STAT 410

Fall 2016

Gamma Distribution:

f (x) =

f (x) =
x 1 e x ,
( )

x 1 e x ,

( )
0x<
1

0x<

E(X) =

E ( X ) = ,

Var ( X ) = 2

Var ( X ) = 2

If T has a Gamma ( , = 1 ) distribution, where is an integer, then

F T ( t ) = P ( T t ) = P ( X t ),

P ( T > t ) = P ( X t 1 ),

where X t has a Poisson ( t = ) distribution.

1.

Alex is told that he needs to take bus #5 to the train station. He misunderstands the
directions and decides to wait for the fifth bus. Suppose that the buses arrive to the bus
stop according to Poisson process with the average rate of one bus per 20 minutes.
X t = number of buses in t hours.

Poisson ( t )

T k = arrival time of the k th bus.

Gamma, = k.

one bus per 20 minutes

a)

= 3.

Find the probability that Alex would have to wait longer than 1 hour for the fifth bus
to arrive.
P ( T 5 > 1 ) = P ( X 1 4 ) = P ( Poisson ( 3 ) 4 ) = 0.815.
OR

P( T5 > 1 ) =

35
t 5 1 e 3 t dt =
(5 )

5
3
4 3t
4! t e dt =
1

b)

Find the probability that the fifth bus arrives during the second hour.
P( 1 < T5 < 2 ) = P( T5 > 1 ) P( T5 > 2 ) = P( X1 4 ) P( X2 4 )
= P ( Poisson ( 3 ) 4 ) P ( Poisson ( 6 ) 4 ) = 0.815 0.285 = 0.530.
OR
2

P( 1 < T5 < 2 ) =

c)

35
t 51 e 3 t dt =
(5 )

35 4
t
4!

e 3 t dt =

Find the probability that the fifth bus arrives during the third hour.
P( 2 < T5 < 3 ) = P( T5 > 2 ) P( T5 > 3 ) = P( X2 4 ) P( X3 4 )
= P ( Poisson ( 6 ) 4 ) P ( Poisson ( 9 ) 4 ) = 0.285 0.055 = 0.230.
OR
3

P( 2 < T5 < 3 ) =

2.

3 5 4 3t
t e
dt =
4!

During a radio trivia contest, the radio station receives phone calls according to Poisson
process with the average rate of five calls per minute.
X t = number of phone calls in t minutes.

Poisson ( t )

T k = time of the k th phone call.

Gamma, = k.

five calls per minute

a)

35
t 5 1 e 3 t dt =
(5 )

= 5.

Find the probability that we would have to wait less than two minutes for the ninth
phone call.

P ( T 9 < 2 ) = P ( X 2 9 ) = 1 P ( X 2 8 ) = 1 P ( Poisson ( 10 ) 8 ) = 1 0.333 = 0.667.

OR
2

P( T9 < 2 ) =

59
t 91 e 5 t dt =
(9 )

5 9 8 5 t
t e
dt =
8!

b)

Find the probability that the ninth phone call would arrive during the third minute.
P( 2 < T9 < 3 ) = P( T9 > 2 ) P( T9 > 3 ) = P( X2 8 ) P( X3 8 )
= P ( Poisson ( 10 ) 8 ) P ( Poisson ( 15 ) 8 ) = 0.333 0.037 = 0.296.
OR
3

P( 2 < T9 < 3 ) =

59
t 9 1 e 5 t dt =
(9 )

5 9 8 5t
t e
dt =
8!

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

( x ) =

x 1 e u du ,

1
=
2

x>0

( n ) = ( n 1 )!

( x ) = (x 1 ) (x 1)

if n is an integer

_ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _

2(r)

The Chi-Square Distribution

f ( y )=

1
(r

2)2 r 2

yr

E(Y) = r

3.

a)

2 1 e y 2 ,

0y<

Var ( Y ) = 2 r

Let X be a random variable with a chi-square distribution with r degrees of

freedom. Show that X has a Gamma distribution. What are and ?

M X ( t ) = M 2( r ) ( t ) =
M Gamma ( , ) ( t ) =

(1 2t ) r 2
1

(1 t )

t<

t<

1
.
2

If X has a chi-square distribution with r degrees of freedom, then X has a

Gamma distribution with = r 2 and = 2.

Let Y be a random variable with a Gamma distribution with parameters and

= 1 . Assume is an integer. Show that 2 Y has a chi-square distribution.

b)

What is the number of degrees of freedom?

M Y ( t ) = M Gamma ( , ) ( t ) =

(1 t )

t<

If W = a Y + b, then M W ( t ) = e b t M Y ( a t ).
M 2 Y/ ( t ) = M Y ( 2 t ) =

(1 2 t )

t<

1
.
2

/ has a chi-square distribution with r = 2 degrees of freedom.

2Y

4.

Let Y be a random variable with a Gamma distribution with = 5 and = 3.

a)

Find the probability P ( Y > 18 )

i)

by integrating the p.d.f. of the Gamma distribution;

P ( Y > 18 ) =

ii)

18 (5) 3

x 5 1 e x 3 dx =
5

18

1
x 4 e x 3 dx =
5,832

by using the relationship between Gamma and Poisson distributions;

where X 18 is Poisson ( 18 = 6 ).

