Advanced Mathematical Techniques Chapter # 1 Error Analysis

CHAPTER # 1

ERROR ANALYSIS

Dr. Jamil Book Series 2 Page 1

Advanced Mathematical Techniques Chapter # 1 Error Analysis

CHAPTER # 1
ERROR ANALYSIS
Nature or types of errors:-

The problems in applied mathematics, engineering and sciences ultimately converge to the
problems for finding the numerical results or numbers satisfying some kind of equations
under prescribed data. The checking procedures are adopted to eliminate all mistakes
whether these are human, due to the computer or technical, due to the calculating device.
Even when all such mistakes have been eliminated from a computation, the solution is not
generally exact on account of inherent errors of various types. The error in numerical
computation can enter in three different ways:
1): Inherent or propagation error,
2): Truncation errors,
3): Round-off errors.
1): Inherent Errors (Propagation or inherited errors):-

These errors are

1): Due to the supplied data which take part in our numerical calculation throughout,
2): Due to the approximate nature of the applied formula used in the solution and
3): It is caused by the use of previous points calculated by the computer which already
have errors.

2): Truncation Error:-

The truncation error may occur as the result of replacing an infinite process by a finite one.
For example, such error appears when a definite integral is computed by Simpson’s rule or
when a differential equation is solved by some difference method. This error is not machine
– caused but is due to the method.
Example #1:-
The infinite series expansion of 𝒆𝒙 in term of Maclaurin’s series is given by

𝒙 𝒙𝟐 𝒙𝒏−𝟏 𝒙𝒏
𝑻𝑽 = 𝒆 = 𝟏 + 𝒙 + 𝟐! + ⋯ + (𝒏−𝟏)!
+ 𝒏! + ⋯ ∴ 𝒇(𝒙) = 𝒆𝒙

If we approximate 𝒆𝒙 by the "𝒏" finite sum

Dr. Jamil Book Series 2 Page 2

Advanced Mathematical Techniques Chapter # 1 Error Analysis

𝒙
𝒙𝟐 𝒙𝒏−𝟏
𝑨𝑽 = 𝒆 ≈ 𝟏 + 𝒙 + + ⋯ +
𝟐! (𝒏 − 𝟏)!

Then
𝑻. 𝑬. = 𝑬𝒓𝒓𝒐𝒓 = 𝑻𝑽 − 𝑨𝑽

𝑻. 𝑬. = (𝒏 + 𝟏)𝒕𝒉 + (𝒏 + 𝟐)𝒕𝒉 + (𝒏 + 𝟑)𝒕𝒉 + ⋯ + ∞

𝒙𝒏 𝒙𝒏+𝟏 𝒙𝒏+𝟐
𝑻. 𝑬. = + + + ⋯+ ∞
𝒏! (𝒏 + 𝟏)! (𝒏 + 𝟐)!

Then 𝑹𝒏 , the remainder after 𝒏 terms (from Maclaurin’s theorem), constitute the
truncation error given by

𝒙𝒏
𝑹𝒏 = (𝒏 + 𝟏 )𝒕𝒉 + (𝒏 + 𝟐 )𝒕𝒉 + (𝒏 + 𝟑 )𝒕𝒉 +⋯+∞ = 𝒇(𝜽𝒙) ∵ 𝒇(𝒙) = 𝒆𝒙
𝒏!

𝒙𝒏 𝒙𝒏 𝜽𝒙
𝑻. 𝑬. = 𝑺𝒖𝒎 𝒐𝒇 (𝒏 + 𝟏) 𝒕𝒐 ∞ 𝒕𝒆𝒓𝒎𝒔 = 𝑹𝒏 = 𝒇(𝜽𝒙) = 𝒆 , 𝟎 < 𝜽 < 𝟏.
𝒏! 𝒏!

Example #2:-
Let 𝒚 = 𝒇(𝒙) be a continuously differentiable function in the domain. Then by definition of
derivative, we have

𝒅𝒚 𝒇(𝒙 + 𝒉) − 𝒇(𝒙)
= 𝒍𝒊𝒎 (𝑰𝒏𝒔𝒕𝒂𝒏𝒕𝒂𝒏𝒆𝒐𝒖𝒔 𝑹𝒂𝒕𝒆)
𝒅𝒙 𝒉→𝟎 𝒉

𝒅𝒚 𝟏𝒔𝒕 𝑺𝒕𝒆𝒑 (𝑪𝒉𝒂𝒏𝒈𝒆 𝒊𝒏 𝒅𝒆𝒑𝒆𝒏𝒕 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒇(𝒙))

= 𝟑𝒓𝒅 𝑺𝒕𝒆𝒑 𝑳𝒊𝒎𝒕
𝒅𝒙 𝟐𝒏𝒅 𝑺𝒕𝒆𝒑(𝑪𝒉𝒂𝒏𝒈𝒆 𝒊𝒏 𝒊𝒏𝒅𝒆𝒑𝒆𝒏𝒕 𝒗𝒂𝒓𝒊𝒂𝒃𝒍𝒆 𝒙)

So, if we write
𝒅𝒚 𝒇(𝒙 + 𝒉) − 𝒇(𝒙)
≈ (𝑨𝒗𝒆𝒓𝒂𝒈𝒆 𝑹𝒂𝒕𝒆)
𝒅𝒙 𝒉

𝒅𝒚 𝒇(𝒙 + 𝒉) − 𝒇(𝒙)
= + 𝑬𝒓𝒓𝒐𝒓
𝒅𝒙 𝒉
We commit an error, arising through the replacement of the infinitesimal process of
calculating the differential coefficient by the finite process of calculating simply the
difference quotient (average): Such an error is also a truncation error 𝑻. 𝑬. We write:

𝒅𝒚 𝒇(𝒙 + 𝒉) − 𝒇(𝒙)
= +
𝒅𝒙 𝒉

Dr. Jamil Book Series 2 Page 3

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Discussion on the Limit:-

Consider the following table:

1.9 1.99 1.999 1.9999 …….

2.9 2.99 2.999 2.9999 ……..
This table tell us that

𝒘𝒉𝒆𝒏 (𝒎𝒆𝒂𝒏 𝒙 ≠ 𝟐) ( ) → 𝟑 (𝒎𝒆𝒂𝒏 𝒇(𝒙) ≠ 𝟑) OR 𝐥𝐢𝐦 𝒇(𝒙) = 𝟑

𝒙→𝒂

Remember 𝒙 → 𝟐 (𝒙 ≠ 𝟐) mean 𝒙 always always moves/approaches/tends to 𝟐. But never

be equal to 2. Because for 𝒙 to be equal to 2, we have to achieve "𝟗", infinity times, which
is impossible to us. Hence 𝒙 never be equal to 2. Similarly when 𝒙 is not equal to 2 this
implies that 𝒇(𝒙) is also never equal to 3.
Therefore, In general, we can say that:
𝒙 → 𝒂(𝒙 ≠ 𝒂), that is 𝒙 always always moves/approaches/tends to 𝒂. But never be equal to
𝒂. Because for 𝒙 equal to "𝒂", we have to achieve "𝒔𝒐𝒎𝒆 𝒏𝒖𝒎𝒃𝒆𝒓(𝒔)", infinity times,
which is impossible to us.

3): Round-off Errors:-

These error are unavoidable in most of the calculations since some of the quantities in the
calculations will be non-terminating decimals and for practical reasons only certain
number of figures will be carried in a calculation. These are due to the fact that in
computational work we have to deal with approximations.
Example #3:-
Consider and round-off it up to 3 decimal places, we get then
The difference that is the number is called
Round-off error. This is occurred due to Rounding-off a number.

Round-off Rules:-
To round-off a number to "𝒏" decimal place, discard all digits to the right of the 𝒏𝒕𝒉 digit,
and if this discarded number is:
1): Less than 𝟓 in the (𝒏 + 𝟏)𝒕𝒉 place, leave the 𝒏𝒕𝒉 digit unaltered.
2): Greater than Or Equal to 𝟓 in the (𝒏 + 𝟏)𝒕𝒉 place, increase the 𝒏𝒕𝒉 digit by unity.

The number thus rounded-off is said to be correct to 𝒏 significant figures.

Dr. Jamil Book Series 2 Page 4

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Rounding off a numerical value to specified numberof decimal places

Numerical Value Rounded-off up to Round-off value

𝟐𝟒. 𝟓𝟔𝟓𝟑 𝟓 𝟒 𝟒𝒅𝒑 𝟐𝟒. 𝟓𝟔𝟓𝟒

𝟐𝟏. 𝟓𝟔𝟒𝟖 𝟖 𝟔 𝟒𝒅𝒑 𝟐𝟏. 𝟓𝟔𝟒𝟗

𝟐𝟕. 𝟒𝟖𝟑𝟓 𝟓 𝟑 𝟒𝒅𝒑 𝟐𝟕. 𝟒𝟖𝟑𝟔

𝟑𝟎. 𝟎𝟑𝟒𝟔 𝟓 𝟑 𝟒𝒅𝒑 𝟑𝟎. 𝟎𝟑𝟒𝟕

𝟏𝟕. 𝟒𝟒𝟓 𝟑 𝟕𝟐 𝟑𝒅𝒑 𝟏𝟕. 𝟒𝟒𝟓

𝟐𝟖. 𝟒𝟓 𝟓 𝟑𝟒𝟐 𝟐𝒅𝒑 𝟐𝟖. 𝟒𝟔

𝟐𝟕. 𝟒𝟖𝟕𝟒𝟔 𝟔 𝟓𝒅𝒑 𝟐𝟕. 𝟒𝟖𝟕𝟒𝟕

𝟐𝟒. 𝟓𝟔𝟒𝟗 𝟖 𝟔 𝟒𝒅𝒑 𝟐𝟒. 𝟓𝟔𝟓𝟎

𝟓𝟎. 𝟎𝟓𝟖𝟒 𝟓 𝟕 𝟒𝒅𝒑 𝟓𝟎. 𝟎𝟓𝟖𝟓

𝟐. 𝟖𝟗𝟖 𝟕 𝟔𝟐 𝟑𝒅𝒑 𝟐. 𝟖𝟗𝟗

𝟏𝟎. 𝟏𝟏𝟑𝟔 𝟓 𝟕𝟑 𝟒𝒅𝒑 𝟏𝟎. 𝟏𝟏𝟑𝟕

𝟏𝟓. 𝟏𝟖𝟕𝟐𝟒 𝟑 𝟓𝟕 𝟓𝒅𝒑 𝟏𝟓. 𝟏𝟖𝟕𝟐𝟒

𝟐𝟖. 𝟑𝟕𝟗 𝟓 𝟐𝟑 𝟑𝒅𝒑 𝟐𝟖. 𝟑𝟖𝟎

𝟒𝟐. 𝟖𝟗𝟗 𝟕 𝟓 𝟑𝒅𝒑 𝟒𝟐. 𝟗𝟎𝟎

𝟑. 𝟕𝟖𝟔𝟒 𝟓 𝟎𝟏 𝟒𝒅𝒑 𝟑. 𝟕𝟖𝟔𝟓

𝟒. 𝟗 𝟔 𝟕𝟖𝟐𝟏 𝟏𝒅𝒑 𝟓. 𝟎

𝟑𝟔. 𝟗𝟗 𝟖 𝟔𝟐𝟏 𝟐𝒅𝒑 𝟑𝟕. 𝟎𝟎

𝟑𝟏𝟐. 𝟗𝟗𝟗 𝟓 𝟑𝟔 𝟑𝒅𝒑 𝟑𝟏𝟑. 𝟎𝟎𝟎

𝟕𝟐. 𝟖𝟓𝟑𝟗 𝟕 𝟒 𝟒𝒅𝒑 𝟕𝟐. 𝟖𝟓𝟒𝟎

𝟏𝟐𝟓. 𝟗𝟗𝟗 𝟗 𝟗𝟗 𝟑𝒅𝒑 𝟏𝟐𝟔. 𝟎𝟎𝟎

Dr. Jamil Book Series 2 Page 5

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Significant Figures

Numerical Value Significant Figures

Rule-1:- Non-zero digits are significant:

𝟔𝟑 𝟐𝑺𝑭

𝟗𝟖𝟓 𝟑𝑺𝑭

𝟐𝟕𝟗𝟑𝟒 𝟓𝑺𝑭

Rule-2:- Zeros between non-zero digits are significant:

𝟐𝟎𝟕 𝟑𝑺𝑭

𝟓𝟖𝟎𝟎𝟗𝟎𝟕 𝟕𝑺𝑭

𝟐𝟒. 𝟎𝟏𝟐𝟑 𝟔𝑺𝑭

Rule-3:- If zeros are not between non-zero digits, then they are not significant:

𝟖𝟎𝟎𝟎𝟎 𝟏𝑺𝑭

𝟕𝟕𝟎𝟎𝟎 𝟐𝑺𝑭

𝟕𝟖𝟔𝟎𝟑𝟎𝟎𝟎 𝟓𝑺𝑭

𝟐𝟔𝟎𝟓𝟎0. 𝟔𝑺𝑭

Rule-4:- Final zeros are significant:

𝟓. 𝟖𝟎𝟎 𝟒𝑺𝑭

𝟑𝟐. 𝟐𝟓𝟎𝟎𝟎 𝟕𝑺𝑭

𝟖. 𝟏𝟎𝟓𝟎 𝟓𝑺𝑭

Rule-5:- Initial zeros are not significant:

