Maths Part B PDF

MA8451-Probability and Random Processes Dept.
of ECE 2019-2020
MA8451 - PROBABILITY AND RANDOM PROCESSES

II year ECE- IV Semester
UNIT I - RANDOM VARIABLES
PART – A
1. Show that for any events A and B in S , P  B   P  B A P  A  P  B A  P  A  .
(A/M 2019)
Solution:
P  B A P  A   P  B A  P  A 
P  B  A P  B  A
 P  A  P  A
P  A P  A
 P  B  A  P  B  A   P  B  By Venn Diagram.
2. A Problem in statistics is given to three students. A, B and C whose chances of
1 1 1
solving it are , and respectively. What is the probability that the problem will
2 3 4
be solved?
Solution:
Let A, B, C Denote the events that the problem is solved by the students A, B, C
respectively.
1 1 1
Then P  A   , P  B  , P C  
2 2 4
P  A   1   , P  B   1   , P C   1  
1 1 1 2 1 3
2 2 3 3 4 4
P(all the three students will not solve the problem)  P  A  P  B  P  C   . . 
1 2 3 1
2 3 4 4
P(all the three students will solve the problem)  P  A  B  C 
 1  P  A  P  B  P C   1 
1 3
 .
4 4
3. Prove that the function p(x) is a legitimate probability mass function of a discrete
 2  1 x
 x  0,1, 2, ...
random variable X, where p( x )   3  3 
,
(April 2017)
 0,
 otherwise
Proof:
The probability distribution function p( x) is a probability mass function if
p( x)  0, x and  p(x) 1
 x
21
 p ( x)    
x 0 3  3 
2  1  1   1  
2 3
 1         ...
3  3  3   3  
1
St. Joseph’s College of Engineering
MA8451-Probability and Random Processes Dept. of ECE 2019-2020
2 1

3  1
1  
 3
2 1 1
 p  x   3  2  1 1  r  r 2  ... 
1 r
if r  1
 
3
x2
4. If f  x   , 1  x  2 is the pdf of the random variable X , then find P  0  x  1 .
3
(April/May 2018)
Solution:
1
P  0  x  1   f  x  dx
0
1
1
x 2
 x3  1
  dx     .
0
3  9 0 9
5. Find the expected value of the discrete random variable X with the probability mass
1
 3 ;x  0
function P ( X  x )   (Nov/Dec 2016)
2 ;x  2
 3
Solution:
Expected value = E  X    x P  x  = 0.  2.  .
1 2 4
3 3 3
1 1 1
6. If a random variable X assumes three values 1, 0,1 with probabilities , ,
3 2 6
respectively, then find the probability distribution of Y  3 X  1.
Solution:
We have Y  3 X 1
when x  1, Y  3 1 1 2
x  0, Y  3 0 11
x  1, Y  31 1 4
The probability distribution of Y  3 X 1 is
y -2 1 4
1 1 1
P(y)
3 2 6
7.  x , 1  x  1
Test whether f ( x)   can be the probability density function of
0, otherwise
a continuous random variable? (April / May 2015)
2
Solution:
The given function f ( x) is a probability density function, if f ( x)  0, x

and  f  x  dx  1

 1
 f  x  dx   x dx
 1
1
 2 x dx
0
1
 x2 
2 
 2 0
 1 0  1
 Yes, f ( x) can be the probability density function of a continuous random variable.
8. A random variable X has the probability density function f  x  given by

a 1 x ,
f ( x)  
2
 2 x5 
Find ‘a’ and P(X < 4). (Nov/Dec 2015)

 0, otherwise
Solution:

Since X is a continuous random variable f ( x)  0, x and  f  x  dx  1

5
5
 x3 
2      1
2
a (1 x ) dx 1 a  x
 3 2
 53   23  
 a  5     2     1
 3  3 
1
 42a  1  a 
42
1  23  
4
(1  x 2 ) 1  x3  43  
4
31
P  X  4   dx   x      4  
  2    .
2
42 42  3 2 42  3  3  63
9. A Continuous random variable X has a probability density function
f  x   3x2 ; 0  x  1. Find ‘a’ such that P  X  a   P  X  a 
Solution:
We know that the total probability =1
Given P  X  a   P  X  a   K (say)
We know that P  X  a   1  P  X  a 
P  X  a  P  X  a  1
1
K  K  1 K 
2
3
a 0 a
1 1 1
Consider P  X  a   i.e.  f  x  dx    0dx   3x 2 dx 
2 
2  0
2
a
1
 3x dx  2
2
0
a
 x3  1
 3  
 3 0 2
1/ 3
1 1
 a3  a   .
2 2
0, x0
10. The CDF of a continuous random variable is given by F ( x)  
 
x
1  e 5 , 0  x  
Find the PDF of X and mean of X.
Solution:
0, x0
d 
PDF = f ( x)   F ( x)   1  x
dx  e , x0
5
5
 
1  5x
E( X )   xf ( x ) dx  0 5 xe dx

  x    x 
1 e 5   e 5  25
 ( x)    (1)     5
5 1   1  5
  5   25   0
11. Let X be a random variable with E(X) = 1, E(X(X-1)) = 4. Find Var (X), Var (3+2X).
Solution:
E  X  1
E  X  X  1   4  E( X 2  X )  4
E( X 2 )  E( X )  4  E ( X 2 ) 1  4  E( X 2 )  5
Var  X   E X 2     E  X 
2
 5 1  4
Var  3  2 X   (2) 2 Var  X   4  4  16 Var  aX  b   a 2 Var  X 
12. Let M x  t   1 , t1 be the MGF of RV X. Find the MGF of Y = 2X +1.
1 t
Solution:
M Y (t )  M 2 X 1 (t )  et M X (2t )  M aX b (t )  ebt M X (at ) 
 1  et
 M X (at )   M X (t )t at 
 et   .
1  t  t 2t 1  2t
4
13. If the random variable has the moment generating function M  t   3 , compute
3t
X
E[X2].
Solution:
1
3 3  t
M X t     1  
3t  t   3
3 1  
 3
2 3
t t t
 1         
3 3 3
1  t  2  t2  2  t3 
1          
3  1!  9  2!  9  3! 
 tr 
 
E X r  r'  coefficient of   in M X  t 
 r!
 t2 
 
E X  2  coefficient of   in M X  t   .
2 ' 2
 2!  9
2e2 x ; x  0
14. A random variable X has density function given by f  x    . Find the
0 ; x  0
MGF of X.
Solution:
M X  t   E e Xt  

  e f  x  dx
tx


  etx 2e 2 x dx
0

 2 e 2t  x dx
0

 e  2 t  x  2  0
 2   [e  e ]
 (2  t )  0 2t
.
2 2
 [0  1]  ,t  2
2t 2t
2 x
15. A continuous RV X has the probability density function f (x)  x e , x  0 . Find the
2
rth moment of X about the origin.
Solution:

r  E[ X r ]  x
r
f ( x)dx

5
 
x 2e x 1
  xr dx   x r  2e x dx
0
2 20

1
  x( r 3)1e x dx
20

1
x
n 1  x
 (r  3) e dx  (n)
2 0
1
 (r  2)! if n is positive int eger (n )  (n  1)!
2
16. For a Binomial distribution with mean 6 and standard deviation 2 , find the first
two terms of the distribution. (May/June 2014)
Solution:
1 2
np  6 and npq  2  npq  2  6q  2  q   p
3 3
2
n 6  n9
3
x 9 x
 2 1
P  X  x   n Cx p q x n x
 9C x     , x  0,1, 2,3,...9
 3 3
0 9 9
 2 1 1
P  X  0   9C0       
 3  3  3
1 8
 2 1 2 1 2
P  X  1  9C1      9   8  7
 3 3 3 3 3
17. Find the second moment about the origin of the Geometric distribution with
parameter p . (April/M 2019)
Solution:

E  x 2    x 2 q x 1 p
x 1

   x 2  x   x  q x 1 p
x 1
 
  x  x  1 q x 1 p   xq x 1 p
x 1 x 1
 1.2 p  2.3qp  3.4q 2 p  ...   p  2qp  3q 2 p  ...

 p 1.2  2.3q  3.4q 2  ...  p 1  q 
2
 p2 1  q   p 1  q 
3 2
 p2  p   p  p 
3 2
2p p 2 p
   2 .
p3 p 2 p
6
18. Suppose the length of life of an appliance has an exponential distribution with mean
10 years. What is the probability that the average life time of a random sample of
the appliances is atleast 10.5 years?
Solution:
Mean of the exponential distribution = E(X) = 1/10 = 1/
x
1 x 1 10
  , f (x)  e , x  0  f (x)  e , x  0
10 10
  x
1 10
P(X  10.5)  f (x)dx  e dx  e 1.05  0.3499
10.5 10.5
10
19. Suppose that the duration X in minutes of long distance calls from your home,
 5x
 x0
follows exponential law with probability density function f ( x )   e ,
 0, otherwise
what is P(X>5)? (Nov/Dec 2017)
Solution:
  x
P( X  5)   f  x  dx   e 5
dx
5 5
x 
 5  e   e 1  
5
e 5

 5
.
1 e
5
20. If X is a normal variate with mean  30 and S.D = 5.Find P  26  X  40
Solution: X follows N(30, 5)   30 &   5

X 
Let Z  be the standard normal variate

 26  30 40  30 
P  26  X  40  P  Z
 5 5 
 P  0.8  Z  2  P 0.8  Z  0  P 0  Z  2
 P 0  Z  0.8  0  Z  2  0.2881  0.4772  0.7653 .

PART-B
1. A given lot of IC chips contains 2% defective chips. Each chip is tested before
delivery. The tester itself is not reliable. Probability of tester says the chip is good
when it is really good is 0.95 and the probability of tester says chip is defective when
it is actually defective is 0.94. If a tested device is indicated to be defective, what is
the probability that it is actually defective?
Solution:
Let A be the event of chip that is actually defective & B be the event of chip that is
actually good.
Let D be the event of tester says it is good.
7
Given P  A  0.02 , P  B   0.98 , P  D / B   0.95

P  D / A  0.94 , P  D / B   1  P  D / B   1  0.95  0.05
By the theorem of total probability
P  D   P  A P  D / A  P  B  P  D / B 
=0.02  0.94+0.98  0.05
=0.0678
By Baye’s rule,
P  A P  D / A
P  A / D 
P  D
0.94  0.02
=
0.0678
P  A / D   0.2772 .
2. The first bag contains 3 white balls, 2 red balls and 4 black balls. Second bag
contains 2 white, 3 red and 5 black balls and third bag contains 3 white, 4 red and 2
black balls. One bag is chosen at random and from it 3 balls are drawn. Out of three
balls two balls are white and one is red. What are the probabilities that they were
taken from first bag, second bag, third bag?
Solution:
Urn I Urn II Urn III
3 2 4 2 3 5 3 4 2
W R B W R B W R B
Where W, R, B denotes white, red & black balls
Let it be A1 , A2 , A3
1
P  A1   P  A2   P  A3  
3
Let D be the event of selecting three balls taken from the selected bag that 2 are white
and 1 is red.
3C  2C1 6
P  D / A1   2 
9C3 84
2C  3C1 3
P  D / A2   2 
10C3 120
3C  4C1 12
P  D / A3   2 
9C3 84
By the theorem of total probability,
P  D  P  A1  P  D / A1   P  A2  P  D / A2   P  A3  P  D / A3 
1 6 1 3 1 12
     
3 84 3 120 3 84
 0.0746.
8
By Baye’s theorem,
6 1
P  A1  P  B / A1  84  3
P  A1 / B     0.319
P  D 0.0746
1 3
P  A2  P  B / A2  3  120
P  A2 / B     0.0428
P  D 0.0746
1 12
P  A3  P  B / A3  3  84
P  A3 / B     0.6380
P  D 0.0746
3. State and Prove Baye’s Theorem. (April/May 2019)
Solution:
Statement:
n
If E1 , E2 ,...En are mutually disjoint event in S such that S  Ei then for any A  S
i 1
P  Ei  P  A 
we have P  i   n
E  Ei 
 A
 P  Ei  P  A 
i 1  Ei 
Proof:
We have A   A  E1    A  E2   ...   A  En 
P  A  P  A  E1    A  E2   ...   A  En 
P  A  P  A  E1   P  A  E2   ...  P  A  En 
n
P  A   P  A  Ei     (1)
i 1
P  A  Ei 
But P  A  
 Ei  P  Ei 
 P  A  Ei   P  Ei  P  A     (2)
 Ei 
Put (2) in (1)
P  A   P  Ei  P  A     (3)
n
i 1  Ei 
P  Ei  A
Now P  i  
E
   (4)
 A P  A
Put (2),(3) in (4)
P  Ei  P  A 
P Ei 
 n  Ei  .

 A
 P  Ei  P  A 
i 1  Ei 
9
1
4. (i) A random variable X has the probability mass function f (x)= , x= 1, 2, 3, ...
2x
Find its (i) M.G.F (ii) Mean (iii) Variance.
Solution:
(i) M.G.F= M X (t )  E (e tX )
x

1   et


=  e f ( x)  e x  
tx tx

x 1 x 1 2 x 1  2 
2 3 4
et  et   et   et 
=           ...
2  2   2   2 
et   et   et   et  
2 3
= 1           ...
2 2 2 2 

1
et  et 
= 1  
2  2
et
M.G.F= M X (t )  …………….. (1)
2  et
d  et   (2  e t )e t  e t (e t ) 
(ii) Mean = E(X)= M X (t )t 0 
d
     2
dx dx  2  e t  t 0  (2  e t ) 2  t 0
(iii) Variance = Var ( X )  E ( X )  E ( X ) 
2 2
d d  2et 
Where E ( X 2 ) =  M 'X (t)    
dx t 0
dx  (2  e t ) 2  t 0
 (2  et )2 et  et 2(2  et )(e t ) 
 2  6
 (2  et )4  t 0
Variance = Var ( X )  E ( X )  E ( X )  = 6-4 = 2
2 2
x 1
 3  1 
(ii) Let P  X  x      , x  1, 2, 3, be the probability mass function of the
 4  4 
R.V. X. Compute  i  P  X  4  ii  P  X  4 / X  2  iii  E  X  (iv ) Var ( X ).
(May/June 2016)
Solution:
(i) P( X  4)  1  P( X  4)
 1   P( X  1)  P( X  2)  P( X  3)  P( X  4)
 3  1 0  3  1 1  3  1  2  3  1 3 
 1                 
 4  4   4  4   4  4   4  4  
10
 3  1   1   1  
2 3
 1    1          
 4    4   4   4  
 3  85  1
 1     
 4  64  256
P( X  4  X  2) P( X  4)
(ii ) P( X  4 / X  2)  
P( X  2) P( X  2)
Now
P( X  2)  1  P( X  2)
 1   P( X  1)  P( X  2) 
 3  1 0  3  1 1 
 1         
 4  4   4  4  
 3    1 
 1    1    
 4    4 
 3  5 
 1    
 4  4 
1

16
1/ 256 1
P( X  4 / X  2)  
1/16 16
iii  E  X    x P( X  x)
 3 
2
1 1
   1  2.    3.    
 4   4 4 
2
 3  1
   1  
 4  4
2
 3 3
   
 4 4
3 16 4
 * 
4 9 3
 iv Var  X   E  X 2   E  x 
2
Now  
E X 2   x 2 P( X  x )
11
  x  x  1 P ( X  x)   x P( X  x)
 3  4
2
1 1
   1.2  2.3    3.4    
 4   4 4  3
3
 3   1 4
   .2. 1   
 4   4 3
3
 3  3 4
   .2.   
 4  4 3
3 64 4
 .2. 
4 27 3
32 4 20
  
9 3 9
2
20  4  4
Var  X     
9 3 9
5. (i) If the density function of a continuous random variable X is given by
 ax , 0  x 1
 a, 1 x  2

f ( x)  
 3a  ax , 2 x3

 0, elsewhere
(1) Find the value of a (2) Find the CDF of X (3) Find P(X  1.5 ).
Solution:

(i) We know that  f  x  dx  1

1 2 3
0 ax dx  1 a dx  2 3a  ax  dx  1
1 3
 x2  2  x2 
a    a  x 1  3ax  a   1
 2  0  2 
2
1   9  4 
a   0   a  2  1    9a  a    6a  a    1
2   2  2 
a   9a  
 a      4a    1
2  2  
a a
 a  1
2 2
1
2a  1  a 
2
12
x
(ii) If x  0, F  x   P  X  x    f  x  dx  0

x
x 1  x2  x2
x
If 0  x  1, F  x   P  X  x    f  x  dx   ax dx    
 0 2  2  4
0
x
If 1  x  2, F  x   P  X  x    f  x  dx

1 x
  ax dx   a dx
0 1
1
 x2  x
 a    a  x 1
 2 0
a
  a  x 1 
2
1 1
   x  1
4 2
x 1
 
2 4
x
If 2  x  3, F  x   P  X  x    f  x  dx

1 2 x
  ax dx   a dx  
0 1 2
3a  ax  dx
1 x
 x2  2  x2 
 a    a  x 1  a 3x  
 2  0  2 
2
1   x2   4 
 a   0   a 2  1  a  3x     6  
2   2   2  
a  x2 
  a  a 3x   4
2  2 
1 x2 
 a   1  3x   4
2 2 
 
1 x2 5 
  3x   
2  2 2 
3x x 2 5
  
2 4 4
x x
If x  3, F  x   P  X  x    f  x  dx   f  x  dx  1
 0
13
0, x0
 2
x , 0  x 1
4
x 1

 F x    , 1 x 2
2 4
 3x x 2 5
   , 2x 3
2 4 4
1, x 3
1.5
P( X 1.5)  
0
f ( x) dx
1 1.5
  f ( x) dx   f ( x) dx
0 1
1 1.5
x 1
  dx   2 dx
0
2 1
2 1 1.5
x x
 
4 0
21
 1   1.5 1  1
  0    
4   2 2 2
(ii) A continuous random variable X has the probability density function
f  x   kx3e  x , x  0. Find the rth order moment of X about the origin. Hence find
MGF, Mean and Variance of X.
Solution:


  e x  2 e
x
  e x   e x  
(ii) Since  kx e dx  1  k  x 3 
3 x
  (3 x )    (6 x )    (6)    1
0   1   1   1   1 0

k  x3e x  3x2e x  6xe x  6  1  k (0)  (6)  1 6 k  1  k  .
1
0 6
   
1 1
E ( X r )  r   x r f  x dx   x r x3e x dx   x r 3e x dx n   e x x n 1dx , n  0
0
60 60 0

1  x  r 311 1  r  3 !
here n  r  4 
60 e x dx  (r  4) 
6 6
n  (n  1)!
4! 24
Putting r  1 , E ( X )  1   4
6 6
5! 120
r  2 , E ( X 2 )  2    20
6 6
 
2
 Mean = E ( X )  1  4 ; Variance = E ( X 2 )   E ( X )   2  1
2
14
2  20   4  20  16  4
2
To find M.G.F

M X (t)  E(e tX )  e
tx
f (x)dx


1 3 x
M X (t)  e
tx
x e dx

6
 
1 3 tx  x 1
 
60
x e dx   x 3e(1 t)x dx
60

1  e(1 t)x  2 e
 (1 t)x   e(1 t)x   e(1 t)x 
 
  x 3    3x       6x   
3
  6    
 (1  t)   (1  t)  (1  t)   (1  t)
2 4
6     0

1 e(1t)x e(1t)x e(1t)x e(1t)x 
 x3  3x 2  6x  6 
6 (1  t) (1  t)2 (1  t)3 (1  t)4  0
1  6  
 (0)   4 
6   (1  t)  
1
 M X (t)  .
(1  t)4
6. Derive the MGF, mean and variance of Binomial distribution.
M X t   E etX  
n
  etx p  x 
x 0
n
  etx ncx px qn  x
x 0
n
 ncx  pet 
x
 qn  x
x 0
 
n
M X t   q  pet
NOTE:
 
n
M X t   q  pet
   
n 1  n 1 
M ' t   n q  pet . pet  np et q  pet 
X 
 np 1  q  p 1    np
n 1
M ' t 
X t 0  
 i.e. Mean  1'  np
15
   
 n 2 n 1 t 
M '' t   np et  n  1 q  pet . pet  q  pet e 
X  
 np 1 n  1  q  p 1  . p 1  q  p 1 
n 2 n 1
M '' t  1 
X t 0 
 
 np  p  n  1  1
 np  pn  p  1
 i.e. 2'  n2 p2  np2  np
2
 Var  X   2'   1' 
 n2 p2  np2  np  n2 p2
 np 1  p 
Var  X   npq .
7. Out of 800 families with 4 children each, how many families would be expected to
have (i) 2 boys and 2 girls; (ii) at least 1 boy; (iii) at most 2 girls and (iv) children of
both genders. Assume equal probabilities for boys and girls. (April/May 2019)
Solution:
Considering each child as a trial, n  4
1 1
Assuming that birth of a boy is a success, p  q 
2 2
Let X denote the number of successes (boys)
(i) P 2 boys and 2 girls  P  X  2
2 2
1  1 
 4c2    
2 2
 0.375
 No. of families having 2 boys and 2 girls  N. P  X  2
 800 0.375
 300
(ii) P at least 1 boy   P  X  1
 1  P  X  1
 1  P  X  0
0 4
1  1 
 1  4c0    
2 2
 0.9375
 No. of families having at least 1 boy  N.P  X  1
 800 0.9375
 750 .
16
(iii) P  atmost 2 girls  P  exactly 0 girl, 1girl, 2 girls
 P  X  4  P  X  3  P  X  2
4 0 3 1 2 2
 1  1  1  1  1  1
 4c4      4c3      4c4    
 2  2  2  2  2  2
= 0.6875.
No. of families having atmost 2 girls  800  0.6875  550 .
(iv) P Children of both genders  1  P Children of the same gender 
 1  P  all are boys  P  all are girls
 1  P  X  4   P  X  0 
  1
4
 1 
4
 1  4c4    4c0   
  2  2 

