Errors and Approximations:

Error in addition:

Let A and B are two inequalities to be added with ∆ A∧∆ B as errors in A and B respectively.

Let the result of adding A and B be expressed by the equation R = A + B.

When the errors are explicitly included then,

R+ ∆ R=( A+ ∆ A ) +( B+ ∆ B)

¿( A +B)+(∆ A+ ∆ B)

So, the error in R is ∆ R=∆ A+ ∆ B

Error in multiplication:

Let R = A . B

R+ ∆ R=( A+ ∆ A ) (B+∆ B)

¿ AB+ A ∆ B +B ∆ AB+ ∆ A . ∆ B

∆ R= A ∆ B+ B ∆ AB+ ∆ A . ∆ B

Assumption:

If ∆ A and ∆ B is small relative to A and B respectively then ∆ A . ∆ B is extremely very small relative to
A.B and can therefore be ignored.

So, ∆ R= A ∆ B+ B ∆ A

Relative error:

∆ R ∆ A ∆B
= +
R A B

Error in quotient:

A
Let R=
B

A +∆ A B−∆ B
R+ ∆ R= .
B+∆ B B−∆ B

AB−A ∆ B+ B ∆ A−∆ A . ∆ B
¿ 2
B
 AB A ∆ B B ∆ A
¿ 2
− 2 + 2
B B B

−A ∆ B B ∆ A
∆ R= 2
+ 2
B B

∆ R −A ∆ B B B ∆ A B
= . + 2 .
R B
2
A B A

∆ R −∆ B ∆ A
= +
R B A

Examples:

Given X= 2.2255, Y= 0.449, correct to the given number of decimal places. State the maximum
possible errors in the values of X and Y. Hence determine;
(i) The absolute error
X
(ii) The limit within which the value of the quotient lies giving your answer to 2 dps.
Y

Solution:

(i)

Value Maximum Minimum ∆ (Value)

X =2.2255 2.22555 2.22545 0.00005 (∆ X)
Y = 0.449 0.4495 0.4485 0.0005 ( ∆ Y )

X ∆ X X ∆Y
Error in = − .
Y Y Y Y

0.00005 2.2255 0.0005

= − .
0.449 0.449 0.449

= -0.005408

Absolute error is 0.005408 0.01(2 dp)

(ii) Limits bounding the quotient are given by;

min ( YX )≤ YX ≤ max ( YX )
X
4.95 ≤ ≤ 4.96
Y
Example:

(a) Given that x=2.5 , y=14.2∧z=8.1 all values given correct to 2 dps. Find the maximum values of;

x+ y x− y 1 1 1
(i) 2 (ii) (iii) + −
z z x y z

correct to 3 decimal places.

(b) Two decimal numbers x∧ y are rounded off to give X ∧Y with errors E1∧E 2 respectively. Show

that the maximum relative error made in approximating x 2 y by X 2 Y is given by 2 | || |

E1 E2
X
+
Y

Solution:

(a)

Value Maximum Minimum

x=2.5 2.55 2.45
y=14.2 14.25 14.15
z=8.1 8.15 8.05

x+ y max ( x )+ max ( y )
(i) Max =
min ( z 2 )
2
z
2.55+14.25
=
( 8.05 )2
= 0.259
x− y max ( x )−min ( y )
(ii) Max =
z min ( z )
2.55−14.15
=
8.05
= -1.441

(iii) Max ( 1x + 1y − 1z )= min1( x ) + min1( y ) − max1 ( z )

1 1 1
= + −
2.45 14.15 8.15
= 0.356
(b) x=X + E1∧ y=Y + E2
2 2 2 2
Error=x y −X Y =( X+ E1 ) ( Y + E2 )− X Y
2 2 2 2 2 2
= X Y + E2 X +2 XY E 1+ E1 Y + E 1 E2 − X Y
2
E2 X +2 XY E 1
 2
E2 X 2 XY E 1
Relative error= 2
+ 2
X Y X Y
E E
= 2 −2 1
Y X

Maximum Relative Error = ||| |

E2
Y
E
+2 1
X

Example:

FORMULATE QUESTION:

Solution:

No: Min: Max:

