( B(6'1t s..;.M.a.elUlic.

s-ll/CC-U
/CBCS]

2024
MATHEMATICS - HONOURS
Paper : CC-14
(Numerical Methods)
Full Marks : 50
The fig ure s in the margin indica
re ful l marks.
Ca ndi dat es are required to give
the ir answers in the ir own 11t•ords
as far as practicable~
Sym bol s have their usu al meani11
g.
Un it - l
1. Each of the following questio
ns has four possible ansv.-ers of wh
correct alternative : ich exactly ooe is correct Choose
the
(a) Ho w many significant digits l x 10
does the floating point number 0.0
4320x t 0 3 have?
(l) 6
(ii) 4
(iii) 5
(iv) 2.
(b) Th e produc t µb is given by
(i) µo = E - 1 - E (ti) ~lO= £ + E - 1
fui) µo = £ - £-1 (iv) µ5 = Y2(E - £- 1) .
(c) If /(.x )=x 2, then the second
order divided difference fo.r the pot
llts x0 , x 1, x2 will be

(i) - l -t
(ti)
'"I - .r.o

(iii) l (iv}
(d) Th e err or tcn n committed in
computing derivatives is
, , !"~..,(~) , J"+2 \n)
(i) Rn+l(x) = ro (x) ( '-.,
n+ l)'
. (ii) Rn+rlx) = ro(x) , ")'
\n+ - -
I , /~+)(~) . /1Jt 2(~ )
(iii) Rn+l(.x) = ffi (x) (11+1)! +ffi
{x) (n+ 2)!

I //'1+1(~) . 2
ffJ+ (c,.J))
(iv) Rn+l ( x) -- ro'(x). ::- -~ + oo(x)
(n + l)! (n + 2)!
• \ <i:: i:: < max {x, x-0, x~•··, x,.) and
where mm '-x,1
x,.,, "'
x0, x 1,.. , ~~ ~(x) = (x- .x-o){x- x,1)...(x- x,,).

Please Turn Over

( 8(61/1 s-J, .Atil l•~-JI/CC-N ICBcs J (2)
(~) The error committed in nume rical integration of the functi
on /(x) == 4.-c3 + x 2 + 2x - l by Simp son ·s
1
~ rd rule betwe en - l and 2 with two sub.fo terval
.) s is

(i) 0 (i,) 1,451 x 1o-3

(lii) - 2,054 x lQ-2 (iv) - 3.289 x l o-3 .
(t) Cons ider the equat ion (x + l )(x - l)(x - 3) =O. If the
bisection algor ithm is appl ied with imtial
interv al [- 4, 4], bow many root(s) of the gjven equat
ion will you be able to comp ute?
(i) 3 (ii)
(iii ) 2 (iv) 0.

g(x) = ~ + ~ . Whic h root-f indin g problem is

{g) Cons ider the fixed point iterati on xk+\ == g(xi) with
3 3x
this equiv alent to?

(i) x2 - 2 = 0 X 4
(ii) - + - = 0
3 3x

Cw1 .!. - __i_2 = o (iv)

l
X - - +-
4
=0.
3 3x 3 3x2
(b) lo the Gaus s elimi nation metho d for solvin
g a system of linear algeb raic equat ions, the
row-r educt ion phase of the coeffi cient matri x leads
to a
(i) diagonal matrix (ii) lower triang ular matri x
(ui) upper triang ular matri x (iv) identity matnx .

(i) Cons ider A=(: ~3 ~

I and i 0
> = [I O 1)7 . The first appro ximat ion to the eigen vecto r
5 0 1)
corre sponding to the domin ant eigen value using Po\-.e
r metho d is
 (3) ( n(6th Sm,)-Matht'matlcs-lf/CC.U ICBCS)

(i) Using Euler's method to opproximute nn ODE, you obtnin the difference cquntion YJ~t = y + J,ayf'
where a is n constnnt nml I, is the lime step. Whnl is the ODE? 1

(I') ..,.:::..""
c/1•
'0' (ii) dy
dt dt

(iii) dy = hai• (iv) dy =ha .

dt · dt

Unit - 2
Answer"">' m,c question.

