1

Statistical Tutorial
(Chapter 5)
Compiled and Presented by
Alana Cordick
Histograms
Example:
0.04 12 8
0.02 7 7
0.04 13 6
0.04 12 5
0.24 71 4
0.15 45 3
0.37 110 2
0.10 30 1
Relative Frequency Frequency Arrivals /
Party
Histogram of Party Size
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
1 2 3 4 5 6 7 8
# in Party
R
e
l
a
t
i
v
e

F
r
e
q
u
e
n
c
y
Statistical Concepts Part 4
Continuous Distributions
Continuous Distributions
Continuous random variables can be used to describe random
phenomena in which the variable can take on any value in some
interval
In this section, the distributions studied are:
Normal
Uniform
Exponential
Gamma
Erlang
Weibull
Triangular
Lognormal
Beta distribution
2
Normal (Gaussian)
Distribution
Continuous
AKA Bell Curve
Most common continuous distribution found in nature (as seen
through the Central Limit Theorem to come)
Distribution:
Mean:
Variance:
Denoted by: X ~ N(,
2
)
~
|
|
|
|
|
|
|
|
j
)

|
\
[ ~
~ = p p x
x
x f ,
2
1
exp
2
1
) (
2
o

o
p p
0
2
f
Normal (Gaussian)
Distribution
Evaluating the distribution:
Use numerical methods (no closed form)
Independent of and , using the standard normal
distribution:
Z ~ N(0,1)
Transformation of variables: let Z = (X - ) / ,

=
z
t
dt e z
2 /
2
2
1
) ( where ,

( )
) ( ) (
2
1
) (
/ ) (
/ ) (
2 /
2
o

o
o

~
~
~
~
~
~
d = =
=
j
)

|
\
[ ~
s = s =
j
j
x
x
x
z
dz z
dz e
x
Z P x X P x F
Normal (Gaussian)
Distribution
Standard Normal Distribution
Has a = 0 and = 1 i.e.
Special properties:
.
f(-x)=f(+x); the pdf is symmetric about
The maximum value of the pdf occurs at x =
The mean and mode are equal
Linearity
~ X N(0, 1)
0 ) ( lim and , 0 ) ( lim = =

x f x f
x x
Normal Distribution
Linearity
Any linear combination of normally distributed
random variable is normally distributed
Y = a
i
X
i
i=1
n

Normally distributed
In General
~ X + Y N(
X
+
Y
,
X
2
+
Y
2
)
~ X N(
X
,
X
2
) If then and ~ Y N(
Y
,
Y
2
)
~ X Y N(
X

Y
,
X
2
+
Y
2
)
3
Normal Distribution
Evaluating the distribution:
Use numerical methods (no closed form)
Independent of and , using the standard normal
distribution:
Z ~ N(0,1)
Transformation of variables: let Z = (X - ) / ,

=
z
t
dt e z
2 /
2
2
1
) ( where ,

( )
) ( ) (
2
1
) (
/ ) (
/ ) (
2 /
2
o

o
o

~
~
~
~
~
~
d = =
=
j
)

|
\
[ ~
s = s =
j
j
x
x
x
z
dz z
dz e
x
Z P x X P x F
Uniform Distribution
A random variable X is uniformly
distributed on the interval (a,b) if its pdf is
given by:
And a cdf of:

=
otherwise , 0
,
1
) (
b x a
a b
x f

=
b x
b x a
a b
a x
a x
x F
, 1
,
, 0
) ( p
p
Uniform Distribution
P(x
1
< X < x
2
) is proportional to the length
of the interval for all x
1
and x
2
such that:
a <= x
1
< x
2
<=b
a b
x x
x F x F x X x P

= = < <
1 2
1 2 2 1
) ( ) ( ) (
Uniform Distribution
Mean:
Variance:
2
) (
b a
X E
+
=
12
) (
) (
2
a b
X V

