Numerical Methods

h na 's
K ris Text Book Solutions
NUMERICAL METHODS
(For B.A. and B.Sc. VIth Semester students of All Colleges affiliated to
Universities in Kumaun)
As per Kumaun University Syllabus
By
A. R. Vasishtha Hemlata Vasishtha

Retired Head, Dep’t. of Mathematics M.Sc. (Gold Medalist), Ph.D.
Meerut College, Meerut (U.P.) C.C.S. University, Meerut (U.P.)
KRISHNA Prakashan Media (P) Ltd.

KRISHNA HOUSE, 11, Shivaji Road, Meerut-250 001 (U.P.), India
Jai Shri Radhey Shyam
Dedicated
to
Lord
Krishna
Authors & Publishers
B rief C ontents
Dedication...............................................................(iII)
Brief Contents.........................................................(IV)
Chapter 1: The Calculus of Finite Differences..........................N-01—N-28
Chapter 2: Interpolation with Equal Intervals.........................N-29—N-43
Chapter 3: Interpolation with Unequal Intervals of the

Argument..........................................................N-44—N-58
Chapter 4: Central Difference Interpolation Formulae............N-59—N-72
Chapter 5: Solution of Algebraic and Transcendental Equations

.........................................................................N-73—N-86
Chapter 6: Numerical Differentiation....................................N-87—N-97
Chapter 7: Numerical Integration.........................................N-98—N-119
Chapter 8: Errors in Numerical Calculations.......................N-120—N-123
Chapter 9: Simultaneous Linear Algebraic Equations.........N-124—N-137
Chapter 10: Curve Fitting...................................................N-138—N-144
(iv)
N-3
Chapter-1
The Calculus of Finite Differences
Comprehensive Problems 1
Problem 1: Evaluate the following :
 2x 
(i) ∆  (ii) ∆ cot 2 x
 ( x + 1) !
(iii) ∆ sinh ( a + bx ) (iv) ∆ cosh ( a + bx )
∆2 x3
(v) ∆ tan ax (vi)
E x3
(vii) ∆4 ae x (viii) ∆ tan−1 ax.
 2x  2 x +1 2x
Solution: (i) ∆ = − , taking the interval of differencing as unity
 ( x + 1) ! ( x + 2) ! ( x + 1) !
2x  2  2x 2 − x − 2  − x . 2 x
=  − 1 =  = ⋅
( x + 1) !  x + 2  ( x + 1) !  x + 2  ( x + 2) !
(ii) ∆ cot 2 x = cot 2 x +1 − cot 2 x , taking the interval of differencing as 1
cos 2 x +1 cos 2 x cos 2 x +1 sin 2 x − cos 2 x . sin 2 x +1

= − =
x +1
sin 2 sin 2 x
sin 2 x +1 . sin 2 x
sin (2 x − 2 x + 1 ) sin { 2 x (1 − 2)} − sin 2 x

= = =
x +1 x +1
x
sin 2 . sin 2 x
sin 2 . sin 2 sin 2 x . sin 2 x +1
= − cosec 2 x +1
(iii) ∆ sinh ( a + bx ) = sinh [ a + ( x + h) b ] − sinh [ a + bx ]

2a + 2bx + bh bh
= 2 cosh sinh
2 2
= 2 cosh  a + + bx  sinh , for h = 1.

b b
 2  2
(iv) ∆ cosh ( a + bx ) = cosh [ a + b ( x + h)] − cosh [ a + bx ]
2a + 2bx + bh bh
= 2 sinh sinh
2 2
= 2 sinh  a + + bx  sinh , for h = 1.

b b
 2  2
sin a ( x + 1) sin ax
(v) ∆ tan ax = tan a ( x + 1) − tan ax = − , for h = 1
cos a ( x + 1) cos ax
sin a ( x + 1) cos ax − sin ax cos a ( x + 1)
=
cos a ( x + 1) cos ax
N-4
sin [ ax + a − ax ] sin a
= = ⋅
cos ax cos a ( x + 1) cos ax cos a ( x + 1)
∆2 x3 ( E − I )2 x3 ( E2 + I − 2 E ) x3 E2 x3 + Ix3 − 2 E x3
(vi) = = =
3 3 3
E x E x E x E x3
( x + 2)3 + x3 − 2 ( x + 1)3
=
( x + 1)3
x3 + 8 + 6 x2 + 12 x + x3 − 2 ( x + 1)3
=
( x + 1)3
6 ( x + 1) 6
= = ⋅
3
( x + 1) ( x + 1)2
(vii) ∆ ae x = a ∆ e x = a [ e x + h − e x ]
= ae x [ e h − 1] = ae x ( e − 1), for h = 1.
Now ∆2 ae x = ∆ [ ∆ ae x ]
= ∆ [ ae x ( e − 1)] = a ( e − 1) [ e x +1 − e x ]
= a ( e − 1) e x ( e − 1) = a ( e − 1)2 e x .
Proceeding like this, we get

∆4 ae x = a ( e − 1)4 e x .
(viii) ∆ tan−1 ax = tan−1 a ( x + 1) − tan−1 ax

 ax + a − ax  a
= tan−1   = tan
−1
⋅
1 + ( ax + a ) ax  1 + a x + a2 x
2 2
Problem 2: Show that

n
b + π
∆n sin ( a + bx ) =  2 sin  sin  a + bx + n 
b
(i)
  ⋅
 2  2  
n
π + b
∆n cos ( a + bx ) =  2 sin  cos  a + bx + n 
b
(ii)
  ⋅
 2  2  
− sin b
(iii) ∆ cot ( a + bx ) = ⋅
sin ( a + bx ) sin ( a + b + bx )
Solution: (i) We have

∆ sin ( a + bx ) = sin [ a + b ( x + 1)] − sin ( a + bx )
a + bx + b + a + bx   a + bx + b − a − bx 
= 2 cos   sin  
 2   2 
= 2 cos  a + bx +  sin
b b
 2 2
π
sin  a + bx + + 
b b
= 2 sin
2  2 2 
b + π
sin  a + bx +
b
= 2 sin ⋅
2  2 
N-5
b + π
∆2 sin ( a + bx ) = ∆ 2 sin sin  a + bx +
b
And
 
2  2  
= 2 sin
b sin a + b ( x + 1) + b + π  − sin  a + bx + b + π  
   
2   2   2  
2b + π 
⋅ 2 cos  a + bx +
b b
= 2 sin  sin
2  2  2
2
2b + π 
=  2 sin  cos  a + bx +
b

 2  2 
2
b + π
=  2 sin  sin  a + bx + 2 
b
  ⋅
 2  2  
Proceeding similarly, we have

n
b + π
∆n sin ( a + bx ) =  2 sin  sin  a + bx + n 
b
  ⋅
 2  2  
(ii) We have ∆ cos ( a + bx ) = cos { a + b ( x + 1)} − cos ( a + bx )
= − 2 sin  a + bx +  sin
b b
 2 2
b + π
cos  a + bx +
b
= 2 sin ⋅
2  2 
b + π
∆2 cos ( a + bx ) = ∆ 2 sin cos  a + bx +
b
And
 
2  2  
= 2 sin
b  cos a + b ( x + 1) + b + π  − cos  a + bx + b + π  
   
2   2   2  
2b + π  
⋅ 2 sin  a + bx +
b b
= 2 sin   − sin 
2  2   2
2
b + π
=  2 sin  cos a + bx + 2 
b
 ⋅
 2   2  
Proceeding similarly, we have

n
π + b
∆n cos ( a + bx ) =  2 sin  cos  a + bx + n 
b
  ⋅
 2  2  
(iii) We have
∆ cot ( a + bx ) = cot [ a + b ( x + 1)] − cot ( a + bx )
cos ( a + bx + b ) cos ( a + bx )
= −
sin ( a + bx + b ) sin ( a + bx )
cos ( a + bx + b ) sin ( a + bx ) − cos ( a + bx ) sin ( a + bx + b )

=
sin ( a + b + bx ) sin ( a + bx )
− sin b
= ⋅
sin ( a + bx ) sin ( a + b + bx )
N-6
Problem 3: Prove that, if f ( x ) and g( x ) are any functions of x,
(i) ∆ [ f ( x ) + g( x )] = ∆ f ( x ) ± ∆ g( x ),
(ii) ∆ [ a f ( x )] = a ∆ f ( x ), a being a constant,
(iii) ∆ [ f ( x ) g( x )] = f ( x + 1)∆ g( x ) + g( x ) ∆ f ( x )
 f ( x )  g ( x ) ∆ f ( x ) − f ( x ) ∆ g( x )
(iv) ∆ = ⋅
 g ( x ) g( x ) g ( x + 1) (Kanpur 2012)
Solution: See article 9.
Problem 4: Evaluate the following :

2
(i) ∆ ( x + e x + 2) (ii) ∆3 [ ax3 + bx2 + cx + d] (iii) ∆2 e x .
Solution: (i) By definition ∆ f ( x ) = f ( x + h) − f ( x ).
∴ ∆ ( x2 ) = ( x + h)2 − x2 = 2hx + h2 ;
∆ ( e x ) = e x + h − e x = e x ( e h − 1) ;
and ∆ (2) = 2 ∆ (1) = 2 . (1 − 1) = 2 . 0 = 0.

Hence ∆ ( x2 + e x + 2) = ∆ ( x2 ) + ∆ ( e x ) + ∆ (2)
= 2hx + h2 + e x ( e h − 1) + 0 = 2hx + h2 + e x ( e h − 1).
(ii) ∆3 [ ax3 + bx2 + c x + d ] = ∆3 ax3 , [ ∵ By fundamental theorem of difference calculus,

∆3 bx2 = 0 ; ∆3 cx = 0 and ∆3 d = 0 ]
= a ∆3 x3 = a . 3 ! h3 = 6 ah3 .
(iii) Proceed as in Ex. 2 (ii). Ans. e x ( e h −1)2 .
Prove that ∆ log x = log 1 +  ⋅

h
Problem 5(i):
 x
Solution: L.H.S. = ∆ log x = log ( x + h) − log x
x+ h
= log 1 +  ⋅
h
= log
x  x
Problem 5(ii): If f ( x ) = e ax , show that f(0 )and its leading differences form a geometrical progression .
Solution: We have f ( x ) = e ax .
∴ f (0 ) = e0 = 1
∆f (0 ) = f (0 + h) − f (0 ) = f ( h) − f (0 ) = e ah − 1
∆2 f (0 ) = f (0 + 2h) − 2 f (0 + h) + f (0 ) = f (2h) − 2 f ( h) + f (0 )
= e2 ah − 2e ah + 1 = ( e ah − 1)2
∆3 f (0 ) = f (0 + 3h) − 3 f (0 + 2h) + 3 f (0 + h) − f (0 )
= f (3h) − 3 f (2h) + 3 f ( h) − f (0 )
N-7
= e3 ah − 3e2 ah + 3e ah − 1
= ( e ah − 1)3
… … … … … … …
… … … … … … …
Thus f (0 ) and its leading differences form a G.P. whose first tem is 1 and common ratio is
( e ah − 1).
Problem 6: Prove that
cos  a  x +  + b  ⋅
ah h
(i) ∆ sin ( ax + b ) = 2 sin
2   2 
∆ cos ( ax + b ) = − 2 sin   sin  a  x +  + b 

ah h
(ii)
2   2 

∆ sin ( ax + b ) = sin { a ( x + h) + b } − sin ( ax + b)
= 2 sin   ⋅ cos  a  x +  + b ⋅
ah h
2   2 
∆ cos ( ax + b ) = − 2 sin   sin  a  x +  + b 

ah h
(ii)
2   2 
and ∆ tan ( ax + b ) = sin ah sec ( ax + b ) sec [ a ( x + h) + b ].
Problem 7: Prove that ∇ = ∆ E −1 = E −1 ∆ = I − E −1.
Solution: We have
∇ f ( x ) = f ( x ) − f ( x − h) ;
∆ E − 1 f ( x ) = ∆ f ( x − h) = f ( x ) − f ( x − h) ;
E −1 ∆ f ( x ) = E −1 [ f ( x + h) − f ( x )] = E −1 f ( x + h) − E −1 f ( x )
= f ( x ) − f ( x − h) ;
( I − E −1 ) f ( x ) = f ( x ) − E −1 f ( x ) = f ( x ) − f ( x − h).
From above we observe,

∇ = ∆ E − 1 = E − 1 ∆ = I − E − 1.
Problem 8: Prove that ∆3 = E3 − 3E 2 + 3E − I.

Solution: We have ∆3 f ( x ) = ∆ [ ∆2 f ( x )]
= ∆ [ f ( x + 2h) − 2 f ( x + h) + f ( x )]
= ∆ f ( x + 2 h) − 2 ∆ f ( x + h) + ∆ f ( x )
= [ f ( x + 3h) − f ( x + 2h)] − 2 [ f ( x + 2h) − f ( x + h)]
+ [ f ( x + h) − f ( x )]
= f ( x + 3 h) − 3 f ( x + 2 h) + 3 f ( x + h) − f ( x )
= E3 f ( x ) − 3E2 f ( x ) + 3E f ( x ) − I f ( x )
= ( E3 − 3E2 + 3E − I ) f ( x ).
N-8
Hence, by the definition of equality of two operators, we have

∆3 ≡ E3 − 3E2 + 3E − I.
Aliter: We have ∆≡E−I [∵ E ≡ I + ∆ ]

3 3
⇒ ∆ ≡ (E − I )
⇒ ∆3 ≡ E3 − 3E 2 I + 3E I 2 − I 3 [By binomial theorem]

3 3 2
⇒ ∆ ≡ E − 3E + 3E − I.
Problem 9: Find the successive differences of f ( x ) = ab cx and sum the first n differences.
Solution: Here f ( x ) = ab cx .
∴ ∆ f ( x ) = ab c ( x + h) − ab cx = ab cx . ( b ch − 1).
Again ∆2 f ( x ) = ∆ ∆ f ( x ) = ∆ [ ab c x . ( b c h − 1)]
= a ( b ch − 1) ∆ b cx = a ( b ch − 1) [ b c ( x + h) − b cx ]
= a ( b ch − 1) . b cx ( b ch − 1) = ab cx ( b ch − 1)2
and so on.
In general, ∆n f ( x ) = ab cx ( b ch − 1)n.
Hence the required sum

= ∆ f ( x ) + ∆2 f ( x ) + … + ∆n f ( x )
= ab cx ( b ch − 1) + ab cx ( b ch − 1)2 + … + ab cx ( b ch − 1)n
= ab cx ( b ch − 1) [1 + ( b ch − 1) + ( b ch − 1)2 + … + ( b ch − 1)n−1 ]
1 − ( b ch − 1)n  ( b ch − 1)n − 1
= ab cx ( b ch − 1)  cx ch
 = ab ( b − 1) ⋅ ⋅
ch
 1 − ( b − 1)  b ch − 2
Problem 10: Find the function whose first difference is

(i) a, (ii) ax + b, (iii) sin x, (iv) e a + bx , (v) 5 x .
Solution: Let the required function be f ( x ) and the interval of differencing h = 1. Then
(i) As ∆ f ( x ) = a (i. e., a constant), therefore f ( x ) will be a polynomial of first degree and hence
f ( x ) = ax + b, where b is a constant.
(ii) As ∆ f ( x ) = ax + b …(1)
i. e., a polynomial of first degree, so f ( x ) will be a polynomial of second degree.
Let f ( x ) = Ax2 + Bx + C.
∴ ∆ f ( x ) = [ A ( x + 1)2 + B ( x + 1) + C ] − [ Ax2 + Bx + C ], by def. for h = 1
or ∆ f ( x ) = 2 Ax + ( A + B ) …(2)
Comparing (1) and (2), we get
2 Ax + ( A + B ) ≡ ax + b
1 1
giving A= a and B=b− a.
2 2
N-9
Hence the required function is

1 2 
a x +  b − a x + C.
1
f ( x) =
2  2 
(iii) Here ∆ f ( x ) = sin x. …(1)
Let us consider f ( x ) = k sin x and h = π.
Then ∆ f ( x ) = ∆ ( k sin x ) = k sin ( x + π ) − k sin x
= k [sin ( x + π ) − sin x ]
π . cos  x + π 
1 1
= k . 2 sin
2  2 
= 2k ( − sin x ). …(2)
1
− 2k = 1 or k=− ⋅
2
1
Thus f ( x) = − sin x.
2
(iv) Here ∆ f ( x ) = e a + bx . …(1)
Let us consider f ( x ) = k . e a + bx and h = 1
so that ∆ f ( x ) = k e a + b ( x +1) − k . e a + bx
= k . e a + bx ( e b − 1). …(2)
1
k ( e b − 1) = 1 or k= ⋅
b
e −1
1
∴ f ( x) = ⋅ e a + bx
eb − 1
e a + bx
or f ( x) = ⋅
eb − 1
(v) Here ∆ f ( x) = 5 x . …(1)
Let f ( x) = k . 5 x and h =1
x +1
so that ∆ f ( x) = k 5 − k 5 = k 5 x [5 − 1] = 4k 5 x .
x
…(2)
1
4k 5 x = 5 x giving k = ⋅
4
1 x
Hence f ( x) = 5 .
4
Problem 11: Find the value of

∆
(i) [3 x(4) − 2 x(3) − 3 x(2) + x − 5]
∆ x
N-10
(ii) ∆ [2 x(−2) − 3 x(3) − 4 x (−1) ]
∆2
(iii) [3 x(4) − 5 x(−2) ]
∆ x2
(iv) Find the relation between α, β and γ in order that α + βx + γ x2 may be expressible in one term in the
factorial notation. (Kanpur 2009)
2 2
(v) Find the value of E x when the values of x vary by a constant increment of 2.
(Gorakhpur 2013)
Solution: We have
∆
(i) [3 x(4) − 2 x(3) − 3 x(2) + x − 5] = 12 x(3) − 6 x(2) − 6 x (1) + 1
∆ x
(ii) ∆ [2 x(−2) − 3 x (3) − 4 x(−1) ] = [ − 4 x (−3) − 9 x(2) + 4 x(−2) ] h
∆2 ∆ ∆ ∆
(iii) [3 x (4) − 5 x(−2) ] = [3 x(4) − 5 x(−2) ] = [12 x (3) + 10 x(−3) ]
∆ x2 ∆ x ∆ x ∆ x
= 36 x (2) − 30 x(−4).
(iv) Let f ( x ) = α + βx + γ x2 = ( a + bx )(2),
where a and b are some unknown constants.

Now ( a + bx )(2) = ( a + bx ) { a + b ( x − 1)} = ( a + bx ) ( a − b + bx )
= ( a + bx )2 − ab − b2 x
= ( a2 − ab ) + (2ab − b2 ) x + b2 x2 .
∴ α + βx + γ x2 = ( a2 − ab ) + (2ab − b2 ) x + b2 x2 .
Comparing the coefficients of various powers of x, we get

α = a2 − ab, β = 2ab − b2 , γ = b2 .
Eliminating a and b from these equations, we get

γ 2 + 4αγ = β2 ,
which is the required relation between α , β and γ.

(v) We have
E2 x2 = EE x2 = E ( x + 2)2 [ ∵ interval of differencing is 2]
= ( x + 2 + 2)2 = ( x + 4)2
= x2 + 8 x + 16.
Problem 12: A second degree polynomial passes through (0, 1), (1, 3), (2, 7) and (3, 13). Find the
polynomial.
Solution: Let us consider the 2nd degree polynomial in factorial notation i. e., let
f ( x ) = Ax(2) + Bx(1) + C.
Now the point (0, 1) ⇒ x = 0, y = 1 i. e., f ( x ) = 1 ;

N-11
(1, 3) ⇒ x = 1, f ( x ) = 3;
(2, 7) ⇒ x = 2, f ( x ) = 7
and (3, 13) ⇒ x = 3, f ( x ) = 13.
Thus the difference table for the given values of x and f ( x ) is as follows :
x f ( x) ∆ f ( x) ∆2 f ( x ) ∆3 f ( x )
0 1
2
1 3 2
4 0
2 7 2
6
3 13
Now f ( x ) = Ax(2) + Bx(1) + C.
∴ ∆ f ( x ) = 2 Ax(1) + B = 2 Ax + B.
Hence ∆ f (0 ) = B.
But ∆ f (0 ) = 2 (from table)
so that B =2.
Also ∆2 f ( x ) = 2 A.
∴ ∆2 f (0 ) = 2 A.
But from table ∆2 f (0 ) = 2.
∴ 2 A = 2 giving A = 1.
Again for x = 0, f ( 0 ) = C giving C = 1, (from table).
Thus f ( x ) = 1 . x(2) + 2 . x (1) + 1 = x ( x − 1) + 2 x + 1
= x2 + x + 1.
Aliter: We have f ( x ) = E x f (0 ) [∵ interval of differencing is 1 and so

E x f (0 ) = f (0 + x . 1) = f ( x )]
= (1 + ∆ ) x f (0 )
x ( x − 1) 2
= f (0 ) + x ∆ f (0 ) + ∆ f (0 )
1. 2
[ ∵ Third and higher differences of f ( x ) are all zero]
x ( x − 1)
=1+ x .2 + ⋅2
2
= 1 + 2 x + x ( x − 1)
= x2 + x + 1.
N-12
Problem 13: A third degree polynomial passes through the points (0, − 1), (1, 1), (2, 1) and (3, 2). Find the
polynomial.
Solution: Let f ( x ) = Ax(3) + Bx(2) + Cx(1) + D.
The difference table for the given values of x and f ( x ) is as follows :
x f ( x) ∆ f ( x) ∆2 f ( x ) ∆3 f ( x )
0 −1
2
1 1 −2
0 −1
2 1 −3
−3
3 −2
Now ∆ f ( x ) = 3 Ax(2) + 2 Bx(1) + C
= 3 Ax ( x − 1) + 2Bx + C.
∴ ∆ f (0 ) = C and from table ∆ f (0 ) = 2.
Comparing the two values, we get C = 2.
Now ∆2 f ( x ) = 6 Ax(1) + 2B = 6 Ax + 2B.
∴ ∆2 f (0 ) = 2B and ∆2 f (0 ) = − 2 (from table).
∴ 2B = − 2 giving B = − 1.
Again ∆3 f ( x ) = 6 A.
∴ ∆3 f (0 ) = 6 A
and from table ∆3 f (0 ) = − 1.
Equating the two values, we get

1
6 A = − 1 giving A = − ⋅
6
Also f (0 ) = D and from table f (0 ) = − 1.
Comparing, we get D = − 1.
Substituting the values of A, B, C and D, the required polynomial
1 (3)
f ( x) = − x − 1 . x(2) + 2 x(1) − 1
6
1
=− x ( x − 1) ( x − 2) − x ( x − 1) + 2 x − 1
6
1
= − [ x3 + 3 x2 − 16 x + 6].
6
N-13
Aliter: We have f ( x ) = E x f (0 ) [ ∵ interval of differencing is 1]
= (1 + ∆ ) x f (0 )
x ( x − 1) 2 x ( x − 1) ( x − 2) 3
= f (0 ) + x ∆ f (0 ) + ∆ f (0 ) + ∆ f (0 )
1. 2 1. 2 . 3
[ ∵ Fourth and higher differences of f ( x ) are all zero]
x ( x − 1) x ( x − 1) ( x − 2)
= −1+ x .2 + ⋅ ( − 2) + ⋅ ( − 1)
2 6
1 3
=− ( x + 3 x2 − 16 x + 6).
6
2
 ∆ Ux   ∆2U x 
Problem 14: What is the difference between   and   ⋅ If U x = x3 and interval of
 E Ux  2
 E Ux 
differencing is one, find expressions for both.
2
 ∆ Ux 
Solution: In   the first forward difference operator ∆ and the shifting operator E are to
 E Ux 
be operated on U x and then we are to find the square of the quotient of the two expressions thus
 ∆2 U x 
found while in   second forward difference operator ∆2 and shifting operator E2 are to be
2 
 E Ux 
operated on U x and then we are to find the quotient of the two expressions thus obtained. For
U x = x3 ,
2 2 2
 ∆ Ux  ∆ x3   ( x + 1)3 − x3 
  =  
 = 
 E Ux  E x3   ( x + 1)
3

2
 x3  x6 2 x3
= 1 −  =1+ − ⋅
 ( x + 1)3  ( x + 1)6 ( x + 1)3
∆2 U x ∆2 x3 ∆ [( x + 1)3 − x3 ]
While, = =
E2 U x E2 x3 [ x + 2]3
{( x + 2)3 − ( x + 1)3 } − {( x + 1)3 − x3 }

=
( x + 2)3
( x + 2)3 − 2 . ( x + 1)3 + x3
=
( x + 2)3
x3 + 8 + 3 . 2 x2 + 3 . 4 x − 2 ( x3 + 1 + 3 x2 + 3 x ) + x3
=
x3 + 8 + 6 x2 + 12 x
6x + 6
=
x3 + 6 x2 + 12 x + 8
6 ( x + 1)
= ⋅
x3 + 6 x2 + 12 x + 8
N-14
Problem 15: (i) Estimate the missing term in the following :
x : 1 2 3 4 5 6 7
y : 2 4 8 ? 32 64 128
Explain why the result differs from 16. (Kanpur 2008; Kumaun 09)
(ii) Estimate the production of cotton in the year 1935 from the data given below :
Year x : 1931 1932 1933 1934 1935 1936 1937
Production f ( x ) : 17 ⋅1 13 ⋅ 0 14 ⋅ 0 9⋅6 — 12 ⋅ 4 18 ⋅ 2
(In millions)
Solution: (i) Here six values of x and y are given, therefore we can assume the given function
expressible as a polynomial of degree five.
Hence ∆6 y = 0 i. e., ∆6 f ( x ) = 0
or ( E − I )6 f ( x ) = 0
or [E6 − 6 C1 E5 + 6 C2 E4 − 6 C3 E 3 + 6 C4 E2 − 6 C5 E + 6 C6 ] f ( x ) = 0
or E6 f ( x ) − 6 E5 f ( x ) + 15 E4 f ( x ) − 20 E3 f ( x ) + 15 E2 f ( x )
− 6 E f ( x) + f ( x) = 0
or f ( x + 6) − 6 f ( x + 5) + 15 f ( x + 4) − 20 f ( x + 3) + 15 f ( x + 2)
− 6 f ( x + 1) + f ( x ) = 0.
For x = 1, we have
f (7) − 6 f (6) + 15 f (5) − 20 f (4) + 15 f (3) − 6 f (2) + f (1) = 0
or 128 − 6 × 64 + 15 × 32 − 20 f (4) + 15 × 8 − 6 × 4 + 2 = 0
or 128 − 384 + 480 − 20 f (4) + 120 − 24 + 2 = 0
or 20 f (4) = 730 − 408 = 322.
∴ f (4) = 16 ⋅1.
Explanation: From the given values we observe that the function is of the form y = 2 x . From
this we get 24 = 16 i. e., y = 16 for x = 4. This value differs from the calculated value as we have
assumed the function as a polynomial while actually the behaviour of the given values shows that
the function is y = 2 x and hence the difference.
(ii) As six observations are given, therefore we can fit a polynomial of fifth degree and hence the
sixth difference will be equal to zero.
∴ ∆6 f ( x ) = 0
or ( E − I )6 f ( x ) = 0
or [E6 − 6 C1 E5 + 6 C2 E4 − 6 C3 E 3 + 6 C4 E2 − 6 C5 E + 6 C6 ] f ( x ) = 0
or E6 f ( x ) − 6 E5 f ( x ) + 15 E4 f ( x ) − 20 E3 f ( x ) + 15 E2
f ( x) − 6 E f ( x) + f ( x) = 0
N-15
or f ( x + 6) − 6 f ( x + 5) + 15 f ( x + 4) − 20 f ( x + 3) + 15 f ( x + 2)
− 6 f ( x + 1) + f ( x ) = 0.
Taking x = 1931, we get
f (1937) − 6 f (1936) + 15 f (1935) − 20 f (1934) + 15 f (1933)
− 6 f (1932) + f (1931) = 0
or 18 ⋅ 2 − 6 × 12 ⋅ 4 + 15 × f (1935) − 20 × 9 ⋅ 6 + 15 × 14 − 6 × 13 + 17 ⋅1 = 0
or 15 f (1935) = 99 ⋅1 giving f (1935) = 6 ⋅ 60.
Problem 16: If u x = 3 x , then prove that ∆u x = 2u x .
Solution: We have ∆u x = 2u x
∴ u x = 3 x for h = 1
∆u x = ∆3 x = 3 x + 1 − 3 x = 3 x (3 − 1) = 3 x ⋅ 2 = 2 ⋅ u x
Problem 17: If u x = x ( x − 1) ( x − 2) prove that ∆u x = 3 x ( x − 1)
Solution: We have u x = x ( x − 1) ( x − 2)
∆u x = ( x + 1) x ( x − 1) − x ( x − 1) ( x − 2)
= x ( x − 1) ( x + 1 − x + 2) = 3 x( x − 1).
Problem 18: Find a function u x for which ∆u x = x ( x − 1).
Solution: We have ∆u x = x ( x − 1) = x(2).
x(3)
∴ ux = + C, where C is an arbitrary constant
3
1
= x ( x − 1) ( x − 2) + C.
3
Problem 19: Obtain the function whose first difference is

(i) x3 + 4 x2 + 9 x + 12. (ii) x3 + 2 x2 + 5 x + 12.
(Kumaun 2009; Garhwal 10; Kanpur 11)
Solution: (i) Converting the given function into factorial notation, we have
x3 + 4 x2 + 9 x + 12 = Ax (3) + Bx(2) + Cx(1) + D.
By synthetic division method, we get

1 1 4 9 12 = D
0 1 5
2 1 5 14 = C
0 2
3 1
7=B
0
1= A
N-16
Thus the given function = x(3) + 7 x(2) + 14 x(1) + 12.
As given ∆ f ( x ) = x(3) + 7 x(2) + 14 x(1) + 12.
∴ f ( x ) = ∆−1 x(3) + 7∆−1 x(2) + 14∆−1 x(1) + ∆−1 (12)
x(4) x (3) x(2) x(1)

= + 7⋅ + 14 ⋅ + 12 ⋅ + k,
4 3 2 1
where k is a constant and the interval of differencing has been taken as 1,
x ( x − 1) ( x − 2) ( x − 3) 7 x ( x − 1) ( x − 2)
or f ( x) = + + 7 x ( x − 1) + 12 x + k
4 3
x4 5 x3 7 x2 49 x
= + − + + k.
4 6 4 6
x(4) 5 x(3)
(ii) Proceed as in part (i). Ans. + + 4 x(2) + 12 x(1) + k.
4 3
Problem 20: Obtain the missing terms in the following table :

x : 0 ⋅1 ⋅2 ⋅3 ⋅4 ⋅5 ⋅6
f ( x) : ⋅135 ? ⋅111 ⋅100 ? ⋅082 ⋅074
(Purvanchal 2009; Rohilkhand 11)
Solution: Here five values of [ x, f ( x )] are known therefore we can assume that f ( x ) can be
represented by a polynomial of degree four.
Hence ∆5 f ( x ) = 0 or ( E − I )5 f ( x ) = 0
or [ E5 − 5 C1 E4 + 5 C2 E3 − 5 C3 E2 + 5 C4 E − 5 C5 ] f ( x ) = 0
or f ( x + ⋅ 5) − 5 f ( x + ⋅ 4) + 10 f ( x + ⋅ 3) − 10 f ( x + ⋅ 2)
+ 5 f ( x + ⋅1) − f ( x ) = 0,
because here the interval of differencing h = ⋅1.
At x = 0 (first tabulated value of x), we have
f (⋅5) − 5 f (⋅4) + 10 f (⋅3) − 10 f (⋅2) + 5 f (⋅1) − f (0 ) = 0 …(1)
and for x = ⋅1 (second tabulated value of x), we have
f (⋅6) − 5 f (⋅5) + 10 f (⋅4) − 10 f (⋅3) + 5 f (⋅2) − f (⋅1) = 0 …(2)
Substituting the known values from the given data, (1) gives
⋅082 − 5 f (⋅4) + 10 × ⋅100 − 10 × ⋅111 + 5 f (⋅1) − ⋅135 = 0
or 5 f (⋅4) − 5 f (⋅1) = − ⋅163 …(3)
and (2) gives
⋅074 − 5 × ⋅ 082 + 10 f (⋅4) − 10 × ⋅100 + 5 × ⋅111 − f (⋅1) = 0
or 10 f (⋅4) − f (⋅1) = ⋅ 881. …(4)
Solving (3) and (4), we get
f (⋅1) = ⋅123 and f (⋅4) = ⋅ 090,
which are the required missing values.
N-17
Problem 21: Given U0 + U8 = 1⋅ 9243, U1 + U7 = 1⋅ 9590
U2 + U6 = 1⋅ 9823, U3 + U5 = 1⋅ 9956 Find U4 .