 0.875 .
No. of families having children of both genders  800 0.875
 700 .
8. (i) Derive the moment generating function of Poisson distribution and hence find its
first three central moments. (April/May 2019)
(ii) Prove that the sum of two independent Poisson variates is a Poisson variate.
Proof:
M X1 t   e

 1 et 1 
M X 2 t   e

 2 et 1 
M X1 t   X2 t   M X1 t   M X2 t 
e
   e 2 et 1
1 et 1
1   2  et 1
e
RHS is the MGF of a Poisson random variable with parameters  1   2 .
 X1  X 2 is a Poisson random variable with parameter  1   2 .
9. Six coins are tossed 6400 times. Using the Poisson distribution, what is the
approximate probability of getting six heads 10 times?
6
1
Probability of getting six heads in one toss of six coins is p   
2
Number of tosses n = 6400
6
1
Therefore   np  6400    =100
2
Let X be the number of times getting 6 heads
17
Hence the probability of getting six heads 10 times is

100 10
e (100)
P( X  10)   1.025  1030
10!
10. The number of monthly breakdowns of a computer is a random variable having a
Poisson distribution with mean equal to 1.8. Find the probability that this computer
will function for a month (1) without a breakdown (2) with only one breakdown.
(Nov/Dec 2017)
Solution:
Let X denotes the number of breakdowns of the computer in a month.
X follows a Poisson distribution with mean 1.8.
x
(i.e.)   1.8 , P  X  x   e  , x  0,1,2,... .
x!
0
e 1.8 1.8  e 1.8
(i) P  X  0    0.1653
0! 1
1
e 1.8 1.8 
(ii) P  X  1   e 1.8 1.8   0.2975
1!
11. A car hire firm has 2 cars. The number of demands for a car on each day is
distributed as Poisson variate with mean 0.5. Calculate the proportion of days on
which (1) Neither car is used (2) Some demand is refused. (April/May 2017)
Solution:
Since cars distribution in each day follows a Poisson pattern with mean 0.5
e   x
   0.5 P( X  x)  , x0
x!
e0.5 0.50
(i) P( X  0)   0.6065
0!
(ii ) P( some demand is refused )  P( X  2)
 1  P( X  2)
 1   P( X  0)  P( X  1)  P ( X  2) 
 0.50 0.51 0.52 
 1  e 0.5    
 0! 1! 2! 
 1  0.9856
 0.0144
12. In a certain factory producing razor blades, there is a small chance of 1/500 for any
blade to be defective. The blades are supplied in packets of 10. Use Poisson
distribution to calculate the approximate number of packets containing
(a) no defective blade
(b) at least 1 defective blade and
18
(c) at most 1 defective blade in a consignment of 10,000 packets.

Solution:
Let X denote the Poisson distribution with
1
p , n  10, N  10000
500
1
  np  10 *  0.02
500
0.02
 0.02 
x
e   x e
P( X  x)  
x! x!
e 0.02  0.02 
0
(i ) P( X  0)   0.9802
0!
For 10000 packets = 10000 * 0.9802 = 9802 packets
(ii ) P(atleast one defective)  P  X  1  1  P  X  1
e 0.02  0.02 
0
 1
0!
 1  0.9802
 0.0198
(iii ) P (atmost one defective)  P  X  1  P  X  0   P  X  1
e 0.02  0.02  e 0.02  0.02 
0 1
 
0! 1!
 0.9802  0.0196
 0.9998
13. If the probability that a target is destroyed on any one shot is 0.5.
(i) What is the probability that it would be destroyed on 6th attempt?
(ii) What is the probability that it takes less than 5 shots?
(iii) What is the probability that it takes him an even number of shots?
Solution:
Given p  0.5 , q  1  p  0.5
Let X denote the no. of attempts required for the destruction of the target (1st success)
Then X Geometric dist.
(i.e.) P  X  x   q x 1 p, x  1,2,...
(i) The probability that the target destroyed on 6th attempt
19
P  X  6   q5 p
5
  0.5   0.5 
 0.0156 .
(ii) Probability that it takes less than 5 shots
P(X < 5) = P(X = 1) + P(X = 2) + P(X = 3) + P(X = 4)
= p + pq + p 2 q + p3q
= 0.5 + (0.5 * 0.5) + (0.52 * 0.5) + (0.53 * 0.5)
= 0.9375
(iii) Probability that it takes even number of shots
P(Xis even) = P(X = 2) + P(X = 4) + P(X = 6) +
= pq + p3q + p5q +
= (0.5*0.5) + (0.53 *0.5) + (0.55 *0.5) +
2 3
1 1 1
=     
4 4 4
1 1 1 1 
2 3
 1         
4  4  4   4  
1 1
1  1 1 3 1 4 1
 1         
4  4 4 4 4 3 3
14. Derive the MGF, mean and variance of Geometric distribution and also state and
prove the special property of it. (May/June 16)
Solution:
Moment Generating Function (MGF)

M X  t   E (etX )   etx p( x)
x 1

  etx q x 1 p
x 1
 p[et  e 2t q  e3t q 2  ]
 pet [1  qet  qet   2
 ]
 pet [1  qet ]1
pet

1  qet
Mean and Variance
 d  pet   
pet
  M (0)   
' '
      1
1 X
 dt  1  qet
     

  t 0   1  qe
t

2  
  t 0
p
20
 d 2  pet    
  M (0)   2 
' "
    
d pet   1  q
2 X
 dt  1  qet
     
 dt 
  t 0   1  qe
t

2


  t 0
p2
1
Mean  1' 
p
2
1 q  1 
  q
2
Variance      2     2
'
2
'
1
p  p p
Memoryless property of geometric distribution.
Statement:
If X is a random variable with geometric distribution, then X lacks memory, in the sense
that P  X  s  t / X  s  P  X  t  s, t  0 .
Proof:
The probability mass function of the geometric random variable X is
x 1
P( X  x )  q p , x  1, 2,3,....
P  X  s  t  X  s P  X  s  t 
P  X  s  t / X  s        (1)
P  X  s P  X  s

P X  t  q x 1
p  qt p  qt 1 p  qt  2 p  ....  qt p 1  q  q 2  q3  ....
x t 1
 q p(1  q)1  qt p( p)1  q t

t
Hence P  X  s  t   qst and P  X  s  qs

q s t
(1)  P  X  s  t / X  s   s  q t  P[ X  t ]
q
 P  X  s  t / X  s  P  X  t  .
15. Let X be a Uniformly distributed R.V over [-5,5]. Determine
 i  P  X  2  ii  P  X  2 (iii) Cumulative distribution function of X
(iv ) Var ( X ). (May/June 16)
Solution:
X is Uniformly distributed R.V over [-5, 5]
 1 1
  5  x  5
 f ( x)   5  (5) 10
0
 otherwise
2 2
1 1
(i ) P( X  2)   dx   dx
5
10 10 5
1 2 1 7
 x 5  (2  (5)) 
10 10 10
21
2
1
(ii ) P( X  2)  1  P( X  2)  1   10 dx
2
2
1
10 2
1  dx
1 2 1
 1 x 2 1  (2  (2))
10 10
4 3
1  
10 5
(iii) Cumulative distribution function of X
Case(i) When x < -5
x
F ( x)   f ( x) dx

x
  0 dx  0

Case(ii)
when 5  x  5
x 5 x
F ( x)  

f ( x) dx  

f ( x) dx   f ( x) dx
5
5 x
1
  0 dx   10 dx
 5
1 x 1
 x 5  ( x  (5))
10 10
1
 ( x  5)
10
Case(iii) When x > 5
x 5 5 x
F ( x)   f ( x) dx   f ( x) dx   f ( x) dx   f ( x) dx
  5 5
5 5 x
1
  0 dx   dx   0 dx
 5
10 5
1 5 1
 x 5  (5  (5))
10 10
1
 (5  5) 1
10
22
(iv) Var(X)
b  a 
2
If f(x) is defined over (a, b) then variance =

12
 5  (5) 
2
100
Therefore Var ( X )    8.33 .
12 12
16. Trains arrive at a station at 15 minutes interval starting at 4 a.m. If a passenger
arrives at a station at a time that is uniformly distributed between 9.00 a.m. and
9.30 a.m., find the probability that he has to wait for the train for (i) less than 6
minutes (ii) more than 10 minutes.
Solution:
Let X denotes number of minutes past 9.00 a.m. that the passenger arrives at the stop till
9.30 a.m.
1
 , 0  x  30
X ~ U[0,30]  f ( x)   30

 0, otherwise
(i) P  that he has to wait for the train for less than 6 minutes 
 P (9  x  15)  (24  x  30)
15 30
  f ( x)dx   f ( x)dx
9 24
15 30
1 1
 9 30 dx   30 dx
24

1
30

 x 9   x 24
15 30

12
  0.4
30
(ii) P  that he has to wait for the train for more than 10 minutes 
 P (0  x  5)  (15  x  20)
5 20
  f ( x)dx   f ( x)dx
0 15
5 20
1 1
 dx   30 dx
0
30 15

1
30

 x 0   x 15
5 20

10 1
  .
30 3
23
17. A component has an exponential time to failure distribution with mean of 10,000
hours.
(i) The component has already been in operation for its mean life. What is the
probability that it will fail by 15,000 hours?
(ii) At 15,000 hours the component is still in operation. What is the probability
that it will operate for another 5000 hours?
(Nov/Dec 2015)
Solution:
Let X be the random variable denoting the time to failure of the component
1 1
following exponential distribution with Mean  10000 hours.  10,000   
 10,000
 1 
x
 e 10,000 , x  0
The p. d. f. of X is f  x   10, 000

 0 , otherwise
(i) Probability that the component will fail by 15,000 hours given that it has already
been in operation for its mean life  P  X  15,000 / X  10,000
P 10,000  X  15,000
        (1)
P  X  10,000
15,000 x

P 10,000  X  15,000 
1
 10000
10,000
e 10000
dx
15000
  x  15000
1  e 10000    10000
x

     e 
10000  1   10000
 10000 10000
  15000  10000   3 
  e 10000  e 10000    e 2  e 1   e 1  e 1.5    (2)
   

  x  


x
 e 10000    10000
x

P  X  10,000  
1 1
e 10000
dx      e 
10,000
10000 10000  1   10000
 10000 10000
  e   e 1   e 1        (3)
Sub (2) & (3) in (1)
e1  e1.5 0.3679  0.2231
(1)  P  X  15,000 / X  10,000    0.3936 .
e1 0.3679
24
(ii) Probability that the component will operate for another 5000 hours given that it is in
operation 15,000 hours  P  X  20,000/ X  15,000
 P  X  5000 [by memoryless property]

  f  x  dx
5000

  x  
1  e 10000    10000 
 x x
1 
 
5000
10000
e 10000
dx  
10000  1 
   e


 5000
 10000  5000
 e0.5  0.6065 .
18. (i) State and prove the memoryless property of exponential distribution.
If X is exponentially distributed, then P  X  s  t X  s  P  X  t  , for any s, t  0 .
Proof:
Since X is exponentially distributed,
f  x   e x , x0

P  X  k   f  x  dx
k

   e   x dx
k

 ex 
   
    k
  e  k  
 0   
  1  
P  X  k  e  k
Now
P  X  s  t X  s
P  X  s  t X  s 
P  X  s
P  X  s  t

P  X  s
e  
  s t
e   s e  t
  s   e  t
e e  s
P  X  s  t X  s  P  X  t  .
25
(ii) Suppose X has an exponential distribution with mean equal to 10. Determine the
value of x such that P (X < x) = 0.95. (April/May 2017)
Solution:
We know that the probability density function for the exponential distribution is
f  x   e x , x0
Here  10
P  X  x   0.95
x
10e
10 x
dx  0.95
0
x
e10 x
10  0.95
10 0
 e10 x 1 
10     0.95
 10 10 
1  e10 x  0.95
 e10 x  0.95  1
 e10 x  0.05
10 x  log(0.05)
10 x  2.9957
2.9957
x  x  0.2996
10
19. Derive moment generating function normal distribution (April/May 2019)
Solution:
Moment generating function of Normal distribution  M X  t   E etx 
1  x 
2
  
1 
e
2  
 tx
e dx
 2 
x
Put z  then  dz  dx,    Z  

 z2
1 t  z    
 M X t   e 2
dz
2 
 z2 

e t  t z 

2 e

 2 
dz
1   2t 2 
  z t 2  
e t  
 e
2  2 
dz
2 
26
 2t 2
t  1
e e 2   z t 2

2 e

2
dz
 1
1   z t 2
the total area under normal curve is unity, we have
2 e

2
dz  1
 2t 2 t2
t 
Hence M X  t   e 2
For standard normal variable N  0,1 , M X  t   e 2
20. The average percentage of marks of candidates in an examination is 42 with a

standard deviation of 10. If the minimum mark for pass is 50% and 1000 candidates
appear for the examination, how many candidates can be expected to get the pass
mark? If it is required, that double the number of the candidates should pass, what
should be the minimum mark for pass?
Solution:
Let X denote the marks of the candidates, then X N  42,102 
X  42
Let z  , P  X  50  P  z  0.8  0.5  P 0  z  0.8  0.5  0.2881  0.2119
10
If 1000 students write the test, 1000P  X  50  212 students would pass the
examination.
If double that number should pass, then the no of passes should be 424.
We have to find z1 , such that P  z  z1   0.424
 P 0  z  z1   0.5  0.424  0.076
50  x1
From tables, z1= 0.19 , z1   x1  50  10 z1  50  1.9  48.1
10
The pass mark should be 48 nearly.
UNIT – II TWO DIMENSIONAL RANDOM VARIABLES
PART-A
1. Given the joint probability density function of X and Y as
1
 , 0  x  2, 0  y  3
f ( x, y )   6 , determine the marginal density functions.

 0, otherwise
Solution:
 3
 y
3
1 1
The marginal function of X is f X ( x)  

f ( x, y )dy   dy     , 0  x  2
0
6  6 0 2
 2
x
2
1 1
The marginal function of Y is fY ( y)  

f ( x, y )dx   dx     , 0  y  3
0
6  6 0 3
2. Find the value of k, if the joint density function of (X , Y) is given by
k (1  x)(1  y ) , 0  x, y  1
f ( x, y )   .
 0 , otherwise
27
Solution:
Given the joint pdf of (X , Y) is f(x , y) = k (1 – x) (1 – y), 0 < x < 4, 1 < y < 5
  1 1
  f ( x, y )dxdy  1    k (1  x)(1  y )dxdy  1
  0 0
1
1
 x2 x2 
 k   x   yx  y  dy  1
0
2 2 0
1 1 
5
 k    y  dy  1
1
2 2 
1
 1 1 y 2 
 k  y    1
  2 2 2 0
1
 k 1 k  4
4
3. The joint probability density function of bivariate random variable (X , Y) is given
4 xy , 0  x  1, 0  y  1
by f ( x, y )   . Find P (X + Y<1 )
 0 , elsewhere
Solution:
4 xy , 0  x  1, 0  y  1
Given the joint pdf of (X , Y) is f ( x, y )   .
 0 , elsewhere
1 1 x 1 x
1
 y2 
 P( X  Y  1)    4 xydydx  4  x   dx
0 0 0  2 0
1
 2  x(1  x) 2 dx
0
1
 2 x(1  2 x  x 2 )dx
0
1
 2 ( x  2 x 2  x3 )dx
0
1
 x2 x3 x 4 
 2 2  
2 3 4 0
1 2 1 1
 2    
2 3 4 6
28
8 xy , 0  x  1, 0  y  x
4. If f ( x, y )   is the joint probability density function of X and
 0 , elsewhere
Y, find f(y/x).
Solution:
x
x  y2 
f X ( x)   f ( x, y )dy   8 xydy  8 x   4 x3 , 0  x  1
y y 0  2 
0
f ( x, y ) 8 xy 2 y
f ( y / x)  ,  3  2 , 0  y  x, 0  x  1
f X ( x) 4x x
K ,0  y  x  1
5. The joint p.d.f. of R.V. (X,Y) is given as f ( x, y )   . Where K is a
 0 , elsewhere
constant Determine the value of K . (A/M 2019)
Solution:
 
   f  x, y  dxdy  1
 
1 1
   kdxdy  1
0 y
1
 k   x 0 dy  1
1
0
1
 k  dy  1
0
 k 1
6. The following table gives the joint probability distribution of X and Y, find the
marginal distribution function of X and Y.
X
1 2 3
Y
1 0.1 0.1 0.2
2 0.2 0.3 0.1
Solution:
X
1 2 3 p(y)
Y
1 0.1 0.1 0.2 0.4
2 0.2 0.3 0.1 0.6
p(x) 0.3 0.4 0.3 1
The marginal distribution of X is The marginal distribution of Y is

X 1 2 3 Y 1 2
p(x) 0.3 0.4 0.3 p(y) 0.4 0.6
29
7. The regression equations are 3x + 2y = 26 and 6x + y = 31. Find the means of X and
Y. (Nov/Dec 2017)
Solution:
Regression lines pass through the mean values of X and Y. Solving the two equations we
get the mean values.
Let 3x + 2y = 26 --------------(1)
6x + y = 31--------------(2)
Multiply equation (2) by 2 and subtract equation (2)
3x  2y  26
12x  2y  62
( ) ( ) ( )

9 x  36
Substitute in equation (1)
3 (4) – 2y = 26  y = 7.
mean value of X = 4 and mean value of Y = 7
8. Let X and Y be two independent R.Vs with Var(X) = 9 and Var(Y) = 3.
Find Var (4X – 2Y + 6)
Solution:
Var (4X – 2Y + 6) = 16 Var(X) + 4 Var(Y) = 16(9) + 4(3) = 156
9. If the joint cumulative distribution function of X and Y is given by
F ( x, y)  (1  e x )(1  e y ), x  0, y  0 , find P ( 1 < X < 2 , 1 < Y < 2 )
Solution:
The joint pdf is
2 F 2 
f ( x, y )  
xy xy
1  e  x 1  e  y   1  e  x  .e  y  e  x .e  y  e  ( x  y ) , x  0, y  0
x
2 2 2 2 2 2
P 1  X  2,1  Y  2     f ( x, y)dxdy    e ( x y )
dxdy    e x .e y dxdy
1 1 1 1 1 1
 1 1   e 1 
2 2 2 2
  e dx. e dy  e  . e   e  e  

2 2 2
x y x y 1 2
   2    2   0.054
1 1
1 1
e e   e 
10. The lines of regression in a bivariate distribution are X + 9Y = 7 and Y + 4X = 49 .