=
4
Uniform Distribution
Applications:
Random numbers, uniformly distributed
between zero and 1, provide the means to
generate a truly random event
Used to generate samples of random variates
from all other distributions
Uniform Distribution
Example
A bus arrives every 20 minutes at a specified
stop beginning at 6:40 a.m. and continuing until
8:40 a.m. A certain passenger does not know
the schedule, but arrives randomly (uniformly
distributed) between 7:00 a.m. and 7:30 a.m.
every morning. What is the probability that the
passenger wait more than 5 minutes for a bus?
Uniform Distribution
Example
First concentrate on the bus arrival times at the stop:
6:40
7:00
7:20
7:40

We are only interested in its arrival between 7:00 and
7:30
Uniform Distribution
Example
So we know the bus will arrive at 7:00 and 7:20
What ranges of time will the passenger not
have to wait more than 5 minutes?
7:00
7:15 7:20
Therefore, we want to find the probability that
the passenger arrives during the other times
P(0 < X < 15) + P(20 < X < 30)
5
Uniform Distribution
Example
Since X is a uniform random variable (0, 30),
we have the following:
6
5
30
20
1
30
15
0 30
0 20
0 30
0 30
0 30
0 15
) 20 ( ) 30 ( ) 15 (
=
+ =

=
+ = F F F
There is an 5/6 =
83% chance the
passenger will
have to wait more
than 5 minutes
a b
a x
x F

= ) (
Exponential Distribution
A random variable X is said to be
exponentially distributed with parameter >
0 if its pdf is given by:
With cdf:

=

otherwise , 0
0 ,
) (
x e
x f
x

=
=

0 , 1
0 0,
) (
0
x e dt e
x
x F
x
x t

p
Exponential Distribution
Application
Model interarrival times (time between arrivals) when
arrivals are completely random
= arrivals / hour
Model service times
= services / minute
Model the lifetime of a component that fails
catastrophically (i.e. light bulb)
= failure rate
Exponential Distribution
is the value where the pdf intersects the y
axis
6
Exponential Distribution
Memoryless property
For all s>= 0 and t >=0
P(X > s + t | X > s) = P(X > t)
In other words, if you know a component has
survived s hours so far, the same distribution of
the remaining amount of time that it survives is
the same as the original distribution
It does not remember the it already has been used
for s amount of time
Exponential Distribution
Example:
Suppose the life of an industrial lamp is
exponentially distributed with failure rate =
1/3 (one failure every 3000 hours on the avg.)
Determine the probability the lamp will last
longer than its mean life
368 . 0 ) 1 ( 1 ) 3 ( 1 ) 3 ( 1 ) 3 (
1 3 / 3
= = = = = >

e e F X P X P

=

0 , 1
0 0,
) (
x e
x
x F
x
p
Exponential Distribution
Example:
The probability that an exponential random
variable is greater than its mean is always 0.368
Regardless of the value of
The probability the industrial lamp will last
between 2000 and 3000 hours is:
145 . 0 487 . 0 632 . 0 ) 1 ( ) 1 (
) 1 ( ) 1 ( ) 2 ( ) 3 ( ) 3 2 (
3 / 2 1
3 / 2 3 / 3
= = =
= =


e e
e e F F X P
Exponential Distribution
Example:
The probability that the lamp will last for another 1000
hours given that it is operating after 2500 hours
717 . 0
) 1 ( 1
) 1 ( 1
) 1 (
) 5 . 2 | 1 5 . 2 (
) 5 . 2 | 5 . 3 (
3 / 1
3 / 1
= =
=
=
> =
> + > =
> >

e
e
X P
X P
X X P
X X P
7
Gamma Distribution
A function used in defining the gamma
distribution is the gamma function, which is
defined for all > 0 as:
Gamma function is a generalization of the
factorial notion to all positive #s (not just
integers)
)! 1 ( ) ( =
Gamma Distribution
A random variable X is gamma distributed
with parameters and if its pdf is given
by:
Where is a shape parameter and is a scale
parameter

>
=

otherwise
x e x
x f
x
, 0
0 , ) (
) ( ) (
1

Gamma Distribution
Mean:
Variance:
Cdf of X:

