1st Semester M.

Tech Microwave & TV Engineering

And Signal Processing
Tutorial 6: Functions of Two Random Variables

1. Let X and Y be independent exponential random variables with common parameter . Find fZ (z) by
min(x; y)
defining Z =
max(x; y)
Solution:

Y ; XY

min(X, Y )
=
max(X, Y )

Y ; X>Y
X
X
Y
FZ (z) = P ( z, X Y ) + P ( z, X > Y )
Y
X
min
Z=
1 z 1
max
Z=

Since X and Y are exponential,x>0,y>0.So 0 z 1

The area is {Y > 0} {Y X} {Y

X
}
Z

This area is {X > 0} {X > Y } {Y XZ}

FZ (z) =

R R yz
y=0 x=0

fZ (z) =

fXY (x, y) dxdy +

R R xz
x=0 y=0

fXY (x, y) dydx

FXY (x, xz)dx

yfXY (yz, y)dy +

fXY (x, y) = 2 e(x+y)

R
0

xfXY (x, xz)dx

FZ (z) =

R R yz

y=0 x=0

2 e(x+y) dxdy +

ex dx

=
y=0
Z

ex dy +

x=0

ey (1 eyz )dy +

y=0

x=0 y=0

2 e(x+y) dydx

ex dy

x=0

xz

ex dx

y=0

ex (1 exz dx

x=0

yz

R R xz

yz

dy +

y(1+z)

ex dx

ex(z+1) dx

x=0

x=0

y=0

Z
dy +

1
1
2
=1
+1
=2
(1 + z)
(1 + z)
(1 + z)
fZ (z) =

d
2
FZ (z) =
2; 0 < z < 1
dz
(1 + z)

2. Let X U(0;1) and Y U(0; 1) be independent random variables. Find the density function of Z,
defining Z = 2lnXcos(2Y )
Solution:
X u(0,1),Y u(0,1)

Z= 2lnXcos(2Y ); W=Y

Z= 2lnXcos(2W );
Range of Z cos(2Y ) from 1 to + 1
r
ln

1
as X f rom 0 to1 to 0
X2

So Z f rom to

 X

 Z

J Y (x, y) = 
 Y


Z






=
Y 

W
X
W


z 2 sec2 (2W )

2
 zsec2 (2W ) e




0

= zsec2 (2W )e

X = e

zsec(2W )2
2

X
W
1

z 2 sec2 (2W )
2

; Y =W

fZW (z, w) = fx (x1 )fy (y1 )|J 1 | = zsec2 (2W )e

z 2 sec2 (2W )
2

fZ (z) =
w=0
Z 1

zsec2 (2W )e
zsec2 (2W )e

w=0

Page 2

z 2 sec2 (2W )
2

dw

z 2 (1+tan2 (2W ))
2

dw

Put
u=z tan(2 W)
du= 2zsec2 (2W )d
w= 0 u= 0
w= 1 u= 0
1
w= u=
4
3
w= u=
4

Hence

1 ( z2 + u2 )
2
e 2
du
u= 2
Z
u2
z2
1
1
e 2 du
= e 2
2
2 u=
1 z2
= e 2 ; <z <
2

fZ (z) =

3. Let X and Y be independent gamma random variables, X G(m;) and Y G(n;). Define Z = X +
Y and W = X/Y . Show that Z and W are independent random variables.
Solution:
X G(m, ); Y G(n, )
Z = X + Y, W =

X
Y

x
x
e 2
fX (x) =
;x > 0
(m)m
y
y n1 e 2
fY (y) =
;y > 0
(n)n
m1




 Z
Z  

1
1 
 X
 

Y
x+y
x 1
 


J(x, y) = 
 =  1 x  = 2 =
 W W  

y
y
y2

 

2


y
y
X
Y
zw
z
;
z = x + y; x = wy; = z = (w + 1)y = y =
w+1 w+1
2

|J(x1 , y1 )| =

| z|
z(w + 1)
(w + 1)
=
=
2
|z|
z
z
|w + 1|

Page 3

fZW (z, w) =

fXY (x1 , y1 )
|J(x1 , y1 )|

fX (x1 ).fY (y1 )

|J(x1 , y1 )|

z
xm1 e y n1 e
.
(m)m (n)n (w + 1)2

n1
(x+y)
zw m1
z
)
(
)
e z
1
+
w
1
+
w
=
2
(m)(n)m+n (w + 1)

z (m+n1) w(m1) e
(m)(n)(m+n) (w + 1)

m+n

z (m+n1) e
(m + n)w(m1)
.
(
(m + n) m + n) (n)(m)(1 + w)(m+n)
|
{z
} |
{z
}
F unction of z