𝟎. 𝟎𝟎𝟔 𝟏𝑺𝑭

𝟎. 𝟎𝟎𝟎𝟐𝟒 𝟐𝑺𝑭

𝟎. 𝟎𝟎𝟖𝟓𝟑𝟐 𝟒𝑺𝑭

Miscellaneous Examples:-

𝟐𝟖. 𝟎𝟕𝟖𝟎 𝟔𝑺𝑭

𝟔𝟔. 𝟎𝟎𝟗𝟎𝟐𝟏𝟎 𝟗𝑺𝑭

𝟐𝟖𝟎𝟔𝟗𝟎𝟎. 𝟕𝑺𝑭

Dr. Jamil Book Series 2 Page 6

Advanced Mathematical Techniques Chapter # 1 Error Analysis

𝟎. 𝟎𝟖𝟐𝟎𝟎𝟗𝟑 𝟔𝑺𝑭

𝟎. 𝟎𝟎𝟖𝟔𝟎 𝟑𝑺𝑭

𝟐𝟎𝟓𝟎. 𝟎𝟖𝟎𝟗𝟎𝟑𝟔 𝟏𝟏𝑺𝑭

𝟖𝟖𝟎𝟐𝟑𝟎𝟎𝟎 𝟓𝑺𝑭

Significant Errors:-
This is also a computational error and enters in a computation due to the loss of significant
digits. This type of error usually occurs in two cases:
1): When almost two equal numbers be subtracted.
2): When a number is divided by a very small number.
Example # 4:-
Consider almost two equal numbers of eight significant figures as:
𝒙 = 𝟏. 𝟐𝟑𝟒 𝟓𝟔𝟐𝟕 and 𝒚 = 𝟏. 𝟐𝟑𝟒 𝟓𝟓𝟖 𝟒, then

𝒙 − 𝒚 = 𝟎. 𝟎𝟎𝟎 𝟎𝟎𝟒𝟑

Now it is clear that 𝒙 − 𝒚 has only two significant figures namely 𝟒 and 𝟑 .
Therefore 𝒙 − 𝒚 losses six significant digits though 𝒙 and 𝒚 are both of eight significant
digits. If this value of 𝒙 − 𝒚 be used in further computation the subsequent result must be
affected by a serious amount of errors.
Example # 5:-
As another example, we consider a function
𝟒
𝒇(𝒙) =
𝟏 − 𝒙𝟑

Now we consider a number 𝒙 = 𝟎. 𝟗𝟗𝟗, such that 𝟏 − 𝒙𝟑 = 𝟎. 𝟎𝟎𝟐𝟗𝟗𝟕 is very small with
respect to dividend 𝟒. Now
𝟒
𝒇(𝟎. 𝟗𝟗𝟗) = = 𝟏𝟑𝟑𝟒. 𝟔𝟔𝟕𝟓
𝟏 − (𝟎. 𝟗𝟗𝟗)𝟑

If we take 𝒙 = 𝟎. 𝟗𝟗𝟗𝟎𝟎𝟎𝟎𝟔 i.e., if 𝒙 has an error in the 𝟖𝒕𝒉 place, then

𝟒
𝒇(𝟎. 𝟗𝟗𝟗𝟎𝟎𝟎𝟎𝟔) = = 𝟏𝟑𝟑𝟒. 𝟕𝟒𝟕𝟔
𝟏 − (𝟎. 𝟗𝟗𝟗𝟎𝟎𝟎𝟎𝟔)𝟑

From above two calculations, it is clear that the error in computed value of 𝒇(𝒙): occur in
the 𝟓𝒕𝒉 significant digit corresponding to the error in 𝒙 at the 𝟖𝒕𝒉 place. This error ccurs
due to the loss of significant figures.

Dr. Jamil Book Series 2 Page 7

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Drill Excercise # 1:-

Round-off the following numbers correct up to 3-decimal places
a): 2.46289 b): 0.46999 c): 0.0035869
d): 0.0015 e): 1.46294 f): 1.35008
Answer:-
a): 2.463 b): 0.470 c): 0.004 d): 0.002 e): 1.463 f): 1.350

Rules of modren Arthmetics:-

In modern arithmetical computation, the following rules should be followed.
1): For addition:-
a): The numbers which are least number of correct significant figures should be kept
unchanged.
b): The rest numbers should be rounded-off to one or two decimal places relative to those
which are least number of correct significant figures.
c): Then the number so obtained be added and then rounded-off the added number by
retaining as many significant figures as there are in the least accurate number among the
given numbers.
2): For subtraction:-
When the subtraction of two approximate number is required, both the number should be
rounded-off to same decimal places at first, then after the numbers should be subtracted.
3): For multiplication:-
In this case the rules are same as was followed for addition.
Example # 6:-
Find the Sum of the approximate numbers:
𝟐𝟒𝟓. 𝟕𝟐 , 𝟐𝟓𝟖. 𝟖 , 𝟏𝟕. 𝟓 , 𝟎. 𝟎𝟎𝟖𝟕 , 𝟎. 𝟎𝟎𝟎𝟑𝟏 , 𝟒𝟏. 𝟓.
Solution:-
The least accurate numbers are 𝟐𝟓𝟖. 𝟖 , 𝟏𝟕. 𝟓 , 𝟒𝟏. 𝟓 . Rounding-off the remaining numbers
to up one or two decimal places. We decide to round-off the large decimal numbers up to
two decimal places. Then

𝟐𝟒𝟓. 𝟕𝟐 + 𝟐𝟓𝟖. 𝟖 + 𝟏𝟕. 𝟓 + 𝟎. 𝟎𝟏 + 𝟎. 𝟎𝟎 + 𝟒𝟏. 𝟓 = 𝟓𝟔𝟑. 𝟓𝟑 ≈ 𝟓𝟔𝟑. 𝟓

Example # 7:-
Subtract the approximate number 𝟓𝟕. 𝟒𝟔𝟔𝟐 from 𝟕𝟖𝟔. 𝟖𝟓 .
Solution:-
We first round-off the large decimal number 𝟓𝟕. 𝟒𝟔𝟔𝟐 correct to two decimal places, then

𝟕𝟖𝟔. 𝟖𝟓 − 𝟓𝟕. 𝟒𝟕 = 𝟕𝟐𝟗. 𝟑𝟖

Dr. Jamil Book Series 2 Page 8

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Example # 8:-
Find the Product of 𝟏𝟕𝟏. 𝟕𝟐 , 𝟓𝟏. 𝟎𝟐𝟕 , 𝟎. 𝟎𝟎𝟖𝟓.
Solution:-
We first round-off the large decimal numbers correct to two or three decimal places. We
decide to round-off the large decimal numbers up to three decimal places. Then

𝟏𝟕𝟏. 𝟕𝟐 × 𝟓𝟏. 𝟎𝟐𝟕 × 𝟎. 𝟎𝟎𝟗 = 𝟕𝟖. 𝟖𝟔𝟏 ≈ 𝟕𝟖. 𝟖𝟔

Drill Excercise # 2:-

The following numbers are correct up to last digit. Find the sum and Product:
a): 1.3526, 2.00462, 1.532, 28.201, 31.0012
b): 0.532, 7.46571, 1.501, 3.62102
c): 5.2146, 20.12, 11.2356, 1.8948
Answer:-
Sum: a): 64.091 b): 13.120 c): 38.46
Product: a): b): c):

Absolute, Relative and Percentage Error

1): Absolute Error 𝑬𝒂 :-
The absolute error is defined as the difference between its true value 𝑽𝑻 and its
approximate value 𝑽𝑨 i.e.

𝑨𝒃𝒔𝒐𝒍𝒖𝒕𝒆 𝒆𝒓𝒓𝒐𝒓 = 𝑬𝒂 = ∆𝑽 = | 𝑽𝑻 − 𝑽𝑨 |

2): The Relative Error 𝑬𝒓 :-

The relative error of a number is known as the absolute error divided by its true value i.e.

∆𝑽 𝑬𝒂 |𝑽𝑻 − 𝑽𝑨 |
𝑹𝒆𝒍𝒂𝒕𝒊𝒗𝒆 𝒆𝒓𝒓𝒐𝒓 = 𝑬𝒓 = = =
𝑽𝑻 𝑽𝑻 𝑽𝑻

3): The Percentage Error 𝑬𝒑 :-

The percentage error is defined as the relative error multiplied by 𝟏𝟎𝟎 i.e

| 𝑽𝑻 −𝑽𝑨 |
𝑷𝒆𝒓𝒄𝒆𝒏𝒕𝒂𝒈𝒆 𝒆𝒓𝒓𝒐𝒓 = 𝑬𝒑 = 𝑬𝒓 × 𝟏𝟎𝟎 = × 𝟏𝟎𝟎
𝑽𝑻

Dr. Jamil Book Series 2 Page 9

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Example:-
If 𝒗𝑻 = 𝟐. 𝟓𝟑𝟕 𝟓𝟖𝟒 𝟓𝟔𝟗𝟐, then its 3-Decimal Places Round-off number is

𝒗𝑨 = 𝟐. 𝟓𝟑𝟖 (𝟑 𝑫𝒆𝒄𝒊𝒎𝒂𝒍 𝑷𝒍𝒂𝒄𝒆𝒔 𝑨𝒑𝒑𝒓𝒐𝒙𝒊𝒎𝒂𝒕𝒆 𝒏𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝒗𝑻 )

Similarly, the 3-Significant Digits Round-off number of 𝒗𝑻 is

𝒗𝑨 = 𝟐. 𝟓𝟒 (𝟑 𝐒𝐢𝐠𝐧𝐢𝐟𝐢𝐜𝐚𝐧𝐭 𝐃𝐢𝐠𝐢𝐭𝐬 𝑨𝒑𝒑𝒓𝒐𝒙𝒊𝒎𝒂𝒕𝒆 𝒏𝒖𝒎𝒃𝒆𝒓 𝒐𝒇 𝒗𝑻 )

This Example yields that in general the 3 Decimal Places Round-off number is NOT equal
to 3 Significant Digits Round-off number.

Theorem - 1:-
If 𝒗𝑻 is a true number and its 𝒏-Decimal Places Rounding-off number is 𝒗𝑨 (𝒕𝒉𝒆𝒏 𝑬𝒂 =
|𝒗𝑻 − 𝒗𝑨 |), the Absolute Error is given by

𝟏
𝑬𝒂 ≤ × 𝟏𝟎−𝒏 = 𝟎. 𝟓 × 𝟏𝟎−𝒏
𝟐

Theorem - 2:-
If 𝒗𝑻 is a true number and its 𝒏-Significant Digits Rounding-off number is 𝒗𝑨 (𝒕𝒉𝒆𝒏 𝑬𝒂 =
𝑬𝒂
|𝒗𝑻 − 𝒗𝑨 | 𝒂𝒏𝒅 𝑬𝒓 = ), the Relative Error is given by
𝒗𝑻

𝟏
𝑬𝒓 ≤ (𝒏 ≠ 𝟏):
𝑲 × 𝟏𝟎𝒏−𝟏

where 𝑲 is the First Significant Digit in the number (either in true 𝒗𝑻 or approximate
number 𝒗𝑨 ).

Note:-
1): 𝟑 < 𝟓 (𝑻𝒓𝒖𝒆) 2): 𝟑 ≤ 𝟓 (𝑻𝒓𝒖𝒆) 3): 𝟑 ≤ 𝟑 (𝑻𝒓𝒖𝒆) 4): 𝟑 ≥ 𝟑 (𝑻𝒓𝒖𝒆)

Here “or” is actually “OR GATE”.

Example # 9:-
Let a number 𝟑𝟒𝟓. 𝟐𝟔𝟏𝟑𝟐(= 𝒗𝑻 ) be rounded to 𝟑𝟒𝟓. 𝟐𝟔𝟏(= 𝒗𝑨 ) correct to 𝟑 decimal
places, then absolute error is
𝑬𝒂 = 𝒗𝑻 − 𝒗𝑨 = 𝟑𝟒𝟓. 𝟐𝟔𝟏𝟑𝟐 − 𝟑𝟒𝟓. 𝟐𝟔𝟏 = 𝟎. 𝟎𝟎𝟎𝟑𝟐.
Verification:-
From theorem -1, we have (here 𝒏 = 𝟑):-
𝟏
𝑬𝒂 ≤ . 𝟏𝟎−𝒏
𝟐
𝟏
𝑬𝒂 ≤ . 𝟏𝟎−𝟑 = 𝟎. 𝟎𝟎𝟎𝟓
𝟐

Dr. Jamil Book Series 2 Page 10

Advanced Mathematical Techniques Chapter # 1 Error Analysis

𝑬𝒂 = 𝟎. 𝟎𝟎𝟎 𝟑𝟐 ≤ 𝟎. 𝟎𝟎𝟎 𝟓
Hence theorem – 1 is verified. Now

𝑬𝒂 𝟎.𝟎𝟎𝟎𝟑𝟐
𝑬𝒓 = = = 𝟗. 𝟐𝟔𝟖 × 𝟏𝟎−𝟕
𝑽𝑻 𝟑𝟒𝟓.𝟐𝟔𝟏𝟑𝟐

From theorem – 2, we have (here = 𝟑, 𝒏 = 𝟔 ):-

𝟏
𝑬𝒓 ≤
𝑲 × 𝟏𝟎𝒏−𝟏
𝟏 𝟏
𝑬𝒓 ≤ 𝟔−𝟏
= 𝟓
= 𝟑. 𝟑𝟑𝟑 × 𝟏𝟎−𝟔
𝟑 × 𝟏𝟎 𝟑 × 𝟏𝟎

⇒ 𝑬𝒓 = 𝟗. 𝟐𝟔𝟖 × 𝟏𝟎−𝟕 ≤ 𝟑. 𝟑𝟑𝟑 × 𝟏𝟎−𝟔

Hence theorem – 2 is verified.

