 16, and the first approximation
x2 =
(2)(16) (3)(1)
x0 f (x1 ) x1 f (x0 )
=
= 2.0588
f (x1 ) f (x0 )
16 + 1
second approximation using x1 = 3, x2 = 2.0588, f (3) = 16, and f (2.0588) =

0.3908, gives
x3 =
(3)(0.3908) (2.0588)(16)
x1 f (x2 ) x2 f (x1 )
=
= 2.0813
f (x2 ) f (x1 )
0.3908 16
and third approximation using x2 = 2.0588, x3 = 2.0813, f (2.0588) = 0.3908, and

f (2.0813) = 0.1472 is
x4 =
(2.0588)(0.1472) (2.0813)(0.3908)
x2 f (x3 ) x3 f (x2 )
=
= 2.0948
f (x3 ) f (x2 )
0.1472 + 0.3908
27. Solve the equation ex x = 0 by using the secant method, starting with x0 = 0
and x1 = 1, accurate to 104 .
Solution: The secant method gives x4 = 0.56714, accurate to within 104 .
28. Use the secant method to find a solution accurate to within 104 for ln(x)+x5 = 0
on [3, 4].
Solution: The secant method gives x3 = 3.6934, accurate to within 104 .
29. Find the root of multiplicity of the function f (x) = (x 1)2 ln(x) at = 1.
A.BENHARI
16
Solution: From the function f (x) = (x 1)2 ln(x) we have f (1) = 0 and
f0
= 2(x 1) ln x +
f 00
= 2 ln x + 4
f 000 =
(x 1)2
:
x
(x 1) (x 1)2
:
x
x2
6 6(x 1) (x 1)2
+
:
x
x2
x3
f 0 (1)
= 0
f 00 (1) = 0
f 000 (1)
= 6 6= 0.
Thus the root of multiplicity of the given function is 3, that is m = 3.

31. Show that the root of multiplicity of the function f (x) = x4 x3 3x2 + 5x 2 is
3 at = 1. Estimate the number of iterations required to solve the problem with
accuracy 104 , start with the starting value x0 = 0.5 by using:
(a) the Newtons method
(b) the first modified Newtons method
(c) the second modified Newtons method.
Solution: From the function f (x) = x4 x3 3x2 + 5x 2 we have f (1) = 0 and
f 0 = 4x3 3x2 6x + 5 :
f 0 (1) = 0
f 00 = 12x2 6x 6 :
f 00 (1)
= 0
000
000
f
= 24x 6 :
f (1)
= 18 =
6 0.
Thus the root of multiplicity of the given function is 3, that is m = 3.
(a) Using the Newtons method gives the first approximation as follows
x1 = x0
x40 x30 3x20 + 5x0 2

= 0.6785714
4x30 3x20 6x0 + 5
Thus the Newtons method gives x21 = 0.999863, accurate to within 104 .
(b) Using the first modified Newtons method gives the first approximation as
follows
x4 x3 3x2 + 5x0 2
= 1.0357143
x1 = x0 3 0 3 0 2 0
4x0 3x0 6x0 + 5
Thus the first modified Newtons method gives x4 = 0.9999998, accurate to within
104 .
(c) Using the second modified Newtons method gives the first approximation as
follows
(x40 x30 3x20 + 5x0 2)(4x30 3x20 6x0 + 5)
x1 = x0
(4x30 3x20 6x0 + 5)2 (x40 x30 3x20 + 5x0 2)(12x20 6x0 6)
that is, x1 = 0.960526. Thus the second modified Newtons method gives x4 =
1.000001, accurate to within 104 .
A.BENHARI
17
32. If f (x), f 0 (x) and f 00 (x) are continuous and bounded on a certain interval containing
x = and if both f () = 0 and f 0 () = 0 but f 00 () 6= 0, show that
xn+1 = xn 2
f (xn )
f 0 (xn )
will converge quadratically if xn is in the interval.

Solution: Let f (x) = (x )2 h(x), h() 6= 0. Then
g(x) = x 2
(x )2 h(x)
(2(x )h(x) + (x )2 h0 (x))
and
g 0 () = 1
2
= 0,
2
g 00 () 6= 0
x2
33. Show that iterative scheme xn+1 = 1 + xn n , n 0 converges to 2. Find the
2
rate of convergence of the sequence.
Solution: Since
So
x2
g(x)
= 1+x
g( 2) = 1 + 2 1 = 2
2 is a fixed-point of g(x). Also,

g 0 (x) = 1
2x
,
2
and |g 0 ( 2)| = | 0.414214| < 1.
Hence the given sequence converges to 2. Also, since g 0 ( 2) = 0.414214 6= 0,

therefore, the rate of convergence of the sequence is linear.
35. The following sequence is linearly convergent. Use the Aitkens method to generate
the first four terms of the sequence {x0n }:
x0 = 0.5,
xn+1 = cos xn ,
n0
Solution: Starting with x0 = 0.5, the first two approximations by using the given
iterative scheme xn+1 = cos xn ,
n 0, gives x1 = 0.8776 and x1 = 0.6390.
Using the Aitkens method
x0 = x0
(x1 x0 )2
= 0.7314
x2 2x1 + x0
Similarly the other three approximations by the Aitkens method are:

x1 = 0.73609,
A.BENHARI
x2 = 0.73765,
18
x3 = 0.73847
37. Use the fixed-point iteration to find the first four approximations to the root of the
1
xn+1 = ln(1 + xn ), n 0, using x0 = 0.5. Use the Aitken accelerated method to
2
speed up the convergence of the iteration method.
Solution: Using the fixed-point method and the Aitkens method, we have the
desired approximations as follows:
Fixed-point method
x1 = 0.20270
x2 = 0.09230
x3 = 0.04414
x4 = 0.02160
Aitken Accelerated method

x0 = 0.02702
x1 = 0.00691
x2 = 0.00175
x3 = 0.00044
39. Solve the following system using the Newtons method:

4x3 + y = 6
x2 y = 1
Start with initial approximation x0 = y0 = 1. Stop when successive iterates differ
by less than 107 .
Solution: To solve the given nonlinear system using the Newtons method, we do
the following:
f1 (x, y) = 4x3 + y 6, f1 x = 12x2 , f1 y = 1
f2 (x, y) = x2 y 1,
f2 x = 2xy, f2 y = x2
At the given initial approximation x0 = 1.0 and y0 = 1.0, we get
f1 (1.0, 1.0) = 1,
f1
= f1 x
x
= 12.0,
f1
= f1 y
y
= 1
f2 (1.0, 1.0) = 0,
f1
= f2 x
x
= 2,
f2
= f2 y
y
= 1.0
The Jacobian matrix J and its inverse J 1 at the given initial approximation can
be calculated as
f1 f1
12.0 1.0
y
x
J =
=
f2 f2
2.0 1.0
x y
and one can find its inverse as
0.1000 0.1000
0.2000 1.2000
Substituting all these values in the Newtons formula to get the first approximation
as follows
A.BENHARI
x1
y1
1.0
1.0
0.1000 0.1000
0.2000
1.2000
19
1
0
1.1000
0.8000
Similarly, the Newtons method gives x(5) = 1.08828,

to within 107 .
y (5) = 0.84434 accurate
40. Solve the following system using the Newtons method

x
+ ey = 68.1
sin x y = 3.6
Start with initial approximation x0 = 2.5, y0 = 4, compute the first three approximations.
Solution: Solving the given nonlinear system using the Newtons method, we do
the following:
f1 (x, y) = x + ey 68.1, f1 x = 1,
f1 y = ey
f2 (x, y) = sin x y + 3.6, f2 x = cos x, f2 y = 1
At the given initial approximation x0 = 2.5 and y0 = 4.0, we get
f1 (2.5, 4.0) = 11.0018,
f1
= f1 x
x
= 1.0000,
f1
= f1 y
y
= 54.5982
f2 (2.5, 4.0) = 0.1985,
f1
= f2 x
x
= 0.8011,
f2
= f2 y
y
= 1.0000
The Jacobian matrix J and its inverse J 1 at the given initial approximation can
be calculated as
f1
J =
f2
f1
y
1.0000
f2
0.8011
54.5982
1.0000
and one can find its inverse as
0.0234 1.2775
0.0187
0.0234
Substituting all these values in the Newtons formula to get the first approximation
as follows
x1
y1
2.5
4.0
0.0234 1.2775
0.0187
0.0234
11.0018
0.1985
2.4961
4.2016
Similarly, second and third approximations by Newtons method are as follows

x(2) = 2.5188, y (2) = 4.1835 and x(3) = 2.5188, y (3) = 4.1833
A.BENHARI
20
Chapter 3
Systems of Linear Equations
1. Determine the matrix C given by the following expression

C = 2A 3B,
if the matrices A and B are
2 1 1
2 3
A = 1
,
2
1 2
Solution: Since
1 1 1
B= 0 1 3
2 1 4
4 2 2
4 6
2A = 2
4
2 4
and
3 3 3
9
3B =
0 3
6 3 12
therefore,
1 5 1
4 3 2 3 2 3
1 3
43 69
C = 2A 3B =
= 2
2 + 0
2 1 8
46
2 3 4 12
2. Find the product AB and BA for the matrices of the Problem 1.
Solution: Multiplying the matrix A with the matrix B, we have
20+2
21+1
23+4
4 2 3
AB = 1 + 0 + 6 1 + 2 + 3 1 + 6 + 12 = 5 4 17
2+0+4
2+1+2
2+3+8
6 5 13
Similarly, multiplying the matrix B with the matrix A, we have
3 2 6
2 1 + 2 1 + 2 + 1 1 + 3 + 2
0 + +3 + 6 = 5 5 9
BA = 0 1 + 6 0 + 2 + 3
11 4 13
4 1 + 8 2 + 2 + 4 2 + 3 + 8
3. Show that the product AB of the following rectangular matrices is a singular
matrix.
!
6 3
2 1 2
4 ,
B=
A= 1
3 4 1
2
1
Solution: First multiplying the matrix A with the matrix B, we have
3
6 9
12 9 6 + 12 12 + 3
2 4
AB =
= 14 17 6
2 + 12 1 16
1 2
3
4 + 3
24
41
A.BENHARI
21
Then the determinant of the matrix AB can be calculated as
3
6
|AB| = 14 17
1
2
9
6
3
= 3(51 12) + 6(6 42) 9(28 17)
or
|AB| = 189 216 + 405 = 0
4. Let
1
2 3
A=
0 1 2 ,
2
0 2
1 1
2
B=
1 1 1 ,
1 0
2
1 0 1
C=
0 1 2 ,
2 0 1
(a) Compute AB and BA and show that AB 6= BA.

(b) Find (A + B) + C and A + (B + C).
(c) Show that (AB)T = B T AT
Solution: (a) Multiplying the matrix A with the matrix B, we have
12+3 1+2+0 22+6

2
3 6
AB = 0 + 1 + 2 0 1 + 0 0 + 1 + 4 = 3 1 5
2+0+2 2+0+0 4+0+4
4
2 8
Now multiplying the matrix B with the matrix A, we have
1+0+4
21+0
3+2+4
5
1
9
BA =
1 + 0 2 2 1 + 0 3 + 2 2 = 3 3 3
1+0+4
2+0+0
3+0+4
5
2
7
which shows that AB =6= BA.
(b) Adding the matrices A and B, we have
2 3 5
1+1 2+1
3+2
A + B = 0 1 1 + 1 2 1 = 1 0 1
3 0 4
2+1 0+0
2+2
Then
3 3 6
2+1
3+0 5+1
(A + B) + C =
1 + 0 0 + 1 1 + 2 = 1 1 3
5 0 5
3+2
0+0 4+1
Similarly,
2 1 3
1+1
1+0 2+1
B + C = 1 + 0 1 + 1 1 + 2 = 1 2 1
3 0 3
1+2
0+0 2+1
Then
A.BENHARI
3 3 6
1+2 2+1
3+3
A + (B + C) =
0 1 1 + 2 2 + 1 = 1 1 3
5 0 5
2+3 0+0
2+3
22
(c) Since we know from part (a) that product AB is equal to
2
3 6
AB =
3 1 5
4
2 8
Now interchanging rows and columns, gives
2
3 4
(AB)T =
3 1 2
6
5 8
Since transpose of the matrix A is
1
0 2
T
A = 2 1 0
3
2 2
and transpose of the matrix B is
1 1
1
T
1 1
B = 1
2 1
2
Then the product B T AT is equal to
2
3 4
12+3 0+1+2 2+0+2
T T
B A = 1 + 2 + 0 0 1 + 0 2 + 0 + 0 = 3 1 2
6
5 8
22+6 0+1+4 4+0+4
Thus (AB)T = B T AT .
5. Let
A=
1 1
0 1
B=
1 0
1 1
then show that (AB)1 = B 1 A1 .

Solution: Multiplying the matrix A with the matrix B, we have
AB =
1+1 0+1
0+1 0+1
2 1
1 1
Now finding the inverse of product AB as

1
(AB)
1
=
(2 1)
1 1
1
2
1 1
1
2
Similarly, finding inverse of A and B, we get

A
A.BENHARI
1
=
(1 0)
1 1
0
1
23
1 1
0
1
and
B
1
=
(1 0)
Then
B A
1 0
1 1
1 + 0 1 + 0
1 + 0
1+1
1 0
1 1
1 1
1
2
Thus (AB)1 = B 1 A1 .
6. Find a value of x and a value of y so that AB T = 0, where A = [1 x 1] and
B = [2 2 y].
Solution: Multiplying the row matrix A and the column matrix B T , we have
AB T =
1 x 1
2 = 2 + 2x + y
y
Since AB T = 0, then it gives, x = (2 y)/2. In particular, if y = 1, then x = 1/2,

and so on.
7. Evaluate the determinant of each matrix
3
1 1
2
0
4
A=
,
1 5
1
4 1 6
3
6 4
B=
,
5 0 9
17 46 7
C=
20 49 8
23 52 9
Solution: First, we find the determinant of the matrix A as follows
3
1
0
det(A) = |A| = 2
1 5
1
4
1
= 3(0 + 20) + 1(4 2) + (1)(10 0) = 72
Similarly, the determinant of the matrix B and the matrix C can be computed as
follows
4 1 6
det(B) = |B| = 3 6 4 = 4(54 0) + 1(20 + 27) + 6(0 30) = 83
5 0 9
and
17 46
det(C) = |C| = 20 49
23 52
7
8
9
= 17(441 416) + 46(184 180) + 7(1040 1127) = 0
8. Find all zeros (values of x such that f (x) = 0) of the polynomial f (x) = det(A)
where
x1 3
2
3 x
1
A=
2 1 x2
A.BENHARI
24
Solution: The determinant of the matrix A can be calculated as
det(A)
x1 3
3 x
=
2 1
2
1
x2
= (x 1)(x2 2x 1) + 3(2 3x + 6) + 2(3 2x) = x3 3x2 12x + 31
Given
f (x) = det(A) = x3 3x2 12x + 31 = 0
Solving this cubic equation, we get, x1 = 4.0787, x2 = 2.2698, x3 = 3.3485.
9. Compute the adjoint of each matrix A, and find the inverse of it if it exists
(a)
1
2
3 4
(b)
1 2 1
4
2 1
,
1 5 7
(c)
1 1 0
1 0 1
0 1 1
Solution: (a) Since we know that for a 1 1 matrix [a11 ], the entry a11 itself is
defined to be the minor and cofactor of a11 , therefore, the minors Mij of all elements
aij of the matrix A are calculated as follows
M11 = 4,
M12 = 3,
M21 = 2,
M22 = 1
Now using these minors we can have the cofactors of the matrix as follows
A11
A12
A21
A22
=
=
=
=
(1)1+1 M11
(1)1+2 M12
(1)2+1 M21
(1)2+2 M22
=
M11
= M12
= M21
=
M22
=
4
=
3
= 2
=
1
So that the matrix of the cofactor is
C=
4 3
2 1
and the transpose of the cofactor matrix C is the adjoint of A, that is
Adj(A) =
4 2
3
1
Now by using the cofactor expansion along the first row, we can find the determinant
of the matrix A as follows
det(A) = a11 A11 + a12 A12 = 4 + 6 = 10
The inverse of the matrix A can be calculated by using
1
A.BENHARI
1
Adj(A)
=
=
det(A)
10
4 2
3
1
25
2/5 1/5
3/10 1/10
(b) Firstly, the minors Mij of all elements aij of the matrix A are calculated as
follows
1
4
M11 =
= 7 20 = 27
5 7
M12
2
4
=
1 7
2 1
1 5
2 1
=
5 7
1 1
=
1 7
1 2
1 5
M13 =
M21
M22
10 1 =
= 14 + 5 =
7 + 1 =
52 =
2 1
=
1
4
8+1 =
1 1
2
4
4+2 =
14 =
M23 =
M31
M32
= 14 4 = 18
M33 =
1 2
2 1
A11
A12
A13
A21
A22
A23
A31
A32
A33
=
=
=
=
=
=
=
=
=
(1)1+1 M11
(1)1+2 M12
(1)1+3 M13
(1)2+1 M21
(1)2+2 M22
(1)2+3 M23
(1)3+1 M31
(1)3+2 M32
(1)3+3 M33
=
=
=
=
=
=
=
=
=
M11
M12
M13
M21
M22
M23
M31
M32
M33
= 27
=
18
=
9
=
9
= 6
= 3
=
9
= 6
= 3
27 18
9
9 6 3
C=
9 6 3
27
9
9
Adj(A) =
18 6 6
9 3 3
A.BENHARI
26
det(A) = a11 A11 + a12 A12 + a13 A13 = 27 + 36 9 = 0
Since the determinant of the matrix is zero, therefore, the inverse of A does not
exist.
(c) The minors Mij of all elements aij of the matrix A are calculated as follows
M11 =
M12 =
M13 =
M21 =
M22 =
M23 =
M31 =
M32 =
M33 =
0 1
1 1
1 1
0 1
1 0
0 1
1 0
1 1
1 0
0 1
1 1
0 1
1 0
0 1
1 0
1 1
1 1
1 0
= 0 1 = 1
= 10 =
= 10 =
= 10 =
= 10 =
= 10 =
= 10 =
= 10 =
= 0 1 = 1
A11
A12
A13
A21
A22
A23
A31
A32
A33
A.BENHARI
=
=
=
=
=
=
=
=
=
(1)1+1 M11
(1)1+2 M12
(1)1+3 M13
(1)2+1 M21
(1)2+2 M22
(1)2+3 M23
(1)3+1 M31
(1)3+2 M32
(1)3+3 M33
=
=
=
=
=
=
=
=
=
27
M11
M12
M13
M21
M22
M23
M31
M32
M33
=
=
=
=
=
=
=
=
=
1
1
1
1
1
1
1
1
1
1 1
1
1 1
C = 1
1 1 1
1 1
1
1 1
Adj(A) = 1
1 1 1
det(A) = a11 A11 + a12 A12 + a13 A13 = 1 1 0 = 2
To find the inverse of the matrix A, we use
A1
1 1
1/2
1/2 1/2
1
1
Adj(A)
1/2
1 1 = 1/2 1/2
= 1
=
det(A)
2
1/2
1/2
1/2
1 1 1
10. Let
2
1 3
2 0
A = 1
,
3 2 1
Show that A(Adj A) = (Adj A)A = det(A)I3 .

Solution: The minors Mij of all elements aij of the matrix A are calculated as
follows
2 0
M11 =
=
20 =
2
2 1
1 0
3 1
1
2
=
3 2
M12 =
M13
M21 =
M22 =
M23 =
M31 =
A.BENHARI
1 3
2 1
2 3
3 1
2
1
3 2
1 3
2 0
= 1 0 = 1
2 6 = 4
1+6 =
2 9 = 7
= 4 3 = 7
28
0 6 = 6

2 3
=
1 0
M32
2 1
1 2
M33
= 0+3 = 3
= 4+1 = 5
A11
A12
A13
A21
A22
A23
A31
A32
A33
=
=
=
=
=
=
=
=
=
(1)1+1 M11
(1)1+2 M12
(1)1+3 M13
(1)2+1 M21
(1)2+2 M22
(1)2+3 M23
(1)3+1 M31
(1)3+2 M32
(1)3+3 M33
=
=
=
=
=
=
=
=
=
M11
M12
M13
M21
M22
M23
M31
M32
M33
=
=
=
=
=
=
=
=
=
2
1
4
7
7
7
6
3
5
2
1 4
7
C = 7 7
6 3
5
2 7 6
Adj(A) = 1 7 3
4
7
5
det(A) = a11 A11 + a12 A12 + a13 A13 = 4 + 1 12 = 7
Now finding the product
7
0
0
0
A(Adj(A)) = (Adj(A))A = 0 7
0
0 7
which can be also written as
1 0
A(Adj(A)) = (Adj(A))A = 7 0 1 0
= det(A)I3
0 0 1
11. Use the matrices of the Problem 9, solve the following systems using matrix inverse
method
(a) Ax = [1, 1]T ,
A.BENHARI
(b) Ax = [2, 1, 3]T ,
29
(c) Ax = [1, 0, 1]T .
Solution: (a) Since the inverse of the matrix A exists, and is
2/5 1/5
3/10 1/10
Using the matrix inverse method, we get
x=A b=
2/5 1/5
3/10 1/10
1
1
1/5
2/5
(b) Since the inverse of the matrix A does not exist, therefore, the matrix inverse
method can not be applicable.
(c) Since the inverse of the matrix A exists, and is
A1
1/2
1/2 1/2
1/2
= 1/2 1/2
1/2
1/2
1/2
Using the matrix inverse method, we get
1/2
1/2 1/2
1
0
1
1/2 0 = 1
x = A b = 1/2 1/2
1/2
1/2
1/2
1
0
13. Solve the following systems using
(a)
x1 +
5x1
2x1 +
the Cramers rule.

3x2 x3 =
4
2x2 x3 = 2
2x2 + x3 =
9
(b)
x1 + x2 + 3x3 =
2
5x1 + 3x2 + x3 =
3
2x1 + 3x2 + x3 = 1
(c)
4x1 + x2 3x3 =
9
3x1 + 2x2 6x3 = 2
x1 5x2 + 3x3 =
1
Solution: (a) Writing the given system in matrix form
1
3 1
4
x1
5 2 1 x2 = 2
2
2
1
x3
9
gives
1
3 1
A = 5 2 1
2
2
1
A.BENHARI
and b = 2
9
30
The determinant of the matrix can be calculated as
1
3 1
|A| = 5 2 1 = 1(2 + 2) + 3(2 5) 1(10 + 4) = 35 6= 0

2
2
1
which showed that the given matrix A is nonsingular. Then the matrices A1 , A2 ,
and A3 can be computed as
4
3 1
1
4 1
1
3
4
A1 =
2 2 1 , A2 = 5 2 1 , A3 = 5 2 2
9
2
1
2
9
1
2
2
9
Now finding the determinant of the matrices A1 , A2 , and A3 , we get
|A1 | =
4(2 + 2) + 3(9 + 2) 1(4 + 18) = 0 21 14 = 35
|A2 | =
1(2 + 9) + 4(2 5) 1(45 + 4) = 7 28 49 = 70
|A3 | = 1(18 + 4) + 3(4 45) + 4(10 + 4) = 14 147 + 56 = 105
Therefore, by using the Cramers rule, we get
x1 =
35
|A1 |
=
=1
|A|
35
x2 =
70
|A2 |
=
=2
|A|
35
x3 =
105
|A3 |
=
=3
|A|
35
(b) The matrix form of the give system is
1 1 3
x1
2
5 3 1 x2 = 3
2 3 1
x3
1
gives
1 1 3
A=
5 3 1
2 3 1
and b =
3
1
1 1
|A| = 5 3
2 3
3
1
1
= 1(3 3) + 1(2 5) + 3(15 6) = 0 3 + 27 = 24 6= 0
1 1
2
1
2 3
2 1 3
3
3 1 , A3 = 5 3
A1 = 3 3 1 , A2 = 5
2 3 1
2 1 1
1 3 1
A.BENHARI
31
The determinant of the matrices A1 , A2 , and A3 can be computed as

|A1 | =
2(3 3) + 1(1 3) + 3(9 + 3) = 0 4 + 36 = 32
|A2 | =
1(3 + 1) + 2(2 5) + 3(5 6) = 4 6 33 = 35
|A3 | = 1(3 9) + 1(6 + 5) + 2(15 6) = 12 + 11 + 18 = 17
Therefore, by the Cramers rule
x1 =
x2 =
32
|A1 |
=
= 1.3333
|A|
24
35
|A2 |
=
= 1.4583
|A|
24
x3 =
17
|A3 |
=
= 0.70833
|A|
24
(c) Writing the given system in matrix form
9
4
1 3
x1
2 6 x2 = 2
3
1
x3
1 5
3
gives
4
1 3
2 6
A= 3
1 5
3
and b = 2
1
4
1
2
|A| = 3
1 5
3
6
3
= 4(6 30) + 1(6 9) 3(15 2) = 60 6= 0
9
1 3
4
9 3
4
1
9
2
2
6
3
2
6
3
2
2
A1 =
, A2 =
, A3 =
1 5
3
1
1
3
1 5
1
Now finding the determinant of the matrices A1 , A2 , and A3 , we get
9(6 30) + 1(6 + 6) 3(10 2) = 216 + 0 24 = 240
|A1 | =
|A2 | =
4(6 + 6) + 9(6 9) 3(3 + 2) = 0 135 15 = 150
|A3 | = 4(2 10) + 1(2 3) + 9(15 2) = 32 5 153 = 190
Therefore, by using the Cramers rule, we get
x1 =
x2 =
x3 =
A.BENHARI
240
|A1 |
=
=4
|A|
60
150
|A2 |
=
= 2.5
|A|
60
190
|A3 |
=
= 3.1667
|A|
60
32
14. Use the simple Gaussian elimination method to show that the following system
does not have a solution
3x1 + x2
= 1.5
2x1 x2 x3 =
2
4x1 + 3x2 + x3 =
0
Solution: The process begins with the augmented matrix form
..
. 1.5
..
.
2
..
1 .
0
3
1
0
2 1 1
Since a11 = 3 6= 0, so we wish to eliminate the elements a21 and a31 by subtracting
from the second and third rows the appropriate multiples of the first row. In this
case the multiples are given
2
m21 = ,
3
Hence
and m31 =
4
3
..
. 1.5
..
.
1
..
1 . 2
3
1
0
0 5/3 1
5/3
(1)
(1)
As a22 = 5/3 6= 0, therefore, we eliminate entry in a32 position by subtracting

the multiple m32 = 1 of the second row from the third row, to get
..
. 1.5
..
.
1
..
0 . 1
3
1
0
0 5/3 1
Obviously, the original set of equations has been transformed to an upper-triangular

form. Since the third diagonal element of the obtaining upper-triangular matrix
is zero, which means that the coefficient matrix of the given system is singular
and therefore, the given system has no unique solution. Now expressing the set in
algebraic form yields
3x1 +
x2
5/3x2
x3
0x3
= 1.5
=
1
= 1
From the third equation, we find that 0 = 1, which is not possible, therefore, this
system has no solution.
15. Solve the following systems using the simple Gaussian elimination method
(a)
x1 x2
= 0
x1 + 2x2 x3 = 1
x2 + 4x3 = 0
A.BENHARI
33
(b)
x1 + x2 +
x3 = 1
2x1 + 3x2 + 4x3 = 3
4x1 + 9x2 + 16x3 = 11
(c)
3x1 + 2x2 x3 = 1
x1 3x2 + 2x3 = 2
2x1 x2 + x3 = 3
(d)
2x1 + x2 + x3
x1
+ 9x2 + 8x3
x1 + 3x2 + 5x3
x2
+
+
+
x4
4x4
2x4
x4
=
=
=
=
3
15
10
2
Solution: (a) The process begins with the augmented matrix form
..
0
.
0
..
1
2
1
.
1
..
0 1
4 . 0
1 1
Since a11 = 1 6= 0, so we wish to eliminate the elements a21 by subtracting from

the second row the appropriate multiple of the first row. In this case the multiple
is given as
1
= 1
m21 =
1
Hence
..
1
1
0
.
0
..
0
1 1 . 1
..
0 1
4 . 0
(1)
(1)
As a22 = 1 6= 0, therefore, we eliminate entry in a32 position by subtracting the

1
= 1 of the second row from the third row, to get
multiple m32 =
1
..
. 0
..
. 1
..
3 . 1
1 1
0
0
1 1

form. Since all the diagonal elements of the obtaining upper-triangular matrix are
nonzero, which means that the coefficient matrix of the given system is nonsingular
and therefore, the given system has a unique solution. Now expressing the set in
x1 x2
= 0
x2 x3 = 1
3x3 = 1
A.BENHARI
34
Now using backward substitution to give