W =

X
=
Y

fW (w) =

R +

|y|fX (yw)fY (y) dy

y

y(yw)
ey y (n1) e
dy
(m)m
(n)n

y (m+n1) w(m1) e
(m)(n)(m+n)

=
0

(m1)

=
Z

F unction of w=G(w)

y(w+1)

w(m1)
.
(n)(m)(m+n)

dy

y (m+n1) e

y(w+1)

dy

y(w + 1)

(w + 1)
Then du =
dy

Let u =

Therefore, fW (w) =

R (m+n1) u
w(m1) (m+n1)
u
e du
(m)(n)(m+n) (w + 1)(m+n1) (w + 1) 0

w(m1) (m + n)
= G(w)
(m)(n)(1 + w)(m+n)
z

Therefore fZ (z) =

z (m+n1) e d
(m + n)(m+n)

= fZW (z, w) = fZ (z)fW (w) = Z and W are independent.

Page 4

4. Paul is vacationing in Monte Carlo. The amount X he takes to the casino each evening is a random
variable with the PDF shown in Fig.1. At the end of each night, the amount Y that he has on leaving
the casino is uniformly distributed between zero and twice the amount he took in

a. Determine the joint PDF fXY (x,y).

b. What is the probability that on any given night Paul makes a profit at the casino? Justify.
c. Find and sketch the probability density function of Pauls profit on any particular night, Z = Y-X.
Solution:

x; 0 x 40
800
fX (x) =

0;elsewhere
a) fXY (x, y) = f Y fX (x) =
X

; 0 y 2x
2x
f Y (y/x) =
X

0;elsewhere
1
1
1
.x.
=
;
800 2x
1600

Page 5

0 x 40, 0 y 2x;

b) P (Y X) =

R 40 R 2x
x=0 y=x

fXY (x, y) dydx

Z 40
1
(2x x)dx
1600 0
1 x2 40
=
. ]
1600 2 0
1
=
2

c) Z=Y-X,W=X

Page 6


 Z

 X

J(x, y) = 
 W


X



 
 
  1
=
1
W 

Y
Z
Y


1 
=1
0 

x = w; y = z + x = z + w
0 x 40 0 < w < 40
0 y 2x x < w < 40
0 y 2x x < z < x
Z

R
w

40

fZW (z, w) dw =
| w=z

Z 40
1
1
1
1
dw +
dw =
(40 z) +
(40 + z)
1600
1600
1600
1600
{z
} | w=z {z
}

(z>0)

1
(40 |z|);
=
1600

(z<0)

x<z <x

5. X and Y are independent uniformly distributed random variables in (0; 1). Let W = max(X, Y ) and Z
= min(X, Y ). Find the pdf of
a. r=w-z
b. s=w+z
Solution:
X u(0,1);Y u(0,1);W max(X,Y);Z=min(X,Y)

a) r=wz

R =
W Z
 = max(X, Y ) min(X, Y )
X Y; X Y
=
FR (r) = P (X Y < r, X Y ) + P (Y X < r, X < Y )
Y X; X < Y
r2
x 0 0 to r = FR (r) =
2
R1 Rx
x r to 1 = FR (r) = x=r y=xr 1.dydx
=

R1
r

rdx = rx]1r = r(1 r)

Page 7

FR (r) = r2 + 2r(1 r) = 2r r2

2(1 r); 0 < r < 1

fR (r) =

0; otherwise

b) S = W + Z
= max(X, Y ) + min(X, Y )
=X +Y
fS (s) = fX (x) fY (y)
s; 0 < s < 1
=

2 s; 1 < s < 2

Page 8

6. Consider independent Gaussian random variables X and Y with the joint probability density function
(x2 +y 2 )
1
e 22 . Random variables Z and W are defined in terms of X and Y by the
fXY (x, y) =
2
2
transformations Z = X 2 + Y 2 and W = Y/X. Find FZW , FZ and FW .
Solution:

Z = X2 + Y 2 ; W =
fXY (x, y) =

1
2 2

Y
X

; < x, y <

Range of z = 0 to
Range of w = to
Taking polar co-ordinates,

X = R cos
= Z = R , = tan1 W
Y = R sin
RR
FZW (z, w) = x2 +y2 z2 fxy (x, y)dxdy
Over this area r can vary from 0 to z.

