Вы находитесь на странице: 1из 14

MCSE-004

Dec 2013 Solutions


Manual

IGNOUUSER
1
ignouuser@gmail.com

1. (a) Verify the distributive property of floating point numbers i.e. prove :
a(b-c) ab – ac
a=.5555E1, b=.4545E1, c=.4535E1
Define : Truncation error, Absolute Error and Relative Error.

{
Addition/Subtraction of floating point numbers
To add/subtract 2 floating point numbers, the exponents of the two numbers must be
equal. If not the mantissa is shifted appropriately . Then the Mantissas are added
or subtracted. After that the result is normalized and the exponent is appropriately
adjusted.

Multiplication of floating point numbers


To multiply 2 floating point numbers, the mantissas are multiplied and exponents
are added. After that the resulting mantissa is normalized and the exponent is
appropriately adjusted.
}

LHS
(b-c) = .0010E1 = .1000E-1
a(b-c) = .5555E1 * .1000E-1 = .05555E0 = .5555E-1

RHS
ab = .5555E1 * .4545E1 = .2524E2
ac = .5555E1 * .4535E1 = .2519E2
ab – ac = .0005E0 = .5000E-3

Clearly LHS RHS

For definitions, please refer block.

1. (b)Find the real root of the equation x = e-x using Newton-Raphson Method. List the cases
where Newton’s Method fail.

Given x=e-x
multiplying both sides by ex and arranging, we get

xex – 1 = 0

f(x) = xex – 1
f’(x) = (1 + x)ex

Numerical and Statistical Computing Solutions Manual – Dec 2013


2
ignouuser@gmail.com

Let x0 = 1

By Newton Raphson method we have,

xn+1 = xn -

x1 = 1 - = 0.6839397

f(x1) = 0.3553424, f’(x1) = 3.337012

x2 = 0.6839397 - = 0.5774545

f(x2) = 0.028733951, f’(x2) = 2.810232

x3 = 0.5774545 - = 0.5672297

f(x3) = 0.000238786, f’(x3) = 2.763614

x4 = 0.5672297 - = 0.567143297

Since x3 and x4 are same upto 3 decimal places, 0.567 is the required root.

Method fails in the cases listed below –

i) Division by zero occurs in some cases.

let f(x) = x3 – 2x2 + 3 = 0


f’(x) = 3x2 – 4x

By Newton Raphson method we have


xn+1 = xn -

If we choose initial value x0 = 0 then we get division by zero error.

ii) Infinite loop case

In some cases, say x0 is the initial value and 1st approximation gives x1 which inturn gives
2nd approximation as x0 again and the loop continues.

Numerical and Statistical Computing Solutions Manual – Dec 2013


3
ignouuser@gmail.com

1. (c) Solve by Gauss-Seidel method.


2x1 – x2 + x3 = -1
x1 + 2x2 – x3 = 6
x1 – x2 + 2x3 = -3
correct to 3 decimal places

The given system of equations can we written as


x1 = (1/2)(-1 + x2 - x3)
x2 = (1/2)(6 - x1 + x3)
x3 = (1/2)(-3 - x1 + x2)

Let initial approximation x0, y0, z0 is 0,0,0

First approximation
x1(1) = (1/2)(-1 + 0 + 0) = - 0.5
x2(1) = (1/2)(6 + 0.5 + 0) = 3.25
x3(1) = (1/2)(-3 + 0.5 + 3.25) = 0.375

Second approximation
x1(2) = (1/2)(-1 + 3.25 – 0.375) = 0.9375
x2(2) = (1/2)(6 – 0.9375 + 0.375) = 2.71875
x3(2) = (1/2)(-2 - 0.9375 + 2.71875) = -0.60938

Similarly, we have
[x1(3), x2(3), x3(3)] = [1.164063, 2.113281, -1.02539]
[x1(4), x2(4), x3(4)] = [1.069336, 1.952637, -1.05835]
[x1(5), x2(5), x3(5)] = [1.005493, 1.968079, -1.01871]
[x1(6), x2(6), x3(6)] = [0.993393, 2.001721, -0.99972]
[x1(7), x2(7), x3(7)] = [0.996836, 2.001721, -0.99756]
[x1(8), x2(8), x3(8)] = [0.999639, 2.001402, -0.99912]
[x1(9), x2(9), x3(9)] = [1.00026, 2.00031, -0.99997]
[x1(10), x2(10), x3(10)] = [1.000143, 1.999941, -1.0001]
[x1(11), x2(11), x3(11)] = [1.000021, 1.999939, -1.00004]

Since last 2 approximations are same upto 3 decimal places.