3x3 = 1,
gives
x2 = x3 + 1 = 1/3 + 1, gives
x1 = x2 ,
gives
x3 = 1/3
x2 = 4/3
x1 = 4/3
(b) The process begins with the augmented matrix form
.
1 .. 1
.
4 .. 3
.
4 9 16 .. 11
1 1
2 3
m21 =
Hence
2
= 2,
1
and m31 =
4
=4
1
.
1 .. 1
.
2 .. 1
..
0 5 12 . 7
1 1
0 1
(1)
(1)
As a22 = 1 6= 0, therefore, we eliminate entry in a32 position by subtracting the

multiple m32 = 5 of the second row from the third row, to get
..
.
1
1
1
1
.
0 1 2 .. 1
..
0 0 2 . 2

x1 + x2 + x3 = 1
x2 + 2x3 = 1
2x3 = 2
2x3 = 2,
gives
x2 = 1 2x3 = 1 2,
gives
x1 = 1 x2 x3 = 1 + 1 1, gives
(c) The augmented matrix form of the given system
..
3
2
1
.
1
..
1 3
2
.
2
..
2 1
1 . 3
A.BENHARI
35
x3 = 1
x2 = 1
x1 = 1
1
m21 = ,
3
Hence
and m31 =
2
3
..
3
2 1 .
1
.
.
0 11/3 7/3 . 5/3
..
0 7/3 5/3 . 7/3
(1)
(1)

the multiple m32 = 7/11 of the second row from the third row, to get
.
1 ..
1
..
7/3 .
5/3
..
0 2/11 . 14/11
3
2
0 11/3

3x1 +
2x2
11/3x2 +
x3 =
1
7/3x3 =
5/3
2/11x3 = 14/11

2/11x3
= 14/11,
gives
11/3x2 = 5/3 7/3x3 = 44/3, gives
3x1
= 1 2x2 + x3 = 0,
gives
x3 = 7
x2 = 4
x1 = 0
(d) The augmented matrix form of the given system
2 1 1
1 9 8
1 3 5
0 1 0
4
2
..
. 3
..
. 15
..
. 10
..
.
2
from the second, third, and fourth rows the appropriate multiples of the first row.
In this case the multiples are given
1
m21 = ,
2
A.BENHARI
and m31 =
36
1
2
Hence
..
.
3
..
. 33/2
..
7/2 11/2 3/2 . 17/2
.
1
0
1 ..
2
1
1 1
2
0 17/2 15/2 9/2
(1)
(1)
(1)
As a22 = 17/2 6= 0, therefore, we eliminate entries in a32 and a42 positions by

subtracting the multiples m32 = 7/17 and m42 = 2/17 of the second row from the
third and fourth rows, to get
1
2
0 17/2
0
0
15/2
9/2
41/17 6/17
0 15/17
8/17
..
.
3
..
. 33/2
..
. 29/17
..
. 1/17
(1)
(1)

the multiple m42 = 15/41 of the third row from the fourth row, to get
1
1
2
0 17/2
15/2
0
0 41/17
1
9/2
6/17
14/41
..
.
3
..
. 33/2
..
. 29/17
..
. 28/41

2x1 +
x2 +
17/2x2 +
x3
15/2x3 +
41/17x3
x4
9/2x4
6/17x4
14/41x4
=
3
= 33/2
= 29/17
= 28/41
Now using backward substitution, we get

x1 = 1,
x2 = 0,
x3 = 1,
x4 = 2
16. Find a value of k so that the following linear system has a non-trivial solution,
and solve it in this case.
2x1 + 2x2 + 3x3 = 0
3x1 + kx2 + 5x3 = 0
x1 + 7x2 + 3x3 = 0
A.BENHARI
37
Solution: The augmented matrix form of the given system
..
2
2
3
.
0
.
3 k 5 .. 0
..
1 7 3 . 0
3
m21 = ,
2
Hence
and m31 =
1
2
..
. 0
..
. 0
..
6 3/2 . 0
2
2
3
0 k 3 1/2
(1)
(1)
If a22 = k 3 6= 0, then, we eliminate entry in a32 position by subtracting the

multiple m32 = 6/k 3 of the second row from the third row, to get
.
3 .. 0
.
1/2 .. 0
.
0 (3k 15)/2(k 3) .. 0
2
2
0 k3
Since we know that a homogeneous system of n equations in n unknowns has a

solution other than trivial solution if and only if the determinant of the coefficient
matrix equal to zero, therefore, we put
(3k 15)/2(k 3) = 0,
gives k = 5
Now expressing the set in algebraic form yields

2x1 + 2x2 +
3x3 = 0
2x2 + 1/2x3 = 0
0x3 = 0
and this underdetermined linear system has infinitely many solutions. If we take
x3 = 4, then we get
x1 = 5, x2 = 1, x3 = 4
17. Determine the rank of each matrix
3
1 1
0
4
A=
2
,
1 5
1
4 1 6
B=
3 6 4 ,
5 0 9
17 46 7
C=
20 49 8
23 52 9
Solution: Applying the forward elimination step of the simple Gaussian elimination on the given matrix A and eliminating the elements below the first pivot (first
diagonal element) to
3
1
1
2/3 14/3
0
0 16/3 4/3
A.BENHARI
38
We finished with the first elimination step. The second pivot is in (2, 2) position
but after eliminating the element below it we find the triangular form to be
3
1
1
0 2/3 14/3
0
0 36
Since the number of pivots are three, therefore, the rank of the given matrix is 3.
Note that the original matrix is nonsingular as the rank of 3 3 matrix is 3.
Now applying the forward elimination step of the simple Gaussian elimination on
the given matrix B and eliminating the elements below the first pivot (first diagonal
element) to
4
1
6
0 27/4 17/2
0 5/4 3/2
4
1
6
17/2
0 27/4
0
0 83/27
Note that this matrix is also nonsingular.
Repeating the same above procedure in finding the rank of the matrix C, we get
17
46
7
87/17 4/17
0
0 174/17 8/17
and
17
46
7
0 87/17 4/17
0
0
0
Since the number of pivots are two, therefore, the rank of the given matrix is 2.
Note that the original matrix C is singular as the rank of 3 3 matrix is 2.
19. Determine the rank of each matrix
2 1
0
1
A = 2 1
,
1
1 1
0.1 0.2
0.3
0.6
B = 0.4 0.5
,
0.7 0.8 0.901
C=
1
2
3
4
2
4
5
6
3 4
6 8
7 9
8 10
Solution: Applying the forward elimination step of the simple Gaussian elimination on the given matrix A and eliminating the elements below the first pivot (first
diagonal element) to
2 1
0
0
1
0
0 3/2 1
A.BENHARI
39
Replacing second row by third row, we get
2 1
0
0 3/2 1
0
0
1
Note that the original matrix is nonsingular as the rank of 3 3 matrix is 3.
Now applying the forward elimination step of the simple Gaussian elimination on
the given matrix B and eliminating the elements below the first pivot (first diagonal
element) to
0.1
0.2
0.3
0.6
0.0 0.3
0.0 0.6 1.199

0.1
0.2
0.3
0.0 0.3 0.6

0.0
0.0 0.001
Note that this matrix is nonsingular.
Repeating the same above procedure in finding the rank of the matrix C, we get
1
2
3
4
0
0
0
0
0 1 2 3
0 2 4 6
Replacing second row by third row, we get
1
2
3
4
0 1 2 3
0
0
0
0
0 2 4 6
but after eliminating the element below it we find the following form
1
2
3
4
0 1 2 3
0
0
0
0
0
0
0
0
Since the number of pivots are two, therefore, the rank of the given matrix is 2.
Note that the original matrix C is singular as the rank of 4 4 matrix is 2.
A.BENHARI
40
20. Solve the Problem 15 using the Gaussian elimination with partial pivoting and
complete pivoting.
Solution: Partial Pivoting
(a)
x1 x2
= 0
x1 + 2x2 x3 = 1
x2 + 4x3 = 0
For the first elimination step, since a11 = 1 and a21 = 1 are both largest absolute
coefficients of first variable x1 , therefore, we can choose a11 = 1 as the first pivotal
element. Eliminate first variable x1 from the second row by subtracting the multiple
1
= 1 of row 1 from row 2, gives
m21 =
1
x1 x2
= 0
x2 x3 = 1
x2 + 4x3 = 0
For the second elimination step, eliminate second variable x2 from the third row
1
by subtracting the multiple m32 =
1
x1 x2
x2
x3
3x3
= 0
= 1
= 1
Obviously, the original set of equations has been transformed to an equivalent

upper-triangular form. Now using backward substitution to get the solution
x1 = 4/3,
x2 = 4/3,
x3 = 1/3
(b)
x1 + x2 +
x3 = 1
2x1 + 3x2 + 4x3 = 3
4x1 + 9x2 + 16x3 = 11
For the first elimination step, since 4 is the largest absolute coefficient of first
variable x1 , therefore, the first row and the third row are interchange, giving us
4x1 + 9x2 + 16x3 = 11
2x1 + 3x2 + 4x3 = 3
x1 + x2 +
x3 = 1
Eliminate first variable x1 from the second and third rows by subtracting the mul1
2
tiples m21 = and m31 = of row 1 from row 2 and row 3 respectively, gives
4
4
4x1 +
9x2 + 16x3 =
11
3/2x2 4x3 = 5/2
5/4x2
x3 = 7/5
A.BENHARI
41
For the second elimination step, 3/2 is the largest absolute coefficient of second
variable x2 , so eliminate second variable x2 from the third row by subtracting the
5
multiple m32 = of row 2 from row 3, gives
6
4x1 +
9x2 +
3/2x2
16x3 =
11
4x3 = 5/2
1/3x3 =
1/3

upper-triangular form. Now using backward substitution to get the solution of the
linear system which is
x1 = 1, x2 = 1, x3 = 1
(c)
3x1 + 2x2 x3 = 1
x1 3x2 + 2x3 = 2
2x1 x2 + x3 = 3
Since 3 is the largest absolute coefficient of first variable x1 in the first row, therefore, eliminate first variable x1 from the second and third rows by subtracting the
2
1
multiples m21 = and m31 = of row 1 from row 2 and row 3 respectively, gives
3
3
3x1 +
2x2
x3 =
1
11/3x2 + 7/3x3 = 5/3
7/3x2 + 5/3x3 = 7/3
variable x2 in the second row, so eliminate second variable x2 from the third row
7
of row 2 from row 3, gives
11
3x1 +
2x2
11/3x2 +
x3 =
1
7/3x3 =
5/3
2/11x3 = 14/11
Since the original set of equations has been transformed to an equivalent uppertriangular form, therefore, using backward substitution to get the solution of the
x1 = 0, x2 = 4, x3 = 7
(d)
2x1 + x2 + x3
x1
+ 9x2 + 8x3
x1 + 3x2 + 5x3
x2
+
+
+
x4
4x4
2x4
x4
=
=
=
=
3
15
10
2
Since 2 is the largest absolute coefficient of first variable x1 in the first row, therefore, eliminate first variable x1 from the second and third rows by subtracting the
A.BENHARI
42
multiples m21 =
1
1
and m31 =
of row 1 from row 2 and row 3 respectively, gives
2
2
2x1 + x2
+ x3
17/2x2 + 15/2x3
7/2x2 + 11/2x3
x2
+
+
+
x4
9/2x4
3/2x4
x4
=
=
=
=
3
33/2
17/2
2
variable x2 in the second row, so eliminate second variable x2 from the third and
2
7
and m42 =
of row 2 from
fourth rows by subtracting the multiples m32 =
17
17
row 3 and row 4 respectively, gives
2x1 + x2
+ x3
17/2x2 + 15/2x3
41/17x3
15/17x3
x4
9/2x4
6/17x4
8/17x4
=
=
=
=
3
33/2
29/17
1/17
For the third elimination step, 41/17 is the largest absolute coefficient of third
variable x3 in the third row, so eliminate third variable x3 from the fourth row by
15
subtracting the multiple m43 =
41
2x1 + x2
+ x3
x4
17/2x2 + 15/2x3 + 9/2x4
41/17x3 6/17x4
14/41x4
=
=
=
=
3
33/2
29/17
28/41
x1 = 1,
x2 = 0,
x3 = 1,
x4 = 2
Complete Pivoting
(a)
x1 x2
= 0
x1 + 2x2 x3 = 1
x2 + 4x3 = 0
Solution: For the first elimination step, since 4 is the largest absolute coefficient
of variable x3 in the given system, therefore, the first row and the third row are
interchange as well as the first column and third column. Then eliminate third
1
of row 1
variable x3 from the second row by subtracting the multiple m21 =
4
from row 2, gives
4x3
x2
= 0
7/4x2 x1 = 1
x2 + x1 = 0
A.BENHARI
43
7
For the second elimination step,
is the largest absolute coefficient of second
4
variable x2 in the remaining system of equations, so eliminate second variable x2
4
of row 2 from row 3,
from the third row by subtracting the multiple m32 =
7
gives
4x3
x2
=
0
7/4x2
x1 =
1
3/7x1 = 4/7
upper-triangular form. Now using backward substitution to get the solution of the
given system as
x1 = 4/3, x2 = 4/3, x3 = 1/3
(b)
x1 + x2 +
x3 = 1
2x1 + 3x2 + 4x3 = 3
4x1 + 9x2 + 16x3 = 11
For the first elimination step, since 16 is the largest absolute coefficient of variable
x3 in the given system, therefore, the first row and the third row are interchange
as well as the first column and third column, we get
16x3 + 9x2 + 4x1 = 11
4x3 + 3x2 + 2x1 = 3
x3 + 9x2 + x1 = 1
Then eliminate third variable x3 from the second and the third rows by subtracting
1
4
and m31 =
of row 1 from rows 2 and 3 respectively,
the multiples m21 =
16
16
gives
16x3 +
9x2 +
4x1 =
11
3/4x2 +
x1 = 1/4
7/16x2 + 3/4x1 = 5/16
For the second elimination step, 1 is the largest absolute coefficient of first variable
x1 in the second row and the third column, so the second and third columns are
interchange, giving us
16x3 +
4x1 +
9x2 =
11
x1 + 3/4x2 = 1/4
3/4x1 + 7/16x2 = 5/16
3
Eliminate first variable x1 from the third row by subtracting the multiple m32 =
4
16x3 + 4x1 +
9x2 = 11
x1 + 3/4x2 = 1/4
1/8x2 = 1/8
The original set of equations has been transformed to an equivalent upper-triangular
form. Now using backward substitution to get the solution of the linear system
which is
x1 = 1, x2 = 1, x3 = 1
A.BENHARI
44
(c)
3x1 + 2x2 x3 = 1
x1 3x2 + 2x3 = 2
2x1 x2 + x3 = 3
x1 in the first row and first column, therefore, no need to do any interchanging.
Eliminate first variable x1 from the second and the third rows by subtracting the
2
1
multiples m21 = and m31 = of row 1 from rows 2 and 3 respectively, gives
3
3
3x1 +
2x2
x3 =
1
11/3x2 + 7/3x3 = 5/3
7/3x2 + 5/3x3 = 7/3
variable x2 in the second row and the second column, so again no need to do any
interchanging. Then eliminate second variable x2 from the third row by subtracting
7
the multiple m32 =
11
3x1 +
2x2
11/3x2 +
x3 =
1
7/3x3 =
5/3
2/11x3 = 14/11
x1 = 0, x2 = 4, x3 = 7
(d)
2x1 + x2 + x3
x1
+ 9x2 + 8x3
x1 + 3x2 + 5x3
x2
+
+
+
x4
4x4
2x4
x4
=
=
=
=
3
15
10
2
x2 in the given system, therefore, the first row and the second row are interchange
as well as the first column and second column, we get
9x2 + x1 + 8x3
x2 + 2x1 + x3
3x2 x1 + 5x3
x2
+
+
4x4
x4
2x4
x4
=
=
=
=
15
3
10
2
Now eliminate second variable x2 from second, third, and fourth rows by subtracting
3
1
1
of row 1 from rows 2, 3, and 4
the multiples m21 = , m31 = , and m41 =
9
9
9
respectively, gives
9x2 + x1
17/9x1
4/3x1
1/9x1
A.BENHARI
+
+
+
8x3
1/9x3
7/3x3
8/9x3
+
+
45
4x4
13/9x4
2/3x4
5/9x4
=
=
=
=
15
14/3
5
1/3
Since 7/3 is the largest absolute coefficient of third variable x3 in the third row and
third column, therefore, the second row and the third row are interchange as well
as second column and third column, we get
9x2 + 8x3
7/3x3
1/9x3
8/9x3
x1
4/3x1
17/9x1
1/9x1
+
+
4x4
2/3x4
13/9x4
5/9x4
=
=
=
=
15
5
14/3
1/3
Now eliminate third variable x3 from the third and the fourth rows by subtracting
8
1
and m42 =
of row 2 from row 3 and row 4 respectively,
the multiples m32 =
21
21
gives
9x2 + 8x3
+ x1
+ 4x4
= 15
7/3x3 4/3x1
+ 2/3x4
= 5
41/21x1 31/21x4 = 103/21
13/21x1 + 17/21x4 = 47/21
For the third elimination step, 41/21 is the largest absolute coefficient of first variable x1 in the third row and the third column, so no need to do any interchanging.
Now eliminate first variable x1 from the fourth row by subtracting the multiple
13
m43 =
41
9x2 + 8x3
+ x1
+ 4x4
7/3x3 4/3x1
+ 2/3x4
41/21x1 31/21x4
14/41x4
=
=
=
=
15
5
103/21
28/41
x1 = 1,
x2 = 0,
x3 = 1,
x4 = 2
21. Solve the following linear systems using the Gaussian elimination with partial
pivoting and without pivoting
(a)
1.001x1 + 1.5x2 = 0
2x1 +
3x2 = 1
(b)
x1 + 1.001x2 = 2.001
x1 +
x2 =
2
(c)
6.122x1 + 1500.5x2 = 1506.622
2000x1 +
3x2 =
2003
Solution: Without Pivoting
A.BENHARI
46
(a) The process begins with the augmented matrix form
1.001 1.5
2
3
..
. 0
..
. 1
Since a11 = 1.001 6= 0, so we wish to eliminate the elements a21 by subtracting from
is given as
2
= 1.9980
m21 =
1.001
Hence
..
.
0
1.001
1.5
..
0
0.003 . 1
1.001x1 + 1.5x2
= 0
0.003x2 = 1

0.003x2 = 1,
gives
1.001x1 = 0 1.5x2 , gives
x2 = 333.6667
x1 = 500.0000
(c) The process begins with the augmented matrix form
6.122 1500.5
2000 3
..
. 1506.622
..
. 2003
Since a11 = 6.122 6= 0, so we wish to eliminate the elements a21 by subtracting from
is given as
2000
= 326.69
m21 =
6.122
Hence
..
6.122
1500.5
.
1506.622
..
0
490196.28 . 490196.28
The original set of equations has been transformed to an upper-triangular form.
Since all the diagonal elements of the obtaining upper-triangular matrix are nonzero,
which means that the coefficient matrix of the given system is nonsingular. Expressing the set in algebraic form yields
A.BENHARI
6.122x1 + 1500.5x2
= 1506.622
490196.28x2 = 490196.28
47

x1 = 1, quadx2 = 1
Partial Pivoting
(a) For the first elimination step, since 2 is the largest absolute coefficient of first
variable x1 , therefore, the first row and the second row are interchange, giving us
2x1
+ 3x2
= 1
1.001x1 + 1.5x2 = 0
Eliminate first variable x1 from the second row by subtracting the multiple m21 =
1.001
2
2x1 +
3x2 =
1
0.0015x2 = 0.5005
which means that the coefficient matrix of the given system is nonsingular. Now
using backward substitution to give
x1 = 500.0000quadand x2 = 333.6667
(c) For the first elimination step, since 2000 is the largest absolute coefficient of
first variable x1 , therefore, the first row and the second row are interchange, giving
us
2000x1 + 3x2
= 2003
6.122x1 + 1500.5x2 = 1506.622
Eliminate first variable x1 from the second row by subtracting the multiple m21 =
6.122
2000
2000x1 + 3x2
= 2003
+ 1500.49x2 = 1500.49
which means that the coefficient matrix of the given system is nonsingular. Now
using backward substitution to give
x1 = 1.0000quadand x2 = 1.0000
22. The elements of the matrix A, the Hilbert matrix, are defined by
aij = 1/(i + j 1),
for i, j = 1, 2, . . . , n
Find the solution of the system Ax = b for n = 3 and b = [1, 2, 3]T , using the
Gaussian elimination by without pivoting, partial pivoting, and complete pivoting.
A.BENHARI
48
Solution: The Hilbert matrix for n = 3 is
1
1/2 1/3
1/2 1/3 1/4

1/3 1/4 1/5
Without Pivoting Writing the augmented matrix form
1
1/2 1/3 1
1/2 1/3 1/4 2

1/3 1/4 1/5 3
from the second row and the third row the appropriate multiples of the first row.
In this case the multiples are given as
m21 =
1
1/2
=
1
2
m31 =
1
1/3
=
1
3
and
Hence
1 1/2 1/3 1
0 1/12 1/12 3/2

0 1/12 4/45 8/3
For the second elimination step, we eliminate second variable x2 from the third row
1/12
1/12
1 1/2 1/3
1
3/2
0 1/12 1/12
0 0
1/180 7/6
Since the original set of equations has been transformed to an upper-triangular
form, therefore, using the backward substitution give
x1 = 27,
23. Solve the following systems using
(a)
x1 +
2x1 +
3x1 +
x2 = 192,
x3 = 210.
the Gauss-Jordan method

4x2 + x3 = 1
4x2 + x3 = 9
5x2 2x3 = 11
(b)
x1 + x 2 + x3 =
1
2x1 x2 + 3x3 =
4
3x1 + 2x2 2x3 = 2
A.BENHARI
49
(c)
2x1
x1
3x1
2x1
3x2
x2
5x2
2x2
+
+
+
+
6x3
2x3
2x3
2x3
+
+
+
x4
4x4
2x4
3x4
=
=
=
=
2
1
11
2
Solution: (a) Writing the given system in the augmented matrix form
.
1 .. 1
.
1 .. 9
.
3 5 2 .. 11
1 4
2 4
The first elimination step is to eliminate elements a21 = 2 and a31 = 3 by subtracting the multiples m21 = 2 and m31 = 3 of row 1 from rows 2 and 3 respectively,
gives
..
1
4
1
.
1
.
0 4 1 .. 7
..
0 7 5 . 8
The second row is now divided by -4 to give
.
1 ..
1
.
1 1/4 .. 7/4
..
0 7 5 .
8
(1)
The second elimination step is to eliminate elements in positions a12 = 4 and

a32 = 7 by subtracting the multiples m12 = 4 and m32 = 7 of row 2 from rows
1 and 3 respectively, gives
.
0 ..
8
..
1/4 . 7/4
.
0 0 13/4 .. 17/4
1 0
0 1
The third row is now divided by 13/4 to give
..
.
8
..
. 7/4
..
1 . 17/13
1 0
0
0 1 1/4
0 0
(1)
The third elimination step is to eliminate elements in positions a23 = 1 by subtracting the multiples m23 = 1/4 and m13 = 2 of row 3 from row 2, gives
..
1
0
0
.
8
.
0 1 0 .. 27/13
..
0 0 1 .
17/13
A.BENHARI
50
Obviously, the original set of equations has been transformed to a diagonal form.
x1 =
8
x2 = 27/13
x3 =
17/13
which is the required solution of the given system.
(c) Writing the given system in the augmented matrix form
.
1 .. 2
.
4 .. 1
..
3 5 2
2 . 11
.
2 2
2 3 .. 2
6
2 3
1 1 2
Divide the first row by 2 to give
.
3 1/2 .. 1
1 3/2
.
1
1 2
4 .. 1
..
3
5 2
2 . 11
.
2
2
2 3 .. 2
The first elimination step is to eliminate elements a21 = 1, a31 = 3, and a41 = 2 by
subtracting the multiples m21 = 1, m31 = 3, and m41 = 2 of row 1 from rows 2, 3,
and 4 respectively, gives
.
3 1/2 .. 1
.
5 7/2 .. 0
..
1/2 11 1/2 . 8
.
1 4 4 .. 0
3/2
1
0 1/2
Multiply the second row by -2 to give
1 3/2
0
1
0 1/2
3 1/2
10
11 1/2
4
..
.
..
.
..
.
..
.
0
(1)
(1)
The second elimination step is to eliminate elements in positions a12 = 3/2, a32 =
(1)
1/2, and a42 = 1 by subtracting the multiples m12 = 3/2, m32 = 1/2, and m42 =
1 of row 2 from rows 1, 3, and 4 respectively, gives
.
11 .. 1
.
7 .. 0
..
0 0 16
4 . 8
.
0 0
6 11 .. 0
1 0 12
0 1
10
A.BENHARI
51
The third row is now divided by 16 to give
1 0
0 1
0 0
12
11
10
0 0
1 1/4
11
..
.
1
..
.
0
..
. 1/2
..
.
0
The third elimination step is to eliminate elements 6, 10, and 3 in the column 3 by
subtracting the multiples m43 = 6, m23 = 10, and m13 = 3 of row 3 from rows 4, 2,
and 1, gives
..
1
0
0
8
.
5
..
0 1 0
9/2 .
5
..
0 0 1
1/4 . 1/2
..
0 0 0 19/2 .
3
Multiply the fourth row by 2/19 to give
1 0 0
0 1 0
0 0 1
8
9/2
1/4
0 0 0
..
.
5
..
.
5
..
. 1/2
..
. 6/19
The fourth elimination step is to eliminate elements 1/4, 9/2, and 8 in the
column 4 by subtracting the multiples m34 = 1/4, m24 = 9/2, and m14 = 8 of
row 4 from rows 3, 2, and 1, gives
1 0 0
0 1 0
0 0 1
0
0
0
0 0 0 1
..
. 47/19
..
.
68/19
..
. 11/19
..
. 6/19
x1
x2
x3
x4
= 47/19
=
68/19
= 11/19
= 6/19
which is the required solution of the given system.