to tan1 w
2

x < 0, varies f rom to f rom to + tan1 w

2

f or x 0, varies f rom to f rom

tan1

FZW (z, w) =

+tan1

fxy (x, y)dxdy +

r=0

=
2

fr (r, )rdrd
r=0

tan1

1 r22
e 2 rdrd
2 2
r=0 =
2
Z z

1 r22
= 2(tan1 + )
e 2 rdr
2 0 2 2
=2

r2
2 2
r
du = 2 dr

Let u =

Page 9

=
2

z2
1 R 2
) 0 2 eu du
2
z2
1
= (tan1 + ) (1 e 22 )
2

FZW (z, w) = (tan1 +

; 0 < z < and < <

z2

FZ (z) = 1 e 22 u(z)
1
1
FZ (z) = + tan1 w ; < <
2
7. Let be a prescribed angle and consider the rotational transformation V = Xcos + Y sin and W =
(x2 +y 2 )
1
2 2
e
. Find fV W (v, w).
Xsin - Y cos with fXY (x, y) =
2 2
fXY
|J|
 
 
  cos
=
sin


Solution: fV W =



J=


v
x
w
x

v
y
w
y


sin 
= 1
cos 

|J| = 1
v = x cos + y sin
w = x sin y cos
fXY (x1 , y1 )
fV W (v, w) =
1
Solving for x1 and y1
v y1 sin
;
cos
w + y1 cos
x1 =
sin
x1 =

Equating ,
v y1 sin
w + y1 cos
=
cos
sin
v sin y1 sin2 = w cos + y1 cos2 = v sin w cos = y1
v x1 cos
;
y1 =
sin
x1 sin w
y1 =
cos
Equating ,
v cos x1 cos2 = x1 sin2 w sin = x1 = v cos + w sin
fV W (v, w) = fXY (v cos + w sin , v sin w cos )
1
=
exp (v 2 cos2 + w2 sin2
2 2
v 2 sin2 + w2 cos2
+ 2vw sin cos +
2vw sin cos )
2 2
 2

2
1 v 2+w
2
=
e
; < v, w <
2
2

Page 10

8. Let X and Y be independent identically distributed exponential random variables with common parameter . Find the pdf of
a. Z = Y/max(X,Y )
b. W = X/min(X,2Y )
Solution:

Y
Y
= X
a. Z =
1
max(X, Y )
FZ (z) = P (Z z) = P (

;X Y

0z1

;X < Y
Y
z, X Y )
X
Z

|x|fX (x)fY (xz)dx

fZ (z) =
x=
Z

xfX (x)fY (xz)dx

Zx=0

xex exz dx

x=0

= 2

xex(1+z) dx

x=0
 x(1+z)

Z x(1+z) 
xe
e

(1 + z)
(1
+ z)
0
n

o
2

x
ex(1+z)
=
0 lim x(1+z) +
x e
(1 + z)
(1 + z) 0


1

0
(0 1)
=
(1 + z)
(1 + z)
1
=
;0 z < 1
(1 + z)2

= 2

P (Z = 1) = P (X < Y ) =

2 (x+y)

y=0 x=y
Z
y y

Z
=

e2y dy

0

2y

e
2 0
1
1
=
(0 1) =
2
2

X
b. W =
min(X, 2Y )
X
; X 2Y
= 2Y
1
; X < 2Y

1w<

Page 11

dxdy =

e
0

dy

ey


dy
y

|2y|fX (2yw)fY (y)dy

fW (w) =

2ye2yw ey dy
Z
2
yey(1+2w) dy
= 2
=

y is exponential

Let u= y(1+2w)
du= (1+2w)dy
y=0 u=0
y= u= .
Therefore
Z

du
ueu
(1 + 2w) (1 + 2w)
22
= 2
(1 + 2w)2 (2)
2
=
;1 < w <
(1 + 2w)2

fW (w) = 22

2 e(x+y) dxdy

P (W = 1) = P (X < 2Y ) =
2

ey

y=0 x=2y

x

ey e2y dy

=
0

Z
=

e3y dy

0

3y

e
3

1
=
3

Page 12

dy
2y

9. Determine fZ (z) for Z = X+Y , where X and Y are two independent random variables with probability
densities as shown in the figure

Solution: Z=X+Y
fZ (z) = fX (x) fY (y)

case 1:

case 2:
.