Hence the required solution is
x1 = 1.000
x2 = 1.999
x3 = -1.000

Numerical and Statistical Computing Solutions Manual – Dec 2013


4
ignouuser@gmail.com

1. (d)Let f(x) = ln(1+x), x0 = 1 and x1 = 1.1. Use linear interpolation to calculate an approximate
value of f(1.04) and obtain a bound on the truncation error.

We have x0 = 1, x1 = 1.1
f(x0) = ln(1+x0) = 0.693147
f(x1) = ln(1+x1) = 0.741937

Lets compute Lagranges fundamental polynomials,


l0(x) = (x-x1) / (x0 – x1) = (x – 1.1)/(1 – 1.1) = (x – 1.1)/(-0.1)
l1(x) = (x-x0) / (x1 – x0) = (x – 1)/(1.1 – 1) = (x – 1)/( 0.1)

Hence Lagranges Interpolating polynomial is given by


P1(x) = l0(x)* f(x0) + l1(x)* f(x1) =- 10(x – 1.1)* 0.693147 + 10(x – 1)* 0.741937

f(1.04) = P1(1.04) = -10(1.04 – 1.1)* 0.693147 + 10(1.04– 1)* 0.741937


= 0.712663

The truncate error in linear interpolation is given by


E1(f;x) = (1/2)(x-x0)(x-x1)f”( ) where lies between 1 and 1.1

= (1/2)(x – 1)(x – 1.1)(-1/(1+x)2)


The max value of | | is 0.892857
and maximum value of (x-1)(x-1.1) occurs at (1+1.1)/2=1.05

| | | – – | * (0.892857) = 0.0011160

1. (e) Consider initial value problem


= x+y; y(0) = 1

Find y(0.2) using Runge-Kutta Method of fourth order. Also compare it with exact solution
y = -(1+x) + 2ex to find the error.

We have, x0=0, y0=1, h=0.2


Then we get,
k1=hf(x0,y0) = 0.2(0+1) = 0.2
k2 = hf(x0 + h/2,y0 + k1/2) = 0.2(0.1+ 1.1) = 0.24
k3 = hf(x0 + h/2,y0 + k2/2) = 0.2(0.1+1.12) =0.244
k4 = hf(x0 + h,y0 + k3) = 0.2(0.2+1.244) = 0.2888

Therefore
y = y0 + 1/6(k1 +2 k2 + 2 k3 + k4)
= 1 + 0.2428= 1.2428

Numerical and Statistical Computing Solutions Manual – Dec 2013


5
ignouuser@gmail.com

Given, y = -(1+x) + 2ex


Putting x = 0.2, we get y = -1.2 + 2.442806 = 1.242806

Error = 1.242806 – 1.2428 = 0.000006

2. (a) Find the interval in which the smallest positive root of the following equation
lies using Bisection Method
x3 – x – 4 = 0

Given f(x) = x3 – x – 4
f(0) = 0 – 0 – 4 = -4
f(1) = 1 – 1 – 4 = -4
f(2) = 8 – 2 – 4 = 2

Clearly f(1)*f(2)<0, hence root lies between 1 and 2


First approximation of root, x1 = (1+2)/2 = 1.5

f(1.5) = -2.125
Clearly f(1.5)*f(2)<0, hence root lies between 1.5 and 2.

Second approximation of root, x2 = (1.5+2)/2 = 1.75

f(1.75) = -0.39063
Clearly f(1.75)*f(2)<0, hence root lies between 1.75 and 2.

Third approximation of root, x3 = 1.875

f(1.875) = 0.716797

Now, f(1.875)*f(2) > 0,

Hence the interval in which smallest root lie is [1.75,2] and the root is 1.875.

2. (b) Solve the following linear system of equations using Gauss Elimination Method.
x1 + x2 + x3 = 3
4x1 + 3x2 + 4x3 = 8
9x1 + 3x2 + 4x3 = 7

Numerical and Statistical Computing Solutions Manual – Dec 2013


6
ignouuser@gmail.com

Given system of equations


x1 + x2 + x3 = 3
4x1 + 3x2 + 4x3 = 8
9x1 + 3x2 + 4x3 = 7 ..........(1)

Eliminate x1 from second equation of (1), multiply first equation by 4 and subtract it from
second equation
(3 – 4) x2 + (4 – 4) x3 = 8 – 4*3
- x2 = 8 – 12
x2 = 4

Eliminate x1 from third equation of (1), multiply first equation by 9 and subtract it from third
equation.
(3 – 9) x2 + (4 – 9) x3 = 7 – 9*3
-6 x2 + (-5 x3) = 7 - 27
6 x2 + 5 x3 = 20

Hence given system of equations become


x1 + x2 + x3 = 3
x2 = 4
6 x2 + 5 x3 = 20 .............(2)

Substituting value of x2 in third equation we get,


x3 = -4 /5 = -0.8

Substituting value of x2 , x3 in the first equation we get,

x1 = 3 – 4 –(-0.8) = -0.2

Hence the required solution is

x1 = - 0.2
x2 = 4
x3 = -0.8

2. (c) Give properties of polynomial equations.

i) A polynomial equation of degree n has n solutions.


ii) Non-real complex solution with real coefficient, if exist, occur in conjugate pairs.
iii) Descarte’s rule of signs
The number of positive real solutions of the given polynomial equation is equal to the
number of variations in signs of the polynomial.

