CHAPTER 1
Introduction
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Introduction
• Numerical methods are techniques by which mathematical
problems are formulated so that they can be solved with
arithmetic operations.
• Engineering includes heavily mathematical models. Main
responsible of engineers are to solve these mathematical
models
• Solutions are obtained by standard mathematical
techniques if the models are simple enough.
• However most of the engineering problems are not simple
enough to solve using standard mathematical techniques.
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Introduction
• For Example, consider the following integral
b
x2
I1 xe
a
dx
• This integral may be integrated using standard
mathematical technique
1 x2 b 1 a2 b2
I1 e (e e )
2 a
2
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Introduction
• On the other hand
b
x2
I2 e
a
dx
• The integration above may not be solved using standard
mathematical techniques. This integral can be evaluated
only numerical analysis.
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Basic Definitions
• Significant figures
• Whenever we employ a number in a computation, we
must have assurance that it can be used with confidence
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Significant figures
• What is the speed of the car?
• It is 48 km/h or 49km/h?
• Or it is 48.8 km/h or 48.9 km/h?
• May one claim its speed is 48.86523 km/h?
• The significant figures of a number are those that can be
used with confidence.
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
How to find number of significant figures
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
How to find number of significant figures
• Usually it is a straightforward procedure, but some cases can lead
to confusion.
• For example, zeros are not always significant figures because
they may be necessary just to locate a decimal point 0.00001845,
0.0001845, and 0.001845 all have four significant figures.
• Similarly, when trailing zeros are used in large numbers, it is not
clear how many, if any, of the zeros are significant. For example, at
face value the number 45,300 may have three, four, or five
significant digits, depending on whether the zeros are known with
confidence.
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
How to find number of significant figures
• Π is represented with 7 significant figures as below
πchop =0.3141592 x 101
πround =0.3141593 x 101
• To count the significant digits (figures), we write the numbers in a
scientific notation (normalized decimal form)
0.d1d 2 d 3d 4 ....d k x10 n d1 0
• Here k represents the number of significant figures.
• For example 0.1032, 0.001324 and 22.53 have four significant
figures.
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Chopping and Rounding
• Many numbers like π require infinitely many number of
digits for exact representation. The representation of a real
number in a computer by the use of finite number of digits is
called its floating point. Floating point is determined
chopping and rounding.
• Chopping
• Π= 3.1415926535897932384626433832795….
• If the π is represented 5 significant digits
• Π=3.1415
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Target digit
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Effect of number of digits used in computations
• Example
1
Calculate the derivative of f ( x ) 2
with 3& 4 digit aritmetics at x 1.22
(3 2 x )
4x
f ( x)
(3 2 x 2 ) 2
3 Digits Arithmetic's
4*1.22 4.88 4.88 4.88
f (1.22) 2 2
2
2
12, 200
(3 2*(1.22) ) (3 2*1.49) (3 2.98) 0.0004
4 Digits Arithmetic's
4*1.22 4.88 4.88 4.88
f (1.22) 2 2
2
2
8472
(3 2*(1.22) ) (3 2*1.488) (3 2.976) 0.000576
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Accuracy and Precision
• ACCURACY
• Accuracy refers to how closely a computed or measured
value agrees with the true value.
• PRECISION
• Precision refers to how closely individual computed or
measured values agree with each other.
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Accuracy and Precision
(a) Inaccurate and imprecise
(b) accurate and imprecise
(c) inaccurate and precise
(d) accurate and precise
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Error
• TRUE ERROR
• The relationship between the exact (true) value and
the approximation can be formulated as:
• True value= Approximate value + Error
• From the above equation, true error is denoted as:
• Et=True value-Approximate value
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Error
• To account for the magnitudes of the quantities being
evaluated is to normalize the error to the true value, as in
True error
Re lative Error
True value
• The relative error can also be multiplied by 100 percent, as
in
True error
t 100%
True value
• where εt designates the true percent relative error.
