Chapter 5: Poisson Processes

 Exponential distribution and properties.

 Counting Process

 Poisson Process
 Definition

 Properties

 Nonhomogeneous Poisson Process

 Compound Poisson Process

1. Exponential Distribution: Basic Facts

 e −  x , x  0
f ( x) =  ,  0
Density 0, x  0
1 − e−  x , x  0
CDF F ( x) = 
0, x  0

MGF  ( t ) = E e  =
tX

1  −t
Mean EX  =

1
Variance Var ( X ) =
2
 Properties of Exponential
Distribution

1. Memoryless:
PX  s + t X  t = PX  s for all s, t  0
2. Constant Hazard Rate:
f (t )  e − t
r (t ) = = =
1 − F (t ) 1 − (1 − e ) − t

3. If X1, X2 are independent random variables

with parameters ,  then:
1
P( X 1  X 2 ) =
1 + 2
 Example of Memoryless

Example: The amount of time one spends in

a bank is exponentially distributed with
mean 10 minutes (1 = 10). Find the
probability that a customer will spends more
than 15 minutes given that he is still in the
bank after 10 minutes?
Solution:
PX  15 X  10 = PX  5= e −5 = 0.604
 Counting Process
A stochastic process {N(t), t ≥ 0} is said to be a counting process if N(t)
represents the total number of events that have occurred up to time t.

A counting process N(t) must satisfy:

 N(t) ≥ 0.

 N(t) is integer number.

 If s ≤ t then N(s) ≤ N(t).

 For s ≤ t, N(t) - N(s) equals the number of events that have occurred
in the interval (s,t].
 Stationary & Independent Increments
 A counting process has independent increments if, for any
0  s  t  u  v, N ( t ) − N ( s ) is independent of N ( v ) − N ( u )

That is, the number of events occurred in disjoint intervals are

independent.
 A counting process has stationary increments if the distribution
of the number of events that occur in any interval of time
depends only on the length of the time interval. In other words,
the process has stationary increments if the number of events
in the interval (s, s + t) has the same distribution for all s.
Poisson Process Definition 1
 A counting process {N(t), t  0} is a Poisson process
with rate ,  > 0, if
 N(0) = 0
 The process has independent increments
 The number of events in any interval of length t
follows a Poisson distribution with mean t
(therefore, it has stationary increments), i.e.,
e − t (  t )
n

P  N ( t + s ) − N ( s ) = n = , n = 0,1,...
n!
 Poisson Process Definition 2
A counting process {N(t), t  0} is a Poisson process with rate  ,  > 0,
if
 N(0) = 0
 The process has stationary and independent increments
P  N ( h ) = 1 =  h + o ( h )
P  N ( h )  2 = o ( h )

A function f(.) is said to be o(h) - order of magnitude h - if

f (h)
lim =0
h→0 h
Definitions 1 and 2 are equivalent!
 Interarrival and Waiting Times
The times between arrivals T1, T2, …
N(t) are independent exponential r.v.’s
with mean 1/ :
P T1  t = P  N ( t ) = 0 = e − t
PT2  t = E PT2  t T1
PT2  t T1 = s = P0 event in (s, s + t ) = e −t

 PT2  t = E PT2  t T1 =   e −t e −x dx = e −t
0

S1 S2 S3 S4 t The (total) waiting time until the nth

event has a gamma distribution:
T1 T2 T3 T4
Sn =  i =1Ti
n
 Example
People immigrate into a territory at a Poisson rate  = 1 per day

 Expected time until the tenth immigrants arrives:

10
E S10  = = 10 days

 Probability that the elapsed time between the tenth and the eleventh
arrivals exceeds two days:

PT11  2 = e −2  = 0.133
Other Poisson Process Properties
 Poisson Splitting
Suppose {N(t), t  0} is a P.P. with rate , and suppose
that each time an event occurs, it is classified as type I
with probability p and type II with probability 1-p,
independently of all other events. Let N1(t) and N2(t),
respectively, be the number of type I and type II events up
to time t.

Then {N1(t), t  0} and {N2(t), t  0} are independent

Poisson processes with respective rates p and (1-p).
 Example
Immigrants to area A arrive at a Poisson rate of
10 per week. Each immigrant is of English
descent with prob. 1/12. Find the prob. that no
people of English descent will emigrate during
one month.
The number of Englishman emigrating to A
during one month is Poisson distributed with
mean tp (10×4×1/12 = 10/3) −10
PN1 (t ) = 0 = e 3
 Competing Poisson Processes
the probability that n events occur in one Poisson
process before m events have occurred in a second and
independent Poisson process
 Suppose {N1(t), t  0} and {N2(t), t  0} are independent Poisson
processes with respective rates 1 and 2.