Solution: Since eight entries are given, we have
∆8 U0 = 0 i. e., ( E − I )8 U0 = 0
or [E8 − 8 C1 E7 + 8 C2 E6 − 8 C3 E5 + 8 C4 E4 − 8 C5 E3
+ 8 C6 E2 − 8 C7 E + 8 C8 ] U0 = 0
or U8 − 8 C1 U7 + 8 C2 U6 − 8 C3 U5 + 8 C4 U4 − 8 C5 U3
+ 8 C6 U2 − 8 C7 U1 + U0 = 0
or (U0 + U8 ) − 8 C1 (U1 + U7 ) + 8 C2 (U2 + U6 ) − 8 C3 (U3 + U5 )
+ 8 C4 U4 = 0.
Putting the given values, we get

1⋅ 9243 − 8 × 1⋅ 9590 + 28 × 1⋅ 9823 − 56 × 1⋅ 9956 + 70 U4 = 0
or 57 ⋅ 4287 − 127 ⋅ 4256 + 70 U4 = 0
or 70 U4 = 69 ⋅ 9969
or U4 = 0 ⋅ 9999.
Problem 22: Evaluate

(i) (2 ∆2 + ∆ − 1) ( x2 + 2 x + 1)
(ii) ( ∆ + 1) (2∆ − 1) ( x2 + 2 x + 1)
(iii) ( E − 2) ( E − 1) (2 x / h + x )
(iv) ( E2 − 3E + 2) (2 x / h + x )
Solution: (i) We have (2 ∆2 + ∆ − 1) ( x2 + 2 x + 1)
= 2 ∆2 ( x2 + 2 x + 1) + ∆ ( x2 + 2 x + 1) − 1 ( x2 + 2 x + 1)
= 2 ∆2 x2 + ∆ ( x2 + 2 x ) − x2 − 2 x − 1 [∵ ∆2 (2 x + 1) = 0, ∆ 1 = 0 ]
= 2∆ [( x + h)2 − x2 ] + [( x + h)2 + 2 ( x + h) − x2 − 2 x ] − x2 − 2 x − 1
= 2 ∆ (2hx + h2 ) + 2hx + h2 + 2h − x2 − 2 x − 1
= 4h ∆x + 2hx + h2 + 2h − x2 − 2 x − 1 [∵ ∆ h2 = 0 ]
= 4h ( x + h − x ) + 2hx + h2 + 2h − x2 − 2 x − 1
= 5h2 + 2hx + 2h − x2 − 2 x − 1 .
(ii) We have ( ∆ + 1) (2∆ − 1) ( x2 + 2 x + 1)
= ( ∆ + 1) [2∆ ( x2 + 2 x + 1) − 1 ( x2 + 2 x + 1)]
= ( ∆ + 1) [2 {( x + h)2 + 2( x + h) + 1 − x2 − 2 x − 1} − x2 − 2 x − 1]
= ( ∆ + 1) [2(2hx + h2 + 2h) − x2 − 2 x − 1]
N-18
= (∆ + 1)(4hx + 2h2 + 4h − x2 − 2 x − 1)
= ∆(4hx − x2 − 2 x ) + 1(4hx + 2h2 + 4h − x2 − 2 x − 1) [∵∆ (2h2 + 4h − 1) = 0 ]
= 4h ∆ x − ∆ x2 − 2∆ x + 4hx + 2h2 + 4h − x2 − 2 x − 1
= 4h( x + h − x ) − [( x + h)2 − x2 ] − 2( x + h − x ) + 4hx + 2h2 + 4h − x2 − 2 x − 1
= 4h2 − 2hx − h2 − 2h + 4hx + 2h2 + 4h − x2 − 2 x − 1
= 5h2 + 2hx + 2h − x2 − 2 x − 1.
(iii) We have ( E − 2) ( E − 1) (2 x / h + x )
= ( E − 2)[ E (2 x / h + x ) − (2 x / h + x )]
= ( E − 2)[2( x + h) / h + x + h − 2 x / h − x ]
= ( E − 2)[2 x / h(2 − 1) + h] = ( E − 2)(2 x / h + h)
= E (2 x / h + h) − 2(2 x / h + h)
= 2( x + h) / h + h − 2( x + h) / h − 2h = − h
(iv) We have ( E2 − 3E + 2) (2 x / h + x )
= E2 (2 x / h + x ) − 3E (2 x / h + x ) + 2(2 x / h + x )
= [2( x +2 h) / h + x + 2h] − 3[2( x + h) / h + x + h] + 2.2 x / h + 2 x
= 4.2 x / h + x + 2h − 3(2.2 x / h + x + h) + 2.2 x / h + 2 x = −h.
Problem 23: Use the method of separation of symbols to prove the following identities :
(i) ∆ nu x = u x + n − n C1 u x + n − 1 + nC2 u x + n − 2 + … + ( − 1)n u x .
1 2 1.3 4 1.3.5 6
(ii) Ux − ∆ U x −1 + ∆ Ux −2 − ∆ Ux −3 + …
8 8.16 8.16.24
1 1
= U x + 1 /2 − ∆U x + 1 /2 + ∆2U x + 1 /2 − …
2 4
(iii) U x = U x − 1 + ∆U x − 2 + ∆2U x − 3 + … + ∆n − 1 U x − n + ∆nU x − n.
(Rohilkhand 2010; Kumaun 10; Bundelkhand 12)
Solution: (i) L.H.S. = ∆ nu x = ( E − 1)n u x

= [ E n − nC1 E n − 1 + nC2 E n − 2 + … + ( − 1)n ] u x
= E nu x − nC1 E n − 1 u x + nC2 E n − 2 u x + … + ( − 1)n u x
= u x + n − nC1 u x + n − 1 + n C2 u x + n − 2 + … + ( − 1)n u x [∵ E n u x = u x + n ]
= R.H.S.
1 1. 3 4 1. 3 . 5 6
(ii) We have U x − ∆2 U x −1 + ∆ U x −2 − ∆ U x −3 + …
8 8 . 16 8 . 16 . 24
 1 1 . 3 4 −2 1 . 3 . 5 6 −3 
= 1 − ∆2 E −1 + ∆ E − ∆ E + … U x
 8 8 . 16 8 . 16 . 24 
N-19
  − 1  − 3
 2 −1     2 −1
2
=  1+  − 1   ∆ E  +  2   2   ∆ E 
 
  2   4  2!  4 
 

 − 1  − 3  − 5 
      3
 2   2   2   ∆2 E −1  
+  
3!  4  + … U x
  

−1 / 2 −1 / 2
 ∆2 E −1   ∆2 
= 1 +  U x = E1 /2 E +  Ux
 4   4 
−1 / 2
 ∆2 
= E1 /2 1 + ∆ +  Ux [∵ E ≡ 1 + ∆ ]
 4 
−1  
∆ ∆ ∆2
= E1 /2 1 +  U x = E1 /2 1 − + − … U x
 2  2 22

1 1
= U x +1 /2 − ∆ U x +1 /2 + ∆2 U x +1 /2 − …
2 4
(iii) We have U x − ∆nU x − n = U x − ∆nE − nU x = (1 − ∆nE − n )U x
 ∆n  E n − ∆n 1  E n − ∆n 
= 1 − U = Ux =   Ux
n x n
 E  E E n  I 
1  E n − ∆n 
=   Ux [∵ I = E − ∆ ]
E n  E − ∆ 
= E − n[ E n − 1 + ∆ E n − 2 + ∆2 E n − 3 + … + ∆n − 1 ] U x , [Dividing Nr. by Dr.]
−1 −2 2 −3 n −1 − n
= [E + ∆E +∆ E +… + ∆ E ]Ux
2 n −1
= U x − 1 + ∆U x − 2 + ∆ U x − 3 + … + ∆ U x − n.
Problem 24: Prove the following identities :

(i) U0 + U1 + U2 + … + U n = n +1 C1 U0 + n +1C2 ∆ U0 + n +1C3 ∆2 U0 + … + ∆n U0
(Kumaun 2008)
x x +1 2 x +2 3
(ii) U x +n = Un + C1 ∆ U n −1 + C2 ∆ U n−2 + C3 ∆ U n − 3 + …
(iii) ∆n U x − n = U x − nC1 U x −1 + nC2 U x −2 − nC3 U x −3 + …
U0 + U1 + U2 + … + U n = U0 + EU0 + E2 U0 + … + E n U0
E n +1 − 1
= (1 + E + E2 + … + E n ) U0 = U0 , summing the G.P.
E −1
(1 + ∆ )n +1 − 1
= U0
∆
1
= [(1 + n +1C1 ∆ + n +1C2 ∆2 + n +1C3 ∆3 + … + ∆n +1 ) − 1] U0
∆
1 n +1
= [ C1 ∆ U0 + n +1C2 ∆2 U0 + n +1C3 ∆3 U0 + … + ∆n +1 U0 ]
∆
= n+1C1 U0 + n+1C2 ∆ U0 + n+1C3 ∆2 U0 + … + ∆n U0 .
N-20
(ii) We have U n + x C1 ∆ U n−1 + x +1C2 ∆2 U n−2 + x +2 C3 ∆3 U n−3 + …
= U n + x C1 ∆ E −1 U n + x +1C2 ∆2 E −2 U n + x +2 C3 ∆3 E −3 U n + …
= [1 + x C1 ∆ E −1 + x +1C2 ∆2 E −2 + x +2 C3 ∆3 E −3 + … ] U n
= [1 − ∆ E −1 ]− x U n
−x −x
∆ E − ∆
= 1 −  Un =  Un
 E   E 
−x
=  
1
U n = E x U n = U n+ x = U x + n .
 E
(iii) We have U x − nC1 U x −1 + nC2 U x −2 − nC3 U x −3 + …
= U x − nC1 E −1 U x + nC2 E −2 U x − nC3 E −3 U x + …
= (1 − nC1 E −1 + nC2 E −2 − nC3 E −3 + … ) U x

n n
E − 1
= (1 − E −1 )n U x = 1 −  u x = 
1
Ux
 E   E 
n
∆ 
=  n −n n
 Un = ∆ E U x = ∆ U x −n .
 E 
Problem 25: Prove the following identities :
(i) U0 + x
C1∆U1 + x C2 + … = U x + x C1 ∆2 U x − 1 + x C2 ∆4 U x − 2 + … (Kumaun 2013)
(ii) U x − U x +1 + U x +2 − U x +3 + …
1
2
=
1 2 1. 3  1 ∆4 U x −5 /2
 U x −1 /2 − ∆ U x −3 /2 +  
2  8 2!  8
1. 3 . 5  1
3 
−   ∆6 U x −7 /2 + … ⋅
3 !  8 
(iii) U0 + nC1 U1 x + nC2 U2 x2 + nC3 U3 x3 + …
= (1 + x )n U0 + nC1 (1 + x )n −1 x ∆ U0 + nC2 (1 + x )n − 2 x2 ∆2U0 + …

U x + x C1 ∆2 U x −1 + x C2 ∆4 U x −2 + …
= [1 + x C1 ∆2 E −1 + x C2 ∆4 E −2 + … ] U x
x x
2 −1 x  E + ∆2   E2 − E + 1
= (1 + ∆ E ) U x =   Ux = 
 
 Ux

 E   E 
1
= [1 + E ( E − 1)] x U x = E − x (1 + ∆E ) x U x
Ex
= (1 + ∆E ) x U0 = (1 + x C1 ∆E + x C2 ∆2 E2 + … ) U0
= U0 + x C1 ∆ U1 + x C2 ∆2U2 + …
N-21
(ii) We have
1
2
1 2 1. 3  1
U x −1 /2 − ∆ U x −3 /2 +   ∆4 U x −5 /2
2  8 2 !  8
1. 3 . 5  1
3 
−   ∆6 U x −7 /2 + …
3!  8 
  1  3
    2
1  −1 / 2  1   1  2 −3 / 2     1
=  E Ux −     ∆ E U x + 2 2   ∆4 E −5 /2 U x
2  2  4 2!  4

 1  3  5 
      3 
      1
− 2 2 2   ∆6 E −7 /2 U x + …
3!  4 

  − 1  − 3
     2
 2   2   1 2 −1 
= E −1 /2 1 +  −   ∆2 E −1  +
1 1 1
 ∆ E 
2   2  4  2! 4 

 − 1  − 3  − 5 
      3 
 2   2   2   1 2 −1 
+  ∆ E  + … U x
3! 4  

−1 / 2
=
1 −1 / 2
E 1 + 1 ∆2 E −1  Ux
2  4 
−1 / 2 −1 /2
1 −1 /2  4E + ∆2  1 −1 /2  4 + 4 ∆ + ∆2 
= E   Ux = E   Ux
2  4E  2  4E 
1 −1 / 2
= E . 2E1 /2 {(2 + ∆ )2 }−1 /2 U x = (2 + ∆ )−1 U x
2
= (1 + E )−1 U x = [1 − E + E2 − E3 + E4 − E5 + … ] U x
= U x − U x +1 + U x + 2 − U x + 3 + U x + 4 − U x + 5 + …
(iii) We have, (1 + x )n U0 + nC1 (1 + x )n−1 x ∆ U0 + nC2 (1 + x )n−2 x2 ∆2 U0 + …
= {(1 + x ) + x ∆ }n U0 = {1 + x (1 + ∆ )}n U0
= (1 + x E )n U0 = [1 + nC1 x E + nC2 x2 E2 + nC3 x3 E3 + … ] U0
= U0 + nC1 x U1 + nC2 x2U2 + nC3 x3U3 + …
Problem 26(i): Prove that

∆ f [ f ( x − 1) ∆ g ( x − 1)] = ∇ [ f ( x ) ∆ g ( x )] = ∆ [ f ( x − 1) ∇ g ( x )]
(Rohilkhand 2010)
Solution: We have ∆ g ( x − 1) = g ( x ) − g ( x − 1)
∴ f ( x − 1) ∆ g ( x − 1) = f ( x − 1) g ( x ) − f ( x − 1) g ( x − 1)
⇒ ∆ [ f ( x − 1) ∆ g ( x − 1)] = ∆ [ f ( x − 1) g ( x )] − ∆ [ f ( x − 1) g ( x − 1)] …(1)
= f ( x ) g ( x + 1) − f ( x − 1) g ( x ) − f ( x ) g ( x ) + f )( x − 1) g ( x − 1)
N-22
= [ f ( x ) g ( x + 1) − f ( x − 1) g ( x )] − [ f ( x ) g ( x ) − f ( x − 1) g ( x − 1)]
= ∇ [ f ( x ) g ( x + 1)] − ∇ [ f ( x ) g ( x )]
= ∇ [ f ( x ) g ( x + 1) − f ( x ) g ( x )]
= ∇ [ f ( x ) { g ( x + 1) − g ( x )}]
= ∇ [ f ( x ) ∆ g ( x )].
Now from (1), we have
∆[ f ( x − 1) ∆ g ( x − 1)] = ∆ [ f ( x − 1) g ( x ) − f ( x − 1) g ( x − 1)]
= ∆ [ f ( x − 1){ g ( x ) − g ( x − 1)}] = ∆ [ f ( x − 1) ∇ g ( x )]
Hence, ∆ [ f ( x − 1) ∆ g ( x − 1)] = ∇[ f ( x ) ∆ g ( x )] = ∆ [ f ( x − 1) ∇ g ( x )].
1 ∆2 y ∆3 y ∆4 y
Problem 26(ii): Show that Dy = [ ∆y − + − + …].
h 2 3 4
(Garhwal 2010)
Solution: Proceed as article 10 part (f ).
Problem 27: Express the function 2 x3 − 3 x2 + 3 x − 10 in factorial notation, the interval of differencing
being unity. Also find its successive differences. (Kanpur 2010)
3 2 (3) (2) (1)
Solution: Let 2 x − 3 x + 3 x − 10 = Ax + Bx + Cx + D.
The coefficients A, B, C, D are easily obtained by the method of synthetic division. The procedure
is given below :
1 2 −3 3 − 10 = D
0 2 −1
2 2 −1 2=C
0 4
3 2 3=B
2= A
Hence f ( x ) = 2 x(3) + 3 x (2) + 2 x(1) − 10.
∴ ∆ f ( x ) = 2 . 3 x(2) + 3 . 2 x(1) + 2 = 6 x(2) + 6 x(1) + 2
= 6 x ( x − 1) + 6 x + 2 = 6 x2 + 2,
∆2 f ( x ) = 6. 2 x(1) + 6 = 12 x + 6,
and ∆3 f ( x ) = 12.
N-23
Problem 28: If
x : 1 2 3 4 5
y : 2 5 10 20 30
find by forward difference table ∆4 y (1).
Solution: The forward difference table is as follows :
x y = f ( x) ∆ f ( x) ∆2 f ( x ) ∆3 f ( x ) ∆4 f ( x )
1 2
2 5 2
5 3
3 10 5 −8
10 −5
4 20 0
10
5 30
From table we observe that ∆4 f (1) = − 8.
Problem 29: If f (0 ) = − 3, f (1) = 6, f (2) = 8 and f (3) = 12, prepare forward difference table to find
∆3 f (0 ).
Solution: Proceed as in problem 28, Ans. ∆3 f (0 ) = 9.

n ( n + 1) n n
∆n 0 n +1 = ∆ 0 .
2 (Gorakhpur 2014, 15)
n m n −1 m −1 n m −1
Solution: Using the relation ∆ 0 = n [∆ 0 + ∆ 0 ], we get
∆n 0 n +1 = n [ ∆n −1 0 n + ∆n 0 n ]
∆n −1 0 n = ( n − 1) [ ∆n − 2 0 n −1 + ∆n −1 0 n −1 ]
∆n − 2 0 n −1 = ( n − 2) [ ∆n − 3 0 n − 2 + ∆n − 2 0 n − 2 ]
… … … … …
… … … … …
∆2 03 = 2 [ ∆ 02 + ∆2 02 ]
∆ 02 = 1 [ ∆0 01 + ∆1 01 ].
N-24
By back substitution of these values, we get

∆n 0 n +1 = n ∆n 0 n + n ( n − 1) ∆n −1 0 n −1 + n ( n − 1) ( n − 2) ∆n − 2 0 n − 2
+ …+ n ( n − 1) ( n − 2) … 2 . 1 ∆1 01
= nn ! + n ( n − 1) ( n − 1) ! + n ( n − 1) ( n − 2) ( n − 2) ! + …
… + n ( n − 1) ( n − 2) … 2 . 1 . 1 !
= n ! { n + ( n − 1) + ( n − 2) + … + 2 + 1}
n ( n + 1) n ( n + 1) n n
= n !⋅ = ⋅∆ 0 . [ ∵ ∆n 0 n = n !]
2 2

u2 n − nC1 2u2 n−1 + nC2 22 u2 n−2 − … + ( −2)n un = ( −1)n ( c − 2an),
where un = an2 + bn + c. (Gorakhpur 2010)

n n 2 n
Solution: L.H.S. = u2 n − C1 2 u2 n−1 + C2 2 u2 n−2 + … + ( −2) un
= E n un − nC1 . 2 E n−1 un + nC2 22 E n−2 un − … + ( −2)n un
= [ E n − nC1 . 2E n−1 + nC2 . 22 E n−2 − … + ( −2)n ] un
= ( E − 2)n un = ( ∆ − 1)n un = ( −1)n (1 − ∆ )n un
 n ( n − 1) 2 
= ( −1)n 1 − n∆ + ∆  un [neglecting higher order differences
 1. 2 
as un is a polynomial of second degree]
n ( n − 1) 2 
= ( −1)n  un − n ∆ un + ∆ un
 2 
 n2 − n 2 
= ( −1)n ( an2 + bn + c ) − n∆ ( an2 + bn + c ) + ∆ ( an2 + bn + c )
 2 
 n2 − n 
= ( −1)n ( an2 + bn + c ) − n { a ∆ n2 + b ∆ n} + ⋅ a ∆2 n2 
 2 

= ( −1)n ( an2 + bn + c ) − n { a ( n + 1)2 − an2 }

n2 − n 
− bn ( n + 1 − n) + ⋅ a ∆ {( n + 1)2 − n2 }
2 
 n2 − n 
= ( −1)n ( an2 + bn + c ) − n (2an + a + b ) + ⋅ a ∆ (2n + 1)
 2 
 n2 − n 
= ( −1)n ( an2 + bn + c ) − n (2an + a + b ) + ⋅ a {2 ( n + 1) − 2n}
 2 
= ( −1)n [( an2 + bn + c ) − n (2an + a + b ) + a ( n2 − n)]
= ( −1)n [ c − 2an]
= R. H. S.
N-25
Problem 32: Taking fifth order differences of U x to be constant and given U0 , U1 , … , U6 , prove that
1 25 ( c − b ) + 3 ( a − c )
U = c+
21 2 256
2
where a = U0 + U5 , b = U1 + U4 , c = U2 + U3 . (Purvanchal 2009; Kumaun 10)
Solution: We have U = U5 /2 = E 5 /2 U0
21
2
= (1 + ∆ )5 /2 U0 [∵ Here, the interval of differencing h is 1]
5 5  5 5  5 
 − 1  − 1  − 2
5 2 2  2 2 2  2  3
= U0 + ∆ U0 + ∆ U0 + ∆ U0
2 2! 3!
5 5  5  5 
 − 1  − 2  − 3
2 2  2  2  4
+ ∆ U0
4!
5  5 − 1  5 − 2  5 − 3  5 − 4
       
2  2  2  2  5
+ 2 ∆ U0
5!
[ ∵ sixth and higher differences are all zero]
5 15 2 5 3 5 4 3 5
= U0 + ∆ U0 + ∆ U0 + ∆ U0 − ∆ U0 + ∆ U0
2 8 16 128 256
5 15 5
= U0 + ( E − 1) U0 + ( E − 1)2 U0 + ( E − 1)3 U0
2 8 16
5 3
− ( E − 1)4 U0 + ( E − 1)5 U0
128 256
5 15 5
= U0 + (U1 − U0 ) + ( E 2 − 2 E + 1) U0 + ( E 3 − 3E 2 + 3E − 1) U0
2 8 16
5
− ( E 4 − 4 E 3 + 6 E 2 − 4 E + 1) U0
128
3
+ [ E 5 − 5E 4 + 10 E 3 − 10 E 2 + 5E − 1] U0
256
5 15 5
= U0 + (U1 − U0 ) + (U2 − 2U1 + U0 ) + (U3 − 3U2 + 3U1 − U0 )
2 8 16
5
− (U4 − 4U3 + 6U2 − 4U1 + U0 )
128
3
+ (U5 − 5U4 + 10U3 − 10U2 + 5U1 − U0 )
256
= U0 1 − +
5 15 5 5 3   5 15 + 15 + 5 + 15 
− − −  + U1  − 
 2 8 16 128 256   2 4 16 32 256 
+ U2 
15 15 15 15   5 + 5 + 15  + U  − 5 − 15  + 3 U
− − −  + U3   4   5
 8 16 64 128   16 32 128   128 256  256
N-26
3 25 75 75 25 3
= U0 − U1 + U2 + U3 − U4 + U5
256 256 128 128 256 256
3 25 75
= (U0 + U5 ) − (U1 + U4 ) + (U2 + U3 )
256 256 128
3 25 75 3a 25b  1 11 
= a− b+ c= − +  +  c
256 256 128 256 256  2 128 
1 3a − 25b + 22 c 1 3a − 25b + 25c − 3c

= c+ = c+
2 256 2 256
1 3 ( a − c ) + 25 ( c − b )
= c+ ⋅
2 256
Problem 33: If ∆3 u x = 0, prove that

1 1 2
u x +(1 /2) = ( u x + u x +1 ) − ( ∆ u x +1 + ∆2 u x ).
2 16 (Bundelkhand 2012)
Solution: We have
u x +(1 /2) = E1 /2 u x = (1 + ∆ )1 /2 u x = 1 + ∆ − ∆2  u x ,

1 1
 2 8 
expanding upto ∆2 only because ∆3 u x = 0

1 1
= ux + ∆ u x − ∆2 u x . …(1)
2 8
Again ∆3 u x = ∆ ( ∆2 u x ) = ∆2 u x +1 − ∆2 u x .
∴ ∆3 u x = 0 ⇒ ∆2 u x +1 − ∆2 u x = 0 ⇒ ∆2 u x = ∆2 u x +1 .
Also ∆ u x = u x +1 − u x . Putting these values in (1), we get
u x +(1 /2) = u x + ( u x +1 − u x ) −  ∆2 u x + ∆2 u x 
1 1 1 1
2 8 2 2 
1 1  ∆2 u x ∆2 u x +1 
= ux + ( u x +1 − u x ) −  + 

2 8 2 2 
1 1 2
= ( u x + u x +1 ) − ( ∆ u x + ∆2 u x +1 ).
2 16
Hints to Objective Type Questions
Multiple Choice Questions

1. We have ∆ f ( x ) = f ( x + h) − f ( x ).
2. We have ∇ f ( x ) = f ( x ) − f ( x − h).
3. If f ( x ) = 2 x3 − 3 x2 + 3 x − 10,
then ∆3 f ( x ) = 2 . 3 ! . 1 = 12, if the interval of differencing is 1.

N-27
4. We have ∆3 (1 − x ) (1 − 2 x ) (1 − 3 x )
= ∆3 (1 − 6 x + 11x2 − 6 x3 )
= ( −6) . 3 ! . 1 = − 36.
5. If ∆ U x = x ( x − 1) = x (2), then
x(3) x ( x − 1) ( x − 2)
Ux = + C= + C.
3 3
6. See article 10 part ( f ), remark.
7. See article 10 part (a).

8. We have ∆n ( ax n + bx n−1 ) = a ∆n x n + b ∆n x n−1
= a . 1 . n ! . 1 + b . 0 = a n !.
9. We have ∆3 f (0 ) = ( E − I )3 f (0 )
= ( E3 − 3E2 + 3E − I ) f (0 ) = E3 f (0 ) − 3E2 f (0 ) + 3E f (0 ) − If (0 )
= f (3) − 3 f (2) + 3 f (1) − f (0 ) [ ∵ the interval of differencing h = 1]
= 12 − 3 × 8 + 3 × 6 − ( −3) = 12 − 24 + 18 + 3 = 9.
10. Refer problem 11 (i).
11. See article 5, the operator ∇.
12. See article 15, part (I), Note.

13. See article 12.
14. See article 12.
Fill in the Blank(s)

1. Let f ( x ) = c.
Then ∆ f ( x ) = f ( x + h) − f ( x ) = c − c = 0.
2. We have (1 + ∆ ) (1 − ∇) = EE −1 = 1.
3. We have E n f ( x ) = f ( x + nh), where h is the interval of differencing.
4. We have ∆2 f ( a + h) = ( E − I )2 f ( a + h)
= ( E2 − 2E + I ) f ( a + h)
= E2 f ( a + h) − 2Ef ( a + h) + I f ( a + h)
= f ( a + 3h) − 2 f ( a + 2h) + f ( a + h).

5. We have ∇2 f ( a + 2h) = ( I − E −1 )2 f ( a + 2h)
= ( I − 2 E − 1 + E − 2 ) f ( a + 2 h)
= If ( a + 2h) − 2E −1 f ( a + 2h) + E −2 f ( a + 2h)
= f ( a + 2h) − 2 f ( a + h) + f ( a).
N-28
6. We have ∆ [ f ( x ) g ( x )] = f ( x + h) g ( x + h) − f ( x ) g ( x )
= f ( x + h) g ( x + h) − f ( x + h) g ( x ) + f ( x + h) g ( x ) − f ( x ) g ( x )
= f ( x + h) [ g ( x + h) − g ( x )] + g ( x ) [ f ( x + h) − f ( x )]
= E f ( x) . ∆ g ( x) + g ( x) . ∆ f ( x)
 f ( x)  g ( x) . ∆ f ( x) − f ( x) . ∆ g ( x)
7. We have ∆  = ⋅
 g ( x)  g ( x) . E g ( x)
8. We have ∆3 = ( E − I )3 = E3 − 3E2 + 3E − I .
9. If f ( x ) = x4 − 12 x3 + 24 x2 − 30 x + 9,
then ∆4 f ( x ) = 1 . 4 ! . 1 = 24, if the interval of differencing is 1.
10. We have E −1 f ( x ) = f ( x − h), where h is the interval of differencing.
True or False
1. We know that both E and ∆ are linear operators.
Hence, the given statement is false.
2. We know that E = 1 + ∆ or ∆ = E − 1.
Hence, the given statement is true.
3. If the interval of differencing h is 1, then
∆ [ f ( x ) g ( x )] = E f ( x ) ∆ g ( x ) + g ( x ) ∆ f ( x )
= f ( x + 1) ∆ g ( x ) + g ( x ) ∆ f ( x ).
Hence, the given statement is true.
4. We have ∆ ( e x ) = e x + h − e x = e x ( e h − 1).
 ex  x  ex 
∴ ∆ =e and ∆ + C = e x .
h h
 e − 1  e − 1 
ex
So, if ∆ f ( x ) = e x , then f ( x ) = + C.
h
e −1
Hence, the given statement is false.

5. The given statement is true.
6. If the interval of differencing is unity, then ∆ x(n) = nx(n −1).
d
Also, D x n = n x n−1, where D ≡ ⋅
dx
Hence, the operator ∆ behaves on a factorial function x(n) in the same way as the
differentiation operator D behaves on x n.
So, the given statement is true.
❍❍❍
N-29
Chapter-2
Interpolation with Equal Intervals
Problem 1: What do you mean by Interpolation ? What are the underlying assumptions for the validity
of the various methods used for interpolation ?
Solution: See article 1 and article 2.
Problem 2: Derive an interpolation formula for equal intervals.

Problem 3: By constructing a difference table, find the 7th term as well as the general term of the sequence
0, 0, 2, 6, 12, 20, …
Solution: The difference table for the given data is as follows :

x f ( x) ∆ f ( x) ∆2 f ( x ) ∆3 f ( x )
1 0
0
2 0 2
2 0
3 2 2
4 0
4 6 2
6 0
5 12 2
8
6 20
Here, the interval of differencing h is 1. We have

6.5 2 6.5.4 3
the 7th term = f (7) = E6 f (1) = (1 + ∆)6 f (1) = f (1) + 6∆ f (1) + ∆ f (1) + ∆ f (1)
2! 3!
6.5 6.5.4
= 0 + 6.0 + .2 + .0 = 30
2 6
The general term = f ( r + 1) = E r f (1) = (1 + ∆ )r f (1)
r ( r − 1) 2 r ( r − 1) ( r − 2) 3
= f (1) + r∆f (1) + ∆ f (1) + ∆ f (1)
2! 3!
r ( r − 1) r ( r − 1) ( r − 2)
= 0 + r ⋅0 + .2 + .0 = r ( r − 1).
2 6
N-30
Problem 4: In an examination the number of candidates who secured marks between certain limits were as
follows :
Marks : 0 – 19 20 – 39 40 – 59 60 – 79 80 – 99
No. of candidates : 41 62 65 50 17
Estimate the number of candidates getting marks less than 70. (Kumaun 2008, 10, 12)
Solution: The forward difference table is as given below :
f ( x)
Marks less than No. of candidates securing ∆ f ( x) ∆2 f ( x ) ∆3 f ( x ) ∆4 f ( x )
x marks less than x
20 41
62
40 103 3
65 − 18
60 168 − 15 0
50 − 18
80 218 − 33
17
100 235
Obviously, the number of candidates getting marks less than 70 = f (70 )i. e., f ( x )when x = 70.
By Newton-Gregory forward difference formula, we have
u ( u − 1) 2
f ( a + uh) = f ( a) + u ∆ f ( a) + ∆ f ( a)
2!
u ( u − 1) ( u − 2) 3 u ( u − 1) ( u − 2) ( u − 3) 4
+ ∆ f ( a) + ∆ f ( a) + … .
3! 4!
Here a = 20, h = 20.

x − a 70 − 20 50
∴ for x = 70, we have u = = = = 2 ⋅ 5.
h 20 20
1 1
∴ f (70 ) = 41 + 2 ⋅ 5 × 62 + × 2 ⋅ 5 × 1⋅ 5 × 3 + × 2 ⋅ 5 × 1⋅ 5 × 0 ⋅ 5 × ( −18)
2 6
1
+ × 2 ⋅ 5 × 1 ⋅ 5 × 0 ⋅ 5 × ( − 0 ⋅ 5) × 0
24
= 41 + 155 + 5 ⋅ 625 − 5 ⋅ 625 = 196.
Hence, the number of candidates getting marks less than 70 = 196.
Problem 5: Find the polynomial which fits the data in the following table :
x : 3 5 7 9 11
y = f ( x) : 6 24 58 108 174
N-31
Solution: The forward difference table for the given data is as given below.
x y = f ( x) ∆ f ( x) ∆2 f ( x ) ∆3 f ( x )
3 6
18
5 24 16
34 0
7 58 16
50 0
9 108 16
66
11 174
Here, the first tabulated value of the argument x is 3 and the interval of differencing h is 2. We have
 x − 3 
f ( x ) = f [3 + ( x − 3)] = f 3 +   ⋅2
  2  
= E ( x −3) /2 f (3) = (1 + ∆ )( x −3) /2 f (3)
 x − 3 1  x − 3  x − 3 
= 1 +   ∆+ ⋅   − 1 ∆2  f (3)
  2  2 !  2   2  
[ ∵ Third and higher differences of f ( x ) are all zero]
x − 3 1  x − 3  x − 5 2
= f (3) +   ∆f (3) + ⋅     ∆ f (3)
 2  2  2   2 
x − 3 ( x − 3) ( x − 5)
= 6 +   ⋅ 18 + ⋅ 16
 2  8
= 6 + 9 ( x − 3) + 2 ( x − 3) ( x − 5) = 2 x2 − 7 x + 9
Problem 6: Prove the following formula due to Newton

x ( x + 1) 2
f ( x ) = f (0 ) + x∆f ( −1) + ∆ f ( −2)
2!
given values of f ( x ) for x = 0, − 1, − 2 , … .
Solution: We have
f ( x ) = E x f (0 ) [ ∵ Here, the interval of differencing h is 1]
= (1 − ∇)− x f (0 ) [ ∵ E −1 = 1 − ∇ or E = (1 − ∇)−1 ]
 x ( x + 1) 2 x ( x + 1) ( x + 2) 3 
= 1 + x ∇ + ∇ + ∇ + … f (0 )
 1 . 2 1 . 2 . 3 
x ( x + 1) 2 x ( x + 1) ( x + 2) 3
= f (0 ) + x ∇f (0 ) + ∇ f (0 ) + ∇ f (0 ) +
2! 3!
x ( x + 1) ( x + 2) … ( x + n − 1) n
…+ ∇ f (0 ) + … …(1)
n!
N-32
Now ∇r f (0 ) = (1 − E −1 )r f (0 )
r
E − 1
=  r −r
 f (0 ) = ( E − 1) E f (0 )
 E 
= ( E − 1)r f [0 + ( − r ) . 1] = ∆r f ( − r ).
∴ ∇f (0 ) = ∆f ( −1), ∇2 f (0 ) = ∆2 f ( −2),
∇2 f (0 ) = ∆3 f ( −3), … .
Hence, from (1), we have

x ( x + 1) 2
f ( x ) = f (0 ) + x ∆f ( −1) + ∆ f ( −2)
2!
x ( x + 1) ( x + 2) 3
+ ∆ f ( −3) + … .
3!
Problem7: The gamma function, Γ ( x) is tabulated for various values of x in the following table :
1⋅15 1⋅16 1⋅17 1⋅18 1⋅19
x :
Γ ( x) : 0 ⋅ 93304 0 ⋅ 92980 0 ⋅ 92670 0 ⋅ 92373 0 ⋅ 92088
Find Γ (1⋅1673).
Solution: The table of finite differences is given below :
x f ( x) = Γ ( x) ∆f ( x ) ∆2 f ( x ) ∆3 f ( x ) ∆4 f ( x )
1⋅15 ⋅93304
− ⋅ 00324
1⋅16 ⋅92980 ⋅00014

− ⋅ 00310 − ⋅ 00001
1⋅17 ⋅92670 ⋅00013 0

− ⋅ 00297 − ⋅ 00001
1⋅18 ⋅92373 ⋅00012

− ⋅ 00285
1⋅19 ⋅92088
Newton-Gregory forward interpolation formula is

f ( a + uh) = f ( a) + uC1 ∆f ( a) + uC2 ∆2 f ( a) + uC3 ∆3 f ( a) + … …(1)
Here a + uh = 1⋅1673, a = 1⋅15 and h = ⋅ 01.
1⋅1673 − 1⋅15
∴ u= = 1⋅ 73.
⋅01
∴ From (1) substituting the values from the table, we have
N-33
f (1⋅1673) = ⋅ 93304 + ( − ⋅ 00324) × 1⋅ 73
⋅00014 × 1⋅ 73 × ⋅ 73 ( − ⋅ 00001) × 1⋅ 73 × ⋅ 73 × ( − ⋅ 27)

+ +
2 ×1 3 × 2 ×1
= ⋅ 93304 − ⋅ 0056052 + ⋅ 0008840 + ⋅ 000000508 = ⋅ 92831.
Problem 8: The following are the marks obtained by 492 candidates in a certain examination.
Marks : No. of candidates
0 – 40 210
40 – 45 43
45 – 50 54
50 – 55 74
55 – 60 32
60 – 65 79
Find out the number of candidates :

(a) who secured more than 48 but not more than 50 marks,
(b) less than 48 but not less than 45 marks.
Solution: The forward difference table is as given below :
Marks f ( x)
less than No. of candidates ∆ f ( x) ∆2 f ( x ) ∆3 f ( x ) ∆4 f ( x ) ∆5 f ( x )
x securing marks
less than x
40 210
43
45 253 11
54 9
50 307 20 − 71
74 − 62 222
55 381 − 42 151
32 89
60 413 47
79
65 492
Here a = 40, h = 5.
u ( u − 1) 2
f ( a + uh) = f ( a) + u ∆ f ( a) + ∆ f ( a)
2!
u ( u − 1) ( u − 2) 3 u ( u − 1) ( u − 2) ( u − 3) 4
+ ∆ f ( a) + ∆ f ( a)
3! 4!
u ( u − 1) ( u − 2) ( u − 3) ( u − 4) 5
+ ∆ f ( a) + … . …(1)
5!
N-34
(a) Obviously, the number of candidates who secured more than 48 but not more than 50 marks
= the number of candidates who secured 49 or 50 marks
= (the number of candidates who secured less than 51 marks)
− (the number of candidates who secured less than 49 marks)
= f (51) − f (49).
To find f (51). When x = 51, we have

x − a 51 − 40 11
u= = = = 2⋅2 .
h 5 5
∴ From (1), we have
2 ⋅ 2 × 1⋅ 2
f (51) = 210 + 2 ⋅ 2 × 43 + × 11
2
2 ⋅ 2 × 1⋅ 2 × 0 ⋅ 2 2 ⋅ 2 × 1 ⋅ 2 × 0 ⋅ 2 × ( − 0 ⋅ 8)
+ ×9+ × ( − 71)
6 24
2 ⋅ 2 × 1 ⋅ 2 × 0 ⋅ 2 × ( − 0 ⋅ 8) × ( − 1 ⋅ 8)
+ × 222
120
= 210 + 94 ⋅ 6 + 14 ⋅ 52 + 0 ⋅ 792 + 1⋅ 2496 + 1⋅ 4066 = 322 ⋅ 5682 = 323.
To find f (49). When x = 49, we have
x − a 49 − 40 9
u= = = = 1⋅ 8.
h 5 5
1⋅ 8 × 0 ⋅ 8
f (49) = 210 + 1⋅ 8 × 43 + × 11
2
1 ⋅ 8 × 0 ⋅ 8 × ( − 0 ⋅ 2) 1 ⋅ 8 × 0 ⋅ 8 × ( − 0 ⋅ 2) × ( − 1 ⋅ 2)
+ ×9+ × ( − 71)
6 24
1 ⋅ 8 × 0 ⋅ 8 × ( − 0 ⋅ 2) × ( − 1 ⋅ 2) × ( − 2 ⋅ 2)
+ × 222
120
= 210 + 77 ⋅ 4 + 7 ⋅ 92 − 0 ⋅ 432 − 1⋅ 0224 − 1⋅ 4066 = 292 ⋅ 459 = 292 .
Hence, the number of candidates who secured more than 48 but not more than 50 marks
= f (51) − f (49) = 323 − 292 = 31.
(b) Obviously, the number of candidates who secured less than 48 but not less than 45 marks
= f (48) − f (45).
To find f (48): When x = 48, we have
x − a 48 − 40 8
u= = = = 1⋅ 6.
h 5 5
1⋅ 6 × 0 ⋅ 6
f (48) = 210 + 1⋅ 6 × 43 + × 11
2
1 ⋅ 6 × 0 ⋅ 6 × ( − 0 ⋅ 4) 1 ⋅ 6 × 0 ⋅ 6 × ( − 0 ⋅ 4) × ( − 1 ⋅ 4)
+ ×9+ × ( − 71)
6 24
1 ⋅ 6 × 0 ⋅ 6 × ( − 0 ⋅ 4) × ( − 1 ⋅ 4) × ( − 2 ⋅ 4)
+ × 222
120
= 210 + 68 ⋅ 8 + 5 ⋅ 28 − 0 ⋅ 576 − 1⋅ 5904 − 2 ⋅ 3869
= 279 ⋅ 5267 = 280.
N-35
Hence, the number of candidates who secured less than 48 but not less than 45 marks
= f (48) − f (45)
= 280 − 253 = 27.
Problem 9: From the following table of yearly premiums for policies maturing at quinquennial ages,
interpolate the premium for policies maturing at the age of 12 years.
Age (years) x : 10 15 20 25 30 35
Premium f ( x ) : 3 ⋅ 54 3 ⋅ 22 2 ⋅ 91 2 ⋅ 60 2 ⋅ 31 2 ⋅ 04
x f ( x) ∆ f ( x) ∆2 f ( x ) ∆3 f ( x ) ∆4 f ( x ) ∆5 f ( x )
10 3 ⋅ 54
− ⋅ 32
15 3 ⋅ 22 ⋅01
− ⋅ 31 − ⋅ 01
20 2 ⋅ 91 ⋅00 ⋅03
− ⋅ 31 ⋅02 − ⋅ 05
25 2 ⋅ 60 ⋅02 − ⋅ 02
− ⋅ 29 ⋅00
30 2 ⋅ 31 ⋅02
− ⋅ 27
35 2 ⋅ 04
Newton-Gregory formula for forward interpolation is

f ( a + uh) = f ( a) + uC1 ∆ f ( a) + uC2 ∆2 f ( a) + … …(1)
Here we have to interpolate f (12).
12 − a 12 − 10
Hence a + uh = 12 or u = = = ⋅ 4,
h 5
because as given a = first tabulated value of x = 10
and h = equal distance between given values = 5.
Hence from (1), putting u = ⋅ 4 and taking upto fifth differences as the sixth and onward
differences are zero, we get
⋅4 (⋅4 − 1) 2
f (12) = f (10 ) + ⋅ 4 ∆ f (10 ) + ∆ f (0 )
2!
⋅4 (⋅4 − 1) (⋅4 − 2) 3 ⋅4 (⋅4 − 1) (⋅4 − 2) (⋅4 − 3) 4
+ ∆ f (0 ) + ∆ f (0 )
3! 4!
⋅4 (⋅4 − 1) (⋅4 − 2) (⋅4 − 3) (⋅4 − 4) 5
+ ∆ f (0 )
5!
N-36
or f (12) = 3 ⋅ 54 + ⋅ 4 ( − ⋅ 32) − ⋅12 (⋅01) + ⋅ 064 ( − ⋅ 01) − ⋅ 0416 (⋅03)