3.551 3.5495 3.5515
2.71635 2.716345 2.716355

Min (3.551X2.71635) = Min (3.551) X Min (2.71635)

= 3.5495 X 2.716345 = 9.6444 (4dp)

Max (3.551X2.71635) = Max (3.551) X Max (2.71635)

= 3.5515 X 2.716355 = 9.6471 (4dp)

9.6444 ≤ True Value ( 9.64575 ) ≤ 9.6471

Alternatively;

1
Error bound = (Max . Value−Min .Value)
2

1
= ( 9.647134783−9.644382923 )=0.00137593
2

Working value = 3.551X2.71635 = 9.64575885

Range of values = ¿ 9.64575885 ± 0. 00137593

Exercise:
1. Given that x=3.57 , y=−4.291∧z =6.7955 are rounded off to the given decimal places
indicated.

(i) State the maximum possible errors in x , y ,∧z .

x
(ii) Find the limits within which the exact value of the expression, lies.
y−z

15.36+27.1−1.672
2. (a) Determine the approximate values between which lies,
2.36 ×1.043
assuming the maximum possible error in each of the figures.

(b) Determine the maximum absolute error in

√ Z given that X=2.4, Y=5.4 and
2 3
X Y
Z=1.8 all numbers rounded off.

3. If the error in each of the values of 2 x and 2−x is ± 0.005. find the exact value of the
x
2
quotient −x , when x=0.40 , giving your answer correct to 3 decimal places.
2

ESTIMATION OF ROOTS:

Newton-Raphson method: y

If point Q is near the point P

X n can be considered to be

the approximation of the exact

root x . Q ( xr , yr )

NQ=f ( x r ) α

NQ NQ
=tan α → RN = O P ( x ,0 ) Q( x r +1 , 0 )N x
RN tan α

tan α is equal to grad at Q.

'
tan α =f ( x r )
 f ( xr )
RN=
f ' ( xr )

f ( xr )
¿=ON – RN →∨¿ x r − '
f ( xr )

f ( xr )
x r +1=x r−
f ' ( xr )

Example:

Using the Newton- Raphson formula, find the root of the equation
3
X −2 X−5=0

giving your answer correct to 2 decimal places.

Solution:

f ( X n)
Newton-Raphson formula is given by X n+1=X n− '
f ( Xn)

' 2
Now,
3
f (X n )=X n −2 X n−5 and f ( X n )=3 X n −2

( X n3−2 X n−5 )
So, X n+1=X n−
3 X n2−2
3 3
3 X n −2 X n− X n +2 X n +5
= 2
3 X n −2
3
2 X n +5
= 2
3 X n −2

Let x 0=2
n Xn X n+1 | X n+1− X n|
0 2 2.1 0.1
1 2.1 2.094 0.0054
2 2.094 2.094 0.0000
Root of equation is 2.09 (2dps)
Example:

(a) Show that the Newton-Raphson iterative formula for finding the cube root of any
number N is given by

X n+1=
1
3 (N
)
2 X n + 2 ; n=0 , 1 , 2, … … … … … …
Xn
1
Use the iterative formula to find 96 3 correct to 3 decimal places.

(b) Show that the equation 3 X 3 + X −5=0 has a real root between X=1 and X=2.

Using the Newton-Raphson formula thrice, find the value of his root correct to 2 d.p.

Solution:
1
(a) Let X =N 3

3
X −N=0
3 ' 2
Now, f ( X n )= X n −N and f ( X n )=3 X n

f ( X n)
Using X n+1=X n− '
f ( Xn)

( X n3−N )
= X n−
3 X n2
3 3
3 X n −X n + N
= 2
3 Xn
3
2 Xn + N 2 1 N
= 2 = 3 . Xn+ 3 . 2
3 Xn Xn

X n+1=
1
3 ( N
2 Xn+ 2
Xn )
Let x 0=5 , N=96

n Xn X n+1 | X n+1− X n|
0 5 4.613 0.3867
1 4.613 4.5791 0.0339
 2 4.5791 4.5789 0.0002
3 4.5789 4.5789 0.0000

1
Therefore, 96 3 =4.579(3 d . p)

(b) f ( x )=3 x 3 + x−5

Now, f ( 1 ) =−1∧f (2 )=+21

The function changes sign, from negative (-) to positive (+) between x=1∧x =2.