Let Sni be the time of the nth event of process i, i = 1,2.

k n + m −1− k
n + m −1
 n + m − 1  1   1 
P S  S
1 2
=      
   
n m
k =n  k + +
 1 2   1 2 
 Conditional Distribution of Arrival Times
 If Y1, Y2, …, Yn are random variables, then Y(1), Y(2), …, Y(n) are
their order statistics if Y(k) is the kth smallest value among Y1, Y2,
…, Yn, k = 1, …, n.

Suppose {N(t), t  0} is a Poisson process with rate 

and for some time t we know that N(t) = n. Then the
arrival times S1, S2, …, Sn have the same conditional
distribution as the order statistics of n independent
uniform random variables on (0, t).
 Example
 Suppose that items arrive at a processing plant in accordance with a
Poisson process with rate λ. At a fixed time T, all items are dispatched
from the system. The problem is to choose an intermediate time, t ∈ (0,
T), at which all items in the system are dispatched, so as to minimize the
total expected wait of all items.
 If we dispatch at time t, 0 < t < T, then the expected total wait of all
items will be

 To see why this is true, we reason as follows: The expected number of

arrivals in (0, t) is λt, and each arrival is uniformly distributed on (0, t),
and hence has expected wait t/2. Thus, the expected total wait of items
arriving in (0, t) is λt2/2. Similar reasoning holds for arrivals in (t, T),
and the preceding follows. To minimize this quantity, we differentiate
with respect to t to obtain

and equating to 0 shows that the dispatch time that minimizes the expected
total wait is t = T/2.
 Nonhomogeneous Poisson Process
 A counting process {N(t), t  0} is a nonhomogeneous Poisson
process with intensity function  (t), t  0, if:
 N(0) = 0
 The process has independent increments (not stationary incr.)
P  N ( t + h ) − N ( t ) = 1 =  ( t ) h + o ( h )
P  N ( t + h ) − N ( t )  2 = o ( h )
t
 Let m ( t ) =   ( y )dy Then
0

( m ( s + t ) − m ( s ))
n

P  N ( t + s ) − N ( s ) = n = e
−  m ( s + t ) − m ( s ) 
, n = 0,1,...
n!
 Advantage of nonhomogeneous
Poisson process
No longer require the stationary increments =>

Allow for the possibility that events may be more

likely to occur in certain time period than other time

periods
 Example (1)
Consider a hot dog stand which opens at 8AM
 From 8AM until 11AM: customers arrive at a steady increasing rate
that starts from 5 customers per hour and reaches 20 customer per
hour at 11AM.
 From 11Am to 1PM: arrival rate seems to remain constant at 20
customers per hour.
 From 1PM until closing time 5PM: arrival rate drops steadily from 20
to 12 customers per hour.
Suppose that the number of customers arriving during disjoint time
periods is independent. Find the prob. that no customer arrives
between 8:30AM and 9:30AM. Find the expected number of customer
arrivals in that period.
 Example (2)
Intensity function: 5 + 5t 0t 3

 (t ) =  20 3t 5
20 − 2(t − 5) 5  t  9

and:  (t ) =  (t − 9 ) t9
The number of arrival between 8:30AM and 9:30AM will be Poisson with
mean: m(t ) = m(3 2 )− m(1 2 )
Hence, the probability that no customer arrive in this period is:
 32 
 
exp −  (5 + 5t )dt  = e −10
 1 2  3
2

Mean number of arrivals:  (5 + 5t )dt = 10

1
2
 Compound Poisson Process
A counting process {X(t), t  0} is a compound Poisson process if:
N (t )
X ( t ) =  Yi , t  0
i =1
where {N(t), t  0} is a Poisson process and {Yi, i = 1, 2, …} are
independent, identically distributed r.v.’s that are independent of {N(t), t
 0}.
By conditioning on N(t), we can obtain:

E  X ( t )  = tE Y1 
Var  X ( t )  = tE Y1 
2
 Example
 Families migrate to an area at Poisson rate λ= 2 per week.
Number of people in each family is independent and takes on
the values 1,2,3,4 with probability 1/6, 1/3, 1/3, 1/6.

 We have: EY  = 5 2 ; E Y 2  = 43 6

 Expected value and variance of the number of individuals

migrating during five-week period:
E  X (5) = 2  5  5 2 = 25 ; Var  X (5) = 2  5  43 6 = 215 3