1 + z < 0 z < 1
fZ (z) < 1 z < 0
1 + z > 0 z > 1
1+z <1z <0


1<z <0

Page 13

1+z

fZ (z) =
0

1 3
. (1 (z t)2 )dt
3 4

1+z

1
(1 z 2 t2 + 2zt)dt
4
0
t3
1
= (t z 2 t + zt2 )1+z
0
4
3
1
(1 + z)3
= [(1 + z)2 z 2 (1 + z)
+ z(1 + z)2 ]
4
3
1
= (2z + z 2 z 3 )
4
=

case 3:
.

1+z >1z >0

1 + z < 0 z < 1

0<z<1
1

1 3
. (1 (z t)2 )dt =
3
4
0
1
t3
= (t z 2 t + zt2 )10
4
3
(1
1
+ z]
= [1 z 2
4
3
1 2
= [ z 2 + z]
4 3

Z
0

1
(1 z 2 t2 + 2zt)dt
4

1+z

fZ (z) =

case 4:
.

1+z >2z >1

1 + z < 1 z < 2
Z


1<z<2

fZ (z) =
1+z

1 3
. (1 (z t)2 )dt +
3 4

2 3
. (1 (z t)2 )dt
3 4

1
t3
1
t3
= (t z 2 t + zt2 )11+z + (t z 2 t + zt2 )1+z
2
4
3
2
3
3
z
7
z
4
=
z2 +
12
4
2 3

case 5:
.

1+z >3z >2

1 + z < 2 z < 3


2<z<3

1+z

1
t3
2 3
. (1 (z t)2 )dt = (t z 2 t + zt2 )1+z
2
3 4
2
3
2
1 z 3
4
= [
+ 2z 2 3z + ]
2 3
3
z 3
3
2
=
+ z2 z +
6
2
3
Z

fZ (z) =

Page 14

case 6:

1 + z > 2 z > 3
1 + z < 3 z < 4


3<z<4

2 3
1
t3
. (1 (z t)2 )dt = (t z 2 t + zt2 )31+z
2
3
1+z 3 4
7
16
= ( z 3 7z 2 + 10z )
3
3
Z

fZ (z) =

10. Let X be the lifetime of a certain electric bulb and Y that of its replacement after the failure of the first
bulb. Suppose X and Y are independent with common exponential density function with parameter .
Find the probability that the combined lifetime exceeds 2. What is the probability that the replacement
outlasts the original component by ?
Solution:
X Exponential(),
Y Exponential()
X and Y are independent
Z=X+Y

Z (s) = X (s).Y (s) =

2

fZ (z) = 2 zez u(z)

z
z
= 2 e u(z)

We need to find P (z > 2),

P (z > 2) =
P ut x =
dx = dz

z
z
e dz
2 2

P (z > 2) =

R
2

xex dx


x

xe
1
2
= 3e
= 0.406

ex ]2

11. The length of time Z, an airplane, can y is given by Z = X where X is the amount of fuel in its tank
and > 0 is a constant of proportionality. Suppose a plane has two independent fuel tanks so that
when one gets empty the other switches on automatically. Because of lax maintenance, a plane takes
of with neither of its fuel can checked. Let X1 be the fuel in the first tank and X2 the fuel in the
second tank. Let X1 and X2 be modelled as independent, identically distributed random variables with
1
fX1 (x) = fX2 (x) = [u(x) u(x b)]. Compute the pdf of Z, the maximum flying time of the plane. If
b
b=100, say in litres, and = one hour/ ten litres, what is the probability that the plane will y atleast
five hours?
Solution: Let Z1 = X1 & Z2 = X2
Z = Z1 + Z2
Then,
1
fZ1 (z1 ) = fX1 ( z1 )

Page 15

i
1 h z1
z1
u( ) u( b)
b

Similarly,
i
1 h z2
z2
u( ) u( b)
b

fZ (z) = fZ1 (z1 ) fZ2 (z2 )

Z1 & Z2 are independent.
1
& b = 100
=
10
fZ2 (z2 ) =

1
[u(10z1 ) u(10z1 100)]
10
1
[u(10z2 ) u(10z2 100)]
fZ2 (z2 ) =
10
fZ1 (z1 ) =

We need to find P (Z 5) which is indicated by shaded area.

1
Area of the unshaded triangle = 12 5 20
= 0.125
P (Z 5) = 1 0.125
= 0.875

Page 16