Because learning changes everything.

Chapter 3

© 2021 McGraw Hill. All rights reserved. Authorized only for instructor use in the classroom. No reproduction or further distribution permitted without the prior written consent of McGraw Hill.
 Approximations and Round-Off Errors
For many engineering problems, we cannot obtain analytical solutions.
Numerical methods yield approximate results, results that are close to the
exact analytical solution. We cannot exactly compute the errors associated
with numerical methods.
• Only rarely given data are exact, since they originate from measurements.
Therefore there is probably error in the input information.
• Algorithm itself usually introduces errors as well, for example, unavoidable round-
offs, etc …
• The output information will then contain error from both of these sources.

How confident we are in our approximate result?

The question is “how much error is present in our calculation and is it
tolerable?”

© McGraw Hill 2
 Accuracy & Precision
• Accuracy. How close is a computed or measured value
to the true value.
• Precision (or reproducibility). How close is a computed
or measured value to previously computed or measured
values.
• Inaccuracy (or bias). A systematic deviation from the
actual value.
• Imprecision (or uncertainty). Magnitude of scatter.

© McGraw Hill 3
 Targets Illustrating Accuracy & Precision

Figure 3.2
Access the text alternative for slide images

© McGraw Hill 4
 Significant Figures
• Number of significant figures indicates precision. Significant digits of a
number are those that can be used with confidence, for example, the number
of certain digits plus one estimated digit.

53,800 How many significant figures?

5.38 × 104 3
5.380 × 104 In the following table, read ‘10 4’ as 10 to the power of 4.
4
5.3800 × 104 5

Zeros are sometimes used to locate the decimal point not significant figures.

0.00001753 4
0.0001753 4
0.001753 4

© McGraw Hill 5
 Error Definitions 1

True Value = Approximation + Error

Et = True value − Approximation (+ / −)

True error

true error
True fractional relative error =
true value

true error
True percent relative error,  t = 100%
true value

© McGraw Hill 6
 Error Definitions 2

• For numerical methods, the true value will be known only

when we deal with functions that can be solved analytically
(simple systems). In real world applications, we usually not
know the answer a priori. Then
Approximate error
a = 100%
Approximation

• Iterative approach, example Newton’s method.

Current approximation −Previous approximation
a = 100%
Current approximation
(+/−)
© McGraw Hill 7
 Error Definitions 3

• Use absolute value.

• Computations are repeated until stopping criterion is satisfied.

a  s
Pre-specified % tolerance based
on the knowledge of your solution

• If the following criterion is met

 s = ( 0.5  10(2-n ) ) %

you can be sure that the result is correct to at least n significant

figures.
© McGraw Hill 8
 Round-off Errors
• Numbers such as  , e, or 7 cannot be expressed
by a fixed number of significant figures.
• Computers use a base-2 representation, they cannot precisely represent
certain exact base-10 numbers.
• Fractional quantities are typically represented in computer using
“floating point” form, for example,

Access the text alternative for slide images

© McGraw Hill 9
 (a) Decimal & (b) Binary Number Systems

Figure 3.5
Access the text alternative for slide images

© McGraw Hill 10
 The signed magnitude method for integer
representation of −173

Figure 3.6

© McGraw Hill 11
 The manner in which a floating-point
(decimal) number is stored in a word.

Figure 3.7
Access the text alternative for slide images

© McGraw Hill 12
 An example how a decimal number would be
stored in a hypothetical base-10 computer
Example: 1/ 34 = 0.029411765
Suppose only 4 decimal places can be stored
0.029411765 → 0.0294 10 0

The first zero to right of decimal point is useless. So,

normalize by multiplying the mantissa by 10 and
lowering the exponent by 1
0.0294 100 → 0.294110−1

Additional significant figure is retained

© McGraw Hill 13
 Normalization means limited range of
mantissa
1
 m 1
b
Therefore
for a base-10 system 0.1 ≤ m <1
for a base-2 system 0.5 ≤ m <1
Floating point representation allows both fractions and very large
numbers to be expressed on the computer. However,
• Floating point numbers take up more room.
• Take longer to process than integer numbers.
• Round-off errors are introduced because mantissa holds only a
finite number of significant figures.
© McGraw Hill 14
 Chopping & Rounding
Example:
 = 3.14159265358 to be stored on a base-10 system
carrying 7 significant digits.
 = 3.141592 chopping error →  t = 0.00000065
If rounded
 = 3.141593 rounding error →  t = 0.00000035
• Some machines use chopping, because rounding adds
to the computational overhead. Since number of
significant figures is large enough, resulting chopping
error is “usually” negligible.
© McGraw Hill 15
 Biggest Sources of Rounding Errors in
Numerical Calculation
• Large interdependent computations (billions of
operations); tiny errors can compound (add up).
• Adding a large & small number.
• Smearing: Occurs whenever the individual terms in a
summation are larger than the summation itself.
• Subtractive cancellation: round-off induced when
subtracting two nearly equal floating-point numbers.

© McGraw Hill 16
 Because learning changes everything. ®

www.mheducation.com

© 2021 McGraw Hill. All rights reserved. Authorized only for instructor use in the classroom. No reproduction or further distribution permitted without the prior written consent of McGraw Hill.