1
) ( = X E
2
1
) (

= X V
0 , ) (
0 , 0
) (
1
) (
1
>
|
|
|

|
s
I
~
=
~ ~

j
x dt e t
x
x F
t
x
0
0
0
Gamma Distribution
Gamma and exponential distributions are
related when is an integer
If the random variable, X, is the sum of
independent, exponentially distributed random
variables, each with parameter , then X has a
gamma distribution with parameters and
Independent if: X= X
1
+ X
2
+ + X

8
Weibull Distribution
A random variable X has a Weibull
distribution if its pdf has the form
3 parameters:
Location parameter: ,
Scale parameter: , ( > 0)
Shape parameter. , (> 0)
|
|
|

|
~
|
|
|
|
|
|
|
|
j
)

|
\
[ ~
~ j
)

|
\
[ ~
=
~
otherwise , 0
, exp
) (
1
v
o
v
o
v
o

x
x x
x f
Weibull Distribution
Example: = 0 and = 1
When = 1,
X ~ exp( = 1/)
Weibull Distribution
Mean:
Variance:
Cdf:
j
j
)

|
|
\
[
+ I + = 1
1
) (

o v X E
|
|
|
|
|
|
|
|
|
|
|
|
|
|
j
j
)

|
|
\
[
+ I ~
j
j
)

|
|
\
[
+ I =
2
2
1
1
1
2
) (

o X V
|
|
|

|
~ ~
<
=
|
|
|
|
|
|
|
|
j
)

|
\
[ ~
~
v x e
v x
x F v x
, 1
, 0
) (

o
Weibull Distribution
Applications
Can model a wide range of failure rates
When multiple components are in a process,
this distribution can be used to govern the time
of the first failure
i.e. ball bearing, capacitator, relay, and strength
material failures
Reference: http://www.itl.nist.gov/div898/handbook/apr/section1/apr162.htm
9
Erlang Distribution
The pdf from the Gamma distribution is
often referred to as the Erlang distribution
of order (or number of phases) k when =
k, an integer

>
=

otherwise
x e x
x f
x
, 0
0 , ) (
) ( ) (
1

Erlang Distribution
Mean:
Variance:
Cdf:

1 1 1 1
) ( = + + =
k k k
X E L
( ) ( ) ( )
2 2 2 2
1 1 1 1
) (
k k k k
X V = + + = L
( )

>
=

0 , 0
0 ,
!
1
) (
1
0
x
x
i
x k e
x F
k
i
i x k

Erlang Distribution
Application
When a customer must complete a series of k
stations
The next customer cannot enter start the first station
until the previous customer has completed all k
stations
Erlang Distribution
Example
A college professor of electrical engineering is leaving
home for the summer, but would like to have a light
burning at all times to discourage burglars. The
professor rigs up a device that will hold two light bulbs.
The device will switch the current to the second bulb if
the first bulb fails. The box in which the light bulbs are
packages says, Average life 1000 hours, exponentially
distributed. The professor will be gone 90 days (2160
hours).
What is the probability that a light will be burning when the
summer is over and the professor returns?
10
Erlang Distribution
Example
Reliability function the probability that the system
will operate at least x hours
R(x) = 1 F(x)
= k = 2
k = 1 / 1000, so = 1/2000 per hour
Determine F(2160):
[ ]
( )
636 . 0
!
16 . 2
1
!
) 2160 )( 2000 / 1 )( 2 (
1 ) 2160 (
1
0
16 . 2
1
0
) 2160 )( 2000 / 1 )( 2 (
=
=
=

i
i
i
i
i
e
i
e
F
Erlang Distribution
Example
Now:
R(x) = 1 F(x)
R(x) = 1 0.636 = 0.364
Therefore, there is a 36% chance that a light
will be burning when the professor returns
Summary
Normal
Exponential
Weibull
~
|
|
|
|
|
|
|
|
j
)