24. The following set of linear equations have common coefficients matrix but different
right side terms.
(a)
2x1 + 3x2 + 5x3 =
0
3x1 + x2 2x3 = 2
x1 + 3x2 + 4x3 = 3
A.BENHARI
52
(b)
2x1 + 3x2 + 5x3 = 1
3x1 + x2 2x3 = 2
x1 + 3x2 + 4x3 = 4
(c)
2x1 + 3x2 + 5x3 = 5
3x1 + x2 2x3 =
6
x1 + 3x2 + 4x3 = 1
the coefficients and three sets of right side terms may be combined into an augmented matrix form
2 3
3 1
1 3
..
.
0 1 5
..
.
. 2 2
6
..
. 3 4 1
If we apply the Gauss-Jordan method to this augmented matrix form and reduce
the first three columns to the unity matrix form, the solution for the three problems are automatically obtained in the fourth, fifth and the sixth columns when
elimination is completed. Calculate the solution in this way.
Solution: Consider the augmented matrix form
2 3
3 1
1 3
..
.
0 1 5
..
. 2 2
6
..
. 3 4 1
The third row is now divided by 16 to give
1 3/2 5/2
3
1 2
..
.
0 1/2 5/2
..
. 2
2
6
..
. 3
4
1
The first elimination step is to eliminate elements a21 = 3 and a31 = 1 by subtracting the multiples m21 = 3 and m31 = 1 of row 1 from rows 2 and 3 respectively,
gives
..
1
3/2
5/2
.
0
1/2
5/2
..
0 7/2 19/2
.
2
1/2
27/2
..
0
3/2
3/2
. 3 7/2
3/2
Multiply the second row by 2/7 to give
1 3/2 5/2
0
1 19/7
0 3/2
A.BENHARI
3/2
..
.
0
1/2 5/2
..
. 4/7 1/7 27/7
..
. 3
7/2
3/2
53
(1)
The second elimination step is to eliminate elements in positions a12 = 3/2 and
(1)
a32 = 3/2 by subtracting the multiples m12 = 3/2 and m32 = 3/2 of row 2 from
rows 1 and 3 respectively, gives
..
1
0
11/7
.
6/7
5/7
23/7
..
0 1
19/7
.
4/7 1/7 27/7
..
. 27/7 26/7
51/7
0 0 18/7
The third row is now divided by 18/7 to give
..
. 6/7
5/7
23/7
1 0 11/7
.
.
0 1
19/7
.
4/7 1/7 27/7
..
0 0
1
.
3/2 13/9 17/6
The third elimination step is to eliminate elements 19/7 and 11/7 in column 3 by
subtracting the multiples m23 = 19/7 and m13 = 11/7 of row 3 from rows 2 and
1 respectively, gives
..
1
0
0
.
3/2
14/9
7/6
..
0 1 0
. 7/2
34/9
23/6
..
.
3/2 13/9 17/6
0 0 1
Thus Now expressing the set in algebraic form for the first linear system yields
x1 =
3/2
x2 = 7/2
x3 =
3/2
for second linear system
x1 = 14/9
x2 =
23/6
x3 = 13/9
and for third linear system
x1 = 7/6
x2 =
34/9
x3 = 17/6
25. Calculate the inverse of each matrix using the Gauss-Jordan method
(a)
3 9 5
0
5 1
,
1
6 3
(b)
1 4 5
2 1 2 ,
(c)
8 1 1
Solution: (a) Consider the augmented matrix
..
3
9
5
. 1
..
0
5 1
. 0
..
1
6 3
. 0
A.BENHARI
54
form
0 0
1 0
0 1
5 2
0
0
2
5 2
0
0 2
5 2
0
0 2
5
Using Gauss-Jordan procedure, we get
..
. 0.3333 0 0
..
.
0 1 0
..
.
0 0 1
1 3 1.6667
..
. 0.3333 0 0
..
.
0 1 0
..
. 0.3333 0 1
1 3 1.6667
0
5
1
3 4.6667
3 4.6667
1 0 2.2667
0 1 0.2000
0 0 4.0667
1 0 0
0 1 0
0 0 1
A1
0.1475
0.9344 0.5574
0.2295 0.0492
= 0.0164
0.0820 0.1475
0.2459
(b) Consider the augmented matrix form
1 4 5
2 1 2
8 1 1
..
. 1 0 0
..
. 0 1 0
..
. 0 0 1
1
4
5
0 7 8
A.BENHARI
..
.
0.1475
0.9344 0.5574
..
. 0.0164
0.2295 0.0492
..
.
0.0820 0.1475
0.2459
Thus
..
. 0.1475
0.9344 0.5574
..
.
0
0.2000
0
..
. 0.0820 0.1475
0.2459
1 0
0
0 1 0.2000
0 0
..
. 0.3333
0.6000 0
..
.
0
0.2000 0
..
. 0.3333 0.6000 1
..
. 0.3333
0 0
..
.
0 0.2000 0
..
. 0.3333
0 1
1 3 1.6667
0
1 0.2000
..
.
1 0 0
..
. 2 1 0
..
.
0 0 1
55

..
.
1 0 0
..
. 2 1 0
..
. 8 0 1
0 31 39
..
. 0.1429
0.5714 0
..
.
0.2857 0.1429 0
..
.
8
0 1
..
. 0.1429
0.5714 0
..
.
0.2857 0.1429 0
..
.
0.8571 4.4286 1
0 0.4286
1 1.1429
0 31
39
1 0
0 1
0.4286
1.1429
0 0 3.5714
0 0
..
. 0.0400
0.0400
0.1200
..
.
0.5600 1.5600
0.3200
..
. 0.2400
1.2400 0.2800
1 0 0
0 1 0
0 0 1
Thus
A1
..
. 0.0400
0.0400
0.1200
..
.
0.2857 0.1429
0
..
. 0.2400
1.2400 0.2800
1 0
0
0 1 1.1429
0.0400
0.0400
0.1200
0.5600
1.5600
0.3200
=
0.2400
1.2400 0.2800
(c) Consider the augmented matrix form
5 2
2
5
0 2
5 2
0 2
..
.
..
.
..
.
..
.
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
1 0.4
0
4.2
0
2
5 2
0 2
0
1 0.4
0
4.2 2
A.BENHARI
5 2
0 2
..
. 0.2000 0 0 0
..
. 0.4000 1 0 0
..
.
0 0 1 0
..
.
0 0 0 1
..
. 0.2000 0 0 0
..
. 0.4000 1 0 0
..
.
0 0 1 0
..
.
0 0 0 1
56
0 0.1905
1
0
1 0.4762
0 2
5
0
0
2
1 0 0.1905
0 1 0.4762
0 0
..
. 0.2381 0.0952 0 0
..
. 0.0952 0.2381 0 0
..
. 0.1905 0.4762 1 0
..
.
0
0 0 1
0
0
4.0476 2
0 0
1 0
0 1
0 0
0.4762
0 0
1 0.4941
5
0 0.2353
1 0.4941
5
..
.
..
.
..
.
..
.
1 0 0 0.0941
0 1 0 0.2353
0 0 1
0.4941
0 0 0
4.0118
1 0
0 1
0 0
..
.
..
.
..
.
..
.
0 0.2353
1 0.4941
0 0 0
1 0
0 1
0 0
0 0 0
1 0 0 0
0 1 0 0
0 0 1
0 0 0 1
A.BENHARI
..
.
..
.
..
.
..
.
1 0.4941
..
.
..
.
..
.
..
.
..
. 0.2471 0.1176 0.0471 0
..
. 0.1176 0.2941 0.1176 0
..
. 0.0471 0.1176 0.2471 0
..
.
0
0
0 1
0 0.0941
0 0 2
..
. 0.2471 0.1176 0.0471 0
..
. 0.0952 0.2381
0 0
..
. 0.0471 0.1176 0.2471 0
..
.
0
0
0 1
0 0.0941
1 0
0 1
0 0
..
. 0.2381 0.0952 0 0
..
. 0.0952 0.2381 0 0
..
.
0
0 1 0
..
.
0
0 0 1
0.2471 0.1176 0.0471 0
0.1176 0.2941 0.1176 0
0.0471 0.1176 0.2471 0

0.0941 0.2353 0.4941 1
0.2493 0.1232 0.0587 0.0235
0.1176 0.2941 0.1176

0
0.0471 0.1176 0.2471
0.0235 0.0587 0.1232 0.2493
0.2493 0.1232 0.0587 0.0235
0.1232 0.3079 0.1466 0.0587
0.0471 0.1176 0.2471
0.0235 0.0587 0.1232 0.2493
0.2493 0.1232 0.0587 0.0235
0.1232 0.3079 0.1466 0.0587
0.0587 0.1466 0.3079 0.1232

0.0235 0.0587 0.1232 0.2493
57
Thus
A1 =
0.2493
0.1232
0.0587
0.0235
0.1232
0.3079
0.1466
0.0587
0.0587
0.0235
0.1466
0.0587
0.3079
0.1232
0.1232 0.2493
26. Find the inverse of the Hilbert matrix of size 44 using the Gauss-Jordan method.
Then solve the linear system Ax = [1, 2, 3, 4]T .
Solution: Since the Hilbert matrix of size 4 4 is
A=
1
0.5
0.3333
0.25
0.5
0.3333
0.25
0.2
0.3333
0.25
0.2
0.1667
0.25
0.2
0.1667
0.1429
Consider the augmented matrix form
..
. 1 0 0 0
1 0.5 0.33 0.25
0.5 0.33 0.25

0.2
..
.
0 1 0 0
0.33 0.25
0.2 0.17
..
.
0 0 1 0
.
0.25
.. 0 0 0 1
0.2 0.17 0.14
1 0.50 0.33 0.25
0 0.08 0.08 0.08
0.33 0.25 0.20 0.17
0.25 0.20 0.17 0.14
1 0.50 0.33 0.25
0 0.08 0.08 0.08

0 0.08 0.09 0.08
0.25 0.20 0.17 0.14
1 0.50 0.33 0.25
0 0.08 0.08 0.08
0 0.08 0.09 0.08
0 0.08 0.08 0.08
A.BENHARI
0 0.17 0.20
0 0.08
0.90
0.09
0.08
0 0.08
0.08
0.08
..
.
1 0 0
..
. 0.50 1 0
..
.
0 0 1
..
.
0 0 0
..
.
1 0 0
..
. 0.50 1 0
..
. 0.33 0 1
..
.
0 0 0
..
.
1 0 0
..
. 0.50 1 0
..
. 0.33 0 1
..
. 0.25 0 0
..
.
4 6
..
.
6 12
..
. 0.33
0
..
. 0.25
0
58
0
1
0 0
0 0
1 0
0 1
0 0.17
1
0
1
1
0 0.01
0.01
0 0.08
0.20
0.90
0.01
0.08
0.08
1 0 0.17 0.20
0 1
1
0.90
0 0
0.01
0.01
0 0
0.01
0.01
1 0
0 1
0 0
0 0.05
1 0.90
1 1.50
0 0 0.01 0.01
..
.
4.00
6
..
. 6.00
12
..
.
0.17
1
..
.
0.20 0.90
0 0
0 0
1 0
0 1
..
.
9
36 30 0
..
. 6
12
0 0
..
.
30 180 180 0
..
. 0.20 0.90
0 1
1 0
0 1
0 0
0.05
0 0.60
1
1.50
0 0 0.01
0.01
0.05
1 0 0
0 1 0 0.60
0 0 1
1.50
0 0 0
0.00
..
.
9
36
30 0
..
. 36
192 180 0
..
.
30 180
180 0
..
. 0.20 0.90
0 1
..
.
9 36
30
..
.
36
192 180
..
.
30 180
180
..
. 0.05 0.60 1.50
0
1
1 0 0
0 1 0
0 0 1
0
0.60
0 0 0
1.00
1.50
..
.
16 120
240 140
..
. 36
192 180
0
..
.
30 180
180
0
..
. 140 1680 4200 2800
1 0
0 1
0 0
1 1.50
0 0 0
..
.
16 120
240 140
..
. 120 1200 2700 1680
..
.
30 180
180
0
..
. 140 1680 4200 2800
1 0 0 0
0 1 0 0
0 0 1 0
0 0 0 1
A.BENHARI
..
.
4 6 0 0
..
.
6 12 0 0
..
.
0.17 1 1 0
..
. 0.25
0 0 1
..
.
16 120
240 140
..
. 120
1200 2700
1680
..
.
240 2700
6480 4200
..
. 140
1680 4200
2800
59
Thus
A1 =
16 120
240 140
120
1200 2700
1680
240 2700
6480 4200
140
1680 4200
2800
Then solving x = A1 b, we get

x = [64, 900, 2520, 1820]T
27. Find the LU decomposition of each matrix A using the Doolittles method, and
then solve the systems.
(a)
2 1
1
4
4 1
A = 3
b = 5 .
,
1 1
1
6
(b)
(c)
b = 4 .
1
2 2 2
A = 1 2 1 ,
3 3 4
(d)
2 4 6
3
A= 1 5
,
1 3
2
(e)
b = 10 .
5
1 1
0
1
A = 2 1
,
2 2 1
(f )
b= 1
.
2
7 6 5
A= 5 4 3
,
3 7 6
1 5 3
A = 2 4 6 ,
1 3 2
b= 4
.
3
b = 11
5
Solution: (a) The factorization of A is
1
0 0
2 1
1
1 0 0 2.5 0.5
LU = 1.5
0.5 0.2 1
0
0 0.6
Now solving the system Ly = b using forward substitution, we get
y = [4.000, 11.000, 6.2000]T
A.BENHARI
60
Again solving the system U x = y using backward substitution, we get

x = [2.00, 2.333, 10.333]T
(b) The factorization of A is
1.0000
0
0
7.0000
6.0000
5.0000
1.0000
0
0 0.2857 0.5714
LU = 0.7143
0.4286 15.5000 1.0000

0
0 5.0000
y = [2.000, 0.4286, 5.5000]T
x = [0.100, 0.700, 1.1000]T
(c) The factorization of A is
1 0 0
2 2 2
LU = 0.5 1 0 0 1 0
1.5 0 1
0 0 1
y = [0, 4, 1]T
x = [3, 4, 1]T
(d) The factorization of A is
1.0
0
0
2 4 6
6
LU = 0.5000 1.0000 0 0 3
0.5000 0.3333 1.0

0 0
3
y = [4, 12, 3]T
x = [3, 2, 1]T
(e) The factorization of A is
1 1
0
1 0 0
1
1
LU =
2 1 0 0
0
0 1
2 0 1
A.BENHARI
61

y = [2, 0, 1]T
x = [1, 1, 1]T
(f ) The factorization of A is
1.0 0
0
1
5
3
0 0 6
0
LU = 2.0 1.0
1.0 0.3333 1.0

0
0 1
y = [4, 3, 0]T
x = [6.50, 0.50, 0]T
28. Solve the Problem 27 by the LU decomposition using the Crouts method.
2
0
0
1 0.5 0.5
2.5
0 0
1 0.2
LU = 3
1 0.5 0.6
0
0
1
y = [2, 4.4, 10.3333]T ,
x = [2, 2.3333, 10.3333]T
2 0 0
1 1 1
LU = 1 1 0 0 1 0
3 0 1
0 0 1
y = [0, 4, 1]T ,
x = [3, 4, 1]T
(e) The factorization of A is
1 0
0
1 1 0
0 0
1 1
LU = 2 1
2 0 1
0
0 1
y = [2, 0, 1]T ,
A.BENHARI
x = [1, 1, 1]T
62
29. Solve the following system by the LU decomposition using the Cholesky method.
(a)
1
2 1
1
2 1
b=
A=
,
2 .
1
3
1 1
2
(b)
(c)
5 2 3
A = 2 4 1 ,
3 1 6
(d)
A=
3
4
6
0
b = 4 .
3
10 2 1
A = 2 10 3 ,
1 3 10
b = 1 .
1
4 6 0
5
3 1
,
3
3 1
1
1 3
b=
4
5
2
3
Solution: (a) The lower-triangular matrix L is
1.414
0
0
1.223
0
L = 0.707
0.707 0.408 1.155

y = [0.7071, 2.0412, 2.8868]T
Again solving the system LT x = y using backward substitution, we get
x = [0.5, 2.5, 2.5]T
(b) The lower-triangular matrix L is
3.1623 0
0
L = 0.6325 3.0984 0
0.3162 0.9037 3.0139

y = [2.2136, 1.7428, 1.2857]T
x = [0.7947, 0.6869, 0.4266]T
(c) The lower-triangular matrix L is
2.236 0
0
0
L=
0.894 1.789
1.342 0.112 2.046
A.BENHARI
63

y = [0.4472, 0.3354, 0.2138]T
x = [0.0597, 0.1940, 0.1045]T
(d) The lower-triangular matrix L is
L=
5.5678
0
0
0
0.7184 2.1175
0
0
1.0776 1.7824 2.9431
0
0 0.4723 0.0538 1.6656

y = [0.7184, 2.1175, 0.3398, 1.2118]T
x = [0.0180, 0.9461, 0.1287, 0.7275]T
30. Solve the following tridiagonal systems using the LU decomposition.
(a)
3 1
0
1
3 1
A = 1
b = 2 .
,
0 1
3
3
(b)
A=
(c)
A=
2
3
0
0
3
2
3
0
0
3
2
3
0
0
3
2
4 1
0
0
1
4 1
0
0 1
4 1
0
0 1
4
b=
6
7
5
3
b=
3
1
2
5
1
0 0
3
1
0
1 0 0 2.6667
1
LU = 0.3333
0 0.375 1
0
0 2.6250
y = [1, 2.3333, 3.8750]T ,
x = [0.8095, 1.4286, 1.4762]T
(b) The factorization of A is
LU =
A.BENHARI
2
3
0
0
0 2.5
3
0
0
0 5.6
3
0
0
0 0.3929
64
1
0
0
1.5
1
0
0 1.2
1
0
0 0.5357
0
0
0
1
y = [6.0, 2.0, 2.60, 1.6071]T ,
x = [4.9091, 1.2727, 1.7273, 4.0909]T
LU =
1
0
0
0.25
1
0
0 0.2667
1
0
0 0.2679
y = [3, 0.25, 1.9333, 5.5179]T ,
0
0
0
1
4 1
0
0
0 3.75
1
0
0
0 3.7333
1
0
0
0 3.7321
x = [0.7943, 0.1770, 0.9139, 1.4785]T
31. Find kxk1 , kxk2 , and kxk for the following vectors:
(a) [2, 1, 6, 3]T , (b) [sin k, cos k, 3k ]T for a fixed integer k, (c) [3, 4, 0, 3/2]T .
Solution: (a)
kxk1 = 12,
kxk2 = 7.0711,
kxk = 6
kxk1 = 8.5,
kxk2 = 5.2202,
kxk = 4
(c)
32. Find k.k1 , k.k and k.ke for the following matrices.
3
1 1
0
4
A= 2
,
1 5
1
4 1 6
B = 3 6 4 ,
5 0 9
17 46 7
C = 20 49 8 .
23 52 9
Solution:
kAk1 = 6,
kAk = 7,
kAke = 7.6158
kBk1 = 19,
kBk = 14,
kBke = 14.8324
kCk1 = 147,
kCk = 84,
kCke = 92.9139
33. Consider the following matrices
11 7 8
5 9
6
A=
,
6 3
7
6
2 7
10 8
B = 12
,
3 15 14
Find k.k1 and k.k for (a) A3 , (b) A2 + B 2 , (c) BC and (d) C 2 + A2
Solution: (a)
kA3 k1 = 2996,
kA3 k = 23
(b)
kA2 + B 2 k1 = 468,
kA2 + B 2 k = 524
(c)
kBCk1 = 427,
kBCk = 531
(d)
kC 2 + A2 k1 = 334,
A.BENHARI
kC 2 + A2 k = 328
65
5 6 4
8 5
C = 7
.
3 9 12
34. The n n Hilbert matrix H (n) defined by

(n)
Hij =
1
,
i+j1
1 i, j n
Find the l -norm of the 10 10 Hilbert matrix.

Solution: The 10 10 Hilbert matrix is
H=
1
1/2
1/3
1/4
1/5
1/6
1/7
1/8
1/9
1/10
1/2
1/3
1/4
1/5
1/6
1/7
1/8
1/9
1/10
1/11
1/3
1/4
1/5
1/6
1/7
1/8
1/9
1/10
1/11
1/12
1/4
1/5
1/6
1/7
1/8
1/9
1/10
1/11
1/12
1/13
1/5
1/6
1/7
1/8
1/9
1/10
1/11
1/12
1/13
1/14
1/6
1/7
1/8
1/9
1/10
1/11
1/12
1/13
1/14
1/15
1/7
1/8
1/9
1/10
1/11
1/12
1/13
1/14
1/15
1/16
1/8
1/9
1/10
1/11
1/12
1/13
1/14
1/15
1/16
1/17
1/9
1/10
1/11
1/12
1/13
1/14
1/15
1/16
1/17
1/18
1/10
1/11
1/12
1/13
1/14
1/15
1/16
1/17
1/18
1/19
and its l -norm is

kHk = 2.9290
35. Solve the following linear systems using the Jacobi method, start with initial
approximation x(0) = 0 and iterate until kx(k+1) x(k) k 105 for each system.
(a)
4x1 x2 + x3 =
7
4x1 8x2 + x3 = 21
2x1 + x2 + 5x3 =
15
(b)
3x1 + x2 + x3 = 5
2x1 + 6x2 + x3 = 9
x1 + x2 + 4x3 = 6
(c)
4x1 + 2x2 +
x3 = 1
x1 + 7x2 +
x3 = 4
x1 + x2 + 20x3 = 7
(d)
5x1 + 2x2 x3 = 6
x1 + 6x2 3x3 = 4
2x1 + x2 + 4x3 = 7
(e)
6x1
2x1
x2 + 3x3 = 2
3x2 + x3 =
1
+ x2 + 5x3 =
5
(f )
4x1 + x2
= 1
2x1 + 5x2 + x3 =
0
x1 + 2x2 + 4x3 =
3
A.BENHARI
66
(g)
1
x 2 + x3 =
3x2 x3 = 1
+ 2x2 + 4x3 =
2
5x1
x1
(h)
9x1 +
x2 +
x3 = 10
2x1 + 10x2 + 3x3 = 19
3x1 + 4x2 + 11x3 = 0
Solution: (a) The approximations of the given system are as follows Thus
k
0
1
2
3
4
5
6
7
8
9
10
11
12
13
(k)
x1
0
1.75000
1.65625
1.92500
1.99063
1.98711
1.99719
1.99965
1.99952
1.99990
1.99999
1.99998
2.00000
2.00000
(k)
(k)
x2
0
2.62500
3.87500
3.85000
3.94844
3.99531
3.99438
3.99807
3.99982
3.99979
3.99993
3.99999
3.99999
4.00000
x3
0
3.00000
3.17500
2.88750
3.00000
3.00656
2.99578
3.00000
3.00025
2.99984
3.00000
3.00001
2.99999
3.00000
x(13) = [2, 4, 3]T .

(c) The approximations of the given system are as follows Thus
x(12) = [0.10156, 0.53906, 0.32813]T .
(e) The approximations of the given system are as follows Thus
x(14) = [0.67308, 0.09616, 0.71154]T
(g) The approximations of the given system are as follows Thus
x(12) = [0.05882, 0.14706, 0.55883]T .
36. Consider the following system of equations
4x1 + 2x2 +
x3 = 1
x1 + 7x2 +
x3 = 4
x1 + x2 + 20x3 = 7
A.BENHARI
67
k
0
1
2
3
4
5
6
7
8
9
10
11
12
(k)
x1
0
0.25000
-0.12321
-0.07009
-0.10542
-0.09858
-0.10208
-0.10127
-0.10163
-0.10153
-0.10157
-0.10156
-0.10156
(k)
x2
0
0.57143
0.48571
0.54490
0.53403
0.53988
0.53857
0.53916
0.53901
0.53907
0.53906
0.53906
0.53906
(k)
x3
0
0.35000
0.30893
0.33188
0.32626
0.32857
0.32794
0.32818
0.32811
0.32813
0.32812
0.32813
0.32813
(a) Show that the Jacobi method converges by using kTJ k < 1.
(b) Compute 2nd approximation x(2) , starting with x(0) = [0, 0, 0]T .
(c) Compute an error estimate kx x(2) k for your approximation.
Solution: (a) The Jacobi iterative matrix is
0 1/2 1/4
0 1/7
TJ = D1 (L + U ) =
1/7
1/20 1/20
0
and its l -norm is
kTJ k = 3/4 = 0.75
(b) Using the Jacobi method, we have
x(1) = [0.25, 0.5714, 0.35]T , x(2) = [0.1232, 0.4857, 0.3089]T
(c) To compute the error bound for the approximation, we have
kx x(2) k
(0.75)2
(0.5714) 1.2857
1 0.75
37. Solve the Problem 35 using the Gauss-Seidel method.

Solution: Using the Gauss-Seidel method for each system, we get
(a) x(8) = [2, 4, 3]T
(c) x(6) = [0.102, 0.539, 0.328]T
(e) x(8) = [0.673, 0.096, 0.711]T
(g) x(7) = [0.588, 0.147, 0.559]T
A.BENHARI
68
k
0
1
2
3
4
5
6
7
8
9
10
11
12
13
14
(k)
x1
0
-0.33333
-0.77778
-0.73333
-0.66667
-0.66272
-0.67259
-0.67460
-0.67337
-0.67289
-0.67300
-0.67310
-0.67309
-0.67307
-0.67308
(k)
x2
0
0.33333
0.00000
0.06667
0.10370
0.10222
0.09580
0.09518
0.09607
0.09629
0.09619
0.09614
0.09615
0.09616
0.09616
(k)
x3
0
1.00000
0.80000
0.68889
0.69333
0.71259
0.71447
0.71180
0.71112
0.71144
0.71159
0.71156
0.71153
0.71153
0.71154
38. Consider the following system of equations

4x1 + 2x2 + x3 = 11
x1 + 2x2
= 3
2x1 + x2 + 4x3 = 16
(a) Show that the Gauss-Seidel method converges by using kTG k < 1.
(b) Compute the second approximation x(2) , starting with x(0) = [1, 1, 1]T .
(c) Compute an error estimate kx x(2) k for your approximation.
Solution: (a) The Gauss-Seidel iterative matrix is
0 1/2 1/4
1
TG = (L + D) U = 0 1/4 1/8
0 5/16 5/32
and its l -norm is
kTG k = 3/4 = 0.75
(b) Using the Gauss-Seidel method, we have
x(1) = [2.75, 2.875, 1.906]T , x(2) = [0.836, 1.918, 3.103]T
(c) To compute the error bound for the approximation, we have
kx x(2) k
(0.75)2
(2.875) 6.469
1 0.75
39. Which of the following matrix is convergent ?
(a)
A.BENHARI
2
2 3
2 1
,
1
2 2 1
(b)
69
1 0
0
1 3
3 2 2
k
0
1
2
3
4
5
6
7
8
9
10
11
12
(k)
(k)
x1
0
0.20000
0.03333
0.04333
0.05944
0.06106
0.05901
0.05855
0.05877
0.05885
0.05884
0.05882
0.05882
x2
0
-0.33333
-0.16667
-0.12778
-0.14167
-0.14898
-0.14801
-0.14692
-0.14692
-0.14706
-0.14708
-0.14706
-0.14706
(k)
x3
0
0.50000
0.61667
0.57500
0.55306
0.55597
0.55923
0.55925
0.55882
0.55877
0.55882
0.55883
0.55883
Solution: (a) Solving the following equation to find the eigenvalues of the given
matrix
|A I| = 3 + 52 2 8 = ( 2)( 4)( + 1) = 0
gives
(A) = 4 > 1
which shows that A is divergent.
(b) Solving the following equation to find the eigenvalues of the given matrix
|A I| = (1 )(3 )(2 ) = 0
gives
(A) = max |i | = 3 > 1
i
which shows that A is divergent.

40. Find the eigenvalues and their associated eigenvectors of the matrix
2 2
3
3 2
A= 0
0 1
2
Also, show that kAk2 > (A).
Solution: Solving the following equation to find the eigenvalues of the given matrix
|A I| = 3 + 72 14 + 8 = ( 2)( 1)( 4) = 0
gives
1 = 4,
A.BENHARI
2 = 2,
70
3 = 1
the eigenvalues of the matrix. The corresponding eigenvectors are

x1 = [1 1 1/2]T ,
Also
x2 = [1 0 0]T ,
x3 = [1 1 1]T
4 4
6
T
14 14
A A = 4
6 14
17
and the characteristic equation of AT A is

|AT A I| = 3 + 352 114 + 64 = 0
which gives eigenvalues of AT A
1 = 31.4386,
Thus
kAk2 =
(AT A) =
2 = 2.8461,
3 = 0.7153
31.4386 = 5.6070 > 4 = (A)
41. Solve the Problem 35 using the SOR method by taking = 1.02 for each system.
Solution: Using SOR method by taking = 1.02 within accuracy 106 for the
following systems, we have:
(a) x(8) = [2, 4, 3]T
(c) x(6) = [0.102, 0.539, 0.328]T
(e) x(8) = [0.673, 0.096, 0.711]T
(g) x(8) = [0.588, 0.147, 0.559]T
42. Use the parameter = 1.543 to solve the linear system by the SOR method within
accuracy 106 in the l -norm, starting x(0) = 0
x1 2x2
2x1 + 5x2
x3
x2 +
2x3 0.5x4
0.5x3 + 1.25x4
= 3
=
5
=
2
= 3.5
Solution: Using SOR method within accuracy 106 , we need 33 iterations and the
33th iteration gives
x(33) = [0.999999, 2.000000, 3.000000, 4.000000]T
43. Find the optimal choice for and use it to solve the linear system by the SOR
method within accuracy 104 in the l -norm, starting x(0) = 0. Also, find how
many iterations need by using the Jacobi method and the Gauss-Seidel method.
A.BENHARI
71
Since
we have
0 1/2
0
1
0 1/2
TJ = D (L + U ) = 1/2
0
1/4
0
1/2
0
1/2
TJ I = 1/2
0
1/4
so
det(TJ I) = ( 3/8)
Thus
3/8 = 0.6124
(TJ ) =
and
=
1+
1 [(TJ )]2
1.117
Using SOR method, we have Thus using SOR method we need 8 iterations
k
0
1
2
3
4
5
6
7
8
(k)
x1
0
1.3963
1.2005
0.6385
0.5679
0.5523
0.5504
0.5500
0.5500
(k)
x2
0
0.0579
1.1053
1.3494
1.3922
1.3989
1.3998
1.4000
1.4000
(k)
x3
0
1.6917
1.7862
1.8433
1.8486
1.8498
1.8500
1.8500
1.8500
x(8) = [0.5500, 1.4000, 1.8500]T

and using Gauss-Seidel method we need 12 iterations
x(12) = [0.5500, 1.4000, 1.8500]T
and using Jacobi method we need 20 iterations
x(20) = [0.5500, 1.3900, 1.8500]T
44. Consider the following system
4x1 2x2 x3
x1 + 4x2
x4
x1
+ 4x3 x4
x2 x3 + 4x4
A.BENHARI
72
=
=
=
=
1
2
0
1
Using x(0) = 0, how many iterations are required to approximate the solution to
within five decimal places using: (a) Jacobi method, (b) Gauss-Seidel method,
(c) SOR method (take = 1.1).
Solution: (a) Using Jacobi method within five decimal places, we need 19 iterations and get
x(19) = [0.73862, 0.81817, 0.31817, 0.53408]T
(b) Using Gauss-Seidel method for same accuracy, we need 12 iterations and get
x(12) = [0.73864, 0.81818, 0.31818, 0.53409]T
(c) Using SOR method with = 1.1, we need only 7 iterations and get
x(7) = [0.73864, 0.81818, 0.31818, 0.53409]T
45. Find the spectral radius of the Jacobi, the Gauss-Seidel and the SOR ( = 1.25962)
iteration matrices when
2 1 0 0
1 2 1 0
A=
0 1 2 1
0 0 1 2
Solution: The spectral radius of the three iteration matrices are as follows
(TJ ) = 0.809017,
(TG ) = 0.654508,
46. Perform only two steps of the conjugate

system, taking x(0) = 0
(a)
3x1 x2
x1 + 3x2
x1 + 2x2
(T ) = 0.25962
gradient method for the following linear
+ x3 = 7
+ 2x3 = 1
+ 5x3 = 5
(b)
3x1 2x2 + x3 = 5
2x1 + 6x2 x3 = 9
x1 x2 + 4x3 = 6
(c)
4x1 2x2 +
x3 = 1
2x1 + 7x2 +
x3 = 4
x1 + x2 + 20x3 = 1
(d)
5x1 3x2 x3 = 6
3x1 + 6x2 3x3 = 4
x1 3x2 + 4x3 = 7
Solution: (a) The first two steps are as follows
x(1) = [1.4957, 0.2137, 1.0683]T , and x(2) = [2.2261, 0.1602, 0.5408]T
A.BENHARI
73
(b) The first two steps are as follows

x(1) = [1.4885, 2.6793, 1.7862]T , and x(2) = [2.4796, 2.4959, 1.9858]T
(c) The first two steps are as follows
x(1) = [0.1385, 0.5539, 0.1385]T , and x(2) = [0.3010, 0.7339, 0.0365]T
(d) The first two steps are as follows
x(1) = [7.9737, 5.3158, 9.3026]T , and x(2) = [16.8605, 19.4381, 0.7106]T
47. Perform only two steps of the
system, taking x(0) = 0
(a)
6x1
2x1
x1
conjugate gradient method for the following linear
+ 2x2 + x3 =
1
+ 3x2 x3 =
0
x2 + 2x3 = 2
(b)
5x1 2x2 + x3 = 3
2x1 + 4x2 x3 = 2
x1 x2 + 3x3 = 1
(c)
6x1
x1
x1
5x1
x2
+ 7x2
+ x2
x2
x3
+ x3
+ 3x3
3x3
+ 5x4
x4
3x4
+ 6x4
=
1
=
2
=
0
= 1
(d)
3x1
2x1
x1
3x1
2x2
+ 7x2
+ x2
x2
x3
+ x3
+ 3x3
3x3
+ 3x4
x4
3x4
+ 6x4
=
=
=
=
1
0
0
0
Solution: (a) The first two steps are as follows

x(1) = [0.5000, 0.0000, 1.0000]T and x(2) = [0.5181, 0.7236, 1.9430]T
(b) The first two steps are as follows
x(1) = [1.0000, 0.6667, 0.3333]T and x(2) = [0.9701, 1.0659, 0.3234]T
(c) The first two steps are as follows
x(1) = [0.2000, 0.4000, 0, 0.2000]T and x(2) = [0.9065, 0.4673, 0.3365, 0.5701]T
(d) The first two steps are as follows
x(1) = [0.3333, 0, 0, 0]T and x(2) = [0.7391, 0.1739, 0.0870, 0.2609]T
A.BENHARI
74
48. Compute the condition numbers of each of the matrix in the Problem 47 relative
to k.k .
Solution: (a) Since the inverse of the given matrix is
A1
0.3333 0.3333 0.3333
0.7333
0.5333
= 0.3333
0.3333
0.5333
0.9333
The condition number of the matrix is