+ ⋅ 02995 ( − ⋅ 05)
= 3 ⋅ 54 − ⋅128 − ⋅ 0012 − ⋅ 00064 − ⋅ 001248 − ⋅ 0001497
= 3 ⋅ 54 − ⋅131237 = 3 ⋅ 4087
or f (12) = 3 ⋅ 41 (rounded up value).
Problem 10: Estimate the sale for 1966 using the following table :
Year : 1931 1941 1951 1961 1971 1981
Sale in thousands : 12 15 20 27 39 52
(Lucknow 2010)
x − 1931
Solution: Let us introduce a new variable u given by u = ;
10
∴ u takes the values 0, 1, 2 , 3, 4.
The forward difference table is as under :
x u f ( u) ∆ f ( u) ∆2 f ( u) ∆3 f ( u) ∆4 f ( u) ∆5 f ( u)
1931 0 12
3
1941 1 15 2
5 0
1951 2 20 2 3
7 3 −10
1961 3 27 5 −7
12 −4
1971 4 39 1
13
1981 5 52
By Newton-Gregory formula, we have

u ( u − 1) 2 u ( u − 1) ( u − 2) 3
f ( u) = f (0 ) + u ∆ f (0 ) + ∆ f (0 ) + ∆ f (0 ) +…
2! 3!
x − a 1966 − 1931 35
Here u= = = = 3 ⋅ 5.
h 10 10
3⋅5 × 2⋅5 3 ⋅ 5 × 2 ⋅ 5 × 1⋅ 5
Thus f (1966) = 12 + 3 ⋅ 5 × 3 + ×2+ ×0
2 6
3 ⋅ 5 × 2 ⋅ 5 × 1⋅ 5 × ⋅ 5 3 ⋅ 5 × 2 ⋅ 5 × 1 ⋅ 5 × ⋅ 5 × ( − ⋅ 5)
+ ×3 + × ( − 10 )
24 120
= 12 + 10 ⋅ 5 + 8 ⋅ 75 + ⋅ 82031 + ⋅ 2734
= 32 ⋅ 3437.
N-37
Problem 11: Given the following table
x : 0 ⋅1 0 ⋅2 0 ⋅3 0 ⋅4 0 ⋅5
x 1⋅10517 1⋅ 2140 1⋅ 34986 1⋅ 49182 1⋅ 64872
e :
Find (a) e 0⋅243 , (b) e 0⋅411.
x f ( x) = e x ∆ f ( x) ∆2 f ( x ) ∆3 f ( x ) ∆4 f ( x )
⋅1 1⋅10517
⋅10883
⋅2 1⋅ 2140 ⋅02703
⋅13586 − ⋅ 02093
⋅3 1⋅ 34986 ⋅00610 ⋅02977

⋅14196 ⋅00884
⋅4 1⋅ 49182 ⋅01494
⋅15690
⋅5 1⋅ 64872
(a) Here a = ⋅1, h = ⋅1 and to interpolate e 0⋅243 we take

x − a ⋅243 − ⋅1
x = ⋅ 243 so that u = = = 1⋅ 43.
h ⋅1
f ( a + uh) = f ( a) + uC1 ∆ f ( a) + uC2 ∆2 f ( a) + … …(1)
u ( u − 1) 2
Thus e 0⋅243 = f (⋅1) + u ∆ f (⋅1) + ∆ f (⋅1)
2
u ( u − 1) ( u − 2) 3 u ( u − 1) ( u − 2) ( u − 3) 4
+ ∆ f (⋅1) + ∆ f (⋅1) + 0
6 24
1
or e 0⋅243 = 1⋅10517 + ⋅10883 × 1⋅ 43 + × ⋅ 02703 × 1⋅ 43 × ⋅ 43
2
1
+ × ( − ⋅ 02093) × 1⋅ 43 × ⋅ 43 × ( − ⋅ 57)
6
1
+ × ⋅ 02977 × 1⋅ 43 × ⋅ 43 × ( − ⋅ 57) × ( − 1⋅ 57)
24
= 1⋅10517 + ⋅1556269 + ⋅ 00831037 + ⋅ 0012226 + ⋅ 00068256 = 1⋅ 2701.
(b) To interpolate e 0⋅411 we take x = ⋅ 411 also a = ⋅1, h = ⋅1 so that

x − a ⋅411 − ⋅1
u= = = 3 ⋅11.
h ⋅1
N-38
∴ From (1),
1
e 0⋅411 = 1⋅10517 + ⋅10883 × 3 ⋅11 + × ⋅ 02703 × 3 ⋅11 × 2 ⋅11
2
1
+ × ( − ⋅ 02093) × 3 ⋅11 × 2 ⋅11 × 1⋅11
6
1
+
× ⋅ 02977 × 3 ⋅11 × 2 ⋅11 × 1⋅11 × ⋅11
24
= 1⋅10517 + ⋅ 3384613 + ⋅ 0886867 − ⋅ 0254087 + ⋅ 0008042 = 1⋅ 5083.
Problem 12: Given log x for x = 310, 320, 330, 340, 350 and 360 according to the logarithmic table.
Find the value of log 3375.
Solution: Obtaining values of log x for given values of x from logarithmic table, the difference
table is as under :
x f ( x ) = log x ∆ f ( x) ∆2 f ( x ) ∆3 f ( x ) ∆4 f ( x ) ∆5 f ( x )
310 2 ⋅ 49136
⋅01378
320 2 ⋅ 50515 − ⋅ 000424

⋅01336 ⋅0000255
330 2 ⋅ 5185 − ⋅ 000398 ⋅0000025

⋅0130 ⋅0000230 0
340 2 ⋅ 5315 − ⋅ 000375 ⋅0000025

⋅0126 ⋅0000205
350 2 ⋅ 5441 − ⋅ 000354

⋅01223
360 2 ⋅ 5563
First we find f (337 ⋅ 5).
Here a = 310, h = 10 ;
x − a 337 ⋅ 5 − 310
∴ u= = = 2 ⋅ 75.
h 10
2 ⋅ 75 × 1⋅ 75 2
∴ log 337 ⋅ 5 = f ( a) + 2 ⋅ 75 × ∆ f ( a) + ∆ f ( a)
2!
2 ⋅ 75 × 1⋅ 75 × ⋅ 75 3 2 ⋅ 75 × 1⋅ 75 × ⋅ 75 × ( − ⋅ 25) 4
+ ∆ f ( a) + ∆ f ( a)
3! 4!
2 ⋅ 75 × 1⋅ 75
= 2 ⋅ 49136 + 2 ⋅ 75 × ⋅ 01378 + × ( − ⋅ 00042),
2
leaving further terms as their value is very small
= 2 ⋅ 49136 + ⋅ 03792 − ⋅ 00101 = 2 ⋅ 52829.
N-39
Now log 3375 = log (337 ⋅ 5 × 10 ) = log 337 ⋅ 5 + log 10

= 2 ⋅ 52829 + 1⋅ 0000 = 3 ⋅ 52829 approx.
Problem 13: The following table gives the amount of a chemical dissolved in water :
Temp : 10° 15° 20° 25° 30° 35°
Solubility : 19 ⋅ 97 21⋅ 51 22 ⋅ 47 23 ⋅ 52 24 ⋅ 65 25 ⋅ 89
Compute the amount dissolved at 22° and 8°.
x° f ( x) ∆ f ( x) ∆2 f ( x ) ∆3 f ( x ) ∆4 f ( x ) ∆5 f ( x )
(in degree)
10 19 ⋅ 97
1⋅ 54
15 21⋅ 51 − ⋅ 58
⋅96 ⋅67
20 22 ⋅ 47 ⋅09 − ⋅ 68
1⋅ 05 − ⋅ 01 ⋅72
25 23 ⋅ 52 ⋅08 ⋅04
1⋅13 ⋅03
30 24 ⋅ 65 ⋅11
1⋅ 24
35 25 ⋅ 89
Newton-Gregory interpolation formula is

f ( a + uh) = f ( a) + uC1 ∆ f ( a) + uC2 ∆2 f ( a) + … …(1)
To evaluate f (22° ).
22 − 10
Here a + uh = 22 gives u = = 2 ⋅ 4.
5
− ⋅ 58 × 2 ⋅ 4 × 1⋅ 4 ⋅67 × 2 ⋅ 4 × 1⋅ 4 × ⋅ 4
∴ f (22° ) = 19 ⋅ 97 + 2 ⋅ 4 × 1⋅ 54 + +
2 ×1 3 × 2 ×1
− ⋅ 68 × 2 ⋅ 4 × 1⋅ 4 × ⋅ 4 × ( − ⋅ 6) ⋅72 × 2 ⋅ 4 × 1⋅ 4 × ⋅ 4 × ( − ⋅ 6) × ( − 1⋅ 6)
+ +
4 × 3 × 2 ×1 5 × 4 × 3 × 2 ×1
= 19 ⋅ 97 + 3 ⋅ 696 − ⋅ 9744 + ⋅15008 + ⋅ 022848 + ⋅ 00677

= 22 ⋅ 87.
Similarly we can find f (8° ) = 18 ⋅ 79.

N-40
Problem 14: Find the value of sin 30 ° 15′ 30 ′ ′ from the following table :
Angle x ° : 30 31 32 33 34
sin x ° : ⋅5000 ⋅5150 ⋅5299 ⋅5446 ⋅5592
x f ( x) ∆ f ( x) ∆2 f ( x ) ∆3 f ( x ) ∆4 f ( x )
30 ⋅5000
⋅0150
31 ⋅5150 − ⋅ 0001
⋅0149 − ⋅ 0001
32 ⋅5299 − ⋅ 0002 ⋅0002

⋅0147 ⋅0001
33 ⋅5446 − ⋅ 0001
⋅0146
34 ⋅5592
Newton-Gregory forward interpolation formula is

f ( a + uh) = f ( a) + uC1 ∆ f ( a) + uC2 ∆2 f ( a) + … …(1)
Here we want to interpolate f (30 ° 15′ 30 ′ ′ ) = f ( a° + uh° ), say.
∴ a° + uh° = 30 ° 15′ 30 ′ ′
or 30 ° + u . 1° = 30 ° 15′ 30 ′ ′ [∵ here a = 30 ° and h = 1° ]
31
= 30 degrees
120
31
giving u= ⋅
120
− ⋅ 0001  31   31
− 1
31
f (30 ° 15′ 30 ′ ′ ) = ⋅ 5 + ⋅ 0150 × +   
120 2!  120   120 
− ⋅ 0001  31   31
− 1  − 2
31
+ ×  
3!  120   120   120 
⋅0002  31   31
− 1  − 2  − 3 + 0
31 31
+   
4 !  120   120   120   120 
= ⋅ 5 + ⋅ 003875 + ⋅ 000009 − ⋅ 000005225 − ⋅ 000000685
= ⋅ 50387 = ⋅ 5039.
Problem 15: Given sin 45° = 0 ⋅ 7071, sin 50 ° = 0 ⋅ 7660
sin 55° = 0 ⋅ 8192 , sin 60 ° = 0 ⋅ 8660.

Find sin 52° by using any method of interpolation.
(Kumaun 2007, 10, 12; Purvanchal 09; Kanpur 11)
N-41
Solution: Here, we have
x: 45° 50 ° 55° 60 °
f ( x) : 0⋅7071 0⋅7660 0⋅8192 0⋅8660
The difference table for the given data is as follows :
x f ( x) ∆ f ( x) ∆2 f ( x ) ∆3 f ( x )
45° 0 ⋅ 7071
0 ⋅ 0589
50 ° 0⋅7660 −0 ⋅ 0057
0⋅0532 −0 ⋅ 0007
55° 0⋅8192 −0 ⋅ 0064

0⋅0468
60 ° 0⋅8660
We want f (52° ) = f ( a + uh), say.

7
∴ 52° = a + uh ⇒ 52° = 45° + u × 5° ⇒ u = = 1⋅ 4.
5
By Newton’s forward interpolation formula, we get
u ( u − 1) 2 u ( u − 1) ( u − 2) 3
f ( a + uh) = f ( a) + u ∆ f ( a) + ∆ f ( a) + ∆ f ( a).
2! 3!
(1⋅ 4) (0 ⋅ 4) 2
∴ f (52° ) = f (45° ) + 1⋅ 4 ∆ f (45° ) + ∆ f (45° )
2!
(1⋅ 4) (0 ⋅ 4) ( − 0 ⋅ 6) 3
+ ∆ f (45° )
3!
(1⋅ 4) (0 ⋅ 4)
= 0 ⋅ 7071 + 1⋅ 4 × 0 ⋅ 0589 + ( − 0 ⋅ 0057)
2!
(1⋅ 4) (0 ⋅ 4) ( − 0 ⋅ 6)
+ × ( − 0 ⋅ 0007)
3!
= 0 ⋅ 7071 + 0 ⋅ 08246 − 0 ⋅ 001596 + 0 ⋅ 0000392 = 0 ⋅ 7880032 .

Thus sin 52° = 0 ⋅ 7880032 = 0 ⋅ 7880 approx.
Problem 16: From the following table, for what value of x, y is minimum? Also find this value of y.
x : 3 4 5 6 7 8
y = f ( x) : 0.205 0.240 0.259 0.260 0.250 0.224
Solution: The difference table of the given data is as follows :

N-42
x f ( x) ∆ f ( x) ∆2 f ( x ) ∆3 f ( x ) ∆4 f ( x )
3 0.205
0.035
4 0.240 − 0.016
0.019 0.001
5 0.259 − 0.016 − 0.001
0.003 0.001
6 0.260 − 0.015 0.000
– 0.012 0.001
7 0.250 − 0.014
– 0.026
8 0.224
Here a = 3 and h = 1.
By Newton’s forward interpolation formula, we get
u ( u − 1) ( u − 2) 2 u ( u − 1) 3
f ( x ) = f ( a + uh) = f ( a) + u∆ f ( a) + ∆ f ( a) + ∆ y,
2! 3!
neglecting the higher order differences
u( u − 1)
f ( x ) = f ( a + uh) = 0 ⋅ 205 + u(0 ⋅ 035) − ( − 0 .016) …(1)
2
Differentiating (1) w.r.t. u, we get
2u − 1
f ′( x ) = hf ′ ( a + uh) = 0 ⋅ 035 + ( − 0 ⋅ 016)
2
For maximum or minimum, we must have f ′( x ) = 0
2u − 1
0 ⋅ 035 + ( − 0 ⋅016) = 0
2
or u = 2 ⋅6875
∴ x = a0 + uh = 3 + 2 ⋅ 6875 × 1 = 5 ⋅ 6875
Hence f ( x ) is minimum at x = 5.6875.
Putting u = 26865
. in (1), we get the minimum value of f ( x )
1
f ( x ) = 0 ⋅ 205 + (2 ⋅ 6875)(0 ⋅ 035) + (2 ⋅ 6875)(2 ⋅ 6875 − 1)( − 0 ⋅ 016)
2
= 0 ⋅ 2628.

1. Here the first tabulated value of the argument x is 0 and the interval of differencing h is 1.
Form the forward difference table. By Newton-Gregory forward difference formula, we
have
f ( x ) = E x f (0 ) = (1 + ∆ ) x f (0 )
x ( x − 1) 2
= f (0 ) + x ∆ f (0 ) + ∆ f (0 ) + …
2!
N-43
2. See article 3. We have

1
f ( x0 , x1 , … , xn ) = ∆n f ( x0 ).
n ! hn
3. In Newton-Gregory formula for backward interpolation, the value of A 0 is f ( a + nh).

1. If f ( x ) is a polynomial of degree n in x then the ( n + 1)th difference of f ( x ) is zero.
2. If f ( x ) is a polynomial of degree n in x then the nth difference of f ( x ) is constant.
3. Newton-Gregory formula for forward interpolation is used for equal intervals.
True or False
1. See article 1.
2. See article 1.
3. See article 1.
❍❍❍
N-44
Chapter-3
Interpolation with Unequal
Intervals of the Argument
Problem 1 (a): Given that f (0 ) = 8, f (1) = 68, f (5) = 123, construct a divided difference table,
using the table determine the value of f (2). (Kumaun 2012)
Solution: The divided difference table is as given below :
x f ( x) ∆| f ( x ) ∆| 2 f ( x )
0 8
68 − 8
= 60
1−0
1 68 13 ⋅ 75 − 60
= − 9 ⋅ 25
123 − 68 5−0
= 13 ⋅ 75
5 −1
5 123
By Newton’s divided difference formula, we get

f ( x ) = f ( x0 ) + ( x − x0 ) f ( x0 , x1 ) + ( x − x0 ) ( x − x1 ) f ( x0 , x1, x2 ).
∴ f (2) = 8 + (2 − 0 ) × 60 + (2 − 0 ) (2 − 1) ( − 9 ⋅ 25) = 109 ⋅ 50.
Problem 1 (b): If f (0 ) = 2 , f (1) = 3, f (2) = 12 , f (5) = 147, prepare Newton’s divided

difference table to find f (4).
Solution: Proceed as in problem 1(a). The answer is 78.
Problem 2: If log 2 = 0 ⋅ 30103, log 3 = 0 ⋅ 47712 ,

log 5 = 0 ⋅ 62897, log 7 = 0 ⋅ 84510,
find the value of log 4 ⋅ 7 to four places of decimal.
x log x = f ( x ) ∆| f ( x ) ∆| 2 f ( x ) ∆| 3 f ( x )
2 ⋅30103
⋅17609
3 ⋅47712 − ⋅ 0333883
⋅075925 ⋅0082846
5 ⋅62897 ⋅008035
⋅108065
7 ⋅84510
N-45
By Newton’s divided difference formula, we have

f (4 ⋅ 7) = ⋅ 30103 + (4 ⋅ 7 − 2) × ⋅17609 + (4 ⋅ 7 − 2) (4 ⋅ 7 − 3)
× ( − ⋅ 0333883) + (4 ⋅ 7 − 2) (4 ⋅ 7 − 3) (4 ⋅ 7 − 5) × ⋅ 0082846
i. e., log (4 ⋅ 7) = ⋅ 30103 + ⋅ 475443 − ⋅153252297 − ⋅ 011407985
= 0 ⋅ 611812 .
1
Problem 3: Show that nth order divided difference f ( x0 , x1, x2 , … , x n ) for the function ux = is equal
x
( − 1)n
to .
x0 x1 x2 … x n (Kumaun 2013)
Solution: We have
f ( x1 ) − f ( x0 ) 1 1 1  1  x0 − x1 
f ( x0 , x1 ) = =  − =  
x1 − x0 x1 − x0  x1 x0  x1 − x0  x0 x1 
( − 1) ( x1 − x0 ) ( − 1)
= = …(1)
( x1 − x0 ) x0 x1 x0 x1
f ( x1, x2 ) − f ( x0 , x1 )
Again f ( x0 , x1, x2 ) =
x2 − x0
1  ( −1) ( −1) 
=  − , [Using (1)]
x2 − x0  x1 x2 x0 x1 
( −1)  x0 − x2  ( − 1)2 ( x2 − x0 )
=  =
x2 − x0  x0 x1 x2  ( x2 − x0 ) x0 x1 x2
( − 1)2
= …(2)
x0 x1 x2
f ( x1, x2, x3 ) − f ( x0 , x1, x3 )

Now f ( x0 , x1, x2 , x3 ) =
x3 − x0
1  ( −1)2 ( − 1)2 
=  − , [(using (2)]
x3 − x0  x1 x2 x3 x0 x1 x2 
( −1)2 ( x0 − x3 ) ( −1)3
= = .
( x3 − x0 ) x0 x1 x2 x3 x0 x1 x2 x3
Proceeding in this way, we get

( −1)n
f ( x0 , x1, x2 , … , x n ) = ⋅
x0 x1 x2 … x n
Problem 4(i): From the following table find f ( x) in powers of ( x − 3):
x : 5 11 27 34 42
f ( x) : 23 899 17315 35606 68510

N-46
x f ( x) ∆| f ( x ) ∆| 2 f ( x ) ∆| 3 f ( x ) ∆| 4 f ( x )
5 23
146
11 899 40
1
1026
27 17315 69 0
1
2613
34 35606 100 0
1
4113
42 68510 a 0
1
b
3 c a′ 0
1
b′
3 c a′ ′
b′ ′
3 c
From the above table we observe that ∆| 3 f ( x ) = constant, therefore f ( x ) is expressible as a

polynomial of degree 3. As f ( x ) is to be expressed as a polynomial in powers of ( x − 3) therefore
argument x = 3 is to be repeated three times in the divided difference table. From the extended
divided difference table, we have
a − 100 a′ − a
= 1 ⇒ a = 76 ; = 1 ⇒ a′ = 45 ;
3 − 27 3 − 34
a′ ′ − a′ b − 4113
= 1 ⇒ a′ ′ = 6 ; = a ⇒ b = 1757 ;
3 − 42 3 − 34
b′ − b c − 68510
= a′ ⇒ b ′ = 2 ; = b ⇒ c = − 13.
3 − 42 3 − 42
Now using the Newton’s divided difference formula,

f ( x ) = f ( x0 ) + ( x − x0 ) f ( x0 , x1 ) + ( x − x0 ) ( x − x1 ) f ( x0 , x1 , x2 )
+ ( x − x0 ) ( x − x1 ) ( x − x2 ) f ( x0 , x1 , x2 , x3 )
taking x0 = x1 = x2 = x3 = 3,
we have f ( x ) = f (3) + ( x − 3) f (3, 3) + ( x − 3)2 f (3, 3, 3)
= c + ( x − 3) b ′ + ( x − 3)2 a′ ′ + ( x − 3)3 × 1
= − 13 + ( x − 3) . 2 + ( x − 3)2 . 6 + ( x − 3)3
= ( x − 3)3 + 6 ( x − 3)2 + 2 ( x − 3) − 13,
which is the required polynomial in powers of ( x − 3).

N-47
Problenm 4(ii): The observed values of a function are respectively 168, 120, 72 and 63 at the four
positions 3, 7, 9 and10 of the independent variable. What is the best estimate you can give for the value of the
function at the position 6 of the independent variable ? (Purvanchal 2011; Bundelkhand 12)
x f ( x) ∆| f ( x ) ∆| 2 f ( x ) ∆| 3 f ( x )
3 168
120 − 168
= − 12
7−3
7 120 − 24 + 12
= −2
72 − 120 9−3
= − 24
9−7 5+2
=1
10 − 3
9 72 − 9 + 24
=5
63 − 72 10 − 7
= −9
10 − 9
10 63
Hence f ( x ) = f ( x0 ) + ( x − x0 ) ∆| f ( x0 ) + ( x − x0 ) ( x − x1 ) ∆| 2 f ( x0 )
+ ( x − x0 ) ( x − x1 ) ( x − x2 ) ∆| 3 f ( x0 )
= 168 + ( x − 3) ( − 12) + ( x − 3) ( x − 7) ( −2) + ( x − 3) ( x − 7) ( x − 9) (1)
= x3 − 21 x2 + 119 x − 27.
When x = 6, the estimate of the function is given by

f (6) = 63 − 21 . 62 + 119 . 6 − 27
= 147.
Problem 5: The following table gives the normal weights of babies during the first twelve months of life :
Age in months ( x ) : 0 2 5 8 10 12
1 1
Weight in lbs f ( x ) : 7 10 15 16 18 21
2 4
Estimate the weight of the baby at the age of 7 months.
Solution: Here x0 = 0, x1 = 2 , x2 = 5, x3 = 8, x4 = 10, x5 = 12 .

By Lagrange’s formula for interpolation, we have
N-48
(7 − 2) (7 − 5) (7 − 8) (7 − 10 ) (7 − 12) 15
f (7) = ×
(0 − 2) (0 − 5) (0 − 8) (0 − 10 ) (0 − 12) 2
(7 − 0 ) (7 − 5) (7 − 8) (7 − 10 ) (7 − 12) 41
+ ×
(2 − 0 ) (2 − 5) (2 − 8) (2 − 10 ) (2 − 12) 4
(7 − 0 ) (7 − 2) (7 − 8) (7 − 10 ) (7 − 12)
+ × 15
(5 − 0 ) (5 − 2) (5 − 8) (5 − 10 ) (5 − 12)
(7 − 0 ) (7 − 2) (7 − 5) (7 − 10 ) (7 − 12)
+ × 16
(8 − 0 ) (8 − 2) (8 − 5) (8 − 10 ) (8 − 12)
(7 − 0 ) (7 − 2) (7 − 5) (7 − 8) (7 − 12)
+ × 18
(10 − 0 ) (10 − 2) (10 − 5) (10 − 8) (10 − 12)
(7 − 0 ) (7 − 2) (7 − 5) (7 − 8) (7 − 10 )
+ × 21
(12 − 0 ) (12 − 2) (12 − 5) (12 − 8) (12 − 10 )
15 287 175 63 21
= − +5+ − +
128 384 12 16 32
= ⋅11718 − ⋅ 7473 + 5 + 14 ⋅ 583 − 3 ⋅ 937 + ⋅ 652 = 15 ⋅ 67 lbs.
Problem 6: Determine by Lagrange’s formula the percentage number of criminals under 35 years :
Age % numbers of criminals

Under 25 years 52
Under 30 years 67⋅3
Solution: Here x = 35, x0 = 25, x1 = 30, x2 = 40, x3 = 50 by Lagrange’s formula, we have

(35 − 30 ) (35 − 40 ) (35 − 50 )
y35 = × 52
(25 − 30 ) (25 − 40 ) (25 − 50 )
(35 − 25)(35 − 40 )(35 − 50 )
+ × 67⋅ 3
(30 − 25)(30 − 40 )(30 − 50 )
((35 − 25)(35 − 30 )(35 − 50 )

+ × 84 ⋅1
(40 − 25)(40 − 30 )(40 − 50 )
(35 − 25)(35 − 30 )(35 − 40 )

+ × 94 ⋅ 4
(50 − 25)(50 − 30 )(50 − 40 )
(5)( −5)( −15) (10 )( −5)( −15)

= × 52 + × 67 ⋅ 3
( −5)( −15)( −25) (5)( −10 )( −20 )
(10 )(5)( −15) (10 )(5)( −5)

+ × 84 ⋅1 + × 94 ⋅ 4
(15)(10 )( −10 ) (25)(20 )(10 )
= −10 ⋅ 4 + 50 ⋅ 475 + 42 ⋅ 05 − 4 ⋅ 72
= 77 ⋅ 405 = 77%
N-49
Problem 7: Find the form of the function y = U x given that

U0 = 8, U1 = 11, U4 = 68, U5 = 123.
x Ux ∆| U x ∆| 2 U x ∆| 3 U x
0 8
11 − 8
=3
1−0
1 11 19 − 3
=4
68 − 11 4−0
= 19
4 −1 9−4
=1
5−0
4 68 55 − 19
=9
123 − 68 5 −1
= 55
5−4
5 123

U x = 8 + ( x − 0 ) . 3 + ( x − 0 ) ( x − 1) . 4 + ( x − 0 ) ( x − 1) ( x − 4) . 1
= 8 + 3 x + 4 x2 − 4 x + x3 − 5 x2 + 4 x = x3 − x2 + 3 x + 8.
Problem 8: Using Lagrange’s formula find the form of the function given by f ( x) :
x : 3 2 1 −1
f ( x) : 3 12 15 −21
(Kanpur 2008)
3 2
Solution: Proceed as in Ex. 10. Ans. x − 9 x + 17 x + 6.
Problem 9: By means of Lagrange’s formula, prove that approximately
( y1 − y−1 ) −  ( y3 − y1 ) − ( y−1 − y−3 ) ⋅

1 1 1 1
y0 =
2 8  2 2 
(Rohilkhand 2002, 09; Kumaun 08; Lucknow 09)
Solution: Here we are required to find y0 i. e., the value of y for x = 0 and given y−3 , y−1 ,
y1 , y3 i. e., the values of y x for x = − 3, − 1, 1 and 3 i. e., x0 = − 3, x1 = − 1, x2 = 1 and x3 = 3.
Using Lagrange’s interpolation formula, we have
N-50
(0 + 1) (0 − 1) (0 − 3) (0 + 3) (0 − 1) (0 − 3)
y0 = ⋅ y− 3 + ⋅ y− 1
( −3 + 1) ( −3 − 1) ( −3 − 3) ( −1 + 3) ( −1 − 1) ( −1 − 3)
(0 + 3) (0 + 1) (0 − 3) (0 + 3) (0 + 1) (0 − 1)
+ ⋅ y1 + ⋅ y3
(1 + 3) (1 + 1) (1 − 3) (3 + 3) (3 + 1) (3 − 1)
1 9 9 1
=− y− 3 + y− 1 + y1 − y3
16 16 16 16
y−3 +  +
1 1 1 1 1 1
=−  y +  +  y − y
16  2 16  −1  2 16  1 16 3
1 1
= ( y1 + y−1 ) − [( y3 − y1 ) − ( y−1 − y−3 )]
2 16
( y1 + y−1 ) −  ( y3 − y1 ) − ( y−1 − y−3 ) ⋅

1 1 1 1
=
2 8  2 2 
Problem 10: Find approximately the real root of the equation

y3 − 2 y − 5 = 0
Solution: Let x = y3 − 2 y − 5 so that we get y = f ( x). We want the value of f(0 ).

Values of f ( x ) at y = 1⋅ 9, 2⋅0, 2⋅1, 2⋅2 are − 1⋅ 941, − 1⋅ 000, 0 ⋅ 061, 1⋅ 248 respectively.
Now
x : –1⋅941 –1⋅000 0⋅061 1⋅248
y = f ( x) : 1⋅9 2⋅0 2⋅1 2⋅2
The difference table for the given data is as follows :
x y = f ( x) ∆| f ( x ) ∆| 2 f ( x ) ∆| 3 f ( x )
– 1⋅941 1⋅9
0⋅1062699
– 1⋅000 2⋅0 – 0⋅0060035
0⋅0942507 0⋅0004869
0⋅061 2⋅1 – 0⋅0044505
0⋅0842459
1⋅248 2⋅2
By Newton’s divided difference formula, we get

f ( x ) = f ( x0 ) + ( x − x0 ) ∆| f ( x0 ) + ( x − x0 )( x − x1 ) ∆| 2 f ( x0 )
+ ( x − x0 ) ( x − x1 ) ( x − x2 ) ∆| 3 f ( x0 )
= 1⋅ 9 + ( x + 1⋅ 941) (0 ⋅ 01062699) + ( x + 1⋅ 941) ( x + 1) ( − 0 ⋅ 0060035)

+ ( x + 1⋅ 941) ( x + 1) ( x − 0 ⋅ 061) (0 ⋅ 0004869) …(1)
Putting x = 0 in (1), we get
N-51
f(0 ) = 1⋅ 9 + 0 ⋅ 2062698 − 0 ⋅ 0116527 − 0 ⋅ 0000576

= 2 ⋅ 09455
Problem 11: If y35⋅0 = 1175, y35⋅5 = 1280, y39⋅5 = 2180, y40⋅5 = 2420, find y40 by
divided differences.
x yx ∆| y x ∆| 2 y x ∆| 3 y x
35 1175
1280 − 1175
= 210
35 ⋅ 5 − 35
35 ⋅ 5 1280 225 − 210

= 3 ⋅ 33
39 ⋅ 5 − 35
2180 − 1280
= 225
39 ⋅ 5 − 35 ⋅ 5 3 − 3 ⋅ 33
= − ⋅ 06
40 ⋅ 5 − 35
39 ⋅ 5 2180 240 − 225

=3
40 ⋅ 5 − 35 ⋅ 5
2420 − 2180
= 240
40 ⋅ 5 − 39 ⋅ 5
40 ⋅ 5 2420
By Newton’s divided difference formula for arguments x0 , x1 , x2 , x3 , we get

y x = y ( x0 ) + ( x − x0 ) y ( x0 , x1 ) + ( x − x0 ) ( x − x1 ) y ( x0 , x1 , x2 )
+ ( x − x0 ) ( x − x1 ) ( x − x2 ) y ( x0 , x1, x2 , x3 ).
Substituting the values from the table and taking x = 40, we get
y40 = 1175 + (40 − 35) × 210 + (40 − 35) (40 − 35 ⋅ 5) × 3 ⋅ 33
+ (40 − 35)(40 − 35 ⋅ 5) (40 − 39 ⋅ 5) × ( − ⋅ 06)

= 1175 + 1050 + 74 ⋅ 925 − ⋅ 675
= 2299 ⋅ 25.
Problem 12: Apply Lagrange’s formula to find f (1⋅ 50 ), using the following values of the function
f ( x) :
x : 1⋅ 00 1⋅ 20 1⋅ 40 1⋅ 60 1⋅ 80 2 ⋅ 00
f ( x) : 0 ⋅ 2420 0 ⋅1942 0 ⋅1497 0 ⋅1109 0 ⋅ 0790 0 ⋅ 0540
Solution: Using the Lagrange’s interpolation formula for the given values, we have
N-52
(1⋅ 5 − 1⋅ 2) (1⋅ 5 − 1⋅ 4) (1⋅ 5 − 1⋅ 6) (1⋅ 5 − 1⋅ 8) (1⋅ 5 − 2 ⋅ 0 )

f (1⋅ 50 ) = × ⋅ 2420
(1 − 1⋅ 2) (1 − 1⋅ 4) (1 − 1⋅ 6) (1 − 1⋅ 8) (1 − 2)
(1⋅ 5 − 1) (1⋅ 5 − 1⋅ 2) (1⋅ 5 − 1⋅ 6) (1⋅ 5 − 1⋅ 8) (1⋅ 5 − 2)
+ × ⋅1497
(1⋅ 4 − 1) (1⋅ 4 − 1⋅ 2) (1⋅ 4 − 1⋅ 6) (1⋅ 4 − 1⋅ 8) (1⋅ 4 − 2)
(1⋅ 5 − 1) (1⋅ 5 − 1⋅ 2) (1⋅ 5 − 1⋅ 4) (1⋅ 5 − 1⋅ 8) (1⋅ 5 − 2)

+ × ⋅1109
(1⋅ 6 − 1) (1⋅ 6 − 1⋅ 2) (1⋅ 6 − 1⋅ 4) (1⋅ 6 − 1⋅ 8) (1⋅ 6 − 2)
(1⋅ 5 − 1) (1⋅ 5 − 1⋅ 2) (1⋅ 5 − 1⋅ 4) (1⋅ 5 − 1⋅ 6) (1⋅ 5 − 2)

+ × ⋅ 0790
(1⋅ 8 − 1) (1⋅ 8 − 1⋅ 2) (1⋅ 8 − 1⋅ 4) (1⋅ 8 − 1⋅ 6) (1⋅ 8 − 2)
(1⋅ 5 − 1) (1⋅ 5 − 1⋅ 2) (1⋅ 5 − 1⋅ 4) (1⋅ 5 − 1⋅ 6) (1⋅ 5 − 1⋅ 8)

+ × ⋅ 0540.
(2 − 1) (2 − 1⋅ 2) (2 − 1⋅ 4) (2 − 1⋅ 6) (2 − 1⋅ 8)
⋅00001089 ⋅00014565 ⋅000336825 ⋅000249525

= − + +
⋅0384 ⋅00768 ⋅00384 ⋅00384
⋅00005925 ⋅0000243
− +
⋅00768 ⋅0384
= ⋅ 000283593 − ⋅ 018964843 + ⋅ 087714843
+ ⋅ 064980468 − ⋅ 007714843 + ⋅ 000632812
= 0 ⋅12693.
Problem 13: Apply Lagrange’s formula to find the cubic polynomial which includes the following values of
x and y x :
x : 0 1 4 6
yx : 1 −1 1 −1
Solution: By Lagrange’s interpolation formula for the given values, we have
( x − 1) ( x − 4) ( x − 6) ( x − 0 ) ( x − 4) ( x − 6)
yx = ⋅1 + ⋅ ( −1)
(0 − 1) (0 − 4) (0 − 6) (1 − 0 ) (1 − 4) (1 − 6)
( x − 0 ) ( x − 1) ( x − 6) ( x − 0 ) ( x − 1) ( x − 4)
+ ⋅1 + ⋅ ( −1)
(4 − 0 ) (4 − 1) (4 − 6) (6 − 0 ) (6 − 1) (6 − 4)
x3 − 11 x2 + 34 x − 24 x3 − 10 x2 + 24 x x3 − 7 x2 + 6 x x3 − 5 x2 + 4 x
= − + −
−24 15 −24 60
17 x3 29 x2 207 x 5
=− + + + ⋅
120 24 120 4
Problem 14: Given the following data, find f ( x) as a polynomial in powers of ( x − 5), by extending the
table to include argument x = 5 repeated as many times as may be necessary :
f (0 ) = 4, f (2) = 26, f (3) = 58, f (4) = 112 , f (7) = 466, f (9) = 922 .
Also find f ′ (5), f ′ ′(5) and f ′ ′ ′(5).
N-53
x f ( x) ∆| f ( x ) ∆| 2 f ( x ) ∆| 3 f ( x )
0 4
26 − 4
= 11
2−0
2 26 32 − 11
=7 11 − 7
3−0 =1
58 − 26 4−0
= 32
3−2
3 58 54 − 32
= 11 16 − 11
4−2 =1
112 − 58 7−2
= 54
4−3
4 112 118 − 54
= 16 22 − 16
7−3 =1
466 − 112 9−3
= 118
7−4
7 466 228 − 118
= 22 a − 22
9−4 =1
922 − 466 5−4
= 228
9−7
9 922 b − 228
=a
c − 922 5−7
=b
5−9 a′ − a
=1
5−7
5 c b′−b
= a′
5−9
a′ ′ − a′
b′ =1
5−9
5 c a′ ′
b ′′
5 c
From divided difference table we observe that the third divided differences of f ( x ) are constant,
therefore f ( x ) is a polynomial of degree 3. As f ( x ) is to be expressed as a polynomial in powers of
( x − 5) therefore argument x = 5 is to be repeated three times in the divided difference table.
In the extended divided difference table the values of c , b, b ′, a, a′, a′ ′ are obtained as follows :
I. Complete the column V by the third difference of f ( x ) which is constant and 1 in this
case.
a − 22
II. From column V, =1 giving a = 23
5−4
a′ − a a′ − 23
then =1 or =1 giving a′ = 21
5−7 −2
a′ ′ − a′ a′ ′ − 21
and =1 or =1 giving a′ ′ = 17.
5−9 −4
N-54
III. From column IV,

b − 228 b − 228
= a or = 23 giving b = 182
5−7 −2
b′ − b b ′ − 182
then = a′ or = 21 giving b′ = 98.
5−9 −4
IV. From column III,

c − 922 c − 922
=b or = 182 giving c = 194.
5−9 −4
Now the Newton divided difference formula with arguments x0 , x1 , x2 , x3 is

f ( x ) = f ( x0 ) + ( x − x0 ) f ( x0 , x1 ) + ( x − x0 ) ( x − x1 ) f ( x0 , x1 , x2 )
+ ( x − x0 ) ( x − x1 ) ( x − x2 ) f ( x0 , x1 , x2 , x3 ).
Putting x0 = x1 = x2 = x3 = 5, we get
f ( x ) = f (5) + ( x − 5) f (5, 5) + ( x − 5)2 f (5, 5, 5) + ( x − 5)3 f (5, 5, 5, 5)
= c + ( x − 5) b ′ + ( x − 5)2 a′ ′ + ( x − 5)3 × 1
= 194 + 98 ( x − 5) + 17 ( x − 5)2 + ( x − 5)3 ,
which is the required polynomial in powers of ( x − 5).