3
Find the coefficient of correlation.
49
Solution: x + 9y – 7 = 0 ----------- (1) y + 4x – = 0 ----------- (2)
3
Let (1) be the regression line of Y on X and let (2) be the regression line of X on Y.
30
1 7 1
y x   b1  
9 9 9
1 49 1 1 1 1 1
x y  b2    r  b1 b2   .    1
4 12 4 9 4 36 6
Since both regression coefficients are negative, correlation coefficient is negative.
11. Determine the value of the constant c if the joint density function of two discrete
random variables X and Y is given by p(x,y) = cxy, x = 1,2,3 and y = 1,2,3.
Solution:
X
1 2 3 p(y)
Y
1 c 2c 3c 6c
2 2c 4c 6c 12c
3 3c 6c 9c 18c
p(x) 6c 12c 18c 36c
Since p (x, y) is the joint pdf of X and Y
p (x, y) ≥ 0 , for all x ,y
 p( x, y)  1  36c  1  c  36 .
1
x y
12. If Y = -2X + 3, find Cov (X, Y).
Solution:
Cov(X,Y) = E(XY) – E(X) E(Y) = E(X(-2X + 3)) – E(X){E(-2X + 3)}
= [E(-2X2 + 3X) - E(X)]{-2E(X) + 3}
= -2E(X2) + 3 E(X) + 2 (E(X))2 - 3E(X)
= 2(E(X))2 – 2 E(X2) = -2 var(X)
13. Let X and Y be two random variables having joint density function
3  1 1
f ( x, y )  ( x 2  y 2 ), 0  x  1, 0  y  1. Determine P X  , Y  
2  2 2
Solution:
1 1

 1
1
 
2 2
1 3 2
P  X  , Y      f ( x, y)dydx    x  y 2 dydx
 2 2  x  1
y x 0 y  1
2
2 2
1 1 1
1
3  2
y3  3   1  1  1 
2
3 2  x2 7 
   x 2 y   dx    x 2 1    1   dx      dx
20 3 1 2 0   2  3  8  2 0  2 24 
2
1
 1 1 3  x 7x  3 1 1 7 1  3  8  1
3 2
P  X  , Y         .  .      .
 2 2 2  6 24  0 2  6 8 24 2  2  48  4
31
14. The joint pdf of a two dimensional random variable (X,Y) is given by
f ( x , y )  kxe  y , 0  x  2, y  0. Find the value of k.
Solution:
Given that f(x,y) is pdf of (X,Y) f(x,y) ≥ 0 , for all x ,y
   2  2
2
 x2  
   f ( x, y )dxdy  1    kxe dxdy  1 k  xdx .  e dy  1  k   . e  y   1
y y
 2 0
0
  0 0 0 0
1
 k (2)(1)  1  k 
2
15. The correlation coefficient of two random variables X and Y is 1 while their
4
variances are 3 and 5. Find the covariance.
Solution:
1
Given rxy = ,  2X  3,  2Y  5 rxy = Cov ( X , Y ) ,  X  0,  Y  0
4  XY
Cov(X, Y) = rxy  X  Y =  1 3 5 = - 0.968.
4
16. If X has mean 4 and variance 9, while Y has mean –2 and variance 5 and the two
are independent find (a) E[XY] (b) E[XY2]
Solution:
Given E[X] = 4, E[Y] = –2,  X2  9,  Y2  5 , X and Y are independent.
(a) E[XY] = E[X] E[Y] = 4(–2) = –8
(b) E[XY2] = E[X] E[Y2]
 Y2  E[Y 2 ]  [ E[Y ]]2  5  E[Y 2 ]  4  E[Y 2 ]  9  E[ XY 2 ]  4(9)  36
17. If the joint distribution function of X and Y is given by
 x  y 

1  e x  e y  e , x  0, y  0
F(x , y) =  . Find the marginal density function of X

 0 , elsewhere
and Y.
Solution:
The joint p.d.f of (X,Y) is given by
f ( x, y ) 
2
xy
F ( x, y )   2
xy

1  e x  e y  e  
 x y

x

 y

e e    e  
 x y  x y
 x  y 

e , x  0, y  0
f ( x, y )  

 0 , elsewhere
18. The two lines of regression are 4x – 5y + 33 = 0 and 20x – 9y = 107. Calculate the
coefficient of correlation between X and Y.
Solution:
4x – 5y + 33 = 0 ----------- (1) 20x – 9y = 107 ----------- (2)
Let (1) be the regression line of Y on X and let (2) be the regression line of X on Y.
32
4 33 4
y  x  b1 
5 5 5
9 107 9
x y  b2 
20 20 20
4 9 9 3
 r  b1b2  .    0.6  1
5 20 25 5
19. Define Covariance and Coefficient of correlation between two random variables x
and y .
(April/May 2019)
Solution:
Covariance between x and y Cov( x, y)  E  x  E  x    y  E  y  
Cov  x, y 
Correlation between x and y rxy 
Var  x  Var  y 
20. State central limit theorem
Solution:
If X1 , X 2 .... X n is a sequence of independent random variable E  X i   i and
Var  X i    i2 , i  1, 2,....n and if Sn  X1  X 2  ......  X n then under several conditions
n n
Sn follows a normal distribution with mean    i and variance  2    i2
i 1 i 1
as n   .
PART-B
1. The joint pdf of the random variables X and Y is defined as
 25e 5 y , 0  x  0.2, y  0
f ( x, y)  
 0 , elsewhere
(i) Find the marginal PDFs of X and Y (ii) covariance of (X,Y) (May / June 2015)
Solution:
The marginal function of X is
  
 e5 y  25
 f ( x, y )dy   25e  e   e0   5, 0  x  0.2
5 y
f X ( x)  dy  25   
 0  5 0 5
The marginal function of Y
 0.2
is fY ( y)   f ( x, y)dx   25e
5 y
dx  25e5 y  x 0  25e5 y 0.2  5e5 y , 0  y  
0.2
 0
 0.2
0.2  x2   0.04 
E(X) =  x f X ( x)dx   x(5)dx  5    5 
  0
2  2   0.1
 0
E(Y)
  
5 y   ye5 y e5 y   1
=  y fY ( y )dy   y (5e )dy  5     5 0    0.2
 0  5 25 
0
 25 
33
 
E  xy     x y f ( x, y ) dx dy
 
 0.2  0.2
5 y 5 y
  x y (25e ) dx dy   25 y e dy .  x dx
0 0 0 0
  ye5 y e 5 y   x2  0.2

 1   0.04 
= 25    .    25 0   .   0.02
 5 25 
0  2  0  25   2 
COV(X,Y) = E(XY) – E(X)E(Y) = 0.02 –(0.1)(0.2) = 0.
2.  k ( x  1)e  y , 0  x  1, y  0
Find the constant k such that f ( x , y )   is a joint p.d.f.
 0 , otherwise
of the continuous random variable (X,Y). Are X and Y independent R.Vs? Explain.
(May / June 2016)
Solution:
To find k :
Given that f(x,y) is pdf of (X,Y)
f(x,y) ≥ 0 , for all x ,y
  1  1
1
 x2  
  f ( x, y )dxdy  1    k ( x  1)e dxdy  1 k  ( x  1)dx .  e dy  1  k   x  . e  y   1
y y
2 0
0
  0 0 0 0
3 2
 k   (1)  1  k 
2 3
2 y
 ( x  1)e , 0  x  1, y  0
 f ( x, y )   3
 0 , otherwise
The marginal function of X is
 
2 2  2
f X ( x)   f ( x, y )dy   ( x  1)e  y dy  ( x  1)  e  y   ( x  1)  e   e 0 
 0
3 3 0 3
2 2

( x  1)(1)  ( x  1), 0  x  1
3 3
The marginal function of Y
 1
2
1
2  x2  2 1 
fY ( y )   f ( x, y)dx   ( x  1)e y dx  e y   x   e  y   1
is  0
3 3 2 0 3 2 
2 3
 . e y  e y , 0  y  
3 2
2
Consider f X ( x ) . fY ( y) = ( x  1) . e y = f(x , y)
3
f X ( x). fY ( y)  f ( x, y)
X and Y are independent.
34
cxy 2 , 0  x  y  1
3. Let the joint p.d.f. of R.V. (X,Y) be given as f ( x, y)   .
 0 , elsewhere
Determine (i) the value of C. (ii) the marginal p.d.fs of X and Y (iii) the
conditional p.d.f. of X given Y = y. (May / June 2016)
Solution:
  1 1 1 1
  f ( x, y )dxdy  1    cxy 2 dxdy  1 c  xdx  y 2 dy  1
  0 0 0 0
 x2   y3  1 1
 c       1
  2  0  3  0 
 1  1 
 c     1  c  6
 2  3 
 1
1
 y3   1 x3 
f X ( x)   
f ( x, y ) dy   6 xy dy  6 x    6 x     2 x 1  x 3 , 0  x  1
2

 x  3 x 3 3 
 y
 x2  2 y 
y 2
fY ( y )   f ( x, y ) dx   6 xy dx  6 y    6 y    3 y 4 , 0  y  1
2 2
 0  2 0  2 
Conditional p.d.f of X on Y
f ( x, y ) 6 xy 2
f(x )   2x .
y fY ( y ) 3y2
4. The joint density function of two random variable X and Y is given

 6  2 xy 
 x   , 0  x  1, 0  y  2
by f ( x, y)   7  2  .

 0 , elsewhere
(i) Compute the marginal density function of X and Y ? (ii) Find E(X) and E(Y)
 1 1
(iii) P  X  , Y   (Nov / Dec 2015)
 2 2
Solution:
The marginal pdf of X is
 2
6 xy  6 x y2  6  2 x 4 6
2
f X ( x)   f ( x, y) dy    x 2   dy   x 2 y    2 x     2 x 2  x  , 0  x  1
 0
7  2  7  2 2 0 7  2 2  7
The marginal pdf of Y is
 y
1
 x2  1
fY ( y )  

f ( x, y ) dx   8 xy dx  8 y    8 y    4 y, 0  y  2
0  2 0 2
 1
6  2 x 4 x3  6 2 1 5
1
6
E  X    x f X ( x) dx   x  2 x 2  x  dx        
 0
7 7 4 3 0 7  4 3  7
35
 2
1
 y3   8  32
E Y    y fY ( y ) dy   y 4 y dy  4    4   
 0  3 0 3 3
1
1/2
 1 6 xy  6  x3 x 2 y  6 1 y
2 2 2 2
1
P  X  , Y       x 2   dx dy      dy     dy
 2 2 1 0 7  2 71 3 4  7 1/2  24 16 
2 2 0
2
6 y y  2
69
    
7  24 32 1/2 448
5. If X,Y and Z are uncorrelated random variables with zero means and standard
deviations 5 , 12 and 9 respectively and if U = X + Y, V = Y + Z, find the correlation
coefficient between U and V. (Nov / Dec 2015)
Solution:
E  X   E Y   E  Z   0
E U   E  X  Y   E  X   E Y   0
Similarly E V   0
Var ( X )  E( X 2 )  ( E( X ))2
Var ( X )  E( X 2 )  25
Similarly
E(Y 2 )  144 E(Z 2 )  81
X , Y , Z are uncorrelated
COV ( X , Y )  0  E( XY )  0 E(YZ )  0 E(ZX )  0
 X  5,  Y  12,  Z  9
   
 E U 2  E X 2  Y 2  2 XY  169  U 2
E V 2   E  Z 2  Y 2  2ZY   225  V 2
E UV   E  XY  XZ  YZ  Y 2   144
 COV U ,V   E UV   E U  E V   144
144 48
ru ,v  
1513 65
e  ( x  y ) , x  0, y  0
6. The joint pdf of the continuous R.V (X,Y) is given as f ( x , y )   .
 0 , elsewhere
X
Find the pdf of the random variable U  . (May / June 2016)
Y
Solution:
x
The transformation functions are u  and v  y
y
36
x
Solving for x and y, we get u   x  uv
v
x x
u v v u
The Jacobian of the transformation is J =  v
y y 0 1
u v
The joint density of U and V is fUV ( u, v )  J f XY ( x , y )  v e  ( x  y )  v e  v ( u 1 )
The range space of (U , V) is obtained from the range space of (X , Y) and the
transformations x = uv, y = v
 x  0 and y  0 we have uv  0 and v  0
 u  0 and v  0
v e v (u 1) , u  0, v  0
fUV (u, v)  
 0 , elsewhere,
The pdf of U is the marginal density function of U,
  
 e v (u 1) e v (u 1) 
 f (u, v)dv   v e  v ( u 1)
fU (u )  dv  v .  1. 
 0  (u  1) (u  1)2  0
1 1
 0  , u0
(u  1) 2
(u  1)2
x y
7. Two random variables have the joint PDF f ( x , y )  , 0  x  1, 0  y  2 . Find
3
the correlation coefficient and regression lines. (Apr/May 2017)
Solution:
Marginal P.d.f of X

x y
2
f X  x    f ( x, y )dy   dy
 0
3
2
1 y2  1
 xy    (2 x  2), 0  x  1
3 2 0 3
Marginal P.d.f of Y

x y
1
fY  y    f ( x, y ) dy   dy
 0
3
1
1  x2  1 1 2y 1
   yx   (  y )  , 0 y2
3 2 0 3 2 6
 1
1
E  X    x f X ( x) dx   x (2 x  2) dx

30
1
1  2 x3  5
   1 
3 3 0 9
37

 2y 1
2
E Y    yfY ( y)dy  0 y  6  dy
2
1  2 y 3 y 2  11
    
6 3 2 0 9
 1
EX2 
1 2


x 2 f X ( x)dx 
3 0
x (2 x  2)dy
1
1  2 x 4 2 x3  7
    
3 4 3 0 18

2  2y 1
2
E Y 2    y fY ( y)dy  0 y  6  dy
2
2
1  2 y 4 y 3  16
    
6 4 3 0 9
 x y 1
2 1 2 1
E  XY     xy      ( x y  xy )dxdy
2 2
0 0  3  300
1 2
1  x3 x2 2  1 1 y2  1  y 2 y3 
2 2
2
  y  y  dy    y  dy     
3 0 3 2 0 3 03 2  3 6 6 0 3
7 25 13
Var ( X )  E ( X 2 )  ( E ( X )) 2   
18 81 162
13
x 
162
16 121 23
Var (Y )  E (Y 2 )  ( E (Y )) 2   
9 81 81
23
y 
81
Cov( X , Y ) E ( XY )  E ( X ) E (Y ) 1 162
rxy   
 x y  x y 9 299
Regression lines Y on X
x
y  y  r. ( x  x)
y
11 1 162 13 9 5
y  (x  )
9 9 299 162 23 9
11 5
y   (0.04)( x  )
9 9
38
Regression lines X on Y
y
( x  x)  r ( y  y)
x
5 1 162 162 23 11
(x  )   (y  )
9 9 299 13 9 9
5 11
( x  )  (0.15)( y  ).
9 9
8. A joint probability mass function of the discrete R.Vs X and Y is given as
x y
 , x  1, 2, y  1, 2 , 3, 4
P(X = x , Y = y) =  32 . Compute the covariance of X and Y.
 0 , otherwise
Solution:
X
1 2 PY(y)
Y
1 2/32 3/32 5/32
2 3/32 4/32 7/32
3 4/32 5/32 9/32
4 5/32 6/32 11/32
PX(x) 14/32 18/32 1
The marginal distribution of X is The marginal

distribution of Y is
X 1 2
Y 1 2 3 4
(x) 14/32 18/32
P(y) 5/32 7/32 9/32 11/32
 14   18  25
E(X) =  xP ( x ) = 1  32   2  32   16
x
 5   7   9   11  45
E(Y) =  yP ( y ) = 1  32   2  32   3  32   4  32   16
y
E(XY)=  xyP ( x , y ) =
x y
 2   3   3   4   4   5   5   6  35
1   2    2    4    3    6    4    8   =
 32   32   32   32   32   32   32   32  8
35  25   45 
Cov(X , Y) = E(XY) – E(X)E(Y) =   0.0195 .
8  16   16 
39
9. If X and Y are independent random variables with PDF e  x , x  0 and e  y , y  0

X
respectively, find the density function of U  and V  X  Y .check whether
X Y
U and V are independent random variables. (April 2017)
Solution:
Given f X ( x)  e  x , x  0 and f Y ( y )  e  y , y  0
Also given that X and Y are independent.
the joint pdf f XY ( x, y)  f X ( x) . f Y ( y )  e  x . e  y  e  ( x  y ) , x  0, y  0
x
The transformation functions are u  and v  x  y
x y
x
Solving for x and y, we get u   x  uv
v
y = v - x  y = v – vu
x x
v u
The Jacobian of the transformation is J = u v   v(1  u )  uv  v
y y  v 1  u
u v
The joint density of U and V is fUV (u, v)  J f XY ( x, y)  v e ( x y )  v e v
The range space of (U , V) is obtained from the range space of (X , Y) and the
transformations x = uv, y = v(1 – u)
 x  0 and y  0 we have uv  0 and v(1  u )  0
 v  0 and u  0 and u  1
v e v , v  0 , 0  u  1
f UV (u, v)  
 0 , elsewhere ,
  
 e v e v 
f U (u )   f (u, v)dv   v e dv  v .
v
 1. 2 
 0 1  1 , 0  u  1
 0   1 ( 1 ) 0
 1
fV (v)  

f (u, v)du   v . e v du  v e v , v  0
0
Now f U (u ) . f V (v)  1. v e  v  f UV (u, v)

U and V are independent.
10. The joint probability density function of a two dimensional random variable (X , Y)
x2
is given by f ( x, y)  xy 2  , 0  x  2, 0  y  1 .
8
1
Compute (i) P(X > 1), (ii) P(Y < ),(iii) P(X < Y),(iv) Are X and Y independent?
2
Solution:
40
 1
 2 x2 
1
 y3 x2 
f X ( x)  

f ( x, y ) dy    xy   dy   x  y 
0
8   3 8 0
x x2 x
   8  3x  , 0  x  2

3 8 24
2

 2 x2 2
 x2 2 x3  1
f Y ( y)   f ( x, y) dx    xy   dx   y    2 y 2  , 0  y  1
 0 8 2 24 0 3
 2
 x x2   x 2 x3 
2
P  X  1   f X ( x) dx      dx    
1
1
3 8   6 24 1
1 1 3 7 19
  4  1   8  1   
6 24 6 24 24
1/2
 1
1/2
 2 1  2 y3 y 
1/2
P  Y     fY ( y ) dy  0  2 y   dy    
 2   3  3 3 0
2 1 1 1 3 1
 .  .  
3 8 3 2 12 4
y
 2 x2 
1 y 1
 x 2 y 2 x3 
P  X  Y      xy   dxdy      dy
0 0
8  0
2 24 0
1
1
 y 4 y3   y5 y 4  53
    dy     
0
2 24   10 96 0 480
 x x2   1  2 xy 2 x 2 y 2 x x 2
f X ( x ) fY ( y )      2 y 2        f ( x, y )
 3 8  3 3 4 9 24
 They are not independent
11. The joint probability mass function of (X,Y) is given by p( x, y)  k (2 x  3 y),
x  0,1, 2 ; y  1, 2,3. Find the Marginal distributions and Conditional probability
distributions. Also find the probability distribution of X +Y.
Solution:
The joint probability mass function table:
Y PX ( x)
X 1 2 3
0 3k 6k 9k 18k
1 5k 8k 11k 24k
2 7k 10k 13k 30k
PY ( y) 15k 24k 33k 72k
3 2
1
We know that   P  xi , y j   1  72k  1 k
72
j 1 i  0
41
MARGINAL DISTRIBUTION OF X:
3
18
P  X  0   P  X  0,Y  j   3k  6k  9k  18k  72
j 1
3
24
P  X  1   P  X  1,Y  j   5k  8k  11k  24k  72
j 1
3
30
P  X  2   P  X  2,Y  j   7k  10k  13k  30k  72
j 1
MARGINAL DISTRIBUTION OF Y:
2
15
P Y  1   P  X  i,Y  1  3k  5k  7k  15k 
i 0 72
2
24
P Y  2   P  X  i,Y  2  6k  8k  10k  24k 
i 0 72
2
33
P Y  3   P  X  i,Y  3  9k  11k  13k  33k 
i 0 72
CONDITIONAL DISTRIBUTION OF X GIVEN Y=1:
P  X  0,Y  1 3k 1
(i) If X  0, P  X  0 Y  1   
P Y  1 15k 5
P  X  1,Y  1 5k 1
(ii) X  1, P  X  1 Y  1   
P Y  1 15k 3
P  X  2,Y  1 7k 7
(iii) X  2, P  X  2 Y  1    .
P Y  1 15k 15
CONDITINAL DISTRIBUTION OF X GIVEN Y=2:
P  X  0,Y  2 6k 1
(i) X  0, P  X  0 Y  2    
P Y  2 24k 4
P  X  1,Y  2 8k 1
(ii) X  1, P  X  1 Y  2    
P Y  2 24k 3
P  X  2,Y  2 10k 5
(iii) X  2, P  X  2 Y  2   
P Y  2 24k 12
CONDITIONAL DISTRIBUTION OF X GIVEN Y=3:
P  X  0,Y  3 9k 3
(i) X  0, P  X  0 Y  3   
P Y  3 33k 11
P  X  1,Y  3 11k 1
(ii) X  1, P  X  1 Y  3   
P Y  3 33k 3
42
P  X  2,Y  3 13k 13
(iii) X  2, P  X  2 Y  3    .
P Y  3 33k 33
CONDITIONAL DISTRIBUTION OF Y GIVEN X=0:
P Y  1, X  0  3k 1
(i) Y  1, P Y  1 X  0    
P  X  0 18k 6
P Y  2, X  0  6k 1
(ii) Y  2, P Y  2 X  0    
P  X  0 18k 3
P Y  3, X  0  9k 1
(iii) Y  3, P Y  3 X  0     .
P  X  0 18k 2
P Y  1, X  1 5k 5
(i) Y  1, P Y  1 X  1   
P  X  1 24k 24
P Y  2, X  1 8k 1
(ii) Y  2, P Y  2 X  1   
P  X  1 24k 3
P Y  3, X  1 11k 11
(iii) Y  3, P Y  3 X  1    .
P  X  1 24k 24
P Y  1, X  2 7k 7
(i) Y  1, P Y  1 X  2   
P  X  2 30k 30
P Y  2, X  2 10k 1
(ii) Y  2, P Y  2 X  2   
P  X  2 30k 3
P Y  3, X  2 13k 13
(iii) Y  3, P Y  3 X  2    .
P  X  2 30k 30
PROBABILITY DISTRIBUTION OF  X  Y  :
3
(i) P  X  Y  1  P  X  0, Y  1  3k 
72
11
(ii) P  X  Y  2  P  X  0, Y  2  P  X  1, Y  1  6k  5k 
72
(iii) P  X  Y  3  P  X  0, Y  3  P  X  1, Y  2
24
 P  X  2, Y  1  9k  8k  7k  24k 
72
(iv) P  X  Y  4   P  X  1, Y  3
21
 P  X  2, Y  2  11k  10k  21k 
72
13
P  X  Y  5  P  X  2, Y  3  13k 
72
43
12. If X and Y are two random variables having joint probability density function
1
  6  x  y  , 0  x  2, 2  y  4
f  x, y    8
 0 otherwise
Find (i) P  X  1  Y  3 (ii) P  X  Y  3 (iii) P  X  1/ Y  3