K(A) = kAk kA1 k = (9.0)(1.8000) = 16.2000
(b) Since the inverse of the given matrix is
A1
0.2558 0.1163 0.0465
0.0698
= 0.1163 0.3256
0.0465 0.0698
0.3721

K(A) = kAk kA1 k = (8.0)(0.5116) = 4.0930
(c) Since the inverse of the given matrix is
A1 =
3.7500
0.3750 3.8750 5.0000
0.3750
0.1875 0.4375 0.5000
3.8750 0.4375
4.6875
5.5000
5.0000 0.5000
5.5000
7.0000

K(A) = kAk kA1 k = (15.0)(18.0000) = 270.0000
(d) Since the inverse of the given matrix is
A1 =
1.0909
0.2727 0.4545 0.7273
0.2727
0.2182 0.1636 0.1818
0.4545 0.1636
0.8727
0.6364
0.7273 0.1818
0.6364
0.8182

K(A) = kAk kA1 k = (13.0)(2.5455) = 33.0909
49. Compute the condition numbers of the following matrices relative to k.k
(a)
A.BENHARI
1
2
1
5
1
3
, (b)
3
0.03 0.01 0.02
0.15 0.51 0.11 , (c)

1.11 2.22
3.33
75
1.11 1.98 2.01
1.01 1.05 2.05

0.85 0.45 1.25
Solution: (a) Since the inverse of the given matrix is
300/1891 4950/1891 2069/1144
300/1891
4950/1891 1351/747
2069/1144
300/1891
4950/1891
K(A) = (31/30)(2971/648) = 289/61 = 4.7377
(b) Since the inverse of the given matrix is
34.2466 1.3699 0.1604
2.1526 0.0053
10.9589
4.1096 0.9785 0.2433

K(A) = (6.6600)(35.7769) = 238.2740
(c) Since the inverse of the given matrix is
0.7755 3.1228
3.8744
0.9544 0.6383 0.4880

0.8709
2.3533 1.6589
K(A) = (5.1000)(7.7726) = 39.6403
50. The following linear systems have x as the exact solution and x is an approximate
kr||
, where r = b Ax is the residual
solution. Compute kx x k and K(A)
kbk
vector. (a)
0.89x1 + 0.53x2 = 0.36
0.47x1 + 0.28x2 = 0.19
x = [1, 1]T
x = [0.702, 0.500]T
(b)
0.986x1 + 0.579x2 = 0.235
0.409x1 + 0.237x2 = 0.107
x = [2, 3]T
x = [2.110, 3.170]T
(c)
1.003x1 + 58.090x2 = 68.12
5.550x1 + 321.8x2 = 377.3
x = [10, 1]T
x = [10, 1]T
A.BENHARI
76
Solution: (a) Since

r = b Ax = [0.00022, 0.00006]T
and
K(A) = 19312.0
therefore,
kx x k = 0.5
and
K(A)
kr||
= (19312)(0.00022)/(0.36) = 11.8018
kbk
(b) Since
r = b Ax = [0.1003, 0.0047]T
and
K(A) = 697.7229
therefore,
kx x k = 0.1700
and
K(A)
kr||
= (697.7229)(0.1003)/(0.235) = 297.7941
kbk
(c) Since
r = b Ax = [20.0600, 111.0000]T
and
K(A) = 339866.0604
therefore,
kx x k = 20.0
and
K(A)
kr||
= (339866.0604)(111.00)/(377.3) = 99987.0997
kbk
51. Discuss the ill-conditioning (stability) of the linear system

1.01x1 + 0.99x2 = 2
0.99x1 + 1.01x2 = 2
If x = [2, 0]T be an approximate solution of the system, then find the residual
vector r and estimate the relative error.
Solution: Since the inverse of the given matrix A is
25.2500 24.7500
24.7500
25.2500
and the l -norms of A and A1 are

kAk = 2 and kA1 k = 50
A.BENHARI
77
Thus
K(A) = (2)(50) = 100 >> 1 (Ill Conditioned)
The residual vector is
r = b Ax = [0.02, 0.02]T
Rel. Error
(100)(0.02)
=1
2
53. Show that if B is singular, then

kA Bk
1
K(A)
kAk
Solution: Since
kxk = kA1 Axk kA1 kkAxk
which gives
kAxk
kxk
kA1 k
With x 6= 0 such that kxk = 1 and Bx = 0, gives

kAx Bxk = k(A B)xk = kAxk
Thus
kxk
kA1 k
1
k(A B)xk
kAk
kAkkA1 k
or
k(A B)xk
1
K(A)
kAk
Since
k(A B)xk kA Bkkxk = kA Bk
therefore,
kA Bk
1
K(A)
kAk
53. Consider the following matrices
0.06 0.01 0.02
0.13
0.05 0.11
A=
,
1.01 2.02 3.03
0.04 0.01 0.05
B = 0.11 0.02 0.03 .

0.89 1.94
2.99
Using the Problem 52, compute the approximation of the condition number of the
matrix A relative to k.k .
Solution: Since kAk = 6.06 and the norm of the following matrix
0.1000 0
0.0700
AB =
0.0200 0.0300 0.1400
0.1200 0.0800 0.0400
A.BENHARI
78
is 0.24. The from the Problem 52, we get

K(A)
6.06
= 25.25
0.24
54. Let A and B are nonsingular n n matrices. Show that

K(AB) K(A)K(B)
Solution: By using definition of the condition number of matrix, we have
K(AB) = kABkk(AB)1 k
Since (AB)1 = B 1 A1 and the norm can be written as
k(AB)1 k = kB 1 A1 k kA1 kkB 1 k
Also
kABk kAkkBk
Thus using these two inequalities, we get the required result.
55. The exact solution of the following linear system
x1 + x2
= 1
x1 + 1.01x2 = 2
is x = [99, 100]T . Change the coefficient matrix slightly to
A =
1 1
1 0.99
and consider the linear system

x1 + x2
= 1
x1 + 0.99x2 = 2
Compute the change solution x of this system. Is the matrix A ill-conditioned ?
Solution: Solving the linear system using simple Gaussian elimination method,
we get
x = [101, 100]T
and the condition number of the matrix is
K(A) = 404.01
which shows that A is ill-Conditioned.
56. Using the Problem 15, compute the relative error and the relative residual.
Solution: (a)
Rel.Error = 0,
Rel.Residual = 0
Rel.Error = 0,
Rel.Residual = 0
(c)
A.BENHARI
79

x1
+ 3x2 = 4
1.0001x1 + 3x2 = 4.0001
is x = [1, 1]T . Change the right-hand vector b slightly to b = [4.0001, 4.0003]T
x1
+ 3x2 = 4.0001
1.0001x1 + 3x2 = 4.0003
Find the change solution x of this system. Is the matrix A ill-conditioned ?
we get
x = [1.9976, 0.6675]T
K(A) = 80002
58. If kAk < 1, then show that the matrix (I A) is nonsingular and
k(I A)1 k
1
1 kAk
Solution: Since kAk < 1 and using matrix geometric series, we have
1
= I + A + A2 + A3 +
1
IA
Taking norm and applying the triangle inequality, we get
k(I A)1 k 1 + kAk + kAk2 + kAk3 + =
1
(1 kAk)

x1 + x2
= 3
x1 + 1.0005x2 = 3.0010
is x = [1, 2]T . Change the coefficient matrix and the right-hand vector b slightly
to
!
1 1
2.99
A =
and b =
1 1.001
3.01
x1 + x2
= 2.99
x1 + 1.001x2 = 3.01
A.BENHARI
80
Find the change solution x of this system. Is the matrix A ill-conditioned ?

we get
x = [17, 20]T
K(A) = 8004
60. Find the condition number of the following matrix
1 1
1
n
T
Solve the linear system A4 x = [2, 2] and compute the relative residual.
An =
1 1
Solution: Since the inverse of the given matrix for n > 0, we have
A1
n
n + 1 n
n
n
Then condition number of the given matrix is

K(A) = kAn kkA1
n k = 4n,
n>0
Solving the linear system A4 x = [2, 2]T , we have

x = [2, 0]T
To compute the relative residual, we first compute the residual vector as
r = b Ax =
Thus the relative residual is
0
0
krk
=0
kbk
61. The following linear system has exact solution x = [10, 1]T , find the approximate
solution of the system
0.03x1 + 58.9x2 = 59.2
5.31x1 6.10x2 = 47.0
by using the simple Gaussian elimination and then use the residual correction
method (one iteration only) to improve the approximate solution.
Solution: Using simple Gaussian elimination method, we get
x(1) = [10, 1]T
Now using residual corrector method, we got
x(2) = x(1) + y(1) = [10, 1]T + [0, 0]T = [10, 1]T
A.BENHARI
81
Chapter 4
Approximating Functions
1. Find the second-degree Taylor polynomial for the function f (x) = (ex + 2)1/2 expanded about x0 = 0. Then use it to approximate f (1).
Solution: The second degree Taylor polynomial expanded about x0 is
(x x0 ) 0
(x x0 )2 00
f (x0 ) +
f (x0 )
1!
2!
Given f (x) = (ex + 2)1/2 , then f (x0 ) = f (0) = 3. Now calculating the derivatives
require for the desired polynomial p2 (x), we get
p2 (x) = f (x0 ) +
f 0 (x) = ex /2(ex + 2)1/2 ,

f 00 (x) = (ex )2 /4(ex + 2)3/2 + ex /2(ex + 2)1/2 ,
f 0 (x0 ) = f 0 (0) = 0.2887

f 00 (x0 ) = f 00 (0) = 0.2406
using these values, we have

p2 (x) = 1.7321 + 0.2887x + 0.1203x2
and the approximation of f (x) at x = 1 is
f (1) p2 (1) = 1.7321 + 0.2887(1) + 0.1203(1)2 = 2.1411
2. Use the Taylor polynomial of third-degree about x0 = 1 to approximate the function
f (x) = x2 + 4ex at x = 1 and x = 2. Find the error bounds for these approximations and compare your results.
Solution: The third-degree Taylor polynomial expanded about x0 is
(x x0 ) 0
(x x0 )2 00
(x x0 )3 000
f (x0 ) +
f (x0 ) +
f (x0 )
p3 (x) = f (x0 ) +
1!
2!
3!
Since f (x) = x2 + 4ex , and the derivatives require for the desired polynomial p3 (x)
can be obtained as
f 0 (x) = 2x + 4ex ,
f 00 (x) = 2 + 4ex ,
f 000 (x) = 4ex ,
f 0 (x0 ) = f 0 (1) = 2 + 4e1

f 00 (x0 ) = f 00 (1) = 2 + 4e1
f 000 (x0 ) = f 000 (1) = 4e1
using all these values, we have

p3 (x) = 1 + 4e1 + (x 1)(2 + 4e1 ) + (x 1)2 (1 + 2e1 ) + (x 1)3 (2/3e1 )
Now take x = 1, we get
f (1) p3 (1) = 1 + 4e1 = 11.8731
The error bound for the approximation of f (1) is
|R4 (1)| = 0
A.BENHARI
82
Note that the exact value of f (1) is 11.8731.

Similarly, at x = 2, we have
f (2) p3 (2) = 1 + 4e1 + (2 + 4e1 ) + (1 + 2e1 ) + 2/3e1 = 32.9950
To compute error bound for the approximation of f (2), we have
|R4 (x)|
1
|x x0 |4 M
24
where
M=
max |4ex | = 29.5562
1.0x2.0
at x = 2.0 and |f (4) ((x))| M . Hence

|R4 (x)|
1
(29.5562) = 1.2315
24
Note that the exact value of f (2) is 33.5562.

3. Find the third-degree Taylor polynomial for the function f (x) = ln(x2 + 2x + 2)
expanded about x0 = 1. Use this polynomial to approximate f (1.1). Compute the
actual error of the problem.
Solution: The third degree Taylor polynomial expanded about x0 = 1 is
p3 (x) = f (1) +
(x 1)2 00
(x 1)3 000
(x 1) 0
f (1) +
f (1) +
f (1)
1!
2!
3!
Since f (x) = ln(x2 + 2x + 2), and the derivatives require for the desired polynomial
p3 (x) can be obtained as
f 0 (x)
00
f (x)
x2
2x + 2
,
+ 2x + 2
f 0 (x0 ) = f 0 (1) = 0.8000
(2x + 2)2
2
,
= 2
x + 2x + 2 (x2 + 2x + 2)2
f 000 (x) =
2(2x + 2)3
6(2x + 2)
+
,
(x2 + 2x + 2)2 (x2 + 2x + 2)3
f 00 (1) = 0.2400
f 000 (1) = 0.0640

f (1.1) p3 (1.1) = 1.6094 + 0.8(1.1 1) 0.12(1.1 1)2 + 0.0107(1.1 1)3 = 1.6882
The actual error is
Error = f (1.1) p3 (1.1) = 1.6882 1.6882 = 0.0000
A.BENHARI
83
4. Find the fourth-degree Taylor polynomial for the function f (x) = (x3 + 1)1 expanded about x0 = 0, and use it to approximate f (0.2). Also, find a bound for the
error in this approximation.
Solution: The fourth degree Taylor polynomial expanded about x0 = 0 is
p4 (x) = f (0) + xf 0 (0) +
x3
x4
x2 00
f (0) + f 000 (0) + f (4) (0)
2
6
24
Since f (x) = 1/(x3 + 1), and the derivatives require for the desired polynomial
p4 (x) can be obtained as
f 0 (x)
3x2
,
(x3 + 1)2
f 0 (x0 ) = f 0 (0) = 0.0000
f (x)
6x
18x4
3
,
=
3
3
(x + 1)
(x + 1)2
f 000 (x)
00
f (4) (x) =
f 00 (0) = 0.0000
162x6
108x3
6
+
(x3 + 1)4 (x3 + 1)3 (x3 + 1)2
1944x5
360x2
1944x8
,
(x3 + 1)5 (x3 + 1)4 (x3 + 1)3
f 000 (0) = 6.0

f (4) (0) = 0

f (0.2) p4 (0.2) = 1 (0.2 0)3 = 0.9920
To compute error bound for the approximation of f (0.2), we have
|R5 (x)|
1
|x x0 |5 M
120
where
M=
max |f (5) (x)| = 120.9884
0.0x0.2
at x = 0.2 and
f (5) (x) =
38880x7
12960x4
720x
29160x10
+
+
(x3 + 1)6 (x3 + 1)5 (x3 + 1)4 (x3 + 1)3
Hence for |f (5) ((x))| M , we have

|R5 (x)|
1
(0.2)5 (120.9884) = 0.0003
120
5. Use the Lagrange interpolation formula based on the points x0 = 0, x1 = 1, x2 = 2.5

2
to find the equation of the quadratic polynomial to approximate f (x) =
at
x+2
x = 2.3.
A.BENHARI
84
Solution: The values of f (x) =

f (0) = 1,
2
at the given points are
x+2
f (1) = 2/3,
f (2.5) = 2/4.5
and using these values in the quadratic Lagrange interpolating formula, we have
the approximation of f (2.3) as follows
p2 (2.3) = L0 (2.3)(1) + L1 (2.3)(2/3) + L2 (2.3)(2/4.5)
=
(2.3 1)(2.3 2.5)

(1)
(0 1)(0 2.5)
(2.3 0)(2.3 2.5)

(2/3)
(1 0)(1 2.5)
(2.3 0)(2.3 1)
(2/4.5)
(2.5 0)(2.5 1)
= 0.4548
6. Let f (x) = cos(x/4), where x is in radian. Use the quadratic Lagrange interpolation formula based on the points x0 = 0, x1 = 1, x2 = 2, and x3 = 4 to find the
polynomial p2 (x) to approximate the function f (x) at x = 0.5 and x = 3.5.
Solution: The values of f (x) = cos(x/4) at the given points are
f (0) = 1,
f (1) = 0.9999,
f (2) = 0.9996
p2 (0.5) = L0 (0.5)(1) + L1 (0.5)(0.9999) + L2 (0.5)(0.9996)
=
(0.5 1)(0.5 2)
(1)
(0 1)(0 2)
(0.5 0)(0.5 2)
(0.9999)
(1 0)(1 2)
(0.5 0)(0.5 1)
(0.9996)
(2 0)(2 1)
= 1.0000
Similarly, using the data points (1, 0.9999), (2, 0.9996), (4, 0.9985) in the quadratic
A.BENHARI
85
Lagrange interpolating formula, we have the approximation of f (3.5) as follows

p2 (3.5) = L0 (3.5)(0.9999) + L1 (3.5)(0.9996) + L2 (3.5)(0.9985)
=
(3.5 2)(3.5 4)
(0.9999)
(1 2)(1 4)
(3.5 1)(3.5 4)
(0.9996)
(2 1)(2 4)
(3.5 1)(3.5 2)
(0.9985)
(4 1)(4 2)
= 0.9988
7. Let f (x) = (x + 2)ln(x + 2). Use the quadratic Lagrange interpolation formula
based on the points x0 = 0, x1 = 1, x2 = 2, and x3 = 3 to approximate f (0.5) and
f (2.8). Also, compute the error bounds for your approximations.
Solution: The values of f (x) = (x + 2)ln(x + 2) at the given points are
f (0) = 1.3863,
f (1) = 3.2958,
f (2) = 5.5452
p2 (0.5) = L0 (0.5)(1.3863) + L1 (0.5)(3.2958) + L2 (0.5)(5.5452)
=
(0.5 1)(0.5 2)
(1.3863)
(0 1)(0 2)
(0.5 0)(0.5 2)
(3.2958)
(1 0)(1 2)
(0.5 0)(0.5 1)
(5.5452)
(2 0)(2 1)
= 2.2986
Similarly, using the data points (1, 3.2958), (2, 5.5452), (3, 8.0472) in the quadratic
A.BENHARI
86
Lagrange interpolating formula, we have the approximation of f (3.5) as follows

p2 (2.8) = L0 (2.8)(3.2958) + L1 (2.8)(5.5452) + L2 (2.8)(8.0472)
=
(2.8 2)(2.8 3)
(3.2958)
(1 2)(1 3)
(2.8 1)(2.8 3)
(5.5452)
(2 1)(2 3)
(2.8 1)(2.8 2)
(8.0472)
(3 1)(3 2)
= 7.5266
Now to compute the error bound, we have
|Ep2 (0.5) | |(0.5 0)(0.5 1)(0.5 2)|
and
|Ep2 (2.8) | |(2.8 1)(2.8 2)(2.8 3)|
0.25
= 0.0156
6
0.1111
= 0.0053
6
8. Consider the function f (x) = ex and x = 0, 0.25, 0.5, 1. Then use the suitable
Lagrange interpolating polynomial to approximate f (0.75). Also, compute an error
bound for your approximation.
2
Solution: The values of f (x) = ex at the given points are

f (0) = 1.0000,
f (0.25) = 1.0645,
f (0.5) = 1.2840,
f (1) = 2.7183
and using these values in the cubic Lagrange interpolating formula, we have the
approximation of f (0.5) as follows
p3 (0.75) = L0 (0.75)(1) + L1 (0.75)(1.0645) + L2 (0.75)(1.284) + L3 (0.75)(5.5452)
=
(0.75 0.25)(0.75 0.5)(0.75 1)

(1)
(0 0.25)(0 0.5)(0 1)
(0.75 0)(0.75 0.5)(0.75 1)

(1.0645)
(0.25 0)(0.25 0.5)(0.25 1)
(0.75 0)(0.75 0.25)(0.75 1)

(1.2840)
(0.5 0)(0.5 0.25)(0.5 1)
(0.75 0)(0.75 0.25)(0.75 0.5)

(2.7183)
(1 0)(1 0.25)(1 0.5)
= 1.7911
A.BENHARI
87

|Ep3 (0.75) | |(0.75 0)(0.75 0.25)(0.75 0.5)(0.75 1)|
206.5894
= 0.2017
24
Note that
|f (4) () M = max |f (4) (x)| = 206.5894
0x1
where
f (4) (x) = 4ex (3 + 16x2 + 4x4 )

9. Let f (x) = x4 2x + 1. Use cubic Lagrange interpolation formula based on the
points x0 = 1, x1 = 0, x2 = 2, and x3 = 3 to find the polynomial p3 (x) to approximate the function f (x) at x = 1.1. Also, compute an error bound for your
approximation.
Solution: The values of f (x) = x4 2x + 1 at the given points are
f (1) = 4.0000,
f (0) = 1.0000,
f (2) = 13.0000,
f (3) = 76.0000
and using these values in the cubic Lagrange interpolating formula, we have the
approximation of f (0.5) as follows
p3 (1.1) = L0 (1.1)(4) + L1 (1.1)(1) + L2 (1.1)(13) + L3 (1.1)(76)
=
(1.1 0)(1.1 2)(1.1 3)

(4)
(1 0)(1 2)(1 3)
(1.1 + 1)(1.1 2)(1.1 3)

(1)
(0 + 1)(0 2)(0 3)
(1.1 + 1)(1.1 0)(1.1 3)

(13)
(2 + 1)(2 0)(2 3)
(1.1 + 1)(1.1 0)(1.1 2)

(76)
(3 + 1)(3 0)(3 2)
= 3.6860
|Ep3 (1.1) | |(1.1 + 1)(1.1 0)(1.1 2)(1.1 3)|
24
= 3.9501
24
Note that
|f (4) () M = max |f (4) (x)| = 24
1x3
10. Construct the Lagrange interpolation polynomials for the following functions and
compute the error bounds for the approximations:
(a) f (x) = x + 2x+1 ,
x0 = 0, x1 = 1, x2 = 2.5, x3 = 3.
3
2
(b) f (x) = 3x + 2x + 1,
x0 = 1, x1 = 2, x2 = 3.
A.BENHARI
88
(c)
f (x) = cos x sin x,
x0 = 0, x1 = 0.25, x2 = 0.5, x3 = 1.
Solution: (a) The cubic Lagrange interpolating polynomial the approximation of

f (x) = x + 2x+1 is
p3 (x) = 0.3661x3 1.3079x2 + 0.2764x + 2
The error bound for the approximation is
|Ep3 (x) | |(x 0)(x 1)(x 2.5)(x 3)|
3.6934
24
11. Consider the following table:

0
1
2
3
x
f (x) 2.0 3.72 8.39 21.06
(a) Construct divided difference table for the tabulated function.
(b) Compute the Newton interpolating polynomials p2 (x) and p3 (x) at x = 2.2.
Solution: (a) The divided difference table for the given tabulated function is as
follows:
2.0000
0
0
0
3.7200 1.7200
0
0
8.3900 4.6700 1.4750
0
21.0600 12.6700 4.0000 0.8417
(b) Firstly, we construct the second degree polynomial p2 (x) by using the quadratic
Newton interpolation formula as follows
p2 (x) = f [x0 ] + f [x0 , x1 ](x x0 ) + f [x0 , x1 , x2 ](x x0 )(x x1 )
then with the help of the divided differences table, we get approximation of f (2.2)
as
f (2.2) p2 (2.2) = 3.7200 + 4.6700(2.2 1) + 4.0000(2.2 1)(2.2 2) = 10.1224
Now to construct the cubic interpolatory polynomial p3 (x) that fits at all four given
points, we use
+ f [x0 , x1 , x2 , x3 ](x x0 )(x x1 )(x x2 )
which gives the approximation of f (2.2) as
f (2.2) p3 (2.2) = 2 + 1.72(2.2) + 1.475(2.2)(1.2) + 0.8417(2.2)(1.2)(0.2)
p3 (2.2) = 10.2840
12. Consider the following table:
1
x
2
3
4
5
f (x) 3.60 1.80 1.20 0.90 0.72
A.BENHARI
89
(a) Construct divided difference table for the tabulated function.

(b) Compute the Newton interpolating polynomials p3 (x) and p4 (x) at
x = 2.5, 3.5.
follows:
3.6000
0
0
0
0
1.8000 1.8000
0
0
0
1.2000 0.6000 0.6000
0
0
0.9000 0.3000 0.1500 0.1500
0
0.7200 0.1800 0.0600 0.0300 0.0300
(b) Firstly, we construct the third degree polynomial p3 (x) by using the cubic
Newton interpolation formula as follows
+ f [x0 , x1 , x2 , x3 ](x x0 )(x x1 )(x x2 )
which gives the approximation of f (2.5) as
f (2.5) p3 (2.5) = 3.6 1.8(1.5) + 0.6(1.5)(0.5) 0.15(1.5)(0.5)(0.5)
p3 (2.5) = 1.4063
Now to construct the fourth degree interpolatory polynomial p4 (x) that fits at all
four points. We only have to add one more term to the polynomial p3 (x) and get
p4 (x) = p3 (x) + f [x0 , x1 , x2 , x3 , x4 ](x x0 )(x x1 )(x x2 )(x x3 )
Using it to get the approximation of f (2.5) as
f (2.5) p4 (2.5) = p3 (2.5) + 0.03(1.5)(0.5)(0.5)(1.5)
p4 (2.5) = 1.4063 + 0.0169 = 1.4232
Similarly, the approximations of f (3.5) using p3 (3.5) and p4 (3.5) are
f (3.5) p3 (3.5) = 1.0687 and f (3.5) p4 (3.5) = 1.0406
13. Consider the following table of the f (x) =
x:
4
x
5
6
7
8
f (x) 2.0000 2.2361 2.4495 2.6458 2.8284
(a) Construct the divided difference table for the tabulated function.
(b) Find the Newton interpolating polynomials p3 (x) and p4 (x) at x = 5.9.
(c) Compute error bounds for your approximations in part (b).
follows:
2.0000
0
0
0
0
2.2361 0.2361
0
0
0
2.4495 0.2134 0.0113
0
0
2.6458 0.1963 0.0086 0.0009
0
2.8284 0.1826 0.0069 0.0006 0.0001
A.BENHARI
90
(b) To find the approximation of f (5.9) using p3 (5.9), we use

f (5.9) p3 (5.9) = 2 + 0.2361(1.9) 0.0113(1.9)(0.9) + 0.0009(1.9)(0.9)(0.1)
p3 (5.9) = 2.4290
Now to find the approximation of f (5.9) using p4 (5.9), we use
f (5.9) p4 (5.9) = p3 (5.9) + 0.03(1.5)(0.5)(0.5)(1.5)
p4 (5.9) = 2.4290 0.0001(1.9)(0.9)(0.1)(1.1)
p4 (5.9) = 2.4290 0.000019 = 2.4290
(c) To compute the error bound, we have
|Ep3 | |(5.9 4)(5.9 5)(5.9 6)(5.9 7)|
0.0073
= 0.000057
24
Note that
|f (4) () M = max |f (4) (x)| = 0.0073
4x7
where
f (4) (x) = 15/16x( 7/2)
Also
|Ep4 | |(5.9 4)(5.9 5)(5.9 6)(5.9 7)(5.9 8)|
0.0064
= 0.000021
120
where
|f (5) () M = max |f (5) (x)| = 0.0064
4x8
and
f (5) (x) = 105/32x( 9/2)
14. Let f (x) = ex sinx, with x0 = 0, x1 = 2, x2 = 2.5, x3 = 4, x4 = 4.5. Then
(a) Construct the divided-difference table for the given data points.
(b) Find the Newton divided difference polynomials p2 (x), p3 (x) and p4 (x) at
x = 2.4.
(c) Compute error bounds for your approximations in part (b).
(d) Compute the actual error.
follows:
0
0
0
0
0
6.7188
3.3594
0
0
0
7.2909
1.1441 0.8861
0
0
41.3200 32.4073 16.7757 3.9724
0
87.9945 93.3489 30.4708 5.4781 0.3346
(b) The approximation of f (2.4) using p2 (2.4), p3 (2.4), and p4 (2.4) are as follows:
p2 (2.4) = 7.8475,
A.BENHARI
p3 (2.4) = 7.4969,
91
p4 (2.4) = 7.5419
(c) Since the higher-order derivatives of the functions are

f 000 (x) = 2ex (cos x sin x),
f (4) (x) = 4ex sin x,
f (5) (x) = 4ex (cos x + sin x)
Then the possible error bounds for each approximations are

EB2 = 0.5456,
EB3 = 1.0578,
EB4 = 0.7421
(d) The exact value of f (2.4) is 7.4457.