Example # 10:-
𝟐
Write down the approximate representation of correct to four significant figures and
𝟑
then find:
1): Absolute error
2): Relative error
3): Percentage error
Solution:-
𝟐 𝟐 𝟐
The four significant figures representation of is = 𝟎. 𝟔𝟔𝟔𝟕 ⇒ 𝑽𝑻 = , 𝑽𝑨 = 𝟎. 𝟔𝟔𝟔𝟕
𝟑 𝟑 𝟑
𝟐
𝟏): 𝑬𝒂 = | 𝑽𝑻 − 𝑽𝑨 | = | − 𝟎. 𝟔𝟔𝟔𝟕| = 𝟎. 𝟎𝟎𝟎 𝟎𝟑𝟑
𝟑

𝑬𝒂 𝟎. 𝟎𝟎𝟎𝟎𝟑𝟑
𝟐): 𝑬𝒓 = = = 𝟎. 𝟎𝟎𝟎𝟎𝟒𝟗𝟓 ≈ 𝟎. 𝟎𝟎𝟎 𝟎𝟓
𝑽𝑻 𝟐/𝟑

𝟑): 𝑬𝒑 = 𝑬𝒓 × 𝟏𝟎𝟎 = 𝟎. 𝟎𝟎𝟓 %

Verification:-
From theorem -1, we have (here 𝒏 = 𝟒):-
𝟏
𝑬𝒂 ≤ . 𝟏𝟎−𝒏
𝟐
𝟏
𝑬𝒂 ≤ . 𝟏𝟎−𝟒 = 𝟎. 𝟎𝟎𝟎𝟎𝟓
𝟐

𝑬𝒂 = 𝟎. 𝟎𝟎𝟎 𝟎𝟑𝟑 ≤ 𝟎. 𝟎𝟎𝟎 𝟎𝟓

Hence theorem -1 is verified

Dr. Jamil Book Series 2 Page 11

Advanced Mathematical Techniques Chapter # 1 Error Analysis

From theorem – 2, we have (here 𝑲 = 𝟔, 𝒏 = 𝟒): −

𝟏
𝑬𝒓 ≤
𝑲 × 𝟏𝟎𝒏−𝟏
𝟏 𝟏
𝑬𝒓 ≤ = = 𝟎. 𝟎𝟎𝟎𝟏𝟔𝟔 ≈ 𝟎. 𝟎𝟎𝟎𝟏𝟕
𝟔 × 𝟏𝟎𝟒−𝟏 𝟔 × 𝟏𝟎𝟑

𝑬𝒓 = 𝟎. 𝟎𝟎𝟎 𝟎𝟓 ≤ 𝟎. 𝟎𝟎𝟎 𝟏𝟕

Hence theorem -2 is verified.

Drill Excercise # 3:-

Find the Relative Percentage Error in approximate representation of 4/3 by 1.33 and verify
theorem-1 and theorem-2.
Answer:-
𝑬𝒑 = 𝟎. 𝟐𝟓%

Example # 11:-
Find the number of Significant Figures in Approximate number 𝑽𝑨 if:
𝐚): 𝑽𝑨 = 𝟏𝟏. 𝟐𝟒𝟔𝟏 given its absolute error as 𝟎. 𝟐𝟓 × 𝟏𝟎−𝟐 .

𝐛): 𝑽𝑻 = 𝟏. 𝟓𝟗𝟐𝟑𝟓𝟖 given its relative error as 𝟎. 𝟏 × 𝟏𝟎−𝟒 .

𝐜): 𝑽𝑨 = 𝟎. 𝟑𝟗𝟒𝟏 given its absolute error as 𝟎. 𝟐𝟓 × 𝟏𝟎−𝟐 .

Solution:-
a): Here
𝟏 𝟏
𝑬𝒂 = 𝟎. 𝟐𝟓 × 𝟏𝟎−𝟐 = × 𝟏𝟎−𝟐 ≤ × 𝟏𝟎−𝟐
𝟒 𝟐

Then by theorem – 1, 𝑽𝑨 is correct to two decimal places that is 𝑽𝑨 = 𝟏𝟏. 𝟐𝟒 ⇒ 𝑽𝑨 is

correct to 𝟒-significant figures.
b):We have
𝑬𝒂 = 𝑽𝑻 × 𝑬𝒓 = 𝟏. 𝟓𝟗𝟐𝟑 × 𝟎. 𝟏 × 𝟏𝟎−𝟒

⇒ 𝑬𝒂 = 𝟎. 𝟏𝟓𝟗𝟐𝟑𝟓𝟖 × 𝟏𝟎−𝟒 ≈ 𝟎. 𝟐 𝐱 𝟏𝟎−𝟒 ≤ 𝟎. 𝟓 × 𝟏𝟎−𝟒

Thus, by theorem -1, the Approximate number is correct to three decimal places is 𝑽𝑨 =
𝟏. 𝟓𝟗𝟐𝟒 . Thus, the number of significant figures in 𝑽𝑨 is 5.

Dr. Jamil Book Series 2 Page 12

Advanced Mathematical Techniques Chapter # 1 Error Analysis

c): Here

𝟏 𝟏
𝑬𝒂 = 𝟎. 𝟐𝟓 × 𝟏𝟎−𝟐 = × 𝟏𝟎−𝟐 < . 𝟏𝟎−𝟐
𝟒 𝟐

Thus, by theorem – 1 the number is correct to two decimal places. Hence the number of
significant figures is 2.

Drill Excercise # 4:-

If 𝑽𝑻 , 𝑽𝑨 , 𝑬𝒂 , 𝑬𝒓 𝐚𝐧𝐝 𝑬𝒑 are true value, approximate value, absolute error, relative error
and percentage error respectively:
a): Given 𝑽𝑻 = 𝟐. 𝟓𝟒𝟔𝟐𝟖𝟐, 𝑽𝑨 = 𝟐. 𝟓𝟒𝟔𝟑. Find 𝑬𝒂 , 𝑬𝒓 , 𝑬𝒑 .
b): Given 𝑬𝒂 = 𝟎. 𝟐 × 𝟏𝟎−𝟑 , 𝑬𝒓 = 𝟎. 𝟑𝟐 × 𝟏𝟎−𝟓 . Find 𝑽𝑻
c): Given 𝑽𝑻 = 𝟐. 𝟕𝟑𝟒𝟓𝟔, 𝑬𝒓 = 𝟎. 𝟐 × 𝟏𝟎−𝟐 . Find 𝑬𝒂 , 𝑬𝒑
Answer:-
a): 𝑬𝒂 = 𝟏. 𝟖 × 𝟏𝟎−𝟓 , 𝑬𝒓 = 𝟕. 𝟎𝟕 × 𝟏𝟎−𝟔 , 𝑬𝒑 = 𝟕. 𝟎𝟕 × 𝟏𝟎−𝟒 %
b): 𝟔𝟐. 𝟓
c): 𝑬𝒂 = 𝟎. 𝟎𝟎𝟓𝟒𝟕, 𝑬𝒑 = 𝟎. 𝟐%
Drill Excercise # 5:-
If 5/6 be represented approximately by 0.8333, find
a): Absolute error
b): Relative error
c): Percentage error and also verify theorem-1 and theorem-2.
Answer:-
𝑬𝒓 = 𝟎.00004, 𝑬𝒑 = 𝟎. 𝟎𝟎𝟒 %
Drill Excercise # 6:-
Determine the number of correct Significant Digits in the numbers 𝒙, given its absolute
error 𝑬𝒂 :
a): 𝒙 = 𝟎. 𝟑𝟔𝟖𝟏, 𝑬𝒂 = 𝟎. 𝟐𝟑 × 𝟏𝟎−𝟐
b): 𝒙 = 𝟖𝟒𝟏. 𝟐𝟓𝟔, 𝑬𝒂 = 𝟎. 𝟏𝟓 × 𝟏𝟎−𝟐
c): 𝒙 = −𝟑𝟑. 𝟕𝟖𝟑, 𝑬𝒂 = 𝟎. 𝟑 × 𝟏𝟎−𝟐
Answer:-
a): 2 b): c): 4

Dr. Jamil Book Series 2 Page 13

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Example # 12:-
If ∆𝒙 = 𝟎. 𝟎𝟎𝟓 and ∆𝒚 = 𝟎. 𝟎𝟎𝟏 be the absolute error in 𝒙 = 𝟐. 𝟏𝟏 and 𝒚 = 𝟒. 𝟏𝟓, find the
relative error in computation of 𝒙 + 𝒚.
Solution:-
Here

𝒙 + 𝒚 = 𝟐. 𝟏𝟏 + 𝟒. 𝟏𝟓 = 𝟔. 𝟐𝟔
The result has an absolute error.

∆(𝒙 + 𝒚) = ∆𝒙 + ∆𝒚 = 𝟎. 𝟎𝟎𝟓 + 𝟎. 𝟎𝟎𝟏 = 𝟎. 𝟎𝟎𝟔

Therefore, relative error in 𝒙 + 𝒚 is

𝑨𝒃𝒔𝒐𝒍𝒖𝒕𝒆 𝑬𝒓𝒓𝒐𝒓 ∆(𝒙 + 𝒚) ∆𝒙 + ∆𝒚 𝟎. 𝟎𝟎𝟔
𝑬𝒓 = = = =
𝑻𝒓𝒖𝒆 𝒗𝒂𝒍𝒖𝒆 𝒙+𝒚 𝒙+𝒚 𝟔. 𝟐𝟔

𝑬𝒓 = 𝟎. 𝟎𝟎𝟎𝟗𝟓𝟖 ≈ 𝟎. 𝟎𝟎𝟏

which is the required error in 𝒙 + 𝒚.

Note(Avearage Rate is Approximation of Instanteneous Rate):-

Remember this:

𝑨𝒗𝒆𝒂𝒓𝒂𝒈𝒆 𝑹𝒂𝒕𝒆 ≅ 𝑰𝒏𝒔𝒕𝒂𝒏𝒕𝒆𝒏𝒆𝒐𝒖𝒔 𝑹𝒂𝒕𝒆

∆𝒚 𝒅𝒚 ∆𝒇(𝒙) 𝒅𝒇(𝒙)
≅ ⇒ ≅
∆𝒙 𝒅𝒙 ∆𝒙 𝒅𝒙

Example # 13:-
If 𝒚 = 𝒚(𝒙) = 𝟒𝒙𝟔 − 𝟓𝒙 , find the Percentage Error in 𝒚 at 𝒙 = 𝟏 , if the error in 𝒙 is
𝟎. 𝟎𝟒.
Solution:-
We have given: 𝒚 = 𝟒𝒙𝟔 − 𝟓𝒙, 𝒙 = 𝟏 and ∆𝒙 = 𝟎. 𝟎𝟒 . Then error in computation of 𝒚 is
given by

∆𝒚 𝒅𝒚 𝒅𝒚 ∆𝒚
∆𝒚 = ∆𝒙 ≈ ∆𝒙 = (𝟐𝟒𝒙𝟓 − 𝟓)∆𝒙 ∴ ≈
∆𝒙 𝒅𝒙 𝒅𝒙 ∆𝒙

At 𝒙 = 𝟏 and ∆𝒙 = 𝟎. 𝟎𝟒, we have

∆𝒚 = (𝟐𝟒(𝟏)𝟓 − 𝟓) × 𝟎. 𝟎𝟒 = 𝟎. 𝟕𝟔

and

𝒚 = 𝟒(𝟏)𝟔 − 𝟓(𝟏) = −𝟏 𝒂𝒕 𝒙 = 𝟏

Dr. Jamil Book Series 2 Page 14

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Thus, percentage error in 𝒚 is (note that we are interested in only +𝒗𝒆 quantity of error.

∆𝒚 𝟎. 𝟕𝟔
𝑬𝒑 = 𝑬𝒓 × 𝟏𝟎𝟎 = | | × 𝟏𝟎𝟎 = | | × 𝟏𝟎𝟎
𝒚 −𝟏

𝑬𝒑 = 𝟕𝟔%

Example # 14:-
If 𝒇(𝒙) = 𝟒 𝐜𝐨𝐬 𝒙 − 𝟔𝒙 , find the relative and percentage error in 𝒇(𝒙) for 𝒙 = 𝟎, if error
in 𝒙 = 𝟎. 𝟎𝟎𝟓.
Solution:-
We have given: 𝒇(𝒙) = 𝟒 𝐜𝐨𝐬 𝒙 − 𝟔𝒙, 𝒙 = 𝟎 and ∆𝒙 = 𝟎. 𝟎𝟎𝟓 . Then error in computation
of 𝒇(𝒙) is given by

∆𝒇(𝒙) 𝒅𝒇(𝒙)
∆𝒇(𝒙) = ∆𝒙 ≈ ∆𝒙
∆𝒙 𝒅𝒙

∆𝒇(𝒙) ≈ ( −𝟒𝒔𝒊𝒏𝒙 − 𝟔)∆𝒙

At 𝒙 = 𝟎, the absolute error is computation 𝒇(𝒙)

∆𝒇(𝒙) = (−𝟒 𝒔𝒊𝒏(𝟎) − 𝟔) × 𝟎. 𝟎𝟎𝟓 = −𝟎. 𝟎𝟑

and

𝒇(𝟎) = 𝟒𝒄𝒐𝒔(𝟎) − 𝟔(𝟎) = 𝟒 at 𝒙 = 𝟎

Therefore, percentage error is

∆𝒇(𝒙) 𝟎. 𝟎𝟑
𝑬𝒑 = 𝑬𝒓 × 𝟏𝟎𝟎 = | | × 𝟏𝟎𝟎 = × 𝟏𝟎𝟎
𝒇(𝟎) 𝟒

𝑬𝒑 = 𝟎. 𝟕𝟓%

Drill Excercise # 7:-

The discharge 𝐐 over a notch for head 𝑯 is calculated by the formula 𝐐 = 𝐊𝑯𝟓/𝟐 , where
𝐊 is a given constant. If the head is 𝟕𝟓 𝐜𝐦 and an error of 𝟎. 𝟏𝟓 𝐜𝐦 possible in its
measurement, estimate the Percentage Error in computing the discharge.
Answer:-
∆𝑸 𝟓 𝟎.𝟏𝟓 𝟏
× 𝟏𝟎𝟎 = × × 𝟏𝟎𝟎 = = 𝟎. 𝟓%
𝑸 𝟐 𝟕𝟓 𝟐

Dr. Jamil Book Series 2 Page 15

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Example # 15:-
The exponential function 𝒇(𝒙) = 𝒆𝒙 can be expand as

𝒙
𝒙𝟐 𝒙𝒏−𝟏 𝒙𝒏
𝒆 = 𝟏 + 𝒙 + + ⋯+ + +⋯
𝟐! (𝒏 − 𝟏)! 𝒏!