Also f ′(5) = f (5, 5) = b ′ = 98 f ′ ′ (5) = 2 ! . f (5, 5, 5) = 2 ! . a′ ′ = 2 × 17 = 34
and f ′ ′ ′(5) = 3 ! f (5, 5, 5, 5) = 3 ! . 1 = 6.
Problem 15: In the following table h is the height above the sea level and p the barometric pressure ;
calculate p when h = 5280.
h : 0 4763 6942 10594
p : 27 25 23 20
Solution: Here x0 = 0, x1 = 4763, x2 = 6942 , x3 = 10594

and f ( x0 ) = 27, f ( x1 ) = 25, f ( x2 ) = 23, f ( x3 ) = 20.
Applying Lagrange’s interpolation formula taking x = 5280, we have
(5280 − 4763) (5280 − 6942) (5280 − 10594)
f (5280 ) = × 27
(0 − 4763) (0 − 6942) (0 − 10594)
(5280 − 0 ) (5280 − 6942) (5280 − 10594)
+ × 25
(4763 − 0 ) (4763 − 6942) (4763 − 10594)
(5280 − 0 ) (5280 − 4763) (5280 − 10594)
+ × 23
(6942 − 0 ) (6942 − 4763) (6942 − 10594)
(5280 − 0 ) (5280 − 4763) (5280 − 6942)
+ × 20
(10594 − 0 ) (10594 − 4763) (10594 − 6942)
or f (5280 ) = ⋅ 344347 + 19 ⋅ 26325 + 6 ⋅ 03776 − ⋅ 402204

= 24 ⋅ 80.
N-55
Problem 16: Apply Lagrange’s formula to find f (5) given that

f (1) = 2 , f (2) = 4, f (3) = 8, f (4) = 16, f (7) = 128
and explain why the result differs from 25 .
Solution: Given x0 = 1, x1 = 2 , x2 = 3, x3 = 4, x4 = 7
and f ( x0 ) = 2 , f ( x1 ) = 4, f ( x2 ) = 8, f ( x3 ) = 16, f ( x4 ) = 128.
∴ From Lagrange’s formula for interpolation, we have
(5 − 2) (5 − 3) (5 − 4) (5 − 7) (5 − 1) (5 − 3) (5 − 4) (5 − 7)
f (5) = ×2+ ×4
(1 − 2) (1 − 3) (1 − 4) (1 − 7) (2 − 1) (2 − 3) (2 − 4) (2 − 7)
(5 − 1) (5 − 2) (5 − 4) (5 − 7) (5 − 1) (5 − 2) (5 − 3) (5 − 7)
+ ×8+ × 16
(3 − 1) (3 − 2) (3 − 4) (3 − 7) (4 − 1) (4 − 2) (4 − 3) (4 − 7)
(5 − 1) (5 − 2) (5 − 3) (5 − 4)
+ × 128
(7 − 1) (7 − 2) (7 − 3) (7 − 4)
2 32 128 128
=− + − 24 + +
3 5 3 15
= − ⋅ 66 + 6 ⋅ 4 − 24 + 42 ⋅ 66 + 8 ⋅ 53
= 57 ⋅ 59 − 24 ⋅ 66 = 32 ⋅ 93.
From the given values of function we observe that the function is of the form 2 x i. e., f ( x ) = 2 x
giving f (5) = 25 = 32 which differs from the interpolated value 32 ⋅ 93 due to the fact that for
applying Lagrange’s formula we have assumed the given function to be expressible as a
polynomial while actually the given function is of the form 2 x .
Problem 17: Four equidistant values u−1 , u0 , u1 and u2 being given, a value is interpolated by
Lagrange’s formula. Show that it may be written in the form
( y2 − 1) 2 x ( x2 − 1) 2
u x = yu0 + xu1 + y ∆ u−1 + ∆ u0 ,
3! 3!
where x + y = 1. (Rohilkhand 2011)
Solution: We have
∆2 u−1 = ( E − 1)2 u−1= ( E 2 − 2 E + 1) u−1
= u1 − 2u0 + u−1 .
Similarly,
∆2 u0 = ( E 2 − 2E + 1) u0 = u2 − 2u1 + u0 .
Now the R.H.S.

y ( y2 − 1) 2 x ( x2 − 1) 2
= yu0 + xu1 + ∆ u−1 + ∆ u0
3! 3!
(1 − x ) x ( x − 2)
= (1 − x ) u0 + xu1 + ( u1 − 2u0 + u−1 )
3!
x ( x + 1) ( x − 1)
+ ( u2 − 2u1 + u0 )
3!
N-56
x ( x − 1) ( x − 2) x ( x − 1) ( x − 2) ( x − 1) x ( x + 1)
=− u−1 +  − ( x − 1) + + u
6  3 6  0
x ( x − 1) ( x − 2) ( x − 1) x ( x + 1) ( x − 1) x ( x + 1)
+ x − − u1 + u2
 6 3  6
x ( x − 1) ( x − 2) ( x − 2) ( x − 1) ( x + 1)
=− u−1 + u0
6 2
( x − 2) x ( x + 1) ( x − 1) x ( x + 1)
− u1 + u2 .
2 6
…(1)
For u−1 , u0 , u1 , u2 as known values, by Lagrange’s formula, we have
( x − 0 ) ( x − 1) ( x − 2) ( x + 1) ( x − 1) ( x − 2)
ux = u−1 + u0
( −1 − 0 ) ( −1 − 1) ( −1 − 2) (0 + 1) (0 − 1) (0 − 2)
( x + 1) ( x − 0 ) ( x − 2) ( x + 1) ( x − 0 ) ( x − 1)
+ u1 + u2
(1 + 1) (1 − 0 ) (1 − 2) (2 + 1) (2 − 0 ) (2 − 1)
x ( x − 1) ( x − 2) ( x + 1)( x − 1) ( x − 2)
=− u−1 + u0
6 2
( x + 1) x ( x − 2) ( x + 1) x ( x − 1)
− u1 + u2 .
2 6
…(2)
From (1) and (2), we have
y ( y2 − 1) 2 x ( x2 − 1) 2
u x = y u0 + xu1 + ∆ u−1 + ∆ u0 .
3! 3!
Problem 18: Prove that the Lagrange’s formula can be put in the form
n φ ( x ) f ( xr )
Pn ( x ) = Σ
r = 0 ( x − xr ) φ′ ( xr )
n
where φ ( x ) = Π ( x − xr ).
r =0 (Rohilkhand 2011)
Solution: We can write Lagrange’s formula as

n ( x − x0 ) ( x − x1 ) … ( x − xn )
Pn ( x ) = Σ f ( xr )
r =0 ( xr − x0 ) ( xr − x1 ) … ( xr − xn )
 φ ( x ) f ( xr )
n  1 
= Σ    ( x − x ) ( x − x ) … ( x − x ) ⋅
r = 0  ( x − xr )   r 0 r 1 r n 
n
Now φ ( x ) = Π ( x − xr ).
r =0
∴ φ′( xr ) = [ φ′( x )] x = x r
= ( xr − x0 ) ( xr − x1 ) … ( xr − xr −1 )( xr − xr +1 ) … ( xr − xn ).
n φ ( x ) f ( xr )
∴ Pn ( x ) = Σ ⋅
r = 0 ( x − xr ) φ′ ( xr )
N-57
Objective Type Questions

1. We know that the nth divided difference can be expressed as the quotient of two
determinants each of order n + 1. So, the third divided difference can be expressed as the
quotient of two determinants each of order 4.
2. An interpolation formula in which both the function and its first derivative are to be
assigned at each point of interpolation is Hermite’s interpolation formula.
3. If f ( x ) = x n, then the first divided difference f ( xr , xr +1 ) is a homogeneous expression in
xr , xr +1 of degree n − 1.
4. We have f (2) = 23 − 2 × 2 = 4, f (4) = 43 − 2 × 4 = 56.
f (4) − f (2) 56 − 4
Now f (2 , 4) = = = 26.
4−2 2
5. See article 2.

1. The first divided difference of f ( x ) for the arguments x0 , x1 is defined as
f ( x1 ) − f ( x0 )
f ( x0 , x1 ) = ⋅
x1 − x0
2. The nth divided difference can be expressed as the quotient of two determinants each of
order ( n + 1).
3. See article 4.
4. See article 7.
f ( b ) − f ( a) (1 / b2 ) − (1 / a2 )
5. We have f ( a, b ) = =
b−a b−a
a2 − b2 (a − b) (a + b) a+ b
= = =−
a b2 ( b − a)
2
a2 b2 ( b − a) a2 b2
6. See problem 18.

7. See article 8.
True or False
1. We know that Lagrange’s formula can be used for both equal and unequal intervals.
2. We know that divided differences are symmetric functions of their arguments. So, the
value of any divided difference is independent of the order of the arguments.
f ( b ) − f ( a)
, then ∆|   =
1 1
3. If f ( x) =
x b  a  b−a
(1 / b ) − (1 / a) a−b 1
= = =− ⋅
b−a ab ( b − a) ab
N-58
f ( x1 ) − f ( x 0 )
4. We have f ( x0 , x1 ) = ⋅
x1 − x 0
5. We know that the nth divided differences of a polynomial of degree n are constant.
6. We know that the divided differences can be expressed as the product of multiple integrals.
7. The given formula is Newton’s divided differences interpolation formula and not
Lagrange’s interpolation formula.
❍❍❍
N-59
Chapter-4
Central Difference Interpolation Formulae
1 2
(i) δ = ∆ E −1 /2 = ∇ E1 /2 (ii) µ 2 = 1 + δ
4
(Rohilkhand 2011;
Gorakhpur 11; Bundelkhand 11; Purvanchal 12)
1 1
(iii) E 1 /2 = µ + δ (iv) E −1 /2 = µ − δ
2 2
(Rohilkhand 2011)
1
(v) √ (1 + δ2µ 2 ) = 1 + δ2 (vi) δ2 y0 = y1 − 2 y0 + y−1
2
(Kumaun 2012)
n n!
n k
(vii) δ y x = Σ ( −1) y x +(n /2)− k .
k =0 k ! (n − k ) !

1 1
δf ( x ) = f ( x + h) − f ( x − h), by def. of the operator δ
2 2
= E 1 /2 f ( x ) − E −1 /2 f ( x ), where E = 1 + ∆
= ( E 1 /2 − E −1 /2 ) f ( x ).
∴ δ = E 1 /2 − E −1 /2 = ( E − 1) E −1 /2
= ∆ E −1 /2 . [ ∵ E = 1 + ∆ or E − 1 = ∆ ]
Again δ = E1 /2 − E −1 /2 ⇒ δ = (1 − E −1 ) E 1 /2 .
∴ δ = ∇ E 1 /2 [∵ 1 − E −1 = ∇]
Hence δ = ∆ E −1 /2 = ∇ E 1 /2 .
(ii) We have δ = E 1 /2 − E −1 /2
1
and µ= ( E 1 /2 + E −1 /2 ).
2
1 1
∴ µ2 = ( E 1 /2 + E −1 /2 )2 = [( E 1 /2 − E −1 /2 )2 + 4 E 1 /2 E −1 /2 ]
4 4
1
= [ 4 + ( E 1 /2 − E −1 /2 )2 ]
4
1 1
= (4 + δ2 ) = 1 + δ2 .
4 4
N-60
(iii) & (iv). We have δ = E1 /2 − E −1 /2 .

1 1 1
∴ δ = E 1 /2 − E −1 /2 . ...(1)
2 2 2
1
Also µ = ( E 1 /2 + E −1 /2 )
2
1 1 /2 1 −1 /2
= E + E …( 2 )
2 2
Adding (1) and (2), we get
1
µ+ δ = E 1 /2 .
2
Subtracting (1) from (2), we get
1
µ− δ = E −1 /2 .
2
1 2
(v) First prove that µ 2 = 1 + δ .
4
1
Now √ (1 + δ2µ 2 ) = √ [1 + δ2 (1 + δ2 )]
4
2 1 4
= √ [1 + δ + δ ]
4
1 2 2 1
= √ [(1 + δ ) ] = 1 + δ2 .
2 2
(vi) We have
δ2 y0 = ( E 1 /2 − E −1 /2 )2 y0
= ( E − 2 E 1 /2 E −1 /2 + E −1 ) y0
= ( E − 2 + E −1 ) y0
= Ey0 − 2 y0 + E −1 y0
= y1 − 2 y0 + y−1.
(vii) We have δ n y x = ( ∆ E −1 /2 )n y x [∵ δ = ∆E −1 /2 ]
= ∆ n E − n /2 y x
= ∆ n y x −(n /2)
= ( E − 1)n y x −(n /2)

 n 
=  Σ ( −1)k nCk E n− k  y x −(n /2)
 k = 0 
n  n! 
= Σ ( −1)k E n− k y x −(n /2) 
k =0  k ! (n − k ) ! 
n n!
= Σ ( −1)k y x −(n /2)+ n− k
k =0 k ! (n − k ) !
n n!
= Σ ( −1)k y x +(n /2)− k .
k =0 k ! (n − k ) !
N-61
Problem 2: Establish the following relations :

1 1 1
(i) µδ = ( ∇ + ∆ ) = ∆ E −1 + ∆
2 2 2 (Kumaun 2009; Rohilkhand 11)
(ii) ∆ + ∇ = ∆ / ∇ − ∇ / ∆.
Solution: (i) We have δ = E 1 /2 − E −1 /2
1
and µ= ( E 1 /2 + E −1 /2 ).
2
1
∴ µδ= ( E 1 /2 + E −1 /2 ) ( E 1 /2 − E −1 /2 )
2
1
= [ ( E 1 /2 )2 − ( E −1 /2 )2 ]
2
1
= ( E − E −1 )
2
1
= [(1 − E −1 ) + ( E − 1)]
2
1
= ( ∇ + ∆ ). [∵ E = 1 + ∆ and E −1 = 1 − ∇ ]
2
1
Again µ δ = [(1 − E −1 ) + ( E − 1)], as proved above
2
1 1
= (1 − E −1 ) + ( E − 1)
2 2
1 1
= ( E − 1) E −1 + ∆ [∵ E − 1 = ∆ ]
2 2
1 1
=∆E −1 + ∆.
2 2
1 1 1
Hence µ δ = ( ∇ + ∆ ) = ∆E −1 + ∆.
2 2 2
(ii) We have ∆ + ∇ = ( E − 1) + (1 − E −1 )
= E − E −1 ...(1)
−1
∆ ∇ E −1 1− E
Also − = −
∇ ∆ 1 − E −1 E −1
E (1 − E −1 ) E −1 ( E − 1)
= −
−1 E −1
1− E
= E − E −1 ...(2)
From (1) and (2), we have ∆ + ∇ = ∆ / ∇ − ∇ / ∆.
Problem 3: If D, E, δ and µ be the operators with usual meanings and if hD = U , where h is the interval of
differencing , prove that
(i) E = eU (ii) e − U = 1 − ∇
(iii) δ = 2sinh (U / 2) (iv) µ = cosh(U / 2)

(v) ( E + 1) δ = 2 µ ( E − 1). (Kumaun 2013)
N-62

E f ( x ) = f ( x + h)
h h2 h3
= f ( x) + f ′ ( x) + f ′ ′ ( x) + f ′ ′ ′( x ) + …
1! 2! 3!
h h2 2 h3 3
= 1 f ( x) + Df ( x ) + D f ( x) + D f ( x) + …
1! 2! 3!
 hD ( hD)2 ( hD)3 
= 1 + + + + … f ( x )
 1! 2! 3! 
= e hD f ( x ).
∴ E = e hD = e U , where hD = U .
(ii) We have eU = E, as proved above.
∴ ( e U )−1 = E −1
or e −U = 1 − ∇ [∵ E −1 = 1 − ∇]
(iii) & (iv) We have E = eU , as proved above.

∴ E 1 /2 = e U /2 and E −1 /2 = e −U /2 .
∴ E 1 /2 − E −1 /2 = e U /2 − e −U /2
or δ = 2 sinh (U / 2).
Also E 1 /2 + E −1 /2 = e U /2 + e −U /2
1 1
or ( E 1 /2 + E −1 /2 ) = ( e U /2 + e −U /2 )
2 2
or µ = cosh (U / 2).
(v) We have ( E + 1) δ = ( E 1 /2 + E −1 /2 ) E 1 /2 δ
1
=2[ ( E 1 /2 + E −1 /2 ) ]E 1 /2 δ
2
= 2µ E 1 /2 δ
= 2µE 1 /2 ( E 1 /2 − E −1 /2 )
= 2µ ( E − 1).
Problem 4: Given y2 = 10, y1 = 8, y0 = 5, y−1 = 10, find y1 /2 by Gauss forward formula.
Solution: Here 0 is origin and h = 1. We require the value of y for u = 1 / 2.

The difference table is given below :
u yu ∆ yu ∆ 2 yu ∆ 3 yu
−1 10
−5
0 5 8
3 −9
1 8 −1
2
2 10
N-63
By Gauss forward formula, we have

u ( u − 1) 2 ( u + 1) u ( u − 1) 3
yu = y0 + u ∆ y0 + ∆ y− 1 + ∆ y− 1 .
2 6
1 1 3 1 1
. ( − 1) . . (− )
1
∴ y1 /2 = 5 + . 3 + 2 2 .8 + 2 2 2 . ( −9)
2 2 6
3 9
= 5 + −1+
2 16
= 5 ⋅ 5 + 0 ⋅ 56 = 6 ⋅ 06 = 6 approximately.
Problem 5: Given that :

√ (12500 ) = 111⋅ 803399 ; √ (12510 ) = 111⋅ 848111 ;
√ (12520 ) = 111⋅ 892806 ; √ (12530 ) = 111⋅ 937483.
Show by Gauss backward formula that
√ (12516) = 111⋅ 874930 . (Kanpur 2009)
Solution: Let us take 12520 as the origin and h = 10 as the unit.

x − x0
We know u = .
h
Here x = 12516, x0 = 12520.
12516 − 12520
∴ u= = − 0 ⋅ 4.
10
We require the value of y for u = −0 ⋅ 4.
u yu ∆ yu ∆ 2 yu ∆ 3 yu
−2 111⋅ 803399
0 ⋅ 044712
−1 111⋅ 848111 −0 ⋅ 000017

0 ⋅ 044695 −0 ⋅ 000001
0 111⋅ 892806 −0 ⋅ 000018

0 ⋅ 044677
1 111⋅ 937483
By Gauss backward formula, we have

( u + 1) u 2 ( u + 1) u ( u − 1) 3
yu = y0 + u ∆ y−1 + ∆ y− 1 + ∆ y− 2 .
2 6
∴ y−0⋅4 = 111⋅ 892806 + ( −0 ⋅ 4) (0 ⋅ 044695)
(0 ⋅ 6) ( −0 ⋅ 4)
+ ( −0 ⋅ 000018)
2
(0 ⋅ 6) ( −0 ⋅ 4) ( −1⋅ 4)
+ ( −0 ⋅ 000001)
6
= 111⋅ 892806 − 0 ⋅ 017878 + ⋅ 000002 + negligible term
= 111⋅ 874930.
N-64
Problem 6: Apply (i) Stirling’s formula , (ii) Bessel’s formula (iii) Gauss forward formula,
(iv) Laplace-Everett’s formula to find the value of f (1⋅ 22) from the following table which gives the values of
1 2
1 t − t
f ( t) =
∫
√ (2 π ) 0
e 2 dt at intervals of t = 0 ⋅ 5 from t = 0 to t = 3 :
t f ( t)
0 ⋅ 00 0 ⋅ 000
0 ⋅ 50 0 ⋅191
1⋅ 00 0 ⋅ 341
1⋅ 50 0 ⋅ 433
2 ⋅ 00 0 ⋅ 477
2 ⋅ 50 0 ⋅ 494
3 ⋅ 00 0 ⋅ 499
Solution: Take 1⋅ 00 as the origin and 0 ⋅ 5 as the unit, thus the new values of t are
1⋅ 22 − 1⋅ 00
−2 , − 1, 0, 1, 2 , 3, 4. When t = 1⋅ 22 , the new value of t = = 0 ⋅ 44 .
0 ⋅5
Construct the following difference table :
t f ( t) ∆ f ( t) ∆2 f ( t) ∆3 f ( t) ∆4 f ( t) ∆5 f ( t) ∆6 f ( t)
−2 0 ⋅ 000
0 ⋅191
−1 0 ⋅191 −0 ⋅ 041
0 ⋅150 −0 ⋅ 017
0 0 ⋅ 341 −0 ⋅ 058 0 ⋅ 027
0 ⋅ 092 0 ⋅ 010 −0 ⋅ 016
1 0 ⋅ 433 −0 ⋅ 048 0 ⋅ 011 −0 ⋅ 001
0 ⋅ 044 0 ⋅ 021 −0 ⋅ 017
2 0 ⋅ 477 −0 ⋅ 027 −0 ⋅ 006
0 ⋅ 017 0 ⋅ 015
3 0 ⋅ 494 −0 ⋅ 012
0 ⋅ 005
4 0 ⋅ 499
(i) The Stirling’s formula in this case is

t ∆ f (0 ) + ∆ f ( −1) t 2
f ( t) = f (0 ) + + ∆ f ( −1)
1! 2 2!
t ( t 2 − 12 ) ∆3 f ( −1) + ∆3 f ( −2) t 2 ( t 2 − 12 ) 4
+ + ∆ f ( −2)
3! 2 4!
⋅44  0 ⋅150 + 0 ⋅ 092  (⋅44)2

∴ f (0 ⋅ 44) = 0 ⋅ 341 +   + ( − 0 ⋅ 058)
1!  2  2!
⋅44 (⋅442 − 1) − 0 ⋅ 017 + 0 ⋅ 010 (⋅44)2 (⋅442 − 1)

+ ⋅ + (0 ⋅ 027)
3! 2 4!
N-65
(⋅44) (⋅8064) (⋅027)

= 0 ⋅ 341 + ⋅ 44 (0 ⋅121) − (⋅1936) (0 ⋅ 029) +
12
(⋅1936) (⋅8064) (0 ⋅ 027)
−
24
= 0 ⋅ 341 + ⋅ 44 (0 ⋅121) − (⋅1936) (0 ⋅ 029) + (⋅44) (⋅0672) (⋅007)
− (⋅0968) (⋅0672) (⋅0027)
= 0 ⋅ 341 + ⋅ 0532 − ⋅ 0056 + ⋅ 0002 − ⋅ 00018
= 0 ⋅ 3886 approx.
(b) The Bessel’s formula is
t ( t − 1) ∆2 f (0 ) + ∆2 f ( −1)
f ( t) = [ f (0 ) + f (1)] +  t −  ∆ f (0 ) −
1 1
2  2 2! 2
 t − 1  t ( t − 1)
 
 2 ( t + 1) t ( t − 1) ( t − 2) ∆4 f ( −1) + ∆4 f ( −2)
+ ∆3 f ( −1) + ⋅
3! 4! 2
 t − 1  ( t + 1) t ( t − 1) ( t − 2)
 
 2
+ ∆5 f ( −2)
5!
f (0 ⋅ 44) = [0 ⋅ 341 + 0 ⋅ 433] + ⋅44 −  (0 ⋅ 092)
1 1
∴
2  2 
 ⋅44 − 1  (⋅44) (⋅44 − 1)
 
⋅44 (⋅44 − 1) − 0 ⋅ 058 − 0 ⋅ 048  2
+ × + (0 ⋅ 010 )
2! 2 3!
(⋅44 + 1) (⋅44) (⋅44 − 1) (⋅44 − 2) 0 ⋅ 027 + 0 ⋅ 011
+ ×
4! 2
 ⋅44 − 1  (⋅44 + 1) (⋅44) (⋅44 − 1) (⋅44 − 2)
 
 2
+ ( −0 ⋅ 016)
5!
= 0 ⋅ 387 − (⋅06) (0 ⋅ 092) + (⋅22) (⋅28) (⋅160 ) + (⋅01) (⋅44) (⋅58) (0 ⋅ 010 )
+ (⋅06) (⋅44) (⋅56) (1⋅ 56) + (0 ⋅ 019) + (⋅06) (⋅44) (⋅56) (1⋅ 56) (0 ⋅ 0032)
= 0 ⋅ 388 approx.
(c) The Gauss’s formula is
f ( t) = f (0 ) + t C1 ∆ f (0 ) + t C2 ∆2 f ( −1) + t +1C3 ∆3 f ( −1)
+ t +1C4 ∆4 f ( −2) + t +1C5 f ( −2).
(⋅44) (⋅44 − 1)
∴ f (⋅44) = 0 ⋅ 341 + (⋅44)(0 ⋅ 092) + ( −0 ⋅ 058)
2!
(1⋅ 44) (⋅44) (⋅44 − 1)
+ (0 ⋅ 010 )
3!
(1⋅ 44) (⋅44) (⋅44 − 1) (⋅44 − 2)
+ (0 ⋅ 027)
4!
(2 ⋅ 44) (1⋅ 44) (⋅44) (⋅44 − 1) (⋅44 − 2)
+ ( −0 ⋅ 016)
5!
= 0 ⋅ 341 + (⋅44) (0 ⋅ 092) + (⋅22) (⋅56) (0 ⋅ 058) − (⋅24) (⋅44) (⋅56) (0 ⋅ 010 )
+ (⋅06) (⋅44) (⋅56) (0 ⋅ 027) − (2 ⋅ 44) (⋅06) (⋅44) (⋅56) (1⋅ 56) (⋅0032)
= ⋅ 389 approx.
N-66
(d) The Laplace-Everett’s formula is
 u ( u2 − 1) 2 u ( u2 − 1) ( u2 − 4) 4 
f ( t) = u f (0 ) + ∆ f ( −1) + ∆ f ( −2)
 3! 5! 
 t ( t2 − 1) 2 t ( t2 − 1) ( t2 − 4) 4 
+ t f (1) + ∆ f (0 ) + ∆ f ( − 1) ,
 3 ! 5 ! 
where u = (1 − t).
 (⋅56) (⋅562 − 1)
∴ f (0 ⋅ 44) = ⋅56 × 0 ⋅ 341 + ( − 0 ⋅ 058)
 6
(⋅56) (⋅562 − 1) (⋅562 − 4) 

+ (0 ⋅ 027)
5! 
 (⋅44) (⋅442 − 1)
+ ⋅44 × 0 ⋅ 433 + ( − 0 ⋅ 048)
 3!
(⋅44) (⋅442 − 1) (⋅442 − 4) 
+ (0 ⋅ 011)
5! 
= [⋅56 × 0 ⋅ 341 + (⋅56) (⋅1144) (0 ⋅ 058) + (⋅56) (⋅1144) (3 ⋅1144) (0 ⋅ 0054)]
+ [⋅44 × 0 ⋅ 43 + (⋅44) (⋅1344) (0 ⋅ 048) + (⋅44) (⋅1344) (3 ⋅1344) (⋅0022)]
= 0 ⋅ 389 approximately.
Problem 7: The values of e − x for certain equidistant values of x are given in the following table. Find the
value of e − x when x = 1.7489 by Bessel’s as well as Stirling formula.
x e− x
1.72 0.1790661479
1.73 0.1772844100
1.74 0.1755204006
1.75 0.1737739435
1.76 0.1720448638
1.77 0.1703329888
1.78 0.1686381473
Solution: Proceed as in Ex. 5 and Ex. 6. Ans. 0.1739652000, 0.1739652000.

N-67
Problem 8: Use Gauss interpolation formula to find y41 with the help of the following data :
y30 = 3678 ⋅ 2, y35 = 2995 ⋅1, y40 = 2400 ⋅1,

y45 = 1876 ⋅ 2, y50 = 1416 ⋅ 3.
x − x0
We know u= . Here x = 41, x0 = 40.
h
41 − 40
∴ u= = 0 ⋅ 2.
5
We require the value of y for u = 0 ⋅ 2 .
u yu ∆ yu ∆ 2 yu ∆ 3 yu ∆ 4 yu
−2 3678 ⋅ 2
−683 ⋅1
−1 2995 ⋅1 88 ⋅1
−595 ⋅ 0 −17
0 2400 ⋅1 71⋅1 9⋅9
−523 ⋅ 9 −7 ⋅1
1 1876 ⋅ 2 64 ⋅ 0
−459 ⋅ 9
2 1416 ⋅ 3

u ( u − 1) 2 ( u + 1) u ( u − 1) 3
yu = y0 + u ∆ y0 + ∆ y− 1 + ∆ y− 1
2 6
( u + 1) u ( u − 1) ( u − 2) 4
+ ∆ y− 2 .
24
(0 ⋅ 2) ( −0 ⋅ 8)
∴ y0⋅2 = 2400 ⋅1 + (0 ⋅ 2) ( −523 ⋅ 9) + . (71⋅1)
2
(1⋅ 2) (0 ⋅ 2) ( −0 ⋅ 8) (1⋅ 2) (0 ⋅ 2) ( −0 ⋅ 8) ( −1⋅ 8)
+ . ( −7 ⋅1) + . (9 ⋅ 9)
6 24
= 2400 ⋅1 − 104 ⋅ 78 − 5 ⋅ 688 + ⋅ 2272 + ⋅14256 = 2290 ⋅ 0.
Problem 9: Apply a central difference formula to obtain y32 , given that
y25 = 0 ⋅ 2707, y30 = 0 ⋅ 3027, y35 = 0 ⋅ 3386, y40 = 0 ⋅ 3794.

x − x0
We know u= .
h
Here x = 32 , x0 = 30.
32 − 30 2
∴ u= = = 0 ⋅ 4.
5 5
We require the value of y for u = 0 ⋅ 4.
N-68
u yu ∆ yu ∆ 2 yu ∆ 3 yu
−1 0 ⋅ 2707
0 ⋅ 0320
0 0 ⋅ 3027 0 ⋅ 0039
0 ⋅ 0359 0 ⋅ 0010
1 0 ⋅ 3386 0 ⋅ 0049
0 ⋅ 0408
2 0 ⋅ 3794

u ( u − 1) 2 ( u + 1) u ( u − 1) 3
yu = y0 + u∆ y0 + ∆ y− 1 + ∆ y− 1.
2 6
(0 ⋅ 4) ( −0 ⋅ 6)
∴ y0⋅4 = 0 ⋅ 3027 + (0 ⋅ 4) (0 ⋅ 0359) + (⋅0039)
2
(1⋅ 4) (0 ⋅ 4) ( −0 ⋅ 6)
+ . (0 ⋅ 001)
6
= 0 ⋅ 3027 + 0 ⋅ 01436 − 0 ⋅ 000468 − 0 ⋅ 000056
= 0 ⋅ 3165.
Problem 10: From the following table find the value of f(⋅5437) by Gauss formula :
x f ( x)
⋅51 ⋅529244
⋅52 ⋅537895
⋅53 ⋅546464
⋅54 ⋅554939
⋅55 ⋅563323
⋅56 ⋅571616
⋅57 ⋅579816
Solution: Let us take ⋅54 as the origin and h = 0 ⋅1 as the unit.

x − x0
We know u= .
h
Here x = ⋅ 5437, x0 = ⋅ 54.
⋅5437 − ⋅ 54
∴ u= = 0 ⋅ 37.
⋅01
We require the value of y for u = 0 ⋅ 37.
N-69
u yu ∆ yu ∆2 yu ∆3 yu ∆4 yu ∆5 yu ∆6 yu
–3 ⋅529244
⋅008651
–2 ⋅537895 – ⋅000082
⋅008569 – ⋅000012
–1 ⋅546464 – ⋅000094 ⋅000015
⋅008475 ⋅000003 – ⋅000018
0 ⋅554939 – ⋅000091 – ⋅000003 ⋅000019
⋅008384 ⋅000000 ⋅000001
1 ⋅563323 – ⋅000091 – ⋅000002
⋅008293
2 ⋅571616 – ⋅000093 – ⋅000002
⋅008200
3 ⋅579816

u ( u − 1) 2 ( u + 1) u ( u − 1) 3
yu = y0 + u ∆ y0 + ∆ y− 1 + ∆ y− 1
2 6
( u + 1) u ( u − 1) ( u − 2) 4 ( u + 2) ( u + 1) u ( u − 1) ( u − 2) 5
+ ∆ y− 2 + ∆ y− 2 .
24 120
∴ y0⋅37 = ⋅ 554939 + (⋅37) (⋅008384)
(0 ⋅ 37) ( − ⋅ 63) (1⋅ 37) (⋅37) ( − ⋅ 63)
+ ( − ⋅ 000091) + .0
2 6
(1⋅ 37) (⋅37) ( − ⋅ 63) ( −1⋅ 63)
+ . ( − ⋅ 000003) + negligible terms
24
= ⋅ 554939 + ⋅ 0031021 + ⋅ 0000106 + negligible terms
= ⋅ 558052 .
Problem 11: Define the operators δ and µ and prove that

δ [ f ( x ) g ( x )] = µ f ( x ) δ g ( x ) + µ g ( x ) δ f ( x ).
Solution: First part: The operators δ and µ are defined by the relations
δ f ( x ) = f  x + h − f  x − h
1 1
 2   2 
1 1 
f  x + h + f  x − h  , where h is the interval of differencing.
1
and µ f ( x) =
2   2   2 
Second part: We have
L.H.S. = µ f ( x ) δ g ( x ) + µ g ( x ) δ f ( x )
1 1 1 1 1
= [ f ( x + h) + f ( x − h)] [ g( x + h) − g( x − h)]
2 2 2 2 2
1 1 1 1 1
+ [ g ( x + h) + g ( x − h)] [ f ( x + h) − f ( x − h)]
2 2 2 2 2
1 1 1 1 1 1
= f ( x + h) g ( x + h) − f ( x − h) g ( x − h)
2 2 2 2 2 2
1 1 1 1 1 1
+ g ( x + h) f ( x + h) − g ( x − h) f ( x − h),
2 2 2 2 2 2
the other terms in the two products cancel each other
N-70
1 1 1 1
= f (x + h) g ( x + h) − f ( x − h) g ( x − h)
2 2 2 2
= δ [ f ( x ) g ( x )].
Problem 12: Apply Bessel’s formula to find a polynomial of degree three or less which takes the following
values of the function u x :
x : 4 6 8 10
f ( x) : 1 3 8 20
x −6
Solution: Let u = .
2
x u f ( x) ∆ f ( x) ∆2 f ( x ) ∆3 f ( x )
4 −1 1
6 0 3 3
5 4
8 1 8 7
12
10 2 20
Bessel’s formula is
1 1 u ( u − 1) ( ∆2 y−1 + ∆2 y0 )
f ( x) = ( y0 + y1 ) + ( u − ) ∆y0 +
2 2 2 2
 u − 1  u u − 1)
 
 2
+ ∆3 y−1 + ...
6
 u − 1  u ( u − 1)
 
1 1 u ( u − 1) (3 + 7)  2
= (3 + 8) + ( u − ) 5 + + ×4
2 2 2 2 6
11 5 5 (2u − 1) u ( u − 1)
= + (2u − 1) + ( u2 − u) +
2 2 2 3
2 3 3 2 17
= u + u + u+ 3
3 2 6
3 2
2  x − 6 3  x − 6 17  x − 6 
=   +   +   +3
3 3  2 2  6  2 
1 3 17
= ( x − 6)3 + ( x − 6)2 + ( x − 6) + 3.
12 8 12
1 3 17
∴ f ( x) = ( x − 6)2 + ( x − 6)2 + ( x − 6) + 3.
12 8 12
N-71
Problem 13: Given:

θ : 0° 5° 10° 15° 20° 25° 30°
tan θ : 0 ⋅ 0000 0 ⋅ 0875 0 ⋅1763 0 ⋅ 2679 0 ⋅ 3640 0 ⋅ 4663 0 ⋅ 5774
Using Stirling formula show that tan16° = 0 ⋅ 28676. (Kumaun 2010)
Solution: Take θ = 15° as the origin and h = 5° as the unit.