15. Show that if x0 , x1 , and x2 are distinct then
f [x0 , x1 , x2 ] = f [x1 , x2 , x0 ] = f [x2 , x0 , x1 ]
Solution: All three divided differences can be expanded as
(x2 x1 )f0 (x2 x0 )f1 + (x1 x0 )f2
(x2 x1 )(x2 x0 )(x1 x0 )
16. The divided difference form of the interpolating polynomial p3 (x) is
p3 (x) = f [x0 ] + (x x0 )f [x0 , x1 ] + (x x0 )(x x1 )f [x0 , x1 , x2 , x0 ]
+ (x x0 )(x x1 )(x x2 )f [x0 , x1 , x2 , x3 ]
By expressing these divided differences in terms of the function values f (xi )(i =
0, 1, 2, 3), verify that p3 (x) does pass through the points (xi , f (xi ))(i = 0, 1, 2, 3).
Solution: Let p3 (x0 ) = f [x0 ] = f (x0 ) and
"
f (x1 ) f (x0 )
= f (x1 )
p3 (x1 ) = f [x0 ] + (x1 x0 )f [x0 , x1 ] = f (x0 )(x1 x0 )
x1 x 0
Similarly, p3 (x2 ) = f (x2 ) and p3 (x3 ) = f (x3 )
17. Let f (x) = x2 + ex and x0 = 0, x1 = 1. Use the divided differences to find the
value of the second divided difference f [x0 , x1 , x0 ].
Solution: Since
f [x0 , x0 ] =
1 0
f (x0 ) = f 0 (x0 )
1!
therefore
f [x0 , x1 , x0 ] = f [x0 , x0 , x1 ] =
f [x0 , x1 ] f [x0 , x0 ]
x1 x0
or
f [x0 , x1 ] f 0 (x0 )
x1 x0
Using definition of the first-order divided difference of f (x) at points x0 and x1 , we
have
f [x1 ] f [x0 ]
f [x0 , x1 ] =
x1 x 0
and it gives
3.7183 1
= 2.7183
f [0, 1] =
10
f [x0 , x1 , x0 ] =
A.BENHARI
92
Also
f 0 (x0 ) = 2x0 + ex0
and f 0 (0) = 1
Using these values, we obtain the value of the second divided difference as
f [0, 1, 0] =
A.BENHARI
2.7183 1
= 1.7183
10
93
Chapter 5
Differentiation and Integration
1. Let f (x) = (x 1)ex and take h = 0.01.

(a) Calculate approximation to f 0 (2.3) using the two-point forward-difference formula. Also, compute the actual error and an error bound for you approximation.
(b) Solve part (a) using the two-point backward-difference formula.
Solution: (a) Given f (x) = (x 1)ex and h = 0.01, then using the two-point
forward-difference formula with x0 = 2.3, we have
f 0 (2.3)
f (2.3 + h) f (2.3)
h
Then for h = 0.01, we get

f 0 (2.3)
f (2.31) f (2.3)
0.01
(2.31 1)e2.31 (2.3 1)e2.3
= 23.10591068
0.01
The actual error is
Error = f 0 (2.3) 23.10591068 = 2.3e2.3 23.10591068 = 0.16529103
To find the error bound, we use the following formula
h
EF (f, h) = f 00 ((x)),
2
where (x) (x0 , x0 + h)
which can be written as
|EF (f, h)| =
0.01 00
|f ((x))|,
2
for (2.3, 2.31)
The second derivative f 00 (x) of the function can be found as

f 0 (x) = xex ,
and f 00 (x) = (1 + x)ex
The value of the second derivative f 00 ((x)) cannot be computed exactly because
(x) is not known. But one can bound the error by computing the largest possible
value for |f 00 ((x))|. So bound |f 00 | on [2.3, 2.31] can be obtain
M=
max
2.3x2.31
|(1 + x)ex | = 33.346346
at x = 2.31. Since |f 00 ((x))| M , therefore, for h = 0.01, we have

|EF (f, h)|
0.01
M = 0.005(33.346346) = 0.16673173
2
which is the possible maximum error in our approximation.
A.BENHARI
94
(b) Using the two-point backward-difference formula with x0 = 2.3, we have

f 0 (2.3)
f (2.3) f (2.3 h)
h

f 0 (2.3)
f (2.3) f (2.29)
0.01
(2.3 1)e2.3 (2.29 1)e2.29
= 22.77675826
0.01
The actual error is

Error = f 0 (2.3) 22.77675826 = 2.3e2.3 22.77675826 = 0.16386139
h
EF (f, h) = f 00 ((x)),
2
where (x) (x0 h, x0 )
|EF (f, h)| =
0.01 00
|f ((x))|,
2
for (2.29, 2.3)

f 0 (x) = xex ,
and f 00 (x) = (1 + x)ex
M=
max
2.29x2.3
|(1 + x)ex | = 32.9148021

|EF (f, h)|
0.01
M = 0.005(32.9148021) = 0.16457401
2

2. Solve the Problem 1 for the f (x) = (x2 + x + 1)e2x with h = 0.05.
Solution: (a) Given f (x) = (x2 + x + 1)e2x and h = 0.05, then using the two-point
forward-difference formula with x0 = 2.3, we have
f 0 (2.3)
A.BENHARI
f (2.35) f (2.3)
0.05
[(2.35)2 + 2.35 + 1]e4.7 [(2.3)2 + 2.3 + 1]e4.6
= 2418.720325
0.05
95
The actual error is

Error = [2(2.3)2 + 4(2.3) + 3]e4.6 2418.720325 = 152.4676147
h
EF (f, h) = f 00 ((x)), where (x) (x0 , x0 + h)
2
0.05 00
|f ((x))|, for (2.3, 2.35)

|EF (f, h)| =
2
f 0 (x) = (2x2 + 4x + 3)e2x
and f 00 (x) = (4x2 + 12x + 10)e2x
M=
max
2.3x2.35
|(4x2 + 12x + 10)e2x | = 6628.715027

0.01
M = 0.025(6628.715027) = 165.7178757
2
|EF (f, h)|
(b) Similarly, approximation of f 0 (2.3) by using the two-point backward-difference

formula is
f 0 (2.3) 2126.1
with actual error
Error = 140.1954
and the maximum error (or the error bound) is
|EF (f, h)| 146.1425
3. By using the following data
(1.2, 11.6), (1.29, 13.8), (1.3, 14), (1.31, 14.3), (1.4, 16.8)
compute the best approximations of f 0 (1.3) using the two-point forward-difference
and backward-difference formulas.
Solution: Given x0 = 1.3 and take x1 = 1.31, gives h = 0.01, then using the
two-point forward-difference formula, we have
f 0 (1.3)
A.BENHARI
f (1.31) f (1.3)
0.01
14.3 14.0
= 30.0
0.01
96
Using the two-point backward-difference formula with x0 = 1.3, x0 h = 1.29, and

h = 0.01, we have
f (1.3) f (1.29)
f 0 (1.3)
0.01
14.0 13.8
= 20.0
0.01
4. Let f (x) = sin(x + 1). Compute the approximations of f 0 ( ) using the two-point
4
forward-difference and backward-difference formulas. Compute the actual errors
and also, find error bounds using the error formulas.
Solution: Given f (x) = sin(x + 1) and h = 0.1, then using the two-point forwarddifference formula with x0 = /4, we have
f 0 (/4)
f (/4 + h) f (/4)
h

sin(/4 + 0.1 + 1) sin(/4 + 1)
0.1
f 0 (/4)
0.9509197 0.9770613
= 0.261416
0.1
The actual error is
Error = f 0 (/4) + 0.261416 = 0.212958 + 0.261416 = 0.048458

h
EF (f, h) = f 00 ((x)),
2
where (x) (x0 , x0 + h)
|EF (f, h)| =
0.1 00
|f ((x))|,
2
for (0.7854, 0.8854)

f 0 (x) = cos(x + 1) and f 00 (x) = sin(x + 1)
M=
max
0.7854x0.8854
| sin(x + 1)| = 0.977061

|EF (f, h)|
A.BENHARI
0.1
M = 0.05(0.977061) = 0.048853
2
97

Now using the two-point backward-difference formula with x0 = /4, we have
f 0 (/4)
f (/4) f (/4 h)
h

sin(/4 + 1) sin(/4 + 1 0.1)
0.1
f 0 (/4)
0.9770613 0.993440
= 0.163787
0.1
The actual error is
Error = f 0 (/4) + 0.163787 = 0.212958 + 0.163787 = 0.049171

h
EF (f, h) = f 00 ((x)),
2
where (x) (x0 h, x0 )
|EF (f, h)| =
0.1 00
|f ((x))|,
2
for (0.6854, 0.7854)

f 0 (x) = cos(x + 1) and f 00 (x) = sin(x + 1)
M=
max
0.6854x0.7854
| sin(x + 1)| = 0.993440

|EF (f, h)|
0.1
M = 0.05(0.993440) = 0.049672
2

5. Use the three-point central-difference formula to compute the approximate value
for f 0 (5) with f (x) = (x2 + 1) ln x, and h = 0.05. Compute the actual error and
the error bound for you approximation.
Solution: Given f (x) = (x2 + 1) ln x and x1 = 5, h = 0.05, then using the
three-point formula, we have
f 0 (5)
A.BENHARI
f (5.05) f (4.95)
f (5 + 0.05) f (5 0.05)
=
2(0.05)
0.1
98
Then
f 0 (5)
[(5.05)2 + 1] ln(5.05) [(4.95)2 + 1] ln(4.95)

0.1
42.917837 40.788382
= 21.294553
0.1
The actual error is

Error = f 0 (5) 21.294553 = 21.294379 21.294553 = 0.000174
To compute the error bound for the approximation, we use the formula
EC (f, h) =
or
(0.05)2 000
f ((x1 )),
6
for (x1 ) (4.95, 5.05)
(0.05)2 000
|f ((x1 ))|, for (x1 ) (4.95, 5.05)
6
The third derivative f 000 (x) of the function can be found as
|EC (f, h)| =
f 0 (x) = 2x ln x + (x2 + 1)/x, f 00 (x) = 2 ln x (x2 + 1)/x2 4, f 000 (x) = 2(x2 + 1)/x3
The value of the third derivative f 000 ((x1 )) cannot be computed exactly because
(x1 ) is not known. But one can bound the error by computing the largest possible
value for |f 000 ((x1 ))|. So bound |f 000 | on [4.95, 5.05] can be obtain
M=
max
4.95x5.05
|2(x2 + 1)/x3 | = 0.4205302

(0.05)2
M = 0.00042(0.4205302) = 0.000175
|EF (f, h)|
6
6. Use the three-point central-difference formula to compute the approximate value
for f 0 (2) with f (x) = ex/2 + 2 cos x, and h = 0.01. Compute the actual error and
the error bound for you approximation.
Solution: Given f (x) = ex/2 + 2 cos x and x1 = 2, h = 0.01, then using the
three-point formula, we have
f 0 (2)
f (2.01) f (1.99)
f (2 + 0.01) f (2 0.01)
=
2(0.01)
0.02
Then
A.BENHARI
f 0 (2)
e2.01/2 + 2 cos(2.01) e1.99/2 + 2 cos(1.99)

0.02
1.88146957 1.89065793
= 0.45941795
0.02
99
The actual error is

Error = f 0 (2) + 0.45941795 = 0.45945394 + 0.45941795 = 0.0000360
EC (f, h) =
or
(0.01)2 000
f ((x1 )),
6
for (x1 ) (1.99, 2)
(0.01)2 000
|f ((x1 ))|, for (x1 ) (1.99, 2)
6
|EC (f, h)| =
f 0 (x) = 1/2ex/2 2 sin x, f 00 (x) = 1/4ex/2 2 cos x, f 000 (x) = 1/8ex/2 + 2 sin x
value for |f 000 ((x1 ))|. So bound |f 000 | on [1.99, 2] can be obtain
M = max |1/8ex/2 + 2 sin x| = 2.164917
1.99x2

(0.01)2
M = 0.00002(2.164917) = 0.0000361
6
|EF (f, h)|
7. Solve the Problem 6 to find the best approximation of f 0 (2) using the three-point
forward-difference and backward-difference formulas.
Solution: Given f (x) = ex/2 + 2 cos x and x0 = 2, h = 0.01, then using the
three-point forward-difference formula, we have
f 0 (2)
3f (2) + 4f (2.01) f (2.02)

2(0.01)
3(e2/2 + 2 cos 2) + 4(e2.01/2 + 2 cos 2.01) (e2.02/2 + 2 cos 2.02)
0.02
5.6579645 + 7.5258783 1.8771043

= 0.4595257
0.02
Now taking x2 = 2, h = 0.01, and then using the three-point backward-difference
formula, we have
f 0 (2)
A.BENHARI
f (1.98) 4f (1.99) + 3f (2)

2(0.01)
(e1.98/2 + 2 cos 1.98) 4(e1.99/2 + 2 cos 1.99) + 3(e2/2 + 2 cos 2)

0.02
1.8954767 7.5626317 + 5.6579645

= 0.4595261
0.02
100
8. By using the following data

(1.0, 2.0), (1.5, 1.94), (2.0, 2.25), (3.0, 3.11)
find the best approximate values for f 0 (1.5), f 0 (1.0), and f 0 (3.0) using suitable threepoint formulas.
Solution: Taking (1.0, 2.0), (2.0, 2.25), and h = 0.5, we get approximation of
f 0 (1.5) by using
Central-difference formula:
2.25 2.0
f (2.0) f (1.0)
=
= 0.25
f 0 (1.5)
2h
1.0
Taking (1.0, 2.0), (1.5, 1.94), (2.0, 2.25), and h = 0.5, we get approximation of f 0 (1.0)
by using
Forward-difference formula:
f 0 (1.0)
3f (1.0) + 4f (1.5) f (2.0)

2h
3(2) + 4(1.94) 2.25

= 0.49
1.0
Taking (1.0, 2.0), (2.0, 2.25), (3.0, 3.11), and h = 1.0, we get approximation of f 0 (1.0)
by using
Backward-difference formula:
f 0 (3.0)
f (1.0) 4f (2.0) + 3f (3.0)

2h
2.0 4(2.25) + 3(3.11)
= 1.165
2.0
9. Use all three-point formulas to compute the approximate value for f 0 (2) for the
derivative of f (x) = ex/2 + x3 , taking h = 0.1. Also, compute the actual errors and
error bounds for your approximation.
Solution: Given f (x) = ex/2 + x3 and h = 0.1, then
(e2.1/2 + (2.1)3 ) (e1.9/2 + (1.9)3 )
(f (2.1) f (1.9))
0
=
= 13.36970729
f (2)
2h
0.2
3f (2) + 4f (2.1) f (2.2)
f 0 (2)
2h
A.BENHARI
3(e2/2 + (2)3 ) + 4(e2.1/2 + (2.1)3 ) (e2.2/2 + (2.2)3 )

= 13.33796482
0.2
101
Backward difference formula:

f 0 (2)
f (1.8) 4f (1.9) + 3f (2)

2h
(e1.8/2 + (1.8)3 ) 4(e1.9/2 + (1.9)3 ) + 3(e2/2 + (2)3 )
= 13.3380498
0.2
Since the exact solution of the first derivative of the given function at x = 2 is
13.35914091, so the corresponding actual errors are, 0.01056638, 0.02117609, and
0.02109111 respectively. This shows that the approximate solution got by using the
central-difference formula is closer to exact solution as compared with the other two
difference formulas.
The error bounds for the approximations got by central-difference, forward-difference,
and backward-difference formulas are as follows:
Central-difference error formula
EC (f, h) =
h2 000
f ((x1 ))
6
or can be written as
h2 000
|f ((x1 ))|
6
Since f 000 (x) = 1/8ex/2 + 6, therefore, taking
|EC (f, h)|
|f 000 ((x1 ))| M =
max |1/8ex/2 + 6| = 6.35720639
1.9x2.1
at x = 2.1 and h = 0.1, we obtain

|EC (f, h)|
(0.1)2
6.35720639 = 0.010595343
6
Forward-difference error formula

EF (f, h) =
h2 000
f ((x0 ))
3
h2 000
|f ((x0 ))|
3
Taking |f 000 ((x0 ))| M = max |1/8ex/2 +6| = 6.37552075 at x = 2.2 and h = 0.1,
|EF (f, h)|
we obtain
2x2.2
(0.1)2
6.37552075 = 0.021251735
3
Backward-difference error formula
|EF (f, h)|
EB (f, h) =
A.BENHARI
h2 000
f ((x2 ))
3
102
|EB (f, h)|
h2 000
|f ((x2 ))|
3
Taking
|f 000 ((x2 ))| M = max |1/8ex/2 + 6| = 6.33978523
1.8x2
at x = 2 and h = 0.1, we obtain

|EB (f, h)|
(0.1)2
6.33978523 = 0.021132617
3
10. Use all three-point formulas to compute the approximate value for f 0 (2.2) for the
derivative of f (x) = x2 ex x + 1, taking h = 0.2. Also, compute the actual errors
and error bounds for your approximation.
Solution: Given f (x) = x2 ex x + 1 and h = 0.2, then
f 0 (2.2)
(f (2.4) f (2.0))
2h
((2.4)2 e2.4 2.4 + 1) ((2.0)2 e2.0 2.0 + 1)
0.4
83.84317889
f 0 (2.2)
3f (2.2) + 4f (2.4) f (2.6)

2h
3((2.2)2 e2.2 2.2 + 1) + 4((2.4)2 e2.4 2.4 + 1) ((2.6)2 e2.6 2.6 + 1)
0.4
78.78979670
Backward difference formula:
f 0 (2.2)
f (1.8) 4f (2.0) + 3f (2.2)

2h
((1.8)2 e1.8 1.8 + 1) 4((2.0)2 e2 2.0 + 1) + 3((2.2)2 e2.2 2.2 + 1)
0.4
80.04789053
Since the exact solution of the first derivative of the given function at x = 2.2 is
82.39112475, so the corresponding actual errors are, 1.45205432, 3.60132803, and
2.3432342 respectively. This shows that the approximate solution got by using the
central-difference formula is closer to exact solution as compared with the other
A.BENHARI
103
two difference formulas.

The error bounds for the approximations got by central-difference, forward-difference,
and backward-difference formulas are as follows:
Central-difference error formula
EC (f, h) =
h2 000
f ((x1 ))
6
h2 000
|f ((x1 ))|
6
Since f 000 (x) = ex (x2 + 6x + 6), therefore, taking
|EC (f, h)|
|f 000 ((x1 ))| M =
max |ex (x2 + 6x + 6)| = 288.36629
2.0x2.4

|EC (f, h)|
(0.2)2
288.36629 = 1.922442
6
Forward-difference error formula

EF (f, h) =
h2 000
f ((x0 ))
3
h2 000
|f ((x0 ))|
3
max |ex (x2 + 6x + 6)| = 381.83161 at x = 2.6 and
|EF (f, h)|

Taking |f 000 ((x0 ))| M =
h = 0.2, we obtain
2.2x2.6
|EF (f, h)|
(0.2)2
381.83161 = 5.091088
3
Backward-difference error formula

EB (f, h) =
h2 000
f ((x2 ))
3
|EB (f, h)|
h2 000
|f ((x2 ))|
3
Taking
|f 000 ((x2 ))| M =
max |ex (x2 + 6x + 6)| = 216.96133
1.8x2.2

|EB (f, h)|
A.BENHARI
(0.2)2
216.96133 = 2.892918
3
104
11. Use the most accurate formula to determine approximations that will complete the
following table.
x
f(x)
f 0 (x)
2.1
-1.709847
2.2
-1.373823
2.3
-1.11921
2.4
-0.916014
Solution: We first find the approximation of f 0 (2.1) using h = 0.1 with the help
of forward-difference formula as follows
f 0 (2.1)
3f (2.1) + 4f (2.2) f (2.3)

2h
3(1.709847) + 4(1.373823) (1.11921)
= 3.7672950
0.2
Now we find the approximation of f 0 (2.2) using h = 0.1 with the help of centraldifference formula as follows
f 0 (2.2)
1.11921 + 1.709847
f (2.3) f (2.1)
=
= 2.9531850
2h
0.2
To find the approximation of f 0 (2.3) using h = 0.1, we use again central-difference

formula as follows
f 0 (2.3)
0.916014 + 1.373823
f (2.4) f (2.2)
=
= 2.2890450
2h
0.2
To find the approximation of f 0 (2.4) using h = 0.1, we use backward-difference

formula as follows
f 0 (2.4)
Thus
f (2.2) 4f (2.3) + 3f (2.4)

2h
1.373823 4(1.11921) + 3(0.916014)
= 1.7748750
0.2
x
2.1
2.2
2.3
2.4
f(x)
-1.709847
-1.373823
-1.11921
-0.916014
f 0 (x)
3.7672950
2.9531850
2.2890450
1.7748750
12. The data in the Problem 11 were taken from the function f (x) = tan(x). Compute
the actual errors in problem 11 and also, find error bounds using the error formulas.
Solution: The actual errors are as follows:
Error = f 0 (2.1) 3.7672950 = 3.9235752 3.7672950 = 0.1562802
Error = f 0 (2.2) 3.7672950 = 2.8873898 2.9531850 = 0.0657952
A.BENHARI
105
Error = f 0 (2.3) 3.7672950 = 2.2526392 2.2890450 = 0.0364058

Error = f 0 (2.4) 3.7672950 = 1.8390822 1.7748750 = 0.0642072
Now finding the error bound for the approximation of f 0 (2.1), we do the following
f 000 (x) = 2sec2 x(sec2 x + 2tan2 x) and |f 000 ((x0 ))| M
where
M=
max |2sec2 x(sec2 x + 2tan2 x)| = 76.6723533
2.1x2.3

|EF (f, h)|
(0.1)2
76.6723533 = 0.2555745
3
The error bound for the approximation of f 0 (2.2), we have

M=
2.1x2.3

|EC (f, h)|
(0.1)2
76.6723533 = 0.1277873
6

M=
2.2x2.4

|EC (f, h)|
(0.1)2
38.4725597 = 0.0641209
6
Finally, the error bound for the approximation of f 0 (2.4), we have

M=
2.2x2.4

|EB (f, h)|
(0.1)2
38.4725597 = 0.1282419
3
13. Use the most accurate formula to determine approximations that will complete the
following table.
x
f(x)
f 0 (x)
8.1
16.94410
8.3
17.56492
8.5
18.19056
8.7
18.82091
A.BENHARI
106
Solution: We first find the approximation of f 0 (8.1) using h = 0.2 with the help
of forward-difference formula as follows
f 0 (8.1)
3f (8.1) + 4f (8.3) f (8.5)

2h
3(16.94410) + 4(17.56492) (18.19056)
= 3.092050
0.4
Now we find the approximation of f 0 (8.3) using h = 0.2 with the help of centraldifference formula as follows
f 0 (8.3)
18.19056 16.94410
f (8.5) f (8.1)
=
= 3.116150
2h
0.4
To find the approximation of f 0 (8.5) using h = 0.2, we use again central-difference

formula as follows
f 0 (8.5)
18.82091 17.56492
f (8.7) f (8.3)
=
= 3.1399750
2h
0.4
To find the approximation of f 0 (8.7) using h = 0.2, we use backward-difference

formula as follows
f 0 (8.7)
Thus
f (8.3) 4f (8.5) + 3f (8.7)

2h
17.56492 4(18.19056) + 3(18.82091)
= 3.1635250
0.4
x
8.1
8.3
8.5
8.7
f(x)
16.94410
17.56492
18.19056
18.82091
f 0 (x)
3.092050
3.116150
3.1399750
3.1635250
14. The data in the Problem 13 were taken from the function f (x) = x ln x. Compute
the actual errors in the Problem 13 and also, find error bounds using the error
formulas.
Solution: The actual errors are as follows:
Error = f 0 (8.1) 3.092050 = 3.0918641 3.092050 = 0.0001859
Error = f 0 (8.3) 3.116150 = 3.1162555 3.116150 = 0.0001055
Error = f 0 (8.5) 3.139975 = 3.1400662 3.139975 = 0.0000912
Error = f 0 (8.7) 3.163525 = 3.163323 3.163525 = 0.000202
Now finding the error bound for the approximation of f 0 (8.1), we do the following
f 000 (x) = 1/x2
A.BENHARI
and |f 000 ((x0 ))| M
107
where
M=
max | 1/x2 | = 0.0152416
8.1x8.3

(0.2)2
0.0152416 = 0.000203
|EF (f, h)|
3
M=
max | 1/x2 | = 0.0152416
8.1x8.5

(0.2)2
0.0152416 = 0.000102
6
|EC (f, h)|

M=
max | 1/x2 | = 0.0138408
8.3x8.7

|EC (f, h)|
(0.2)2
0.0138408 = 0.0000923
6
Finally, the error bound for the approximation of f 0 (8.7), we have

M=
max | 1/x2 | = 0.0138408
8.3x8.7

|EB (f, h)|
(0.1)2
0.0138408 = 0.000185
3
15. Let f (x) = x + ln(x + 2), with h = 0.1. Use the five-point formula to approximate
f 0 (2). Find error bound for your approximation and compare the actual error to
the bound.
Solution: Given f (x) = x+ln(x+2), x1 = 2 and h = 0.1, then using the five-point
formula to get
f 0 (2)
f (2 2(0.1)) 8f (2 0.1) + 8f (2 + 0.1) f (2 + 2(0.1))

12(0.1)
or
(1.8 + ln(3.8)) 8(1.9 + ln(3.9)) + 8(2.1 + ln(4.1)) (2.2 + ln(4.2))
1.2
1.24999992
f 0 (2)
Now to compute the error bound for our approximation, we use the following error
formula
h4
EC (f, h) = f (5) ((x1 ))
30
A.BENHARI
108
or it can be written as
h4
(5)
|f ((x1 ))|
|EC (f, h)| =
30
Since f (5) = 24/(x + 2)5 and f (5) ((x1 )) cannot be computed exactly because (x1 )
is not known. But one can bound the error by computing the largest possible value
for |f (5) ((x1 ))|. So bound |f (5) | on [1.8, 2.2] can be obtain
M = max |f (5) (x)| = max |24/(x + 2)5 | = 0.03028958
1.82.2
1.82.2
Since |f (5) ((x1 ))| M , therefore

|EC (f, h)|
(0.1)4
(0.1)4
M=
(0.03028958) = 0.000000101
30
30
The actual error is as follows:

Error = f 0 (2) 1.24999992 = 1.250000000 1.24999992 = 0.0000001
16. Consider the following data
(1.1, 0.607), (1.2, 0.549), (1.4, 0.497), (1.5, 0.449), (1.7, 0.427), (1.8, 0.387)
Use the five-point formulas to compute the best approximations for f 0 (1.3), and
f 0 (1.6).
Solution: To find the approximation of f 0 (1.3) by five-point formula, we use
f 0 (1.3)
f (1.3 2h) 8f (1.3 h) + 8f (2 + h) f (2 + 2h)

12h
Taking h = 0.1, gives

f 0 (1.3)
f (1.1) 8f (1.2) + 8f (1.4) f (1.5)

12(0.1)
or
(0.607 8(0.549) + 8(0.497) 0.449)

= 0.215000
1.2
Similarly, to find the approximation of f 0 (1.6) with h = 0.1 using five-point formula,
we get
f (1.4) 8f (1.5) + 8f (1.7) f (1.8)
f 0 (1.6)
1.2
or
(0.497 8(0.449) + 8(0.427) 0.387)
f 0 (1.6)
= 0.055000
1.2
f 0 (1.3)
17. Let f (x) = x2 + 1, with h = 0.2. Use the five point formula to approximate f 0 (1.8).
A.BENHARI
109
Solution: Given f (x) = x2 + 1, x1 = 1.8, and h = 0.2, then finding the approximation of f 0 (1.8) by five-point formula, we use
f 0 (1.8)
f (1.8 0.4) 8f (1.8 0.2) + 8f (1.8 + 0.2) f (1.8 + 0.4)

12(0.2)
It gives
f 0 (1.8)
f (1.4) 8f (1.6) + 8f (2.0) f (2.2)

12(2.4)
or
f 0 (1.8)
(((1.4)2 + 1) 8((1.6)2 + 1) + 8((2.0)2 + 1) ((2.2)2 + 1))

= 3.6000000
2.4
18. Let f (x) = e2x , with points x = 0.25, 0.5, 0.75, 1.25, 1.50. Use the three-point
central-difference formula and the five point formula to approximate f 0 (1.0). Also,
compute the error bounds for your approximations.
Solution: Using the given data points to find the approximation of f 0 (1.0) by the
three-point formula central-difference formula with h = 0.25, we have
f 0 (1.0)
f (1.25) f (0.75)
f (1.0 + 0.25) f (1.0 0.25)
=
2(0.25)
0.5
Then
f 0 (1.0)
e2(1.25) e2(0.75)
0.5
0.082085 0.223130
= 0.2820903
0.5
EC (f, h) =
or
(0.25)2 000
f ((x1 )),
6
for (x1 ) (0.75, 1.25)
(0.25)2 000
|f ((x1 ))|, for (x1 ) (0.75, 1.25)
6
|EC (f, h)| =
f 0 (x) = 2e2x , f 00 (x) = 4e2x , f 000 (x) = 8e2x

value for |f 000 ((x1 ))|. So bound |f 000 | on [0.75, 1.25] can be obtain
M=
max
0.75x1.25
| 8e2x | = 1.7850413

|EF (f, h)|
A.BENHARI
(0.25)2
M = 0.01042(1.7850413) = 0.0185942
6
110

Now using the five-point formula to get the approximation of f 0 (1.0), we have
f 0 (1.0)
f (1.0 2(0.25)) 8f (1.0 0.25) + 8f (1.0 + 0.25) f (1.0 + 2(0.25))