Find 𝒏, the number of terms, such that their sum yields the value of 𝒆𝒙 correct to 8 decimal
places at 𝒙 = 𝟏.
Solution:-
The error term (i.e., remainder term) in 𝒇(𝒙) = 𝒆𝒙 is given by

𝒙𝒏 𝒙𝒏 𝜽𝒙
𝑹𝒏 = 𝒇(𝜽𝒙) = 𝒆 , 𝟎<𝜽<𝟏
𝒏! 𝒏!

At 𝒙 = 𝟏 and for maximum absolute error, we put 𝜽 = 𝟏, then

𝒆
𝑨𝒃𝒔𝒐𝒍𝒖𝒕𝒆 𝑬𝒓𝒓𝒐𝒓 = 𝑹𝒏 =
𝒏!

Then relative error in 𝒆𝒙 when 𝒙 = 𝟏, is given by

𝒆
𝑨𝒃𝒔𝒐𝒍𝒖𝒕𝒆 𝒆𝒓𝒓𝒐𝒓 𝑹𝒏 𝒏!
𝑬𝒓 = = =
𝑬𝒙𝒂𝒄𝒕 𝒗𝒂𝒍𝒖𝒆 𝒆 𝒆

𝟏
⇒ 𝑬𝒓 =
𝒏!

For an 𝟖 decimal accuracy at 𝒙 = 𝟏, we must have, by theorem-2, as:

𝟏
𝑬𝒓 ≤ 𝑲×𝟏𝟎𝒏−𝟏 ∴ 𝒆 = 𝟐. 𝟕𝟏𝟖 ⇒ 𝑲 = 𝟐, 𝒏 = 𝟗 (𝑭𝒐𝒓 𝒂𝒏 𝟖 𝒅𝒆𝒄𝒊𝒎𝒂𝒍 𝒂𝒄𝒄𝒖𝒓𝒂𝒄𝒚 𝒊𝒏 𝒆)

Then

𝟏 𝟏 𝟏 𝟏
< ⇒ < × 𝟏𝟎−𝟖 ⇒ 𝒏 = 𝟏𝟐
𝒏! 𝟐 × 𝟏𝟎𝟗−𝟏 𝒏! 𝟐

The least value of the 𝒏 that satisfy above equation is 𝒏 = 𝟏𝟐.Thus, we need to take 𝟏𝟐
terms of the exponential series in order that its sum is correct to 𝟖 decimal places.

Dr. Jamil Book Series 2 Page 16

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Total Differential Coefficient (As an Error):-

If 𝒛 = 𝒇(𝒙, 𝒚), then its total Differential coefficient is defined as:

𝝏𝒇 𝝏𝒇
𝒅𝒛 = 𝒅𝒙 + 𝒅𝒚
𝝏𝒙 𝝏𝒚

With the help of total differential coefficient, we can calculate change, error, accuracy and
increment etc. For example in the language of Error analysis, we can write it as:
𝝏𝒇 𝝏𝒇
∆𝒛 = ∆𝒙 + ∆𝒚
𝝏𝒙 𝝏𝒚

Now, we will discuss it applications in Error analysis in the following.

General Error Formula

Let

𝒖 = 𝒇(𝒙𝟏 , 𝒙𝟐 , 𝒙𝟑 , … , 𝒙𝒏 )

Be a function of several variables 𝒙𝒊 (𝒊 = 𝟏, 𝟐, … , 𝒏) and let the error in each 𝒙𝒊 be ∆𝒙𝒊 .

Then the Absolute Error ∆𝒖 in 𝒖 is given by total differential coefficient as
𝝏𝒖 𝝏𝒖 𝝏𝒖
∆𝒖 = ∆𝒙𝟏 + ∆𝒙𝟐 + ⋯ + ∆𝒙
𝝏𝒙𝟏 𝝏𝒙𝟐 𝝏𝒙𝒏 𝒏
𝒏
𝝏𝒖
∆𝒖 = ∑ ∆𝒙
𝝏𝒙𝒊 𝒊
𝒊=𝟏

The formula for the relative error follows immediately, we get

∆𝒖 𝝏𝒖 ∆𝒙𝟏 𝝏𝒖 ∆𝒙𝟐 𝝏𝒖 ∆𝒙𝒏
𝑬𝒓 = = + +⋯+
𝒖 𝝏𝒙𝟏 𝒖 𝝏𝒙𝟐 𝒖 𝝏𝒙𝒏 𝒖

𝒏
𝝏𝒖 ∆𝒙𝒊
𝑬𝒓 = ∑
𝝏𝒙𝒊 𝒖
𝒊=𝟏

𝒏
𝝏𝒖 ∆𝒙𝒊
𝑷𝒓 = 𝑬𝒓 × 𝟏𝟎𝟎 = ∑ × 𝟏𝟎𝟎
𝝏𝒙𝒊 𝒖
𝒊=𝟏

Example # 16:-
𝟓𝒙𝒚𝟐
Find the Maximum Percentage Error in 𝒖 = , where ∆𝒙 = ∆𝒚 = ∆𝒛 = 𝟎. 𝟎𝟎𝟏 and 𝒙 =
𝒛𝟑
𝒚 = 𝒛 = 𝟏.
Solution:-

Dr. Jamil Book Series 2 Page 17

Advanced Mathematical Techniques Chapter # 1 Error Analysis

We have given
𝟓𝒙𝒚𝟐
𝒖= , 𝒙 = 𝒚 = 𝒛 = 𝟏 and ∆𝒙 = ∆𝒚 = ∆𝒛 = 𝟎. 𝟎𝟎𝟏
𝒛𝟑
∆𝒖
and we have to determine 𝑬𝒑 = 𝑬𝒓 × 𝟏𝟎𝟎 = 𝒖
× 𝟏𝟎𝟎 =?

From general formula of error, we have

𝝏𝒖 𝝏𝒖 𝝏𝒖
∆𝒖 = ∆𝒙 + ∆𝒚 + ∆𝒛
𝝏𝒙 𝝏𝒚 𝝏𝒛

where,
𝝏𝒖 𝟓𝒚𝟐 𝝏𝒖 𝟏𝟎𝒙𝒚 𝝏𝒖 𝟏𝟓𝒙𝒚𝟐
= 𝟑 , = 𝟑 , =−
𝝏𝒙 𝒛 𝝏𝒚 𝒛 𝝏𝒛 𝒛𝟒
Then

𝟓𝒚𝟐 𝟏𝟎𝒙𝒚 𝟏𝟓𝒙𝒚𝟐

∆𝒖 = 𝟑 ∆𝒙 + 𝟑 ∆𝒚 − ∆𝒛
𝒛 𝒛 𝒛𝟒

In general, the error ∆𝒙, ∆𝒚 and ∆𝒛 may be +𝒗𝒆 or −𝒗𝒆 , and hence, we take the absolute
values of the terms on right side and using
(|𝒙 + 𝒚 + 𝒛| ≤ |𝒙| + |𝒚| + |𝒛| 𝒂𝒏𝒅 |𝒙𝒚𝒛| = |𝒙||𝒚||𝒛|).
This gives
𝟓𝒚𝟐 𝟏𝟎𝒙𝒚 𝟏𝟓𝒙𝒚𝟐
|∆𝒖| = | 𝟑 ∆𝒙 + 𝟑 ∆𝒚 − |
𝒛 𝒛 𝒛𝟒

𝟓𝒚𝟐 𝟏𝟎𝒙𝒚 𝟏𝟓𝒙𝒚𝟐

∆𝒖 ≤ | 𝟑 ∆𝒙| + | 𝟑 ∆𝒚| + | 𝟒 ∆𝒛|
𝒛 𝒛 𝒛

At 𝒙 = 𝒚 = 𝒛 = 𝟏 and ∆𝒙 = ∆𝒚 = ∆𝒛 = 𝟎. 𝟎𝟎𝟏 we have

∆𝒖 ≈ (𝟓 + 𝟏𝟎 + 𝟏𝟓) × 𝟎. 𝟎𝟎𝟏 = 𝟎. 𝟎𝟑

Since
𝟓𝒙𝒚𝟐 𝟓(𝟏)(𝟏𝟐 )
𝒖= 𝟑 = = 𝟓 𝒂𝒕 𝒙 = 𝒚 = 𝒛 = 𝟏
𝒛 𝟏𝟑
Then the relative error 𝑬𝒓 is given by

∆𝒖 𝟎. 𝟎𝟑
𝑬𝒓 = = = 𝟎. 𝟎𝟎𝟔
𝒖 𝟓

⇒ 𝑬𝒑 = 𝑬𝒓 × 𝟏𝟎𝟎 = 𝟎. 𝟎𝟎𝟔 × 𝟏𝟎𝟎 = 𝟎. 𝟔%

Example # 17:-
𝒎𝒗𝟐
If the kinetic energy 𝑻 = , find the Absolute Error and Percentage Error in 𝑻 as 𝒎 have
𝟐
true value 49.5 and approximate value 49 and 𝒗 have true value 𝟏𝟓𝟗𝟎 and approximate
value 𝟏𝟔𝟎𝟎.

Dr. Jamil Book Series 2 Page 18

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Solution:-
We have given
𝒎𝒗𝟐
𝑻= , 𝒎𝑻 = 𝟒𝟗. 𝟓, 𝒎𝑨 = 𝟒𝟗, 𝒗𝑻 = 𝟏𝟓𝟗𝟎, 𝒗𝑨 = 𝟏𝟔𝟎𝟎 and
𝟐
∆𝒎 𝟎.𝟓
∆𝒎 = 𝒎𝑻 − 𝒎𝑨 = 𝟒𝟗. 𝟓 − 𝟒𝟗 = 𝟎. 𝟓 ⇒ × 𝟏𝟎𝟎 = × 𝟏𝟎𝟎 = 𝟏. 𝟎𝟏% and
𝒎𝑻 𝟒𝟗.𝟓

∆𝒗 𝟏𝟎
∆𝒗 = |𝒗𝑻 − 𝒗𝑨 | = |𝟏𝟓𝟗𝟎 − 𝟏𝟔𝟎𝟎| = 𝟏𝟎 ⇒ × 𝟏𝟎𝟎 = × 𝟏𝟎𝟎 = 𝟎. 𝟔𝟐𝟗%
𝒗𝑻 𝟏𝟓𝟗𝟎
∆𝑻
and we have to determine 𝑬𝒑 = 𝑬𝒓 × 𝟏𝟎𝟎 = 𝑻
× 𝟏𝟎𝟎 =?

From general formula of error, we have

𝝏𝑻 𝝏𝑻 𝒗𝟐
∆𝑻 = ∆𝒎 + ∆𝒗 = ( ) ∆𝒎 + (𝒎𝒗)∆𝒗
𝝏𝒎 𝝏𝒗 𝟐

𝒗𝟐
∆𝑻 = ( ) ∆𝒎 + (𝒎𝒗)∆𝒗
𝟐
Then the relative error 𝑬𝒓 is given by

𝒗𝟐 𝒗𝟐
∆𝑻 ( 𝟐 ) ∆𝒎 + (𝒎𝒗)∆𝒗 ( ) ∆𝒎 (𝒎𝒗)∆𝒗 ∆𝒎
𝟐 ∆𝒗
𝑬𝒓 = = = + = + 𝟐
𝑻 𝒎𝒗𝟐 𝒎𝒗𝟐 𝒎𝒗𝟐 𝒎 𝒗
𝟐 𝟐 𝟐
Then the Percentage error 𝑬𝒑 is given by
∆𝑻 ∆𝒎 ∆𝒗
𝑬𝒑 = 𝑬𝒓 × 𝟏𝟎𝟎 = × 𝟏𝟎𝟎 = ( × 𝟏𝟎𝟎) + 𝟐 ( × 𝟏𝟎𝟎) = 𝟏. 𝟎𝟏% + 𝟐(𝟎. 𝟔𝟐𝟗%)
𝑻 𝒎 𝒗
∆𝑻
𝑬𝒑 = 𝑬𝒓 × 𝟏𝟎𝟎 = × 𝟏𝟎𝟎 = 𝟐. 𝟐𝟔𝟖%
𝑻

Example # 18:-
The power 𝑷 required to propel a ship of length 𝒍 with a velocity 𝒗 is given by
𝑷 = 𝒌𝒗𝟑 𝒍𝟐 where 𝒌 is constant. Find the Percentage Error in Power if percentage error in
velocity is 𝟑% and in length is 𝟒% .