16° − 15°
We are to find the value of tan θ for u = = 0 ⋅2 .
5°
θ u yu ∆ yu ∆2 yu ∆3 yu ∆4 yu ∆5 yu ∆6 yu
0° −3 0 ⋅ 0000
⋅0875
5° −2 0 ⋅ 0875 ⋅0013
⋅0888 ⋅0015
10° −1 0 ⋅1763 ⋅0028 ⋅0002
⋅0916 ⋅0017 − ⋅ 0002
15° 0 0 ⋅ 2679 ⋅0045 ⋅0000 ⋅0011
⋅0961 ⋅0017 ⋅0009
20° 1 0 ⋅ 3640 ⋅0062 ⋅0009
⋅1023 ⋅0026
25° 2 0 ⋅ 4663 ⋅0088
⋅1111
30° 3 0 ⋅ 5774
Stirling’s formula is
1 u2 2
yu = y0 + u . [ ∆y0 + ∆y−1 ] + ∆ y− 1
2 2
u ( u2 − 1) 1 3 u2 ( u2 − 1) 4
+ [ ∆ y−1 + ∆3 y−2 ] + ∆ y− 2
3! 2 4!
u ( u2 − 1) ( u2 − 4) 1 5 u2 ( u2 − 1) ( u2 − 4) 6
+ [ ∆ y−2 + ∆5 y−3 ] + ∆ y− 3 .
5! 2 6!
⋅2 (⋅2)2
∴ y⋅2 = 0 ⋅ 2679 + [⋅0961 + ⋅ 0916] + (⋅0045)
2 2
(⋅2) (⋅04 − 1) [⋅0017 + ⋅ 0017] (⋅04) (⋅04 − 1)
+ + (⋅0000 )
6 2 24
(⋅2) (⋅04 − 1) (⋅04 − 4) ( − ⋅ 0002 + ⋅ 0009)
+ ⋅
120 2
(⋅04) (⋅04 − 1) (⋅04 − 4)
+ (⋅0011)
720
= 0 ⋅ 2679 + ⋅ 01877 + ⋅ 00009 + negligible quantities
= 0 ⋅ 28676.
Thus the estimated value of tan 16° = 0 ⋅ 28676.
N-72

1. We know that the operator δ is defined by the operator equation
δ = E 1 /2 − E −1 /2 .
2. We know that the operator µ is defined by the operator equation

1
µ= ( E 1 /2 + E −1 /2 ).
2
3. We know that the relation between the operators µ and δ is
δ2
µ2 = 1 + ⋅
4
1
4. We know that σ = ⋅
δ
E 1 /2
5. We know that σ = ⋅
E −1
6. We know that µ = cosh (U / 2).
7. See article 1. Hence, the correct answer is (a).

8. See problem 3 (iii).
9. We know that Stirling’s interpolation formula is the mean of Gauss’s forward formula and
Gauss’s backward formula.
10. See article 3.
11. See article 3.

1. The operator δ is called the central difference operator.
2. The operator µ is called the mean operator.
3. The mean of Gauss’s forward formula and Gauss’s backward formula gives Stirling’s
interpolation formula.
4. The mean of Gauss’s forward formula and the third formula due to Gauss gives Bessel’s
formula.
5. Laplace-Everett’s interpolation formula contains differences only of even order.
True or False
1. The given formula is Gauss’s forward formula.
2. We know that Bessel’s interpolation formula is the mean of Gauss’s forward formula and
the third formula due to gauss.
3. We know that ∆ − ∇ = δ2 . See Ex. 1 part (iii).
4. See article 1.
❍❍❍
N-73
Chapter-5
Solution of Algebraic and
Transcendental Equations
Problem 1: Find the root of x2 − 5 x + 2 = 0 correct to five decimal places which lies between 4 and 5 by
using Newton-Raphson method. (Kumaun 2007, 08, 11, 13; Kanpur 08; 11; Purvanchal 08)
Solution: The given equation is

f ( x ) = x2 − 5 x + 2 = 0. …(1)
2
We have f ( x ) = x − 5 x + 2 and f ′ ( x ) = 2 x − 5.
Now f (4) = 4 × 4 − 5 × 4 + 2 = − 2 , which is −ve

and f (5) = 5 × 5 − 5 × 5 + 2 = 2 , which is +ve.
So, a root of (1) lies between 4 and 5.
To shorten this interval we observe the signs of the values of
f (4 ⋅1), f (4 ⋅ 2), f (4 ⋅ 3) etc.
We have f (4 ⋅ 5) = (4 ⋅ 5)2 − 5 × 4 ⋅ 5 + 2 = 20 ⋅ 25 − 22 ⋅ 5 + 2
= 22 ⋅ 25 − 22 ⋅ 5, which is −ve
and f (4 ⋅ 6) = (4 ⋅ 6)2 − 5 × 4 ⋅ 6 + 2 = 21⋅16 − 23 + 2
= 23 ⋅16 − 23, which is +ve.

So, a root of (1) lies between 4 ⋅ 5 and 4 ⋅ 6.
We take x0 = 4 ⋅ 5 and obtain successive approximations x1 , x2 , x3 etc., using Newton-Raphson
formula
f ( xn ) x 2 − 5 xn + 2
xn +1 = xn − = xn − n
f ′ ( xn ) 2 xn − 5
2 xn2 − 5 xn − xn2 + 5 xn − 2 x 2 −2
= = n ⋅
2 xn − 5 2 xn − 5
xn2 − 2
Thus, xn +1 = ⋅ …(2)
2 xn − 5
Putting n = 0 in (2), the first approximation

x 2 −2 (4 ⋅ 5)2 − 2 18 ⋅ 25
x1 = 0 = = = 4 ⋅ 5625.
2 x0 − 5 2 × 4 ⋅ 5 − 5 4
Putting n = 1 in (2), the second approximation

x2 −2 (4 ⋅ 5625)2 − 2 18 ⋅ 816406
x2 = 1 = = = 4 ⋅ 561553.
2 x1 − 5 2 × 4 ⋅ 5625 − 5 4 ⋅125
N-74
Putting n = 2 in (2), the third approximation

x 2 −2 (4 ⋅ 561553)2 − 2
x3 = 2 =
2 x2 − 5 2 × 4 ⋅ 561553 − 5
18 ⋅ 807766
= = 4 ⋅ 56155, correct to five decimal places.
4 ⋅123106
We see that up to five decimal places, x2 = x3 . So, we take x2 = 4 ⋅ 56155 as the value of the
required root.
Hence the root of x2 − 5 x + 2 = 0 correct to five decimal places which lies between 4 and 5 is
4 ⋅ 56155.
Problem 2: Find the root of log x = cos x , by Newton-Raphson method up to five decimal places.
(Lucknow 2009)
Solution: Let f ( x) ≡ log e x − cos x = 0 so that f ′ ( x ) = (1 / x ) + sin x. We see that f ( x ) = − ive
for x = 1⋅ 3 and f ( x ) = + ive for x = 1⋅ 4, therefore the root lies between 1⋅ 3 and 1⋅ 4.
Taking x0 = 1⋅ 3 and applying Newton-Raphson formula, we have
f ( x0 ) log e 1⋅ 3 − cos 1⋅ 3
x1 = x0 − = 1⋅ 3 −
f ′ ( x0 ) {1 / (1⋅ 3)} + sin 1⋅ 3
⋅11394 − ⋅ 26750
= 1⋅ 3 − = 1⋅ 3029.
⋅76923 + ⋅ 96356
Now taking x1 = 1⋅ 3029 as approximate root and repeating Newton-Raphson formula, we have
f ( x1 ) log e 1⋅ 3029 − cos (1⋅ 3029)
x2 = x1 − = 1⋅ 3029 − = 1⋅ 30295.
f ′ ( x1 ) 1 / (1⋅ 3029) + sin (1⋅ 3029)
f ( x2 ) log e 1⋅ 30295 − cos (1⋅ 30295)
x3 = x2 − = 1⋅ 30295 −
f ′ ( x2 ) 1 / (1⋅ 30295) + sin (1⋅ 30295)
= 1⋅ 30295.
As x2 = x3 , therefore the root is 1⋅ 30295.
1
Problem 3: Find a root of x = + sin x near x = 1⋅ 5.
2
Solution: The given transcendental equation is

1
f ( x) ≡ x − − sin x = 0.
2
∴ f ′ ( x ) = 1 − cos x.
Taking x0 = 1⋅ 5 and applying Newton-Raphson formula, we have
1
1⋅ 5 − − sin (1⋅ 5)
f ( x0 ) 2
x1 = x0 − = 1⋅ 5 −
f ′ ( x0 ) 1 − cos (1⋅ 5)
1⋅ 5 − ⋅ 5 − sin (85 ⋅ 98° ) 1⋅ 0 − ⋅ 9974
= 1⋅ 5 − = 1⋅ 5 −
1 − cos (85 ⋅ 98° ) 1 − ⋅ 0715
(Using trigonometric tables)

⋅0026
= 1⋅ 5 − = 1⋅ 5 − ⋅ 0028 = 1⋅ 4972 .
⋅9285
N-75
f ( x1 ) 1⋅ 4972 − ⋅ 5 − sin (1⋅ 4972)
x2 = x1 − = 1⋅ 4972 −
f ′ ( x1 ) 1 − cos (1⋅ 4972)
1⋅ 4972 − ⋅ 5 − sin 85 ⋅ 82°  ∵ 22 = π = 180 ° 
= 1⋅ 4972 −
1 − cos 85 ⋅ 82°  7 
1⋅ 4972 − ⋅ 5 − ⋅ 9973 − ⋅ 0001

= 1⋅ 4972 − = 1⋅ 4972 −
1 − ⋅ 0726 ⋅9273
= 1⋅ 4972 + ⋅ 0001078 = 1⋅ 4973.
Problem 4: The equation x4 − 5 x3 − 12 x2 + 76 x − 79 = 0 has two roots close to x = 2 . Find the roots
correct upto four decimals.

f ( x ) ≡ x4 − 5 x3 − 12 x2 + 76 x − 79 = 0.
By synthetic division for x = 2 , we have
2 1 −5 −12 76 −79
2 −6 −36 80
2 1 −3 −18 40 1 = f (2)
2 −2 −40
2 1 −1 −20 0 = f ′(2)
2 2
1
1 1 −18 = f ′ ′(2) ⇒ f ′ ′(2) = − 36
2
Newton’s modified formula for nearly equal roots i. e.,

f ′( xi ) f ′( x0 )
xi +1 = xi − ⇒ x1 = x0 − ⋅
f ′ ′( xi ) f ′ ′( x0 )
0
∴ For x0 = 2 , we have x1 = 2 − = 2 = c (say).
36
 −2 f ( c )   −2 f (2)
Now x=c±   =2±  ⋅
 f ′ ′ (c )   f ′ ′ (2) 
 − 2 × 1
∴ x=2±   = 2 ± √ (⋅05555) = 2 ± ⋅ 2357.
 − 36 
∴ The roots are 2 + ⋅ 2357 = 2 ⋅ 2357 and 2 − ⋅ 2357 = 1⋅ 7642 .
Problem 5: Describe the method of false position to solve an equation f ( x) = 0. Workout two steps of this
method to solve x3 + 1⋅ 2 x2 − 4 x − 4 ⋅ 8 = 0.
N-76
Solution: For method of false position refer article 2 part (ii).

Here f ( x ) ≡ x3 + 1⋅ 2 x2 − 4 x − 4 ⋅ 8 = 0.
Trying x = 1, we find
f (1) = 1 + 1⋅ 2 − 4 − 4 ⋅ 8 = − 6 ⋅ 6 = − ive
and for x = 2 , we have
f (2) = 23 + 1⋅ 2 × 22 − 4 × 2 − 4 ⋅ 8 = 8 + 4 ⋅ 8 − 8 − 4 ⋅ 8 = 0.
As x = 2 satisfies the equation so x = 2 is a real root of the given equation.
Problem 6: Determine the real root of tan x = x by iteration method.
Solution: Proceed as in problem 3. The answer is 4 ⋅ 43464.
Problem 7: Find the roots of x3 − 2 x2 − 5 x + 6 = 0, by Graeffe’s method, squaring three times.

(Kumaun 2012)

x3 − 2 x2 − 5 x + 6 = 0 …(1)
Let the roots of (1) be p1 , p2 , p3 .

The equation (1) can be written as
x3 − 5 x = 2 x2 − 6 or x ( x2 − 5) = (2 x2 − 6).
Squaring, we get x2 ( x2 − 5)2 = (2 x2 − 6)2 .
Putting x2 = y, we have
y ( y − 5)2 = (2 y − 6)2 or y3 − 14 y2 + 49 y − 36 = 0
or y ( y2 + 49) = 14 y2 + 36
or y2 ( y2 + 49)2 = (14 y2 + 36)2
or z ( z + 49)2 = (14z + 36)2 [Replacing y2 by z ]
or z 3 − 98 z 2 + 1393 z − 1296 = 0
or z ( z 2 + 1393) = 98 z 2 + 1296
or z 2 ( z 2 + 1393)2 = (98 z 2 + 1296)2
or t ( t + 1393)2 = (98 t + 1296)2 [Replacing z 2 by t ]
or t 3 − 6818 t 2 + 1686433 t − 1679616 = 0 …(2)
The roots of (2) are 8th powers of the roots of (1).

Let the roots of (2) be q1 , q2 , q3 .
Then q1 = p18 = | p1|8 , q 2 = | p2|8 , q 3 = | p3|8 .
From (2), we have

q1 ≈ 6818,
q1 q2 ≈ 1686433
q1 q2 q3 ≈ 1679616 .
N-77
∴ | p1|8 = q1 ≈ 6818
1686433
| p2|8 = q2 ≈
6818
1679616
| p3|8 = q3 ≈ ⋅
1686433
Taking square roots three times, we get
| p1| = (6818)1 /8 = 3 ⋅ 0144433
| p2| = (1686433 / 6818)1 /8 = 1⋅ 9914253
| p3| = (1679616 / 1686433)1 /8 = 0 ⋅ 9994938.
Now | p1| = 3 ⋅ 0144433 ⇒ p1 = ± 3 ⋅ 0144433.

By actually substituting in (1), we see that p1 = − 3 ⋅ 0144433 does not satisfy (1). So
p1 = 3 ⋅ 0144433.
Similarly, we see that the admissible values of p2 and p3 are
p2 = − 1⋅ 9914253, p3 = 0 ⋅ 9994938.
Hence, the required roots of the given equation (1) are
3 ⋅ 0144433, − 1⋅ 9914253, 0 ⋅ 9994938.
We see that the exact values of the roots of (1) are 3, − 2 , 1. The approximate values
3 ⋅ 014, − 1⋅ 99, 0 ⋅ 99949
of the roots given by Graeffe’s root square method are sufficiently close to them.
Problem 8: Use Graeffe’s method to solve the equation given below, squaring three times :
x3 − 5 x2 − 17 x + 20 = 0. (Kumaun 2010; Gorakhpur 13)
Solution: The given equation is x3 − 5 x2 − 17 x + 20 = 0 …(1)

First we shall form an equation whose roots are the squares of the roots of (1) i. e., are p12 , p22 , p32 .
Collecting all terms containing even powers of x to one side and all terms containing odd powers of
x to the other side and squaring both sides, we get from (1)
x2 ( x2 − 17)2 = 25 ( x2 − 4)2 …(2)
From (2) replacing x2 by y, we get
y ( y − 17)2 = 25 ( y − 4)2
or y3 − 59 y2 + 489 y − 400 = 0, whose roots are p12 , p22 , p32 .
Now y ( y2 + 489) = 59 y2 + 400.
Squaring both sides and replacing y2 by z , we get
z ( z + 489)2 = (59 z + 400 )2
or z 3 + (978 − 592 ) z 2 + (4892 − 118 × 400 ) z − 160000 = 0

N-78
or z 3 − 2503 z 2 + 191921 z − 160000 = 0
or z ( z 2 + 191921) = (2503 z 2 + 160000 )
or z 2 ( z 2 + 191921)2 = (2503 z 2 + 160000 )2
or t ( t + 191921)2 = (2503 t + 160000 )2 [Replacing z 2 by t ]
or t 3 + (383842 − 25032 ) t 2 + (1919212 − 5006 × 160000 ) t
− 25600000000 = 0
3 2
or t + (383842 − 6265009) t + (36833670000 − 800960000 ) t
− 25600000000 = 0
3 2
or t − 5881167 t + 36032710000 t − 25600000000 = 0 …(3)
Let the roots of (3) be q1 , q2 , q3 .
Then q1 = p18 = | p1|8 , q 2 = | p2|8 , q 3 = | p3|8 .
From equation (3), we have

q1 ≈ 5881167
q1 q 2 ≈ 36032710000
q1 q 2 q 3 ≈ 25600000000.
∴ | p1|8 = q1 ≈ 5881167
36032710000
| p2|8 = q 2 ≈ ≈ 6126 ⋅ 795
5881167
25600000000 2560000
| p3|8 = q 3 ≈ = ≈ ⋅ 7104655.
36032710000 3603271
Taking square roots 3 times, we get
| p1| ≈ 7 ⋅ 01751
| p2| ≈ 2 ⋅ 97443
| p3| ≈ 0 ⋅ 95817.
Now | p1| = 7 ⋅ 01751 ⇒ p1 = ± 7 ⋅ 01751.
By actually substituting in (1), we see that p1 = 7 ⋅ 01751 satisfies (1) while p1 = − 7 ⋅ 01751 does
not satisfy (1). So, p1 = 7 ⋅ 01751.
Similarly, we see that the admissible values of p2 and p3 are p2 = − 2 ⋅ 97443, p3 = 0 ⋅ 95817.
Hence, the required roots of the given equation are 7 ⋅ 01751, − 2 ⋅ 97443, 0 ⋅ 95817.
Problem 9: Find 30 by using iterative process.

Solution: Newton’s iterative formula for obtaining the square root of b i. e., the root of the
equation
f ( x ) ≡ x2 − b = 0
1 b
is xn +1 =  xn + ⋅ …(1)
2 xn 
Let x = 30 so that x2 = 30.
Take f ( x ) = x2 − 30.
N-79
As f (5 ⋅ 4) = − ⋅ 84 ( − ive ) and f (5 ⋅ 5) = ⋅ 25 ( + ive ) therefore the root lies between x = 5 ⋅ 4 and

x = 5 ⋅ 5.
Taking x0 = 5 ⋅ 5, from (1) we have
1 b  1 30  1
x1 =  x0 +  = 5 ⋅ 5 +  = (5 ⋅ 5 + 5 ⋅ 45)
2 x0  2  5⋅5 2
1
= × 10 ⋅ 95 = 5 ⋅ 475.
2
Again taking approximate root as 5 ⋅ 475 and repeating Newton’s formula (1), we have
1 b 1
x2 =  x1 +  =  5 ⋅ 475 +
30 

2 x1  2  5 ⋅ 475 
1 1
= (5 ⋅ 475 + 5 ⋅ 479) = × 10 ⋅ 954 = 5 ⋅ 477.
2 2
Proceeding in a similar way, we have
1 b  1  5 ⋅ 477 + 30  = 5 ⋅ 4772
x3 =  x2 +  =  
2 x2  2  5 ⋅ 477 
1 b  1 30 
and x4 =  x3 +  =  5 ⋅ 4772 +  = 5 ⋅ 477225.
2 x3  2  5 ⋅ 4772 
Problem 10: Obtain the cube root of 12 to five decimal places by Newton-Raphson method.
Solution: Let x = (12)1 /3 .
Then x3 = 12 or x3 − 12 = 0.
Therefore the cube root of 12 is nothing but the root of the equation
f ( x ) ≡ x3 − 12 = 0. …(1)
We have f ( x ) = x3 − 12 and f ′ ( x ) = 3 x2 .
Also f (2) = 23 − 12 = 8 − 12 = − 4 which is −ve
and f (3) = 33 − 12 = 27 − 12 = 15 which is +ve.
So, a root of (1) lies between 2 and 3.

To shorten this interval we observe the signs of the values of f (2 ⋅1), f (2 ⋅ 2), f (2 ⋅ 3) etc.
We have f (2 ⋅ 2) = (2 ⋅ 2)3 − 12 = 10 ⋅ 648 − 12 , which is −ve
and f (2 ⋅ 3) = (2 ⋅ 3)3 − 12 = 12 ⋅167 − 12 , which is +ve.
So, (12)1 /3 i. e., the root of (1) lies between 2 ⋅ 2 and 2 ⋅ 3.
We take x0 = 2 ⋅ 2 and obtain successive approximations x1, x2 , x3 etc., using Newton-Raphson

formula
f ( xn )
xn +1 = xn −
f ′ ( xn )
x 3 − 12 3 xn3 − xn3 + 12
= xn − n =
3 xn2 3 xn2
2 xn3 + 12 1  
= =  2 xn + 12  ⋅
3 xn2 3  2
 xn 
N-80
1  
Thus, xn +1 =  2 xn + 12  ⋅ …(2)
3  2
 xn 
Taking n = 0, we have from (2), the first approximation
1    
x1 =  2 x0 + 12  = 1 2 × 2 ⋅ 2 + 12 
3  2 3 2
(2 ⋅ 2) 
 x0  
=
1  4 ⋅ 4 + 12  = 1 (4 ⋅ 4 + 2 ⋅ 4793388) = 2 ⋅ 2931129.
 
3  4 ⋅ 84  3
Taking n = 1, we have from (2), the second approximation
1    
x2 =  2 x1 + 12  = 1 2 × 2 ⋅ 2931129 + 12
 2 3 2
3  x1   (2 ⋅ 2931129) 
=
1  4 ⋅ 5862258 + 12 
 
3  5 ⋅ 2583668 
1
= (4 ⋅ 5862258 + 2 ⋅ 2820774) = 2 ⋅ 2894344.
3
Taking n = 2 , we have from (2), the third approximation
1 12  1  12 
x3 =  2 x2 + = 2 × 2 ⋅ 2894344 + 
3  2 
x2  3  2
(2 ⋅ 2894344) 
= 4 ⋅ 5788688 +  = 1 (4 ⋅ 5788688 + 2 ⋅ 2894166)

1 12
3  5 ⋅ 2415099  3
= 2 ⋅ 2894285 = 2 ⋅ 28943 up to five places of decimal.
We see that up to five places of decimal, x2 = x3 .
Hence, up to five places of decimal, the cube root of 12 = 2 ⋅ 28943.
Problem 11: Determine the real root of 2 x − 3sin x − 5 = 0.
Solution: Let f ( x) = 2 x − 3 sin x − 5 so that f ′ ( x) = 2 − 3 cos x.

Here f (2) = − ive and f (3) = + ive so the root lies between 2 and 3. Taking x0 = 3 and applying
Newton-Raphson formula, we have
f ( x0 ) 2 × 3 − 3 sin 3 − 5
x1 = x0 − =3−
f ′ ( x0 ) 2 − 3 cos 3
1 − 3 sin (171⋅ 97° )
=3−
2 − 3 cos (171⋅ 97° ) [ ∵ π = 3 ⋅14 = 180 ° ⇒ 3 = 171⋅ 97° ]
1 − 3 sin (90 ° + 81⋅ 97° )
=3−
2 − 3 cos (90 ° + 81⋅ 97° )
1 − 3 cos 81⋅ 97°
=3−
2 + 3 sin 81⋅ 97°
1 − 3 × ⋅1409 ⋅5773
=3− =3− [From trigonometric tables]
2 + 3 × ⋅ 990 4 ⋅ 97
= 3 − ⋅116 = 2 ⋅ 8838.
Now taking x1 = 2 ⋅ 8838 as approximate root and again applying Newton-Raphson formula, we
obtain x2 = 2 ⋅ 8832 .
N-81
Problem 12: Show that the square root of N = AB is given by

S N
√N ≈ + , where S = A + B.
4 S (Rohilkhand 2012)
Solution: Here let x = N so that x2 = N. Take f ( x) = x2 − N.

∴ f ′( x ) = 2 x.
Applying Newton-Raphson formula,
f ( xn ) x2 − N x2n + N x N
xn + 1 = xn − = xn − n = = n+ ⋅ ...(1)
f ′( xn ) 2 xn 2 xn 2 2 xn
Now taking approximate root

A+ B
xn = ⋅
2
∴ From (1), we get
A+ B N
xn + 1 = +
4 A+ B
S N
xn + 1 = + . [∵A + B = S ]
4 S
x
Problem 13: Find the real root of the equation e − 3 x = 0, using iteration method.
x
Solution: Let f ( x) ≡ e − 3 x = 0 so that f ′ ( x ) = e x − 3.
We have f (0 ) = e 0 − 3 × 0 = 1 = + ive
and f (1) = e − 3 = 2 ⋅ 7182 − 3 = − ⋅ 2817 = − ive.
As f ( x ) is +ive for x = 0 and −ive for x = 1, therefore root lies between x = 0 and x = 1.
Again f (⋅5) = e ⋅5 − 3 × ⋅ 5 = 1⋅ 6487 − 1⋅ 5 = ⋅14872 = + ive
∴ Root lies between x = ⋅ 5 and x = 1.

Again f (⋅6) = e ⋅6 − 3 × ⋅ 6 = 1⋅ 8221 − 1⋅ 8 = ⋅ 0221 = + ive.
∴ Root lies between x = ⋅ 6 and x = 1.

Now taking approximate root x0 = ⋅ 6 we can use Newton’s iterative formula
f ( xn )
xn+1 = xn − ⋅
f ′ ( xn )
f ( x0 ) ⋅0221 ⋅0221
Thus x1 = x0 − = ⋅6 − = ⋅6 −
f ′ ( x0 ) 1⋅ 8221 − 3 − 1⋅ 7779
= ⋅ 6 + ⋅ 0187 = ⋅ 6187.
Repeating the formula taking approximate root as ⋅6187, we have
f ( x1 ) 1⋅ 8565 − 1⋅ 8561
x2 = x1 − = ⋅ 6187 −
f ′( x1 ) 1⋅ 8563 − 3
⋅0004
= ⋅ 6187 + = ⋅ 61904.
1⋅1435
Thus the approximate real root of the given equation is 0 ⋅ 61904.
N-82
Problem 14: Solve x3 − 8 x2 + 17 x − 10 = 0 by Graeffe’s method, squaring two times.

(Kumaun 2013)
3 2
Solution: The given equation is x − 8 x + 17 x − 10 = 0. …(1)
The equation (1) can be written as
x ( x2 + 17) = 8 x2 + 10.
Squaring, we get
x2 ( x2 + 17)2 = (8 x2 + 10 )2 .
Putting x2 = y, we get
y ( y + 17)2 = (8 y + 10 )2
or y3 − 30 y2 + 129 y − 100 = 0
or y3 + 129 y = 30 y2 + 100
or y ( y2 + 129) = 30 y2 + 100.
Squaring, we get
y2 ( y2 + 129)2 = (30 y2 + 100 )2 .
Putting y2 = z , we have
z ( z + 129)2 = (30 z + 100 )2
or z 3 − 642 z 2 + 10641 z − 10000 = 0 …(2)

Let the roots of (2) be q1 , q 2 , q 3 .
Then q1 = p14 = | p1|4 , q 2 = | p2|4 , q 3 = | p3|4 .
From (2), we have

q1 ≈ 642
q1q 2 ≈ 10641
q1 q 2 q 3 ≈ 10000.
∴ | p1|4 = q1 ≈ 642
10641
| p2|4 = q2 ≈
642
10000
| p3|4 = q3 ≈ ⋅
10641
Taking square roots two times, we get
| p1| = (642)1 /4 = 5 ⋅ 03366
| p2| = (10641 / 642)1 /4 = 2 ⋅ 01772
| p3| = (10000 / 10641)1 /4 = 0 ⋅ 98459.
Now | p1| = 5 ⋅ 03366 ⇒ p1 = ± 5 ⋅ 03366.

N-83
By actually substituting in (1), we see that p1 = 5 ⋅ 03366 satisfies (1) while p1 = − 5 ⋅ 03366 does
not satisfy (1).
So, p1 = 5 ⋅ 03366.
Similarly, we see that the admissible values of p2 , p3 are p2 = 2 ⋅ 01772 , p3 = 0 ⋅ 98459.
Hence, the approximate values of the roots of the given equation are
5 ⋅ 03366, 2 ⋅ 01772 , 0 ⋅ 98459.
1
Problem 15: The equation sin x = 5 x − 2 can be put as x = sin−1 (5 x − 2)and also as x = ( sin x + 2)
5
suggesting two iterative procedures for its solution. Which of these, if any, would succeed and which would fail
to give the root in the neighbourhood of 0 ⋅ 5.
Solution: As given sin x = 5 x − 2 ;
∴ x = sin−1 (5 x − 2).
Let us first consider x = sin−1 (5 x − 2).
Therefore first approximation of x is given by
x1 = sin−1  5 × − 2 , taking

1
 2 
1
x = ⋅5 = approx., (as given)
2
1 π
= sin−1 = ⋅
2 6
π
Hence first approximation is ⋅
6
We can apply the procedure again to get another approximation and so on.
1
Now let us consider x = (sin x + 2).
5
1
In this case taking x = ⋅ 5 = on (R.H.S.), the first approximation is given by
2
x1 =
1  sin 1 + 2 ⋅
 
5  2 
1
Now sin is not known, hence this method fails even for first approximation. Hence we would
2
succeed from first one.
Problem 16: Find the cube root of 10.

Solution: Let x be the cube root of N = 10. Then as (8)1 /3 = 2 and (27)1 /3 = 3, we may take
initial approximate value of the root x0 = 2 ⋅ 5.
The Newton’s iterative formula for obtaining cube root is
1 N 
xn +1 = 2 xn + 2  [ Prove this formula before using it]
3 xn 
N-84
Here N = 10.
1 10 
Therefore, x1 = first approximation = 2 × 2 ⋅ 5 +  = 2⋅2 ;
3  (2 ⋅ 5)2 
1  10 
x2 = second approximation = 2 × 2 ⋅ 2 +  = 2 ⋅15537 ;
3  (2 ⋅ 2)2 
and x3 = third approximation
1  10 
= 2 × 2 ⋅15537 + = 2 ⋅15444 .
2
3  (2 ⋅ 15537 ) 
Hence the desired cube root is 2 ⋅15444.
Problem 17: Evaluate by Newton-Raphson method

1 /4
(ii)  
1
(i) 29, ⋅
 3
Solution: (i) Let x = 29. Then x2 = 29 so that f ( x) ≡ x2 − 29 = 0 and f ′( x) = 2 x.

Now f (5) = 25 − 29 = − 4 i. e., negative
and f (6) = 36 − 29 = 7 i. e., positive.
So the root lies between x = 5 and x = 6. Let x0 = 6.
By Newton-Raphson formula, we have
f ( xn ) f ( x0 )
xn +1 = xn − giving x1 = x0 − ⋅
f ′ ( xn ) f ′ ( x0 )
62 − 29 7
For x0 = 6, we get x1 = 6 − =6− = 5 ⋅ 417.
2×6 12
Again applying formula for x1 = 5 ⋅ 417, we have

f ( x1 ) (5 ⋅ 417)2 − 29
x2 = x1 − = 5 ⋅ 417 −
f ′ ( x1 ) 2 × 5 ⋅ 417
= 5 ⋅ 417 − ⋅ 03173 = 5 ⋅ 3852 .

1 /4
x =  
1
(ii) Let ⋅
 3
1 1
∴ x4 = so that f ( x ) = x4 −
3 3
and f ′ ( x ) = 4 x3 .
Now f (0 ) = 0 − ⋅ 33 = − ive.
and f (1) = 1 − ⋅ 33 = + ive.
∴ Root lies between 0 and 1. Taking x0 = 1 and applying Newton-Raphson method, we have
f ( x0 ) (1)4 − ⋅ 33 ⋅67
x1 = x0 − =1− =1−
f ′ ( x0 ) 4 × (1)3 4
= ⋅ 8325.
N-85
Now taking x1 = ⋅ 8325 as approximate root and again applying Newton-Raphson formula, we
have
f ( x1 ) (⋅8325)4 − ⋅ 33
x2 = x1 − = ⋅ 8325 −
f ′ ( x1 ) 4 × (⋅8325)3
⋅1503268
= ⋅ 8325 − = 0 ⋅ 7673.
2 ⋅ 307876
Problem 18: By using Newton-Raphson method, find the root of x4 − x − 10 = 0, which is nearer to
x = 2 , correct to three places of decimals. (Rohilkhand 2002; Agra 12; Bundelkhand 12;
Kumaun 12)
Solution: Here f ( x) = x4 − x − 10.
∴ f ′( x ) = 4 x3 − 1.
f ( xn ) x 4 − xn − 10 3 xn4 + 10
∴ xn +1 = xn − = xn − n = ⋅
f ′( xn ) 4 xn3 − 1 4 xn3 − 1
The approximate value of the root is given to be 2 . Taking x 0 = 2 , we get

3 x 04 + 10 3 ⋅ 24 + 10
x1 = = = 1⋅ 871
4 x 03 − 1 4 ⋅ 22 − 1
3 x14 + 10 3 (1⋅ 871)4 + 10

x2 = = = 1⋅ 856
4 x13 − 1 4 (1⋅ 871)3 − 1
3 x 24 + 10 3 (1⋅ 856)4 + 10
x3 = = = 1⋅ 856.
3
4 x2 − 1 4 (1⋅ 856)3 − 1
Since x 2 = x 3 , so the required root is 1⋅ 856.

1. We know that a is a root of the equation f ( x ) = 0 if and only if f ( a) = 0.
2. We know that an algebraic equation of degree n has exactly n roots.
3. The real roots of the equation f ( x ) = 0 are the abscissae of the points where the curve
y = f ( x ) crosses the x-axis.
4. See article 2, part (ii), Regula-Falsi method.
5. See article 2, part (iv), Iteration method.

f ( x0 ) f ( x0 )
6. The improved value of the root is x 0 − ⋅ So, h = − ⋅
f ′ ( x0 ) f ′ ( x0 )
7. Newton’s iterative formula for obtaining a−1 is xn+1 = xn (2 − axn ). See article 2, part (v).
8. Newton’s iterative formula for obtaining the square root of a is
1 a
xn +1 =  xn + ⋅
2 xn 
N-86
9. The method which does not require any prior information of the roots such as approximate
value of the root etc., is Graeffe’s root squaring method.
10. Let f ( x ) = x3 − x − 1. Then f (1) = 1 − 1 − 1 = − 1 and f (2) = 8 − 2 − 1 = 5.
Since f (1) and f (2) are of opposite signs, therefore at least one root or an odd number of
roots of f ( x ) = 0 lie between 1 and 2.
11. All the three numbers1, 2 , 5 satisfy the equation x3 − 8 x2 + 17 x − 10 = 0. Hence, the exact
roots of the equation x3 − 8 x2 + 17 x − 10 = 0 are 1, 2 , 5.
12. See article 2 (v).
13. Lagrange’s method is used to find any value of a function when the given values of the
independent variable are not equidistant, not for finding roots.
14. See article 2 (ii).

1. In an algebraic equation f ( x ) = 0 with real coefficients complex roots occur in conjugate
pairs. So if a + ib is a root of f ( x ) = 0, then a − ib is also a root of f ( x ) = 0.
2. Since complex roots occur in conjugate pairs, therefore every equation of odd degree has at
least one real root.
3. We know that x = a is a root of the equation f ( x ) = 0 if and only if f ( x ) is exactly divisible
by x − a.
4. If f ( a) and f ( b ) are of opposite signs, then the curve y = f ( x ) must cross the x-axis at least
once or at an odd number of points between x = a and x = b. So at least one or an odd
number of real roots of the equation f ( x ) = 0 lie between a and b.
5. See article 2, part (iv), Iteration method.
6. In Newton-Raphson method successive approximations for a root of the equation
f ( xn )
f ( x ) = 0 are given by xn + 1 = xn − ⋅
f ′ ( xn )
7. See article 2, part (vi), Graeffe’s root squaring method.
True or False
1. If f ( x ) is exactly divisible by x − a, then a is a root of the equation f ( x ) = 0.
2. By Descarte’s rule of signs. See article 1.