This implies that the real root of the function f ( x )=3 x 3 + x−5 lies between the two

x-values.

Note: Graphical method is also appropriate.

Example:

FORMULATE QUESTION

Solution:
−3
(i) a=23.037 ; error ∈a ( e1 )=±0.5 × 10 =± 0.0005
−4
b=8.4658 ; error ∈b ( e 2 )=± 0.5 ×10 =± 0.00005

a
(ii) Absolute error in
b

Absolute error ¿
|e 1| |e 2|
+
a b

0.0005 0.00005
¿ +
23.037 8.4658

¿ 0.0000276

(iii) max ( ab )= max a 23.0375

=
min b 8.46575
=2.7213

min ( ba )= max
min a 23.0365
=
b 8.46585
=2.7211
 a
Range is given by: 2.7211< <2.7213
b

Example:

FORMULATE QUESTION:

Solution:
' x
Let (x n )=e −2 x n−1 , f ( x n )=e −2
x n n

f ( X n)
From; X n+1=X n− '
f ( Xn)

( e x −2 x n−1 )
n

¿ X n−
e x −2
n

xn
e ( X n−1 ) +1
¿ xn
e −2

n Xn X n+1 | X n+1 −X n|
0 1.000 1.389 0.389
1 1.389 1.271 0.108
2 1.271 1.259 0.012
3 1.259 1.256 0.003
4 1.256 1.256 0.000

xn
Root of equation e −2 x n−1=0 is x=1.256

Example:

FORMULATE QUESTION:

Solution

(a) Error: ( X + ∆ X ) ( Y +∆ Y ) −XY = XY + X ∆ Y +Y ∆ X + ∆ X ∆ Y −XY

¿ X ∆ Y +Y ∆ X +∆ X ∆ Y ; Neglect error in ∆ X ∆ Y

¿ X ∆ Y +Y ∆ X
 X ∆ Y +Y ∆ X ∆ X ∆ Y
Relative error: = +
XY X Y

Percentage error = Relative error × 100%

[| | | |]
∆ X ∆Y
X
+
Y
×100 %

Exercise:

1. By plotting graphs of y=e x and y=− x+ 4 on the same co-ordinate axes, estimate the
only real root of the equation e x + x−4=0 correct to one decimal place.

2. (a) Show that the equation x=ln ( 8−x ) has a root between x=1 and x=2.

(b) Use the Newton- Raphson method to find the root of the equation correct to 2
dp.

3. (a) Given two iterative formulae I and II for calculating the positive root of the
quadratic equation f(x) = 0 as:

I II

[ ]
2
1 1 X n +1
X n+1= ( X 2−1 )
2 n
X n+1=
2 X n−1

For n = 1, 2, 3, ….

Taking X 0=2.5, use each formula thrice to 2 dps to decide which is the more suitable
formula. Give a reason for your answer.

(b) If α is an approximate root of the equation X 3 =n, show that the iterative
formula for finding the root reduces to

n
+α
α
2

Hence, taking α =4 estimate √ 17 correct to 3 dps.

4. (i) Show that the iterative formula for approximating the root of F(x) = 0 by the
Newton-Raphson process for the equation x e x +5 x−10=0 is
 2 Xn
X n e +10
X n+1= Xn
e ( X n +1 ) +5

(ii) Show that the root of the equation in (i) above lies between 1 and 2. Hence, find
the root of the equation correct your answer to 2 dps.

5. Using the same graph show that the curve X 3 and 2X+5 have a common real root. Using
the Newton-Raphson formula twice, find the positive root of X 3 −2 X−5=0, giving your
answer correct to 2 dps.

6. (a) Show that the Newton-Raphson formula for finding the root of the equation
x −3 x+ 4=0 is
3

[ ]
3
2 x n −2
, n = 0, 1, 2, ……………………
3 x n2−1

(b) Taking the first approximatrion to the root of the above equation as -2.0, draw a
flow diagram which reads and prints the number of iterations and the root. Carry
out a dry run of the flow chart and obtain the root with an error of less than
0.005.