Solution:-

We have given
∆𝒗 ∆𝒍 ∆𝑷
𝑷 = 𝒌𝒗𝟑 𝒍𝟐 , × 𝟏𝟎𝟎 = 𝟑% , × 𝟏𝟎𝟎 = 𝟒% and × 𝟏𝟎𝟎 =?
𝒗 𝒍 𝑷

∆𝑷
In order to calculate Percentage Error in Power 𝑷, that is × 𝟏𝟎𝟎, first we calculate
𝑷

Absolute Error in 𝑷 as

𝝏𝑷 𝝏𝑷
∆𝑷 = ∆𝒗 + ∆𝒍 = 𝟑𝒌𝒗𝟐 𝒍𝟐 ∆𝒗 + 𝟐𝒌𝒗𝟑 𝒍∆𝒍
𝝏𝒗 𝝏𝒍

Dr. Jamil Book Series 2 Page 19

Advanced Mathematical Techniques Chapter # 1 Error Analysis

For Relative in 𝑷, we have

∆𝑷 𝟑𝒌𝒗𝟐 𝒍𝟐 ∆𝒗 + 𝟐𝒌𝒗𝟑 𝒍∆𝒍 𝟑𝒌𝒗𝟐 𝒍𝟐 𝟐𝒌𝒗𝟑 𝒍 ∆𝒗 ∆𝒍

= = ∆𝒗 + ∆𝒍 = 𝟑 + 𝟐
𝑷 𝒌𝒗𝟑 𝒍𝟐 𝒌𝒗𝟑 𝒍𝟐 𝒌𝒗𝟑 𝒍𝟐 𝒗 𝒍

Then the Percentage error 𝑬𝒑 is given by

∆𝑷 ∆𝒗 ∆𝒍
𝑬𝒑 = × 𝟏𝟎𝟎 = 𝟑 ( × 𝟏𝟎𝟎) + 𝟐 ( × 𝟏𝟎𝟎)
𝑷 𝒗 𝒍

∆𝑷
𝑬𝒑 = × 𝟏𝟎𝟎 = 𝟑(% 𝑬𝒓𝒓𝒐𝒓 𝒊𝒏 𝒗) + 𝟐(% 𝑬𝒓𝒓𝒐𝒓 𝒊𝒏 𝒍)
𝑷

∆𝑷
𝑬𝒑 = × 𝟏𝟎𝟎 = 𝟑(𝟑%) + 𝟐(𝟒%) = 𝟗% + 𝟖% = 𝟏𝟕%
𝑷

Example # 19:-
If 𝐓 = 𝟒𝐱 𝟐 𝐲 𝟑 𝐳 −𝟒 , find the Maximum Percentage Error in 𝐓 when error in 𝐱 = 𝟏 , 𝐲 =
𝟐 , 𝐳 = 𝟑 respectively are equal to 𝟎. 𝟎𝟎𝟏, 𝟎. 𝟎𝟎𝟐, 𝟎. 𝟎𝟎𝟑.
Solution:-
We have given
𝐓 = 𝟒𝐱 𝟐 𝐲 𝟑 𝐳 −𝟒 , 𝐱 = 𝟏 , 𝐲 = 𝟐, 𝐳 = 𝟑 and ∆𝐱 = 𝟎. 𝟎𝟎𝟏, ∆𝐲 = 𝟎. 𝟎𝟎𝟐, ∆𝐳 = 𝟎. 𝟎𝟎𝟑.
∆𝑻
and we have to determine 𝑬𝒑 = 𝑬𝒓 × 𝟏𝟎𝟎 = 𝑻
× 𝟏𝟎𝟎 =?

From general formula of error, we have

𝝏𝑻 𝝏𝑻 𝝏𝑻
∆𝑻 = ∆𝒙 + ∆𝒚 + ∆𝒛
𝝏𝒙 𝝏𝒚 𝝏𝒛

where,

𝝏𝑻 𝟖𝒙𝒚𝟑 𝝏𝑻 𝟏𝟐𝐱 𝟐 𝐲 𝟐 𝝏𝑻 𝟏𝟔𝐱 𝟐 𝐲 𝟑

= 𝟒 , = , =−
𝝏𝒙 𝒛 𝝏𝒚 𝒛𝟒 𝝏𝒛 𝒛𝟓

Then
𝟖𝒙𝒚𝟑 𝟏𝟐𝐱 𝟐 𝐲 𝟐 𝟏𝟔𝐱 𝟐 𝐲 𝟑
∆𝑻 = 𝟒 ∆𝒙 + ∆𝒚 − ∆𝒛
𝒛 𝒛𝟒 𝒛𝟓
In general, the error ∆𝒙, ∆𝒚 and ∆𝒛 may be +𝒗𝒆 or −𝒗𝒆 , and hence, we take the absolute
values of the terms on right side and using
(|𝒙 + 𝒚 + 𝒛| ≤ |𝒙| + |𝒚| + |𝒛| 𝒂𝒏𝒅 |𝒙𝒚𝒛| = |𝒙||𝒚||𝒛|).
This gives

Dr. Jamil Book Series 2 Page 20

Advanced Mathematical Techniques Chapter # 1 Error Analysis

𝟖𝒙𝒚𝟑 𝟏𝟐𝐱 𝟐 𝐲 𝟐 𝟏𝟔𝐱 𝟐 𝐲 𝟑

|∆𝑻| = | 𝟒 ∆𝒙 + ∆𝒚 − ∆𝒛|
𝒛 𝒛𝟒 𝒛𝟓

𝟖𝒙𝒚𝟑 𝟏𝟐𝐱 𝟐 𝐲 𝟐 𝟏𝟔𝐱 𝟐 𝐲 𝟑

∆𝑻 ≤ | 𝟒 ∆𝒙| + | ∆𝒚| + | ∆𝒛|
𝒛 𝒛𝟒 𝒛𝟓

At 𝒙 = 𝟏, 𝒚 = 𝟐, 𝒛 = 𝟑 and ∆𝒙 = 𝟎. 𝟎𝟎𝟏, ∆𝒚 = 𝟎. 𝟎𝟎𝟐, ∆𝒛 = 𝟎. 𝟎𝟎𝟑 we have

(𝟖)(𝟏)(𝟐𝟑 ) (𝟏𝟐)(𝟏𝟐 )(𝟐𝟐 ) (𝟏𝟔)(𝟏𝟐 )(𝟐𝟑 )

∆𝑻 ≈ [ ] (𝟎. 𝟎𝟎𝟏) + [ ] (𝟎. 𝟎𝟎𝟐) [ ] (𝟎. 𝟎𝟎𝟑)
𝟑𝟒 𝟑𝟒 𝟑𝟓

∆𝑻 ≈ (𝟎. 𝟕𝟗𝟎)(𝟎. 𝟎𝟎𝟏) + (𝟎. 𝟓𝟗𝟑)(𝟎. 𝟎𝟎𝟐) + (𝟎. 𝟓𝟐𝟕)(𝟎. 𝟎𝟎𝟑) = 𝟑. 𝟓𝟓𝟕 × 𝟏𝟎−𝟑

and

𝟐 𝟑 −𝟒
𝟒𝒙𝟐 𝒚𝟑 (𝟒)(𝟏𝟐 )(𝟐𝟑 )
𝑻 = 𝟒𝒙 𝒚 𝒛 = = = 𝟎. 𝟑𝟗𝟓 𝒂𝒕 𝒙 = 𝟏 , 𝒚 = 𝟐, 𝒛 = 𝟑.
𝒛𝟒 𝟑𝟒

Then the Relative error 𝑬𝒓 is given by

∆𝑻 𝟑. 𝟓𝟓𝟕 × 𝟏𝟎−𝟑
𝑬𝒓 = = = 𝟗. 𝟎𝟏 × 𝟏𝟎−𝟑
𝑻 𝟎. 𝟑𝟗𝟓

and Percentage error 𝑬𝒑 is given by

∆𝑻
𝑬𝒑 = 𝑬𝒓 × 𝟏𝟎𝟎 = × 𝟏𝟎𝟎 = 𝟗. 𝟎𝟏 × 𝟏𝟎−𝟑 × 𝟏𝟎𝟎 = 𝟎. 𝟗𝟎𝟏%
𝑻

Drill Excercise # 8:-

Suppose 𝑵 is a function of five different measurable quantities 𝒖, 𝒗, 𝒘, 𝒚, 𝒛 and is given by
𝒖𝒑 𝒗 𝒒 𝒘 𝒓
𝑵=
𝒙𝒔 𝒚𝒕
Find an upper limit to the Relative Error in measure of 𝑵.
Note:- The upper limit of 𝒙 is 𝒂 if 𝒙 < 𝒂 or 𝒙 ≤ 𝒂 .
Answer:-
∆𝒖 ∆𝒗 ∆𝒘 ∆𝒙 ∆𝒚
The upper limit of the relative error in 𝑵 is: 𝒑 | | + 𝒒 | | + 𝒓 | | + 𝒔 | | + 𝒕 | |, that is
𝒖 𝒗 𝒘 𝒙 𝒚

∆𝑵 ∆𝒖 ∆𝒗 ∆𝒘 ∆𝒙 ∆𝒚
≤ 𝒑| | + 𝒒| | + 𝒓| | + 𝒔| | + 𝒕| |
𝑵 𝒖 𝒗 𝒘 𝒙 𝒚

Dr. Jamil Book Series 2 Page 21

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Example # 20:-
Find the possible Percentage Error in computing the parallel resistance 𝐫 of these resistance
𝟏 𝟏 𝟏 𝟏
𝐫𝟏 , 𝐫𝟐 , 𝐫𝟑 from the formula = + + . If 𝐫𝟏 , 𝐫𝟐 , 𝐫𝟑 are each in the error by plus 𝟒. 𝟓 %.
𝒓 𝒓𝟏 𝒓𝟐 𝒓𝟑

Solution:-
We have given
𝟏 𝟏 𝟏 𝟏 ∆𝒓𝟏 ∆𝒓𝟐 ∆𝒓𝟑
= + + , × 𝟏𝟎𝟎 = × 𝟏𝟎𝟎 = × 𝟏𝟎𝟎 = 𝟒. 𝟓%
𝒓 𝒓𝟏 𝒓𝟐 𝒓𝟑 𝒓𝟏 𝒓𝟐 𝒓𝟑
∆𝒓
and we have to determine 𝑬𝒑 = 𝑬𝒓 × 𝟏𝟎𝟎 = 𝒓
× 𝟏𝟎𝟎 =?

From general formula of error, we have

𝝏𝒓 𝝏𝒓 𝝏𝒓
∆𝒓 = ∆𝒓𝟏 + ∆𝒓𝟐 + ∆𝒓
𝝏𝒓𝟏 𝝏𝒓𝟐 𝝏𝒓𝟑 𝟑

Since
𝝏 𝟏 𝝏 𝟏 𝟏 𝟏 𝟏 𝝏𝒓 𝟏 𝝏𝒓 𝒓𝟐
( )= ( + + )⇒− 𝟐 =− 𝟐⇒ =
𝝏𝒓𝟏 𝒓 𝝏𝒓𝟏 𝒓𝟏 𝒓𝟐 𝒓𝟑 𝒓 𝝏𝒓𝟏 𝒓𝟏 𝝏𝒓𝟏 𝒓𝟏 𝟐

Similarly, we get
𝝏𝒓 𝒓𝟐 𝝏𝒓 𝒓𝟐
⇒ = 𝒂𝒏𝒅 =
𝝏𝒓𝟐 𝒓𝟐 𝟐 𝝏𝒓𝟑 𝒓𝟑 𝟐
Then

𝒓𝟐 𝒓𝟐 𝒓𝟐 𝟏 ∆𝒓 𝟏 𝟏 𝟏
∆𝒓 = 𝟐 ∆𝒓𝟏 + 𝟐 ∆𝒓𝟐 + 𝟐 ∆𝒓𝟑 ⇒ = 𝟐 ∆𝒓𝟏 + 𝟐 ∆𝒓𝟐 + 𝟐 ∆𝒓𝟑
𝒓𝟏 𝒓𝟐 𝒓𝟐 𝒓 𝒓 𝒓𝟏 𝒓𝟐 𝒓𝟑

𝟏 ∆𝒓 = 𝟏 𝟏 ∆𝒓 + 𝟏 𝟏 ∆𝒓 + 𝟏 𝟏 ∆𝒓
𝒓 𝒓 𝒓𝟏 𝒓𝟏 𝟏 𝒓𝟐 𝒓𝟐 𝟐 𝒓𝟑 𝒓𝟑 𝟑

𝟏 ∆𝒓 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏
⇒ ( × 𝟏𝟎𝟎) = ( ∆𝒓𝟏 × 𝟏𝟎𝟎) + ( ∆𝒓𝟐 × 𝟏𝟎𝟎) + ( ∆𝒓𝟑 × 𝟏𝟎𝟎)
𝒓 𝒓 𝒓𝟏 𝒓𝟏 𝒓𝟐 𝒓𝟐 𝒓𝟑 𝒓𝟑

𝟏 ∆𝒓 𝟏 𝟏 𝟏
⇒ ( × 𝟏𝟎𝟎) = (𝟒. 𝟓%) + (𝟒. 𝟓%) + (𝟒. 𝟓%)
𝒓 𝒓 𝒓𝟏 𝒓𝟐 𝒓𝟑

𝟏 ∆𝒓 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏 𝟏
⇒ ( × 𝟏𝟎𝟎) = ( + + ) (𝟒. 𝟓%) ∴ = + +
𝒓 𝒓 𝒓𝟏 𝒓𝟐 𝒓𝟑 𝒓 𝒓𝟏 𝒓𝟐 𝒓𝟑

𝟏 ∆𝒓 𝟏
⇒ ( × 𝟏𝟎𝟎) = (𝟒. 𝟓%)
𝒓 𝒓 𝒓

∆𝒓
⇒ × 𝟏𝟎𝟎 = 𝟒. 𝟓%
𝒓

Dr. Jamil Book Series 2 Page 22

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Example # 21:-
The voltage 𝑽 across a resistor is measured with error 𝑯, and the resistance 𝑹 is measured
𝑽𝟐
with an error 𝑲. Show that the error in calculating the power 𝑾 = generated in the
𝑹
𝑽𝟐 𝑯 𝑲
resistor is (𝟐 − ). If 𝑽 can be measured to an accuracy of 𝟎. 𝟓% and 𝑹 to an accuracy
𝑹 𝑽 𝑹
of 𝟏%. What is the approximately possible Percentage Error in 𝑾.

Solution:-
We have given
𝑽𝟐 𝑽𝟐 𝑯 𝑲
𝑾= , ∆𝑽 = 𝑯, ∆𝑹 = 𝐊, we have to show that: ∆𝑾 = (𝟐 − ) and we have to
𝑹 𝑹 𝑽 𝑹
∆𝑾
determine 𝑬𝒑 = 𝑬𝒓 × 𝟏𝟎𝟎 = 𝑾
× 𝟏𝟎𝟎 =? when ∆𝑽 = 𝟎. 𝟓%, ∆𝑹 = 𝟏%.