Numerical and Statistical Computing Solutions Manual – Dec 2013


7
ignouuser@gmail.com

3. (a) The table below gives the values of tanx for 0.10 x 0.30.
x 0.10 0.15 0.20 0.25 0.30
y=tanx 0.1003 0.1511 0.2027 0.2553 0.3093

Find i) tan0.12 ii) tan0.26

First lets construct the difference table

x y=tanx Δ Δ2 Δ3 Δ4
0.1 0.1003
0.0508
0.15 0.1511 0.0008
0.0516 0.0002
0.2 0.2027 0.001 0.0002
0.0526 0.0004
0.25 0.2553 0.0014
0.054
0.3 0.3093

i) tan0.12
here, h=0.05, a=0.10, x=0.12
u=(0.12-0.10)/0.05 = 0.4

By Newton’s Forward Interpolation formula we have


tan(0.12) = f(a) + u Δ f(a) + Δ 2f(a) + Δ 3f(a) + ...
= 0.1003 + 0.4*0.0508 + 0.0008 + 0.0002
+ 0.0002

= 0.1003 + 0.02032 – 0.000096 + 0.0000128 – 0.00000832


= 0.1205

ii) tan0.26
here, h=0.05, a=0.10, x=0.26
u=(0.26-0.10)/0.05 = 3.2

By Newton’s Forward Interpolation formula we have


tan(0.26) = f(a) + u Δ f(a) + Δ 2f(a) + Δ 3f(a) + ...
= 0.1003 + 3.2*0.0508 + 0.0008 + 0.0002
+ 0.0002
= 0.1003 + 0.16256 + 0.002816 + 0.0002816 + 0.00001408
= 0.2659

Numerical and Statistical Computing Solutions Manual – Dec 2013


8
ignouuser@gmail.com

3. (b) Evaluate I = ∫ dx, correct to three decimal places using

i) Trapezoidal, and
ii) Simpson’s rule with h=0.5 and h=0.25

Here, f(x) =

The value of f(x) at x=0,0.25,0.5,0.75,1 are tabulated below.

x 0 0.25 0.50 0.75 1


f(x) 1 0.8 0.67 0.57 0.5
f0 f1 f2 f3 f4

i) Trapezoidal rule
Taking h=0.25, we have
I = (0.25/2)[ f0 + f4 + 2(f1 + f2 + f3 )]
= 0.125*(1.5 + 2*2.04) = 0.6975

ii) Simpsons rule


h = 0.25,
I = (0.25/3)[ f0 + f4 + 4(f1 + f3 ) + 2 f2]
= 0.0833 * (1.5 + 4*1.37 + 2*0.67) = 0.6930

h=0.5
we have to consider only f0, f2, f4
I = (0.5/3)[ f0 + f4 + 4* f2]
= 0.16667*(1.5 + 4*0.67) = 0.6967

3. (c) Determine the value of y when x=0.1 given that y(0) = 1 and y’ = x2 + y.

let h = 0.1/5 = 0.02


x0 = 0, y0 = 1
By Euler’s method we have
y1 = y0 + hf(x0,y0) = 1 + 0.1*(0+1) = 1.1

y2 = y1 + hf(x1,y1) = 1.1 + 0.1*(0.02*0.02 + 1.1) = 1.21004

y3 = y2 + hf(x2,y2) = 1.21004 + 0.1*(0.04*0.04 + 1.21004) = 1.331204

Numerical and Statistical Computing Solutions Manual – Dec 2013


9
ignouuser@gmail.com

y4 = y3 + hf(x3,y3) = 1.331204 + 0.1*(0.06*0.06 + 1.331204) = 1.464684

y5 = y5 + hf(x4,y4) = 1.464684 + 0.1*(0.08*0.08 + 1.464684) = 1.611792

Hence the value of y at 0.1 is 1.611792

4. (a) A problem in statistics is given to the three students A,B and C whose chances of
solving it are 1/2, 3/4 and 1/4 respectively. What is the probability that the problem will be
solved.