I
2. Deline operators fl nnd 6. llcnce prove that p[/(x).g(x)] = µ/( x).µg(x) + 46g(x).6/(x) . 2+3

J . Derive Newton's forward interpolation fonnuln for (n + I) equispaced lobular data. Slate how it differs
from the derivation of Hennite interpolation fom1ula. 4+ I

Unlt-3
Answer any two questions.

4. Explain the principle of numerical diffcrcnlintion. Deduce Lagrange's numerical differentiation fonnula
and deduce it in particular for computing the derivative at an interpolating point. 1+2+2
b
5. Deduce Newton-Cotes fonnula lo approximate J /(x)dr by selecting the nodes x0, x 1, x such that
2
a
a = x 0• x 1 = x 0 + h, .,·2 = x 0 + 2h = b. State, wi th justification , whether it is a closed type fonnula.
4+1

I
6. (a) The quadrature formula f/(x)d-c = ¼/ (.ro) + c,f(x1) has the highest degree of precision.
0 -

I
(b) Give geometrical interpretation of Si mpson's
3 Ruic. 3+2

b
7. Establish the Boole's rul e for computing the value of the integral f /(x)dx by tnl<lng four equal
a
sub-intervals of [a, b]. State the degree of precision of the method by writing U1e error tcnn. 4+1

Please Turn Over

 9

( B((itlt SM.)•Mtllltt>JlfatJt:,s.JIJCC. J4/CBCS)

(4)

Unll - 4
Answ er uny two qucs1i om1.

8. Deriv e lhc Rcgul a-Fals i metho d for findin g a real simple

root of a given equation /(:c) = 0 and give
its geometrical interp retatio n.
3 12
9. Show that, in order to !ind a doubl e root a, of the equal ion
f (x ) = 0, if we apJJIY the NcWkm-Ra r,hson

method , x n+ l = x II _ f(xn)
f'(xn) , 11 -- O, I , 2, 3, •• • t I1cn 1t
' converges I'inearI b t 'f'
y u 1 we app J,v
✓ modi fied

Newt on-Rn phson mcth.od , x,,+ 1 =:x,, - 2 f(x )

, " , = O, I, 2, 3,... lhen
11 1L con v<.:rgcs quadrauc.ally, 3+2
.f(x,, )

10. Using
• Newto n-Rap hson metho d, establish the tlcrati ve formu
l a, .x,,i 1 = I [ 2;cn
3 1- N] to calculate the
2
x,,
cube root of N.
5
l l. Discuss B isectio n metho d lO find a simple root of an equatfo n f(x)
bisect ion metho d is alway s conve rgent" .
=0. l Jenee prove o r di sprov e ·'Toe
3+2
U nit - 5
Answ er any two questi ons.

12. Descr ibe Gauss Elimin at ion metho d for numer ical solut
ion of a system of linear equati ons, AX = B,
where A = (alj)n,..n• B = (bp b , .••, bn)T and X = (x , x
2 1 2, .. . , x n) r . Why is pivoLi ng process neces sary?
4+1
13 . Descr ibe the Gauss -Jacob i iterati on metho d fo r so lving
a system of lmear cquatr nns and express the
i,ter.1tive schem e in the form ;c(k+ l) = J!xf.k) + C (k = 0,
I. 2,...) dcscn bms the iterati on matnx JI a nd the
colum n vecw r C
3+2
14. Write down the Crout ' s algori thm for LU decom pos itio
n of a square matrLx Facto rize the follow ing
I I
matrix by thjs algori thm : 3 I
(
I -2 -5
-~)- 2+3

15. PrO'Ve that a sufficienit condition of convergence for the

Gauss-Seide l metho d is that the coefficient
matrix of the sys lem is 'strict ly dia gonall y dom~nant ' .
5
 (S)
u.,, . . ,
~-u-,.,, ~ .
I~ ~ e 1k ,..iuc ofJ(l.l) fbt the following IVP ~ ~ =r-:/ ·,y(l} =l by Euler's method. correct
tb:
to tbm- ~ifitanl t1prcs ~ $tCp M~ It = (l,1 . '.5

l 7.. Soh~ 11'Umttical~ the tollowil\ll ditlcrential cqualion ?

·~
~ , : : l~
.......

bx = Ul(tl..'i)l~ St'-~ lhal J(I.O) = 1.() using fourth order Runge.Kuna method, correct up to 4D.
5