Solution:
1 3
1
(i) P  X  1  Y  3    f  x, y  dydx     6  x  y dydx
1 3
0 2
0 2
8
3
1  y2  1  9 1 7  2x
1 1 1
1 3
  6 y  xy   dx   8  x  dx  
1
dx  7 x  x 2  
80 2 2 80 2 80 2 16 0 8
3 y
1  
3
3 y 1 x2
(ii) P  X  Y  3    
3
2 
0 8
6  x  y dydx 
8 2 
 6 x 
2
 xy  dy
0
3
1  (3  y)2  1  27 y2  1  27 y y 2 y3 
3 3
5
  6(3  y)   (3  y) y dy     6 y  dy   6   
82 2  82 2 2 8 2 2 6  2 24
P  X  1  Y  3
(iii ) P  X  1/ Y  3 
P  Y  3
2 3
1 5
P Y  3     6  x  y dydx 
0 2
8 8
3/8 3
P  X  1/ Y  3 
 .
5/8 5
13. Two random variables X and Y have the joint probability density function
f ( x, y)  kxye( x  y ) , x  0, y  0 . Find the value of k and also prove that X and Y are
2 2
independent. (April/May 2009,

2010, 2011)
Solution:
Given the joint pdf f ( x, y)  kxye( x  y ) , x  0, y  0
2 2
 
  f ( x, y) dx dy  1
 


 x2  y 2 
   kxye dx dy  1
0 0
 
 k  xe  x2
dx. ye y dy  1
2
0 0
44
Put x 2 = u, Put y 2 = v
 2xdx = du,  2ydy = dv
When x = 0,u = 0 When y = 0,v = 0
x = ,u   y = ,v = 
 
k 1
  e  u du .  e  v dv  1
20 20
 
k  eu   ev 
   .   1
4   1  0  1  0
k
  0  1 . 0  1 1
4
k 4
 the joint pdf is f ( x, y)  4 xye

 x2  y 2  , x  0, y  0
 

 x2  y 2  dy
f X ( x)  

f ( x, y)dy   4 xye
0

 4 xe x  ye
2
 y2
dy
0
2 1
 4 xe x    2 xe x , x  0
2
2
 

 x2  y 2  dy
fY ( y )  

f ( x, y)dx   4 xye
0

 4 ye y  xe
2
 x2
dy
0
2 1
 4 ye y    2 ye y , y  0
2
2
Now,

f X ( x). fY ( y )  2 xe  x . 2 ye  y
2
 2

 4 xye

 x2  y 2 
f X ( x). fY ( y )  f ( x, y )
 X and Y are independent
45
14. The joint pdf of random variables X and Y is given by

2  x  y, 0  x  1, 0  y  1
f ( x, y )   . Find Cov ( X , Y ) and the correlation
 0 , elsewhere
coefficient of X and Y. (April/May 2009, 2010, 2011)

Solution:
 1
1
 y2   1 3
f X ( x)   f ( x, y ) dy    2  x  y  dy   2 y  xy    2  x     x, 0  x  1
 0  2 0  2 2
 1
1
 x2   1  3
fY ( y )   f ( x, y ) dx    2  x  y  dx  2 x   yx   2   yx    y, 0  y  1
 0  2 0  2  2
 1
3
1

1
3   3 x 2 x3  3 1 5
E ( X )   xf X ( x) dx   x .   x  dx    x  x 2  dx   .     
 0 2  0
2   2 2 3  0 4 3 12
 1
3 
1 1
3   3 x3 x 4  1 1 1
E ( X )   x f X ( x) dx   x .   x  dx    x 2  x 3  dx   .     
2 2 2
 0 2  0
2   2 3 4 0 2 4 4
 1
1
3 
1
3 2  3 y 2 y3  3 1 5
E (Y )   Y
yf ( y ) dy  0  2  0  2
y .  y dy  y  y 

dx   .     
 2 2 3 0 4 3 12
 1
1
2 3 
1
3 2 3  3 y3 y 4  1 1 1
E (Y 2 )   Y  0  2  0  2
     .     
2
y f ( y ) dy y . y dy y y  dy
 2 3 4 0 2 4 4
2
1 5 1 25 11 11
Var ( X )  E ( X 2 )  ( E ( X ))2       X 
4  12  4 144 144 12
2
1 5 1 25 11 11
Var (Y )  E (Y )  ( E (Y ))      
2 2
 X 
4  12  4 144 144 12
  1 1 1
 1

E ( XY )    xyf ( x, y ) dx dy    xy (2  x  y )dx dy   y   (2 x  x 2  xy )dx  dy
  0 0 0 0 
1
 x 2 x3 x 2
1
 1
 1 y
1
2 y2 
  y 2   y  dy   y 1   dy    y   dy
0  2 3 2 0 0  3 2 0
3 2 
1
 2 y2 y 3  1 1  1
     
 3 2 6 0  3 6  6
1 5 5 1
Cov(X , Y) =  .  
6 12 12 144
1

 rXY  144   1  0.0909
11 11 11
.
12 12
46
15. Two dimensional random variable (X, Y) have the joint probability density
8 xy , 0  x  y  1
function f ( x, y )   .
 0 , elsewhere
 1 1
Find (i) P X   Y   (ii) the marginal and conditional distributions.
 2 4
(iii) Are X and Y independent? (April/May 2006, 2010, 2012)
Solution:
1 1
y
 x2 
y
 1 4 4
1
(i) P  X   Y      8 xy dx dy  8  y  dy
 2 4 0 0 0 
2 0
1
4 1
1
 4 y 3dy   y 4  4 
0
0 256
(ii) The marginal pdf of X is
1
 1
 y2  1 x2 
f X ( x)   f ( x, y) dy   8 xy dy  8 x    8 x     4 x(1  x 2 ), 0  x  1
 x  2 x 2 2 
y
 y
 x2   y2 
f Y ( y)   f ( x, y) dx   8 xy dx  8 y    8 y    4 y 3 , 0  y  1
 0  2 0 2
The conditional probability density function of X given Y is
f ( x, y) 8 xy 2 x
f ( x / y)    , 0  x  y , 0  y 1
f Y ( y) 4 y 3 y 2
The conditional probability density function of Y given X is
f ( x, y) 8 xy 2y
f ( y / x)    , x  y 1 , 0  x 1
f X ( x) 4 x(1  x ) (1  x 2 )
2
(iii) To check whether X and Y are independent.

f X ( x). fY ( y)  4x(1  x 2 ). 4 y 3  16xy3 (1  x 2 )  f ( x, y)
X and Y are not independent.
16. If the independent random variables X and Y have variances 36 and 16 respectively,
find the correlation coefficient between X + Y and X – Y. (May/June 2014)
Solution:
Given Var  X   36, Var Y   16
Given X and Y are independent random variable
 Cov( X , Y )  0
Let U  X  Y , V  X  Y
47
C ov U ,V 
ruv 
U V
C ov U ,V   C ov( X  Y , X  Y )
 C ov( X  Y , X )  C ov( X  Y , Y )
 C ov( X , X )  C ov(Y , X )  C ov( X , Y )  C ov(Y , Y )
 Var ( X )  Var (Y )
 36  16  20
 C ov U ,V   20
U2  Var U   Var  X  Y 
 Var  X   Var Y   36  16  52
U  52
V2  Var V   Var  X  Y 
 Var  X   Var Y   36  16  52
V  52
20
rUV   0.385
52 52
17. Obtain the equations of the lines of regression from the following data.
X 22 26 29 30 31 33 34 35
Y 20 20 21 29 27 24 27 31
Also estimate the value of Y when X = 38 and the value of X when Y = 18.
Solution:
X Y U = X – 30 V = Y – 27 U2 V2 UV
22 20 –8 –7 64 49 56
26 20 –4 –7 16 49 28
29 21 –1 –6 1 36 6
30 29 0 2 0 4 0
31 27 1 0 1 0 0
33 24 3 –3 9 9 –9
34 27 4 0 16 0 0
35 31 5 4 21 16 20
0 –17 132 163 101
n  8, U  0, V  17, U 2  132, V 2  163, UV  101 U     0 ,

U 0
n 8
V
V  17  2.125 ,  2  U  U 2  132  0  16.5 ,   4.062 ,
2
 
U U
n 8 n 8
48
 V2  
V2
 
163 2
 V  (2.125)2  15.86 ,  V  3.9825

n 8
Cov(U , V) =
UV  U V  101  0  12.625
n 8
Cov(U ,V ) 12.625
 rUV    0.7804
 U . v (4.062).(3.9825)
 rXY  0.7804
X  U  30  X  0  30  30
Y  V  27  Y  2.125  27  24.875
 X   U   X  4.062
 Y   V   Y  3.9825
The regression line of X on Y is
r X
X X  Y Y
Y
 
(0.7804).(4.062)
 X  30  Y  24.875
3.9825
 X  0.796Y  10.2
When Y = 18,
X = 0.796 (18) + 10.2 = 24.528
The regression line of Y on X is
r
Y Y  Y X  X
X
 
(0.7804).(3.9825)
 Y  24.875   X  30 
(4.062)
 Y  0.765 X  1.925
When X = 38,
Y = 0.765 (38) +1.925 = 30.995
18. In a partially destroyed laboratory record only the lines of regressions and variance
of X are available. The regression equations are 8x – 10y + 66 = 0 and 40x – 18y =
214 and variance of X = 9. Find (a) the correlation coefficient between X and Y
(b) Mean values of X and Y
(c) variance of Y.
Solution:
Since both the lines of regression passes through the mean values
49
8 x  10 y  66.............(1)
40 x  18 y  214............(2)

(1) * 5 40 x  50 y  330..........(3)
(2)  (3)  32 y  544  y  17
(1)  8 x  104  x  13
8 66
10 y  8 x  66  y  x 
10 10
8
 bYX 
10
18 214
40 x  18 y  214  x  y
40 40
18
 bXY 
40
8 18 9
 r 2  bYX bXY  
10 40 25
3
r  
5
Since both the regression coefficients are positive, r must be positive
 r  0.6
b  8*3
 y  YX x  4
r 10 * 0.6
Var ( y )  16
19. Assume that the random variables S is the sum of 48 independent experimental
1
 , 1 x  4
values of random variable X whose PDF is given by f x  x    3 . Find the
0, otherwise
Probability that S lies in the range 108,126 . (A/M 2019)
Solution:
 4  1  3
2
4 1
E  x   2.5, V  x  
2 12 4
E  S   48E  x  120,V  S   48V  x   36
50
108  120 126  120 

P 108  S  126  P  z 
 6 6
 P  2  z  1
 0.8125
20. A random sample of size 100 is taken from a population whose mean is 60 and
variance is 400. Using central limit theorem, with what probability can we assent
that the mean of the sample will not differ from   60 by more than 4.
Given that n  100 ,   60 ,  2  400

Since the probability statement is with respect to mean, we use the Linderberg-levy form
of central limit Theorem.
 2  2
X N   ,  i.e. X follows normal distribution with mean '  ' and variance .
 n  n
 400 
i.e. X N  60, 
 100 
X N  60, 4
 mean of the sample will not   X will not differ from 
P   P 
 differ from 60 by morethan 4    60 by morethan 4 

 P X  4 
 P  4  X    4 
 P  4  X  60  4 
 56  60 64  60 
 P 56  X  64  P  z
 2 2 
 P 2  Z  2
 2P 0  Z  2  2  0.4773  0.9446
UNIT – III RANDOM PROCESSES

PART-A
1. Define random process. Give an example. (N/D 2014) , (N/D 2013)
Solution:
A random process (Stochastic process) is a function X which assigns a real valued
function of time ' t ' to every outcome ' s ' of a random experiment. It is denoted by
X  s, t  or X  t  .
Example: X t   A cos t    , where A and  are constants and t is a random variable.
51
2. Write down the classification of random process. ( N/D 2010)

Solution:
Based on the nature of random variable X and the time t we can classify the random
process as follows
(i) Discrete random process ( X is discrete and t is continuous )
(ii)Continuous random process ( X is continuous and t is continuous )
(iii)Discrete random sequence ( X is discrete and t is discrete )
(iv)Continuous random sequence ( X is continuous and t is discrete)
3. Define first &second order stationary Process.
Solution:
I Order Stationary Process:
A random process is said to be first order stationary if its first order density function does
not change with a shift in time origin.
i.e., f X  x ; t   f X  x ; t    for any time t and any real number .
II Order Stationary Process:

A random process is said to be second order stationary if its second-order density
functions does not change with a shift in time origin.
i.e., f X  x1, x2 ; t1, t2  =f X  x1, x2 ; t1   , t2    for all t1 , t2 and .
4. Prove that a first order stationary process has a constant mean. (A/M 2011)
Solution:
Let X  t  be a first order stationary process.  f X  x, t   f X  x, t    .

We know that, E  X  t     x f  x, t  dx
X

 
Now, E  X  t       x f X  x, t    dx   x f  x, t  dx  E  X t 
X
 
 E  X  t   E  X  t    . Hence, E  X  t  is constant. i.e. Mean is constant
5. Define wide sense stationary process. (N/D 2017)(A/M 2017) (M/J 2013)
Solution:
A random process X  t  is said to be wide sense stationary (WSS) or weak sense
stationary process if the following conditions are satisfied
(i) E  X  t   constant .i.e. mean is a constant
52
(ii) E  X  t1  X  t2   RXX  t2  t1  or E  X  t  X  t     RXX  

i.e., autocorrelation function depends only on the time difference.
6. Define a strict sense stationary random process. (M/J 2016), (N/D 2015), (N/D 2012)
Solution:
A random process is called a strongly stationary process (SSS) or strict sense stationary if
all its statistical properties are invariant to a shift of time origin.
This means that X t and X t have the same statistics for any and any t.
Example: Bernoulli process is a SSS process.
7. Give an example of evolutionary random process. (A/M 2015)
Solution:
A random process that is not stationary in any sense is called an Evolutionary random
process.
Consider a random process X (t )  coswt    , where w is a real constant and  is a
uniform variable in 0 ,   2
.
Since  is uniformly distributed in 0 ,   2
 then f ( )   / 21  0  2 , 0  

2
 
2
2 2
2 
E  X (t )   X (t ) f ( )d   cos( wt   )d  sin(wt    0 2
 
0 0
2   2 2

 sin( wt  )  sin wt  0  cos wt  sin wt )  function of t
 2  
So, X(t) is not evolutionary random process
8. Define Markov Processes . (April/May 2019)
Solution:
A random process X(t) is said to be a Markov process if for every t1  t2  t3  ...  tn  t ,
P  X  tn   xn / X  tn 1   xn 1 , X  tn  2   xn  2 ,..., X  t1   X 1
 P  X  tn   xn / X  tn 1   xn 1
9. Define Markov chain. (Nov/Dec 2010)

Solution:
Let X(t) be a Markov process with states X(tr) = X r, t0  t1  t2  ...  tn .
If for all n , P  X n  an / X n1  an1, X n2  an2 ,..., X 0  a0   P X n  an / X n1  an1
then the sequence X n , n 0,1,... is called a Markov chain, where a0 , a1 ,..., an are
called the states of the Markov chain.
53
10. Define one step and n step transition probability.

Solution:
 
The conditional probability P X n1  a j / X n  ai is called one step transition probability
from the state ai to the state a j in one step. It is denoted by Pi j (n, n  1).

The conditional probability P X mn  a j / X m  ai  is called n - step transition
probability from the state ai to the state a j in n step. It is denoted by Pi (jn ) .
11. Define Homogeneous Markov chain and regular Markov chain.

Solution:
A Markov chain is said to be homogeneous in time if the one-step transition probabilities
are independent of the step. i.e. Pi j (n, n  1)  Pi j (m, m  1) for all m,n and all i,j.
A Markov chain is said to be regular if its one-step transition matrix P is regular. i.e. if for
some positive integer m , all the entries of Pm are positive.
12. What is (i) transition probability matrix? (ii)stochastic matrix
Solution:
When a Markov chain is homogeneous, the square matrix of one-step transition
probabilities Pij is called the transition probability matrix. It is denoted by P   Pi j  .
A square matrix P   Pi j  is called a stochastic matrix if each row of P satisfies
n
every Pi  0 and  P 1.
i 1
i
13. State the postulates of a Poisson process. (Nov/Dec 2010,2017)

Solution:
If X  t  represents the number of occurrences of a certain event in  0,t  then the discrete
random process X  t  is called the Poisson process, provided the following postulates
are satisfied
(i) P 1 occurrence in  t , t  t   t  o  t 
(ii) P no occurrence in  t , t  t   1  t  o  t 
(iii) P 2 or more occurrences in  t , t  t   o  t 
(iv) X  t  is independent of the number of occurrences of the event in any interval prior
and after the interval  0,t  .
(v) The probability that the event occurs in a specified number of times t0 , t0  t 
depends only on t , but not on t0 .
54
14. If the customers arrive at a bank according to a Poisson process with a mean rate of
2 per minute, find the probability that, during one-minute interval no customer
arrives. (April/May 2017)
Solution:
Here   2 , t 1.
e  t  t 
n
 P  X  t   n  , n  0,1, 2,...
n!
Probability during 1-min interval, no customer arrives  P  X t   0  e 2 .
15. Prove that sum of two independent Poisson processes is again a Poisson process
(Nov/Dec 2012)
Solution:
The moment generating function of the Poisson process is M X t   u   e
 
t eu 1
Let X1  t  and X 2  t  be two independent Poisson processes.
Their moment generating functions are M X1 t   u   e 1

  and
 t eu 1  
2t eu 1
M X 2 t   u   e
 
1t eu 1  
2t eu 1  1 2 t  eu 1
M X1 t  X 2 t   u   M X1 t  u  M X 2 t  u   e e e
By uniqueness of moment generating function, the process X1  t   X 2 t  is a Poisson
16. Prove that difference of two independent Poisson process is not a Poisson process.
Solution:
Let X  t   X1  t   X 2 t 
E  X t   E  X1 t   E  X 2 t    1  2  t
E  X 2  t   E  X12  t   E  X 22  t   2E  X 1 t  E  X 2 t 
  12t 2  1t    22t 2  2t   2  1t  2t 
  1  2  t   1  2  t 2
2
  1  2  t   1  2  t 2
2
Recall that E  X 2  t  for a Poisson process with parameter  is given

by E  X 2  t   t   2t 2
 X1 t   X 2 t  is not a Poisson process.
17. Define random telegraph process.