P ( Y > 18 ) = P ( X 18 4 ) = 0.285

EXCEL:

= POISSON( x , , 0 )

gives

P( X = x )

= POISSON( x , , 1 )

gives

P( X x )

A
1
2

=POISSON(4,18/3,1)

1
2

0.285057

iii)

by using the relationship between Gamma and Chi-square distribution.

2
2

P ( Y > 18 ) = P Y > 18 = P ( X > 12 )

3
3

= CHIINV ( , v )

gives

= CHIDIST ( x , v )

gives

A
1

where X is

2 ( 5 2 = 10 ).

2 (v ) for 2 distribution with v degrees

of freedom, x s.t. P ( 2 (v ) > x ) = .
2 distribution
with v degrees of freedom, P ( 2 (v ) > x ).
the upper tail probability for

=CHIDIST(12,10)

0.285057

b)

Find a and b such that P ( a < Y < b ) = 0.90.

2
Y is
3

2 ( 5 2 = 10 degrees of freedom ).

P ( 3.940 < 2 ( 10 ) < 18.31 ) = 0.95 0.05 = 0.90.

P ( 3.940

3
3
< Y < 18.31 ) = P ( 5.91 < Y < 27.465 ) = 0.90.
2
2

P ( 0 < 2 ( 10 ) < 15.99 ) = 0.90 0.00 = 0.90.

P ( 0

3
3
< Y < 15.99 ) = P ( 0 < Y < 23.985 ) = 0.90.
2
2

P ( 4.865 < 2 ( 10 ) < ) = 1.00 0.10 = 0.90.

P ( 4.865

3
3
< Y < ) = P ( 7.2975 < Y < ) = 0.90.
2
2

Let T 7 be a random variable with a Gamma distribution with = 7 and = 5.

Find the probability P ( 20 < T 7 < 30 ).

5.

[ Text messages arrive according to Poisson process, on average once every 5 minutes. Find the
probability that we would have to wait more than 20 minutes but less than 30 minutes for the
7th text message. ]

30

P ( 20 < T 7 < 30 ) =

20

30

1
(7 ) 5

t 7 1 e t 5 dt =

20

1
6! 5

t 6 e t 5 dt =

OR
P ( 20 < T 7 < 30 ) = P ( T 7 > 20 ) P ( T 7 > 30 ) = P ( X 20 6 ) P ( X 30 6 )
= P ( Poisson ( 4 ) 6 ) P ( Poisson ( 6 ) 6 ) = 0.889 0.606 = 0.283.
[ If the 7th text message arrives after 20 minutes, then we could have received at most 6 text
messages during the first 20 minutes. If the average time between the text messages is 5
minutes, then the expected number of text messages in 20 minutes is 4.
If the 7th text message arrives after 30 minutes, then we could have received at most 6 text
messages during the first 30 minutes. If the average time between the text messages is 5
minutes, then the expected number of text messages in 30 minutes is 6. ]
A
1

=POISSON(6,4,1)

=POISSON(6,6,1)

=A1-A2

0.889326

0.606303

0.283023

4
OR
2
2
2

P ( 20 < T 7 < 30 ) = 20 < T 7 < 30 = P ( 8 < 2 ( 2 7 = 14 ) < 12 )

5
5
5

A
1

=CHIDIST(8,14)

=CHIDIST(12,14)

=A1-A2

0.889326

0.606303

0.283023

Let Y be a random variable with a Gamma distribution with = 5 and = 4

6.

( i.e., = 0.25 ). Find the probability P ( Y 15 )

a)

by integrating the p.d.f. of the Gamma distribution;

P ( Y 15 ) =

b)

15

0 (5) 4

15

x 51 e x 4dx =
5

1
x 4 e x 4 dx =
24,576

by using the relationship between Gamma and Poisson distributions;

where X 15 is Poisson ( 15 = 3.75 ).

P ( Y 15 ) = P ( X 15 5 ) = 1 P ( X 15 4 )

EXCEL:

= POISSON( x , , 0 )

gives

P( X = x )

= POISSON( x , , 1 )

gives

P( X x )

A
1

=1-POISSON(4,15/4,1)

c)

0.322452

by using the relationship between Gamma and Chi-square distribution.

2
2

P ( Y 15 ) = P Y 15 = P ( X 7.5 )
4
4

= CHIINV ( , v )

gives

= CHIDIST ( x , v )

gives

A
1

=1-CHIDIST(7.5,10)

where X is

2 ( 5 2 = 10 ).

2 (v ) for 2 distribution with v degrees

of freedom, x s.t. P ( 2 (v ) > x ) = .
2 distribution
with v degrees of freedom, P ( 2 (v ) > x ).
the upper tail probability for

1
2

0.322452

7.