3. We know that an equation of odd degree has at least one real root.
4. See article 1, point (v).
5. See article 2, part (v). We know that Newton-Raphson method is not suitable in cases
when the graph of f ( x ) is nearly horizontal where it crosses the x-axis.
6. Newton-Raphson method can also be used when the roots are complex.
7. By Newton-Raphson method, the iterative formula for computing p th root of a is
( p − 1) xnp + a
xn +1 = ⋅
pxnp −1
❍❍❍
N-87
Chapter-6
Numerical Differentiation
dy d2 y
Problem 1: Find and at x = 1 from the following table :
dx dx2
x : 1 2 3 4 5 6
y : 198669 295520 389418 479425 564642 644217
Solution: Here the values of the argument are equally spaced and the derivatives are required at
x = 1, which is near the beginning of the table so we shall use Newton’s forward formula.
The forward difference table is given below :
x y = f ( x) ∆ f ( x) ∆2 f ( x ) ∆3 f ( x ) ∆4 f ( x ) ∆5 f ( x )
1 198669
96851
2 295520 − 2953
93898 − 938
3 389418 − 3891 39
90007 − 899 8
4 479425 − 4790 47
85217 − 852
5 564642 − 5642
79575
6 644217
Newton’s forward formula is

x ( x − 1) 2
f ( a + xh) = f ( a) + x ∆ f ( a) + ∆ f ( a)
2!
x ( x − 1) ( x − 2) 3 x ( x − 1) ( x − 2) ( x − 3) 4
+ ∆ f ( a) + ∆ f ( a)
3! 4!
x ( x − 1) ( x − 2) ( x − 3) ( x − 4) 5
+ ∆ f ( a)
5!
N-88
( x2 − x ) 2 ( x3 − 3 x2 + 2 x ) 3
= f ( a) + x ∆ f ( a) + ∆ f ( a) + ∆ f ( a)
2 6
( x4 − 6 x3 + 11x2 − 6 x ) 4
+ ∆ f ( a)
24
( x5 − 10 x4 − 35 x3 − 50 x2 + 24 x ) 5
+ ∆ f ( a).
120
Differentiating with respect to x twice, we get
(2 x − 1) 2 (3 x2 − 6 x + 2) 3
h f ′ ( a + xh) = ∆ f ( a) + ∆ f ( a) + ∆ f ( a)
2 6
(4 x3 − 18 x2 + 22 x − 6) 4
+ ∆ f ( a)
24
(5 x4 − 40 x3 + 105 x2 − 100 x + 24) 5

+ ∆ f ( a)
120
(12 x2 − 36 x + 22) 4
and h2 f ′ ′ ( a + xh) = ∆2 f ( a) + ( x − 1) ∆3 f ( a) + ∆ f ( a)
24
(20 x3 − 120 x2 + 210 x − 100 ) 5

+ ∆ f ( a).
120
Putting a = 1, h = 1, x = 0 and the values of various differences in these equations, we get
f ′ (1) = 96851 +  −  × ( − 2953) + × ( − 938) − × 39 + × 8

1 1 1 1
 2 3 4 5
= 96851 + 1476 ⋅ 5 − 312 ⋅ 6667 − 9 ⋅ 75 + 1⋅ 6
= 98329 ⋅1 − 322 ⋅ 4167 = 98006 ⋅ 683 ;
11 5
and f ′ ′ (1) = − 2953 + ( −1) × ( − 938) + × 39 − × 8
12 6
= − 2953 + 938 + 35 ⋅ 75 − 6 ⋅ 6667 = − 1985 ⋅ 9167.
dy d2 y
Hence, at x = 1, = 98007 and = − 1986.
dx dx2
Problem 2: Compute f ′ ′ ′ (15), given
x : 2 4 9 13 16 21 29
f ( x) : 57 1345 66340 402052 1118209 4287844 21242820
Solution: In this case the values of the argument are unequally spaced, so we may use Newton’s
divided difference formula.
The divided difference table is as given below :

N-89
x f ( x) ∆| f ( x ) ∆| 2 f ( x ) ∆| 3 f ( x ) ∆| 4 f ( x ) ∆| 5 f ( x )
2 57
644
4 1345 1765
12999 556
9 66340 7881 45
83928 1186 1
13 402052 22113 64
238719 2274 1
16 1118209 49401 89
633927 4054
21 4287844 114265
2119372
29 2124282
Since the fifth divided differences are constant, Newton’s divided difference formula is
f ( x ) = f ( a) + ( x − a) ∆| f ( a) + ( x − a) ( x − b ) ∆| 2 f ( a)
+ ( x − a)( x − b ) ( x − c ) ∆| 3 f ( a) + ( x − a) ( x − b ) ( x − c ) ( x − d ) ∆| 4 f ( a)
+ ( x − a) ( x − b ) ( x − c ) ( x − d ) ( x − e ) ∆| 5 f ( a)
[Upto fifth differences]
Differentiating with respect to x thrice, we have
f ′ ′ ′ ( x ) = 6 ∆| 3 f ( a) + [24 x − 6 ( a + b + c + d )] ∆| 4 f ( a)
+ [60 x2 − 24 x ( a + b + c + d ) + 6 ( ab + cd ) + 6 ( a + b ) ( c + d )
− e {24 x − 6 ( a + b + c + d )}] ∆| 5 f ( a).
Putting a = 2 , b = 4, c = 9, d = 13, e = 16, x = 5, we get

f ′ ′ ′ (5) = 6 × 556 + [120 − 168] × 45 + [1500 − 3360
+ 750 + 792 − 16 (120 − 168)] × 1
= 3336 − 2160 + 450 = 1626.
Problem 3: Find the first three derivatives of the function tabulated below at the point x = 2 ⋅ 5 :
x : 1⋅ 5 1⋅ 9 2⋅5 3⋅2 4⋅3 5⋅9
f ( x) : 3 ⋅ 375 6 ⋅ 059 13 ⋅ 625 29 ⋅ 368 73 ⋅ 907 196 ⋅ 579
Solution: In this case the values of the argument are unequally spaced, so we may use here
Newton’s divided difference formula.
N-90
x f ( x) ∆| f ( x ) ∆| 2 f ( x ) ∆| 3 f ( x )
1⋅ 5 3 ⋅ 375
6 ⋅ 71
1⋅ 9 6 ⋅ 059 5 ⋅ 90
12 ⋅ 61 1
2⋅5 13 ⋅ 625 7 ⋅ 60
22 ⋅ 49 1
3⋅2 29 ⋅ 368 10 ⋅ 00
40 ⋅ 49 1
4⋅3 73 ⋅ 907 13 ⋅ 40
76 ⋅ 67
5⋅9 196 ⋅ 579
We observe that the third divided differences are constant.

Again Newton’s divided difference formula is given by
f ( x ) = f ( a) + ( x − a) ∆| f ( a) + ( x − a) ( x − b ) ∆| 2 f ( a)
+ ( x − a) ( x − b ) ( x − c ) ∆| 3 f ( a)
[Taking upto third differences]
On differentiating successively with respect to x thrice, we get

f ′ ( x ) = ∆| f ( a) + [2 x − a − b ] ∆| 2 f ( a)
+ [3 x2 − 2 x ( a + b + c ) + ab + ac + bc ] ∆| 3 f ( a),
f ′ ′ ( x ) = 2 ∆|2 f ( a) + [6 x − 2 ( a + b + c )] ∆| 3 f ( a),
f ′ ′ ′ ( x ) = 6 ∆| 3 f ( a).
Here a = 1⋅ 5, b = 1⋅ 9, c = 2 ⋅ 5.
On putting x = 2 ⋅ 5, we get
f ′ (2 ⋅ 5) = 6 ⋅ 710 + [5 − 3 ⋅ 4] × 5 ⋅ 9
+ [3 × 6 ⋅ 25 − 5 (1⋅ 5 + 1⋅ 9 + 2 ⋅ 5) + 11⋅ 35] × 1
= 6 ⋅ 710 + 9 ⋅ 44 + ⋅ 600 = 16 ⋅ 750,
f ′ ′ (2 ⋅ 5) = 2 [5 ⋅ 9] + [15 ⋅ 00 − 2 × 5 ⋅ 9] × 1 = 15 ⋅ 00,
f ′ ′ ′ (2 ⋅ 5) = 6 × 1 = 6 ⋅ 00.
N-91
Problem 4: From the following data, find f ′(10 ) :
x : 3 5 11 27 34
f ( x) : − 13 23 899 17315 35606
(Purvanchal 2008)
Solution: Here the values of the argument x are not equally spaced, therefore we shall use
Newton’s divided difference formula. The divided difference table is given below :
x f ( x) ∆| f ( x ) ∆| 2 f ( x ) ∆| 3 f ( x ) ∆| 4 f ( x )
3 − 13
18
5 23 16
146 1
11 899 40 0
1026 1
27 17315 69
2613
34 35606

f ( x ) = f ( x0 ) + ( x − x0 ) ∆| f ( x0 ) + ( x − x0 ) ( x − x1 ) ∆| 2 f ( x0 )
+ ( x − x0 ) ( x − x1 ) ( x − x2 ) ∆| 3 f ( x0 )
[Neglecting terms of fourth and higher order differences]
= f (3) + ( x − 3) ∆| f (3) + ( x − 3) ( x − 5) ∆| 2 f (3)
+ ( x − 3) ( x − 5) ( x − 11) ∆| 3 f (3)
= − 13 + ( x − 3) × 18 + ( x − 3) ( x − 5) × 16 + ( x − 3) ( x − 5) ( x − 11) × 1
= − 13 + 18 ( x − 3) + 16 ( x2 − 8 x + 15) + ( x3 − 19 x2 + 103 x − 165).
∴ f ′ ( x ) = 18 + 16 (2 x − 8) + (3 x2 − 38 x + 103).
Putting x = 10, we get
f ′ (10 ) = 18 + 16 (2 × 10 − 8) + (3 × 102 − 38 × 10 + 103)
= 18 + 192 + 300 − 380 + 103 = 233.

Hence, f ′ (10 ) = 233.
N-92
dy
Problem 5: Find at x = 1 from the following table :
dx
x: ·7 ·8 ·9 1·0 1·1 1·2 1·3
f (x): 0·644218 0·717356 0·783327 0·841471 0·891207 0·932039 0·963558
Solution: Proceed as in Ex. 5. Ans. 0·54030.
Problem 6: Find the first and second derivatives of log e x at x = 500.
x : 500 510 520 530 540 550
log x : 6·214608 6·234411 6·253829 6·272877 6·291569 6·309918
Solution: Proceed as in Ex. 1. Ans. 0 ⋅ 002000, − 0 ⋅ 000040.
dy d2 y
Problem 7: Find and of y = x1 /3 at x = 50 from the following table :
dx dx2
x: 50 51 52 53 54 55 56
1 /3
y= x : 3⋅6840 3⋅7084 3⋅7325 3⋅7563 3⋅7798 3⋅8030 3⋅8259
Solution: Since the derivatives are required at x = 50, which is near the beginning of the table,
therefore we shall use Newton’s forward formula. The difference table is given below :
x y = x1 /3 ∆ y ∆2 y ∆3 y
50 3⋅6840
0⋅0244
51 3⋅7084 – 0⋅0003
0⋅0241 0
52 3⋅7325 – 0⋅0003
0⋅0238 0
53 3⋅7563 – 0⋅0003
0⋅0235 0
54 3⋅7798 – 0⋅0003
0⋅0232 0
55 3⋅8030 – 0⋅0003
0⋅0229
56 3⋅8259
N-93
Newton’s forward formula is

x ( x − 1) 2 x( x − 1) ( x − 2) 3
f ( a + xh) = f ( a) + x ∆ f ( a) + ∆ f ( a) + ∆ f ( a),
2 6
taking upto third differences only
Differentiating w.r.t. x twice and then putting x = 0 in the equations obtained, we get
1 1
h f ′( a) = ∆ f ( a) − ∆2 f ( a) + ∆3 f ( a)
2 3
h2 f ′ ′( a) = ∆2 f ( a) − ∆3 f ( a).
Putting a = 50, h = 1. Then
1 1 1
f ′(50 ) = 0 ⋅ 0244 − ( −0 ⋅ 0003) + (0 ) = 0 ⋅ 02455
1  2 3 
1
f ′ ′(50 ) = [( −0 ⋅ 0003) − 0 ] = − 0 ⋅ 0003.
(1)2
Problem 8: Find the first two derivatives of f ( x) at x = 1 from the following table :
x : −2 −1 0 1 2 3 4
f ( x) : 104 17 0 −1 8 69 272
Solution: Proceed as in Ex. 1. Ans. 1, 6.
Problem 9: Using divided differences, find the value of f ′ (8), given that
f (6) = 1⋅ 556, f (7) = 1⋅ 690, f (9) = 1⋅ 908, f (12) = 2 ⋅158 .
Solution: Divided difference formula is given by
f ( x ) = f ( x 0 ) + ( x − x 0 ) ∆| f ( x 0 ) + ( x − x 0 ) ( x − x1 ) ∆| 2 f ( x 0 )
+ ( x − x 0 ) ( x − x1 ) ( x − x 2 ) ∆| 3 f ( x 0 )
+ ( x − x 0 ) ( x − x1 ) ( x − x 2 ) ( x − x 3 ) ∆| 4 f ( x 0 )
+ ( x − x 0 ) ( x − x1 ) ( x − x 2 ) ( x − x 3 ) ( x − x 4 ) ∆| 5 f ( x 0 ) + … …(1)
x f ( x) ∆| f ( x ) ∆| 2 f ( x ) ∆| 3 f ( x )
6 1⋅ 556
⋅134
7 1⋅ 690 − ⋅ 0083
⋅109 ⋅00051
9 1⋅ 908 − ⋅ 0052
⋅083
12 2 ⋅158
Differentiating (1) with respect to x, we get

f ′ ( x ) = ∆| f ( x 0 ) + [2 x − x 0 − x1 ] ∆| 2 f ( x 0 ) + [3 x2 − 2 x ( x 0 + x1 + x 2 )
+ x 0 x1 + x1 x 2 + x 2 x 0 ] ∆| 3 f ( x 0 ).
N-94
Putting x 0 = 6, x1 = 7, x 2 = 9, x = 8 and the values of the divided differences from the table, we
get
f ′ (8) = ⋅134 + [2 × 8 − 6 − 7] ( − ⋅ 0083)
+ [3 × 64 − 2 × 8 (6 + 7 + 9) + 6 × 7 + 7 × 9 + 9 × 6] (⋅00051)
= ⋅134 − ⋅ 0249 + (192 − 352 + 159) (⋅00051)
= ⋅134 − ⋅ 0249 − ⋅ 00051
= ⋅134 − ⋅ 0254 = 0 ⋅10859.
Problem 10: Find f ′(6) from the following table :
x: 0 1 3 4 5 7 9
f ( x) : 150 108 0 – 54 – 100 – 144 – 84
Solution: Proceed as in Ex. 11. Ans. −23.
Problem 11: Show that the expressions given below are approximations to the third derivative of y x :
∆3 y0 +  x −  ∆4 y0
3
(i)
 2
∆3 y−1 +  x −  ⋅ ( ∆3 y−2 + ∆4 y−1 )

1 1
(ii)
 2 2
Solution: (i) By Newton’s Gregory forward formula, we have

x ( x − 1) 2 x ( x − 1) ( x − 2) 3
y x = y0 + x ∆y0 + ∆ y0 + ∆ y0
2! 3!
x ( x − 1) ( x − 2) ( x − 3) 4
+ ∆ y0 + …
4!
( x2 − x ) 2 x3 − 3 x2 + 2 x 3
= y0 + x ∆y0 + ∆ y0 + ∆ y0
2! 6
x4 − 6 x3 + 11x2 − 6 x 4
+ ∆ y0 , …(1)
24
neglecting terms of higher order differences.
Differentiating this w.r.t. x, 3 times, we get
3 ⋅ 2 ⋅1 3 (4 ⋅ 3 ⋅ 2 x − 6 ⋅ 3 ⋅ 2 ⋅1) 4
y3′′′ = ∆ y0 + ∆ y0
6 24
3 4
y ′′′ 3
x = ∆ y0 + ( x − )∆ y0
2
(ii) Differentiating Bessel’s formula 3 times gives us the second result.
Problem 12: Use Stirling’s formula to find the first derivatives of the function y = 2e x − x − 1
tabulated below at the point x − 0 .6.
x : 0.4 0.5 0.6 0.7 0.8
y : 1.5836 1.7974 2.0442 2.3275 2.6510
N-95
Solution: Here we require the derivative at the point x = 0 .6 which lies near the middle of the
table, so we use Stirling’s formula.
Take x = 0 .6, x0 = 0 .6, h = 1
x − x0 0 .6 − 0 .6
∴ u= = =0
h 1
The Stirling’s formula is given by
∆y + ∆y−1  u2 2 u3 − u  ∆3 y−1 + ∆3 y−2 

y = y0 + u  0 + ∆ y− 1 +  
 2  2 6  2 
u4 − u2 4
+ ∆ y− 2 + … …(1)
24
x u yu ∆ yu ∆2 yu ∆3 yu ∆4 yu
0.4 –2 1.5836
0.2138
0.5 –1 1.7974 0.0330
0.2468 0.0035
0.6 0 2.0442 0.0365 0.0002
0.2833 0.0037
0.7 1 2.3275 0.0402
0.3235
0.8 2 2.6510
Differentiating (1) w.r.t. u and putting u = 0, we get

 ∆ y0 + ∆ y−1  1  ∆3 y−1 + ∆3 y−2  
y ′ (0 ) =   −  
 2  6 2 

leaving higher order differences
 2
∆ y0 + ∆ y−1  1  ∆3 y−1 + ∆3 y−2  
=   −  
 2  6 2  
 
0 .2468 + 0 .2833  1  0 .0035 + 0 .0037  
=   −  
  2  6 2  
= [0 .26505 − 0 .0006] = 0 ⋅ 26445.

x − x0 du 1
We have u= ⇒ =
h dx h
d d du 1
∴ ( y) = { y} = yu ′
dx du dx h
1 1
Now y ′ (0 .6) = y ′(0) = (0 .26445) = 2.6445.
h 0 .1
N-96

µ 
δ y − …
1 3 1 5
y′=  δy − δ y +
h  6 30 
1  1 1 4
and y0 ′ = ( y1 − y−1 ) − (δ2 y1 − δ2 y−1 ) + (δ y1 − δ4 y−1 ) − … ⋅
2h  6 30 
(Bundelkhand 2012)
2
δ
Solution: We know that µ 2 = 1 + ⋅
4
−1 / 2
 δ2  δ2 3 4
∴ µ −1 = 1 +  =1− + δ −…
 4 8 128
Also from the equation (2) Example 8, we have

hD δ2 3 4
=1− + δ −…
δ 24 640
hD hD −1  δ2 3 4   δ2 3 4 
Now = µ = 1 − + δ − … 1 − 8 + 128 δ − …
δµ δ  24 640   
δ2 1 4
=1− + δ −…
6 30
δµ  δ2 1 4  µ  δ3 δ5 
or D= 1− + δ − … = δ − + − … ⋅
h  6 30  h  6 30 
Operating these operators on y, we get
µ  δ3 δ5 
Dy = δ − 6 + 30 − … y
h  
µ  δ3 y δ5 y 
⇒ y′= δy − + − …
h  6 30 …(1)

By definition, we have
1
δ = E 1 /2 − E −1 /2 and µ = ( E 1 /2 + E −1 /2 ).
2
1
∴ δµ = ( E − E −1 ) = µδ.
2
1 1
∴ (δµ ) y0 = ( E − E −1 ) y0 = ( y1 − y−1 )
2 2
…(2)
= µδy0
From (1), we have
1  µ δ3 µ δ5 
y0 ′ = µ δy0 − 6 y0 + 30 y0 − …
h  
1 ( y − y ) − 1 (δ2 y − δ2 y ) + 1 (δ4 y − δ4 y ) − … ⋅
= −1 −1 −1
2h  1 6
1
30
1 
N-97

dy
1. See Ex. 5. The value of at x = 4 is 2 ⋅ 8326.
dx
2. If the values of the argument are unequally spaced, then to represent the function we
should use Newton’s divided difference formula.
3. See Ex. 9, here y = f ( a).
4. See Ex. 9, here y = yn .

1. If the values of the argument are equally spaced and we require to find the derivative of the
function at a point near the beginning of a set of tabular values, we use Newton-Gregory
forward formula.
2. If the values of the argument are equally spaced and we require to find the derivative of the
function at a point near the end of a set of tabular values, we use Newton-Gregory
backward formula.
3. If the values of the argument are equally spaced, then to find the derivative at a point near
the middle of the table, we should use some central difference formula.
4. Newton’s divided difference formula should be used to represent the function if the values
of the argument are unequally spaced.
True or False
1. See article 1.
2. See article 1.
3. In case the values of the argument are unequally spaced, and we desire to find the derivative
of the function at a point, we should use Newton’s divided difference formula.
4. See Ex. 9.
❍❍❍
N-98
Chapter-7
Numerical Integration
3
Problem 1: Calculate the approximate value of ∫ x4 dx by using
−3
(i) Trapezoidal rule (Kanpur 2012)
(ii) Simpson’s rule, by dividing the range in six equal parts.
Solution: Divide the range of integration ( − 3, 3) into six equal parts each of width
3 − ( − 3)
= = 1. Hence h = 1. The values of the function for each point of sub-division are given
6
below :
x y = x4
x0 = − 3 ( − 3)4 = 81
x0 + h = − 2 ( − 2)4 = 16
x 0 + 2h = − 1 ( − 1)4 = 1
x 0 + 3h = 0 04 = 0
x 0 + 4h = 1 14 = 1
x 0 + 5h = 2 24 = 16
x 0 + 6h = 3 34 = 81
By Trapezoidal rule, we get

3 h
∫ −3 x
4
dx = [ y0 + y6 + 2 ( y1 + y2 + y3 + y4 + y5 )]
2
1
= [162 + 2 × 34] = 115.
2
By Simpson’s rule, we get
3 h
∫ −3 x
4
dx = [ y0 + y6 + 4 ( y1 + y3 + y5 ) + 2 ( y2 + y4 )]
3
1
= [162 + 4 × 32 + 2 × 2]
3
1
= × 294 = 98.
3
N-99
Problem 2(i): Use Simpson’s rule dividing the range into ten equal parts, to show that
1 log (1 + x2 )
∫0 2
1+ x
dx = 0 ⋅173.
Solution: Dividing the range of integration (0, 1) into 10 equal parts by taking h = 0 ⋅1, we
log (1 + x2 )
compute the values of the function y = at each point of subdivision.
1 + x2
log (1 + x2 )
x x2 1 + x2 log (1 + x2 ) y=
1 + x2
0 0 1 0 0 = y0
⋅1 ⋅01 1⋅01 ·0099503 ·0098518 = y1
⋅2 ⋅04 1⋅04 ·0392207 ·0377122 = y2
⋅3 ⋅09 1⋅09 ·0861776 ·0790621 = y3
⋅4 ⋅16 1⋅16 ·14842 ·1279482 = y4
⋅5 ⋅25 1⋅25 ·2231435 ·1785148 = y5
⋅6 ⋅36 1⋅36 ·3074847 ·2260916 = y6
⋅7 ⋅49 1⋅49 ·3987761 ·2676349 = y7
⋅8 ⋅64 1⋅64 ·4946962 ·301644 = y8
⋅9 ⋅81 1⋅81 ·5933268 ·3278048 = y9
1 1 2 ·6931471 ·3465735 = y10
1
By Simpson’s ‘ ’ rule, we have
3
1 log ( 1 + x2 ) h
∫0 1+ x 2
dx = [ y0 + y10 + 4 ( y1 + y3 + y5 + y7 + y9 )
3
+ 2 ( y2 + y4 + y6 + y8 )]
⋅1
= [0 + ⋅ 3465735 + 4 (⋅0098518 + ⋅ 0790621 + ⋅1785148
3
+ ⋅ 2676349 + ⋅ 3278048 + 2 (⋅0377122 + ⋅1279482
+ ⋅ 2260916 + ⋅ 301644)]
⋅1
= [⋅3465735 + 3 ⋅ 4514736 + 1⋅ 386792]
3
⋅1
= (5 ⋅184839) = ⋅17282793
3
= ⋅1728 correct to four decimal places
= ⋅173 correct to three decimal places.
N-100
6 dx
Problem 2(ii): Using Simpson’s one-third rule, find ∫0 (1 + x )2
⋅
(Bundelkhand 2011)
Solution: Dividing the range of integration (0, 6) into 6 equal parts by taking h = 1, we compute
the values of the function y = 1 / (1 + x )2 at each point of subdivision.
x 0 1 2 3 4 5 6
(1 + x )2 1 4 9 16 25 36 49
y 1 ⋅25 ⋅11111 ⋅0625 ⋅04 ⋅02778 ⋅02041

y0 y1 y2 y3 y4 y5 y6
1
3
6 dx h
∫0 (1 + x )2
= [ y0 + y6 + 4 ( y1 + y3 + y5 ) + 2 ( y2 + y4 )]
3
1
= [1 + ⋅ 02041 + 4 (⋅25 + ⋅0625 + ⋅02778) + 2 (⋅11111 + ⋅ 04)]
3
1
= [1⋅ 02041 + 1⋅ 36112 + ⋅ 30222] = 0 ⋅ 8945833
3
= 0 ⋅ 8946 correct to four decimal places.
1 x2 1
Problem 2(iii): Find the value of log 2 from ∫ dx, using Simpson’s ‘ ’ rule, by dividing the
0 1 + x3 3
range of integration into four equal parts. Also find the error.
Solution: Dividing the range of integration (0, 1) into four equal parts by taking h = ⋅ 25, we
compute the values of the function y = x2 / (1 + x3 ) at each point of sub-division.
x 0 ⋅25 ⋅50 ⋅75 1
x2 0 ·0625 ⋅25 ⋅5625 1

3
x +1 1 1·015625 1⋅125 1·421875 2
y 0 0·0615384 0·222222 0·3956044 ⋅5

y0 y1 y2 y3 y4
1
3
1 x2 h
∫0 1+ x 3
dx =
3
[ y0 + y4 + 4 ( y1 + y3 ) + 2 y2 ]
⋅25
= [0 + ⋅ 5 + 4 (⋅0615384 + ⋅ 3956044) + 2 (⋅2222222)]
3
⋅25
= [⋅5 + 1⋅ 8285712 + ⋅ 4444444]
3
⋅25
= (2 ⋅ 7730156) = 0 ⋅ 2310846
3
= 0 ⋅ 23108 correct to five decimal places.
N-101
Again actual value of the integral

1 x2 1 1
=
∫0 1+ x 3
dx =
3
[log (1 + x3 )]10 = log 2
3
1
= (0 ⋅ 6931471) = 0 ⋅ 23105 correct to five decimal places.
3
Therefore the error = 0 ⋅ 23105 − 0 ⋅ 23108 = − ⋅ 00003.
1
Remark: We have log 2 = 0 ⋅ 2310846 ⇒ log 2 = 0 ⋅ 6932538.
3
Also, from tables, log e 2 = 0 ⋅ 6931471.
Therefore, the error = 0 ⋅ 6932538 − 0 ⋅ 6931471 = 0 ⋅ 0001067.
12 dx
Problem 3(i): Evaluate ∫ by numerical methods.
5 x (Rohilkhand 2011)
1
Solution: Let y = f ( x ) = ⋅
x
Dividing the range of integration (5, 12) into seven equal parts by taking h = 1, we compute the
values of the function y = 1 / x at each point of sub-division.
x: 5 6 7 8 9 10 11 12
y: ⋅2 ⋅166667 ⋅142857 ⋅125 ⋅111111 ⋅1 ⋅090909 ⋅083333

y0 y1 y2 y3 y4 y5 y6 y7
By Trapezoidal rule, we have
12 1 h
∫5 x
dx =
2
[( y0 + y7 ) + 2 ( y1 + y2 + y3 + y4 + y5 + y6 )]
1
= [⋅2 + ⋅ 083333 + 2 (⋅166667 + ⋅142857 + ⋅125
2
+ ⋅111111 + ⋅1 + ⋅ 090909) ]
1
= (⋅283333 + 1⋅ 473088) = ⋅ 87821.
2
Also by actual integration
12 1 12
∫5 x
dx = [log x ]12
5
= log 12 − log 5 = log
5
= log 2 ⋅ 4
= ⋅ 87547.
∴ The error = ⋅ 87547 − ⋅ 87821 = − ⋅ 00274.
21
Problem 3(ii): Evaluate ∫ dx by Simpson’s rule with 4 strips (i. e., 5 ordinates)
1 x
and 8 strips (i. e., 9 ordinates) respectively. Determine the error by direct integration.
Solution: Let y = f ( x ) = 1 / x.
N-102
Case I: The range of integration (1, 2)is divided into four equal parts. In this case, h = 1 / 4 = ⋅ 25.
x 1 5 /4 3 /2 7 /4 2
y 1 ⋅8 ⋅666667 ⋅571429 ⋅5
y0 y1 y2 y3 y4
1
In this case by Simpson’s ‘ ’ rule,
3
21 h
∫
1 x
dx = [( y0 + y4 ) + 4 ( y1 + y3 ) + 2 y2 ]
3
1
= [1 + ⋅ 5 + 4 (⋅8 + ⋅ 571429) + 2 (⋅666667)]
12
1
= (1⋅ 5 + 5 ⋅ 485716 + 1⋅ 333334)
12
1
= (8 ⋅ 31905) = 0 ⋅ 693254. …(1)
12
Case II: The range of integration (1, 2) is divided into eight equal parts. In this case,
h = 1 / 8 = ⋅125.
When x = 1, y =1= y ;
0
when x = 1⋅125, y = ⋅ 888889 = y1 ;
when x = 1⋅ 25, y = ⋅ 8 = y2 ;
when x = 1⋅ 375, y = ⋅ 727273 = y3 ;
when x = 1⋅ 5, y = ⋅ 666667 = y4 ;
when x = 1⋅ 625, y = ⋅ 615385 = y5 ;
when x = 1⋅ 75, y = ⋅ 571429 = y6 ;
when x = 1⋅ 875, y = ⋅ 533333 = y7 ;
and when x = 2, y = ⋅ 5 = y8 .
1
3
21 h
∫1 x
dx = [ y0 + y8 + 4 ( y1 + y3 + y5 + y7 ) + 2 ( y2 + y4 + y6 )]
3
1/8
= [1 + ⋅ 5 + 4 (⋅888889 + ⋅ 727273 + ⋅ 615385 + ⋅ 533333)
3
+ 2 (⋅8 + ⋅ 666667 + ⋅5571429)]
1
= (1⋅ 5 + 11⋅ 05952 + 4 ⋅ 076192)
24
= 0 ⋅ 693155 …(2)
2 1
By actual integration
∫1 x
dx = [log x ]12 = log 2 = 0 ⋅ 693147.
∴ Error in case I = 0 ⋅ 693147 − ⋅ 693254 = − ⋅ 000107

and error in case II = ⋅ 693147 − ⋅ 693155 = − ⋅ 000008.
We observe that the absolute value of error in case of division into 8 strips is less than the absolute
value of error in case of division into 4 strips. Hence we can say that an
increase in the number of strips gives a better approximation.
N-103
7 dx
Problem 3(iii): Use Simpson’s rule to prove that loge 7 is approximately 1⋅ 9587 using ∫ ⋅
1 x
Solution: Divide the range (1, 7) into six equal parts each of width 1. Hence h = 1. The values
of y for each point of sub-division are given below :
x: x0 = 1 x1 = 2 x2 = 3 x3 = 4 x4 = 5 x5 = 6 x6 = 7
y: y0 = 1 y1 = 1 / 2 y2 = 1 / 3 y3 = 1 / 4 y4 = 1 / 5 y5 = 1 / 6 y6 = 1 / 7
By Simpson’s ‘1/3’ rule, we have

7 dx h
∫
1 x
= [ y0 + y6 + 4 ( y1 + y3 + y5 ) + 2 ( y2 + y4 )]
3
1
= [1⋅1428571 + 3 ⋅ 6666667 + 1⋅ 0666667]
3
1
= × 5 ⋅ 8761905 = 1⋅ 9587302 = 1⋅ 9587.
3
7 dx
The exact value of
1 x∫ = [log x ]17 = log e 7.
Thus log e 7 = 1⋅ 9587.
1 dx
Problem 4(i): Show that ∫ = log 2 = 0 ⋅ 69315.
0 1+ x (Lucknow 2008; Purvanchal 09)
Solution: To use Simpson’s ‘1/3’ rule, divide the range of integration (0, 1) into ten equal
parts each of width 1/10. Hence h = 1 / 10 = ⋅1, n = 10.
The values of y at each point of sub-division are given below :
1
x y=
1+ x
x0 = 0 1
x 0 + h = ⋅1 1 / 1⋅1 = ⋅ 9090909
x 0 + 2h = ⋅ 2 1 / 1⋅ 2 = ⋅ 8333333
x 0 + 3h = ⋅ 3 1 / 1⋅ 3 = ⋅ 7692307
x 0 + 4h = ⋅ 4 1 / 1⋅ 4 = ⋅ 7142857
x 0 + 5h = ⋅ 5 1 / 1⋅ 5 = ⋅ 6666666
x 0 + 6h = ⋅ 6 1 / 1⋅ 6 = ⋅ 6250000
x 0 + 7h = ⋅ 7 1 / 1⋅ 7 = ⋅ 5882352
x 0 + 8h = ⋅ 8 1 / 1⋅ 8 = ⋅ 5555555
x 0 + 9h = ⋅ 9 1 / 1⋅ 9 = ⋅ 5263157
x 0 + 10 h = 1 1 / 2 = ⋅ 5000000
By Simpson’s ‘1 / 3’ rule, we get

N-104
1 dx h
∫0 1 + x = 3 [ y0 + y10 + 4 ( y1 + y3 + y5 + y7 + y9 )
+ 2 ( y2 + y4 + y6 + y8 )]
⋅1
= [1⋅ 5 + 4 × 3 ⋅ 4595391 + 2 × 2 ⋅ 7281745]
3
⋅1
= × 20 ⋅ 794505 = ⋅ 6931501
3
= ⋅ 69315.
1 dx
The actual value of
∫0 1+ x
= [log (1 + x )]1
0
= log 2 − log 1 = log 2.

1
dx
Hence
∫0 1+ x
= log 2 = ⋅ 69315.
10 dx
Problem 4(ii): Calculate (upto 4 places of decimal ) ∫ by dividing the range into eight equal
2 1+ x
parts.
Solution: Divide the range of integration (2, 10 ) into eight equal parts each of width
10 − 2
= 1. Hence h = 1. The values of y for each point of sub-division are given below :
8
1
x y=
1+ x
x0 = 2 1/3
x0 + h = 3 1/4
x 0 + 2h = 4 1/5
x 0 + 3h = 5 1/6
x 0 + 4h = 6 1/7
x 0 + 5h = 7 1/8
x 0 + 6h = 8 1/9
x 0 + 7h = 9 1 / 10
x 0 + 8h = 10 1 / 11
By Simpson’s ‘1/3’ rule, we get

10 dx h
∫2 = [ y0 + y8 + 4 ( y1 + y3 + y5 + y7 ) + 2 ( y2 + y4 + y6 )]
1+ x 3
N-105
1 1 1 1 1 
+ 4  + + +
1
 + 2  + + 
1 1 1 1
= +
3  3 11  4 6 8 10   5 7 9  
1
= [⋅4242424 + 4 (⋅6416666) + 2 (⋅4539682)]
3
1
= × 3 ⋅ 8988453 = 1⋅ 2996151 = 1⋅ 2996.
3
0⋅7 1 /2
Problem 5(i): Evaluate ∫ x e − x dx approximately by using a suitable formula.
0⋅5
Solution: Divide the range of integration (⋅5, ⋅ 7) into 4 equal parts each of width
⋅7 − ⋅ 5
= ⋅ 05. Hence h = ⋅ 05. The values of y for each point of sub-division are given below :
4
x y = x1 /2 e − x
x 0 = ⋅ 50 ⋅4288818
x 0 + h = ⋅ 55 ⋅4278774
x 0 + 2h = ⋅ 60 ⋅4251076
x 0 + 3h = ⋅ 65 ⋅4208867
x 0 + 4h = ⋅ 70 ⋅4154730
By Simpson’s ‘1/3’ rule, we have

0⋅7 1 /2 h
e − x dx =
∫0⋅5 x
3
[ y0 + y4 + 4 ( y1 + y3 ) + 2 y2 ]
⋅05
= [⋅8443548 + 3 ⋅ 3950564 + ⋅ 8502152]
3
= ⋅ 0848271.
4
Problem 5(ii): Evaluate ∫ e x
dx, by Simpson’s rule, given that
0
e = 2 ⋅ 72, e2 = 7 ⋅ 39, e3 = 20 ⋅ 09, e4 = 54 ⋅ 60,
and compare it with the actual value.