FLOW CHARTS/DIAGRAMS:

1. Flow charts/diagrams are schematic representations of algorithms.

2. They show the logical or sequential flow of processes/programs using arrows and
symbols.

Symbols:

START/END
 OVAL

INPUT (READ)/
PARALLELOGRAM
OUTPUT (PRINT)

RECTANGLE
PROCESS/ALGORITM TO
BE EXECUTED

DECISION
TO BE YES RHOMBUS/DIAMOND BOX
TAKEN

NO

DIRECTION OF FLOW ARROW

Example: [1990 (1)]

Study the flow chart and perform a dry run of the flow chart. Hence, state the purpose of the
flow chart.

START
Solution:

Dry run: (Look at the variables)

S=0
S N (S , N)

0 1 …………

N=1
1 3 1 , 1

4 5 4 , 3

9 7 9 , 5

16 9 16 , 7

25 11 25 , 9

36 13 36 , 11

49 15 49 , 13

The flow chart prints S or squares of the first seven

natural numbers and N sequence of first seven odd

numbers.

Example:

Draw a flow chart for computing and printing the mean of the square roots of the first one
hundred natural numbers.

N=N+1
START END

N=0 S=S+N
S=0
 IS
N=100 NO
PRINT A
S ?
A=
100

START
YES

Flow Chart: N=0

S=0
Notes:

 Start
 Inputs: (N=0 ; S=0)/PRINT S=S+N
 Counters: (N=N+1)
S
 Algorithm for Average: (A= ) IS
100
N=100 N=N+1
 Algorithm for Sum: (S=S+N)
?
 Decision box
 End

S
A=
100

PRINT A

Example: [1995 (1)] START

Given the flow diagram/chart shown.

What is the outcome in words? A=0

N=0
After the dry run, state the relationship between N and T T=0

Dry Run:

N=N+1
 A N T N , T
0 0 0 …………..
1 1 1 1 , 1
3 2 8 2 , 8
T=T+6A+1
6 3 27 3 , 27
10 4 64 4 , 64
15 5 125 5 , 125
21 6 216 6 , 216 A=A+N
28 7 343 7 , 343
36 8 512 8 , 512
45 9 729 9 , 729 PRINT; N,T
55 10 1000 10 , 1000

The flow chart prints the first natural numbers and their NO
IS
corresponding cube roots. T =N 3 N=10
?

YES

STOP
Example: [2008(2)]

Study the flow chart below. Using the flow chart, perform a dry run and complete the table
below for X=22 and Y=7.

Solution:

P Q
START
0 22

1 15 Read: X,Y

2 8
P=0
3 1

The flow chart is used for finding the first four terms of
Q=X
whose common difference is -7 and first term is 22.

IS
Q<Y
?
 YES

WRITE: P,Q
NO
P=P+1

STOP

Q=Q-Y

Example: [2009 (9)]

Given below are parts of a flow diagram not arranged in order.

READ: A C=B÷ 2
START
IS
B ≤0 YES
?

NO

PRINT: C
IS NO
B DIVISIBLE STOP
BY 2 ?

A ← A +1
YES B= A−5
 (a) Rearrange them and draw a complete logical flow chart.
(b) State the purpose of the flow chart.
(c) Perform a dry run of your arranged flow chart by coping and completing the table
below.
A B C
46 ……….. ………..
77 ……..... ………..
120 ………… ………..
177 ………… …………

Solution:

(b) Completes and prints positive natural

START
numbers divisible by 2.

(c) A B C
READ: A
46 41 ……

77 72 36
B= A−5
120 115 ……

177 172 86

YES
IS
B ≤0 A ← A +1
?
 NO

NO

YES

Example: [1989 (2)]

In the flow diagram C=Counter; SSN = Social Security Number; W=Monthly Wage of a factory
employee in shillings.
START
Copy and complete the following
table:
SSN W T C=1
01-86-003 8400 ………
03-86-095 8200 ………
04-86-064 7500 ………. SSN, W
02-86-035 8000 ………
04-86-066 6400 ………
01-87-098 4800 ………
02-87-105 6300 ……….
IS
03-87-135 5500 ………. W ≤ 4000 YES T =0
01-88-215 3800 ………. ?
01-88-217 3500 ………..
For how many employees is this program
Designed? Among these employees there NO
are some senior staff whose monthly
EXCESS=W −4000
salary is more than Sh.10,000 each.
These pay tax as follows: 12.5% for salary
Above Sh.4,000 but not more than Sh. 10,000;
T =EXCESS ¿0.125
Then 25% for salary above Sh. 10,000. Modify
The given flow diagram in order to incorporate
the senior staff.
WRITE: SSN,W,T

IS
C=C+ 1 NO C=270
?