Solution:
A random telegraph process is a discrete random process X  t  satisfying the following
55
conditions
(i) X  t  assumes only one of the two possible values 1 or -1 at any time t.
1
(ii) X  0  1 or  1 with equal probability .
2
(iii) The number of flips , N   , from one value to another occurring in any interval of
length  is a Poisson process with rate  so that the probability of exactly r flips is
e    
r
P  N    r  , r  0,1, 2,...
r!
18. Write down the probability law, mean and variance of Poisson process.
Solution:
Let X(t) denote the number of times an event occur in the time interval (0,t) and >0 be
the rate of occurrence, then the probability of exactly n occurrences is
e  t  t 
n
P  X  t   n  , n  0,1, 2,...
n!
Mean   t , variance   t .
19. State any two properties of Poisson process. (A/M 2018), (N/D 2015)
Solution:
(i) The Poisson process is a Markov process.
(ii) Sum of two independent Poisson processes is a Poisson process.
0 1
20. Let A   1 
1  be a stochastic matrix. Check whether it is regular. (N/D 2016)

2 2
Solution:
Matrix A is said to be regular, if for some positive integer m, all the entries of Am are
positive.
1 1
0 1  0
1  
A 1
2   2 2
1  
1  1
 1 3
2 2 
2  2 
4 4
Since all the entries of A2 are positive then the given matrix is regular.
56
PART-B
1. Show that the process X  t  whose probability distribution under certain conditions
  at n 1
 , n  1, 2,...
 1  at n 1
is given by P  X  t   n   is not stationary. (N/D 2016)
 at
 1  at , n0

Solution:
The probability distribution is given by
X t n : 0 1 2 3 . . . .
2
at 1 at at
P X t n : 2 3 4
. . . .
1 at 1 at 1 at 1 at

E  X  t     nP  X  t   n 
n 0
 at 
2
1 at
 (1)  (2)  (3)  ...
1  at  1  at  1  at 
2 3 4
1   at   at 
2

 1  2    3   ...
1  at   1  at   1  at 
2
 
2
1 at
2
1
1 at 1 at
2
1 1 at at
2
1 at 1 at
2
1 1
2
1
1 at 1 at
E  X  t   1  Constant
 
E  X 2  t     n 2 P  X  t   n    (n 2  n  n) P  X  t   n 
n 0 n 0
 
=  (n 2  n) P  X  t   n    n P  X  t   n 
n 0 n 0
 at 
n 1

  n  n  1 1
1  at 
n 1
n 0
 (at )0 (at )1 (at )2 

 1.2  2.3  3.4  ...  1
 (1  at ) (1  at ) (1  at )
2 3 4

57
2   at   at 
2

  1  3  
 6   ...  1
1  at   1  at   1  at 
2
 
3
2  at 
 1  1  at  –1
1  at 
2
3
2  1 
 1  at   1
1  at 
2
 2 1  at  1
E  X 2  t   1  2at  Constant (it is a function of time t)
The given process X t is not stationary
2. Given a random variable Y with characteristic function  ( )  E[ei y ] and a random
process defined by X (t )  cos(t  y), show that X t is stationary in the wide
sense if  (1)   (2)  0.

Solution:
Given
 ( )  E cos Y  i sin Y   E cos Y ]  iE[sin Y 
X (t )  cos(t  Y ) &  (1)  0   (2)
We have to prove that X (t ) is a WSS process
E[ X (t )]  E[cos( t  Y )]  E[cos  t.cos Y  sin  t sin Y ]
 cos  tE[cos Y ]  sin  tE[sin Y ]
given  (1)  0
E[cos Y ]  iE[sin Y ]  0
E[cos Y ]  0 & E[sin Y ]  0
 E[ X (t )]  0
RXX (t , t   )  E[ X (t ) X (t   )]
 E[cos( t  Y ).cos( t    Y )]
1
 {E[cos(2 t    2Y )  cos( t    Y   t  Y )]}
2
58
1
 {E[cos(2 t    2Y )  cos( t )]}
2
1
 {E[cos(2 t   ).cos 2Y )  sin(2 t   )sin 2Y  cos( t )]}
2
1
 {[cos(2 t   ).E (cos 2Y )  sin(2 t   ) E (sin 2Y )]  cos( t )}
2
given  (2)  0  put   2
E[cos 2Y ]  iE[sin 2Y ]  0
 E[cos 2Y ]  0 & E[sin 2Y ]  0
1
RXX (t , t   )  cos 
2
The autocorrelation function is a function of time difference only. So, both the condition
are satisfied. Hence X (t ) is a WSS.
3. Show that the random process X  t   A sin t    is first order stationary, if A
and  are constants and  is uniformly distributed random variable in  0,2  .
Solution:
Since  is uniformly distributed in  0,2 
1
 f    , 0    2
2
E  X t   E  A sin t   
2
  A sin t    f   d
0
2
1
  A sin t    d
2 0
A 2
 cos t    
2 0
A
 cos t  2   cos t  
2
A
 cos t   cos t    0
2
 E  X t   0 (independent of t)
 
 X t  is First order stationary process.
59
4. Verify whether the sine wave process X t , where X t Y cos t , where Y is
uniformly distributed in  0,1 is a SSS process.
Solution:
If X t is a SSS process then all its finite dimensional distributuin should be invariant
with respect to the time to the time parameter ‘t’ . Therefore, we first consider the
distribution of X t
1
Since Y is uniformly distributed in 0,1 , f y , 0 y 1
1 0
1
EY E Y cos t y cos t f y dy
0
1
1
 y2  1 
  y cos  t dy  cos t    cos t   0 
0  2  0 2 
1
 cos  t funtion of t
2
X t is not a SSS process.
5. If the process  X  t  : t  0 is a Poisson process with parameter  , obtain
P  X  t   n . Is the process first order stationary? (May/June 2014)
Solution:
Let  be the number of occurrences of the event in unit time.
Let Pn  t  represent the probability of n occurrences of the event in the interval 0,t .
i.e., Pn  t   P  X t   n
 Pn t  t   P  X t  t   n
 Pn occurences in the time  0, t  t 
n occurences in the interval  0, t  and no occurences in  t , t  t  or

 P n  1occurences in the interval  0, t  and 1 occurences in  t , t  t  or

n  2 occurences in the interval  0, t  and 2 occurences in  t , t  t  or...
 Pn t 1  t   Pn1 t  t  0  ...
Pn  t  t   Pn  t 
   Pn1  t   Pn  t 
t
Taking the limits as t 0
60
Pn  t    Pn 1  t   Pn  t  -------- (1)

d
dt
d
Pn  t    Pn  t    Pn 1  t 
dt
This is a linear differential equation.
t
Pn  t  e    Pn1  t et dt ------------- (2)
t
0
Now taking n 1 we get
t
et P1  t     P0  t et dt ---------------- (3)
0
Now, we have,
P0 t t P 0 occurences in 0, t t
P 0 occurences in 0, t and 0 occurences in t, t t
P0 t 1 t
P0 t t P0 t P0 t t
P0 t t P0 t
P0 t
t
Taking limit t 0
Lt P0 t t P0 t
P0 t
t 0 t
dP0 t
P0 t
dt
dP0 t
dt
P0 t
log P0 t t c
t c
P0 t e
P0 t e t ec
P0 t e t A ----------------- (4)
Putting t 0 we get
P0 0 e0 A A
i.e., A 1
(4) we have
P0 t e t
substituting in (3) we get
t
t
e P1 t e t e t dt
0
61
dt t
0
t
P1 t e t
Similarly n 2 in (2) we have,
t
t
P2 t e P1 t e t dt
0
t
t
e te t dt
0
2 t2
2
t 2
t e t
P2 t e
2!
Proceeding similarly we have in general
e  t  t 
n
Pn  t   P  X  t   n  , n  0,1,...
n!
Thus the probability distribution of X t  is the Poisson distribution with parameter  t .
Mean  E  X  t    
n  0,1,2,...
nP  X  t   n 
 t 
n
 
n  0,1,2,...
ne  t
n!
 t
  t 1  t 
2

e  2  ....
 1! 2! 
  t   t 2
 t

 e t 1    ....  e  t t (et )
 1! 2! 
 t  constant
Therefore Poisson process is not stationary.
6. Show that the process X t   A cos t  B sin t (where A and B are random
variables) is a wide sense stationary if E  A   E  B   0, E A2  E B2     and
E  AB   0 .
Solution:
E  X t   E  A cos t  B sin t   cos t E  A   sin t E  B 
 0 (constant)
62
R t1 ,t2   E  X t1  X t2 
 E  A cos t1  B sin t1  A cos t2  B sin t2 
 E  A 2 cos t1 cos t2  AB cos t1 sin t2
 AB sin t1 cos t2  B 2 sin t1 sin t2 


 
 E A 2 cos t1 cos t2  E  AB  sin  t1  t2 
 
 E B 2 sin t1 sin t2
 
 E A2  cos t1 cos t2  sin t1 sin t2 
 E  A   E  B  , E  AB  0
2 2
 E  A  cos  t  t 
2
1 2
 k cos  t1  t2 
.
Thus, {X(t)} is a WSS process.
7. Show that the random process X  t   A cos  t    is WSS if A &  are constants
and  is uniformly distributed random variable in  0, 2  . (Nov/Dec 2017)
Solution:
Since  is uniformly distribution in  0,2  .
1
f    , 0    2
2
Ensemble average same,
2
1
E  X  t   A  cos t    d
0
2
2
A
 cost cos  sint sin  d
2 0

A
 cos  t sin   sin t cos  0
2

2
A
 sin t  sin t   0
2
E  X  t    0 .
63
RXX  t , t     E  X  t  X  t   
 E  A cos t    A cos t     
A2
 E cos  2t    2   cos  
2 
A2 A2 cos 
 E cos  2t    2   
2 2
2 2
A 1 A2 cos 
 
2 0 2
 cos 2t    2 d 
2
2
A2  sin  2t    2   A2 cos 
   
4  2 0 2
A2 cos 
RXX  t , t     =a function of 
2
 X t  is WSS.
8. If the process  X t  is given by X t   10cos 100t    where  is uniformly
distributed over   ,   check whether  X  t  is WSS or not. (May/June 2014)
Solution:
Given X (t )  10cos(100t   ),  ( ,  ) is uniformly distributed
1
 f ( )  ,     
2

1
E  X  t    2 10cos 100t    d


10
 cos100t cos  sin100t sin  d
2 

5 
  cos100t cos  sin100t sin   

5
   sin100t  sin100t   0

RXX (t , t   )  E  X (t ) X (t   ) 
 E 10cos(100t   )10cos(100t  100   )
100
 E  cos(200t  100  2 )  cos100 
2
 50E cos(200t  100  2 )]  E[cos100 
64

1
 50  cos(200t  100  2 ) d  50cos100

2

25  1 
  sin(200t  100  2 )   50cos100
 2  
25
 [sin(200t  100  2 )  sin(200t  100  2 )]  50 cos100
2
25
 [sin(200t  100 )  sin(200t  100 )]  50 cos100
2
 50cos100  RXX ( )
X (t ) is a WSS process.
Note: If the process  X t  is given by X t   A cos t    where  is uniformly

distributed over   ,   check whether  X  t  is WSS or not. (A/M 2018)

1
Proceed as above we get, E  X  t    2 Acos t    d  0 and

A2
RXX (t , t   )  E  X (t ) X (t   )  cos   RXX ( )
2
 X (t ) is a WSS process.
9. Two Random processes  X t  and Y t are defined by
X t   A cos t  B sin t and Y t   B cos t  A sin t . Show that X  t  and
Y  t  are jointly WSS if A and B are uncorrelated random variables with zero
means and the same variance and is constant.

SOLUTION:
Given E  A   E  B   0
Var  A   Var  B   k (say)
  2
 
Var  A   E A2  E  A   E A2  0  E A2  
Var  B   E  B   E  B   E B   0  E B 
2 2 2 2
 E  A   E  B   k (say)
2 2
65
Since A and B are uncorrelated, E  AB   0 .
Let us prove X  t  and Y  t  are individually WSS process.
To prove X  t  is WSS process, the following conditions are to be verified.
(i) E  X t   must be a constant.
(ii) R t1 ,t2  is a function of t1  t2  .

Now,
E X t E A cos t B sin t cos t E A sin t E B 0 0
E X t 0
R t1 ,t2 E X t1 X t2 E A cos t1 B sin t1 A cos t2 B sin t2
E A 2 cos t1 cos t2 AB cos t1 sin t2
AB sin t1 cos t2 B 2 sin t1 sin t2
E A2 cos t1 cos t2 0 0 E B2 sin t1 sin t2
k cos t1 t2
R t1 ,t2 is a function of t1 t2
Thus, it is a WSS process.
Similarly, we can prove that Y t is also a WSS process.
To prove X t and Y t are jointly WSS,
RXY t1 ,t2 E X t1 Y t2 E A cos t1 B sin t1 B cos t2 A sin t2
E AB cos t1 cos t2 A 2 cos t1 sin t2
B 2 sin t1 cos t2 AB sin t1 sin t2
0 k cos t1 sin t2 k sin t1 cos t2 0 k sin t1 t2
RXY t1 ,t2 is a function of t1 t2
Thus, X t and Y t are jointly WSS.
66
10. Define semi-random telegraph signal process and random telegraph signal process
and also prove that the former is evolutionary and the latter is wide sense stationary.
(Nov/Dec 2017)
Solution:
The random process X t ( 1) N (t ) , where N t is a Poisson process with rate  , is
called semi-random telegraph process which starts with initial value X 0 1.
1, N (t ) is even
Here, X  t    .
1, N (t ) is odd
By the above definition X  t  can take the values 1 and –1 only.
P  X  t   1  P  N  t  is even 
e  t  t 
r
 
r even r!
 t
  t 2  t 4 
 e 1    ...
 2! 4! 
 et cosht
P  X  t   1  P  N  t  is odd 
e  t  t 
r
 
r  odd r!
 t   t   t 
 1 3

e    ...
 1! 3! 
 e sinht
 t
Mean  X  t    r P  X  t   r 
r 1,1
 t
 1 e cosht  (1)  e t sinht
 et cosht  sinht 
 e t e t  cosht  sinht  e t 
 e2t  constant it isa function of time t 
Therefore, semi-random telegraph signal process is evolutionary (not stationary).
Definition: A random telegraph process is a discrete random process X  t  satisfying the
following conditions
(i) X  t  assumes only one of the two possible values 1 or  1 at any time t.
1
(ii) X  0  1 or  1 with equal probability .
2
(iii) The number of flips , N   , from one value to another occurring in any interval of
67
length  is a Poisson process with rate  so that the probability of exactly r flips is
e    
r
P  N    r  , r  0,1, 2,...
r!
Mean  E  X  t     X(t ) P  X  t  
X ( t ) 1,1
 (1) P  X  t   1  (1) P  X  t   1

1 1
 (1)  (1)
2 2
 0, a constant
RXX  t , t     E  X  t  X  t   
 1 P  X  t  X  t     1  1 P  X  t  X  t     1
     
n n
 e
n even
 t
n!
 e
n odd
 t
n!
 e cosh  e sinh
   
 e cosh  sinh 

 e e
RXX  , t     e2  function of 
Therefore, random telegraph process is wide sense stationary.
11. Prove that the random telegraph signal process y t    x t  is a wide sense
stationary process where  is a random variable which is independent of x  t  and

2 t1 t2
it assumes values  1 and 1 with equal probability and Rxx  t1 , t2   e .
(Apr2017)
Solution:
1
Given P   1  P   1 
2
1 1
Now, E    1.   1 .  0.
2 2
Also, E  2   1 .   1 .  1
2 1 2 1
2 2
E Y  t   E  X  t   E   E  X  t    , X  t  areindependent 
E Y  t   0, a constant  E    0
RYY  t1 , t2   E Y  t1  Y  t2 
 E  X  t1   X  t2 
68
 E  2 X  t1  X  t2 
 E  2  E  X  t1  X  t2 
 (1) E  X  t1  X  t2 
 Rxx  t1 , t2 
2  t1 t2
e , function of time difference.
Therefore the random telegraph signal process y t    x t  is a wide sense stationary.
12. A random process X  t  defined by X t   A cos t  B sin t for all t , where A & B are
independent binary random variables. Each of which assumes the values – 2&1
1 2
with probabilities & respectively, prove that  X  t  is WSS. (Apr/May 2018)
3 3
Solution:
Given A and B are independent binary random variables and each of which assumes the
1 2
values – 2&1 with probabilities & respectively.
3 3
Aor B : 1 2
2 1
P A or P( B) :
3 3
2 1
E  A   E  B   1  2   0
3 3
E  A2   E  A2   1    2       2
2 2 2 1 2 4 6
3 3 3 3 3
Since A & B are independent, then E AB E A E B 0 ---- (1)
Hence E  X  t   E  A cos t  B sin t 
 E  A cos t  E  B sin t
E  X  t   0  is a constant.  E  A  E  B   0
RXX  t , t     E  X  t  X  t   
 E  A cos t  B sin t   A cos  t     B sin  t    
 E  A2 cos t cos  t     AB cos t sin(t   )  BA sin t cos  t     B 2 sin t sin  t   
 E  A2  cos t cos  t     E  AB cos t sin  t     E[BA]sin t cos  t     E  B 2  sin t sin t   
 E  A2  cos t cos  t     E  B 2  sin t sin  t   
 2 cos t cos  t     sin t sin  t      E  A2   E  B 2   2 
 2cos  is a function of time difference.  X t  is WSS.
69
13. If  X1  t  and  X 2 t  represent two independent Poisson process with parameter
1t and 2t respectively, then prove that P  X1  t   x X1  t   X 2  t   n  is Binomial
1
with parameter n and p . Where P  . (Apr/May 2019)
1  2
Solution:
P  X 1  t   k  .P  X 2  t   n  k 
P  X 1  t   x X 1  t   X 2  t   n  
P  X 1  t   X 2  t   n 
P  X 1  t   k  .P  X 2  t   n  k 

P  X 1  t   X 2  t   n 
e  1t  1t  e  2t  2t 

k nk
.

k!  n  k !
e  1  2 t   1  2  t 
n
n!
 nCk p k q n  k .
14. A hard disk fails in a computer system and it follows a Poisson distribution with
mean rate of 1 per week. Find the probability that 2 weeks have elapsed since last
failure. If we have 5 extra hard disks and the next supply is not due in 10 weeks, find
the probability that the machine will not be out of order in next 10 weeks.
Solution:
Here the unit time is 1 week.
Mean failure rate = mean number of failures in a week
1
P No failures in the 2 weeks since last failure P X 2 0
2 0
e 2 2
e
0!
0.135
(ii) There are only 5 extra hard disk and the computer should not go out of order in the
next 10 weeks
P For this event P X 10 5
5 10
e 10n
=
n 0 n!
0.068
70
15. A machine goes out of order, whenever a component fails. The failure of this part
follows a Poisson process with a mean rate of 2 per week. Find the probability that 2
or 3 weeks have elapsed since last failure. If there are 4 spare parts of this
component in an inventory and that the next supply is not due in 5 weeks, find the
probability that the machine will not be out of order in the next 5 weeks.
Solution:
Mean failure rate = mean number of failures in a week
2
P No failures in the 2 or 3 weeks since last failure P X 2 X (3) 0
2 0 3 0
e 2 e 3
0! 0!
4 6 3
e e 2.48 10
(ii) There are only 4 spare parts and the machine should not go out of order in the next 5
weeks
P For this event P X 5 4
4 10
e 10n
= 0.029
n 0 n!
16. Prove that the inter arrival time of a Poisson process with parameter has an
1
exponential distribution with mean . (Apr/May 2019)
Proof:
Let two consecutive occurrences of the event be Ei and Ei 1.
Let Ei take place at time ti and T be the interval between the occurrences of Ei and
Ei 1 . The T is a continuous random variable.
P T t P Ei 1 did not occur in ti , ti t

P No event occurs in an interval of length t
P X t 0
t 0
e t
0!
t
e
The cdf of T is given by
71
F t P T t 1 P T t
1 e t
The pdf of T is given by
t d
f t 0 e F t f t
dt
t
e , t 0
1
Which is exponential distribution with mean .
17. The transition probability matrix of a Markov chain X t , n  1, 2,3,... having
 0.1 0.5 0.4 

 
three states 1,2,3 is P   0.6 0.2 0.2  , and the initial distribution
 0.3 0.4 0.3 
 
is P 0  0.7 0.2 0.1 , Find P  X 2  3 and P  X 3  2, X 2  3, X1  3, X 0  2
(Nov/Dec 2016), (Apr/May 2010)

Solution:
 0.43 0.31 0.26 
 
P  P.P   0.24 0.42 0.34 
2
 0.36 0.35 0.29 

 
3
P  X 2  3   P  X 2  3/ X 0  i  P  X 0  i 
i 1
 P  X 2  3 / X 0  1 P  X 0  1  P  X 2  3 / X 0  2  P  X 0  2 
 P  X 2  3 / X 0  3 P  X 0  3
 P132 P  X 0  1  P232 P  X 0  2  P332 P  X 0  3
 0.26  0.7  0.34  0.2  0.29  0.1  0.279
P  X 3  2, X 2  3, X1  3, X 0  2
 P  X 3  2 / X 2  3 P  X 2  3/ X1  3 P  X1  3/ X 0  2 P  X 0  2
  0.4 0.3 0.2 0.2  0.0048
18. Find the nature of the states of the Markov Chain with three states 0,1,2 and with
72
0 1 0
 
1 1
one step transition probability matrix P   0 . (Apr/May 2019)
2 2
0 1 0 

Solution:
0 1 0
 
1 1
P 0
2 2
0 1 0 

1 1
 0 1 0  0 1 0   0
   2 2
1 1 1 1 
P 
2
0  0 0 1 0
2 2  2 2  
 0 1 0  0 1 0   1 0 1 
  
2 2
1 1
 2 0 2  0 1 0   0 1 0 
  1  
1 1

1
P  P .P   0 1 0 
3 2  0  0 P
1  2 2 2 2
1 
 0   0 1 0   0 1 0 
 2 2 
P  P .P  P.P  P
4 3 2
P5  P4 .P  P2 .P  P3  P and so on
P 2 n  P 2 and P2n1  P for all n  1, 2,3,...
1 1
We note that P00    0, P01   1  0, P02    0
2 1 2
2 2
1 1
P101   0, P11 2  1  0, P12 3   0
2 2
1 1
P20 2   0, P211  1  0, P22 2   0
2 2
 n
 We find Pij  0 for all i, j  0,1, 2 and integers n  1, 2,3,...
 the Markov chain is irreducible.
Further since P2n  P2 we get Pii   Pii   Pii   ...  0 i and
2 4 6
g.c.d of 2, 4,6,...  2 for all i.