12(0.25)
or
(e2(0.5) 8(e2(0.75) ) + 8(e2(1.25) ) e2(1.5) )
= 0.2700896
f (1.0)
3.0
0
Now to compute the error bound for our approximation, we use the following error
formula
h4
EC (f, h) = f (5) ((x1 ))
30
h4
(5)
|f ((x1 ))|
|EC (f, h)| =
30
Since f (5) = 32e2x and f (5) ((x1 )) cannot be computed exactly because (x1 ) is
not known. But one can bound the error by computing the largest possible value
for |f (5) ((x1 ))|. So bound |f (5) | on [0.5, 1.5] can be obtain
M = max |f (5) (x)| = max | 32e2x | = 11.7721421
0.51.5
0.51.5

|EC (f, h)|
(0.25)4
(0.25)4
M=
(11.7721421) = 0.0015328
30
30
19. Let f (x) = ex . Use the five-point formula to compute the best approximations
d2 f
at 0.5 and h = 0.1. Find bound on the error for your approximation.
for
dx2
2
Solution: Given f (x) = ex , x1 = 0.5, h = 0.1, then using five-point formula for
finding the approximation of f 00 (0.5), we have
f 00 (0.5)
f (0.5 0.2) + 16f (0.5 0.1) 30f (0.5) + 16f (0.5 + 0.1) f (0.5 + 0.2)
12(0.01)
f (0.3) + 16f (0.4) 30f (0.5) + 16f (0.6) f (0.7)
0.12
2
e(0.3) + 16e(0.4) 30e(0.5) + 16e(0.6) e(0.7)
= 3.8515964
0.12
To compute the error bound for our approximation, we use the following error
formula
h4
EC (f, h) = f (5) ((x1 ))
30
A.BENHARI
111
h4
(5)
|f ((x1 ))|
|EC (f, h)| =
30
Since f (5) (x) = 8xex (15+20x2 +32x4 ) and f (5) ((x1 )) cannot be computed exactly
because (x1 ) is not known. But one can bound the error by computing the largest
possible value for |f (5) ((x1 ))|. So bound |f (5) | on [0.3, 0.7] can be obtain
M=
max |f (5) (x)| =
0.3x0.7
max |8xex (15 + 20x2 + 32x4 )| = 44.2021901
0.3x0.7

|EC (f, h)|
(0.1)4
(0.1)4
M=
(44.2021901) = 0.0001473
30
30
20. Let f (x) = x+ln(x+2), with h = 0.1. Use the three-point formula to approximate
f 00 (2). Find error bound for your approximation and compare the actual error to
the bound.
Solution: Given f (x) = x + ln(x + 2), x1 = 2, h = 0.1, then using three-point
formula for finding the approximation of f 00 (2), we have
f 00 (2)
or
f 00 (2)
f (2 + 0.1) 2f (2) + f (2 0.1)

(0.1)2
f (2.1) 2f (2) + f (1.9)

0.01
(2.1 + ln(4.1)) 2(2 + ln(4.0)) + (1.9 + ln(3.9))
0.01
3.5109870 6.7725887 + 3.2609766

= 0.0625195
0.01
formula
h2 (4)
EC (f, h) = f ((x1 )), for (x1 ) (1.9, 2.1)
12
or
h2
|EC (f, h)| = |f (4) ((x1 ))|, for (x1 ) (1.9, 2.1)
12
The fourth derivative of the given function at (x1 ) is
f (4) ((x1 )) = 6/((x1 ) + 2)4

and it cannot be computed exactly because (x1 ) is not known. But one can bound
the error by computing the largest possible value for |f (4) ((x1 ))|. So bound |f (4) |
on the interval (1.9, 2.1) is
M=
A.BENHARI
max | 6/(x + 2)4 | = 0.0259354
1.9x2.1
112
at x = 1.9, Thus, for |f (4) ((x))| M , we have

h2
|EC (f, h)| M
12
Taking M = 0.0259354 and h = 0.1, we obtain
|EC (f, h)|
0.01
(0.0259354) = 0.0000216
12

The actual error is
Error = f 00 (2) + 0.0625195 = 0.0625000 + 0.0625195 = 0.0000195
21. Let f (x) = x sin x, with h = 0.1. Use five-point formula to approximate f 00 (1).
Find error bound for your approximation.
Solution: Given f (x) = x sin x, x1 = 1, h = 0.1, then using five-point formula
for finding the approximation of f 00 (1), we have
f 00 (1)
f (1 0.2) + 16f (1 0.1) 30f (1) + 16f (1 + 0.1) f (1 + 0.2)

12(0.01)
f (0.8) + 16f (0.9) 30f (1) + 16f (1.1) f (1.2)

0.12
(0.8 sin 0.8) + 16(0.9 sin 0.9) 30(1 sin 1) + 16(1.1 sin 1.1) (1.2 sin 1.2)
0.12
0.23913096
formula
h4
EC (f, h) = f (5) ((x1 ))
30
h4
(5)
|f ((x1 ))|
|EC (f, h)| =
30
Since f (5) (x) = 5 sin x+x cos x and f (5) ((x1 )) cannot be computed exactly because
value for |f (5) ((x1 ))|. So bound |f (5) | on [0.8, 1.2] can be obtain
M=
max |f (5) (x)| =
0.8x1.2
max |5 sin x + x cos x| = 5.0950247
0.8x1.2

|EC (f, h)|
A.BENHARI
(0.1)4
(0.1)4
M=
(5.0950247) = 0.0000170
30
30
113
22. Consider the following data

(0.2, 0.39), (0.4, 1.08), (0.6, 1.49), (0.8, 1.78), (1, 2), (1.2, 2.18), (1.4, 2.34)
Use the five-point formula to compute the best approximation for f 00 (0.6) and
f 00 (1.0). The data in this problem were taken from the function f (x) = ln x + 2.
Compute the actual errors and also, find error bounds of your approximations.
Solution: Using five-point formula for finding the approximation of f 00 (0.6) with
h = 0.2, we have
f 00 (0.6)
f (0.2) + 16f (0.4) 30f (0.6) + 16f (0.8) f (1.0)

12(0.04)
0.39 + 17.28 44.70 + 28.48 2.0

= 2.7708333
0.48
Now using five-point formula for finding the approximation of f 00 (1.0) with h = 0.2,
we have
f 00 (1.0)
f (0.6) + 16f (0.8) 30f (1.0) + 16f (1.2) f (1.4)

12(0.04)
1.49 + 28.48 60.0 + 34.88 2.34

= 0.9791667
0.48
Since the exact solution of the second derivative f 00 (x) = 1/x2 of the given function f (x) = ln x + 2 at x = 0.6 is 2.7777778, so the corresponding actual error
is, 0.0069445. Also, the exact solution of f 00 (x) at x = 1.0 is 1.0, so the corresponding actual error is, 0.0208333.
To compute the error bound for the approximation of f 00 (0.6), we have
|EC (f, h)|
(0.2)4
M
30
where
M=
max |24/x5 | = 75000
0.2x1.0
So
(0.2)4
75000 = 4.0
30
Similarly, we have for the approximation of f 00 (1.0)
|EC (f, h)|
|EC (f, h)|
(0.2)4
M
30
where
M=
max |24/x5 | = 308.64198
0.6x1.2
and
|EC (f, h)|
A.BENHARI
(0.2)4
308.64198 = 0.0164609
30
114
23. Let f (x) = 1/x, with h = 0.2. Use the five-point formula to approximate f 00 (2).
Find error bound for your approximation.
Solution: Given f (x) = 1/x, x1 = 2, h = 0.2, then using five-point formula for
finding the approximation of f 00 (2.0), we have
f 00 (2)
f (2 0.4) + 16f (2 0.2) 30f (2) + 16f (2 + 0.2) f (2 + 0.4)

12(0.04)
f (1.6) + 16f (1.8) 30f (2) + 16f (2.2) f (2.4)

0.48
0.625 + 8.8889 15.0 + 7.2727 0.4167

= 0.2497917
0.48
To compute the error bound for the approximation of f 00 (2), we have

(0.2)4
M
|EC (f, h)|
30
where
M = max | 120/x6 | = 7.1525574
1.62.2
So
(0.2)4
7.1525574 = 0.0003812
|EC (f, h)|
30
24. For the three-point central-difference formula for f 0 (x), perform an error analysis
to show that the optimum stepsize hopt is given by
s
h = hopt =
3
10t
2M
where M = max |f 000 (x)|.

Solution: Since
f (x2 ) f (x)
2h
where x2 = x1 + h and x0 = x1 h. Taking
(1)
f 0 (x1 ) Dh =
f (xi ) fi = i ,
Then
i = 0, 1, 2
h2
2 0
(1)
f 0 (x1 ) Dh f (x1 ) f 000 () +
6
2h
and
2
h2
(1)
|f 0 (x1 ) Dh f (x1 )| M +
6
2h
000
where M = max|f (x)|, 0 , 1 , 2 are generally random in some interval [, ] and
is a bound on the experimental error. Let
E(h) =
A.BENHARI
2
h2
M+
6
2h
115
and E 0 (h) = 0 gives
h = hopt =
where =
3
10t
2M
1
10t .
2
25. Approximate the integral

n = 4, and n = 6.
Z 2
0
x2 ex dx, using composite Trapezoidal rules with
Solution: Given f (x) = x2 ex , then for n = 4 and h = 0.5, we have the Trapezoidal rule of the form
i
0.5 h
f (0) + 2[f (0.5) + f (1.0) + f (1.5)] + f (2) = 14.4980
T4 (f ) =
2
which gives
i
0.5 h
2
2
2
2
0+2[(0.5)2 e(0.5) +(1.0)2 e(1.0) +(1.5)2 e(1.5) ]+(2.0)2 e(2.0) = 0.4182
2
For n = 6 and h = 1/3, we get
T4 (f ) =
1/3 h
2
2
2
2
0 + 2[(1/3)2 e(1/3) + (2/3)2 e(2/3) + (1.0)2 e(1.0) + (4/3)2 e(4/3)
2
i
2
2
+ (5/3)2 e(5/3) ] + (2.0)2 e(2.0) = 0.4207
T6 (f ) =
26. The following values of a function f (x) = tan x/x are given
x
1.00
1.10
1.20
1.30
Find
Z 1.6
1.0
f(x)
1.5574
1.7862
2.1435
2.7709
x
1.40
1.50
1.60
f(x)
4.1342
9.4009
-21.3953
f (x)dx, using the Trapezoidal rule with h = 0.1.
Solution: Given f (x) = tan x/x, then for h = 0.1 and n = 6, we have the
Trapezoidal rule of the form
T6 (f ) =
i
0.1 h
1.5574+2[1.7862+2.1435+2.7709+4.1342+9.4009]+21.3953 = 1.0316750
2
27. Use a suitable composite integration formula to approximate the integral

with n = 5.
Z 1
0
dx
,
2ex 1
Solution: Given f (x) = 1/(2ex 1), then for n = 5 and h = 0.2, we have the
0.2 h
1/(2e0 1) + 2[1/(2e0.2 1) + 1/(2e0.4 1) + 1/(2e0.6 1)
2
i
+ 1/(2e0.8 1)] + 1/(2e1.0 1) = 0.49557207
T5 (f ) =
A.BENHARI
116
28. ZUse a suitable composite integration formula for the approximation of the integral
2 dx
, with n = 5. Compute an upper bound for your approximation.
1 3x
Solution: Given f (x) = 1/(3 x), then for n = 5 and h = 0.2, we have the
T5 (f ) =
i
0.2 h
1/(31)+2[1/(31.2)+1/(31.4)+1/(31.6)+1/(31.8)]+1/(32)
2
it gives
T5 (f ) = 0.6956349
To compute error bound for the approximation, we use the error formula
ET5 (f ) =
h2
(b a)f 00 ((x)),
12
for (x) (1, 2)
or
h2 00
|ET5 (f )| = |f ((x))|, for (x) (1, 2)
12
00
3
Since f (x) = 2/(3 x) and f 00 ((x)) cannot be computed exactly because (x) is
not known. But one can bound the error by computing the largest possible value
for |f 00 ((x))|. So bound |f 00 | on [1, 2] can be obtain
M = max |f 00 (x)| = max |2/(3 x)3 | = 2.0
1x2
1x2
Since |f 00 ((x))| M , therefore

(0.2)2
(2.0) = 0.006667
|ET5 (f )|
12
29. Use the composite Trapezoidal rule for the approximation of the integral
with h = 0.5. Also, compute an error term.
Z 3
1
dx
7 2x
Solution: Given f (x) = 1/(7 2x), then for h = 0.5 and n = 4, we have the
Trapezoidal rule of the
T5 (f ) =
i
0.5 h
1/(7 2) + 2[1/(7 3) + 1/(7 4) + 1/(7 5)] + 1/(7 6) = 0.8416667
2
To compute error bound for the approximation, we use

|ET4 (f )|
(0.5)2
(3 1)M
12
where
M = max |f 00 (x)| = max |8/(7 2x)3 | = 8.0
1x3
1x3
and
(0.5)2
(2)(8.0) = 0.3333333
12
|ET4 (f )|
A.BENHARI
117
30. Find the step size h so that the absolute value of the error for the composite
Trapezoidal rule is less than 5 104 when it is used to approximate the integral
Z 7
dx
.
2 x
Solution: To find the step size h for the given accuracy, we use the composite
Trapezoidal error formula as
|ETn (f )|
| (b a)| 2
h M 5 104
12
As
M = max |f 00 (x)| = max |2/x3 | = 0.25
2x7
2x7
so solving for h2 , we obtain

h2 0.0048,
31. Estimate the integral
Z 1
1
gives h 0.06928 or h = 0.05
dx
using the Simpsons rules with n = 8.
1 + x2
Solution: Given f (x) = 1/(1 + x2 ),then for n = 8 and h = 0.25, we have the
Simpsons rule of the form
Z x8
x0
h
[f (x0 ) + 4(f (x1 ) + f (x3 ) + f (x5 ) + f (x7 ))
3
+ 2(f (x2 ) + f (x4 ) + f (x6 )) + f (x8 )]
f (x)dx
or
Z 1
1
0.25
[f (1) + 4[f (0.75) + f (0.25) + f (0.25) + f (0.75)]
3
+ 2[f (0.5) + f (0.0) + f (0.5)] + f (1)]
1/(1 + x2 )dx
It gives
Z 1
1
0.25 h
0.5 + 4[0.64 + 0.9412 + 0.9412 + 0.64)]
3
i
+ 2[0.8 + 1.0 + 0.8] + 0.5 = 1.570800
1/(1 + x2 )dx
32. Repeat the Problem 26 by using the composite Simpsons rule.

Solution: The Simpsons rule for n = 6 can be written as
Z x6
x0
f (x)dx
h
[f (x0 ) + 4(f (x1 ) + f (x3 ) + f (x5 ))) + (f (x2 ) + f (x4 )) + f (x6 )]
3
Using the given values, we have

Z 1.6
1.0
A.BENHARI
0.1 h
1.5574 + 4(1.7862 + 2.7709 + 9.4009)
3
i
+ 2(2.1435 + 4.1342) 21.3953 = 1.6183167
tan x/x dx
118
33. Repeat the Problem 25 by using the composite Simpsons rule.

2
Solution: Given f (x) = x2 ex , then for n = 4 and h = 0.5, we have the Simpsons
rule of the form
i
0.5 h
f (0) + 4(f (0.5) + f (1.5)) + 2f (1.0) + f (2) = 14.4980
S4 (f ) =
3
which gives
S4 (f ) =
i
0.5 h
2
2
2
2
0+4((0.5)2 e(0.5) +(1.5)2 e(1.5) )+2(1.0)2 e(1.0) +(2.0)2 e(2.0) = 0.4227
3
Similarly, for n = 6 and h = 1/3, we get

1/3 h
2
2
2
2
0 + 4[(1/3)2 e(1/3) + (1.0)2 e(1.0) + (5/3)2 e(5/3) ] + 2[(2/3)2 e(2/3)
3
i
2
2
+ (4/3)2 e(4/3) ] + (2.0)2 e(2.0) = 0.4227
S6 (f ) =
34. ZUse the composite Trapezoidal and the Simpsons rules to approximate the integral
2
dx
such that the error does not exceed 102 .
1 4x + 1
dx
a = 1, b = 2, and = 102 , then using Trapezoidal
4x + 1
and Simpsons rules for the given accuracy, we need n = 2. Thus for n = 2, we get
the approximations from both rules as follows:
Solution: Given f (x) =
T2 (f ) = 0.1492
Z 1
and S2 (f ) = 0.1471
35. Evaluate
ex dx by the Simpsons rule choosing h small enough to guarantee five
0
decimal accuracy. How large can h be ?
2
Solution: Given f (x) = ex , a = 0, b = 1, and = 105 , then using the error

formula of the Simpsons rule, we have
|E|
where
M=
1 4
h M 105
180
max |4ex (3 + 16x2 + 4x4 )| = 206.5894
0.0x1.0
x2
and f (4) (x) = 4e (3 + 16x2 + 4x4 ). So

1 4
h (206.5894) 105
180
and solving for h, we get
h4 8.7129 106 ,
gives h 0.0543
Thus for h = 0.05 and n = 20, we have approximate value of the given integral
using Simpsons rule is
S20 = 1.4627
A.BENHARI
119
composite integration rule to find the best approximate value of the

36. Use a suitable
Z 2
integral
1 + sin x dx, with h = 0.1. Estimate the error bound.
1
Solution: Given f (x) = 1 + sin x, a = 1, b = 2, h = 0.1, then for n = 10, the

suitable composite integration rule is the Simpsons rule. Using Simpsons rule for
h = 0.1, we get
S10 (f ) = 1.3987
To compute the error bound, we use
|E|
1 4
hM
180
where
M=
So
|E|
max |f (4) (x)| = 0.1037
1.0x2.0
1
(0.1)4 (0.1037) = 5.7611 108
180
Z 2
37. Evaluate the integral

ln x dx using closed Newton-Cotes formulas for n = 3 and
1
n = 4. Also, compute the error bounds for your approximations.
Solution: Given f (x) = ln x, a = 1, b = 2, then for n = 3 (Simpsons 3/8 rule),
h = 1/3, we have
ISR =
i
3h h
f (x0 ) + 3f (x1 ) + 3f (x2 ) + f (x3 )
8
i
1h
f (1) + 3f (4/3) + 3f (5/3) + f (2)
8
i
1h
ln 1 + 3 ln 4/3 + 3 ln 5/3 + ln 2
8
= 0.3861
To compute the error bound for the approximation, we have
|E|
3
(1/3)5 M
80
where
M=
max | 6/x4 | = 6.0
1.0x2.0
and f (4) (x) = 6/x4 . So

|E|
A.BENHARI
3
(1/3)5 (6.0) = 0.0009
80
120
Now for n = 4(Booles rule) and h = 1/4, we have

i
2h h
7f (x0 ) + 32f (x1 ) + 12f (x2 ) + 32f (x3 ) + 7f (x4 )
45
IB =
=
i
1h
7f (1) + 32f (5/4) + 12f (6/4) + 32f (7/4) + 7f (2)
90
i
1h
7 ln 1 + 32 ln 5/4 + 12 ln 3/2 + 32 ln 7/4 + 7 ln 2
90
= 0.3862
|E|
8
(1/4)7 M
945
where
M=
max | 120/x6 | = 120
1.0x2.0
and f (6) (x) = 120/x6 . So

|E|
8
(1/4)7 (120) = 0.00006
945
38. Evaluate the following integrals using closed Newton-Cotes formulas for n = 4, 5,
and 6. Also, compute the error bounds for your approximations:
(a)
Z 1
0
2x
(x + e ) dx (b)
Z /2
0
cos x dx (c)
Z 3
2
(x + ln x) dx.
Solution: (a) Given f (x) = (x + e2x ), a = 0, b = 1, then for n = 4 ((Booles

rule)), h = 1/4, we have
IB =
i
2h h
7f (x0 ) + 32f (x1 ) + 12f (x2 ) + 32f (x3 ) + 7f (x4 )
45
i
1h
7f (0) + 32f (1/4) + 12f (2/4) + 32f (3/4) + 7f (1)
90
i
1h
7(1) + 32(1/4 + e1/2 ) + 12(1/2 + e1 ) + 32(3/4 + e3/2 ) + 7(1 + e2 )
90
= 3.69462
|E|
8
(1/4)7 M
945
where
M=
A.BENHARI
max |64e2x | = 472.8996
0.0x1.0
121
and f (6) (x) = 64e2x . So

|E|
8
(1/4)7 (472.8996) = 0.00024
945
For n = 5 ((six-point rule)), h = 1/5, we have

Isp =
i
5h h
19f (x0 ) + 75f (x1 ) + 50f (x2 ) + 50f (x3 ) + 75f (x4 ) + 19f (x5 )
288
i
1 h
19(f (0) + 75f (1/5) + 50f (2/5) + 50f (3/5) + 75f (4/5) + 19f (1)
288
1 h
19(1) + 75(1/5 + e2/5 ) + 50(2/5 + e4/5 ) + 50(3/5 + e6/5 )
288
i
+ 75(4/5 + e8/5 ) + 19(1 + e2 )

= 3.69458

|E|
275
(1/5)7 M
12096
where
M=
max |64e2x | = 472.8996
0.0x1.0
and f (6) (x) = 64e2x . So

|E|
275
(1/5)7 (472.8996) = 0.00014
12096
Z 2
x + 1 dx using the Weddles rule and compute the error

0
bound for your approximation.
IW =
=
x + 1, a = 0, b = 2, then for n = 6, h = 1/3, we have
i
3h h
(f (x0 ) + f (x2 ) + f (x4 ) + f (x6 )) + 5(f (x1 ) + f (x5 )) + 6f (x3 )
10
1h
(f (0) + f (2/3) + f (4/3) + f (2))
10
i
+ 5(f (1/3) + f (5/3)) + 6f (1)
q
q
1h
( 0 + 1 + 2/3 + 1 + 4/3 + 1 + 2 + 1)
10
q
q
i
+ 5( 1/3 + 1 + 5/3 + 1) + 6 1 + 1
= 2.7974
A.BENHARI
122

1
(1/3)7 M
|E|
140
where
M = max | 945/64(x + 1)11/2 | = 14.765625
0.0x2.0
and f
(6)
(x) = 945/64(x + 1)11/2 . So

|E|
1
(1/3)7 (14.765625) = 0.00005
140
40. Evaluate the following integrals using closed Newton-Cotes formulas for n =
1, 2, 3, 4, 5, and 6. Also, compute the error bounds for your approximations:
Z 2
Z /2
Z 1
( x + 1 + ln x) dx.
(sin x x) dx (c)
(x2 + x + 1) dx (b)
(a)
1
Solution: (a) Given f (x) = x2 + x + 1, a = 0, b = 1, then for n = 3 (Simpsons

3/8 rule), h = 1/3, we have
ISR =
i
3h h
f (x0 ) + 3f (x1 ) + 3f (x2 ) + f (x3 )
8
i
1h
f (0) + 3f (1/3) + 3f (2/3) + f (1)
8
q
q
i
1h
(1 + 3((1/3)2 + 1/3 + 1) + 3((2/3)2 + 2/3 + 1) + (1 + 2)
8
= 1.55223
3
|E| (1/3)5 M
80
where
M = max | 15/16(x + 1)7/2 | = 0.9375
0.0x1.0
and f
(4)
(x) = 15/16(x + 1)7/2 . So
3
(1/3)5 (0.9375) = 0.00015
80
Now for n = 4(Booles rule) and h = 1/4, we have
|E|
IB =
=
i
2h h
7f (x0 ) + 32f (x1 ) + 12f (x2 ) + 32f (x3 ) + 7f (x4 )
45
i
1h
7f (0) + 32f (1/4) + 12f (2/4) + 32f (3/4) + 7f (1)
90
q
q
1h
7(1) + 32((1/4)2 + 1/4 + 1) + 12((1/2)2 + 1/2 + 1) + 32((3/4)2
90
q
i
3/4 + 1) + 7(1 + 2)
+
= 1.39134
A.BENHARI
123

|E|
8
(1/4)7 M
945
where
M=
max | 945/64(x + 1)11/2 | = 14.7656
0.0x1.0
and f (6) (x) = 945/64(x + 1)11/2 . So

|E|
and n = 2.
8
(1/4)7 (14.7656) = 0.000008
945
Z 1
x + 4 dx using open Newton-Cotes formulas for n = 0, 1,
x + 4, a = 0, b = 1, then for n = 0, h = 0.5, we have
= 2hf (x0 ) = 2(0.5)[ 0.5 + 4] = 2.1213

IM 0
For n = 1, h = 1/3, we have

q
i
i
3h h
1 hq
1/3 + 4 + 2/3 + 4
f (x0 ) + f (x1 ) =
2
2
IM 1 =
= 2.1210
IM 2 =
4h
[2f (x0 ) f (x1 ) + 2f (x2 )]
3
1
[2f (0.25) f (0.5) + 2f (0.75)]
3
1
[2 0.25 + 4 0.5 + 4 + 2 0.75 + 4] = 2.1202
3
42. Evaluate the following integrals using open Newton-Cotes formulas for n = 0, 1, 2, 3, 4,
and 5. Also, compute the error bounds for your approximations:
(a)
Z 2
1
(x + e ) dx (b)
Z
0
(x + sin x) dx (c)
Z 3
2
ln(x2 + 4) dx.
Solution: Given f (x) = (x + ex ), a = 1, b = 2, then for n = 0, h = 0.5, we have

IM 0 = 2hf (x0 ) = 2(0.5)[1.5 + e1.5 ] = 5.9817
1
|E| (0.5)3 M
3
A.BENHARI
124
where
M=
max |ex | = 7.3891
1.0x2.0
and f 00 (x) = ex . So
1
|E| (0.5)3 (7.3891) = 0.3079
3
IM 1 =
=
i
3h h
f (x0 ) + f (x1 )
2
i
1h
(4/3 + e4/3 ) + (5/3 + e5/3 )
2
= 6.0441
3
|E| (1/3)3 M
4
where
M=
max |ex | = 7.3891
1.0x2.0
and f 00 (x) = ex . So
3
|E| (1/3)3 (7.3891) = 0.2053
4
IM 2 =
4h
[2f (x0 ) f (x1 ) + 2f (x2 )]
3
1
[2f (1.25) f (1.5) + 2f (1.75)]
3
1
[2(1.25 + e1.25 ) (1.5 + e1.5 ) + 2(1.75 + e1.75 ] = 6.1694
3

|E|
14
(1/4)5 M
45
where
M=
and f (4) (x) = ex . So
|E|
A.BENHARI
max |ex | = 7.3891
1.0x2.0
14
(1/4)3 (7.3891) = 0.0023
45
125

IM 3 =
5h
[11f (x0 ) + f (x1 ) + f (x2 ) + 11f (x3 )]
24
1
[11f (1.2) + f (1.4) + f (1.6) + 11f (1.8)]
24
1
[11(1.2 + e1.2 ) + (1.4 + e1.4 ) + (1.6 + e1.6 + (1.8 + e1.8 )]
24
= 6.1698
|E|
95
(1/5)5 M
144
where
M=
max |ex | = 7.3891
1.0x2.0
and f (4) (x) = ex . So

|E|
95
(1/5)3 (7.3891) = 0.0016
144

IM 4 = (6h/20)[f (x0 ) + f (x2 ) + f (x4 )]
= (1/20)[11(f (7/6) + f (11/6)) 14(f (8/6) + f (10/6)) + 26f (9/6)]
= (1/20)[11[(7/6 + e7/6 ) + (11/6 + e11/6 )] 14[(8/6 + e8/6 ) + (10/6 + e10/6 )]
+ 26(9/6 + e9/6 )] = 6.1707
|E|
41
(1/6)7 M
144
where
M=
max |ex | = 7.3891
1.0x2.0
and f (6) (x) = ex . So

|E|
41
(1/6)7 (7.3891) = 0.000008
144
Z 2

ln(x + 1) dx using the four-point open Newton-Cotes
1
formula and compute the error bound for your approximation.
A.BENHARI
126
Solution: Given f (x) = ln(x + 1), a = 1, b = 2, n = 4, then for h = 1/6, we

have
IM 4 = (6h/20)[f (x0 ) + f (x2 ) + f (x4 )]
= (1/20)[11(f (7/6) + f (11/6)) 14(f (8/6) + f (10/6)) + 26f (9/6)]
= (1/20)[11(ln(7/6 + 1) + ln(11/6 + 1)) 14(ln(8/6 + 1) + ln(10/6 + 1))
+ 26 ln(9/6 + 1)] = 0.90954304
|E|
41
(1/6)7 M
140
where
M=
max | 120/(x + 1)6 | = 1.8750
1.0x2.0
and f (6) (x) = 120/(x + 1)6 . So

|E|
A.BENHARI
41
(1/6)7 (1.8750) = 0.000002
140
127
Chapter 6
Ordinary Differential Equations
1. Find the general solution of the differential equation

y0 =
x
y
Solution: Since the given differential equation is separable, therefore, we can write
in the form
x
dy
=
or ydy = xdx
dx
y
from which it follows that
ydy =
xdx
x2
+ c1
2
y2
2
y2
= x2 + c
2. Put the following differential equations into a form for numerical solution by the
Eulers method.
(a) y + 2yy 0 y 0 = 0
(b) ln y 0 = x2 y 2
(c) y 0 x2 y 0 = y
Solution:
(a) y 0 = f (x, y) =
y
1 2y
2 y 2
(b) y 0 = f (x, y) = ex
(c) y 0 = f (x, y) =
y
1 x2
3. Solve the following initial-value problems using the Eulers method.