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Example
• Suppose that you have the task of measuring the lengths of a bridge
and a rivet and come up with 9999 and 9 cm, respectively. If the true
values are 10,000 and 10 cm, respectively, compute (a) the true error
and (b) the true percent relative error for each case.
• Solution:
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Approximate Error
• In real-world applications, we will obviously not know
the true answer a priori. Therefore, an alternative error
(approximate error) should be defined
• Certain numerical methods use an iterative approach to
compute answers. In such an approach, a present
approximation is made on the basis of a previous
approximation. The approximate error is then defined:
Current Approximation Pr evious Approximation
a 100%
Current Approximation
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Approximate Error
• The signs of approximate error may be either positive or
negative, therefore, it is often useful to employ the absolute
value
• For such cases, the computations repeated until
a s
• It is also convenient to relate these errors to the number of
significant figures in the approximation, we can be assured
that the result is correct to at least n significant figures.
s 0.5 x102 n %
• n= number of significant figure
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Classification of Errors
• 1. TRUNCATION ERROR
• This error results from the discrepancy introduced by the
fact that the numerical method employs approximation to
represent exact mathematical operations and quantities
• 2. ROUND OFF ERRORS
• This kind of errors occurs due to the fact that computers
can only represent quantities with a finite number of digits
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Taylor Series
• The Taylor Series provide a means to predict a function value
at one point in terms of the function value and its derivatives
at another point.
• If a function f(x) and its first n+1 derivatives are continuous
on an interval containing ‘’a’’ and ‘’x’’ then the value of the
function at x is given by:
f n (a)
f ( x) ( x a) n
n 0 n!
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Taylor Series
( x a) ( x a)2 ( x a )3 ( x a)n
f ( x) f (a ) f (a) f (a) f (a) n
...... f (a ) Rn
1! 2! 3! n!
where, the remainder Rn is defined as:
n 1 ( x a ) n 1
Rn f (a ).
( n 1)!
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Example
• Expand f(x)=ex to the Taylor Series at a=0. Then calculate
true values and approximate values at x=0.5 (for n=2, 3
and 4) (n= number of terms in expansion)
• Solution
f ( x) e x f (0) e0 1
f ( x) e x f (0) e0 1
f ( x) e x f (0) e0 1
f ( x) e x f (0) e0 1
2 3
( x 0) ( x 0) ( x 0)
f ( x) e x f (0) f (0) f (0) f (0) .....
1! 2! 3!
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
2 3 4 n
x x x x x
f ( x) e x 1 ...
1! 2! 3! 4! n!
• Then calculate true and approximate values of function at
x=0.5
• True value =
f (0.5) e0.5 1.6487
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Example
• Approximate values:
• n=2
x
f ap ( x) 1 f ap (0.5) 1 0.5 1.5
1!
• Relative True error
True value Approximate value 1.6487 1.5
t 0.0902
True value 1.6487
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
• n=3
x2 0.52
f ap ( x) 1 x f ap (0.5) 1 0.5 1.625
2! 2
• Relative True error
1.6487 1.625
t 0.0144
1.6487
• n=4
x 2 x3 0.52 0.53
f ap ( x) 1 x f ap (0.5) 1 0.5 1.6458
2! 3! 2 6
• Relative True error
1.6487 1.6458
t 0.0018
1.6487 CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN
Matlab Code
n=input('enter n');
Outputs
x=input('enter x');
enter n=4
eax=0; enter x=0.5
ex=exp(x); n eax error
disp(['n eax error']) 1.0000 1.6487 0.3935
for i=1:n; 2.0000 1.6487 0.0902
k=i-1;
eax=eax+(x^k)/factorial(k); 3.0000 1.6487 0.0144
error=abs((ex-eax)/ex); 4.0000 1.6487 0.0018
disp([i eax error])
end
CENG 202: NUMERICAL ANALYSIS, DR. SEMIH ERHAN