From general error formula, we have

𝝏𝑾 𝝏𝑾 𝑽 𝑽𝟐
∆𝑾 = ∆𝑽 + ∆𝑹 = 𝟐 ∆𝑽 − 𝟐 ∆𝑹
𝝏𝑽 𝝏𝑹 𝑹 𝑹

Since ∆𝑽 = 𝑯, ∆𝑹 = 𝐊, we have

𝑽𝟐
𝑽𝑯 𝑽𝟐 𝑲 𝑽𝟐 𝟐𝑯 𝑲 ∴𝑾= , ∆𝑽 = 𝑯, ∆𝑹 = 𝐊
∆𝑾 = 𝟐 − 𝟐 = ( − ) 𝑹
𝑹 𝑹 𝑹 𝑽 𝑹

This prove the result. The above expression can be written as

𝟐𝑯 𝑲 ∆𝑾 𝟐𝑯 𝑲 𝟐 𝟏
∆𝑾 = 𝑾 ( − )⇒ = − = ∆𝑽 − ∆𝑹
𝑽 𝑹 𝑾 𝑽 𝑹 𝑽 𝑹

∆𝑾 ∆𝑽 ∆𝑹
⇒ × 𝟏𝟎𝟎 = 𝟐 × 𝟏𝟎𝟎 − × 𝟏𝟎𝟎
𝑾 𝑽 𝑹

∆𝑾
⇒ × 𝟏𝟎𝟎 = 𝟐(𝟎. 𝟓%) − 𝟏% = 𝟎%
𝑾

Dr. Jamil Book Series 2 Page 23

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Drill Excercise # 9:-

𝑬𝟐
The Power dissipated in a resistor is given by 𝑷 = . Find the Percentage Error in 𝑷 when
𝑹

percentage error in 𝑬 is 𝟑% and 𝑹 is 𝟐%.

Answer:-
𝜟𝑷
× 𝟏𝟎𝟎 = 𝟖%.
𝑷

Drill Excercise # 10:-

The Deflection 𝐃 at the centre of a road of length 𝒍 and the diameter 𝒅 of supported at its
ends and loaded at the centre with a weight 𝐰 varies as 𝒘𝒍𝟑 𝒅−𝟒 (𝑫 =
𝒌𝒘𝒍𝟑 𝒅−𝟒 , 𝐰𝐡𝐞𝐫𝐞 𝐤 𝐢𝐬 𝐩𝐫𝐨𝐩𝐨𝐫𝐭𝐨𝐧𝐚𝐥𝐢𝐭𝐲 𝐜𝐨𝐧𝐬𝐭𝐚𝐧𝐭. ).What is the Percentage Error in the
deflection corresponding to the percentage error in 𝐰, 𝐥 and 𝐝 of 𝟑% , 𝟐% and 𝟏%
respectively.
Answer:-
∆𝐃
× 𝟏𝟎𝟎 = 𝟓 %
𝐃
Drill Excercise # 11:-
The diameter and altitude of a can in the shape of a right circular cylinder are measured as
𝟒𝟎𝐜𝐦 and 𝟔𝟒𝐜𝐦 respectively. The possible error in each measurement is ±𝟓%. Find
approximately the Maximum Percentage Error in the computed value for the Volume and
the Lateral Surface. (Note:- 𝑽 = 𝝅𝒓𝟐 𝒉 , 𝑺 = 𝟐𝝅𝒓𝒍)
Answer:-
𝜟𝑽 𝜟𝒔
× 𝟏𝟎𝟎 = ±𝟏𝟓% , × 𝟏𝟎𝟎 = ±𝟏𝟎%.
𝑽 𝒔

Dr. Jamil Book Series 2 Page 24

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Numerical stability & ill-condition

Sensitivity of the solution w.r.t. errors in the data
1): Well-conditioned or stability:-
A problem is well- conditioned if small errors in the data produce small errors in the
solution.

2):Ill-conditioned or instability:-
A problem is well- conditioned if small errors in the data produce large errors in the
solution.

The condition:-
The term condition is often used to describe the sensitivity of a function 𝒇(𝒙) to small
changes in the argument 𝒙 . It is measured by the maximum relative change in 𝒇(𝒙) due to
unit relative change/error in the argument.
Let ∆𝒇(𝒙) be the increment/error in the 𝒇(𝒙) due to small increment ∆𝒙 in 𝒙. The
condition of 𝒇(𝒙) at 𝒙, may be expressed as:

𝑹𝒆𝒍𝒂𝒕𝒊𝒗𝒆 𝒆𝒓𝒓𝒐𝒓 𝒊𝒏 𝒇(𝒙)

𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝐦𝐚𝐱 | |
𝑹𝒆𝒍𝒂𝒕𝒊𝒗𝒆 𝒆𝒓𝒓𝒐𝒓 𝒊𝒏 𝒙

∆𝒇(𝒙) /𝒇(𝒙)
⇒ 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝐦𝐚𝐱 | |
∆𝒙/𝒙

Dr. Jamil Book Series 2 Page 25

Advanced Mathematical Techniques Chapter # 1 Error Analysis

∆𝒇(𝒙) 𝒙
⇒ 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝐦𝐚𝐱 | |
∆𝒙 𝒇(𝒙)

Since

∆𝒇(𝒙) 𝒅𝒇(𝒙)
≈ = 𝒇′(𝒙)
∆𝒙 𝒅𝒙

𝒙𝒇′ (𝒙)
⇒ 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 ≈ 𝐦𝐚𝐱 | |
𝒇(𝒙)

A large value of the condition indicates that 𝒇(𝒙) is highly sensitive near 𝒙, and it is termed
ill-condition.
Note that a function may be ill-condition near a particular point but well-conditioned at
other points.
Example # 22:-
𝝅 𝝅
The function 𝒇(𝒙) = 𝒕𝒂𝒏𝒙, is ill-conditioned near 𝒙 = and well-conditioned at 𝒙 = .
𝟐 𝟒

At 𝒙 = 𝝅/𝟐:-
Since 𝒇′ (𝒙) = 𝒔𝒆𝒄𝟐 𝒙 then

𝒙𝒇′ (𝒙) 𝒙𝒔𝒆𝒄𝟐 𝒙 𝟐𝒙𝒄𝒐𝒔𝒙

𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝐦𝐚𝐱 | | = 𝐦𝐚𝐱 | | = 𝐦𝐚𝐱 | |
𝒇(𝒙) 𝒕𝒂𝒏𝒙 𝟐 𝐜𝐨𝐬 𝟐 𝒙 𝒔𝒊𝒏𝒙

𝟐𝒙 𝝅 𝝅 𝝅
𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝐦𝐚𝐱 | | = 𝒎𝒂𝒙 | | = 𝒎𝒂𝒙 | | ⟶ ∞ 𝒂𝒔 𝒙 ⟶
𝐬𝐢𝐧(𝟐𝒙) 𝒔𝒊𝒏(𝝅) 𝟎 𝟐

𝝅 𝝅
Which grows unbounded as 𝒙 ⟶ . Hence 𝒇(𝒙) = 𝒕𝒂𝒏𝒙 is ill-conditioned at = .
𝟐 𝟐

At 𝒙 = 𝝅/𝟒:-

We have

𝟐𝒙 𝝅/𝟐 𝝅 𝝅 𝝅
𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝒎𝒂𝒙 | | = 𝒎𝒂𝒙 | | = 𝒎𝒂𝒙 | | = 𝒂𝒕 𝒙 = .
𝐬𝐢𝐧(𝟐𝒙) 𝒔𝒊𝒏(𝝅/𝟐) 𝟐 𝟐 𝟒

𝝅 𝝅
Which is finite as 𝒙 ⟶ . Hence 𝒇(𝒙) = 𝒕𝒂𝒏𝒙 is well-conditioned at 𝒙 = .
𝟒 𝟒

Example # 23:-
Check whether 𝒇(𝒙) = 𝒔𝒊𝒏𝒙 and 𝒇(𝒙) = 𝒄𝒐𝒔𝒙 are well-conditioned or ill-conditioned at 𝒙 =
𝝅
𝟎 and 𝒙 = .
𝟐

Solution:-
1): For 𝒇(𝒙) = 𝒔𝒊𝒏𝒙:-
We have 𝒇′ (𝒙) = 𝒄𝒐𝒔𝒙, then

Dr. Jamil Book Series 2 Page 26

Advanced Mathematical Techniques Chapter # 1 Error Analysis

𝒙𝒇′ (𝒙) 𝒙𝒄𝒐𝒔𝒙

𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝐦𝐚𝐱 | | = 𝐦𝐚𝐱 | |
𝒇(𝒙) 𝒔𝒊𝒏𝒙
𝒙
𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝐦𝐚𝐱 | |
𝒕𝒂𝒏𝒙
At 𝒙 = 𝟎:-
At = 𝟎 , 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝟎/𝟎 form therefore apply L’hopital theorem, we get
𝟏
𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝐦𝐚𝐱 | 𝟐
| = 𝐦𝐚𝐱 | 𝐜𝐨𝐬 𝟐 𝒙|
𝒔𝒆𝒄 𝒙

𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝐦𝐚𝐱 | 𝐜𝐨𝐬 𝟐 𝒙| = 𝒎𝒂𝒙|𝟏| → 𝟏 𝒘𝒉𝒆𝒏 𝒙 → 𝟎.

Hence 𝒇(𝒙) = 𝒔𝒊𝒏𝒙 is well-conditioned at 𝒙 = 𝟎.

At 𝒙 = 𝝅/𝟐:-
𝒙 𝝅/𝟐 𝝅
𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝐦𝐚𝐱 | | = 𝒎𝒂𝒙 | | ⟶ 𝟎 𝒂𝒔 𝒙 ⟶
𝒕𝒂𝒏𝒙 ∞ 𝟐

𝝅
Hence 𝒇(𝒙) = 𝒔𝒊𝒏𝒙 is well-conditioned at 𝒙 = .
𝟐

2): For 𝒇(𝒙) = 𝒄𝒐𝒔𝒙:-

We have 𝒇′ (𝒙) = −𝐬𝐢𝐧 𝒙, then
𝒙𝒇′ (𝒙)
𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝐦𝐚𝐱 | |
𝒇(𝒙)
𝒙(−𝒔𝒊𝒏𝒙)
⇒ 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝐦𝐚𝐱 | | = 𝐦𝐚𝐱|−𝒙𝒕𝒂𝒏𝒙|
𝒄𝒐𝒔𝒙

⇒ 𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝐦𝐚𝐱|𝒙𝒕𝒂𝒏𝒙|

At 𝒙 = 𝟎:-
𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝐦𝐚𝐱|𝒙𝒕𝒂𝒏𝒙| ⟶ 𝟎 𝒂𝒔 𝒙 ⟶ 𝟎

Hence 𝒇(𝒙) = 𝒄𝒐𝒔𝒙 is well-conditioned at 𝒙 = 𝟎.

At 𝒙 = 𝝅/𝟐:-
𝝅
𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 = 𝐦𝐚𝐱|𝒙𝒕𝒂𝒏𝒙| ⟶ ∞ 𝒂𝒔 𝒙 ⟶ .
𝟐
𝝅
Hence 𝒇(𝒙) = 𝒄𝒐𝒔𝒙 is ill-conditioned at 𝒙 = .
𝟐

Dr. Jamil Book Series 2 Page 27

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Drill Excercise # 12:-

Check whether the function 𝒇(𝒙) = 𝒄𝒐𝒕𝒙 and 𝒇(𝒙) = 𝒔𝒊𝒏𝟐𝒙 is well-conditioned or ill-
conditioned at 𝒙 = 𝟎 and 𝒙 = 𝝅⁄𝟐.

Answer:-

Numerical Algorithms & Flow charts

1): Algorithms:-

An algorithm for solving a problem is a Finite Sequence of Simple Instruction which, when
followed yields the solution of the problem.

2): Flow charts:-

Graphical or Diagrammatic Representation of an Algorithm is a flow chart, which its
representation in a programming language is a program.

3): Flowchart Keys:-

1. Start/End

2. Process or
task

3.
Decision
s

Input/
4.
Outpu
t

5. Sequence

Dr. Jamil Book Series 2 Page 28

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Example # 24:-
Given a real number 𝒙. Write the Algorithm and make Flow chart of the absolute value
𝑬 = |𝒙|.
Solution:-
1): Algorithm: Absolute Value Computation:-
Input :𝒙
Output : |𝒙|
Method :
1. Test 𝒙 ≥ 𝟎 ? if true, then 𝑬 = 𝒙, terminate.
Otherwise
2. 𝑬 = −𝒙

2):“Flow chart:- Absolute value computation”

Start

Input 𝒙

NO YES
𝒙 ≥ 𝟎?

𝑬 = −𝒙 𝑬=𝒙

Print 𝑬

End

Dr. Jamil Book Series 2 Page 29

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Example # 25:-
“Input Output flow chart”

Start

Input
Temp

NO YES
Temp≤ 𝟑𝟐

Print “below Print “above

freezing” freezing”
”

End

Dr. Jamil Book Series 2 Page 30

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Example # 26:-
“Order Processing in E-Commerce”

Start
End

Receive order
via e-mail

Copy and paste e-mail

data into database.

Shipping YES
involved

Print invoice
NO and UPS lable.

Send e-mail to confirm

shipping.