Let X be a random variable with a Gamma distribution with = 3 and = 5

( i.e., = 0.2 ). Find the probability P ( X > 31.48 )

a)

by integrating the p.d.f. of the Gamma distribution;

P ( X > 31.48 ) =

31.48

1
25

1
25

x 2 e x 5 dx

5 x 2 e x 5 2 5 2 x e x 5 2 5 3 e x 5

31.48

= 0.04999.

b)

by using the relationship between Gamma and Poisson distributions;

Hint: If X has a Gamma ( , = 1 ) distribution, where is an integer, then

F X ( t ) = P ( X t ) = P ( Y ) and P ( X > t ) = P ( Y 1 ), where

Y has a Poisson ( t ) distribution.
P ( X > 31.48 ) = 1 P ( X 31.48 ) = 1 P ( Y 3 ) = P ( Y 2 )
where Y has a Poisson ( 31.48 5 = 6.296 ) distribution.

6.296 0 e 6.296 6.296 1 e 6.296 6.296 2 e 6.296

+
+
0!
1!
2!

= 0.00184 + 0.01161 + 0.03654 = 0.04999.

c)

by using the relationship between Gamma and Chi-square distribution.

Hint: If X has a Gamma ( , = 1 ) distribution, where is an integer, then

2X
has a chi-square distribution with 2 degrees of freedom.

P ( X > 31.48 ) = P ( 2 X 5 > 12.592 ) = P ( 2 ( 6 ) > 12.592 ) = 0.05.

8.

a)

Mistakes that Alex makes in class occur according to Poisson process with the
average rate of one mistake per 10 minutes. Find the probability that the third
mistake Alex makes occurs during the last 15 minutes of a 50-minute class.

Notations:

X t = number of server failures in t days.

T k = time of the k th server failure.

1 min

= 10,

1
= 0.10.
10

P ( 35 < T 3 < 50 ) = P ( T 3 > 35 ) P ( T 3 > 50 ) = P ( X 35 2 ) P ( X 50 2 ) =

5 min

= 2,

1
= 0.50.
2

P ( 7 < T 3 < 10 ) = P ( T 3 > 7 ) P ( T 3 > 10 ) = P ( X 7 2 ) P ( X 10 2 ) =

10 min

= 1,

= 1.

P ( 3.5 < T 3 < 5 ) = P ( T 3 > 3.5 ) P ( T 3 > 5 ) = P ( X 3.5 2 ) P ( X 5 2 ) =

= P ( Poisson ( 3.5 ) 2 ) P ( Poisson ( 5.0 ) 2 ) = 0.3208 0.1247 = 0.1961.

OR

50
1 min

P ( 35 < T 3 < 50 ) =

35

10
5 min

P ( 7 < T 3 < 10 ) =

0.50 3 3 1 0.50 t
t
e
dt =
(3 )

5.0
10 min

P ( 3.5 < T 3 < 5 ) =

0.10 3 3 1 0.10 t
t
e
dt =
(3 )

3.5

1
t 3 1 e t dt =
(3 )

5.0

3.5

50

35

10

1
t 2 e 0.10 t dt =
2,000

16

t 2 e 0.50 t dt =

1 2 t
t e dt =
2

b)

1 min

Students ask questions in class according to Poisson process with the average
rate of one question per 20 minutes. Find the probability that the third question
is asked during the last 10 minutes of a 50-minute class.

= 20,

1
= 0.05.
20

P ( 40 < T 3 < 50 ) = P ( T 3 > 40 ) P ( T 3 > 50 ) = P ( X 40 2 ) P ( X 50 2 ) =

5 min

= 4,

1
= 0.25.
4

P ( 8 < T 3 < 10 ) = P ( T 3 > 8 ) P ( T 3 > 10 ) = P ( X 8 2 ) P ( X 10 2 ) =

10 min

= 2,

1
= 0.50.
2

P( 4 < T3 < 5 ) = P( T3 > 4 ) P( T3 > 5 ) = P( X4 2 ) P( X5 2 ) =

20 min

= 1,

= 1.

P ( 2 < T 3 < 2.5 ) = P ( T 3 > 2 ) P ( T 3 > 2.5 ) = P ( X 2 2 ) P ( X 2.5 2 ) =

= P ( Poisson ( 2.0 ) 2 ) P ( Poisson ( 2.5 ) 2 ) = 0.6767 0.5438 = 0.1329.
OR
50
1 min

P ( 40 < T 3 < 50 ) =

40
10
5 min

P ( 8 < T 3 < 10 ) =

8
5
10 min

P( 4 < T3 < 5 ) =

0.25 3 3 1 0.25 t
t
e
dt =
(3 )

0.50 3 3 1 0.50 t
t
e
dt =
(3 )
2.5

20 min

P ( 2 < T 3 < 2.5 ) =

2.0

50

0.05 3 3 1 0.05 t
t
e
dt =
(3 )

1
t 3 1 e t dt =
(3 )

40

10

1
t 2 e 0.05 t dt =
16,000

128

t 2 e 0.25 t dt =

16

t 2 e 0.50 t dt =

2.5

2.0

1 2 t
t e dt =
2