|
\
[ ~
~ = p p x
x
x f ,
2
1
exp
2
1
) (
2
o

=

otherwise , 0
0 ,
) (
x e
x f
x

=
=

0 , 1
0 0,
) (
0
x e dt e
x
x F
x
x t

p
|
|
|

|
~ ~
<
=
|
|
|
|
|
|
|
|
j
)

|
\
[ ~
~
v x e
v x
x F v x
, 1
, 0
) (

o
|
|
|

|
~
|
|
|
|
|
|
|
|
j
)

|
\
[ ~
~ j
)

|
\
[ ~
=
~
otherwise , 0
, exp
) (
1
v
o
v
o
v
o

x
x x
x f
Review
The time to failure of a nickel-cadmium battery is
Weibull distributed with parameters v=0, =1/4,
and =1/2 years
What fraction of batteries are expected to last longer
than the mean life?
368 . 0 1 1
2
1
1
2
1
1
2
1
1 1 2 / 1
2 / 1
4 / 1
4 / 1
= = =
|
|
|
|
|
|
|
|
~ ~ = j
)

|
\
[
~ = j
)

|
\
[
s ~ = j
)

|
\
[
>
~ ~
j
)

|
\
[
~
e e e F X P X P
11
Triangular Distribution
A random variable X has a triangular distribution if
its pdf is given by
( )
( )( )
( )
( )( )
c b a where
otherwise , 0
,
2
1
,
2
) (

<

= c x b
a c b c
x c
b x a
a c a b
a x
x f
F(x)
x b a c
( ) a c
=
2
Height
Triangular Distribution
Mean:
Mode: occurs at x = b
Variance exercise in text
This distribution uses the mode more than the
mean
3
) (
c b a
X E
+ +
=
( ) c a X E b Mode + = = ) ( 3
Triangular Distribution
Cumulative Density Function
( )
( )( )
( )
( )( )

>
<

<

=
c x , 1
, 1
,
, 0
) (
2
2
c x b
a c b c
x c
b x a
a c a b
a x
a x
x F
Lognormal Distribution
A random variable X has a lognormal distribution
if its pdf has the form
Mean E(X) = e
+
2
/2
Variance V(X) = e
2+
2
/2 (
e

2

- 1)
( )
|
|
|

|
|
|
|
|
|
|
|
|
~
~
=
otherwise 0,
0 ,
2
ln
exp
2
1
) (
2
2
f x

x
x
x f
=1,
2
=0.5,1,2.
12
Lognormal Distribution
Relationship with normal distribution
When Y ~ N(,
2
), then X = e
Y
~ lognormal(,

2
)
Parameters and
2
are not the mean and
variance of the lognormal
=1,
2
=0.5,1,2.
Beta Distribution
A random variable X is beta-distributed
with parameters
1
> 0 and
2
> 0 if its pdf
is given by:
Range: (0, 1)
( )
( )
( )
( ) ( )
( )
2 1
2 1
2 1
2 1
1 1
,
otherwise , 0
1 0 ,
,
1
) (
2 1





+

=

< <

=

B
x
B
x x
x f
Beta Distribution
Range (0, 1) is restrictive
May want a range (a, b), where a < b
To do this let Y be the random variable such that
Y = a + (b a) X
Mean (of Y):
Variance (of Y):
( )

+
+ =
2 1
1
) (

a b a Y E
( )
( ) ( )
j
j
)

|
|
\
[
+ + +
~ =
1
) (
2 1
2
2 1
2 1
2 2


o a b Y
Poisson Distribution
Definition: N(t) is a counting function that
represents the number of events occurred in [0,t]
A counting process {N(t), t>=0} is a Poisson
process with mean rate if:
1. Arrivals occur one at a time
2. {N(t), t>=0} has stationary increments (completely
random, without rush or slack periods)
3. {N(t), t>=0} has independent increments
13
Poisson Distribution
Properties
Equal mean and variance: E[N(t)] = V[N(t)] = t (=
as before)
Since stationary increments, the number of arrivals in
time s to t, such that s < t, is also Poisson-distributed
with mean (t-s)
,... 2 , 1 , 0 and 0 for ,
!
) (
] ) ( [ = = =

n t
n
t e
n t N P
n t

( ) [ ]
[ ] [ ] ) ( ) ( ) ( ) ( ) (
,... 2 , 1 , 0 ,
!
] ) ( ) ( [
) (
s N t N V s t s N t N E
n
n
s t e
n s N t N P
n s t
= =
=