Assemble Package and ship

End

Dr. Jamil Book Series 2 Page 31

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Drill Excercise # 13:-

Draw a flow chart of software development.
Answer:-
Flow chart of software development

Start
End

Design

Coding

Testing

YES Error

NO
Design
Error NO
YES

End

Dr. Jamil Book Series 2 Page 32

Advanced Mathematical Techniques Chapter # 1 Error Analysis

EXCERCISE -1
Question # 1:-
Round off the following numbers correct up to 4-significant figures:
a): 2.56235 b): 0.34026 c): 4.50089 d): 0.000243468 e): 2456.782
f): 1.3456 g): 2.5555 h): 8.0008
Answer:-
a): 2.562 b): 0.3403 c): 4.501 d): 0.0002435 e): 2457
f): 1.346 g): 2.226 h): 8.001

Question # 2:-
Round-off the following numbers correct up to 3-decimal places
a): 2.46289 b): 0.46999 c): 0.0035869 d): 0.0015
e): 1.46294 f): 1.35008
Answer:-
a): 2.463 b): 0.470 c): 0.004 d): 0.002
e): 1.463 f): 1.350

Question # 3:-
The following numbers are correct up to last digit. Find the sum
Assignment
a): 1.3526, 2.00462, 1.532, 28.201, 31.0012 SIR : 4

SIR : 4
b): 0.532, 7.46571, 1.501, 3.62102
SIR :2
c): 5.2146, 20.12, 11.2356, 1.8948
Answer:-
a): 64.091 b): 13.120 c): 38.46

Question # 4:-
If 𝑽𝑻 , 𝑽𝑨 , 𝑬𝒂 , 𝑬𝒓 𝐚𝐧𝐝 𝑬𝒑 are true value, approximate value, absolute error, relative error
and percentage error respectively:
a): Given 𝑽𝑻 = 𝟐. 𝟓𝟒𝟔𝟐𝟖𝟐, 𝑽𝑨 = 𝟐. 𝟓𝟒𝟔𝟑. Find 𝑬𝒂 , 𝑬𝒓 , 𝑬𝒑 .

b): Given 𝑬𝒂 = 𝟎. 𝟐 × 𝟏𝟎−𝟑 , 𝑬𝒓 = 𝟎. 𝟑𝟐 × 𝟏𝟎−𝟓 . Find 𝑽𝑻

c): Given 𝑽𝑻 = 𝟐. 𝟕𝟑𝟒𝟓𝟔, 𝑬𝒓 = 𝟎. 𝟐 × 𝟏𝟎−𝟐 . Find 𝑬𝒂 , 𝑬𝒑

Answer:-
a): 𝑬𝒂 = 𝟏. 𝟖 × 𝟏𝟎−𝟓 , 𝑬𝒓 = 𝟕. 𝟎𝟕 × 𝟏𝟎−𝟔 , 𝑬𝒑 = 𝟕. 𝟎𝟕 × 𝟏𝟎−𝟒 %
b): 𝟔𝟐. 𝟓
c): 𝑬𝒂 = 𝟎. 𝟎𝟎𝟓𝟒𝟕, 𝑬𝒑 = 𝟎. 𝟐%

Dr. Jamil Book Series 2 Page 33

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Question # 5:-
Find the Relative Error in computation of 𝒙 + 𝒚 for 𝒙 = 𝟏𝟏. 𝟕𝟓 𝐚𝐧𝐝 𝒚 = 𝟕. 𝟐𝟑 having
absolute errors ∆𝒙 = 𝟎. 𝟎𝟎𝟐 𝐚𝐧𝐝 ∆𝒚 = 𝟎. 𝟎𝟎𝟓 respectively.
Answer:-
𝑬𝒓 = 𝟎. 𝟎𝟎𝟎𝟑𝟒

Question # 6:-
If 5/6 be represented approximately by 0.8333, find a): Absolute error b): Relative
error c): Percentage error and also verify theorem-1 and theorem-2.
Answer:-
𝑬𝒓 = 𝟎.00004, 𝑬𝒑 = 𝟎. 𝟎𝟎𝟒 %

Question # 7:-
Find the Relative Percentage Error in approximate representation of 4/3 by 1.33 and also
verify theorem-1 and theorem-2.
Answer:-
𝑬𝒑 = 𝟎. 𝟐𝟓%

Question # 8:- Assignment

Let a number 345.26132 be rounded to 345.261 correct to 3-decimal places, then find
1): Absolute error 2): Relative error 3): Percentage errors and also verify theorem-1
and theorem-2.
Answer:-
𝑬𝒂 = 𝟎. 𝟎𝟎𝟎𝟑𝟐, 𝑬𝒓 = 𝟗. 𝟐𝟔𝟖 × 𝟏𝟎−𝟕 , 𝟗. 𝟐𝟔𝟖 × 𝟏𝟎−𝟓

Question # 9:-
Write down the approximate representation of 2/3 correct to 4 significant figures and then
find: 1): Absolute error 2): Relative error 3): Percentage error and verify theorem-1
and theorem-2.
Answer:-
𝑬𝒂 = 𝟎. 𝟎𝟎𝟎𝟎𝟑𝟑, 𝑬𝒓 = 𝟎. 𝟎𝟎𝟎𝟎𝟓, 𝑬𝒑 = 𝟎. 𝟎𝟎𝟓%
Question # 10:-
Find the number of Significant Figures in:
a): 𝑽𝑨 = 𝟏𝟏. 𝟐𝟒𝟔𝟏 given its absolute error as 𝟎. 𝟐𝟓 × 𝟏𝟎−𝟐
b): 𝑽𝑻 = 𝟏. 𝟓𝟗𝟐𝟑 given its relative error as 𝟎. 𝟏 × 𝟏𝟎−𝟑
c): 𝑽𝑨 = 𝟎. 𝟑𝟗𝟒𝟏 given its absolute error as 𝟎. 𝟐𝟓 × 𝟏𝟎−𝟐
Answer:-
a): 4-significant figures b): 4-significant figures c): 2-significant figures

Dr. Jamil Book Series 2 Page 34

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Question # 11:-
If ∆𝒙 = 𝟎. 𝟎𝟎𝟓 𝐚𝐧𝐝 ∆𝒚 = 𝟎. 𝟎𝟎𝟏 be the absolute error in 𝒙 = 𝟐. 𝟏𝟏 𝒂𝒏𝒅 𝒚 = 𝟒. 𝟏𝟓, find the
Relative Error in computation of 𝒙 + 𝒚.
Answer:-
𝑬𝒓 = 𝟎. 𝟎𝟎𝟎𝟗𝟓𝟖 ≈ 𝟎. 𝟎𝟎𝟏

Question # 12:-
Determine the number of correct Significant Digits in the numbers 𝒙, given its absolute
error 𝑬𝒂 :
a): 𝒙 = 𝟎. 𝟑𝟔𝟖𝟏, 𝑬𝒂 = 𝟎. 𝟐𝟑 × 𝟏𝟎−𝟐
b): 𝒙 = 𝟖𝟒𝟏. 𝟐𝟓𝟔, 𝑬𝒂 = 𝟎. 𝟕 × 𝟏𝟎−𝟑
c): 𝒙 = −𝟑𝟑. 𝟕𝟖𝟑, 𝑬𝒂 = 𝟎. 𝟑 × 𝟏𝟎−𝟐
Answer:-
a): 2 b): ??? c): 4

Question # 13:- Assignment

SIGNIFICANT DIGIT
Determine the number of correct digits in the number 𝒙 given its relative error 𝑬𝒓 :
a): 𝒙 = 𝟎. 𝟒𝟕𝟖𝟓, 𝑬𝒓 = 𝟎. 𝟐 × 𝟏𝟎−𝟐
b): 𝒙 = 𝟑𝟖𝟔. 𝟒, 𝑬𝒓 = 𝟎. 𝟑 × 𝟏𝟎−𝟐
c): 𝒙 = 𝟖𝟔. 𝟑𝟒, 𝑬𝒓 = 𝟎. 𝟏 × 𝟏𝟎−𝟑
Answer:-
a): 3 b): 3 c): 4

Question # 14:-
Find the number of Significant Figures in 𝑽𝑨 = 𝟏. 𝟖𝟗𝟐𝟏, given its relative error as 𝟎. 𝟏 ×
𝟏𝟎−𝟐 .
Answer:-
3-signifcant figures

Question # 15:-
If 𝒚 = 𝟒𝒙𝟔 − 𝟓𝒙, find the Percentage Error in 𝒚 at 𝒙 = 𝟏 if the error in 𝒙 is 𝟎. 𝟎𝟒
Answer:-
𝑬𝒑 = 𝟕𝟔%
Question # 16:-
If 𝒇(𝒙) = 𝟒𝒄𝒐𝒔𝒙 − 𝟔𝒙, find the Relative Percentage Error in 𝒇(𝒙) for 𝒙 = 𝟎 if error in 𝒙 =
𝟎. 𝟎𝟎𝟓
Answer:-
𝑬𝒑 = 𝟎. 𝟕𝟓%

Dr. Jamil Book Series 2 Page 35

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Question # 17:-
𝟓𝒙𝒚𝟐
Find the Percentage Relative Error in 𝒖 = , where ∆𝒙 = ∆𝒚 = ∆𝒛 = 𝟎. 𝟎𝟎𝟏 and 𝒙 =
𝒛𝟑
𝒚 = 𝒛 = 𝟏.
Answer:-
𝑬𝒑 = 𝟎. 𝟔%

Question # 18:- Assignment

𝒍
The period 𝑻 of a simple pendulum is 𝑻 = 𝟐𝝅√ . Find the Maximum Percentage Error in
𝒈
𝑻 due to possible error up to 𝟏% in 𝒍 and 𝟐. 𝟓% in 𝒈.
Answer:-
∆𝑻 𝟏
𝑬𝒑 = × 𝟏𝟎𝟎 = (𝟏 ± 𝟐. 𝟓)(Note: +ve For Maximum and –ve For Minimum)
𝑻 𝟐

Question # 19:-
𝒎𝒗𝟐
If the kinetic energy 𝑻 = , find the Absolute Error and Percentage Error in 𝑻 as 𝒎 have
𝟐
true value 49.5 and approximate value 49 and 𝒗 have true value 𝟏𝟓𝟗𝟎 and approximate
value 𝟏𝟔𝟎𝟎.
Answer:-
∆𝑻
𝑬𝒑 = 𝑬𝒓 × 𝟏𝟎𝟎 = × 𝟏𝟎𝟎 = 𝟐. 𝟐𝟔𝟖%
𝑻
Question # 20:-
The power 𝑷 required to propel a ship of length 𝒍 with a velocity 𝒗 is given by
𝑷 = 𝒌𝒗𝟑 𝒕𝟐 , where 𝒌 is Proportionality constant. Find the Percentage Error in Power if
percentage error in velocity is 𝟑% and in length is 𝟒% .

Answer:-
Percentage increase in 𝑷 = 𝑬𝒑 = 𝟏𝟕%

Question # 21:-
Suppose 𝑵 is a function of five different measurable quantities 𝒖, 𝒗, 𝒘, 𝒚, 𝒛 and is given by
𝒖𝒑 𝒗 𝒒 𝒘 𝒓
𝑵=
𝒙𝒔 𝒚𝒕
Find an upper limit to the Relative Error in measure of 𝑵.
Note:- The upper limit of 𝒙 is 𝒂 if 𝒙 < 𝒂 or 𝒙 ≤ 𝒂 .
Answer:-
∆𝒖 ∆𝒗 ∆𝒘 ∆𝒙 ∆𝒚
The upper limit of the relative error in 𝑵 is: 𝒑 | | + 𝒒 | | + 𝒓 | | + 𝒔 | | + 𝒕 | |, that is
𝒖 𝒗 𝒘 𝒙 𝒚

∆𝑵 ∆𝒖 ∆𝒗 ∆𝒘 ∆𝒙 ∆𝒚
≤ 𝒑| | + 𝒒| | + 𝒓| |+ 𝒔| | + 𝒕| |
𝑵 𝒖 𝒗 𝒘 𝒙 𝒚

Dr. Jamil Book Series 2 Page 36

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Question # 22:-
Assume that an electrical function 𝑬 is expressible in the four different measurable
quantities , 𝒃, 𝒄, 𝒅 . The Mathematical modelling leads to the following expression:
𝑬 = 𝒂𝒌 𝒃𝒍 𝒄−𝒎 𝒅−𝒏
Determine the upper bond of the Percentage Error in 𝑬.
Answer:-
The upper bond of the Percentage Error in 𝑬 is:

% 𝑬𝒓𝒓𝒐𝒓 𝒊𝒏 𝑬 ≤ 𝒌 % 𝑬𝒓𝒓𝒐𝒓 𝒊𝒏 𝒂 + 𝒍 % 𝑬𝒓𝒓𝒐𝒓 𝒊𝒏 𝒃 + 𝒎 % 𝑬𝒓𝒓𝒐𝒓 𝒊𝒏 𝒄 + 𝒏 % 𝑬𝒓𝒓𝒐𝒓 𝒊𝒏 𝒅

Question # 23:-
Assignment
𝒙
The Exponential function 𝒇(𝒙) = 𝒆 can be expanded as

𝒙
𝒙𝟐 𝒙𝒏−𝟏 𝒙𝒏
𝒆 = 𝟏 + 𝒙 + + ⋯+ + +⋯
𝟐! (𝒏 − 𝟏)! 𝒏!
Find 𝒏, the number of terms, such that their sum yields the value of 𝒆𝒙 correct up to 9
decimal places at 𝒙 = 𝟏.
Answer:-
We need to take 𝒏 =? ? ? terms of the exponential series in order that its sum is correct to 9
decimal places.

Question # 24:-
The discharge 𝐐 over a notch for head 𝑯 is calculated by the formula 𝐐 = 𝐊𝑯𝟓/𝟐 , where
𝐊 is a given Proportionality constant. If the head is 𝟕𝟓 𝐜𝐦 and an error of
𝟎. 𝟏𝟓 𝐜𝐦 possible in its measurement, estimate the Percentage Error in computing the
discharge.

Answer:-
∆𝑸 𝟓 𝑲 (𝟕𝟓)𝟑/𝟐 (𝟎.𝟏𝟓) 𝟏
× 𝟏𝟎𝟎 = × × 𝟏𝟎𝟎 = = 𝟎. 𝟎𝟓%
𝑸 𝟐 𝑲 (𝟕𝟓)𝟓/𝟐 𝟐

Question # 25:-
𝒒𝟏 𝒒𝟐
Coulomb's law gives the magnitude of the force between point charges. It is 𝐅 = 𝚱 .
𝒓𝟐

where 𝒒𝟏 and 𝒒𝟏 are two point charges separated by a distance 𝒓. Find the (1): Percentage
Error (2): Maximum Percentage Error in the Force 𝐅, if the percentage errors in 𝒒𝟏 , 𝒒𝟏
and 𝒓 are 𝟐%, 𝟐% and 𝟑% respectively.