So, all the states of the chain are periodic with same period 2.
Since the chain is finite and irreducible all the states are persistent and non-null.
But the states are periodic.
Hence all the states are not ergodic.
73
19. A man either drives a car or catches a train to go to office each day. He never goes 2
days in a row by train but he drives one day, then the next day he is just as likely to
drive again as he is to travel by train. Now suppose that one the first day of the
week, the man tossed a fair dice and drove to work if and only if a 6 appeared. (1)
Find the probability that he takes a train on the third day and (2) the probability
that he drives to work in the long run. (Nov/Dec 2017)
Solution:
State Space = (train, car)
The TPM of the chain is
T C
PT 0 1
 
1 1
C 
2 2
1
P (traveling by car) = P (getting 6 in the toss of the die)=
6
5
& P (traveling by train) =
6
5 1
P (1)   , 
6 6
0 1
 5 1   1 11 
P (2)
 P P  ,  1 1 , 
(1)
 6 6    12 12 
2 2
0 1
 1 11    11 13 
P  3  P P ,  1 1 , 
 2
 12 12     24 24 
2 2
11
P (the man travels by train on the third day) =
24
Let   1,  2  be the limiting form of the state probability distribution or stationary state
distribution of the Markov chain.
By the property of  ,  P  
0 1
1  2   1 1   1  2 
2 2
1
 2  1
2
1
1   2   2
2
& 1   2  1
74
1 2
Solving  1  & 2 
3 3
2
P{The man travels by car in the long run }= .
3
20. Three boys A, B and C are throwing a ball to each other. A always throws the ball to
B and B always throws the ball to C, but C is just as likely to throw the ball to B as
to A. Show that the process is Markovian. Find the transition matrix and classify the
states.
Solution:
The tpm of the process Xn is given by
State of X n
A B C
A 0 1 0
P State of X n 1B 0 0 1
C 1/ 2 1/ 2 0
The states of X n depend only on state of X n 1, but not on states of X n 2, Xn 3 ,... or
earlier states. Xn is a Markov chain.
0 1 0 0 1 0 0 0 1
2
Now, P 0 0 1 0 0 1 1/ 2 1/ 2 0
1/ 2 1/ 2 0 1/ 2 1/ 2 0 0 1/ 2 1/ 2
0 0 1 0 1 0 1/ 2 1/ 2 0
3 2
P P P 1/ 2 1/ 2 0 0 0 1 0 1/ 2 1/ 2
0 1/ 2 1/ 2 1/ 2 1/ 2 0 1/ 4 1/ 4 1/ 2
1
Here P11 3 0, P13 2 0, P21 2 0, P22 2 0, P33 2 0 and all other Pij 0
the chain is irreducible.
0 1/ 2 1/ 2 1/ 4 1/ 4 1/ 2 1/ 4 1/ 2 1/ 4
4 5 6
P 1/ 4 1/ 4 1/ 2 , P 1/ 4 1/ 2 1/ 4 , P 1/ 4 3 / 8 1/ 2 and so on.
1/ 4 1/ 2 1/ 4 1/ 8 3 / 8 1/ 2 1/ 8 3 / 8 3 / 8
We note that Pii 2 , Pii 3 , Pii 5 , Pii 6 etc are 0 for i 2,3 and GCD of
2,3,5,6,... 1 .
the states 2 and 3 (i.e. B and C) are periodic with period 1. (i.e.) aperiodic.
We note that P11 3 , P11 5 , P11 6 etc are 0 and GCD of 3,5,6,... 1 .
the state 1 (i.e. state A) is periodic with period 1. (i.e. aperiodic)
Since the chain is finite and irreducible, all its states and non-null persistent.
Moreover all the states are ergodic.
75
UNIT IV CORRELATION AND SPECTRAL DENSITIES

PART-A
1. Define auto correlation function and write any two of its properties. (Apr/May 2019)
Definition: The auto-correlation of the random process  X  t  is defined by
RXX  t1 , t2   E  X  t1  X  t2  or RXX  t , t     E  X  t  X  t   
Property.1 For WSS process, the Mean Square value = Rxx  0 i.e. E  X 2  t   Rxx  0 
Property.2 For WSS process, Autocorrelation function is an even function i.e.
RXX     RXX   .
2. Define cross correlation function and state any three of its properties. (May 2014)
Definition: The cross-correlation of the two processes  X  t  and Y  t  is defined by
RXY  t1 , t2   E  X  t1  Y  t2  or RXY  t , t     E  X  t  Y  t   
Property.1 For Jointly WSS Process, RXY    RYX   
Property.2 For Jointly WSS Process, Rxy    Rxx  0  Ryy  0  .
Property.3 If Rxy    E  X  t  E Y  t  then X  t  and Y  t  are independent jointly
WSS Process.
3. A random process X  t  is defined by X  t   K cos  t , t  0 where  is a constant
and K is uniformly distributed over  0, 2 . Find the auto correlation function
of X  t  . (May/June 2013)
1
Solution: Since K is uniformly distributed over  0, 2 then f  K   , 0  K  2
2
RXX  t , t     E  X  t  X  t   
 E  K cos t.K cos   t   
 cos t.cos   t    E  K 2 
2
 cos t.cos   t     K 2 f  K  dK
0
2
1
 cos t.cos   t     K 2 dK
0
2
2
1  K3 
 cos t.cos   t     
2  3 0
76
1  23 
 cos t.cos   t      0
2 3 
4
RXX  t , t     cos t.cos   t   
3
4. Prove that auto correlation function is an even function of  . (N/D 2016) (N/D 2012)
Proof: We have to Prove RXX     RXX  
For a WSS process  X  t  , RXX    E  X  t  X  t   
 RXX     E  X  t  X  t    Put t1  t    t  t1  

 E  X  t1    X  t1 
 E  X  t1  X  t1     RXX  
5. Find the variance of the stationary process  X  t  whose autocorrelation function is

4
given by Rxx    25  . (Apr/May 2017, 2018)
1  6 2
Solution:
E  X  t 
Lt
R  
2

 
Lt  4 
  25    25
   1  6 2 
 E  X  t   5
E  X 2  t    RXX  0   25  4  29

Var  X  t    E  X 2  t   E  X t  
2
 29  25  4 .
6. Find the variance of the stationary process  X  t  whose auto correlation function
is given by Rxx    2  4e
2 
. (Nov/Dec 2017)
Solution:
E  X t   
Lt Lt
R   
2 2 
 2  4e 2
   
 E  X  t   2
E  X 2  t    RXX  0   2  4  6

Var  X  t    E  X 2 t   E  X t  
2
 62  4
77
7. Find the mean and variance of the stationary process  X  t  whose autocorrelation
25 2  36
function R    . (May /June 2015)
6.25 2  4
Solution:
36
25  2
25 2  36 
R    
6.25  4 6.25  4
2
2
E  X t 
Lt 25 2500
R   
2
  4
  6.25 625
E  X  t   2
36
E  X 2  t    RXX  0   9
4
 
Var  X  t   E  X 2  t   E  X t 
2
 94  5
8. Define power spectral density function. (Nov /Dec 2013)

Definition: Let X  t  be a stationary random process with autocorrelation function
RXX   . Then the power spectral density S XX   is defined as the Fourier transform of
RXX   .

i.e. S XX     R   e
XX
 i
d

9. State any two properties of the power spectral density function.

(Apr/May 2018)
Solution:
(i). S XX    S XX   

(ii). S XX  0    R   d
XX

10. Define cross power spectral density function

Definition: Let X  t  and Y  t  be two jointly stationary processes with cross correlation
function RXY   . Then the cross power spectral density S XY   is defined as the

Fourier transform of RXY   . i.e. S XY     R  e

XY
 i
d

78
11. State any two properties of the cross power spectral density function. (May 2017)
Solution:
(i) S XY     SYX  
(ii) If X  t  and Y  t  are orthogonal then S XY    0 and SYX    0
12. Prove that S XY     SYX   . (May /June 2012)

Proof:
 
We know that S XY     RXY  e  i d and S XY      R  e d
XY
 i
 
 
Now SYX     RYX  e  i d   R   e
XY
 i
d
 
Put   u d  du  d   du
when     u   and     u  

SYX      RXY  u eiu du

 R  u e du  SXY    .

 XY
i u

13. Find the autocorrelation function RXX   for the following power spectral density
functions
1
(i).
 4  i 
2
1
(ii).
25   2
Solution:
We know that RXX    F 1  S XX  
 1   1 
(i). RXX    F 1  2
 u   e 4 F 1  2
 u   e 
  4  i      i  
5   
 1  e  1  e
(ii). RXX    F 1   F 1  2 
 25    10      2
2 2
79
 
14. If R    e 2  is the autocorrelation function of a random process X  t  , obtain
the spectral density of  X  t  .

Solution:
  
S     R   e  i d  e  cos  isin  d  2 e2 cos d
2  
  0

4 a
S    e
 ax
cos bxdx 
4   2
2
0
a  b2
2
15. Define average power in a random process

Definition:

The average power in a random process  X  t  is defined as 
1
2  S   d 

XX
16. Find the mean-square value of the process  X  t  if its autocorrelation function is
 2
given by R    e 2
.
  
2
Solution: Mean-Square value  E  X 2

 t    RXX  0    e 2  1

  0
17. The Power spectral density of a random process  X  t  is given by
 ,   1
S XX      Find its autocorrelation function. (May/June 2016)
 0, elsewhere
Solution:
 1
1 1
RXX     XX     cos   i sin   d
2 1
S  e i
d  
2 
sin  
1
 cos   i sin  d   cos  d  

1 1
1
 
2 1 0    0
sin
Rxx    .

18. Find the auto correlation function of a stationary process whose power spectral
 2 for   1
density function is given by   
S   .
 0 for   1
Solution:
 1 1
1 1 1
R     S  e d  
i
   cos   i sin  d    2 cos  d
2
2  2 1 0
80
1
1   sin     cos     sin   
  2    2    2 
     
2
  
3
0
1  sin  2cos 2sin  
R      3 .
   2  
19. Check whether S XX     2 2  6

is a valid power spectral density function
8 4  3 2  4
Solution:
2     6
2
2 2  6
S XX       S XX  
8     3     4 8 4  3 2  4
4 2
Therefore the given function is a valid power spectral density function.

1
20. Given the power spectral density: S xx    , find the average power of the
4  2
process.
Solution:
 
1   i 1  1
R     S  e  i d  
e d    4  

2 
2 2
Hence the average power of the processes is given by

 
1 d 1 d
E  X  t   R  0  
2
 4   2  2 2
2 2 0 2   2

1 1    1   1
  tan 1       0  .
 2  2 0   4  4
PART-B
1. If the power spectral density of a WSS process is given by
b
 a   ;   a
S     a find its autocorrelation function of the process.
0 ;  a

(Nov/Dec 2017)
Solution:

1
R     S   e
i
d
2 
a
 a a    e
1 b
 i
d
2 a
81
 a  a     cos   i sin   d
1 b

2 a
a
2 b
 a    cos  d
2 0 a

sin  cos  
a
b 
  a  
a    2  0
b  cos a 1 
   2
 a   2  
b b  a 
2 
 1  cos a   2sin 2  
 a  a 2
 2 
2. The random processes  X  t  and Y  t  defined by
X  t   A cos  t  B sin  t , Y  t   B cos  t  A sin  t , where A & B are uncorrelated
zero mean random variables with same variance find its autocorrelation function.
Solution:
 
Given E  A  E  B   0 & E A2  E B 2   2  
Also A and B are uncorrelated i.e., E  AB   0
RXY    E  X  t  Y  t   

 E  Acost  B sint   B cos  t     Asin  t     
 ABcostcos  t     A2 costsin  t    
E 2 
  B sintcos  t     BAsintsin  t    
 E  AB  costcos  t     E  A2  costsin  t   
 E  B 2  sintcos  t     E  BA  sintsin  t   
 E  B2  sintcos  t     E  A2  costsin  t     E  AB   0
  2  sintcos  t     costsin  t      E  A2   E  B 2    2 
  2 sin t  t  
  2 sin   
RXY     2 sint  sin     sin 
3. Find the auto correlation function of the periodic time function {X(t)} = Asin
Solution:
Given X(t)= Asin t
X(t) X(t+ ) = Asin t Asin (t+ )
=A2 sin t sin (t+ )
82
= A2 sin( t+ ) sin t
=
The time autocorrelation function
Rxx(t,t+ ) =
=
=
= .
4. 24
The power spectrum of WSS process {X(t)} is given by Sxx(  )= . Find the
  16
2
mean square value of the process by Brute Force method. (Apr/May 2019)
Solution:
Autocorrelation Function

1
Rxx     S xx   ei d 
2 
Mean Square Value of the Process

1
E  X  t    Rxx  0  
2
 S xx   d
2 

1 24
E  X  t    Rxx  0  
2
 d
2 
2
 16

12 1

 

2
 42
d

3   
  tan 1   
  4   
3
  tan 1     tan 1    

3   
   
 2 2
E  X  t   Rxx  0   3
2
5. The power spectral density function of a WSS process  X  t  is given by
 S ;   0

S     0 . Find R   (Apr/May 2019)

0 ; elsewhere
Solution:

1
R     S   e
i
d
2 
83
0
1
 S e
i
 d
2
0
 0
0
S0

2 0
  cos   i sin   d
0
S0

  cos  d
0

S0  sin 0  0

    0
S sin 0
R    0 .

 
6. Consider two random processes X  t   3cos  t    and Y  t   2cos   t    
 2
where  is a random variable uniformly distributed in  , 2  . Prove
that RXY    RXX  0  RYY  0  . (Apr/May 2018)
Solution:
Given X t   3cos t   
 
Y  t   2 cos  t      2sin t   
 2
RXX    E  X  t  X  t   
 E 3cos t    3cos t     
9
 E cos  2t    2   cos   
2 
1
 is uniformly distributed in  0, 2  , f   
2
9 9
 E cos  2t    2    cos 
2 2
2
9 1 9
  cos  2t    2  d  cos 
20 2 2
2
9 1  sin  2t    2   9
    cos 
2 2  2 0 2
2
9  sin  2t    2  
 cos t   0
2  2 0
84
9
RXX    cos 
2
9
RXX  0 
2
RYY    E Y  t  Y  t   
 E 2sin t    2sin t     
 E 2sin t    2sin t     
 4E sin t    sin t     
4
 E cos    cos  2t    2  
2 
 2E cos    2E cos  2t    2 
2
1
 2cos   2  cos  2t    2  d
0
2
2
1  sin  2t    2  
 2 cos    
 2 0
 2 cos 
RYY  0  3
9
RXX  0  RYY  0  
.2  3
2
Similarly, RXY    3sin 
RXY    3sin   3
 RXY    RXX  0  RYY  0 
7. Two random process  X  t  and Y  t  are given by
X  t   Acos  t    , Y  t   A sin  t    where A &  are constants and  is a
uniform random variable over 0 to 2  . Find the cross-correlation function.

(M/J 2013).
Solution:
RXY    E  X  t  Y  t   
 E  Acos t    Asin   t      
 A2 E  sin t      cos t   
 sin  2t    2   sin   
 A2 E  
 2 
85
A2
 E sin  2t    2   sin  
2 
A2 A2
 E sin  2t    2   E sin  
2 2 
A2 A2
 E sin  2t    2   sin 
2 2
1
 is a uniform random variables f    , 0    2
2
2
A2 A2
 RXY     
4 0
sin 2t    2 d  sin 
2
2
A2  cos  2t    2   A2 A2
     sin   sin  .
4  2 0 2 2
8. Find the Cross-correlation function of W  t   A  t   B  t  & Z  t   A  t   B  t  ,
where A  t  and B  t  are statistically independent random variables with zero
means and autocorrelation function RAA   e  ,       ,
RBB   3e  ,       respectively.

Solution:
Given E  A  t   0, E  B  t   0
At  & B t  are independent
RAB    E  A  t  B  t    
 E  A  t  E  B  t     0
Similarly
RBA    E  B  t  A  t     E  B  t  E  A  t     0
 RWZ    E W  t  Z  t   
 
 E  A  t   B  t    A  t     B  t    
 E  A  t  A  t     A  t  B  t     B t  A t     B t  B t   
 E  A  t  A  t     E  A  t  B  t     E  B  t  A t     E  B t  B t   
 RAA    RAB    RBA    RBB  
 RAA    RBB  
 
 e  3e
RWZ    2e

86
9. If the cross power spectral density of X  t  and Y  t  is
 jb
a  ,     
S XY      where ' a '&' b ' are real constants. Find the cross
 0 ,elsewhere
correlation. (May/June 2017, Nov /Dec 2017)
Solution:

1
RXY     S   e d 
 j
2
XY


1  jb  j

2   a 

 
e d
 
a  e j jb  e j e j 
   2 2
2  j 
  2  j
 j   
a  e j e j  jb   e j e j  e j e j 
 
 2  j   2  j   2 
2  j   
a sin   b  e  e j  j
 j
jb  e  e  j

     2  
   2    2 
a sin   b cos   b sin  
  
   2
1  b 
 2  a   sin    b cos   
   
Note: If the cross power spectral density of X  t  and Y  t  is
a  jb ,     
S XY     where ' a '&' b ' are real constants. Find the cross
0 ,elsewhere
correlation. (A/M 2018)
1
Proceed as above with   1 we get RXY     a  b  sin    b cos   
 2 
10. If  X  t  is a random process with mean 3 and autocorrelation R    9  4e 0.2  .
xx
Determine the mean, variance and covariance of the random variables Y  X  5
and Z  X  8 .
Solution:
Given Y  X  5 & Z  X 8 , E  X  t   3 --------- (1)
87
Mean of Y  E Y   E  X  5  3
Mean of Z  E  Z   E  X 8  3
We know that
Var Y   E Y 2    E Y  
2
E Y 2   E  X 2  5  
But E  X 2  t   RXX  0 
 9  4e0.2101
 9  4  13
Thus Var Y   13   3  13  9  4
2
Var  Z   E  Z 2    E  Z  
2
E  Z 2   E  X 2 8  Z  X 8
 RXX  0
 9  4  13
Hence Var  Z   13   32   13  9  4
E YZ   R  5,8  9  4e
0.2 58
 R  t1 , t2   9  4e
0.2 t1 t2

0.6
 9  4e
Covariance  R  t1 , t2   E  X  t1  E  X  t2 
 R 5,8  E 5 E 8
 9  4e0.6  3 3  4e0.6  2.195
11. The autocorrelation function of random binary transmission  X  t  is given by
 
1  ,  T
R     T , where T is a constant. Find the power spectrum of the
0
 , elsewhere
process  X  t  . (Nov /Dec 2017)
Solution:

S     R  e
 i
d

T
 
 1  T e d
 i

T
T
  
  1  T  cos  isin  d
T
88
T
  
  1  T  cos d
T
  
T
 2 1   cos d
0
T
 
T
 2 1   cos d
0
T
T
   sin  1  cos  
 2  1      
 T       0
2
 cosT 1 
 2  2 
 T  T
2
2
 2 1  cos T 
T
2  T 
 2 2 sin2  
T  2 
 T 
4sin2  
S     2 
.
 2T
12. Find the power spectral density function of a WSS process with autocorrelation
function R    e  .
2
(Nov/Dec 2014)
Solution:

S XX     R   e
XX
 i
d


e
 2
 e  i d



  2  i  d
 e

  i 
  2 
 
 e


d
 i  i   i  
2 2
   2      
   2   2  
 e


d
 i  i    2 
2
   2      2 

   2   4 
 e


d
 i 2  2 
      2
 2  4 
 e

d
89
 i 
2
2     
 
 S XX    e 4
e  2 
d

i  i 
2
 
Now let,      u i.e. u    
2

 2   2 
 du   d and as,   ,u   and as   ,u  
2 
 du
 S XX    e 4
e
u 2
 
2
 
4
e
.2 e u du e u is even
2

2
 0
 2

2e 4
 x
 Standard Result :  e x dx 
2
 2 0
2
 2

 S XX    e 4

13. A random process  X  t  is given by X  t   A cos  t  B sin  t , where A & B are
independent random variables such that E  A  E  B   0 and
E  A2   E  B 2    2 . Find the power spectral density of the process. (Dec 2014)
Solution:
Given X t   A cos t  B sin t ,
E  A  E  B   0 , E  AB   0 , E  A2   E  B2    2
E  X  t   cos t E  A  sin t E  B 
E  X  t   0  is a constant. E  A  E  B   0
R  t , t     E  X  t  X  t   
 E  A cos t  B sin t   A cos   t     B sin   t   
 E  A2 cos  t cos   t     B 2 sin  t sin   t     AB sin t cos   t     AB cos  t sin   t    
 E  A2  cos  t cos   t     E  B 2  sin  t sin   t   
 E  AB  sin t cos   t     cos  t sin   t    
  2 cos t cos  t      2 sin t sin  t    0
  2 cos   a function of time difference.
 X t  is WSS.
90
14. If the cross-correlation of two processes {X(t)} & {Y(t)} is

AB
R xy ( t , t  )  [sin( 0 )  cos( 0 (2t  ))], where A, B and  0 are constant. Find
2
the cross power spectrum.
Solution:
We know that the time average is given by
T
PXX  limT  R
T
XY ( t,t   )d 
T
1 AB
 limT 
2T 
T
2
[sin(0 )  cos 0 (2t  ) dt
AB
 sin(0 )  0
2
[ Max value of sinθ is1]
The cross power spectrum is given by
AB
SXY ()  Fourier transform of sin(0 )
2

AB
 sin(0 )e i d 

2
iAB

[2(  0 )  2(  0 )]
4
Where is the dirac delta function such that
iAB
Hence = [(  0 )  (  0 )] .
2
15. If {X(t)} is a WSS process with autocorrelation R XX ( ) and if
Y (t )  X (t  a )  X (t  a). Show that RYY ( )  2 RXX ( )  RXX (  2a )  RXX (  2a ) .