YES

STOP

Solution:

The program is designed for 270 employees.

START

C=1

SSN, W

IS
W ≤ 4000 T =0
? YES

IS
W ≤ 4000 EXCESS=W −4000 T =EXCESS ¿0.125
?
 NO

YES

NO

NO

YES
STOP

Example:

In the flow chart given shows

the SSN and the monthly wage

(W/=) of an employee. P START

represents the net pay.

Copy and complete the following: SSN, W

SSN W T P

280 – 04 380 0 380 IS YES

W ≤ 400 T =0
180 – 34 840 55 785 ?

179 – 93 4500 625 3875 NO

380 – 06 5580 895 4685

385 – 03 8000 1500 6500 IS YES

EXCESS I = W - 400
400 ≤ W ≤ 3000 EXCESS II =0
?
 NO

Example:

Derive a Newton-Raphson formula for finding;

(a) The fifth root of any number, N.

(b) Draw a flow chart that can be used to work out the root.
1
(c) Perform a dry run to find ( 165 ) 5 . Use x 0=2.000

Solution:
1
Let X =N 5
5
X =N
5
X −N=0
5
f ( X n )= X n−N
' 4
f ( X n )=5 X n

Using the Newton –Raphson Iterative formula;

 f ( X n)
X n+1=X n− '
f ( Xn)

5
(X n −N)
¿ X n− 4
5 Xn

X n+1=
1
5( N
4 X n− 4
Xn )
(b)
READ: X 0
END
START

n =0 n = n+1
IS
| X n+1− X n|≤0.5 × 10−3
?

X n+1=
1
5( N
4 X n− 4
Xn )
Flow chart:

START

READ: X 0

n =0

X n+1=
1
5 ( N
4 X n− 4
Xn )

IS
| X n+1− X n|≤0.5 × 10−3 n = n+1
?
 NO

YES

(c)

Dry run:

n X n+1 X n+1 | X n+1− X n|

0 2.0000 3.6625 1.6625
1 3.6625 3.1134 0.5491
2 3.1134 2,8419 0.2715
3 2.8419 2.7794 0.0625
4 2.7794 2.7765 0.0029
5 2.7765 2.7765 0.0000

Example:

FORMULATE QUESTION:

Solution:

(a) A = X – a B = f(A) B=Y–b

Y = f(X) Y = f (A + a)

f(A) = f (A + a) – b

b = f (A + a) - f(A)

f ( A+ a)−f (A )
but f ' ( A)=
a
'
af ( A )=f ( A +a )−f ( A )

b=af ' ( A )
 |b|=|a||f ' ( A )|
p
(b) f ( A )= A
' p−1
f ( A ) =p A

|b|=|a| p A p−1

|b|
Relative error =
B

|a| p A p−1
=
f ( A)

|a| p A p−1
=
Ap

= |a| p A−1

Example: [1988 (2b)]

An iterative method for approximating a root of the equation f ( x )=0 is described in the flow
chart below.

Given that A = 1.6875, B = 1.8750 and Tol = 0.5 ×10−3, perform a dry run to find the values of B
and X.

START

READ: A, B, TOL

A+ B
X=
2
LINEAR INTERPOLATION:

Interpolation = Interval estimation

If point P and Q are close enough we

can approximate the length of curve PQ f ( x 1 ) Q

to be equal to the length on line PQ.