Solution: Divide the whole range (0, 4) into 4 equal parts taking h = 1.
4 h
∫0 e x dx = [ y0 + y4 + 4 ( y1 + y3 ) + 2 y2 ]
3
1
= [1 + 54 ⋅ 60 + 4 (2 ⋅ 72 + 20 ⋅ 09) + 2 × 7 ⋅ 39]
3
1
= × 161⋅ 62 = 53 ⋅ 873333 = 53 ⋅ 87.
3
N-106
The actual value of

4 4
∫0 e x dx = [ e x ] = e4 − e0
0
= 54 ⋅ 60 − 1 = 53 ⋅ 60.
π /2 sin x 1
Problem 6(i): Calculate ∫ e dx correct to four decimal places by Simpson’s ‘ ’ rule, dividing
0 3
the range of integration (0, π / 2) into four equal parts.
π
Solution: Dividing the range of integration (0, π / 2) into four equal parts by taking h = , we
8
compute the values of the function y = esin x at each point of sub-division.
x sin x y = esin x
x0 = 0 0 y0 = 1
x0 + h = π / 8 0.3826834 y1 = 1⋅ 4662138
x 0 + 2h = π / 4 0.7071067 y2 = 2 ⋅ 028115
x 0 + 3h = 3 π / 8 0.9238795 y3 = 2 ⋅ 5190442
x 0 + 4h = π / 2 1 y4 = 2 ⋅ 7182818
1
By Simpson’s ‘ ’ rule, we get
3
π /2 sin x h
∫0 e dx =
3
[ y0 + y4 + 4 ( y1 + y3 ) + 2 y2 ]
π /8
= [1 + 2 ⋅ 7182818 + 4 (1⋅ 4662138 + 2 ⋅ 5190442)
3
+ 2 (2 ⋅ 028115)]
π
= [23 ⋅ 7155438] = 3 ⋅1043574
24
= 3 ⋅1044, correct to four decimal places.
π /2 3
Problem 6(ii): Calculate ∫ e sin x dx correct to four decimal places by Simpson’s ‘ ’ rule, dividing
0 8
the range of integration (0, π / 2) into 3 equal parts.
Solution: Dividing the range of integration (0, π / 2) into 3 equal parts by taking h = π / 6, we
compute the values of the function y = esin x at each point of sub-division.
x 0 π /6 π /3 π /2
sin x 0 0 ⋅5 0 ⋅ 8660254 1
y = esin x 1 1⋅ 6487213 2 ⋅ 377427 2 ⋅ 7182818

y0 y1 y2 y3
N-107
3
By Simpson’s ‘ ’ rule we have
8
π / 2 sin x 3h
∫0 e dx =
8
[ y0 + y3 + 3 ( y1 + y2 )]
3 π
= ⋅ [1 + 2 ⋅ 7182818 + 3 (1⋅ 6487213 + 2 ⋅ 377427)]
8 6
π
= [3 ⋅ 7182818 + 12 ⋅ 078445]
16
π
= [15 ⋅ 796727] = 3 ⋅10168
16
= 3 ⋅1017 correct to four decimal places.
1⋅6 x
Problem 7: Find by Weddle’s rule the value of the integral I = ∫ dx by taking 12
0⋅4 sinh x
sub-intervals.
1⋅ 6 − 0 ⋅ 4
Solution: Here h = = 0 ⋅1.
12
x
Let y= ⋅ We make the following calculations :
sinh x
x sinh x y
0 ⋅4 ⋅4107523 ⋅9738229 = y0
0 ⋅5 ⋅5210953 ⋅9595173 = y1
0 ⋅6 ⋅6366535 ⋅9424278 = y2
0 ⋅7 ⋅7585837 ⋅9227722 = y3
0 ⋅8 ⋅8881059 ⋅9007934 = y4
0 ⋅9 1⋅ 0265167 ⋅8767514 = y5
1⋅ 0 1⋅1752012 ⋅8509181 = y6
1 ⋅1 1⋅ 3356475 ⋅8235705 = y7
1⋅ 2 1⋅ 5094614 ⋅7949855 = y8
1⋅ 3 1⋅ 6983824 ⋅7654342 = y9
1⋅ 4 1⋅ 9043015 ⋅7351777 = y10
1⋅ 5 2 ⋅1292795 ⋅7044636 = y11
1⋅ 6 2 ⋅ 375568 ⋅6735231 = y12

N-108
By Weddle’s rule
1⋅6 x 3h
∫0⋅4 sinh x
dx =
10
[ y0 + y12 + 5 ( y1 + y5 + y7 + y11
+ y2 + y4 + y8 + y10 + 2 y6 + 6 ( y3 + y9 )]
3 (⋅1)
= [1⋅ 647346 + 16 ⋅ 821514 + 3 ⋅ 3733844
10
+ 1⋅ 7018362 + 10 ⋅129238]
3
= (33 ⋅ 673319) = 1⋅ 0101996.
100
π /2
Problem 8: Calculate an approximate value of ∫ sin x dx by the
0
(i) Trapezoidal rule,
1
(ii) Simpson’s ‘ ’ rule, using 11 ordinates. (Purvanchal 2011)
3
Solution: First we divide the range of integration into ten equal parts by taking the interval of
differencing h = π / 20 and then we compute the values of the function f ( x ) = sin x for each point
of sub-division. These computed values are as shown in the following table.
x y = sin x x y = sin x
x0 = 0 ⋅00000 x 0 + 6h = 6 π / 20 ⋅80902
x 0 + h = π / 20 ⋅15643 x 0 + 7h = 7 π / 20 ⋅89101
x 0 + 2h = 2 π / 20 ⋅30902 x 0 + 8h = 8 π / 20 ⋅95106
x 0 + 3h = 3 π / 20 ⋅45399 x 0 + 9h = 9 π / 20 ⋅98769
x 0 + 4h = 4 π / 20 ⋅58779 x 0 + 10 h = 10 π / 20 1⋅ 00000
x0 + 5h = 5 π / 20 ⋅70711
(i) By Trapezoidal rule, we have

π /2 h
0 ∫ sin x dx = [ y0 + y10 + 2 ( y1 + y2 + ... + y9 )]
2
( π / 20 )
= [1⋅ 00000 + 2 (5 ⋅ 85312)]
2
π
= [12 ⋅ 70624] = ⋅ 9979.
40
1
(ii) By Simpson’s ‘ ’ rule, we have
3
π /2 h
∫0
sin x dx = [ y0 + y10 + 4 ( y1 + y3 + ... + y9 )
3
+ 2 ( y2 + y4 + ... + y8 )]
π
= [1⋅ 00000 + 4 (3 ⋅19623) + 2 (2 ⋅ 65689)]
60
π
= [1⋅ 00000 + 12 ⋅ 78492 + 5 ⋅ 31378]
60
π
= [19 ⋅ 09870 ] = 1⋅ 00001.
60
N-109
Again the exact value of the integral

π /2 π /2
∫0 sin x dx = [ − cos x ]
0
= 1⋅ 0000.
Therefore the error

(i) Due to Trapezoidal rule = ⋅ 0021.
1
(ii) Due to Simpson’s ‘ ’ rule = − ⋅ 00001.
3
6 dx
Problem 9: Evaluate ∫ by using
0 1 + x2
(i) Simpson’s ‘1 / 3’ rule , (Bundelkhand 2011)
(ii ) Simpson’s ‘3/8’ rule, (Rohilkhand 2009)
(iii) Weddle’s rule. (Gorakhpur 2013)
Solution: Divide the range of integration (0, 6) into six equal parts each of width
6−0
= 1.
6
Hence h = 1. The values of y for each point of sub-division are given below :
1
x y=
1 + x2
x0 = 0 1 / 1 = 1⋅ 0000000
x0 + h = 1 1 / 2 = ⋅ 5000000
x 0 + 2h = 2 1 / 5 = ⋅ 2000000
x 0 + 3h = 3 1 / 10 = ⋅1000000
x 0 + 4h = 4 1 / 17 = ⋅ 0588235
x 0 + 5h = 5 1 / 26 = ⋅ 0384615
x 0 + 6h = 6 1 / 37 = ⋅ 0270270

6 dx h
∫0 1+ x 2
=
3
[ y0 + y6 + 4 ( y1 + y3 + y5 ) + 2 ( y2 + y4 )]
1
= [1⋅ 0270270 + 2 ⋅ 5538462 + ⋅ 517647] = 1⋅ 3661734.
3
6 dx 3h
∫0 1 + x2
=
8
[ y0 + y6 + 3 ( y1 + y2 + y4 + y5 ) + 2 y3 ]
3
= [1⋅ 0270270 + 2 ⋅ 391855 + ⋅ 200000 ]
8
3
= × 3 ⋅ 618882 = 1⋅ 3570808.
8
N-110
By Weddle’s rule, we get

6 dx 3h
∫0 =
1 + x2 10
[ y0 + y6 + 5 ( y1 + y5 ) + y2 + y4 + 6 y3 ]
3
= [1⋅ 0270270 + 2 ⋅ 6923075 + ⋅ 8588235]
10
= 1⋅ 3734474.
⋅
11
Problem 10: Using the data of the following table compute the integral ∫ xy dx, by Simpson’s rule :
0⋅5
x : 0 ⋅5 0 ⋅6 0 ⋅7 0 ⋅8 0 ⋅9 1⋅ 0 1 ⋅1
y : 0 ⋅ 4804 0 ⋅ 5669 0 ⋅ 6490 0 ⋅ 7262 0 ⋅ 7985 0 ⋅ 8658 0 ⋅ 9281
Solution: Here h = 0 ⋅1. We make the following calculations.
x : 0⋅5 0⋅6 0⋅7 0⋅8 0⋅9 1⋅0 1⋅1
xy : ⋅2402 ⋅34014 ⋅4543 ⋅58096 ⋅71865 ⋅8658 1⋅02091
1
By Simpson’s ‘ ’ rule,
3
⋅
11 h
∫0⋅5 xy dx = 3 [⋅2402 + 1⋅ 02091 + 4 (⋅34014 + ⋅ 58096 + ⋅ 8658)
+ 2 (⋅4543 + ⋅ 71865)]
⋅1
= [1⋅ 26111 + 7 ⋅1476 + 2 ⋅ 3459] = ⋅ 358487.
3
Problem 11(i): Find, from the following table, the area bounded by the curve and the x-axis from
x = 7 ⋅ 47 to x = 7 ⋅ 52.
x 7 ⋅ 47 7 ⋅ 48 7 ⋅ 49 7 ⋅ 50 7 ⋅ 51 7 ⋅ 52
f ( x) 1⋅ 93 1⋅ 95 1⋅ 98 2 ⋅ 01 2 ⋅ 03 2 ⋅ 06
7⋅52
Solution: The required area = ∫ f ( x ) dx. …(1)
7⋅74
With h = 0 ⋅ 01, by Trapezoidal rule, the value of the integral (1)

0 ⋅ 01
= [1⋅ 93 + 2 ⋅ 06 + 2 (1⋅ 95 + 1⋅ 98 + 2 ⋅ 01 + 2 ⋅ 03)] = 0 ⋅ 09965.
2
Hence the required area = 0 ⋅ 09965.
Problem 11(ii): A rocket is launched from the ground, its acceleration is registered during the first 80
seconds and is given in the table below :
t ( sec ) : 0 10 20 30 40 50 60 70 80
acc (m/sec ) 2
30 ⋅ 00 31⋅ 63 33 ⋅ 44 35 ⋅ 47 37 ⋅ 75 40 ⋅ 33 43 ⋅ 25 46 ⋅ 69 50 ⋅ 67
Find the velocity of the rocket at t = 80.

N-111
Solution: Let v denote the velocity and f denote the acceleration.

dv
Then = f.
dt
80
∴ The velocity v at t = 80 is given by v=
∫0 f dt.
Here h = 10.
1
3
80 h
0 ∫ f dt = [30 ⋅ 00 + 50 ⋅ 67 + 4 (31⋅ 63 + 35 ⋅ 47 + 40 ⋅ 33 + 46 ⋅ 69)
3
+ 2 (33 ⋅ 44 + 37 ⋅ 75 + 43 ⋅ 25)]
10
= [80 ⋅ 67 + 616 ⋅ 48 + 228 ⋅ 88] = 3086 ⋅ 77.
3
Hence velocity at time t = 80 is 3086 ⋅ 77 m / sec.
Problem 12: Using the following table of values, approximate by Simpson’s rule the arc length of the
graph y = 1 / x between the points (1, 1) and (5, 1 / 5).
x : 1 2 3 4 5
1 + x 4
  : 1⋅414 1⋅031 1⋅007 1⋅002 1⋅001
 x4 
Solution: The given curve is y = 1 / x.

dy 1
∴ =− ⋅
dx x2
 2  1 + x4 
∴
ds
=  dy   1 + 1  =
1 +    = 
 ⋅
dx  dx  x4   x4 
∴ The arc length of the curve between the points (1, 1) and (5, 1 / 5)
5  1 + x4 
=
∫1 
 x4 
 dx
h
= [1⋅ 414 + 1⋅ 001 + 4 (1⋅ 031 + 1⋅ 002) + 2 (1⋅ 007)],
3
1
by Simpson’s ‘ ’ rule, where h = 1
3
1
= (2 ⋅ 415 + 8 ⋅132 + 2 ⋅ 014) = 4 ⋅187.
3
Problem 13: Prove Simpson’s formula
b b−a
∫a f ( x ) dx =
6n
[ f ( x 0 ) + 4 f ( x1 ) + 2 f ( x2 ) + ... + f ( x2 n )],
where x 0 = a, x2 n = b (Bundelkhand 2012)

2 dx
and use it to evaluate
∫1 x
and give estimates of error for n = 1 and 2, given that
loge 2 = 0 ⋅ 69315.
N-112
Solution: We know that Simpson’s ‘1/3’ rule is

x0 + nh h
∫x0 f ( x ) dx =
3
[ f ( x 0 ) + f ( x 0 + nh)
+ 4 { f ( x 0 + h) + f ( x 0 + 3h) + …}
+ 2 { f ( x 0 + 2h) + f ( x 0 + 4h) + …}] …(1)
b−a
Putting x 0 = a, n = 2n, x 0 + 2nh = x2 n = b, h = in (1), we get the required form of
2n
Simpson’s rule.
When n = 1, we should divide the whole range (1, 2) into 2 equal parts by three points
x 0 , x1, x 2 . Thus the above formula becomes
2 dx x2 1
∫1 x
=
∫ x0 f ( x ) dx where f ( x ) =
x
, x2 = 2, x0 = 1
b−a 2 − 1 1 1 1
= [ f ( x0 ) + 4 f ( x1 ) + f ( x2 )] = + 4⋅ + 
6n 6 × 1 1 3 / 2 2 
= ⋅ 69444.
When n = 2, divide the whole range (1, 2) into 4 equal parts by the points x 0 , x1, x 2 , x 3 , x 4 .
2 dx x4 b−a
We have
∫1 x
=
∫ x0 f ( x ) dx =
6n
[ f ( x 0 ) + 4 f ( x1 ) + 2 f ( x 2 ) + 4 f ( x 3 ) + f ( x 4 )]
1 1 1 1 1 1
= + 4⋅ + 2⋅ + 4⋅ +  = ⋅ 69325.
12 1 5 /4 3 /2 7 / 4 2 
2 dx 2
The actual value of
∫1 x
= [log e x ] = log e 2 = ⋅ 69315.
1
Hence the error when n = 1 is = ⋅ 69315 − ⋅ 69444 = − ⋅ 00129.

The error for n = 2 is given by ⋅69315 − ⋅ 69325 = − ⋅ 0001.
2⋅5 1
Problem 14: Prove that ∫1⋅5 U x dx =
24
[U1 + 22U2 + U3 ] ⋅
Solution: Changing the scale to 2, we have to prove that

5 2 1
∫3 U x dx = 24 [U2 + 22U4 + U6 ] = 12 [U2 + 22U4 + U6 ]⋅
Again shifting the origin to 2 we have to prove that
3 1
∫1 U x dx =
12
[U0 + 22U2 + U4 ] ⋅
Now proceed as in solved example 19.
Problem 15: If f ( x) is a polynomial of degree 2, prove that

1 1
(i)
∫0 f ( x ) dx =
12
[5 f (0 ) + 8 f (1) − f (2)] ⋅
1 1
(ii)
∫0 f ( x ) dx =
12
[5 f (1) + 8 f (0 ) − f ( − 1)] ⋅
N-113
1
Solution: (i) Here we are to express ∫ f ( x ) dx in terms of f (0 ), f (1) and f (2).
0
By Lagrange’s interpolation formula, we have
( x − 1) ( x − 2) ( x − 0 ) ( x − 2) ( x − 0 ) ( x − 1)
f ( x) = f (0 ) + f (1) + f (2)
(0 − 1) (0 − 2) (1 − 0 ) (1 − 2) (2 − 0 ) (2 − 1)
1 2 1
= ( x − 3 x + 2) f (0 ) − ( x2 − 2 x ) f (1) + ( x2 − x ) f (2).
2 2
1 1
1 1  x3 3 2   x3 
∴
∫0 f ( x ) dx =
2
f (0 ) 
 3
− x + 2 x  − f (1) 
2 0  3
− x2 
0
1
1  x3 x2 
+ f (2)  −
2 3 2 
0
5 2 1
= f (0 ) + f (1) − f (2)
12 3 12
1
= [5 f (0 ) + 8 f (1) − f (2)].
12
1
(ii) Here we are to express
∫0 f ( x ) dx in terms of f (1), f (0 ) and f ( − 1).

( x − 0 ) ( x + 1) ( x − 1) ( x + 1) ( x − 0 ) ( x − 1)
f ( x) = f (1) + f(0 ) + f ( − 1)
(1 − 0 ) (1 + 1) (0 − 1) (0 + 1) ( − 1 − 0 ) ( − 1 − 1)
1 2 1
= ( x + x ) f (1) − ( x2 − 1) f (0 ) + ( x2 − x ) f ( − 1).
2 2
1 1 1 1
∴
∫0
f ( x ) dx =
2
f (1)
0 ∫
( x2 + x ) dx − f (0 )
0 ∫
( x2 − 1) dx
1 1 2
+
2
f ( − 1)
∫0
( x − x ) dx
1 1 1
1  x3 x2   x3  1  x3 x2 
= f (1)  +  − f (0 )  − x + f ( − 1)  −
2 3 2 3 0 2 3 2 
0 0
5 2 1
= f (1) + f (0 ) − f ( − 1)
12 3 12
1
= [5 f (1) + 8 f (0 ) − f ( − 1)] ⋅
12
Problem 16(i): Prove that

2 1   1
f  −  + 23 f   + 23 f   + f    ,
1 3 5
∫0 f ( x ) dx =
24   2   2  2  2  
if third differences are constant.

2
f ( x ) dx in terms of f  −  , f   , f   ,
1 1 3
Solution: Here we are to obtain a formula for ∫
0  2  2  2
f   ⋅ Since we are known four pairs of values of x and f ( x ), therefore f ( x ) can be represented as
5
 2
N-114
a polynomial in x of degree 3. Consequently third differences of f ( x ) are constant and fourth and
higher differences of f ( x ) are all zero.
1
We have x0 = − and h = 1.
2
f ( x ) = f  − + x + 
1 1
Now [Form f ( x0 + uh)]
 2 2
= E x + (1 /2) f  −  = (1 + ∆ ) x + (1 /2) f  − 
1 1
 2  2
 x + 1   x − 1 
 1  1  1  2  2  ∆2 f  − 1 
= f  −  +  x +  ∆f  −  +  
 2  2  2 1⋅ 2  2
 x + 1   x − 1   x − 3 
 2  2  2
∆3 f  −  ⋅
1
+
1⋅ 2 ⋅ 3  2
[∵ Fourth and higher differences of f ( x ) all vanish]

2  2
2  1  1  1  x − 1 / 4 ∆2 f  − 1 
∴
∫0 f ( x ) dx =
∫0  f  − 2  +  x + 2  ∆ f  − 2  +
 2  2
1 3 3 2 1 1 
 x − x − x +  ∆ f  −  dx
3 3
+
6 2 4 8  2  
2 2
1  x2 1  1  x3 1 
= f  −  [ x]2 + ∆ f  −   + x  + ∆2 f  −  
1 1
− x
 2 0  2 2 2  2  2 3 4 
 0  0
2
1 3  1   x4 x3 x2 3 
+ ∆ f −   − − + x
6  2 4 2 8 8 
 0
= 2 f  −  + 3 ∆ f  −  +
1 1 13 2  1  1 3  1
∆ f −  + ∆ f − 
 2  2  12  2  24  2
= 2 f  −  + 3  f   − f  −   +
1 1 1 13   3   1  + f  − 1  
f   −2f
 2   2   2   12   2   2  2  
1   5  3  + 3 f  1  − f  − 1  
+ f   −3f
24   2   2  2  2  
f  −  +
23  1  23  3 
f  
1 1 1 5
= f   + f   +
24  2  24  2  24  2  24  2
1   1
f  −  + 23 f   + 23 f   + f    ⋅
1 3 5
=
24   2   2  2  2  
Problem 16(ii): Obtain the approximate quadrature formula

1 /2 1   1  1  1 ∆ f  − 3  − ∆ f  1   ⋅
∫−1 /2 f ( x ) dx =  f  −  + f    +
2   2 2  24

  2  2  
Solution: We have ∆ f  −  = f  −  − f  − 
3 1 3
 2  2  2
N-115
∆ f   = f   − f   ⋅
1 3 1
and
 2  2  2
1 /2
So here we are to obtain a formula for
∫−1 /2 f ( x ) dx in terms of
f  −  , f  −  , f   , f   ⋅
3 1 1 3
 2  2  2  2
Since we are known four pairs of values of x and f ( x ), therefore f ( x ) can be represented as a
polynomial in x of degree 3. Consequently third differences of f ( x ) are constant.
3
Here x0 = − and h = 1.
2
f ( x ) = f  − + x + 
3 3
Now [Form f ( x 0 + uh)]
 2 2
= E x + (3 /2) f  − 
3
 2
= (1 + ∆ ) x + (3 /2) f  − 
3
 2
 x + 3   x + 1 
 3  3  3  2  2  ⋅ ∆2 f  − 3 
= f −  +  x +  ∆ f −  +  
 2  2  2 1⋅ 2  2
 x + 3   x + 1   x − 1 
 2  2  2  ∆3 f  − 3 
+  
1⋅ 2 ⋅ 3  2
[∵ Fourth and higher differences of f ( x ) all vanish]
= f  −  +  x +  ∆ f  −  +  x2 + 2 x +  ∆2 f  − 
3 3 3 1 3 3
 2  2  2 2  4  2
1 3 3 2 1
 x + x − x −  ∆ f  −  ⋅
3 3 3
+
6 2 4 8  2
1 /2 3 1 /2 3 1 /2 
f ( x ) dx = f  −  dx + ∆ f  −   x +  dx
3
∴
∫−1 /2 ∫
 2  −1 / 2  2  −1 / 2  ∫
2
1 2  3  1 /2  2
 x + 2 x +  dx
3
+
2
∆ f − 
 2  −1 / 2  ∫ 4
1 3  3  1 /2  3 3 2 1
 x + x − x −  dx
3
+
6
∆ f − 
∫
 2  −1 / 2  2 4 8
3 1 /2 3 1 /2 3
= 2 f  −  dx + 2 ∆ f  − 
 2 0 ∫  2 0 2
dx
∫
1 2  3  1 /2  2 3  3 1 /2  3 2 3 
 x +  dx +2 ⋅ ∆ f  − 
1 3
+ 2⋅
2
∆ f − 
 2  0  ∫ 4  6 
 x −  dx
2 0  2 8 ∫
 0, if f ( − x ) = − f ( x )
∵ a  
 ∫ f ( x ) dx =  a
−a
 ∫
2 0 f ( x ) dx, if f ( − x ) = f ( x ) 

N-116
= f  −  + ∆ f  −  +
3 3 3 10 2  3  1 3  3
∆ f −  − ∆ f − 
 2 2  2  24  2  24  2
= f  −  +  f  −  − f  −   + f   − 2 f  −  + f  −  
3 3 1 3 5   1 1 3
 2  2   2   2   12   2   2  2  
1   3
f   − 3 f   + 3 f  −  − f  −  
1 1 3
−
24   2   2  2  2  
f  −  +
13  1  13  1 
f  
1 3 1 3
=− f −  + f   −
24  2  24  2  24  2  24  2 
12  1  12  1 
f  −  + f   − f  −  − f  
1 1 1 1 1 3 1 3
= f −  + f   +
24  2  24  2  24  2  24  2  24  2  24  2 
=
1   1  1   1   f  − 1  − f  − 3   −  f  3  − f  1  
 f  −  + f    +  
 2     2   2  
2   2 2  24    2 
1   1  1  1   3  1 
=  f  −  + f    + ∆ f  −  − ∆ f    ⋅
2   2 2  24  2 2 
Problem 17: Derive the following quadrature formula :

b a+ b
∫ −a f ( x ) dx =
6ab
[ b (2a − b ) f ( − a) + ( a + b )2 f (0 ) + a (2b − a) f ( b )] .
b
Solution: Here we are to express ∫ f ( x ) dx in terms of f ( − a), f (0 ) and f ( b ).
−a

( x − 0 ) ( x − b) ( x + a) ( x − b ) ( x + a) ( x − 0 )
f ( x) = f ( − a) + f (0 ) + f (b)
(− a − 0 ) (− a − b) (0 + a) (0 − b ) ( b + a) ( b − 0 )
1 1 2
= ( x2 − bx ) f ( − a) − [ x + ( a − b ) x − ab ] f (0 )
a (a + b) ab
x2 + ax
+ f ( b ).
b ( b + a)
b
b 1  x3 bx2 
∴
∫−a f ( x ) dx =
a (a + b)
f ( − a) 
3
−
2 
−a
b b
1  x3 x2  1  x3 ax2 
− f (0 )  + (a − b) − abx  + f (b)  +
ab 3 2  − a b ( b + a) 3 2 
−a
1  b3 b3 a3 ba2 
=  − + + f ( − a)
a (a + b)  3 2 3 2 
1  b3 b2 a3 a2 ( a − b ) 
−  + (a − b) − ab2 + − − a2 b  f (0 )
ab  3 2 3 2 
1  b3 ab2 a3 a3 
+  + + − f (b)
b ( b + a)  3 2 3 2 
N-117
1 1
= (2a3 + 3a2 b − b3 ) f ( − a) + ( a3 + 3a2 b + 3ab2 + b3 ) f (0 )
6a ( a + b ) 6ab
1
+ (2b3 + 3b2 a − a3 ) f ( b )
6 b ( b + a)
1 1
= ( a + b )2 (2a − b ) f ( − a) + ( a + b )3 f (0 )
6a ( a + b ) 6ab
( b + a)2 (2b − a)
+ f (b)
6 b ( b + a)
a+ b
= [ b (2a − b ) f ( − a) + ( a + b )2 f (0 ) + a (2b − a) f ( b )].
6ab
Problem 18: Prove that if f ( x) is a continuous function whose fifth differences are constant, then
1 8 5
∫−1 f ( x ) dx =
9
f (0 ) + [ f { √ (0 ⋅ 6)} + f { − √ (0 ⋅ 6)}] ⋅
9
Solution: Proceed as in problem 17.
Problem 19: Obtain the approximate quadrature formula

n 3 1
f ( x ) dx = n  f (0 ) + {19 f (n) − 5 f (2n) + f (3n)} ⋅
∫0  8 24 
n
Solution: Here we are to obtain a formula for ∫0 f ( x ) dx in terms of f (0 ), f ( n), f (2n),
f (3n). Since we are given four pairs of values of x and f ( x ) i. e.,(0, f (0 )), ( n, f ( n)), (2n, f (2n)) and
(3n, f (3n)), therefore f ( x ) can be represented as a polynomial in x of degree 3. Consequently
fourth and higher differences of f ( x ) are all zero.
We have x 0 = 0 and h = n.
n
Let I =
∫0 f ( x ) dx.
x−0
Put = u i. e., x = nu.
n
Then dx = n du.
1
∴ I =n
∫0 f ( nu) du
1
=n
∫0 E u f (0 ) du [∵ E u f (0 ) = f (0 + hu) = f (0 + nu) = f ( nu)]
1
=n
∫0 (1 + ∆ )u f (0 ) du
1  u ( u − 1) 2 u ( u − 1) ( u − 2) 3
∆ f (0 )  du
=n
∫0 

f (0 ) + u ∆ f (0 ) +
1⋅ 2
∆ f (0 ) +
1⋅ 2 ⋅ 3 
[∵ Fourth and higher differences of f ( x ) are all zero]

1 1
 u2   u3 u2 
= n  f (0 ) ⋅ ( u) + ∆ f (0 ) ⋅   + ∆2 f (0 ) ⋅ 
1 1
− 
 0  2 0 2  3 2
0
1
 u4 
− u3 + u2  
1 3
+ ∆ f (0 ) ⋅ 
6  4  0 
N-118
1 1 2 1 3
= n  f (0 ) + ∆ f (0 ) − ∆ f (0 ) + ∆ f (0 )
 2 12 24 
1 1
= n  f (0 ) + { f ( n) − f (0 )} − { f (2n) − 2 f (n) + f (0 ) }
 2 12
{ f (3n) − 3 f (2n) + 3 f (n) − f (0 )} 

1
+
24 
= n  1 − −
1 1 1  1 + 1 + 1  f ( n) −  1 + 1  f ( 2n) + 1 f (3n)
−  f (0 ) +
  2 12 24   2 6 8  12 8  24 
3 1
= n  f (0 ) + {19 f ( n) − 5 f (2n) + f (3n)} ⋅
 8 24 

1. In trapezoidal rule we assume that y is a polynomial in x of degree one.
2. In Simpson’s one-third rule we assume that y is a polynomial in x of degree two.
3. In Simpson’s three-eighth rule we assume that y is a polynomial in x of degree
three.
4. In Weddle’s rule we assume that y is a polynomial in x of degree six.
5. If Ckn are Cote’s numbers, then
n
Σ Ckn = 1.
k =0
6. In Simpson’s one-third rule we assume y is a polynomial in x of degree two.
7. See article 6.
8. Newton Raphson method is used for finding successively better approximations to the
roots of real-valued function.
9. See article 5.
10. See article 3.

1. The general quadrature formula is
x0 + nh  n2  n3 n2  ∆2 y0
+ … upto ( n + 1) terms .
∫x0 y dx = h  ny0 +
 2
∆ y0 + 
 3
−
2

 2 ! 
See article 2.
2. The trapezoidal rule is applicable to any number of sub-intervals whether even or odd.
3. The formula obtained by putting n = 2 in the general quadrature formula is Simpson’s
one-third rule.
N-119
4. Weddle’s rule is obtained by putting n = 6 in the general quadrature formula.
5. In Cote’s method we integrate Lagrange’s interpolation formula.

6. The numbers Ckn , 0 ≤ k ≤ n, are called Cote’s numbers.
n
7. If Ckn are Cote’s numbers, then Σ Ckn = 1.
k= 0
True or False
1. We know that the process of computing the value of a definite integral from a set of
numerical values of the integrand is called Numerical integration.
2. Simpson’s one-third rule requires the division of the whole range of integration into an
even number of sub intervals of width h.
3. Trapezoidal rule is
xn h
∫x0 y dx =
2
[ y0 + 2 ( y1 + y2 + … + yn −1 ) + yn ].
4. In general quadrature formula we integrate Newton’s formula for forward integration.

5. In Cote’s method we integrate Lagrange’s interpolation formula.
6. Simpson’s three-eighth rule is obtained by putting n = 3 in the general quadrature formula.
❍❍❍
N-120
Chapter-8
Errors in Numerical Calculations
Problem 1: Compute an approximate value of (1⋅ 04) 3⋅01.
Solution: Let f ( x, y ) = x y. ...(1)

Also let x = 1, ∂x = ⋅ 04, y = 3, δ y = ⋅ 01. ...(2)
Differentiating (1) partially w.r.t. x and y respectively, we get
∂f ∂f
= y x y − 1 and = x y log x.
∂x ∂y
∂f ∂f
By total differentiation, df = dx + dy = y x y −1 dx + x y log x dy
∂x ∂y
= (3) (1) (⋅04) + (1)3 log (1) (⋅01) = 0 ⋅12. [Using (2)]
Hence, (1⋅ 04) 3⋅01 = f (1, 3) + df = 1 + 0 ⋅12 = 1⋅12.
Problem 2: The deflection at the centre of a rod of length l and diameter d supported at its ends and loaded at
the centre with a weight w varies as wl3 d − 4 . What is the percentage increase in the deflection corresponding to
the percentage increase in w, l and d of 3, 2 and 1 respectively ?
w l3
Solution: Let D be the deflection of the rod at the centre, then D = k
d4
or log D = log k + log w + 3 log l − 4 log d.
δD δw δl δd
Differentiating, we get = +3 −4
D w l d
100 δ D 100 δ w 100 δ l δd
or = +3× − 4 × 100 = 3 + 3 × 2 − 4 × 1 = 5.
D w l d
Hence, the required percentage increase in the deflection is 5%.
Problem 3: Show that the acceleration due to gravity is reduced nearly by 1% at an altitude equal to
0 ⋅ 5% of earth’s radius. Given that at an external point x kilometers from the earth’s centre, such an
2
acceleration is given by g   , where r is the radius of the earth.
r
 x
Solution: Let a be the acceleration due to gravity, then we have a = g ( r / x)2 . Taking log , we
get log a = log g + 2 log r − 2 log x .
δa δx
∴ = − 2 , since g and r are constants
a x
δa δx Given δx × 100 = 0 ⋅ 5
i. e., × 100 = − 2 × 100 = − 2 × 0 ⋅ 5 = − 1.
a x  x 
Thus the acceleration is nearly reduced by 1 percent.

N-121
Problem 4: In estimating the number of bricks in a pile which is measured to be 5m × 10 m × 5m, count of
bricks is taken as 100 bricks per m3 . Find the error in the cost when the tape is stretched 2% beyond its
standard length. The cost of bricks is ` 2,000 per thousand bricks.
Solution: Let the length, breadth and height of the pile be x, y and z respectively so that its
volume V = x y z .
Taking log, we get log V = log x + log y + log z
δV δ x δ y δ z
∴ = + +
V x y z
δ V 100 δ x 100 δ y 100 δ z
or 100 = + + =2+ 2+ 2=6
V x y z
6V 6 (5 × 10 × 5)
or δV = = = 15 m3 .
100 100
Number of bricks in δ V = 15 × 100 = 1500.
1500 × 2000
Thus the error in the cost = = ` 3000
1000
which results in a loss to the brick seller.
Problem 5: Find expression for the remainder term in Bessel's formula when terminated with a difference
of even order.
Problem 6: Use the series

1+ x  x3 x5 
log e   = 2  x + + + …
1 − x  3 5 
 
to compute the value of log (1⋅2) correct to seven decimal places and find the number of terms retained.
Solution: We have
 1+ x   x3 x5 x2 n −1 
log e   = 2  x + + + …+  + Rn( x )
 1− x   3 5 2n −1 
If we retain n terms then, we have

2 x2 n +1  1+ θ 
Rn( x ) = log e   ; 0 < θ < x.
2n + 1  1− θ 
This gives the maximum absolute error (at θ = x )

2 x2 n +1  1+ x 
= log e  ⋅
2n + 1  1− x 
Hence the maximum relative error

2
= x2 n +1.
2n + 1
1+ x 1
Let =1⋅ 2 i. e., x = ⋅
1− x 11
N-122
2 n +1
1 2 1
Now ( ER )max. at x = is =   .
11 2n + 1  11
For seven decimal accuracy, we have

2 n +1
2 1 1
.  < × 10 −7
2n + 1  11 2
or (2n + 1)(11)2 n +1 > 4 × 10 −7,
which gives n ≥ 3.
Hence retaining the first three terms of the given series, we have
 x3 x5 
 at  x =  = 0 ⋅1823215.
1
log e(1⋅ 2) = 2 x + +
 3 5   11
Problem 7: Derive the expression for the remainder in Newton's forward formula

1. True value of 2 = 1.41421356.
Approximate value of 2 = 1.414.

∴ Error = True value − Approximate value
= 2 −1.414 = 1.41421356 − 1.414
= 0 .00021356.
Absolute error E A = Error
= 0 .00021356 = 0 .21356 × 10 −3
EA 0 .21356 × 10 −3
Relative error ER = =
True Value 2
= 0 .151 × 10 −3 .
2. See article 4.
2
3. True value of = 0 .666666
3
2
Approximate value of = 0 .667
3
∴ Absolute error E A = True value − Approximate value
= .666666 − .667 = . 000334
EA .000334
Relative error ER = = = 0 .0005.
True value .666666
N-123
1. See article 4.
2. See article 4.
True or False
1. See article 3.
2. Here E A = 625483
. − 625 = 0.483
EA .483
and ER = = =.000772.
625.483 625.483
EP = ER × 100 = .077%.
❍❍❍
N-124
Chapter-9
Simultaneous Linear Algebraic Equations
Problem 1: Solve the following equations by factorization method i. e., by triangularization method :
x1 + x 2 + x 3 = 1, 3 x1 + x 2 − 3 x 3 = 5, x1 − 2 x 2 − 5 x 3 = 10.
Solution: The given equations can be written as a single matrix equation

 1 1 1   x1   1 
3 1 −3   x2  =  5 
    
 1 −2 −5   x3  10 
1 1 1  x1   1
i. e., AX = B, where A =  3 1 −3  , X =  x2 , B =  5 ⋅
     
 1 −2 −5   x3   10 
 1 0 0   u11 u12 u13 

Let A = LU , where L =  l21 1 0  and U =  0 u22 u23 .
   
 l31 l32 0   0 0 u33 
 1 0 0   u11 u12 u13  1 1 1

Then l 1 0  0 u22 u23 = 3 1 −3  ⋅
 21    
 l31 l32 1   0 0 u33   1 −2 −5 
 u11 u12 u13 

∴ l u l21 u12 + u22 l21 u13 + u23 
 21 11 
 l31 u11 l31 u12 + l32 u22 l31 u13 + l32 u23 + u33 
1 1 1
=3 1 −3  ⋅
 
 1 −2 −5 
Equating the corresponding elements on both sides, we get

u11 = 1, u12 = 1, u13 = 1,
l21 u11 = 3 ⇒ l21 = 3, l21 u12 + u22 = 1 ⇒ u22 = 1 − 3 = − 2 ,
l21 u13 + u23 = − 3 ⇒ u23 = − 3 − 3 = − 6,
3,
l31 u11 = 1 ⇒ l31 = 1, l31 u12 + l32 u22 = − 2 ⇒ l32 =
2
l31 u13 + l32 u23 + u33 = − 5 ⇒ u33 = 3.
 
1 0 0   1 1 1
∴ L =  3 1 0  ,U =  0 −2 −6  ⋅
   
 3   0 0 3 
 1 1 
2
N-125
Now AX = B
⇒ LUX = B [ ∵ A = LU ]
 y1 
⇒ LY = B, where UX = Y =  y2 
 
 y3 
 
 1 0 0   y1   1 
⇒ 3 1 0  y = 5
  2   
 1 3 1   y3   10 
 
2
 
 y1   1
⇒  3 y1 + y2 = 5
   
 3   10 
 y1 + y 2 + y3 
2
3
⇒ y1 = 1, 3 y1 + y2 = 5, y1 + y2 + y3 = 10
2
⇒ y1 = 1, y2 = 2, y3 = 6.
Again UX = Y
 1 1 1  x1   1
⇒ 0 −2 −6   x2 =2
     
 0 0 3   x3   6 
 x1 + x2 + x3   1
⇒  − 2 x − 6x =2
 2 3   
 3 x3   6 
⇒ x1 + x2 + x3 = 1, − 2 x2 − 6 x3 = 2 , 3 x3 = 6
⇒ x3 = 2 , x2 = − 7, x1 = 6.
Hence, the required solution of the given system of equations is
x1 = 6, x2 = − 7, x3 = 2 .
Problem 2: Solve the system

x + (1 / 2) y + (1 / 3)z = 1
(1 / 2) x + (1 / 3) y + (1 / 4)z = 0
(1 / 3) x + (1 / 4) y + (1 / 5)z = 0
by Gauss elimination method.
Solution: The given equations are

x + (1 / 2) y + (1 / 3)z = 1
(1 / 2) x + (1 / 3) y + (1 / 4)z = 0
(1 / 3) x + (1 / 4) y + (1 / 5)z = 0
To eliminate x from the second and third equations, performing

1 1
R2 → R2 − R1 , R3 → R3 − R1 ,
2 3
N-126
the given equations are equivalent to

x + (1 / 2) y + (1 / 3)z = 1
(1 / 12) y + (1 / 12)z = ( −1 / 2)
(1 / 12) y + (4 / 45) = ( −1 / 3)
To eliminate y from the third equation, performing R3 → R3 − R2 , the given equations are
equivalent to
x + (1 / 2) y + (1 / 3)z = 1 ...(1)
(1 / 12) y + (1 / 12)z = ( −1 / 2) ...(2)
+ (1 / 180 )z = (1 / 6) ...(3)
From (3), we have z = 30.