Solution:
Ryy    E  y  t  y  t   

 E  X  t  a   X  t  a    X  t    a   X  t    a  
 E  X  t  a  X  t    a   E  X  t  a  X t    a 
 E  X  t  a  X  t    a    E  X  t  a  X t    a 
 Rxx    E  X  t  a  X  t  a    2a  
 E  X  t  a  X  t  a    2a    Rxx  
 2Rxx    Rxx   2a   Rxx   2a  .
91
16. Given that a process {(X(t)} has the autocorrelation function
RXX ( )  Ae  | | cos( w0 ) , where A>0,   0 and w 0 are real constants, find the
Power spectrum of {(X(t)}. (April/May 2018)

Solution:

By definition S     R   e
 i
d


 Ae cos 0  e i d

 



Ae cos 0     cos  isin  d
 

 
Ae cos0 cos  A  e cos0   sin  d
   
 

 2A  e cos0 cos d
0

 cos 0     cos   0  
 2 A  e   d
  2 

a
Using  e ax cosbxdx  2 2
0
a b
   
S    A  2  2 2
   0        0  
2
 1 1 
 A  2  2
   0        0  
2 2
17. The autocorrelation function of the Poisson increment process is given by

 2 for  

R     2     . Prove that its spectral density is given by
   1   for  
  
t
4 sin 2
S     2 2     2 . (Nov/Dec 2011,2013)
 
2 2
92
Solution:
 2 for     or  

Given the autocorrelation function R     2    
    1    for    
  

The spectral density is given by S     R   e

 i
d , by definition

  
  R   ei d   R   ei d   R   ei d
  

 2      i
 
     e d
2  i
 e d      1   e d  2  i
       
   
   i
 e d    2ei d    2ei d  
2  i
 1   e d
   
  

  
 
 e  
2  i  i
 d  1   e d
 
 
 
 F  2   1    cos  i sin   d
 

Where F   2
 is the Fourier transform of  2
   
  
   

S    F   2    1   cos  d  i  1   sin  d  .....(1)

     
 

     
But 1   cos is an even function of  and 1   sin  is an odd function of 
   

   
   
 
  1   sin d  0 and  1   cosd  2 1   cosd    0;    
 
  
 0


  
S    F   2   2 1   cosd
 0  

2    sin    1   cos  
 F    2
1        [by Bernoulli’s formula]
          2   0

2  1    
 F   
1
1   sin   cos 
2

      2
0
2  cos   cos 0 
 F  2    
    2   2 
2 1
 F  2   1  cos   
  2
93
2 
 F  2   .2sin 2
  2 2
2
  
S    F   2   2 2 sin 2
4
....... 1
  2
To find the value of F   2  , we shall find the inverse Fourier transform of S   ,
R    F 1  S   

1
 S   e
i
 d
2 
Consider S    2 2   , where    the unit impulse function is.


1
R     2    e
2 i 
d
2 

     e
i 
2
d


  .1 2
    t    t  dt    0 


 2  e
i 
   d  e0  1 as     ei 

Thus   R  
2
Taking Fourier transform
F   2   F  R     S    2 2  
4   
Substituting in (1) we get S    2 2    sin 2  .
 
2 2
 2 
18. Show that the power spectrum of the autocorrelation function e  || [1   |  |] is
4 3
.
( 2  w 2 ) 2
Solution:

By definition, S     R  e
 i
d


e
 
 1     e  i d

0 
 e
 i 
1    d   e i  1    d
 0
0 
 e i  e i    e i  e i  
 1        2
 1      
 2
   i   i        i    i   0
94
 1    1  
     0   2
   i   i      i   i  
2
 1 1   1 1 
      2
   i   i     i    i  
2
   i 2    i 2 
 2
2
 
    
       
2 2 2 2
 
2 2    
2 2
 2 
   2  2   2 2
2  2   2   2  2   2 

  2 
2 2
2  2   2   2   2 

  2 
2 2
4 3

 2 
2 2
19. Given the power spectral density of a continuous process

2  9
S XX     4 find the mean square value of the process.
  5 2  4
Solution:
We know that mean square value of  X  t 

 E  X 2  t  
1
2  S   d 

XX

1 2  9

2   4  5 2  4 d

1 2  9
 .2 4 d
2 0   5 2  4

1 2  9
 0  4   2  4 2  4
 d

1 2  9
 0  2  2  1  4  2  1
 d

2  9
E  X 2  t  
1
 0  2  4  2  1
i.e., d
95
let 2  u
2  9 u 9
 We have 
  4  1  u  4u  1
2 2
4  9 1  9
 4  1  1  4  
5 8
 ….Partial fractions
u4 u 1 3  u  4  3  u  1
2  9 5 8
i.e.,  
  4  1 3   4 3  2  1
2 2 2
From (1),
1  5 8 

E  X 2  t  
1
 0 3   2  4   2  1 
  d
 

 1 1 
1 
 5. 2 tan 2  8 tan  
1

3 0
1  5    
     8   0
3  2  2   2  
1  5  1  11 
 .    8   
3 2  2  6  2 
E  X 2  t  
11
12 .
20. A stationary random process X  t  with mean 2 has the autocorrelation function
 1
RXX    4  e 10 . Find the mean and variance of Y   X  t  dt .
0
(May/June 2012)
Solution:
Given RXX    9  2e

Mean of X  t  is given by
Lt
X  E  X  t    R  
2 2
   xx
 
Lt 
 9  2e
 
2
X 9
X 3
Also E  X 2  t   RXX  0   9  2e  9  2  11
0
96
Var  X  t   E  X 2  t     E  X  t   
2
 11  32  11  9  2
Mean of Y  t   E Y  t 
2 
 E   X  t  dt 
0 
2
  E  X  t  dt
0
2
  3dt
0
 3  t 0  6
2
 E Y  t   6
.
UNIT V LINEAR SYSTEMS WITH RANDOM INPUTS
PART-A
1. Define a system. When is it called a linear system? (M/J 2014)
Definition:
Mathematically, a system is a functional relation between input x  t  and output y t  .
Symbolically, y  t   f  x  t  ,   t  . .
The system is said to be linear if for any two inputs x1  t  and x2  t  the output
f a1 x1  t   a2 x2  t   a1 f  x1  t   a2 f  x2 t  where a1 , a2 are constants
2. Give an example of a linear system.

Solution:
Consider the system y  t   f  x  t   t x  t  .
For any two inputs x1  t  , x2  t  the outputs are tx1  t  and tx2 t 
Now f a1 x1  t   a2 x2  t   t a1 x1  t   a2 x2  t 
 a1tx1  t   a2tx2 t   a1 f  x1 t   a2 f  x2 t 
the system is linear.
3. Define time-invariant system. (Apr/May 2010)
Definition:
The system Y (t )  f X (t ) is said to be time invariant if Y (t   )  f  X (t   ) for
any    , 
97
4. Check whether the system y  t   tx  t  is time invariant .

Solution:
Given y  t   tx  t   f  x  t 
Now f  x  t      t    x t     y t    . Therefore the system is time invariant.
5. Check whether the the system y  t   x  t   x  t  1 is time invariant.

Solution:
Given y  t   x  t   x  t  1  f  x t 
Now f  x  t     x  t     x  t    1  y t    for any  . Therefore the system is
time invariant.
6. Define Linear time invariant system. (Nov/Dec 2013) (May/June 2013)
Definition: A Linear system Y (t )  f X (t ) which satisfies Y (t  )  f  X (t  ) ,
   ,  , is called Linear time invariant system

7. Define Causal System (Nov /Dec 2015)
Definition: The system Y (t )  f X (t ) is said to be causal if the value of the output
y t  at t  t1 depends only on the past values of the input x t  , t  t1 .
8. Define Memoryless System and give an example.

Definition: The system Y (t )  f X (t ) is said to be memoryless if Y  t1   f  X  t1  .
i.e the output y(t ) at t  t1 depends only on x(t1 ) and not on any other values.
Example: y t   1  x t  cos t 
9. When a system is said to be stable? (N/D 2016)

Solution: The system Y (t )  f X (t ) is said to be stable if its output to any bounded
input is bounded. i.e. if x  t    then y  t    , where  and  are constants.
10. What is unit impulse response of a system? Why is it called so?

(Nov/Dec 2017)

Solution: If a system is of the form Y  t    h u  X t  u  du then the system weighting

function h  t  is also called unit impulse response of the system. It is called so because
Y t   h t  and X t   unit impulse function  t .
98
11. State the convolution form of the output of a Linear time invariant system.
(May/June 2013)
Solution: If X  t  is the input and h  t  is the system impulse function then the output

Y  t  can be expressed as the convolution form, Y t   h t   X t    h  u  X  t  u  du


12. Define transfer function of a system. (Nov/Dec 2015)

Definition: The Fourier transform of system impulse function h  t  is called system

transfer function H   i.e. H    F  h  t     h t  e

 it
H   is called the
2
dt .

power transfer function of the system.

13. Find the system transfer function if a Linear Time Invariant system has a impulse
1
 , t c
function h  t    2c . (Apr/May 2019)
 0, t  c

Solution:
 c c
1 2
H     h  t  e  i  t dt  c 2c  cos t  i sin t  dt  2c 0 cos t dt sin t is an odd

function
1 sin c 
 sin t 0 
c
c c
14. Prove that Y  t   2 X  t  is linear.
Proof:
Given y  t   2 x  t   f  x t 
Now, f a1 x1  t   a2 x2  t   2 a1 x1  t   a2 x2 t   2a1x1 t   2a2 x2 t 
 a1  2 x1  t    a2  2 x2  t    a1 f  x1  t   a2 f  x2  t 
Therefore the system is linear.

15. Prove that the system Y  t  
 h  u X  t  u du is a linear time-invariant system.

(Apr/May 2017)
Proof: Let X (t )  a1 X 1 (t )  a2 X 2 (t )

Now, Y (t )   h(u) X (t  u) du

99

Y (t )   h(u) a X (t  u)  a X

1 1 2 2 (t  u ) du
 
 a1  h(u) X (t  u) du  a  h(u) X

1 2

2 (t  u ) du  a1Y1 (t )  a2Y2 (t )
Hence the system is linear.

Replacing t by t   , we have
 
Y (t   )   h(u) X (t    u) du   h(u) X (t  u) du  Y t 
 
X (t   )  X t 
Hence the system is linear time invariant system.

16. State any two properties of Linear time-invariant system. (Nov/Dec 2011)
Solution:
(i) If the input of linear time invariant system is a WSS process then the output will also
be WSS process
(ii) In a linear time invariant system the input X  t  and the output Y  t  are related by a

convolution integral Y  t    h u  X t  u  du .

17. 
If  X  t  and Y  t  in the system Y  t    h  u X  t  u du are WSS process

how are their auto correlation function related. (Nov/Dec 2011)

Solution:
We know that RXY    RXX   * h   and RYY    RXY   * h   
 RYY    RXX   * h   * h   
18. Prove that the mean output of a linear time invariant system given by
 y  H  0 .x , where X  t  is WSS. (Apr/May 2018)
Solution:
The mean output of
 
Y  t   E Y  t    E   h(u ) X (t  u ) du 
  
 
  h(u ) x du  x  h(u ) du X  t  is WSS then mean is constant
 

 x H  0 since by replacing   0 in H     h(t )e
it
dt .

100
19. Prove that Y  t   X 3  t  is linear. (Nov/Dec 2017)

Solution: Given y  t   x3  t   f  x t 
Now, f  a1 x1  t   a2 x2  t     a1 x1  t   a2 x2  t    a1 x1  t   a2 x2  t 
3 3 3 3
 a1 f  x1  t   a2 f  x2  t 
Therefore the system is linear.
20. Define Band-Limited white noise. (Nov/Dec 2017)
Definition:
Noise with non-zero and constant density over a finite frequency band is called band-
 N0
,  W
limit white noise i.e., S NN     2 .

0 , otherwise
PART-B
1. If the input  x  t  of a linear time invariant system is a WSS process, show that the
output  y  t  is also a WSS process. (Nov/Dec 2016) (May/June 2013, 2014)
Solution:
Let X  t  be a WSS process for a linear time variant stable system with Y  t  as the
output process.

Then Y  t    h  u  X  t  u  du where h  t  is weighting function or unit impulse


response.

 E Y  t     E h  u  X  t  u   du


  h  u  E  X  t  u  du

Since X  t  is a WSS process, E  X  t   is a constant  X for any t.

 E  X  t  u    X

 E Y  t     h u   X du


 X  h  u  du


Since the system is stable ,  h  u  du is finite


101
 E Y  t   is a constant.
Now RYY  t , t     E Y  t  Y  t   
 

 E   h  u1  X  t  u1  du1  h  u2  X t    u2  du2 
   
  

 E    h  u1  h  u2  X  t  u1  X  t    u2  du1u2 
   
 
   h  u  h  u  E  X t  u  X t    u  du u
 
1 2 1 2 1 2
Since X  t  is a WSS process, auto correlation function is a function of time difference.

 
 RYY  t , t       h  u  h  u R   u
1 2 XX 1  u2  du1 du2
 
When this double integral is evaluated by integrating w.r. to u1 , u2 , the R.H.S is only a
function of  .
 RYY t, t    is only a function of time difference  .
Hence Y  t  is a WSS process.
2. If  X  t  is the input voltage to a circuit and  y  t  is the output voltage. If
 x  t  is a stationary random process with  X  0 and RXX    e

 
. Find
Y , SYY   & RYY  

R
if the power transfer function is H   
R  iL
(Nov /Dec 2017)
Solution:

We know that Y  t    h  u  X  t  u  du


E Y  t     h  u  E  X  t  u  du

E Y  t   0

S XX     R   e
XX
 i
d


S XX    e
 
e  i d


S XX    e  cos   i sin   d
 

102

S XX    2 e cos  d
0
2
S XX   
  2
2
SYY    S XX   H  
2
2 R2

 2   2 R 2  L2 2
2 R 2 A B
By partial fractions  2  2
    R  L      R  L2 2
2 2 2 2 2 2
2 R2  A  R 2  L2 2   B  2   2 
2 R 2
Put  2   2  A 
R 2  L2 2
R2 2 R 2
Put  2   2  B 
L R2
  2
2
L
 
 R2 
 R2  2  2 R 2 
2  2 2 2     
  2 
SYY     R L   L 
 2  2 R 2  L2 2
2 2
R R
2   2  
 L . 1  L . 1
R
2
 
2 2 2
R R
2
    2       2
2
L L L

1 1
 . 2  .
  2
R
2
  
2
L
RYY    F 1  SYY  
 
 ei  ei
RYY      2   2 d  2  R d
2 2
  
 2
L
By contour integration technique we know that
eiaz
2 2
dz e ab , a 0
z b b
103
2 2
R R
      R  
 RYY     L
e
 
   2 e L
L
2
R R
   2  
2
L L .
1, 0  t T

3. A circuit has unit impulse response given by h (t )  T . Evaluate
0 , elsewhere

SYY () in terms of S XX () .
Solution:
2
We know that S ()  S () H ()
YY XX
 T
1
H ()   h(t )e d 
 j 
T e
 j 
Now d
 0
T
1 e  j 
   j 
T  0
1 e  jT  1 

T   j 

1
 1  e  jT 
T j
1
 1  (cos T  j sin T )
T j
1
H ()  (1  cos T )  j sin T )
T j
2 1
 H ()  (1  cos T )2  sin2 T 
T  2 2
1
 1  cos2 T  2 cos T  sin2 T 
T  2 2
1
 2  2 cos T 
T 2
2
2
 1  cos T 
T 2
2
104
2 T 
 2 sin2  
T  
2 2
 2 
4 T 
 sin2  
T2 2
 2 
4 T 
 SYY ()  sin2   .S ()
T  2 2
 2  XX
4. Assume a random process X(t) is given as input to a system with transfer function
H ( )  1 for 0    0 . If the autocorrelation function of the input process is
N0
  t  , find the Autocorrelation function of the process.
2
Solution:
N0
Given RXX ( )   ( ) (1)
2
and H ()  1; 0    0 .
To find RYY ( ) (i.e.) to find the auto correlation of the output process.

We know that R ( )  1 S ()e i d
YY  YY
2 
(2)
Taking Fourier transform on both sides of (1), we get,

N 
F [RXX ( )]  F  0  ( )
 2 
N0 N
S XX ( )  F [ ( )]  0 (1) ( F [ ( )]  1)
2 2
N0
 S XX () 
2
2
lllly W.K.T. SYY ()  SXX () H ()
N0 N
 (1)  0
2 2

 (2)  RYY ( )  1 S ()e i d
2 
YY
0
1 N 0 i

2 
 2
e d
0
105
0
N
 0
4 
 e i d
0

N 0 e i  0

4  i  
0
N 0  e i0  e i0 

4  i 

N0  2i sin 0 
  
4  i 
N0
RYY ( )  sin 0
2
5. A system has an impulse response h( t )  2e 7 t , t  0. The auto correlation function of
the process is R XX (  )  e 4|| . Find the power spectral density of the output Y  t  .
Solution:
Given X(t) is a WSS process which is the input to a linear system and so the output
process Y(t) is also a WSS process (by property autocorrelation function)
Further the spectral relationship is SYY    H   S XX  
2
Where S XX   = Fourier transform of RXX  


 R  e
 i
 XX d


e
4 
 e i d

0 
 e e d  e e d
4  i 4  i

 0
0 
  4i 
 e d   e  4i  d
 0

  4 i 
 e   4i  
0
e 
   
 4  i    4  i  0
1 1
 e0  e    e   e0 
4  i 4  i
1 1
 
4  i 4  i
4  i  4  i 2i
S XX    
 4  i  4  i  16   2
106
H   = Fourier transform of h  t 

 h  t e
 i t
 dt


  2e7t .eit dt  h  t   0 if t  0
0

 2 e 7i t dt
0

 e 7i t 
 2 
   7  i   0
2 2
 e   e0  
7  i 7  i
2 2
 H    
7  i 49   2
4
 H   
2
49   2
Substituting in (1) we get the power spectral density of Y(t),
4 2i 8i
SYY    . 
49   16  
2 2
 49   16   2 
2
6. If Y  t   A cos 0 t     N  t  , where A is a constant,  is a random variable with
uniform distribution in   ,   and N  t  is a band-limited Gaussian white noise
 N0
, for   0   B
with a power spectral density S NN      2 . Find the power

 0, elsewhere
spectral density of Y  t  . Assume that N  t  and  are independent. (May 2014)
Solution:
Given Y t   A cos 0t     N t 
N  t  is a band-limited Gaussian white noise process with power spectral density
N0
S NN    ,   0  B ie. 0  B    0  B
2

Required SYY     R   e
YY
 i
d

Now RYY    E Y  t  Y  t   

 
 E  A cos 0t     N  t    A cos 0t  0     N  t    
107
 A2 cos 0t    cos 0t  0     N  t  N  t     A cos 0t    N  t    

 E 
 A cos 0t  0    N  t  
 A2 E cos 0t    .cos 0t  0     E  N  t  N  t   
 AE cos 0t    E  N  t   
 AE cos 0t  0    E  N  t    and N  t  are independent]

A2
2
 
E cos  20t  0t  2   cos 0   RNN  
 A E cos 0t    E  N  t     A E cos 0t  0  2   E  N  t   

A2 A2
 E cos  20t  0t  2   cos 0  RNN  
2 2
 A E cos 0t    E  N  t     A E cos 0t  0     E  N  t   
Since  is uniformly distributed in   ,   the pdf of  is f   
1
,     
2