Given, probability that A will solve the problem P(A) = 1/2


probability that B will solve the problem P(B) = 3/4
probability that C will solve the problem P(C) = 1/4

Hence, probability that A will not solve the problem P( ̅) = 1 – 1/2 = 1/2
probability that B will not solve the problem P( ̅ ) = 1 – 3/4 = 1/4
probability that C will not solve the problem P( ̅ ) = 1 – 1/4 = 3/4

Since all events are independent, so


P(no one will solve the problem) = P( ̅)* P( ̅ )* P( ̅ ) = (1/2)*(1/4)*(3/4) = (3/32)
P(that prob will be solved) = 1 – (3/32) = 29/32.

4. (b) Calculate the correlation coefficient for the following heights (in inches) of fathers (x) and
their sons (y) :
x: 65 66 67 67 68 69 70
y: 67 68 65 68 72 72 69

The various calculations on the given data are tabulated below.

x y x-x' y-y' (x-x')*(y-y') (x-x')*(x-x') (y-y')*(y-y')


65 67 -2.42857 -1.71429 4.16326531 4225 4489
66 68 -1.42857 -0.71429 1.02040816 4356 4624
67 65 -0.42857 -3.71429 1.59183673 4489 4225
67 68 -0.42857 -0.71429 0.30612245 4489 4624
68 72 0.571429 3.285714 1.87755102 4624 5184
69 72 1.571429 3.285714 5.16326531 4761 5184
70 69 2.571429 0.285714 0.73469388 4900 4761
14.8571429 31844 33091

Numerical and Statistical Computing Solutions Manual – Dec 2013


10
ignouuser@gmail.com

where
x’ = mean value of x = 67.42857
y’ = mean value of y = 68.71429

We have the correlation coefficient



r= = = 0.000457
√∑ ∑ √

4. (c) Three identical bags have the following proportion of balls.


First bag: 2 black 1 white
Second bag: 1 black 2 white
Third bad: 2 black 2 white

One of the bag is selected and one ball is drawn. It turns out to be white. What is the
probability of drawing a white ball again. The first one not been returned.

Let E = event of selection of ith bag(i=1,2,3)


A = event of drawing white ball

Then P(E1) = P(E2) = P(E3) = 1/3


Also, P(A | E1) = C(1,1)/C(3,1) = 1/3
P(A | E2) = C(2,1)/C(3,1) = 2/3
P(A | E3) = C(2,1)/C(4,1) = 1/2

Let C denote the further event of drawing another white ball.


P(C | E1 A) = 0/2 = 0
P(C | E2 A) = 1/2
P(C | E3 A) = 1/3

Hence by Bayes theorem


| | | | | |
P(C|A) = | | |

= = = 1/6

Numerical and Statistical Computing Solutions Manual – Dec 2013


11
ignouuser@gmail.com

5. (a) Evaluate ∫ [ ( √ )] using Simpsons rule with 11 points.

Here, h = (6-1)/11 = 5/11

x f(x)
1 2.0349
1+(5/11) 2.04209
1+(10/11) 2.04821
1+(15/11) 2.05364
1+(20/11) 2.05857
1+(25/11) 2.06311
1+(30/11) 2.06734
1+(35/11) 2.07132
1+(40/11) 2.07509
1+(45/11) 2.07868
1+(50/11) 2.08211
6 2.0854

By Simpsons one-third rule,

∫ [ ( √ )] = (5/33) [(2.0349 + 2.0854) + 4(2.04209 + 2.05364 + 2.06311 + 2.07132 +


2.07868) + 2(2.04821 + 2.05857 + 2.06734 + 2.07509 + 2.08211 )]

= (5/33) [4.1203 + 4(10.30884) + 2(10.33132) ]


= 10.003

Numerical and Statistical Computing Solutions Manual – Dec 2013


12
ignouuser@gmail.com

5. (b) Estimate the sale of a particular quantity for 1966 using the following table.
Year 1931 1941 1951 1961 1971 1981
Sale in 12 15 20 27 39 52
thousands

First lets construct the difference table.

Year Sale ∇ ∇2 ∇3 ∇4 ∇5
1931 12
3
1941 15 2
5 0
1951 20 2 3
7 3 -10
1961 27 5 -7
12 -4
1971 39 1
13
1981 52

here, h=1981-1971=10
a=1981
x=1966

hence, u=(1966 – 1981)/10 = - 1.5

By Newton’s Backward Interpolation formula we have

f(1966) = f(a) + u∇f(a) + ∇2f(a) + ∇3f(a) + ∇4f(a) + ...

= 52 + (-1.5)*13 + 1+ +
(-7) + (-10)

= 52 - 19.5 + 0.375 - 0.25 – 0.164 – 0.117


= 32.344

Numerical and Statistical Computing Solutions Manual – Dec 2013


13
ignouuser@gmail.com

For more solutions keep visiting

www.googledrive.com/host/0B3kSUnlvgaY4bEFVc25oNnRaVU0

or contact me at

ignouuser@gmail.com

Numerical and Statistical Computing Solutions Manual – Dec 2013