(a) y 0 = y + x2 , x = 0(0.2)1, y(0) = 1.
(b) y 0 = (x 1)(x + y + 1), x = 0(0.2)1, y(0) = 1.
(c) y 0 = y + sin(x), x = 0.2(0.01)0.25, y(0.2) = 1.5.
Solution: (a) Since f (x, y) = y + x2 , and x0 = 0, y0 = 1, then
yi+1 = yi + hf (xi , yi ),
for i = 0, 1, . . . , 4
For h = 0.2 and taking i = 0, we have

y1 = y0 + hf (x0 , y0 ) = y0 + h(y0 + x20 ) = 1 + (0.2)[1 + 0]
which gives
y1 = 1.2000
A.BENHARI
128
This and other approximations by taking xi = xi1 + h, i = 1, 2, . . . , 4, are as

follows
xi
yi
0.2000
0.4000
0.6000
0.8000
1.0000
1.2000
1.4480
1.7696
2.1955
2.7626
(b) Since f (x, y) = (x 1)(x + y + 1), and x0 = 0, y0 = 1, then

yi+1 = yi + hf (xi , yi ),
for i = 0, 1, . . . , 4

y1 = y0 + hf (x0 , y0 ) = y0 + h[(x0 1)(x0 + y0 + 1)] = 1 + (0.2)[(0 1)(0 + 1 + 1)]
which gives
y1 = 0.6000
follows
xi
yi
0.2000 0.6000
0.4000 0.3120
0.6000 0.1066
0.8000 0.0300
1.0000 0.1008
(c) Since f (x, y) = y + sin(x), and x0 = 0.2, y0 = 1.5, then
yi+1 = yi + hf (xi , yi ),
for i = 0, 1, . . . , 4

y1 = y0 + hf (x0 , y0 ) = y0 + h[y0 + sin(x0 )] = 1.5 + (0.01)[1.5 + 0.1987]
which gives
y1 = 1.5170
follows
xi
yi
0.2100
0.2200
0.2300
0.2400
0.2500
A.BENHARI
1.5170
1.5342
1.5518
1.5696
1.5876
129
4. Solve the following initial-value problems and compare the numerical solutions obtained with the Eulers method using the values of h = 0.1 and h = 0.2. Compare
the results to the actual values.
(a) y 0 = 1 + x2 ,
0 x 1, y(0) = 0, y(x) = tan x.
0
(b) y = 2(y + 1), 0 x 1, y(0) = 0, y(x) = e2x 1.
(c) y 0 = 2(y 1)2 , 1 x 2, y(1) = 0.5, y(x) = (2x 1)/2x.
Solution: (a) Since f (x, y) = 1 + x2 , and x0 = 0, y0 = 0, then
yi+1 = yi + hf (xi , yi ),
for i = 0, 1, . . . , 9

y1 = y0 + hf (x0 , y0 ) = y0 + h(1 + x20 ) = 0 + (0.1)[1 + 0]
which gives
y1 = 0.1000
follows
xi
yi
y(xi ) y(xi ) yi
0.1000
0.2000
0.3000
0.4000
0.5000
0.6000
0.7000
0.8000
0.9000
1.0000
0.1000
0.2010
0.3050
0.4140
0.5300
0.6550
0.7910
0.9400
1.1040
1.2850
0.1003
0.2027
0.3093
0.4228
0.5463
0.6841
0.8423
1.0296
1.2602
1.5574
0.0003
0.0017
0.0043
0.0088
0.0163
0.0291
0.0513
0.0896
0.1562
0.2724

y1 = y0 + hf (x0 , y0 ) = y0 + h(1 + x20 ) = 0 + (0.2)[1 + 0]
which gives
y1 = 0.2000
follows
xi
yi
y(xi ) y(xi ) yi
0.2000
0.4000
0.6000
0.8000
1.0000
0.2000
0.4080
0.6400
0.9120
1.2400
0.2027
0.4228
0.6841
1.0296
1.5574
0.0027
0.0148
0.0441
0.1176
0.3174
(b) Since f (x, y) = 2(y + 1), and x0 = 0, y0 = 0, then

yi+1 = yi + hf (xi , yi ),
A.BENHARI
for i = 0, 1, . . . , 9
130

y1 = y0 + hf (x0 , y0 ) = y0 + h(2y0 + 2) = 0 + (0.1)[0 + 2]
which gives
y1 = 0.2000
follows
xi
yi
y(xi ) y(xi ) yi
0.1000
0.2000
0.3000
0.4000
0.5000
0.6000
0.7000
0.8000
0.9000
1.0000
0.2000
0.4400
0.7280
1.0736
1.4883
1.9860
2.5832
3.2998
4.1598
5.1917
0.2214
0.4918
0.8221
1.2255
1.7183
2.3201
3.0552
3.9530
5.0496
6.3891
0.0214
0.0518
0.0941
0.1519
0.2300
0.3341
0.4720
0.6532
0.8899
1.1973

y1 = y0 + hf (x0 , y0 ) = y0 + h(2y0 + 2) = 0 + (0.2)[0 + 2]
which gives
y1 = 0.4000
follows
xi
yi
y(xi ) y(xi ) yi
0.2000
0.4000
0.6000
0.8000
1.0000
0.4000
0.9600
1.7440
2.8416
4.3782
0.4918
1.2255
2.3201
3.9530
6.3891
0.0918
0.2655
0.5761
1.1114
2.0108
(c) Since f (x, y) = 2(y 1)2 , and x0 = 1, y0 = 0.5, then

yi+1 = yi + hf (xi , yi ),
for i = 0, 1, . . . , 9

y1 = y0 + hf (x0 , y0 ) = y0 + 2h[(y0 2)2 ] = 0.5 + (0.2)[(0.5 1)2 ]
which gives
y1 = 0.2000
A.BENHARI
131

follows
xi
yi
y(xi ) y(xi ) yi
1.1000
1.2000
1.3000
1.4000
1.5000
1.6000
1.7000
1.8000
1.9000
2.0000
0.5500
0.5905
0.6240
0.6523
0.6765
0.6974
0.7157
0.7319
0.7463
0.7591
0.5455
0.5833
0.6154
0.6429
0.6667
0.6875
0.7059
0.7222
0.7368
0.7500
0.0045
0.0072
0.0087
0.0095
0.0098
0.0099
0.0098
0.0097
0.0094
0.0091

y1 = y0 + hf (x0 , y0 ) = y0 + 2h[(y0 2)2 ] = 0.5 + (0.4)[(0.5 1)2 ]
which gives
y1 = 0.6000
follows
xi
yi
y(xi ) y(xi ) yi
1.2000
1.4000
1.6000
1.8000
2.0000
0.6000
0.6640
0.7092
0.7430
0.7694
0.5833
0.6429
0.6875
0.7222
0.7500
0.0167
0.0211
0.0217
0.0208
0.0194
5. Solve the following initial-value problems using the Taylors method of order two.
(a) y 0 = 2x2 y, x = 0(0.2)1, y(0) = 1.
(b) y 0 = 3x2 y, x = 0(0.2)1, y(0) = 1.
(c) y 0 = x/y x, x = 0(0.2)1, y(0) = 2.
Solution: (a) Since f (x, y) = 2x2 y, f 0 (x, y) = 4x2x2 +y, and x0 = 0, y0 = 1,
then using second-order Taylors method
yi+1 = yi + hf (xi , yi ) +
h2 0
f (xi , yi ),
2
for i = 0, 1, 2, 3, 4
for h = 0.2 and taking i = 0, we have

y1 = y0 + hf (x0 , y0 ) +
h2
h2 0
f (x0 , y0 ) = y0 + 2h[2x20 y0 ] + (4x0 2x20 + y0 )
2
2
which gives
y1 = 1 + (0.2)[0 + 1] + 0.02[0 0 1] = 0.8200
A.BENHARI
132

follows
xi
yi
0.2000
0.4000
0.6000
0.8000
1.0000
0.8200
0.6420
0.4368
0.1806
0.1463
(b) Since f (x, y) = 3x2 y, f 0 (x, y) = 6xy + 9x4 y, and x0 = 0, y0 = 1, then using
h2 0
yi+1 = yi + hf (xi , yi ) + f (xi , yi ),
2
y1 = y0 + hf (x0 , y0 ) +
for i = 0, 1, 2, 3, 4
h2
h2 0
f (x0 , y0 ) = y0 + 2h[3x20 y0 ] + [6x0 y0 + 9x40 y0 ]
2
2
which gives
y1 = 1 + (0.2)[0] + 0.02[0 + 0] = 1.0000
follows
xi
yi
0.2000
0.4000
0.6000
0.8000
1.0000
1.0000
1.0483
1.2041
1.5789
2.4532
(c) Since f (x, y) = x/y x, f 0 (x, y) = 1/y x2 /y 3 + x2 /y 2 1, and x0 = 0, y0 = 2,

then using
h2 0
f (xi , yi ),
2
yi+1 = yi + hf (xi , yi ) +
y1 = y0 +hf (x0 , y0 )+
for i = 0, 1, 2, 3, 4
h2 0
h2
f (x0 , y0 ) = y0 +2h[x0 /y0 x0 ]+ [1/y0 x20 /y03 +x20 /y02 1]
2
2
which gives
y1 = 2 + (0.2)[0] + 0.02[1/2 0 + 0 1] = 1.9900
follows
xi
yi
0.2000
0.4000
0.6000
0.8000
1.0000
A.BENHARI
1.9900
1.9603
1.9117
1.8458
1.7651
133
6. Solve the initial-value problems by using the Taylors method of order three of the
Problem 3.
Solution: (a) Since f (x, y) = y+x2 , f 0 (x, y) = y+x2 +2x, f 00 (x, y) = y+x2 +2x+2,
and x0 = 0, y0 = 1, then using third-order Taylors method
yi+1 = yi + hf (xi , yi ) +
h2 0
h3
f (xi , yi ) + f 00 (xi , yi ),
2
6
for i = 0, 1, 2, 3, 4

y1 = y0 + hf (x0 , y0 ) +
= y0 + h[y0 + x20 ] +
h2 0
h3
f (x0 , y0 ) + f 00 (x0 , y0 )
2
6
h3
h2
[y0 + x20 + 2x0 ] + [y0 + x20 + 2x0 + 2]
2
6
which gives
y1 = 1 + (0.2)[1 + 0] + 0.02[1 + 0 + 0] + 0.0013[1 + 0 + 0 + 2] = 1.2240
follows
xi
yi
0.2000
0.4000
0.6000
0.8000
1.0000
1.2240
1.5150
1.9054
2.4351
3.1525
(b) Since
f (x, y) = (x 1)(x + y + 1)
f 0 (x, y) = 2x + y + (x 1)2 (x + y + 1)
f 00 (x, y) = 2 + 3(x 1)(x + y + 1) + (x 1)2 (1 + (x 1)(x + y + 1))
and x0 = 0, y0 = 1, then using
yi+1 = yi + hf (xi , yi ) +
h3
h2 0
f (xi , yi ) + f 00 (xi , yi ),
2
6
for i = 0, 1, 2, 3, 4

y1 = y0 + hf (x0 , y0 ) +
h2 0
h3
f (x0 , y0 ) + f 00 (x0 , y0 )
2
6
= y0 + h[(x0 1)(x0 + y0 + 1)] +

+
A.BENHARI
h2
[2x0 + y0 + (x0 1)2 (x0 + y0 + 1)]
2
h3
[2 + 3(x0 1)(x0 + y0 + 1) + (x0 1)2 (1 + (x0 1)(x0 + y0 + 1))]
6
134
which gives
y1 = 1 + (0.2)(2) + 0.02(3) + 0.00133(5) = 0.6533
follows
xi
yi
0.2000
0.4000
0.6000
0.8000
1.0000
0.6533
0.3979
0.2174
0.1060
0.0666
(c) Since f (x, y) = y + sin(x), f 0 (x, y) = y + sin(x) + cos(x), f 00 (x, y) = y + cos(x),

and x0 = 0.2, y0 = 1.5, then using
h3
h2 0
f (xi , yi ) + f 00 (xi , yi ),
2
6
yi+1 = yi + hf (xi , yi ) +
y1 = y0 + hf (x0 , y0 ) +
for i = 0, 1, 2, 3, 4
h2 0
h3
f (x0 , y0 ) + f 00 (x0 , y0 )
2
6
= y0 + h[y0 + sin(x0 )] +
h3
h2
[y0 + sin(x0 ) + cos(x0 )] + [y0 + cos(x0 )]
2
6
which gives
y1 = 1.5 + (0.01)[1.5 + 0.1987] + 0.00005[1.5 + 0.1987 + 0.9801]
+ 0.0000002[1.5 + 0.9801] = 1.5171
follows
xi
yi
0.2100
0.2200
0.2300
0.2400
0.2500
1.5171
1.5345
1.5522
1.5701
1.5883
7. Solve the following initial-value problems using the Modified Eulers method.
(a) y 0 = y 2 x2 , x = 1(0.2)2,
y(1) = 1.
0
(b) y = x y/2x, x = 1(0.02)1.10, y(1) = 0.25.
(c) y 0 = 1/y 2 yx,
x = 1(0.2)2, y(1) = 1.
Solution: (a) Since f (x, y) = y 2 x2 , and x0 = 1, y0 = 1, h = 0.2, then for i = 0,
we have
k1 = f (x0 , y0 ) = (y02 x20 ) = 1.0000
k2 = f (x1 , y0 + hk1 ) = ((y0 + hk1 )2 x21 ) = (0.8)2 (1.2)2 = 0.9216
A.BENHARI
135
Now using these values in the Modified formula, we have

y1 = y0 +
h
[k1 + k2 ] = 1 + 0.1(1 + 0.9216) = 0.8078
2

follows
xi
yi
1.2000 0.8078
1.4000 0.6385
1.6000 0.5000
1.8000 0.3911
2.0000 0.3075
(b) Since f (x, y) = x y/2x, and x0 = 1, y0 = 0.25, h = 0.2, then for i = 0, we
have
k1 = f (x0 , y0 ) = (x0 y0 /2x0 ) = 0.8750
k2 = f (x1 , y0 + hk1 ) = (x1 (y0 + hk1 )/2x1 ) = 1.02 0.1311 = 0.8889
y1 = y0 +
h
[k1 + k2 ] = 0.25 + 0.01(0.8750 + 0.8889) = 0.2676
2

follows
xi
yi
1.0200 0.2676
1.0400 0.2856
1.0600 0.3037
1.0800 0.3222
1.1000 0.3410
(c) Since f (x, y) = 1/y 2 yx, and x0 = 1, y0 = 1, h = 0.2, then for i = 0, we
have
k1 = f (x0 , y0 ) = (1/y02 y0 x0 ) = 0
k2 = f (x1 , y0 + hk1 ) = (1/(y0 + hk1 )2 (y0 + hk1 )x1 ) = 1 1.2 = 0.2
y1 = y0 +
h
[k1 + k2 ] = 1 + 0.1(0 0.2) = 0.9800
2

follows
xi
yi
1.2000
1.4000
1.6000
1.8000
2.0000
A.BENHARI
0.9800
0.9432
0.9013
0.8611
0.8252
136
8. Solve the following initial-value problems and compare the numerical solutions obtained with the Modified Eulers method using the values of h = 0.05 and h = 0.1
and compare the results with the actual values.
3y
(a) y 0 = x + , 1 x 2, y(1) = 0, y(x) = x3 x2 .
x
(b) y 0 = y, 0 x 1, y(0) = 1, y(x) = (x + 2)2 /4.
(c) y 0 = 4 3y, 0 x 1, y(0) = 5, y(x) = 4/3 + 11/3e3x .
Solution: (a) Since f (x, y) = x + 3y/x, and x0 = 1, y0 = 0, h = 0.05, then for
i = 0, we have
k1 = f (x0 , y0 ) = (x0 + (3y0 )/x0 ) = 1
k2 = f (x1 , y0 + hk1 ) = (x1 + 3(y0 + hk1 )/x1 ) = 1.05 + 0.4129 = 1.1929
y1 = y0 +
h
[k1 + k2 ] = 0 + 0.025(1 + 1.1929) = 0.0548
2
This and other approximations by taking xi = xi1 + h, i = 1, 2, . . . , 19, and the

possible errors, are as follows
A.BENHARI
xi
yi
y(xi )
y(xi ) yi
1.0500
1.1000
1.1500
1.2000
1.2500
1.3000
1.3500
1.4000
1.4500
1.5000
1.5500
1.6000
1.6500
1.7000
1.7500
1.8000
1.8500
1.9000
1.9500
2.0000
0.0548
0.1203
0.1973
0.2864
0.3885
0.5043
0.6345
0.7799
0.9412
1.1192
1.3146
1.5282
1.7607
2.0128
2.2854
2.5792
2.8948
3.2331
3.5948
3.9806
0.0551
0.1210
0.1984
0.2880
0.3906
0.5070
0.6379
0.7840
0.9461
1.1250
1.3214
1.5360
1.7696
2.0230
2.2969
2.5920
2.9091
3.2490
3.6124
4.0000
0.0003
0.0007
0.0011
0.0016
0.0021
0.0027
0.0034
0.0041
0.0049
0.0058
0.0068
0.0078
0.0089
0.0102
0.0115
0.0128
0.0143
0.0159
0.0176
0.0194
137
Similarly, for h = 0.1, we get
(b) Since f (x, y) =
xi
yi
y(xi )
y(xi ) yi
1.1000
1.2000
1.3000
1.4000
1.5000
1.6000
1.7000
1.8000
1.9000
2.0000
0.1186
0.2824
0.4973
0.7692
1.1040
1.5077
1.9862
2.5453
3.1910
3.9293
0.1210
0.2880
0.5070
0.7840
1.1250
1.5360
2.0230
2.5920
3.2490
4.0000
0.0024
0.0056
0.0097
0.0148
0.0210
0.0283
0.0368
0.0467
0.0580
0.0707
y, and x0 = 0, y0 = 1, h = 0.05, then for i = 0, we have
k1 = f (x0 , y0 ) = y0 = 1
k2 = f (x1 , y0 + hk1 ) = y0 + hk1 = 1.05 = 1.0247

y1 = y0 +
h
[k1 + k2 ] = 1 + 0.025(1 + 1.0247) = 1.0506
2

A.BENHARI
xi
yi
y(xi )
y(xi ) yi
0.0500
0.1000
0.1500
0.2000
0.2500
0.3000
0.3500
0.4000
0.4500
0.5000
0.5500
0.6000
0.6500
0.7000
0.7500
0.8000
0.8500
0.9000
0.9500
1.0000
1.0506
1.1025
1.1556
1.2100
1.2656
1.3225
1.3806
1.4399
1.5006
1.5624
1.6255
1.6899
1.7555
1.8224
1.8905
1.9599
2.0305
2.1024
2.1755
2.2498
1.0506
1.1025
1.1556
1.2100
1.2656
1.3225
1.3806
1.4400
1.5006
1.5625
1.6256
1.6900
1.7556
1.8225
1.8906
1.9600
2.0306
2.1025
2.1756
2.2500
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0002
138

xi
yi
y(xi )
y(xi ) yi
0.1000
0.2000
0.3000
0.4000
0.5000
0.6000
0.7000
0.8000
0.9000
1.0000
1.1024
1.2099
1.3223
1.4398
1.5622
1.6896
1.8221
1.9595
2.1020
2.2494
1.1025
1.2100
1.3225
1.4400
1.5625
1.6900
1.8225
1.9600
2.1025
2.2500
0.0001
0.0001
0.0002
0.0002
0.0003
0.0004
0.0004
0.0005
0.0005
0.0006
(c) Since f (x, y) = 4 3y, and x0 = 0, y0 = 5, h = 0.05, then for i = 0, we have

k1 = f (x0 , y0 ) = 4 3y0 = 11
k2 = f (x1 , y0 + hk1 ) = 4 3(y0 + hk1 ) = 4 3(4.45) = 9.3500
y1 = y0 +
h
[k1 + k2 ] = 5 + 0.025(11 9.3500) = 4.4913
2

A.BENHARI
xi
yi
y(xi )
y(xi ) yi
0.0500
0.1000
0.1500
0.2000
0.2500
0.3000
0.3500
0.4000
0.4500
0.5000
0.5500
0.6000
0.6500
0.7000
0.7500
0.8000
0.8500
0.9000
0.9500
1.0000
4.4913
4.0531
3.6757
3.3507
3.0708
2.8297
2.6221
2.4433
2.2893
2.1566
2.0424
1.9440
1.8593
1.7863
1.7235
1.6693
1.6227
1.5826
1.5480
1.5182
4.4893
4.0497
3.6713
3.3456
3.0653
2.8241
2.6164
2.4377
2.2839
2.1515
2.0375
1.9394
1.8550
1.7823
1.7198
1.6660
1.6196
1.5798
1.5454
1.5159
0.0020
0.0034
0.0044
0.0051
0.0055
0.0056
0.0057
0.0056
0.0054
0.0052
0.0049
0.0046
0.0043
0.0040
0.0037
0.0034
0.0031
0.0028
0.0026
0.0023
139

xi
yi
y(xi )
y(xi ) yi
0.1000
0.2000
0.3000
0.4000
0.5000
0.6000
0.7000
0.8000
0.9000
1.0000
4.0650
3.3684
2.8495
2.4629
2.1748
1.9602
1.8004
1.6813
1.5926
1.5265
4.0497
3.3456
2.8241
2.4377
2.1515
1.9394
1.7823
1.6660
1.5798
1.5159
0.0153
0.0228
0.0254
0.0251
0.0234
0.0208
0.0180
0.0153
0.0128
0.0106
9. Solve the following initial-value problems using the Heuns method and the Midpoint method.
(a) y 0 = (x + 1)y, x = 0.5(0.2)1.5, y(0.5) = 1.
(b) y 0 = xy 2 , x = 0(0.2)1, y(0) = 2.
(c) y 0 = x2 + y 2 , x = 1(0.2)2, y(1) = 1.
Solution: (a) Given f (x, y) = (x + 1)y, and x0 = 0.5, y0 = 1, h = 0.2 :
Heuns Method: First we find the values of k1 and k2 by taking i = 0, we have
k1 = f (x0 , y0 ) = (x0 + 1)y0 = 1.5000
2h
2h
k2 = f (x0 + , y0 + k1 ) = [(0.6333 + 1)(1.2)] = 1.9600
3
3
Now using these values in the Heuns formula, we get
y1 = y0 +
h
[k1 + 3k2 ] = 1 + 0.05(1.5000 + 3(1.9600)) = 1.3690
4

follows
xi
yi
0.7000
0.9000
1.1000
1.3000
1.5000
1.3690
1.9472
2.8765
4.4120
7.0236
Midpoint Method: First we find the values of k1 and k2 by taking i = 0, we

have
k1 = f (x0 , y0 ) = (x0 + 1)y0 = 1.5000
h
h
k2 = f (x0 + , y0 + k1 ) = (1.7)(1.15) = 1.8400
2
2
Now using these values in the Midpoint formula, we get
y1 = y0 + hk2 = 1 + (0.2)(1.8400) = 1.3680
A.BENHARI
140

follows
xi
yi
0.7000
0.9000
1.1000
1.3000
1.5000
1.3680
1.9442
2.8696
4.3974
6.9937
(b) Given f (x, y) = xy 2 , and x0 = 0, y0 = 2, h = 0.2 :

k1 = f (x0 , y0 ) = x0 y02 = 0
2h
2h
k2 = f (x0 + , y0 + k1 ) = [(0.1333)(4)] = 0.5333
3
3
y1 = y0 +
h
[k1 + 3k2 ] = 2 + 0.05(0 + 3(0.5333)) = 1.9200
4

follows
xi
yi
0.2000
0.4000
0.6000
0.8000
1.0000
1.9200
1.7172
1.4636
1.2156
0.9998

have
k1 = f (x0 , y0 ) = x0 y02 = 0
h
h
k2 = f (x0 + , y0 + k1 ) = (0.1)(4) = 0.4000
2
2
y1 = y0 + hk2 = 2 + (0.2)(0.4000) = 1.9200
follows
xi
yi
0.2000
0.4000
0.6000
0.8000
1.0000
A.BENHARI
1.9200
1.7155
1.4602
1.2117
0.9962
141
(c) Given f (x, y) = x2 + y 2 , and x0 = 1, y0 = 1, h = 0.2 :

k1 = f (x0 , y0 ) = x20 + y02 = 2
k2 = f (x0 +
2h
2h
, y0 + k1 ) = [(1.1333)2 + (0.7334)2 ] = 1.8222
3
3

y1 = y0 +
h
[k1 + 3k2 ] = 1 + 0.05(2 + 3(1.8222)) = 0.6267
4

follows
xi
yi
1.2000 0.6267
1.4000 0.2464
1.6000 0.2073
1.8000 0.8343
2.0000 1.8688
have
k1 = f (x0 , y0 ) = x20 + y02 = 2
h
h
k2 = f (x0 + , y0 + k1 ) = (1.1)2 + (0.8)2 = 1.8500
2
2
y1 = y0 + hk2 = 1 + (0.2)(1.8500) = 0.6300
follows
xi
yi
1.2000 0.6300
1.4000 0.2522
1.6000 0.1983
1.8000 0.8183
2.0000 1.8328
10. Solve the initial-value problems in the Problem 8 and compare the numerical solutions obtained with the Heuns method and the Midpoint method.
Solution: (a) Given f (x, y) = x + 3y/x, and x0 = 1, y0 = 0, h = 0.05 :
k1 = f (x0 , y0 ) = x0 + 3y0 /x0 = 1
k2 = f (x0 +
A.BENHARI
2h
2h
, y0 + k1 ) = [1.0333 + 3(0.0333)/1.0333] = 1.1301
3
3
142

h
[k1 + 3k2 ] = 0 + 0.0125(1 + 3(1.1301)) = 0.0549
4
follows
xi
yi
y(xi ) y(xi ) yi
y1 = y0 +
1.0500
1.1000
1.1500
1.2000
1.2500
1.3000
1.3500
1.4000
1.4500
1.5000
1.5500
1.6000
1.6500
1.7000
1.7500
1.8000
1.8500
1.9000
1.9500
2.0000
0.0549
0.1205
0.1975
0.2868
0.3889
0.5048
0.6352
0.7807
0.9422
1.1204
1.3160
1.5298
1.7625
2.0149
2.2878
2.5818
2.8977
3.2364
3.5984
3.9846
0.0551
0.1210
0.1984
0.2880
0.3906
0.5070
0.6379
0.7840
0.9461
1.1250
1.3214
1.5360
1.7696
2.0230
2.2969
2.5920
2.9091
3.2490
3.6124
4.0000
0.0002
0.0005
0.0009
0.0012
0.0017
0.0022
0.0027
0.0033
0.0039
0.0046
0.0054
0.0062
0.0071
0.0081
0.0091
0.0102
0.0114
0.0126
0.0140
0.0154
xi
yi
y(xi )
y(xi ) yi
1.1000
1.2000
1.3000
1.4000
1.5000
1.6000
1.7000
1.8000
1.9000
2.0000
0.1191
0.2835
0.4991
0.7720
1.1080
1.5132
1.9933
2.5544
3.2024
3.9432
0.1210
0.2880
0.5070
0.7840
1.1250
1.5360
2.0230
2.5920
3.2490
4.0000
0.0019
0.0045
0.0079
0.0120
0.0170
0.0228
0.0297
0.0376
0.0466
0.0568
Similarly, for h = 0.1, we have

have
k1 = f (x0 , y0 ) = (x0 + 3y0 /x0 ) = 1
h
h
k2 = f (x0 + , y0 + k1 ) = 1.025 + 3(0.025)/1.025 = 1.0982
2
2
A.BENHARI
143

y1 = y0 + hk2 = 0 + (0.05)(1.0982) = 0.0549
follows
xi
yi
y(xi ) y(xi ) yi
1.0500
1.1000
1.1500
1.2000
1.2500
1.3000
1.3500
1.4000
1.4500
1.5000
1.5500
1.6000
1.6500
1.7000
1.7500
1.8000
1.8500
1.9000
1.9500
2.0000
0.0549
0.1205
0.1976
0.2869
0.3892
0.5051
0.6355
0.7811
0.9427
1.1210
1.3167
1.5306
1.7634
2.0160
2.2890
2.5831
2.8992
3.2380
3.6002
3.9866
0.0551
0.1210
0.1984
0.2880
0.3906
0.5070
0.6379
0.7840
0.9461
1.1250
1.3214
1.5360
1.7696
2.0230
2.2969
2.5920
2.9091
3.2490
3.6124
4.0000
0.0002
0.0005
0.0008
0.0011
0.0015
0.0019
0.0024
0.0029
0.0034
0.0040
0.0047
0.0054
0.0062
0.0070
0.0079
0.0089
0.0099
0.0110
0.0121
0.0134
xi
yi
y(xi )
y(xi ) yi
1.1000
1.2000
1.3000
1.4000
1.5000
1.6000
1.7000
1.8000
1.9000
2.0000
0.1193
0.2840
0.5001
0.7735
1.1101
1.5160
1.9970
2.5591
3.2083
3.9505
0.1210
0.2880
0.5070
0.7840
1.1250
1.5360
2.0230
2.5920
3.2490
4.0000
0.0017
0.0040
0.0069
0.0105
0.0149
0.0200
0.0260
0.0329
0.0407
0.0495
(b) Given f (x, y) = y, and x0 = 0, y0 = 1, h = 0.05 :

k1 = f (x0 , y0 ) = y0 = 1
2h
2h
k2 = f (x0 + , y0 + k1 ) = 1.0333 = 1.0165
3
3
A.BENHARI
144

y1 = y0 +
h
[k1 + 3k2 ] = 1 + 0.0125(1 + 3(1.0165)) = 1.0506
4

follows
xi
yi
y(xi ) y(xi ) yi
0.0500
0.1000
0.1500
0.2000
0.2500
0.3000
0.3500
0.4000
0.4500
0.5000
0.5500
0.6000
0.6500
0.7000
0.7500
0.8000
0.8500
0.9000
0.9500
1.0000
1.0506
1.1025
1.1556
1.2100
1.2656
1.3225
1.3806
1.4400
1.5006
1.5624
1.6256
1.6899
1.7556
1.8224
1.8905
1.9599
2.0305
2.1024
2.1755
2.2499
1.0506
1.1025
1.1556
1.2100
1.2656
1.3225
1.3806
1.4400
1.5006
1.5625
1.6256
1.6900
1.7556
1.8225
1.8906
1.9600
2.0306
2.1025
2.1756
2.2500
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
xi
yi
y(xi )
y(xi ) yi
0.1000
0.2000
0.3000
0.4000
0.5000
0.6000
0.7000
0.8000
0.9000
1.0000
1.1025
1.2099
1.3224
1.4398
1.5623
1.6898
1.8222
1.9597
2.1021
2.2496
1.1025
1.2100
1.3225
1.4400
1.5625
1.6900
1.8225
1.9600
2.1025
2.2500
0.0000
0.0001
0.0001
0.0002
0.0002
0.0002
0.0003
0.0003
0.0004
0.0004

have
k1 = f (x0 , y0 ) = y0 = 1
h
h
k2 = f (x0 + , y0 + k1 ) = 1.025 = 1.0124
2
2
A.BENHARI
145

y1 = y0 + hk2 = 1 + (0.05)(1.0124) = 1.0506
follows
xi
yi
y(xi ) y(xi ) yi
0.0500
0.1000
0.1500
0.2000
0.2500
0.3000
0.3500
0.4000
0.4500
0.5000
0.5500
0.6000
0.6500
0.7000
0.7500
0.8000
0.8500
0.9000
0.9500
1.0000
1.0506
1.1025
1.1556
1.2100
1.2656
1.3225
1.3806
1.4400
1.5006
1.5625
1.6256
1.6900
1.7556
1.8224
1.8906
1.9599
2.0306
2.1024
2.1756
2.2499
1.0506
1.1025
1.1556
1.2100
1.2656
1.3225
1.3806
1.4400
1.5006
1.5625
1.6256
1.6900
1.7556
1.8225
1.8906
1.9600
2.0306
2.1025
2.1756
2.2500
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
xi
yi
y(xi )
y(xi ) yi
0.1000
0.2000
0.3000
0.4000
0.5000
0.6000
0.7000
0.8000
0.9000
1.0000
1.1025
1.2099
1.3224
1.4399
1.5623
1.6898
1.8223
1.9598
2.1022
2.2497
1.1025
1.2100
1.3225
1.4400
1.5625
1.6900
1.8225
1.9600
2.1025
2.2500
0.0000
0.0001
0.0001
0.0001
0.0002
0.0002
0.0002
0.0002
0.0003
0.0003
(c) Given f (x, y) = 4 3y, and x0 = 0, y0 = 5, h = 0.05 :