Putting z = 30 in (2), we get
 1  y +  5 = −  1 or  1  y = −3 or y = − 36.
       
 12   2  2  12 
Putting z = 30, y = − 36 in (1), we get

x − 18 + 10 = 1 or x = 9.
Hence, the required solution of the given system of equations is x = 9, y = − 36, z = 30.
Problem 3: Solve by Gauss’s elimination method the following :

6x + 3y + 2z = 6,
6 x + 4y + 3z = 0,
20 x + 15 y + 12 z = 0.
6 x + 3y + 2z = 6
6 x + 4y + 3z = 0
20 x + 15 y + 12 z = 0.
20
Performing R2 → R2 − R1 , R3 → R3 − R1 , the given equations are equivalent to
6
6 x + 3y + 2z = 6
y + z = −6
5y + 16 = −20
z
3
or
6 x + 3y + 2z = 6 …(1)
y + z = −6 …(2)
1 = 10, …(3)
z
3
by R3 → R3 − 5R2
From (3), we have z = 30.
Putting z = 30 in (2), we get
y = − 6 − 30 = − 36.
Putting y = − 36, z = 30 in (1), we get
6 x = 6 + 108 − 60 = 54 or x = 9.
x = 9, y = − 36, z = 30.
N-127
Problem 4: Factorize the matrix

5 −2 1
A =7 1 −5 
 
 3 7 4 
into the form LU, where L is unit lower triangular matrix and U is upper triangular matrix and hence solve
the system of equations
5 x − 2 y + z = 4, 7 x + y − 5z = 8, 3 x + 7 y + 4z = 10.
Solution: The given system of equations in matrix form AX = B is
5 −2 1  x   4
7 1 −5   y  =  8 ⋅
     
 3 7 4   z   10 
Now let A = LU
5 −2 1  1 0 0   u11 u12 u13 
i. e., 7 1  
−5 = l21 1 0  0 u22 u23 
    
 3 7 4   l31 l32 1   0 0 u33 
5 −2 1   u11 u12 u13 

or 7 1 −5  =  u11 l21 u12 l21 + u22 l21 u13 + u23 ⋅
   
 3 7 4   u11 l31 u12 l31 + u22 l32 u13 l31 + l32 u23 + u33 

u11 = 5 ; u12 = − 2 ; u13 = 1
7 3
u11 l21 = 7 ⇒ l21 = ; u11 l31 = 3 ⇒ l31 = ;
5 5
7 19
u12 l21 + u22 = 1 ⇒ − 2 × + u22 = 1 ⇒ u22 = ;
5 5
3 19 41
u12 l31 + u22 l32 = 7 ⇒ − 2 × + × l32 = 7 ⇒ l32 = ;
5 5 19
7 32
l21 u13 + u23 = − 5 ⇒ ⋅ 1 + u23 = − 5 ⇒ u23 = − ;
5 5
3 41  32  327
u13 l31 + l32 u23 + u33 = 4 ⇒ 1 ⋅ + ⋅ −  + u33 = 4 ⇒ u33 = ⋅
5 19  5  19
 1 5 −2 
0 0  
1

  19 32
7   − 
∴ L= 1 0  and U =  0 5 5 ⋅
5  
  327
 3 41  0 0 
 5 19 1   19 
 
∵ AX = B and A = LU , therefore
LUX = B.
Now let
UX = Y so that LY = B
N-128
 
 1 0 0   y1   4 
7 
i. e.,  1 0   y2   8 
5   
 3 41   y  10 
1    
3

 5 19 
7 12
giving y1 = 4 ; y1 + y 2 = 8 ⇒ y 2 =
5 5
3 41 3 41 12 46
and y1 + y 2 + y 3 = 10 ⇒ ⋅ 4 + ⋅ + y3 = 0 ⇒ y3 = ⋅
5 19 5 19 5 19
Now UX = Y implies
   
 5 −2 1  4 
   x   
 0 19 − 32   y  =  12 
 5 5     5 
  z  
327  46 
0 0   
 19   19 
327 46 46
giving z= or z = ;
19 19 327
19 32 12 19 32 46 12 284
y− z= or y− × = or y = ;
5 5 5 5 5 327 5 327
284 46 122
and 5 x − 2 y + z = 4, or 5 x − 2 × + = 4 or x = ⋅
327 327 109
∴ The solution of the given system of equations is
122 284 46
x= , y= and z = ⋅
109 327 327
Problem 5: By Gauss’s elimination method, solve

5 x − y − 2 z = 142 , x − 3 y − z = − 30, 2 x − y − 3z = − 5.
Solution: The given equations can be written as
x − 3y − z = − 30
2x − y − 3z = −5
5x − y − 2z = 142 .
To eliminate x from the second and third equations with the help of first equation we perform
R2 → R2 − 2 R1 and R3 → R3 − 5R1 . So the given equations are equivalent to
x −3y − z = − 30
5y − z = 55
14 y + 3z = 292
or
x −3y − z = − 30
1
y − (1 / 5) z = 11 , by R2 → R2
5
14 y + 3z = 292.
N-129
Now to eliminate y from the third equation with the help of second equation we perform
R3 → R3 − 14 R2 . So the given equations are equivalent to
x −3y − z = − 30 …(1)
y − (1 / 5) z = 11 …(2)
(29 / 5) z = 138. …(3)
The coefficient matrix of these equations is upper triangular.

5 × 138 690
From (3), we have z= = = 23 ⋅ 79.
29 29
Putting z = 23 ⋅ 79 in (2), we get
1
y = 11 + × 23 ⋅ 79 = 15 ⋅ 76.
5
Putting z = 23 ⋅ 79, y = 15 ⋅ 76 in (1), we get
x = 3 × 15 ⋅ 76 + 23 ⋅ 79 − 30 = 41⋅ 07.
x = 41⋅ 07, y = 15 ⋅ 76, z = 23 ⋅ 79.
Problem 6: Solve the system

2 x1 + 4 x 2 + x 3 = 3,
3 x1 + 2 x 2 − 2 x 3 = − 2 ,
x1 − x 2 + x 3 = 6
by Gauss’s elimination method.

x1 − x2 + x3 = 6
2 x1 + 4 x2 + x3 = 3
3 x1 + 2 x2 − 2 x3 = −2
or x1 − x2 + x3 = 6
6 x2 − x3 = −9 , by R2 → R2 − 2 R1
5 x2 − 5 x3 = −20 R3 → R3 − 3R1
or
x1 − x2 + x3 = 6
by R2 → 1 / 6 R2
x2 − (1 / 6) x3 = ( −3 / 3) ,
R3 → 1 / 5 R3
x2 − x3 = −4
or
x1 − x2 + x3 = 6 …(1)
x2 − (1 / 6) x3 = − 3 /2, …(2)
− (5 / 6) x3 = − 5 /2 …(3)
by R3 → R3 − R2
From (3), we have x 3 = 3.

Putting x 3 = 3 in (2), we get x 2 = − 1.
Putting x 2 = − 1, x 3 = 3 in (1), we get x1 = 2 .
x1 = 2 , x 2 = − 1, x 3 = 3.
N-130
Problem 7: Solve by Gauss-elimination method

2 x + y + 4 z = 12 , 8 x − 3 y + 2 z = 23, 4 x + 11 y − z = 33.

2x + y + 4z = 12
8 x − 3y + 2z = 23
4 x + 11 y − z = 33
To eliminate x from the second and third equations with the help of first equation we perform
R2 → R2 − 4 R1 and R3 → R3 − 2 R1 . So the given equations are equivalent to
2x + y + 4z = 12
− 7y − 14 z = −25
9y − 9z = 9
or
2x + y + 4z = 12
by R2 → ( −1 / 7) R2
y + 2z = (25 / 7) ,
R3 → (1 / 9) R3
y − z = 1
or
2x + y + 4z = 12 …(1)
y + 2z = (25 / 7) …(2)
− 3z = ( −18 / 7) , by R3 → R3 − R2 . …(3)
From (3), we have

18 6
z= = = 0 ⋅ 857.
21 7
Putting z = 0 ⋅ 857 in (2), we get
25
y= − 1⋅ 714 = 3 ⋅ 571 − 1⋅ 714 = 1⋅ 857.
7
Putting y = 1⋅ 857, z = 0 ⋅ 857 in (1), we get
2 x = 12 − 1⋅ 857 − 3 ⋅ 428 = 6 ⋅ 715.
∴ x = 3 ⋅ 358.
x = 3 ⋅ 358, y = 1⋅ 857, z = 0 ⋅ 857.
Problem 8: Solve the system of equations

2 x − 3 y + 10 z = 3, − x + 4 y + 2 z = 20, 5 x + 2 y + z = − 12
by factorization method.
Solution: Let us put the given system of equations in the matrix form AX = B
 2 −3 10   x   3 
i. e.,  −1 4 2  y  =  20 ⋅
     
 5 2 1   z   −12 
Now let
A = LU
N-131
 2 −3 10   0 0   u11 u12
1 u13 
i. e.,  −1 4 2=  1 0   0 u22
l21 u23 
    
 5 2 1  l31 l32 1   0
 0 u33 
 2 −3 10   u11 u12 u13 
or  −1 4 2  =  l21 u11 l21 u12 + u22 l21 u13 + u23 
   
 5 2 1   l31 u11 l31 u12 + l32 u22 l31 u13 + l32 u23 + u33 
Equating the corresponding elements on both sides,

u11 = 2 ; u12 = − 3 ; u13 = 10 ;
1 5
l21 u11 = − 1 ⇒ l21 = − ; l31 u11 = 5 ⇒ l31 = ;
2 2
l21 u12 + u22 = 4 ⇒  −  × ( − 3) + u22 = 4 ⇒ u22 = ,

1 5
 2 2
l21 u13 + u23 = 2 ⇒  −  × 10 + u23 = 2 ⇒ u23 = 7,

1
 2
5 5 19
l31 u12 + l32 u22 = 2 ⇒ × ( − 3) + l32 × = 2 ⇒ l32 = ,
2 2 5
5 19 −253
l31 u13 + l32 u23 + u33 = 1 ⇒ × 10 + × 7 + u33 = 1 ⇒ u33 = ⋅
2 2 5
 1 0 0  2 −3 10 
∴ 
L = −1/2  
1 0 and U = 0 5 / 2 7 ⋅
   
 5 / 2 19 / 5 1   0 0 − 253 / 5 
Since AX = B and A = LU , therefore LUX = B.

Now let UX = Y so that LY = B
 1 0 0   y1   3 
i. e.,  −1/2 1 0   y2  =  20 
    
 5 / 2 19 / 5 1   y3   −12 
43 1012
giving y1 = 3, y2 = and y3 = − ⋅
2 10
Now UX = Y implies
2 −3 10   x   3 
 0 5 /2 7   y= 43 / 2 
     
 0 0 − 253 / 5   z   − 1012 / 10 
giving 2 x − 3 y + 10 z = 3,
5 43 253 −1012
y + 7z = and − z= ⋅
2 2 5 10
Hence z = 2 , y = 3 and x = − 4.
∴ The solution of the given system of equations is
x = − 4, y = 3 and z = 2 .
N-132
Problem 9: Solve the system of equations

x1 + 2 x 2 + 3 x 3 = 14 ,
2 x1 + 5 x 2 + 2 x 3 = 18,
3 x1 + x 2 + 5 x 3 = 20 ;
by factorization method.
Solution: The given system of equations can be written in the form AX = B i. e.,
1 2 3   x1   14 
2 5 2   x   18  ⋅
   2  =  
 3 1 5   x3   20 
←A→ ← X→ ←B→
Now let A = LU , where L is a unit lower triangular matrix and U is an upper triangular matrix.
Thus we have
1 2 3  1 0 0  u11 u12 u13 
2 5 2  = l 1 0  0 u22 u23 
   21   
 3 1 5   l31 l32 1   0 0 u33 
1 2 3   u11 u12 u13 
or 2 5 2 =l u l21 u12 + u22 l21 u13 + u23 ⋅
   21 11 
 3 1 5   l31 u11 l31 u12 + l32 u22 l31 u13 + l32 u23 + u33 

u11 = 1 ; u12 = 2 ; u13 = 3 ;
l21 u11 = 2 ⇒ l21 = 2 ; l21 u12 + u22 = 5 ⇒ 2 . 2 + u22 = 5 ⇒ u22 = 1 ;
l21 u13 + u23 = 2 ⇒ 2 . 3 + u23 = 2 ⇒ u23 = − 4 ;
l31 u11 = 3 ⇒ l31 . 1 = 3 ⇒ l31 = 3 ;
l31 u12 + l32 u22 = 1 ⇒ 3 . 2 + l32 . 1 = 1 ⇒ l32 = − 5 ;
and l31 u13 + l32 u23 + u33 = 5 ⇒ 3 . 3 + ( − 5) . ( − 4) + u33 = 5
⇒ u33 = 5 − 9 − 20 = − 24.
Hence the given system AX = B is equivalent to
LUX = B i. e.,
1 0 0  1 2 3   x1   14 
2 1 0  0 1 −4   x2  =  18 ⋅
       
 3 −5 1   0 0 −24   x3   20 
 y1 
Now let UX = Y =  y2  ⇒ LY = B i. e.,
 
 y3 
 1 0 0   y1  14 
2 1 0   y2  = 18  , giving
    
3 −5 1  y3  20 
y1 = 14, 2 y1 + y2 = 18, 3 y1 − 5 y2 − y3 = 20
and y1 = 14, y2 = − 10, y3 = − 72.
N-133
Now UX = Y implies
1 2 3   x1   14 
0   
1 −4  x 2  =  −10  ⋅ , giving
   
 0 0 −24   x 3   −72 

− 24 x 3 = − 72 i.e., x 3 = 3,
x 2 − 4 x 3 = − 10 ⇒ x 2 = 4 × 3 − 10 = 2
and x1 + 2 x 2 + 3 x 3 = 14 ⇒ x1 + 2 ⋅ 2 + 3 ⋅ 3 = 14 ⇒ x 3 = 1.
Hence the solution of the given equations is
x1 = 1, x 2 = 2 and x 3 = 3.
Problem 10: Solve the following system

10 x + 2 y + z = 9
2 x + 20 y − 2 z = − 44
− 2 x + 3 y + 10 z = 22
by using Gauss-Seidel method.
Solution: Dividing the given equations by respective pivot elements,

1
10 x + 2 y + z = 9 ⇒ x= (9 − 2 y − z ) …(1)
10
1
2 x + 20 y − 2 z = − 44 ⇒ y= ( − 44 − 2 x + 2 z ) …(2)
20
1
− 2 x + 3 y + 10 z = 22 ⇒ z= (22 + 2 x − 3 y ) …(3)
10
To start with let y = 0, z = 0.
9
Then from (1), we get x = x (1) = = ⋅ 9 (first approximation for x).
10
Now put x = ⋅ 9 and z = 0 in (2), so that
1
y = y (1) = ( − 44 − 2 × ⋅ 9 + 0 ) = − 2 ⋅ 2 − 0 ⋅ 09 = − 2 ⋅ 29
20
(first approximation for y).
Again putting x = ⋅ 9 and y = − 2 ⋅ 29 in (3), we get
1 1
z = z (1) = [22 + 2 × ⋅ 9 − 3 × ( − 2 ⋅ 29)] = [22 + 1⋅ 8 + 6 ⋅ 87]
10 10
1
= [30 ⋅ 67] = 3 ⋅ 07 (first approximation for z).
10
With these values of x (1), y (1) and z (1) we proceed to obtain second approximations as
1 1
x (2) = [9 − 2 y (1) − z (1) ] = [9 − 2 × ( − 2 ⋅ 29) − 3 ⋅ 07] = 1⋅ 05,
10 10
1 1
y (2) = [ − 44 − 2 x (2) + 2 z (1) ] = ( − 44 − 2 × 1⋅ 05 + 2 × 3 ⋅ 07) = − 1⋅ 998,
20 20
1 1
z (2) = [22 + 2 x (2) − 3 y (2) ] = [22 + 2 × 1⋅ 05 − 3 × ( − 1⋅ 998)] = 3 ⋅ 009.
10 10
N-134
Repeating the process with x (2), y (2) and z (2), we get

1 1
x (3) = [9 − 2 y (2) − z (2) ] = [9 − 2 × ( − 1⋅ 998) − 3 ⋅ 009] = 0 ⋅ 9987,
10 10
1
y (3) = [ − 44 − 2 x (3) + 2 z (2) ]
20
1
= [ − 44 − 2 × 0 ⋅ 9987 + 2 × 3 ⋅ 009] = − 1⋅ 999,
20
1
z (3) = [22 + 2 x (3) − 3 y (3) ]
10
1
= [22 + 2 × 0 ⋅ 9987 − 3 × ( − 1⋅ 999)] = 2 ⋅ 999.
10
Problem 11: Find the solution of the system

83 x + 11 y − 4 z = 95, 7 x + 52 y + 13 z = 104, 3 x + 8 y + 29z = 71
using Jacobi and Gauss-Seidel iteration methods.
Solution: Since the requirement for iteration is satisfied by these equations, we solve each
equation for the unknown having the largest coefficient and the new equations are
1
x= (95 − 11 y + 4 z ), …(1)
83
1
y= (104 − 7 x − 13z ), …(2)
52
1
and z= (71 − 3 x − 8 y ). …(3)
29
Jacobi Iterative method: Starting with x = 0, y = 0, z = 0, we get
95 104 71
x (1) = = 1⋅14 ; y (1) = = 2 and z (1) = = 2 ⋅ 45.
83 52 29
These are first approximations. We proceed to obtain the second approximations as follows :
1 1
x (2) = [95 − 11 y (1) + 4 z (1) ] = (95 − 11 × 2 + 4 × 2 ⋅ 45) = ⋅ 998,
83 83
1 1
y (2) = [104 − 7 x (1) − 13z (1) ] = (104 − 7 × 1⋅14 − 13 × 2 ⋅ 45) = 1⋅ 234,
52 52
1 1
and z (2) = [71 − 3 x (1) − 8 y (1) ] = (71 − 3 × 1⋅14 − 8 × 2) = 1⋅ 779.
29 29
Proceeding in a similar way third approximations are given by
1 1
x (3) = [95 − 11 y (2) + 4 z (2) ] = (95 − 11 × 1⋅ 234 + 4 × 1⋅ 779) = 1⋅ 07,
83 83
1 1
y (3) = [104 − 7 x (2) − 13 z (2) ] = (104 − 7 × ⋅ 998 − 13 × 1⋅ 779) = 1⋅ 421,
52 52
1 1
and z (3) = [71 − 3 x (2) − 8 y (2) ] = (71 − 3 × ⋅ 998 − 8 × 1⋅ 234) = 2 ⋅ 005.
29 29
Continuing in this way further iterations can be obtained.
N-135
Gauss-Seidel iteration method:

Starting with y = 0, z = 0, we get from (1)
1 95
x (1) = (95 − 11 × 0 + 4 × 0 ) = = 1⋅14.
83 83
Putting x = x (1) = 1⋅14, z = 0 in (2), we get
1 104 7 × 1⋅14
y = y (1) = [104 − 7 x (1) − 13 z ] = − = 1⋅ 85.
52 52 52
Now putting x = x (1) = 1⋅14 and y = y (1) = 1⋅ 85 in (3), we get
1 1
z = z (1) = [71 − 3 x (1) − 8 y (1) ] = (71 − 3 × 1⋅14 − 8 × 1⋅ 85) = 1⋅ 82 .
29 29
Now we obtain the second approximations :
1 1
x (2) = [95 − 11 y (1) + 4 z (1) ] = (95 − 11 × 1⋅ 85 + 4 × 1⋅ 82) = ⋅ 99,
83 83
1 1
y (2) = [104 − 7 x (2) − 13 z (1) ] = (104 − 7 × ⋅ 99 − 13 × 1⋅ 82) = 1⋅ 41,
52 52
1 1
z (2) = [71 − 3 x (2) − 8 y (2) ] = (71 − 3 × ⋅ 99 − 8 × 1⋅ 41) = 1⋅ 96.
29 29
Similarly the third approximations are
1 1
x (3) = [95 − 11 y (2) + 4 z (2) ] = (95 − 11 × 1⋅ 41 + 4 × 1⋅ 96) = 1⋅ 05,
83 83
1 1
y (3) = [104 − 7 x (3) − 13z (2) ] = (104 − 7 × 1⋅ 05 − 13 × 1⋅ 96) = 1⋅ 37
52 52
1 1
and z (3) = [71 − 3 x (3) − 8 y (3) ] = (71 − 3 × 1⋅ 05 − 8 × 1⋅ 37) = 1⋅ 96.
29 29
Again the fourth approximations are
1
x(4) = [95 − 11 y (3) + 4 z (3) ]
83
1
= [95 − 11 × 1⋅ 37 + 4 × 1⋅ 96] = 1⋅ 06,
83
1
y(4) = [104 − 7 x (4) − 13 z (3) ]
52
1
= [104 − 7 × 1⋅ 06 − 13 × 1⋅ 96] = 1⋅ 37,
52
1 1
and z (4) = [71 − 3 x (4) − 8 y (4) ] = [71 − 3 × 1⋅ 06 − 8 × 1⋅ 37] = 1⋅ 96.
29 29
Since x (4), y (4), z (4) are sufficiently close to x (3), y (3), z (3), so the values x = 1⋅ 06, y = 1⋅ 37,
z = 1⋅ 96 can be taken as the solution of the given system.
Problem 12: Apply Jacobi iteration method to solve

10 x + y + z = 12 ,
2 x + 10 y + z = 13,
2 x + 2 y + 10 z = 14.
N-136
Solution: To apply Jacobi iteration method we start with x = 0, y = 0, z = 0 as the initial values
of x, y, z .
The given system of equations can be written as
12 y z 12
x= − − ⇒ x (1) = (first approximation for x)
10 10 10 10
13 x z 13
y= − − ⇒ y (1) = (first approximation for y)
10 5 10 10
14 x y 14
and z= − − ⇒ z (1) = (first approximation for z).
10 5 5 10
12 y (1) z (1) 12 13 14 93
Again x (2) = − − = − − =
10 10 10 10 10 × 10 10 × 10 100
13 x (1) z (1) 13 12 14 92
y (2) = − − = − − =
10 5 10 10 10 × 5 10 × 10 100
14 x (1) y (1) 14 12 13 90
z (2) = − − = − − = ⋅
10 5 5 10 10 × 5 10 × 5 100
Repeating the process with x (2), y (2), z (2), we get
12 y (2) z (2) 12 92 90
x (3) = − − = − − = 1⋅ 018
10 10 10 10 100 × 10 100 × 10
13 x (2) z (2) 13 93 90
y (3) = − − = − − = 1⋅ 024
10 5 10 10 100 × 5 100 × 10
14 x (2) y (2) 14 93 92
z (3) = − − = − − = 1⋅122 .
10 5 5 10 100 × 5 100 × 10
Similarly x (4) = 1⋅ 2 − ⋅1024 − ⋅1122 = 0 ⋅ 9854
y (4) = 1⋅ 3 − ⋅ 2036 − ⋅1122 = ⋅ 9842
z (4) = 1⋅ 4 − ⋅ 2036 − ⋅ 2048 = ⋅ 9916.
Again x (5) = 1⋅ 2 − 0 ⋅ 09842 − 0 ⋅ 09916 = 1⋅ 00242
y (5) = 1⋅ 3 − 0 ⋅19708 − 0 ⋅ 09916 = 1⋅ 00376
z (5) = 1⋅ 4 − 0 ⋅19708 − 0 ⋅19884 = 1⋅ 00408.
Proceeding in this way we obtain x = 1, y = 1 and z = 1 as the most appropriate values.

1. We have A = LU
5 −2 1  1 0 0   u11 u12 u13 

⇒ 7 1 −5 = l 1 0  0 u22 u23 
   21  
3 7 4  l31 l32 1  0 0 u33 
⇒ u11 = 5, u12 = − 2 , u13 = 1.

Thus the elements of the first row of U are 5, − 2 , 1.
2. See article 2(I), part (b).
N-137

2. See article 2(I), part (c). In the method of factorization for solving the linear equations
written in the matrix form as AX = B, we write the coefficient matrix A as A = LU, where L
is a lower triangular matrix with diagonal elements as unity and U is an upper triangular
matrix.
3. See article 2(II).
4. See article 2(II), part (a) and part (b).
True or False
1. See article 2(II), part (c).
3. We know that the linear equations expressed in the matrix form AX = B are consistent if
and only if rank A = rank [ A B].
❍❍❍
N-138
Chapter-10
Curve Fitting
Problem 1: Find the line of fit to the following data :
x : 0 5 10 15 20 25
y : 12 15 17 22 24 30
Solution: Here the number of values of x is 6 i. e. even and the values of x are equidistant at an
interval of 5 i. e., h = 5, so we take
{x − (10 + 15) / 2} x − 12 ⋅ 5
u= =
2⋅5 2⋅5
v = y − 20
Let the least square line of v on u be v = a + bu. Then the normal equations are
Σv = ma + bΣu
and Σvu = aΣu + bΣu2
x y u v uv u2
0 12 −5 −8 40 25
5 15 −3 −5 15 9
10 17 −1 −3 3 1
15 22 1 2 2 1
20 24 3 4 12 9
25 30 5 10 50 25
0 0 122 70
Substituting the values in the normal equations, we get

122 = 0 ⋅ a + 70 b
0 = 6a + 0 ⋅ b
These give a = 0 and b = 1⋅ 743.
Hence the least square line of v on u is
v = 0 + 1⋅ 743 u
∴ The line of best fit of y on x is
x − 12 ⋅ 5 
y − 20 = 1⋅ 743  
 2⋅5 
or y = 11⋅ 285 + 0 ⋅ 7 x.
N-139
Problem 2: Fit a second degree parabola to the following :
x : 0 1 2 3 4
y : 1 1⋅8 1⋅3 2⋅5 6⋅3
Solution: Here the number of values of x is odd and the values of x are equidistant. So we take the
origin at the middle value 2 for the x series.
Let us take X = x − 2 and Y = y. Let the curve of best fit be Y = a + bX + cX2 . Then the normal
equations are
ΣY = na + b Σ X + c Σ X2 

Σ XY = a Σ X + b Σ X2 + c Σ X3  .
2 2 3 4
Σ X Y = aΣ X + bΣ X + c Σ X

x y X Y XY X2 X2Y X3 X4
0 1 –2 1 –2 4 4 –8 16
1 1⋅8 –1 1⋅8 –1⋅8 1 1⋅8 –1 1
2 1⋅3 0 1⋅3 0 0 0 0 0
3 2⋅5 1 2⋅5 2⋅5 1 2⋅5 1 1
4 6⋅3 2 6⋅3 12⋅6 4 25⋅2 8 16
Total 0 12⋅9 11⋅3 10 33⋅5 0 34
Substituting the values from the table in the normal equations, we get
12 ⋅ 9 = 5a + 10 c,11⋅ 3 = 10 b and 33 ⋅ 5 = 10 a + 34c.
Solving these equations simultaneously, we get
a = 1⋅ 48, b = 1⋅13, c = 0 ⋅ 55.
Hence the curve of best fit is
Y = 1⋅ 48 + 1⋅13 X + 0 ⋅ 55 X2 .
Changing the origin i. e., putting X = x − 2, Y = y, we have

y = 1⋅ 48 + 1⋅13 ( x − 2) + 0 ⋅ 55 ( x − 2)2
or y = 1⋅ 42 − 1⋅ 07 x + 0 ⋅ 55 x2 .
Problem 3: Fit a second degree parabola to the following data taking x as the independent variables :
x : 1 2 3 4 5 6 7 8 9
y : 2 6 7 8 10 11 11 10 9
Solution: Changing the origin by X = x − 5; Y = y − 7, let the curve of best fit be

Y = a + bX + cX2 .
N-140
Then the normal equations are

ΣY = na + b Σ X + c Σ X2
Σ XY = a Σ X + b Σ X2 + c Σ X3
Σ X2Y = a Σ X2 + b Σ X3 + c Σ X4
x y X Y XY X2 X2Y X3 X4
1 2 −4 −5 20 16 −80 −64 256
2 6 −3 −1 3 9 −9 −27 81
3 7 −2 0 0 4 0 −8 16
4 8 −1 1 −1 1 1 −1 1
5 10 0 3 0 0 0 0 0
6 11 1 4 4 1 4 1 1
7 11 2 4 8 4 16 8 16
8 10 3 3 9 9 27 27 81
9 9 4 2 8 16 32 64 256
0 11 51 60 −9 0 708
11 = 9a + 0 + 60 c,
51 = 0 + 60 b, − 9 = 60 a + 0 + 708c
a = 3, b = 0 ⋅ 85, c = − 0 ⋅ 27
Y = 3 + 0 ⋅ 85 X − 0 ⋅ 27 X2
Changing the origin i. e., putting X = x − 5, Y = y − 7, we have

y − 7 = 3 + 0 ⋅ 85 ( x − 5) − 0 ⋅ 27 ( x − 5)2
= 3 + 0 ⋅ 85 x − 4 ⋅ 25 − 0 ⋅ 27 x2 + 2 ⋅ 7 x − 6 ⋅ 75
or y = − 1 + 3 ⋅ 55 x − 0 ⋅ 27 x2
Problem 4: Fit a parabolic curve to the following data :
x : 1 2 3 4 5
y : 1090 1220 1390 1625 1915

N-141
Solution: Changing the origin by u = x − 3, v = ( y − 1450 ) / 5, let the curve of best fit
be v = a + bu + cu2 .
Then the normal equations are
Σ v = na + b Σ u + c Σ u2 

Σ uv = a Σu + b Σ u2 + c Σ u3 
Σ u2 v = a Σu2 + b Σ u3 + c Σ u4 

x y u v u2 u4 uv u2 v
1 1090 –2 –72 4 16 144 –288
2 1220 –1 –46 1 1 46 –46
3 1390 0 –12 0 0 0 0
4 1625 1 35 1 1 35 35
5 1915 2 93 4 16 186 372
Total 0 –2 10 34 411 73
Putting the values in the normal equations, we get

− 2 = 5a + 10 c, 411 = 10 b, 73 = 10 a + 34c.
Solving these, we get a = − 11⋅ 4, b = 41⋅1, c = 5 ⋅ 5.
v = − 11⋅ 4 + 41⋅1 u + 5 ⋅ 5 u2 .
Changing the origin back i. e, putting u = x − 3 and v = ( y − 1450 ) / 5, the required curve of best fit
is
1
( y − 1450 ) = − 11⋅ 4 + 41⋅1 ( x − 3) + 5 ⋅ 5 ( x − 3)2
5
or y = 1024 + 40 ⋅ 5 x + 27 ⋅ 5 x2 .
Problem 5: If P is a pull required to lift a load W by means of a pulley block, find a linear law of the form
P = mW + c connecting P and W, using the following data :
P : 12 15 21 25
W : 50 70 100 120
Compute P when W = 150 kg.
Solution: We have
P = mW + c.
The normal equations are
Σ P = 4c + m Σ W
ΣWP = c ΣW + m ΣW 2
N-142
W P W2 WP
50 12 2500 600
70 15 4900 1050
100 21 10000 2100
120 25 14400 3000
340 73 31800 6750
73 = 4c + 340 m; 6750 = 340 c + 31800 m
m = 0 ⋅1879, c = 2 ⋅ 2785
P = 2ɺ ⋅ 2759 + 0 ⋅1879W
Now, when W = 150 kg, then P = 30 ⋅ 4635 kg.
Problem 6: Fit an exponential curve of the form y = ab x to the following data :
x : 1 2 3 4 5 6 7 8
y : 1⋅0 1⋅2 1⋅8 2⋅5 3⋅6 4⋅7 6⋅6 9⋅1
Solution: y = ab x takes the form Y = A + Bx,

where Y = log y, A = log a and B = log b.
The normal equations are
ΣY = nA + B Σ x and Σ xY = A Σ x + B Σ x2 .
x y Y = log y xY x2
1 1⋅0 0⋅0000 0⋅0000 1
2 1⋅2 0⋅0792 0⋅1584 4

3 1⋅8 0⋅2553 0⋅7659 9
4 2⋅5 0⋅3979 1⋅5916 16
5 3⋅6 0⋅5563 2⋅7815 25
6 4⋅7 0⋅6721 4⋅0326 36
7 6⋅6 0⋅8195 5⋅7365 49
8 9⋅1 0⋅9590 7⋅6720 64
Total 36 30⋅5 3⋅7393 22⋅7385 204
Putting the values in the normal equations, we get

3 ⋅ 7393 = 8 A + 36 B and 22 ⋅ 7385 = 36 A + 204B.
N-143
Solving these, we get B = 0 ⋅1406, A = 1 ⋅ 8336.

∴ b = antilog B = 1⋅ 38 and a = antilog A = 0 ⋅ 68.
Hence the required curve of best fit is
y = (0 ⋅ 68) (1⋅ 38) x .

1. See article 3.
2. See article 3.
3. See Problem 6.
4. See article 6.
❍❍❍

Numerical Methods

Uploaded by

Copyright:

Available Formats

Numerical Methods

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 Methods

Uploaded by

Copyright:

Available Formats

h na 's

K ris Text Book Solutions

A. R. Vasishtha Hemlata Vasishtha

KRISHNA Prakashan Media (P) Ltd.

Chapter 1: The Calculus of Finite Differences..........................N-01—N-28

Chapter 2: Interpolation with Equal Intervals.........................N-29—N-43

Chapter 3: Interpolation with Unequal Intervals of the

Chapter 4: Central Difference Interpolation Formulae............N-59—N-72

Chapter 5: Solution of Algebraic and Transcendental Equations

Chapter 6: Numerical Differentiation....................................N-87—N-97

Chapter 7: Numerical Integration.........................................N-98—N-119

Chapter 8: Errors in Numerical Calculations.......................N-120—N-123

Chapter 9: Simultaneous Linear Algebraic Equations.........N-124—N-137

Chapter 10: Curve Fitting...................................................N-138—N-144

(ii) ∆ cot 2 x = cot 2 x +1 − cot 2 x , taking the interval of differencing as 1

cos 2 x +1 cos 2 x cos 2 x +1 sin 2 x − cos 2 x . sin 2 x +1

sin (2 x − 2 x + 1 ) sin { 2 x (1 − 2)} − sin 2 x

(iii) ∆ sinh ( a + bx ) = sinh [ a + ( x + h) b ] − sinh [ a + bx ]

= 2 cosh  a + + bx  sinh , for h = 1.

= 2 sinh  a + + bx  sinh , for h = 1.

Proceeding like this, we get

(viii) ∆ tan−1 ax = tan−1 a ( x + 1) − tan−1 ax

Problem 2: Show that

Solution: (i) We have

Proceeding similarly, we have

(ii) We have ∆ cos ( a + bx ) = cos { a + b ( x + 1)} − cos ( a + bx )

Proceeding similarly, we have

∆ cot ( a + bx ) = cot [ a + b ( x + 1)] − cot ( a + bx )

cos ( a + bx + b ) sin ( a + bx ) − cos ( a + bx ) sin ( a + bx + b )

Problem 3: Prove that, if f ( x ) and g( x ) are any functions of x,

Solution: See article 9.

Problem 4: Evaluate the following :

Solution: (i) By definition ∆ f ( x ) = f ( x + h) − f ( x ).

and ∆ (2) = 2 ∆ (1) = 2 . (1 − 1) = 2 . 0 = 0.

= 2hx + h2 + e x ( e h − 1) + 0 = 2hx + h2 + e x ( e h − 1).

(ii) ∆3 [ ax3 + bx2 + c x + d ] = ∆3 ax3 , [ ∵ By fundamental theorem of difference calculus,

(iii) Proceed as in Ex. 2 (ii). Ans. e x ( e h −1)2 .

Prove that ∆ log x = log 1 +  ⋅

Problem 6: Prove that

∆ cos ( ax + b ) = − 2 sin   sin  a  x +  + b 

Solution: (i) We have

∆ cos ( ax + b ) = − 2 sin   sin  a  x +  + b 

and ∆ tan ( ax + b ) = sin ah sec ( ax + b ) sec [ a ( x + h) + b ].

Problem 7: Prove that ∇ = ∆ E −1 = E −1 ∆ = I − E −1.

From above we observe,

Problem 8: Prove that ∆3 = E3 − 3E 2 + 3E − I.

Hence, by the definition of equality of two operators, we have

Aliter: We have ∆≡E−I [∵ E ≡ I + ∆ ]

⇒ ∆3 ≡ E3 − 3E 2 I + 3E I 2 − I 3 [By binomial theorem]

Hence the required sum

Problem 10: Find the function whose first difference is

∴ ∆ f ( x ) = [ A ( x + 1)2 + B ( x + 1) + C ] − [ Ax2 + Bx + C ], by def. for h = 1

Hence the required function is

Let us consider f ( x ) = k . e a + bx and h = 1

(v) Here ∆ f ( x) = 5 x . …(1)

Problem 11: Find the value of

(ii) ∆ [2 x(−2) − 3 x(3) − 4 x (−1) ]

(ii) ∆ [2 x(−2) − 3 x (3) − 4 x(−1) ] = [ − 4 x (−3) − 9 x(2) + 4 x(−2) ] h

(iv) Let f ( x ) = α + βx + γ x2 = ( a + bx )(2),

where a and b are some unknown constants.