Let P[ x 0 , f ( x0 ) ] and Q[ x 1 , f ( x 1 ) ] f (x) R

be points on the curve PQ whose f ( x0 ) P

coordinates are known. Suppose

R[ x , f ( x ) ] is a point between
P and Q in which either the ordinates

or abscissa is unknown by considering

the fact that Grad PQ=Grad RP

f ( x 1 )−f ( x 0 ) f ( x ) −f ( x 0 )
=
( x 1 )−( x 0 ) ( x ) −( x 0 )

From which;

f ( x ) −f ( x 0 )
x=x 0 + ( x 1−x 0 )
f ( x 1 )−f ( x 0)

( x−x 0 )
Or f ( x )=f ( x 0 ) +
( x1 −x 0)
[ f ( x 1 ) −f ( x 0 ) ]

Example:

In an experiment the following observations were recorded

T: (x) 0 12 20 30
Q: 6.6 2.9 -0.1 -2.9
Use linear interpolation to find

(i) Q when T = 16

(ii) T when Q -1

Solution:

f ( x 1 )−f ( x 0 ) f ( x ) −f ( x 0 )
=
( x 1 )−( x 0 ) ( x ) −( x 0 )

(i) f ( x )=? When x=16 , x 0=12, x 1=20, f (x ¿¿ 0)=2.9 ¿, f ( x 1 ) =−0.1

16−12
f ( x )=2.9+ (−0.1−2.9)
20−12

¿ 1.4

(ii) x=? f ( x )=−1, f (x ¿¿ 0)=−0.1 ¿, f ( x 1 ) =−2.9, x 0=20 , x 1=30

 (−1−−0.1 )
x=20+ (30−20)
(−2.9−−0.1 )

¿ 23.2

Example:

FORMULATE QUESTION:

Solution:

Let log 10 x=f (x)

x 80.00 80.20 80.50 80.80

log 10 x or f (x) 1.9031 1.9042 1.9058 1.9074
(a) Using linear interpolation

( x−x 0 )
f= ( f 1−f 0 ) +f 0
( x 1−x 0 )
( 80.759−80.50 )
f= ( 1.9074−1.9058 ) +1.9058
( 80.80−80.50 )

f =1.9072(4 dp)

( f −f 0 )
(b) x= ( x 1−x 0 ) + x 0
( f 1−f 0 )
( 1.90388−1.9031 )
x= ( 80.20−80.00 )+ 80.00
( 1.9042−1.9031 )

x=80.14 (2 dp)

Exercise:

1. The table below shows the values of a function f(x) at a set of points.

x 0.9 1.0 1.1 1.2

f(x) 0.266 0.242 0.218 0.192

Use linear interpolation to find;

 (i) the value of f(1.04)

(ii) the value of x corresponding to f(x)=0.25

2. Given the following table of values:

X 0 5 10 15
t 0 13 24.1 38.7
Use linear interpolation or extrapolation to find:

(i) t when x = 8

(ii) x when t = 44

TRAPEZIUM RULE:
3

Use the trapezium rule with 7 ordinates to estimate ∫ x+1 1 dx correct to 3 dps.
0

Solution:
3

∫ x+1 1 dx 1
2
h [ y 0+ 2 ( y 1+ y 2 +…+ y n−1 )+ y n ]
0

x 0 0.5 1.0 1.5 2.0 2.5 3.0

y 1 0.667 0.5 0.4 0.333 0.286 0.25

h=0.5
3

∫ x+1 1 dx= 0.5

2
[1+ 2 ( 0.667+ 0.5+0.4 +0.333+0.286 )+ 0.25 ]
0

¿ 1.406 3 dp

Example:

FORMULATE QUESTION:

Solution:
 xn
1
∫ f (x )dx 2 [ 0
h y +2 ( y 1+ y 2+ …+ y n−1 ) + y n ]
x0

n 0 1 2 3 4 5
x 0 0.2 0.4 0.6 0.8 1.0
yn 1 0.9606 0.8518 0.6978 0.5274 0.3678

∫ e−x dx = 12 ×0.2 [ 1+ 2 ( 0.9606 +0.8518+0.6978+ 0.5274¿+ 0.3678 ) ]

2

¿ 0.7443

Exercise:
π
4
1. Use the trapezium rule with 5 ordinates to evaluate ∫ sec x dx correct to 4 dps.
0