= =

Interarrival Times
Consider the interarrival times of a Poisson process (A
1
, A
2
, ), where
A
i
is the elapsed time between arrival i and arrival i+1
The 1
st
arrival occurs after time t iff there are no arrivals in the interval
[0,t], hence:
P{A
1
> t} = P{N(t) = 0} = e
-t
Since: P{A
1
<= t} = 1 e
-t
[cdf of exp()]
Interarrival times, A
1
, A
2
, , are exponentially distributed and
independent with mean 1/
Stationary & Independent Memoryless
Arrival counts
~ Poi()
Interarrival time
~ Exp(1/)
Splitting
Suppose each event of a Poisson process can be
classified as Type I (high priority arrival), with
probability p and Type II (low priority arrival),
with probability 1-p
N(t) = N
1
(t) + N
2
(t), where N
1
(t) and N
2
(t) are
both Poisson processes with rates p and (1-p)
N(t) ~ Poi()
N
1
(t) ~ Poi[p]
N
2
(t) ~ Poi[(1-p)]

p
(1-p)
Pooling
Pooling:
Suppose two Poisson processes are pooled
together
N
1
(t) + N
2
(t) = N(t), where N(t) is a Poisson
processes with rates
1
+
2
N(t) ~ Poi(
1
+
2
)
N
1
(t) ~ Poi[
1
]
N
2
(t) ~ Poi[
2
]

1
+
2

2
14
Nonstationary Poisson Process
(NSPP)
Poisson Process without the stationary increments,
characterized by (t), the arrival rate at time t
Useful when arrival rates vary i.e. restaurant
The expected number of arrivals by time t:
Relating stationary Poisson process n(t) with rate =1 and
NSPP N(t) with rate (t):
Let arrival times of a stationary process with rate = 1
be t
1
, t
2
, , and arrival times of a NSPP with rate (t)
be T
1
, T
2
, , we know: t
i
= (T
i
) & T
i
=
1
(t
i
)

=
t
(s)ds (t)
0
Nonstationary Poisson Process
(NSPP)
Example: Suppose arrivals to a Post Office have
rates 2 per hour from 8 am until 12 pm, and then
0.5 per hour until 4 pm
Let t = 0 correspond to 8 am, NSPP N(t) has rate
function:
Expected number of arrivals by time t:

=
8 4 , 5 . 0
4 0 , 2
) (
p
p
t
t
t
( )

+ = + = +

=

8 4 , 6
2
5 . 0 * 4 5 . 0 * 4 * 2 5 . 0 2
4 0 , 2
) (
4
0 4
p
p
t
t
t ds ds
t t
t
t
Nonstationary Poisson Process
(NSPP)
Hence, the probability distribution of the number
of arrivals between 11 am and 2 pm.
11 am = 3
rd
hour
2 pm = 6
th
hour
P[N(6) N(3) = k]
= P[N((6)) N((3)) = k]
= P[N(9) N(6) = k]
= e
(9-6)
(9-6)
k
/k!
= e
3
(3)
k
/k!
Empirical Distributions
A distribution whose parameters are the observed values in
a sample of data.
May be discrete or continuous
May be used when it is impossible or unnecessary to establish that
a random variable has any particular parametric distribution.
Advantage: no assumption beyond the observed values in the
sample.
Disadvantage: sample might not cover the entire range of possible
values.
15
References
Discrete-Event System Simulation, 4
th
Ed.
Banks, Carson, Nelson, Nicol
Published by Prentice-Hall, 2005