Answer:-

(1): Percentage Error Force 𝐅 = 𝟐% (2): Maximum Percentage Error Force 𝐅 = 𝟏𝟎%

Dr. Jamil Book Series 2 Page 37

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Question # 26:-
The very famous gravitational formula says that: the magnitude of the attractive force 𝐅 is
equal to 𝐆 (the gravitational constant, a number the size of which depends on the system of
units used and which is a universal constant) multiplied by the product of the masses
𝒎𝟏 𝒎𝟐
(𝒎𝟏 and 𝒎𝟐 ) and divided by the square of the distance 𝒓 that is 𝐅 = 𝐆 . Find the (1):
𝒓𝟐

Percentage Error (2): Maximum Percentage Error in the Force 𝐅, if the percentage errors
in 𝒎𝟏 , 𝒎𝟏 and 𝒓 are 𝟏%, 𝟐% and 𝟑% respectively.

Answer:-

(1): Percentage Error Force 𝐅 = 𝟑% (2): Maximum Percentage Error Force 𝐅 = 𝟗%

Question # 27:-
𝟏
The Resonant Frequency of an RLC series circuit is 𝐟 = 𝐆 , where 𝑳 is the inductance
𝟐𝝅√𝑳𝑪

and 𝑪 represent the capacitance of circuit. This is also the natural frequency at which the
circuit would oscillate if not driven by the voltage source. Find the Maximum Percentage
Error in the Force 𝐟, if 𝑳 = 𝟐𝟎𝒎𝑯 and 𝑪 = 𝟓𝝁𝑭 and absolute error in 𝑳 and 𝑪
respectively are 𝟏𝒎𝑯 and 𝟎. 𝟐𝟓𝝁𝑭 . At the end explain and conclude your results.

Answer:-
∆𝒇
× 𝟏𝟎𝟎 =? ? ? %
𝒇
Question # 28:- Assignment

If 𝒓 = 𝟑𝒉(𝒉𝟐 − 𝟐). Find the Percentage Error in 𝐫 at 𝐡 = 𝟏, if the percentage error in 𝐡 is 𝟓.

Answer:-
∆𝒓
× 𝟏𝟎𝟎 = 𝟐𝟓 % 5%
𝒓
Question # 29:-
The error in the measurement of the Area of the circle is not allowed to exceed 𝟎. 𝟏%. How
accurately should the Diameter be measured?
Answer:-
∆𝒅
× 𝟏𝟎𝟎 = 𝟎. 𝟎𝟓
𝒅

Question # 30:-
If 𝐓 = 𝟒𝐱 𝟐 𝐲 𝟑 𝐳 −𝟒 , find the Maximum Percentage Error in 𝐓 when error in 𝐱 = 𝟏 , 𝐲 =
𝟐 , 𝐳 = 𝟑 respectively are equal to 𝟎. 𝟎𝟎𝟎𝟏, 𝟎. 𝟎𝟎𝟐, 𝟎. 𝟎𝟎𝟑.
Answer:-
∆𝑻
× 𝟏𝟎𝟎 = 𝟎. 𝟗𝟎𝟏%
𝑻

Dr. Jamil Book Series 2 Page 38

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Question # 31:-
If 𝑽 = 𝟏𝟎𝒙𝟑 𝒚𝟐 𝒛𝟐 and error in 𝐱, 𝐲, 𝐳 are respectively, 𝟎. 𝟎𝟎𝟑, 𝟎. 𝟎𝟏 , 𝟎. 𝟎𝟐 at 𝐱 = 𝟑, 𝐲 = 𝟏, 𝐳 =
𝟐. Calculate the Absolute and Percentage Error in the calculation of it.
Answer:-
∆𝑽
× 𝟏𝟎𝟎 = %
𝑽
Question # 32:-
Find the number of terms of the Exponential Series such that their sum gives the value of 𝐞𝐱
correct to 𝟓 decimal places for all the value of 𝐱 in the range 𝟎 ≤ 𝐱 ≤ 𝟏.
Answer:-
The number of terms in the exponential series should be 𝟗.

Question # 33:- Assignment

Find the possible Percentage Error in computing the parallel resistance 𝐫 of these resistance
𝟏 𝟏 𝟏 𝟏
𝐫𝟏 , 𝐫𝟐 , 𝐫𝟑 from the formula = + + . If 𝐫𝟏 , 𝐫𝟐 , 𝐫𝟑 are each in the error by plus 𝟏. 𝟐 %.
𝒓 𝒓𝟏 𝒓𝟐 𝒓𝟑

Answer:-
∆𝒓
× 𝟏𝟎𝟎 = 𝟏. 𝟐 %
𝒓

Question # 34:-
The Deflection 𝐃 at the centre of a road of length 𝐥 and the diameter of supported at its
ends and loaded at the centre with a weight 𝐰 varies as 𝒘𝒍𝟑 𝒅−𝟒 (𝑫 =
𝒌𝒘𝒍𝟑 𝒅−𝟒 , 𝒘𝒉𝒆𝒓𝒆 𝒌 𝒊𝒔 𝒑𝒓𝒐𝒑𝒐𝒓𝒕𝒐𝒏𝒂𝒍𝒊𝒕𝒚 𝒄𝒐𝒏𝒔𝒕𝒂𝒏𝒕. ).What is the Percentage Error in the
deflection corresponding to the percentage error in 𝐰, 𝐥 and 𝐝 of 𝟑% , 𝟐% and 𝟏%
respectively.
Answer:-
∆𝐃
× 𝟏𝟎𝟎 = 𝟓 %
𝐃

Question # 35:-
𝑬𝟐
The Power dissipated in a resistor is given by 𝑷 = . Find the Percentage Error in 𝑷 when
𝑹
percentage error in 𝑬 is 𝟑% and 𝑹 is 𝟐%.
Answer:-
𝜟𝑷
× 𝟏𝟎𝟎 = 𝟖%.
𝑷

Question # 36:-
The diameter and altitude of a can in the shape of a right circular cylinder are measured as
𝟒𝟎𝐜𝐦 and 𝟔𝟒𝐜𝐦 respectively. The possible error in each measurement is ±𝟓%. Find
approximately the Maximum Percentage Error in the computed value for the Volume and
the Lateral Surface. (Note:- 𝑽 = 𝝅𝒓𝟐 𝒉 , 𝑺 = 𝟐𝝅𝒓𝒍)

Dr. Jamil Book Series 2 Page 39

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Answer:-
𝜟𝑽 𝜟𝒔
× 𝟏𝟎𝟎 = ±𝟏𝟓% , × 𝟏𝟎𝟎 = ±𝟏𝟎%.
𝑽 𝒔

Question # 37:-
𝑨
In determining the Specific Gravity by the formula = , where 𝐀 is the weight in air and
𝑨−𝒘
𝒘 is the weight in water. 𝑨 can be read within 𝟎. 𝟎𝟏 𝒈𝒎 and 𝒘 within 𝟎. 𝟎𝟐𝒈𝒎. Find the
Maximum Error in 𝑺 if the error are 𝑨 = 𝟏. 𝟏𝒈𝒎 and 𝒘 = 𝟎. 𝟔𝒈𝒎.
Answer:-
𝜟𝑺 = −𝟎. 𝟏𝟏𝟐.

Question # 38:- Assignment

The voltage 𝑽 across a resistor is measured with error 𝑯, and the resistance 𝑹 is measured
𝑽𝟐
with an error 𝑲. Show that the error in calculating the power 𝑾 = generated in the
𝑹
𝑽𝟐 𝑯 𝑲
resistor is (𝟐 − ). If 𝑽 can be measured to an accuracy of 𝟎. 𝟓% and 𝑹 to an accuracy
𝑹 𝑽 𝑹
of 𝟏%. What is the approximately possible Percentage Error in 𝑾.
Answer:-
% error in 𝑾 = 𝟎%

Question # 39:-
Find the possible Percentage Error in computing parallel resistance 𝒓 of two resistance 𝒓𝟏
𝟏 𝟏 𝟏
and 𝒓𝟐 from the formula = + , where 𝒓𝟏 and 𝒓𝟐 are both in error by 𝟐% each.
𝒓 𝒓𝟏 𝒓𝟐

Answer:-
∆𝒓
× 𝟏𝟎𝟎 = 𝟐%
𝒓

Question # 40:-
Check whether the function 𝒇(𝒙) = 𝒕𝒂𝒏𝒙 is well-conditioned or ill-conditioned at 𝒙 = 𝟎 and
𝒙 = 𝝅⁄𝟐.

Answer:-
At 𝒙 = 𝟎:-

Well-conditioned, since the condition=𝟏 at 𝒙 = 𝟏

𝝅
At 𝒙 = : −
𝟐

Ill-conditioned since condition 𝒙 ⟶ ∞ as 𝒙 = 𝝅⁄𝟐

Question # 41:-

Check whether the function 𝒇(𝒙) = 𝒔𝒊𝒏𝒙 and 𝒇(𝒙) = 𝒄𝒐𝒔𝒙 are well-conditioned or ill-
𝝅
conditioned at 𝒙 = 𝟎 𝒂𝒏𝒅 𝒙 =
𝟐

Dr. Jamil Book Series 2 Page 40

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Answer:-
For 𝒇(𝒙) = 𝒔𝒊𝒏𝒙:-
Well-conditioned at 𝒙 = 𝟎 and
𝝅
Well-conditioned at 𝒙 =
𝟐

For 𝒇(𝒙) = 𝒄𝒐𝒔𝒙:-

Well-conditioned at 𝒙 = 𝟎 and
𝝅
Ill-conditioned at 𝒙 =
𝟐

Question # 42:-
Write:
1): Algorithm
2): Draw flow charts for each of the following:
a): Absolute value computation
b): Input and output flow chart
c): Order processing
Question # 43:- Assignment
Draw a flow chart of 𝑵!(Factorial):
Answer:-
“Flow chart of Factorial”

Dr. Jamil Book Series 2 Page 41

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Question # 44:-
Draw a flow chart of software development.
Answer:-

Flow chart of software development

Dr. Jamil Book Series 2 Page 42

Advanced Mathematical Techniques Chapter # 1 Error Analysis

Question # 45:
Draw a flow chart of “Accounts Receivable.”
Answer:
“Flow chart of Accounts Receivable”

Dr. Jamil Book Series 2 Page 43

Advanced Mathematical Techniques Chapter # 1 Error Analysis

SUMMARY OF THE CHAPTER-1

Nature or types of errors:-
1): Inherent or propagation error,
2): Truncation errors,
3): Round-off errors.
4): Significant error
The infinite series expansion of 𝒆𝒙 in term of Maclaurin’s series is given by

𝒙
𝒙𝟐 𝒙𝒏−𝟏 𝒙𝒏
𝒆 = 𝟏 + 𝒙 + + ⋯+ + +⋯
𝟐! (𝒏 − 𝟏)! 𝒏!

𝒙𝒏 𝒙𝒏 𝜽𝒙
𝑻. 𝑬 = 𝑺𝒖𝒎 𝒐𝒇 (𝒏 + 𝟏) 𝒕𝒐 ∞ 𝒕𝒆𝒓𝒎𝒔 = 𝑹𝒏 = 𝒇(𝜽𝒙) = 𝒆 , 𝟎 < 𝜽 < 𝟏.
𝒏! 𝒏!

Absolute, Relative and Percentage Error:-

𝑨𝒃𝒔𝒐𝒍𝒖𝒕𝒆 𝒆𝒓𝒓𝒐𝒓 = 𝑬𝒂 = ∆𝑽 = | 𝑽𝑻 − 𝑽𝑨 |

∆𝑽 𝑬𝒂 |𝑽𝑻 − 𝑽𝑨 |
𝑹𝒆𝒍𝒂𝒕𝒊𝒗𝒆 𝒆𝒓𝒓𝒐𝒓 = 𝑬𝒓 = = =
𝑽𝑻 𝑽𝑻 𝑽𝑻

| 𝑽𝑻 −𝑽𝑨 |
𝑷𝒆𝒓𝒄𝒆𝒏𝒕𝒂𝒈𝒆 𝒆𝒓𝒓𝒐𝒓 = 𝑬𝒑 = 𝑬𝒓 × 𝟏𝟎𝟎 = × 𝟏𝟎𝟎
𝑽𝑻

Theorem - 1:-
If a number be rounded to 𝒏 decimal places, the absolute error is given by

𝟏
𝑬𝒂 ≤ × 𝟏𝟎−𝒏 = 𝟎. 𝟓 × 𝟏𝟎−𝒏
𝟐

Theorem - 2:-
If a number be rounded to 𝒏 correct significant figures, then the relative error is given by

𝟏
𝑬𝒓 ≤ (𝒏 ≠ 𝟏):
𝑲 × 𝟏𝟎𝒏−𝟏

Dr. Jamil Book Series 2 Page 44

Advanced Mathematical Techniques Chapter # 1 Error Analysis

A General Error Formula:-

𝒖 = 𝒇(𝒙𝟏 , 𝒙𝟐 , 𝒙𝟑 , … , 𝒙𝒏 )

𝒏
𝝏𝒖
∆𝒖 = ∑ ∆𝒙
𝝏𝒙𝒊 𝒊
𝒊=𝟏
𝒏
𝝏𝒖 ∆𝒙𝒊
𝑷𝒓 = 𝑬𝒓 × 𝟏𝟎𝟎 = ∑ × 𝟏𝟎𝟎
𝝏𝒙𝒊 𝒖
𝒊=𝟏

Note:-
1): |𝒙 + 𝒚 + 𝒛| ≤ |𝒙| + |𝒚| + |𝒛|
2): |𝒙𝒚𝒛| = |𝒙||𝒚||𝒛|
Numerical stability & ill-condition:-
Well-conditioned or stability:-
A problem is well- conditioned if small errors in the data produce small errors in the
solution.
Ill-conditioned or instability:-
A problem is well- conditioned if small errors in the data produce large errors in the
solution.
The condition:-

𝒙𝒇′ (𝒙)
𝒄𝒐𝒏𝒅𝒊𝒕𝒊𝒐𝒏 ≈ 𝐦𝐚𝐱 | |
𝒇(𝒙)
Algorithms:-

An algorithm for solving a problem is a finite sequence of simple instruction which, when
followed yields the solution of the problem.

Flow charts:-
Graphical or diagrammatic representation of an algorithm is a flow chart, which its
representation in a programming language is a program.

Dr. Jamil Book Series 2 Page 45