E cos 0t      cos  t    f   d
0


1
  cos  t.cos  sin  t.sin   2 d

0 0
 
1  
 cos 0t  cos  d   sin 0t  sin  d  
2    
1 cos 0t sin    sin 0t   cos   

2    
1
  cos 0t  0  sin 0t 1  1   0
2

1
Similarly E cos  20t  0  2    cos  2 t     2  2 d
0 0


1  sin  20t  0  2  
  
2  2  
1  sin  20t  0  2  sin  20t  0  2  
   
2  2 2 
1  sin  2  20t  0  sin  2   20t  0   
   
2  2 2 
1  sin  20t  0  sin  20t  0  
   0
2  2 2 
108

1
Similarly E cos 0t  0      cos  t       2 d
0 0

1 
 sin 0t  0       0
2 
A2
RYY    cos 0  RNN  
2

 A2 
 SYY      cos 0  RNN   ei d 
  
2
 
A2
 cos   . e  R   e
 i
 0 d  NN
 i 
d
2  
 A2

2
            S  
0 0 NN
 A2
           
N0
 SYY    0 0 , 0  B    0  B
2 2
7. A system has an impulse response h  t   e   t U  t  , find the power spectral density of
the output Y  t  corresponding to the input X  t  . (Apr /May 2018)
Solution:
Given X  t  is the input process and Y  t  is output process of a linear system
We know that SYY    S XX   H  
2
Where H    Fourier transform of the function h  t 

0 if t  0
Unit step function U  t   
 1 if t  0
 0 if t  0
 h t    t
e if t  0
 
H     h t  e
 it
dt   e  t eit dt
 0


 e  i t 
  i t 1
 e dt    
      i  t  0   i
1 1
H    
  i  2  2
1
SYY    S  
   2 XX
2
109
8. A random process X  t  is the input to a linear system whose impulse response is
h  t   2e t , t  0 . If the autocorrelation function of the process is RXX    e

2 
,
find the power spectral density of the output process Y  t  . (Apr /May 2018, 2019)
Solution:
Given X  t  is the input process to the linear system with impulse response
h t   2et , t  0
So the transfer function of the linear system is its Fourier transform

H     h t  e
 it
dt


 2e
 t  it
 e dt  2e  t , t  0 


 2 e1i t dt
0

1i t
 e 
 2 
  1  i   0
2 2
  0  1 
1  i 1  i
Given RXX    e 2 
 the spectral density of the input is


S XX     R  e
XX
 i
dt


e
2 
 e i d


 e  cos   i sin   d
 2



4
 2 e2 cos  d 
0
4  2
We know the power spectral density of the output process Y  t  is given by
SYY    H   S XX  
2
2
2 4

1  i 4   2
4 4 16
 
1    4   1    4   2 
2 2 2
110

9. If {X (t )} is a WSS process and if Y (t ) 
 h(u ) X(t  u ) du

then prove that the
following: (i) RXY ( )  RXX ( )* h( ) (ii) RYY ( )  RXY ( )* h( ) , where *
denotes the convolution, (iii) SXY ( )  SXX ( ) H ( ),

2
(iv) SYY ( )  SXX ( ) H ( ) . (Apr /May 2017, 2018)
Proof:
(i) RXY ( )  E[X (t ) Y (t   )]
 

 E X (t )  h(u ) X (t    u )du 
  
 
 E   h(u ) X (t ) X (t    u )du 
  

  h(u) E X (t ) X (t    u) du


  h(u) R

XX
(  u )du
RXY ( )  RXX ( )* h( )
(ii) W.K.T. RYY ( )  E[Y (t ) Y (t   )]
 
 E   h(u ) X (t  u )Y (t   )du 
  

  E X (t  u)Y (t   ) h(u)du


  E X (t  u)Y (t  u  u   ) h(u)du


 R

XY
(  u ) h(u )du
Put u  , du  d If u  ,  

u  ,   
111

  R

XY
(   ) h( )(d )

 R

XY
(   ) h( )d

  h( ) R

XY
(   ) d
 RYY ( )  RXY ( )* h( )
(iii) Since RXY ( )  RXX ( )* h( ) ( by property (i))

Talking Fourier transform on both sides,
F [RXY ( )]  F [RXX ( )* h( )]
 F [RXX ( )] F [ h( )]
SXY ()  SXX () F[ h( )]
 SXX () H (), where H ()  F[h( )]
(iv) Since RYY ( )  RXY ( )* h( )

Taking Fourier transform on both sides,
F [RYY ( )]  F [RXY ( )* h( )]
 F [RXY ( )] F [h( )]
 SYY ()  SXY () H (), where H ()  F[h( )]

SYY ( )  S XX () H * () H ( ) ( by (iii ))
 SYY ()  S XX () H () ,2

where H () is the conjugate of H * ()
10. If X  t  is the input voltage to a circuit and Y  t  is the output voltage,  X  t  is a
stationary random process with  x  0 and RXX    e2  . Find  y , S XX  
1
and SYY   , if the system function is given by H    .
  22
2
Solution:
Given Mean  X  t     X  0
112

Y t    h   X  t    d 


E Y  t     h   E  X  t    d  0


S XX     R   eXX
 i
d

e
2 
 e  i d


e
2 
 cos d


 2 e2 cos d
0
 2  4
S XX    2  2 
   4    4
2
1
H    2
 4
 
1  4
SYY    H   S XX    
2
  4
   2  4
2 2
 
4
SYY   
 
3
2  4
11. A random process X ( t ) having the auto correlation function RXX ( )  pe  | | , where
p and  are real positive constants, is applied to the input of the system with impulse
 e   t ,t0
response h( t )   , where  is a positive constant. Find the auto
0 ,t0
correlation function of the networks response Y (t ) .
Solution:
 e t ,t 0
Given h(t )   and RXX ( )  pe | | ,
0 ,t0
First we find S XX   :

S XX     RXX   ei d

113

  pe
 
ei d


  pe

 
 cos  i sin d
 
  pe cos d  i  pe
   
sin  d
 

 2 pe cos d  i  0
 
0

 2 p  e cos d
0
2 p

 2  2
And also find H   :

H     h  t  eit dt


   et eit dt
0
   i t
  e dt
0

 e  i t 
 
     i   0


  i
Since SYY    H   S XX  
2
 2 p
2

  i  2   2
 2   2 p 
 2  2 2 
       
2
2 p 2

 2
 2   2
 2 
 RYY    F 1  SYY    
1 
 S XX   ei d 
2 
114
1  2 p 2

2  
 2
 2   2   2 
ei d 
2 p 2  1

2  
 2
 2
  2
 2

ei d 
2 p 2   1 1 1 1 

2    2   2
   2
 2 

 2
2   2
2 
 ei d 

 
2 2  1  1 1   1  
  ei d  

2  2  
 2   2   2 
ei d 
2    2   2
   
   
2 p 2  1  1   1 
 2 2 
F  2 2 
 F 1  2 
       
2

2 p 2  1  1   1 
  F  2 2 
 F 1  2 2 
2  2         
2 p 2  1    1   
  2 e  e 
2  2 2
p 2  1   1    
 e  e 
 2   2    
12. A linear system is described by the impulse response h(t )  1 e  t / RC u( t ). Assume

RC
an input process whose Autocorrelation function is B( ) .Find the mean and
Autocorrelation function of the output process.
Solution:
Let X  t  be the input process, whose autocorrelation function is given as
RXX    B ( )
1  t / RC
The system impulse response is h  t   e u t 
RC
 1  t / RC
 e u  t  , if t  0
  RC
0, if t  0
115
Let Y  t  be the output process.


Now H     h  t  eit dt

 1  t / RC  it
 e e dt
0 RC
 1 
1   RC i t
RC 0
 e dt

  1 i t 
1  e  RC  
 
  
RC   1 
   RC  i  
 0
1

1  i RC
The mean E Y  t   H  0   X
But H  0   1 and  X  lim RXX    lim B    0

2
   
 E Y  t   0
We know that SYY    H   S XX  

2

S XX     R  e
XX
 i
dt

 
  B   e i dt  B    e  i dt  B 1
 
1
H   
1   2 R2C 2
1
 SYY    B
1   R2C 22
 B 
RYY    F 1  2 2 2 
1  R C 
116
 
 B 
 F 1  
 2 2 1 2 
 R C  R2C 2    
  
 
B  B 
 2 2 F 1  
R C  1 2 
  R2C 2    
 
B  B  1
 F 1
  , where  
2 2
RC      
2 2
RC
 
1
   
  e
1
1 e RC RC  RC 
We know  F 1  2    e
   2
2
1 2
2
RC
1 1
B RC  RC  B  RC 
RYY    2 2 e  e
RC 2 2 RC
13. If {X (t )} is a band limited process such that S ( )  0,    , prove that
XX
2[RXX (0)  RXX ( )]   2 2RXX (0) .
Solution:

W.K.T. R ( )  1
XX
2 S

XX
()e i d

1

2 S

XX
()(cos   i sin  )d
 
1 i

2  SXX ()cos  d  2 S

XX
()sin  d

1

2 S

XX
()cos  d  0 ( S XX () is even)

1
RXX ( ) 
2 S

XX
()cos  d (1)
( SXX ()  0,    )
Now

1
RXX (0) 
2 S

XX
() d (2)
117

(2) – (1)  1
RXX (0)  RXX ( ) 
2  (1  cos  )S

XX
() d

1    S () d
  2 sin
2
  (3)
2 
 2  XX
We know that sin   
 sin2    2
2      
2
 sin   
 2   2 
2     2 2
(i.e.) sin   (4)
 2  4
Using (4) in (3), we get,

1   2 2 
RXX (0)  RXX ( )   S XX ( )   d
   4 

1  ( )2 2 
  XX
 S ( )   d (       
 4 
         )

 2 2

4 S

XX
() d
 2 2
 2 RXX (0) ( Eqn.(2))
4
 2 2
(i.e.) RXX (0)  RXX ( )  R (0)
2 XX
 2 RXX (0)  RXX ( )   2 2RXX (0)
.
14. If X  t  is a wide sense stationary process that is the input to a linear system with
1
the transfer function H    where   0 . If X  t  is a zero mean whit noise
  j
N
with power spectral density 0 , determine the following
2
(i). The impulse response h  t  of the system
(ii). The cross power spectral density S XY   of the input process X  t  and the
output process Y  t  .
(iii). The cross-correlation function RYX   of Y  t  and.
(iv). The power spectral density SYY   of the output process.
(April/May 2019)
Solution:
118
N0
Rxx      
2
N0
S xx   
2
(i). The impulse response h  t  of the system

h t    H   e
jt
d


1 1
 
2  a  j
e jt d 
 e at , a  0, t  0 .
N0 1
(ii). The Cross Power Spectral density S xy    H   S xx 
2 a  j
(iii). The Cross Correlation function is the inverse Fourier Transform Sxy  
N 0 a
Rxy    e ,  0
2
N 0 a
The Cross Correlation function Ryx    Rxy     e ,  0
2
(iv). The Power spectral density S yy    H   S xx  

2
N0 1
S yy    .
2 a  2
2
15. Consider a system with transfer function 1

. An input signal with
1  j
autocorrelation function m    m2 is fed as input to the system. Find the mean
and mean-square value of the output. (May /June 2012) (April /May 2011)
Solution:
and RXX    m    m2
1
Given, H   
1  j
S XX    m  2 m2  
We know that,
SYY    H   S XX  
2
119
2
1
  m  2 m2   
1  j
1
  m  2 m 2   
1 2 
RYY   is the Fourier inverse transform of SYY   .
m 
So, RYY    e  m2
2
lim
RXX    X
2
We know that
 
2
So X  m 2
X m
Also H  0  1
We know that Y  1, m  m
m
Mean-square value of the output  Y  RYY  0  
2
 m2
2 .
16. A stationary random process X  t  having the autocorrelation function
RXX    A   is applied to a linear system at time t  0 where    represent the

impulse function. The linear system has the impulse response of
h  t   e  bt u  t  where u  t  represents the unit step function. Find RYY   . Also find
the mean and variance of Y  t  . (May /June 2012) (April /May 2011)
Solution:
Given RXX    A   and
1 , t  0 ebt , t  0
h  t   ebt u  t   ebt  
0 , t  0  0 , t  0
The autocorrelation of the output Y  t  of the system
is RYY    F 1  SYY ( ) , where SYY ()  H ( ) S XX ()
2

Now, H     h t  e
 it
dt


  e  bt e  it dt
0
 
 e  (b i )t  1
e  ( b  i ) t
dt    (b i )   b i
0  0
120
2
1
2
 1  1
H ( )     2
2
b  i
 b  
2 2 b  2

Now, S XX     R   e
XX
 i
d


 A   e
 i
 d


 A  e  i    d  A  e0   A .


since        d    0 

A
 SYY ( ) 
b  2 2
 A   1 
RYY    F 1  2 2
 AF 1  2
b     b   
2
a 
 1  e
we know that F 1  2 2 

 a    2a
b 
e
RYY   
2b
17. A wide sense stationary noise process N(t) has an auto correction function
RNN ( )  Pe 3| | where P is a constant. Find its power spectrum.
Solution:
Given the autocorrelation function of the noise process N  t  is RNN    Pe3 .

S NN     R   e
XX
 i
d


 Pe
3 
 e  i d

 0 3 i 

 P   e e d   e3 ei d 
 0 
 0 

 P   e3i  d   e3i  d 
 0 

 e
 3  i  

0
 e  
 3  i 

 P    
  3  i      3  i   0

 1
 0  1
1
 P 1  0  
 3  i 3  i 
121
 1 1 
 P  
 3  i 3  i 

 3  i  3  i   6P
 P 
  3  i  3  i  
 9
2

18. If X(t) is the input voltage to a circuit and Y(t) is the output voltage.{X(t)} is a
stationary random process with  x  0 and RXX    e . Find the mean  y and
2 
power spectrum SYY   of the output if the system transfer function is given by
1
H ( )  .
  2i
Solution:
Given Mean  X  t     X  0

Y t    h  u  X  t  u  du


E Y  t     h  u  E  X  t  uu  du  0


S XX     R   eXX
 i
d


e
2 
 e  i d


 e  cos   i sin   d
 2



 2 e2 cos  d
0
 2  4
S XX    2  2   2
  4    4
1
H    2
 4
 1  4
SYY    H   S XX     2
2
 2
 4  4
4
SYY   
 
2
2  4
19. A wide sense stationary random process X(t) with autocorrelation
122
RXX    Ae
a 
where A and a are real positive constants, is applied to the input of
an Linear transmission input system with impulse response h  t   e  bt u  t  where b
is a real positive constant. Find the auto correlation of the output Y  t  of the system.
(A/M 2010)
Solution:
Given
RXX    Ae , a  0 and
a 
1 , t  0 ebt , t  0
h t   e u t   e
 bt  bt
 
0 , t  0  0 , t  0
The autocorrelation of the output Y  t  of the system
is RYY    F 1  SYY ( ) , where SYY ()  H ( ) S XX ()
2

Now, H     h t  e
 it
dt


  e  bt e  it dt
0
 
 e  (b i )t  1
e  ( b  i ) t
dt    (b i )   b i
0  0
2
1
2
 1  1
H ( )    2
2
b  i 
 b   b 
2 2 2

Now, S XX     R   e
XX
 i
d


 Ae
a 
 e  i d


 A e
a 
e  i d


 A e
a 
 cos   i sin   d


2 Aa
 2 A e a cos  d 
0
a  2
2
 1  2 Aa  2 Aa
 SYY ( )   2 2  2 2 
 2
 b    a     a   2  b2   2 
123
 2 Aa 
RYY    F 1  2 
  a    b    
2 2 2
1 A B
Let 2  2  2 .
 a   b     a    b   2 
2 2 2 2
1 1
By partial fraction A  and B  2 .
b a 2 2
b  a2
 2 Aa  1 1 
RYY    F 1  2 2  2
 2 2 
 b  a  a   b   
2
2 Aa  1  1  1  1 
  F  a 2   2   F  b2   2 
b2  a 2     
a 
 1  e
we know that F 1  2 2 

 a    2a
2 Aa  e e    a  a eb  
a  b 
A A
RYY    2 2 
  2 e   be  a   a e b  
 b b  a 
2  
b  a  2a 2b  b  a 2  b 2
20. Consider the white Gaussian noise of zero mean, power spectral density N 0 applied
2
to a low pass RC filter, where transfer function is H (f )  1 . Find the

1  2 jfRC
output spectral density and auto correlation function of the output.
Solution:
The transfer function is H (f )  1
1  2 jfRC
1
H ()  (   2 f )
1  jRC
N
Spectral density function at the input of filter is N 0 . (i.e.) S XX ( )  0 .
2 2
2
N0 1
We know that SYY ()  SXX () H () 2 
2 1  jRC
2
N0 1  1 
   
2 1  j 1  RC 

2
 
N0 1
 | zz  z 2
2 2
1 2

124
N0   2 
SYY ( )   
2   2  2 

1 N0   2  j
 RYY ( ) 
2  2   2   2

 e d

 2N 0 
 e j 

4 

 2
    2 

d
 2N 0      
 e j   a  

4 
e 



 a 2   2  d  a e 
  
N0
RYY ( )  e 
4 .
125

Maths Part B PDF

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

Maths Part B PDF

Hochgeladen von

Copyright:

Verfügbare Formate

MA8451-Probability and Random Processes Dept.

MA8451 - PROBABILITY AND RANDOM PROCESSES

 1.2 p  2.3qp  3.4q 2 p  ...   p  2qp  3q 2 p  ...

20. If X is a normal variate with mean  30 and S.D = 5.Find P  26  X  40

Solution: X follows N(30, 5)   30 &   5

 P 0  Z  0.8  0  Z  2  0.2881  0.4772  0.7653 .

Given P  A  0.02 , P  B   0.98 , P  D / B   0.95

(iii) P  atmost 2 girls  P  exactly 0 girl, 1girl, 2 girls

Hence the probability of getting six heads 10 times is

(c) at most 1 defective blade in a consignment of 10,000 packets.

 q p(1  q)1  qt p( p)1  q t

Hence P  X  s  t   qst and P  X  s  qs

If f(x) is defined over (a, b) then variance =

20. The average percentage of marks of candidates in an examination is 42 with a

The marginal distribution of X is The marginal distribution of Y is

  e dx. e dy  e  . e   e  e  

10. The lines of regression in a bivariate distribution are X + 9Y = 7 and Y + 4X = 49 .

  ye5 y e 5 y   x2  0.2

4. The joint density function of two random variable X and Y is given

The marginal distribution of X is The marginal

9. If X and Y are independent random variables with PDF e  x , x  0 and e  y , y  0

Now f U (u ) . f V (v)  1. v e  v  f UV (u, v)

Find (i) P  X  1  Y  3 (ii) P  X  Y  3 (iii) P  X  1/ Y  3

independent. (April/May 2009,

14. The joint pdf of random variables X and Y is given by

coefficient of X and Y. (April/May 2009, 2010, 2011)

(iii) To check whether X and Y are independent.

n  8, U  0, V  17, U 2  132, V 2  163, UV  101 U     0 ,

Probability that S lies in the range 108,126 . (A/M 2019)

108  120 126  120 

Given that n  100 ,   60 ,  2  400

UNIT – III RANDOM PROCESSES

2. Write down the classification of random process. ( N/D 2010)

II Order Stationary Process:

 E  X  t   E  X  t    . Hence, E  X  t  is constant. i.e. Mean is constant

(ii) E  X  t1  X  t2   RXX  t2  t1  or E  X  t  X  t     RXX  

9. Define Markov chain. (Nov/Dec 2010)

called the states of the Markov chain.

10. Define one step and n step transition probability.

probability from the state ai to the state a j in n step. It is denoted by Pi (jn ) .

11. Define Homogeneous Markov chain and regular Markov chain.

13. State the postulates of a Poisson process. (Nov/Dec 2010,2017)

Let X1  t  and X 2  t  be two independent Poisson processes.

Their moment generating functions are M X1 t   u   e 1

  12t 2  1t    22t 2  2t   2  1t  2t 

Recall that E  X 2  t  for a Poisson process with parameter  is given

17. Define random telegraph process.

 (at )0 (at )1 (at )2 

2. Given a random variable Y with characteristic function  ( )  E[ei y ] and a random

process defined by X (t )  cos(t  y), show that X t is stationary in the wide

sense if  (1)   (2)  0.

and  are constants and  is uniformly distributed random variable in  0,2  .

4. Verify whether the sine wave process X t , where X t Y cos t , where Y is

uniformly distributed in  0,1 is a SSS process.

5. If the process  X  t  : t  0 is a Poisson process with parameter  , obtain

P  X  t   n . Is the process first order stationary? (May/June 2014)

Pn  t    Pn 1  t   Pn  t  -------- (1)

variables) is a wide sense stationary if E  A   E  B   0, E A2  E B2     and

R t1 ,t2   E  X t1  X t2 

 E  A 2 cos t1 cos t2  AB cos t1 sin t2

 AB sin t1 cos t2  B 2 sin t1 sin t2 

and  is uniformly distributed random variable in  0, 2  . (Nov/Dec 2017)

8. If the process  X t  is given by X t   10cos 100t    where  is uniformly