Heuns Method: The approximations by taking xi = xi1 + h, i = 1, 2, . . . , 19,
A.BENHARI
146
are as follows
(c) h = 0.05
xi
yi
y(xi )
y(xi ) yi
0.0500
0.1000
0.1500
0.2000
0.2500
0.3000
0.3500
0.4000
0.4500
0.5000
0.5500
0.6000
0.6500
0.7000
0.7500
0.8000
0.8500
0.9000
0.9500
1.0000
4.4913
4.0531
3.6757
3.3507
3.0708
2.8297
2.6221
2.4433
2.2893
2.1566
2.0424
1.9440
1.8593
1.7863
1.7235
1.6693
1.6227
1.5826
1.5480
1.5182
4.4893
4.0497
3.6713
3.3456
3.0653
2.8241
2.6164
2.4377
2.2839
2.1515
2.0375
1.9394
1.8550
1.7823
1.7198
1.6660
1.6196
1.5798
1.5454
1.5159
0.0020
0.0034
0.0044
0.0051
0.0055
0.0056
0.0057
0.0056
0.0054
0.0052
0.0049
0.0046
0.0043
0.0040
0.0037
0.0034
0.0031
0.0028
0.0026
0.0023
Similarly, for h = 0.1, the approximations by taking xi = xi1 + h, i = 1, 2, . . . , 9,

are as follows
h = 0.1
xi
yi
y(xi ) y(xi ) yi
0.1000
0.2000
0.3000
0.4000
0.5000
0.6000
0.7000
0.8000
0.9000
1.0000
4.0650
3.3684
2.8495
2.4629
2.1748
1.9602
1.8004
1.6813
1.5926
1.5265
4.0497
3.3456
2.8241
2.4377
2.1515
1.9394
1.7823
1.6660
1.5798
1.5159
0.0153
0.0228
0.0254
0.0251
0.0234
0.0208
0.0180
0.0153
0.0128
0.0106
Midpoint Method: The approximations by taking xi = xi1 +h, i = 1, 2, . . . , 19,
A.BENHARI
147
are as follows
xi
yi
y(xi )
y(xi ) yi
0.0500
0.1000
0.1500
0.2000
0.2500
0.3000
0.3500
0.4000
0.4500
0.5000
0.5500
0.6000
0.6500
0.7000
0.7500
0.8000
0.8500
0.9000
0.9500
1.0000
4.4913
4.0531
3.6757
3.3507
3.0708
2.8297
2.6221
2.4433
2.2893
2.1566
2.0424
1.9440
1.8593
1.7863
1.7235
1.6693
1.6227
1.5826
1.5480
1.5182
4.4893
4.0497
3.6713
3.3456
3.0653
2.8241
2.6164
2.4377
2.2839
2.1515
2.0375
1.9394
1.8550
1.7823
1.7198
1.6660
1.6196
1.5798
1.5454
1.5159
0.0020
0.0034
0.0044
0.0051
0.0055
0.0056
0.0057
0.0056
0.0054
0.0052
0.0049
0.0046
0.0043
0.0040
0.0037
0.0034
0.0031
0.0028
0.0026
0.0023

are as follows
xi
yi
y(xi ) y(xi ) yi
0.1000
0.2000
0.3000
0.4000
0.5000
0.6000
0.7000
0.8000
0.9000
1.0000
4.0650
3.3684
2.8495
2.4629
2.1748
1.9602
1.8004
1.6813
1.5926
1.5265
4.0497
3.3456
2.8241
2.4377
2.1515
1.9394
1.7823
1.6660
1.5798
1.5159
0.0153
0.0228
0.0254
0.0251
0.0234
0.0208
0.0180
0.0153
0.0128
0.0106
11. Solve the following initial-value problems using the fourth-order Runge-Kutta formula using h = 0.2
(a)
y
y0 = 1 + ,
x
(b)
y 0 = y tan x,
(c)
y 0 = (1 x)y 2 y,
A.BENHARI
1 x 2 y((1) = 1.
0 x 1,
y(0) = 2.
1 x 2 y(0) = 1.
148
Solution: (a) We set f (x, y) = 1 + y/x and x0 = 1, y0 = 1, and h = 0.2, then for
i = 0, the calculated values of k1 , k2 , k3 , and k4 are as follows
k1 = f (x0 , y0 ) = (1 + y0 /x0 ) = 2.0000
h
h
k2 = f (x0 + , y0 + k1 ) = [1 + 1.2/1.1] = 2.0909
2
2
h
h
k3 = f (x0 + , y0 + k2 ) = [1 + 1.2091/1.1] = 2.0992
2
2
k4 = f (x0 + h, y0 + hk3 ) = [1 + 1.4198/1.2] = 2.1832
Hence the first approximate solution is
y1 = y0 +
i
hh
k1 + 2k2 + 2k3 + k4 = [1 + (0.2/6)(12.5634)] = 1.4188
6

follows
xi
yi
1.2000
1.4000
1.6000
1.8000
2.0000
1.4188
1.8710
2.3520
2.8580
3.3863
(b) We set f (x, y) = y tan x and x0 = 0, y0 = 2, and h = 0.2, then for i = 0, the
calculated values of k1 , k2 , k3 , and k4 are as follows
k1 = 0,
k2 = 0.2007,
k3 = 0.2027,
k4 = 0.4136

y1 = y0 +
i
hh
k1 + 2k2 + 2k3 + k4 = [2 + (0.2/6)(1.2204)] = 2.0407
6

follows
xi
yi
0.2000
0.4000
0.6000
0.8000
1.0000
2.0407
2.1714
2.4233
2.8707
3.7017
(c) We set f (x, y) = (1 x)y 2 y and x0 = 1, y0 = 1, and h = 0.2, then for i = 0,

the calculated values of k1 , k2 , k3 , and k4 are as follows
k1 = 1,
A.BENHARI
k2 = 0.9810,
k3 = 0.9832,
149
k4 = 0.9324

y1 = y0 +
i
hh
k1 + 2k2 + 2k3 + k4 = [1 + (0.2/6)(5.8608)] = 0.8046
6

follows
xi
yi
1.2000
1.4000
1.6000
1.8000
2.0000
0.8046
0.6315
0.4892
0.3772
0.2910
12. Solve the following initial-value problems and compare the numerical solutions obtained with the fourth-order Runge-Kutta formula and the fourth-order Taylors
method by using the values of h = 0.1 and h = 0.2, over the interval [a, b].
(a)
y 0 = 4 3y,
(b)
y 0 = (2 x)y,
(c)
y0 =
[0, 1],
y(0) = 5,
[2, 3],
1
2y 2 ,
1 + x2
y(2) = 1,
y(x) = 4/3 + 11/3e3x .

2
y(x) = e1/2(x2) .
[0, 1], y(0) = 0,
y(x) =
x
.
1 + x2
Solution: (a) We set f (x, y) = 4 3y and x0 = 1, y0 = 5, and h = 0.1 :

Fourth-order Runge-Kutta Formula: For i = 0, the calculated values of
k1 , k2 , k3 , and k4 are as follows
k1 = f (x0 , y0 ) = (4 3y0 ) = 11.0000
h
h
k2 = f (x0 + , y0 + k1 ) = [4 3(4.45)] = 9.3500
2
2
h
h
k3 = f (x0 + , y0 + k2 ) = [4 3(4.5325)] = 9.5975
2
2
k4 = f (x0 + h, y0 + hk3 ) = [4 3(4.04025)] = 8.1208
y1 = y0 +
i
hh
k1 + 2k2 + 2k3 + k4 = [5 + (0.2/6)(57.0158)] = 4.0497
6
A.BENHARI
150
follows
xi
yi
y(xi )
y(xi ) yi
0.1000
0.2000
0.3000
0.4000
0.5000
0.6000
0.7000
0.8000
0.9000
1.0000
4.0497
3.3457
2.8242
2.4378
2.1516
1.9395
1.7824
1.6660
1.5798
1.5159
4.0497
3.3456
2.8241
2.4377
2.1515
1.9394
1.7823
1.6660
1.5798
1.5159
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0000

are as follows
xi
yi
y(xi ) y(xi ) yi
0.2000
0.4000
0.6000
0.8000
1.0000
3.3478
2.4401
1.9414
1.6674
1.5169
3.3456
2.4377
1.9394
1.6660
1.5159
0.0022
0.0024
0.0020
0.0014
0.0010
Fourth-order Taylors Method: Since f (x, y) = 4 3y, f 0 (x, y) = 12 +

9y, f 00 (x, y) = 36 27y, f 000 (x, y) = 108 + 81y, and x0 = 0, y0 = 5, h = 0.1, then
using fourth-order Taylors method
yi+1 = yi +hf (xi , yi )+
h3
h4
h2 0
f (xi , yi )+ f 00 (xi , yi )+ f 000 (xi , yi ),
2
6
24
for i = 0, 1, 2, 3, 4

y1 = y0 + hf (x0 , y0 ) +
= y0 + h[4 3y0 ] +
h2 0
h3
h4
f (x0 , y0 ) + f 00 (x0 , y0 ) + f 000 (x0 , y0 )
2
6
24
h3
h4
h2
[12 + 9y0 ] + [36 27y0 ] + [108 + 81y0 ]
2
6
24
which gives
y1 = 5 + (0.1)[4 15] + 0.005[12 + 45] + 0.0002[36 135]
+ 0.000004[108 + 405] = 4.0497
A.BENHARI
151
follows
xi
yi
y(xi )
y(xi ) yi
0.1000
0.2000
0.3000
0.4000
0.5000
0.6000
0.7000
0.8000
0.9000
1.0000
4.0497
3.3457
2.8242
2.4378
2.1516
1.9395
1.7824
1.6660
1.5798
1.5159
4.0497
3.3456
2.8241
2.4377
2.1515
1.9394
1.7823
1.6660
1.5798
1.5159
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0001
0.0000

are as follows
xi
yi
y(xi ) y(xi ) yi
0.2000
0.4000
0.6000
0.8000
1.0000
3.3478
2.4401
1.9414
1.6674
1.5169
3.3456
2.4377
1.9394
1.6660
1.5159
0.0022
0.0024
0.0020
0.0014
0.0010
(b) We set f (x, y) = (2 x)y and x0 = 2, y0 = 1, and h = 0.1 :

Fourth-order Runge-Kutta Formula: For i = 0, the calculated values of
k1 , k2 , k3 , and k4 are as follows
k1 = 0,
k2 = 0.0500,
k3 = 0.0499,
k4 = 0.0995

y1 = y0 +
i
hh
k1 + 2k2 + 2k3 + k4 = [1 + (0.2/6)(0.2993)] = 0.9950
6

follows
xi
yi
y(xi ) y(xi ) yi
2.1000
2.2000
2.3000
2.4000
2.5000
2.6000
2.7000
2.8000
2.9000
3.0000
A.BENHARI
0.9950
0.9802
0.9560
0.9231
0.8825
0.8353
0.7827
0.7261
0.6670
0.6065
0.9950
0.9802
0.9560
0.9231
0.8825
0.8353
0.7827
0.7261
0.6670
0.6065
152
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000

are as follows
xi
yi
y(xi ) y(xi ) yi
2.2000
2.4000
2.6000
2.8000
3.0000
0.9802
0.9231
0.8353
0.7261
0.6065
0.9802
0.9231
0.8353
0.7261
0.6065
0.0000
0.0000
0.0000
0.0000
0.0000
Fourth-order Taylors Method: Since f (x, y) = (2 x)y, f 0 (x, y) = y +

(2 x)2 y, f 00 (x, y) = 3(2 x)y + (2 x)3 y, f 000 (x, y) = 3y 6(2 x)2 y + (2
x)4 y, and x0 = 2, y0 = 1, h = 0.1, then using fourth-order Taylors method, the
approximations by taking xi = xi1 + h, i = 1, 2, . . . , 9, are as follows
xi
yi
y(xi )
y(xi ) yi
2.1000
2.2000
2.3000
2.4000
2.5000
2.6000
2.7000
2.8000
2.9000
3.0000
0.9950
0.9802
0.9560
0.9231
0.8825
0.8353
0.7827
0.7262
0.6670
0.6065
0.9950
0.9802
0.9560
0.9231
0.8825
0.8353
0.7827
0.7261
0.6670
0.6065
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000

are as follows
xi
yi
y(xi ) y(xi ) yi
2.2000
2.4000
2.6000
2.8000
3.0000
0.9802
0.9231
0.8353
0.7262
0.6066
0.9802
0.9231
0.8353
0.7261
0.6065
0.0000
0.0000
0.0000
0.0000
0.0000
1
2y 2 and x0 = 0, y0 = 0, and h = 0.1 :
1 + x2
Fourth-order Runge-Kutta Formula: The approximations by taking xi =
(c) We set f (x, y) =
A.BENHARI
153
xi1 + h, i = 1, 2, . . . , 9, are as follows

xi
yi
y(xi )
y(xi ) yi
0.1000
0.2000
0.3000
0.4000
0.5000
0.6000
0.7000
0.8000
0.9000
1.0000
0.0990
0.1923
0.2752
0.3448
0.4000
0.4412
0.4698
0.4878
0.4972
0.5000
0.0990
0.1923
0.2752
0.3448
0.4000
0.4412
0.4698
0.4878
0.4972
0.5000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000
0.0000

are as follows
xi
yi
y(xi ) y(xi ) yi
0.2000
0.4000
0.6000
0.8000
1.0000
A.BENHARI
0.1923
0.3448
0.4411
0.4877
0.4999
0.1923
0.3448
0.4412
0.4878
0.5000
154
0.0000
0.0000
0.0001
0.0001
0.0001
Bibliography
General Texts in Numerical Analysis
K. Atkinson: An Introduction to Numerical Analysis, Wiley, (2nd ed.), 1989.

P.G. Ciarlet and J. L. Lions (eds), Handbook of Numerical Analysis,
North Holland, 1990.
E. W. Cheney and D. R. Kincaid: Numerical Mathematics and Computing,
Brooks Cole, 6 edition, 2007.
S. Conte and C. de Boor: Elementary Numerical Analysis, McGraw-Hill, 1980.
G. Dahlquist and A. Bjorck: Numerical Methods, Prentice Hall, 1974.
P. Deuflhard and A. Hohmann, Numerical Analysis in Modern Scientific
Computing, 2nd ed., Springer, 2003.
W. Gautschi: Numerical Analysis: an introduction, Birkhauser, 1997.
M. T. Heath, Scientific Computing: An Introductory Survey, 2nd ed.,
McGraw-Hill, 2002.
E. Isaacson and H. Keller: Analysis of Numerical Methods, Wiley, 1966
(or Dover 1994).
D. Kahaner, C. Moler, and S. Nash: Numerical Methods and Software,
Prentice-Hall, 1989.
D. Kincaid and W. Cheney: Numerical Analysis: Mathematics of Scientific
Computing, American Mathematical Society, 3rd edition, 2001.
A. Quarteroni, R. Sacco, and F. Saleri: Numerical Mathematics, 2nd Edition,
Springer, 2004.
A. Ralston and P. Rabinowitz: A First Course in Numerical Analysis,
McGraw-Hill, 1978.
L. Shampine, Allen, and Pruess: Fundamentals of Numerical Computing,
G. W. Stewart: Afternotes on Numerical Analysis, SIAM, 1996.
G. W. Stewart: Afternotes Goes to Graduate School:
Lectures on Advanced Numerical Analysis, SIAM, 1998
J. Stoer and R. Bulirsch: Introduction to Numerical Analysis, Springer,
3rd edition, 2002.
E. Suli and D. Mayers: An Introduction to Numerical Analysis, Cambridge, 2003.
A.BENHARI
154
Numerical Solution of Ordinary Differential Equations
U. M. Ascher and L. R. Petzold, Computer Methods for Ordinary Differential

Equations and Differential-Algebraic Equations, SIAM, 1998
J. C. Butcher, Numerical Methods for Ordinary Differential Equations, 2nd
ed., Wiley, 2003.
J. C. Butcher: The Numerical Analysis of Ordinary Differential
Equations: Runge-Kutta and general linear methods, Wiley, 1987.
K. Dekker and J. G. Verwer, Stability of Runge-Kutta methods for stiff
nonlinear differential equations, North Holland, 1984.
P. Deuflhard and F. Bornemann, Scientific Computing with Ordinary
Differential Equations, Springer, 2002
S. O. Fatunla: Numerical Methods for Initial Value Problems in Ordinary
Differential Equations, Academic Press, 1988.
C. W. Gear: Numerical Initial Problems in Ordinary Differential Equations,
Prentice Hall, 1971.
E. Hairer, S. P. Norsett, and G. Wanner, Solving Ordinary Differential
Equations, Springer-Verlag, I: nonstiff problems (1993), II (1991).
P. Henrici: Discrete Variable Methods in Ordinary Differential Equations,
Wiley, 1962.
I. Iserles, A First Course in the Numerical Analysis of Differential
Equations, Cambridge University Press, 1996.
J. D. Lambert, Numerical Methods for Ordinary Differential Systems:
The Initial Value Problem, Wiley, 1991.
J. Lambert, Computational Methods in Ordinary Differential Equations, 1973.
L. Shampine: Numerical solution of ordinary differential equations,
Chapman & Hall, 1994.
L. Shampine and M. Gordon: Computer Solution of Ordinary Differential
Equations, Freeman, 1975.
L. Shampine, I. Gladwell, and S. Thompson, Solving ODEs with MATLAB,
Cambridge, 2003.
A.BENHARI
155
Solution of Nonlinear Systems of Equations and Optimization
J. E. Dennis and J. More: Quasi-Newton methods, motivation, and theory,

SIAM Review, vol 19 #1, Jan. 1977.
J. E. Dennis and R. B. Schnabel: Numerical Methods for Unconstrained
Optimization and Nonlinear Equations, Prentice-Hall, 1983, SIAM, 1996.
P. Deuflhard: Newton Methods for Nonlinear Problems: Affine Invariance and
Adaptive Algorithms, Springer, 2004.
C. T. Kelley: Iterative Methods for Linear and Nonlinear Equations, SIAM, 1995.
C. T. Kelley, Iterative Methods for Optimization, SIAM, 1999.
C. T. Kelley: Solving Nonlinear Equations with Newton's Method, SIAM, 2003.
J. Nocedal and S.J. Wright: Numerical Optimization, Springer, 2000.
J. Ortega and W. Rheinboldt: Iterative Solution of Nonlinear Equations
in Several Variables, Academic Press, 1970.
Numerical Linear Algebra
O. Axelsson: Iterative Solution Methods, Cambridge University Press, 1994.

J. W. Demmel: Applied Numerical Linear Algebra, SIAM, 1997.
G. Forsythe and C. Moler: Computer Solution of Linear Algebraic Systems,
Prentice Hall, 1967.
G. Golub and C. Van Loan: Matrix Computations, (3rd ed.), Johns Hopkins
University Press, 1996
A. Gourley and G. Watson: Computational Methods for Matrix Eigenproblems,
Wiley, 1973.
W. W. Hager: Applied Numerical Linear Algebra, Prentice-Hall, 1988.
G. W. Stewart: Introduction to Matrix Computations, Academic Press, 1973.
G. W. Stewart: Matrix Algorithms: Basic Decompositions, SIAM, 1998
G. W. Stewart: Matrix Algorithms, Volume II: Eigensystems, SIAM, 2001
G. W. Stewart, J-q. Sun: Matrix Perturbation Theory, Academic Press, 1990
L. N. Trefethen and D. Bau, Numerical Linear Algebra, SIAM, 1997.
J. H. Wilkinson: The Algebraic Eigenvalue Problem, Oxford, 1988.
J. H. Wilkinson and C. Reinsch: Linear Algebra
A.BENHARI
156
Finite Element, Finite Volume, and Spectral Methods

D. Braess: Finite Elements: Theory, fast solvers, and applications in
solid mechanics, Cambridge University Press, 3 edition, 2007.
S. Brenner and L. R. Scott: The Mathematical Theory of Finite Element
Methods, Springer-Verlag, 3rd edition, 2007.
P. G. Ciarlet: The Finite Element Method for Elliptic Problems,
SIAM, 2002 (Originally published by North Holland, 1980).
K. Eriksson, D. Estep, P. Hansbo, C. Johnson: Computational Differential
A. Ern and J-L. Guermond: Theory and Practice of Finite Elements,
Springer, 2004.
C. Grossmann, H-G. Roos, and M. Stynes: Numerical Treatment of Partial
Differential Equations, Springer, 2007.
T. J. R. Hughes, The Finite Element Method: Linear Static and Dynamic
Finite Element Analysis, Prentice Hall, 1987.
C. Johnson: Numerical Solutions of Partial Differential Equations by
the Finite Element Method, Cambridge University Press, 1987.
R. J. LeVeque, Finite Volume Methods for Hyperbolic Problems, Cambridge, 2002.
G. Strang and G. Fix: An Analysis of the Finite Element Method,
Prentice-Hall, 1973.
B. Szabo and I. Babuska: Finite Element Analysis, Wiley, 1991.
L. N. Trefethen: Spectral Methods in MATLAB, SIAM 2000.
O. C. Zienkiewicz and R. L. Taylor, The Finite Element Method
(4th edition), McGraw Hill, 1989.
Finite Difference Methods for Two-Point Boundary Value Problems and Partial Differential Equations
S. K. Godunov and V. S. Ryabenki, Difference Schemes: an introduction
to the underlying theory, North Holland, 1987.
C. Grossmann, H-G. Roos, and M. Stynes: Numerical Treatment of Partial
Differential Equations, Springer, 2007.
C. Hall and T. Porsching: Numerical Analysis of Partial Differential
Equations, Prentice Hall, 1990.
R. J. LeVeque, Numerical methods for conservation laws, Birkhauser Verlag,
1992.
I. Iserles, A First Course in the Numerical Analysis of Differential
H. Keller: Numerical Methods for Two-Point Boundary Value Problems,
SIAM, 1976.
R. D. Richtmyer and K. W. Morton: Difference Methods for Initial Value
Problems, Wiley-Interscience, 1967.
G. Sod: Numerical Methods in Fluid Mechanics: Initial and
Initial Boundary Value Problems, Cambridge University Press, 1985.
J. Strikwerda: Finite Difference Schemes and Partial Differential
Equations, Second Edition, SIAM, 2004.
MATLAB
T. Davis and K. Sigmon: MATLAB Primer, Seventh Edition, Chapman and Hall, 2004.
D. J. Higham and N. J. Higham: MATLAB Guide, Second Edition, SIAM, 2005.
C. Moler: Numerical Computing with MATLAB, SIAM, 2004.
A. Quarteroni and F. Saleri, Scientific Computing with Matlab, Springer, 2003.

Numerical Analysis Exercises

Uploaded by

Copyright:

Available Formats

Numerical Analysis Exercises

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Numerical Analysis Exercises

Uploaded by

Copyright:

Available Formats

Abdelkader BENHARI

Numerical anlysis exercises

Number Systems and Error

Solution of Nonlinear Equations

Systems of Linear Equations

Differentiation and Integration

Ordinary Differential Equations

Number Systems and Error

1. Convert the following binary numbers to decimal form

Thus the binary representation for 165 is

Thus the binary representation for 111 is

Thus the binary representation for 278.5 is

(11.1)16 = 1(16)1 + 1(16)0 + 1(16)1

Thus, the result lies in the range S Abs. Error

The answer may be rounded meaningfully to 18.9 which is correct to 3 significant

Then the relative error is

and the absolute error is

which is correct to 3 sd (3 dp).

Since 8.3221 104 <

2 = 1.4142 and p = 1.414, then the absolute error is define as

Solution: The floating decimal point representation of 13.2 is

Similarly, writing 12.532 in decimal floating point form

and the number 2/125 in decimal floating point form

Solution: (a) Let x1 = 187.2, x2 = 93.5, and S = x1 + x2 = 280.7. Also, let e1

which is correct to 3 significant digits.

Then the relative error is

and the absolute error is

or 1.04 which is correct to 3 sd (2 dp).

or 5.321 which is correct to 4 sd (3 dp).

12 104 12 105 12 103

The normalized binary floating point representation of 1/5 can be

Solution of Nonlinear Equations

Thus a bound for the number of iterations is

So using the bisection method, we have x14 = 0.5943

5. Use the bisection method for f (x) = x3 3x + 1 in [1, 3] to find:

6. Solve the Problem 1 by the false position method.

(1., 8301)(0.4907) (2.4)(0.0095)

(c) x = (3x + 20)1/3 .

(c) The error bound for the approximation is

where k = max |g 0 (x)| = g 0 (1) = 0.41218. Thus

Solution: Since g(x) =

15. Which of the following iterations

we have x1 = 10/3, we have and

Thus a bound for the number of iterations is

Solution: Since g 0 (x) =

Thus, the Newtons method gives x2 = 1.8414, accurate to within 104 .

(xk 1)exk + 3x2k

(b) For x0 = 4, x4 = 0.1215

Solution: Let N be a positive number and x = 1/ N . If f (x) = 0, then

gives f 0 (x) = 2/x3

x1 = x0 (3 N x20 )/2 = 0.05(3 (15)(0.05)2 )/2 = 0.0741

Thus g 0 () = 0, gives at least quadratic convergence.

23. Given the iterative scheme xn+1 = xn

Solution: Since the Newtons iteration function is

Since f 0 () = 0, so by using the Hopital rule, it gives us

f 0 (x)f 00 (x) + f (x)f 000 (x)

[f 00 (x)]2 + f (x)f 000 (x) + f 0 (x)f 000 (x) + f (x)f (4)(x)

24. The Halleys iterative method

then f 0 (x) = 2x,

and third approximation is