Scribd
Upload
48 views

04 Exponential Poisson

This document discusses exponential distributions, Poisson processes, and their properties. It provides examples of how to calculate expected costs, times, and probabilities for systems that can be modeled by these distributions. Key concepts covered include the memoryless property of exponential distributions, the minimum/sum of exponential random variables, and that inter-arrival times in a Poisson process are independent and exponentially distributed.

Uploaded by

Fadila Nastiti
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
48 views

04 Exponential Poisson

This document discusses exponential distributions, Poisson processes, and their properties. It provides examples of how to calculate expected costs, times, and probabilities for systems that can be modeled by these distributions. Key concepts covered include the memoryless property of exponential distributions, the minimum/sum of exponential random variables, and that inter-arrival times in a Poisson process are independent and exponentially distributed.

Uploaded by

Fadila Nastiti
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 29

TKI 3206 Operations Research 2

04 Exponential Distribution and Poisson Process

S in t a R S u lis t yo
s in t a .s ulis tyo @ugm.a c.id

Last Lecture
o Unconditional Probability
o Classification of States
o Long Run Behavior of DTMC

Outline
o Cost Calculation
o Exponential Distribution
o Poisson Process

Cost Calculation

State Based Costs


= long run (L-R) average fraction of time DTMC spends in state .
State based costs
()

= expected cost per visit to state


= expected cost per time unit spent in state

Long run average (state based) costs per unit time:


Transition Based Costs


= long run (L-R) average fraction of time DTMC spends in state .
Transition based costs
(, )

= expected cost per transition from state to state

Long run average (transition based) costs per unit time:


,
S

Example
=

0 0.7
1 0.4

0
1

0.3
0.6

0 = 4/7

1 = 3/7

if machine down on day


if machine up on day

= 0,1
Assume is a DTMC with transition probability matrix
When up, machine makes nothing w.p 0.1, 1 item w.p 0.3, and 2
items w.p 0.6. Each item can be sold for $ 200. Operating cost per
day $50 only if the machine is up. When down, repair costs $ 100
per day. Expected cost per repair is $ 80.
How much is the L-R average daily profit?

Exponential Distribution

Exponential Distribution
~()

~()

is a continuous RV
Probability Density Function (PDF)

=
0

0
<0

Cumulative Density Function (CDF)

= =

= 1/

= 1/2

0
otherwise

Properties of Exponential
Property 1: memoryless property

~ ()
, 0
> + , >
> + > } =
== >
>

Example
The amount of time a person spends in a bank is exponentially
distributed with mean 10 minutes.
a.

What is the probability that a customer will spend more than


15 minutes in the bank?

b. What is the probability that a customer will spend more than


15 minutes in the bank given that he is still in the bank after 10
minutes?

Properties of Exponential
Property 2: minimum of some exponential RVs
1 , 2 , are independent
~ ( )

= min

~ (

=1 )

= 1,2, ,

Properties of Exponential
Property 3: sum of some exponential RVs
1 , 2 , independent
~ ()
=

= 1,2, ,

=1

~ ,

Erlang Distribution
~ Erlang (, )
1
=
1 !
0
1

=1
=0

= /
= /2

0
<0
0

Example
The lifetime of an electronic tube is exponentially distributed with
mean 1,000 hours. There are three of these tubes (original and 2
spares). When one burns out, it is replaced by a spare.
What is the expected amount of time Y until the last spare burns
out?

Properties of Exponential
Property 4: between two exponential RVs
~
~ ()

< =
+
= =0

> =
+

Example
Suppose a stereo system consists of two main parts, a radio and a
speaker. If the lifetime of the radio is exponential with mean 1000
hours and the lifetime of the speaker is exponential with mean
500 hours, independent of the radios lifetime, then
What is the probability that the systems failure (when it occurs)
will be caused by the radio failing?

Counting Process

Counting Process
{ } is a counting process if () is the number of events by time

>

= increment from to
= number of events in (, ]

The Poisson Process (PP)

The Poisson Process (PP)


A PP with rate > 0 is a counting process with
i.
ii.
iii.

0 =0

has independent increments

The increment over any period of length is Poisson


distributed with parameter
~ + ~ ()

Poisson Distribution
A Poisson () RV is discrete with PMF
= =

=
=

0,1,2,
otherwise

Example
The arrival of customers at a caf is a Poisson process with rate 4
per minute. Find the probability that there is no less than 3 arrivals
between time (0,2] (the first two minutes)!

Inter-arrival Time
()

= number of events by time

= time when first event occur

= time between events and 1 for 2

1 , 2 , are all () distribution independent


Hence,
Time between events in a PP() are iid () RVs

In a PP, we cannot have two events take place at same time.


0 = 0
= time when event takes place ( 1)
1 = 1
2 = 1 + 2
=

=1

~ (, )
Note that,
=
= = 1 < = 1
=1

1
=0

= = 1

1
=0 {

=0 !

= }

Example
Suppose that people immigrate into a territory at a Poisson rate
=1 per day.
a.

What is the expected time until the tenth immigrant arrives?

b. What is the probability that the elapsed time between the


tenth and the eleventh arrival exceeds two days?

Next Lecture
o a little bit more on Poisson Process
o Semi Markov Processes
o Continuous Time Markov Chain

Question?

thankyou

Acknowledgement:
Prof. Sigrn Andradttir

You might also like