CHAPTER: 4

POISSON PROCESSES

In the previous chapter, we discussed a general theory of random

processes. In this chapter, we will focus on some specific random
processes that are used frequently in applications. More specifically,
we will discuss the Chapter 4: Poisson process, chapter 3: Markov
chains.

1
 The Bernoulli Process
independent identically distributed (IID)
binary Random Variables.
We often visualize a Bernoulli process as evolving in discrete time with the event
representing an arriving customer at time i and {Zi = 0} representing no arrival.

IID means: if each random variable has

the same probability distribution as the
others and all are mutually  Independent.

2
 Counting Processes

 In some problems, we count the occurrences of some types of events. In such

scenarios, we are dealing with a counting process.

 For example, you might have a random process N(t) that shows the number of
customers who arrive at a supermarket by time t starting from time 0.

 For such a processes, we usually assume N(0)=0, so as time passes and

customers arrive, N(t) takes positive integer values.

3
Definition:

A random process {N(t),t∈[0,∞)} is said to be a counting process if

N(t) is the number of events occurred from time 0 up to and including
time t. For a counting process, we assume

N(0)=0;
N(t)∈{0,1,2,⋯}, for all t∈[0,∞);
for 0≤s<t, N(t)−N(s) shows the number of events that occur in the
interval (s,t].

 Since counting processes have been used to model arrivals (such as

the supermarket example above), we usually refer to the occurrence
of each event as an "arrival".

4
 A counting process has independent increments if the numbers of arrivals in non-
overlapping (disjoint) intervals are independent.

 A counting process has stationary increments if, for all t2>t1≥0, N(t2)−N(t1) has
the same distribution as N(t2−t1).

5
 Basic Concepts of the Poisson Process

The Poisson process is one of the most widely-used counting processes. It is

usually used in scenarios where we are counting the occurrences of certain events
that appear to happen at a certain rate, but completely at random (without a certain
structure).

 For example, suppose that from historical data, we know that earthquakes occur
in a certain area with a rate of 2 per month. Other than this information, the timings
of earthquakes seem to be completely random. Thus, we conclude that the Poisson
process might be a good model for earthquakes.

Example:

 The number of car accidents at a site or in an area;

The location of users in a wireless network;
The requests for individual documents on a web server;
The outbreak of wars;
Photons landing on a photodiode.
6
Poisson random variable:
A discrete random variable X is said to be a Poisson random variable with parameter
μ, shown as X∼ Poisson(μ), if its range is RX={0,1,2,3,...}, and its PMF is given by

Poisson Process as the Limit of a Bernoulli Process:

Suppose that we would like to model the arrival of events that happen completely at
random at a rate λ per unit time. Here is one way to do this. At time t=0, we have no
arrivals yet, so N(0)=0. We now divide the half-line [0,∞) to tiny subintervals of
length δ as shown in Figure 2.
7
 Figure 11.2 - Dividing the half-line [0,∞) to tiny subintervals of length δ.

 Each subinterval corresponds to a time slot of length δ. Thus, the intervals are
(0,δ], (δ,2δ], (2δ,3δ], ⋯. More generally, the kth interval is ((k−1)δ, kδ].

 We assume that in each time slot, we toss a coin for which P(H)=p=λδ. If the
coin lands heads up, we say that we have an arrival in that subinterval.

 Otherwise, we say that we have no arrival in that interval. Figure 11.3 shows
this process. Here, we have an arrival at time t=kδ, if the kth coin flip results in
a heads.

8
 Now, let N(t) be defined as the number of arrivals (number of heads) from time 0
to time t. There are n≈ t/δ time slots in the interval (0,t].
 Thus, N(t) is the number of heads in n coin flips. We conclude that
N(t)∼Binomial(n,p). Note that here p=λδ, so

The PMF of N(t) converges to a Poisson distribution with rate λt.

9
 The Poisson Process
Let λ>0 be fixed. The counting process {N(t),t∈[0,∞)} is called a Poisson process
with rates λ if all the following conditions hold:
N(0)=0;
N(t) has independent increments;
The number of arrivals in any interval of length τ>0 has Poisson(λτ) distribution.

Example
The number of customers arriving at a grocery store can be modeled by a Poisson
process with intensity λ=10 customers per hour.

Find the probability that there are 2 customers between 10:00 and 10:20.
Find the probability that there are 3 customers between 10:00 and 10:20 and 7 customers
between 10:20 and 11.

Solution:
1. Here, λ=10 and the interval between 10:00 and 10:20 has length τ=1/3 hours. Thus, if X
is the number of arrivals in that interval, we can write X∼Poisson(10/3). Therefore,

10
2, Here, we have two non-overlapping intervals I1=(10:00 a.m., 10:20 a.m.] and I2= (10:20
a.m., 11 a.m.]. Thus, we can write

Since the lengths of the intervals are τ1=1/3 and τ2=2/3 respectively, we obtain
λτ1=10/3 and λτ2=20/3. Thus, we have

 The arrival rate is the number of arrivals per unit of time.

 The inter arrival time is the time between each arrival into the system and the next.
 inter arrival time = 1/arrival rate
If 12 customers enter a store per hour, the time between each arrival is;

inter arrival time = 1/arrival rate = 1/12= 0.083(hours)== 0.083 x 60 minutes= 5 (minutes)
11
Therefore from the arrival rate of 12 per hour, the time between each arrival is 5 minutes.
 Arrival and Interarrival Times:
Let N(t) be a Poisson process with rate λ. Let X1 be the time of the first interarrival.
Then,

Therefore, X1∼Exponential(λ). Let X2 be the time elapsed between the first and
the second arrival (Figure 11.4).

Figure 11.4 - The random variables X1, X2, ⋯ are called the interarrival times of the counting process
N(t).

Let s>0 and t>0. Note that the two intervals (0,s] and (s,s+t] are disjoint. We can
write

12
We conclude that X2∼Exponential(λ), and that X1 and X2 are independent. The random
variables X1, X2, ⋯ are called the interarrival times of the counting process N(t). Similarly, we
can argue that all Xi's are independent and Xi ∼Exponential(λ) for i=1,2,3, ⋯.

 Interarrival Times for Poisson Processes

If N(t) is a Poisson process with rate λ, then the interarrival times X1, X2, ⋯ are
independent and
Xi∼Exponential(λ), for i=1,2,3,⋯.

 Remember that if X is exponential with parameter λ>0, then X is a

memoryless random variable, that is

13
Example:
Let N(t) be a Poisson process with intensity λ=2, and let X1, X2, ⋯ be the
corresponding interarrival times.
a. Find the probability that the first arrival occurs after t=0.5, i.e., P(X1>0.5).
b. Given that we have had no arrivals before t=1, find P(X1>3).
c. Given that the third arrival occurred at time t=2, find the probability that the
fourth arrival occurs after t=4.
d. I start watching the process at time t=10. Let T be the time of the first arrival that
I see. In other words, T is the first arrival after t=10. Find ET and Var(T).

(a)

14
 (c)

(d) When I start watching the process at time t=10, I will see a Poisson process.
Thus, the time of the first arrival from t=10 is Exponential (2). In other words, we
can write

15
Now that we know the distribution of the interarrival times is exponential , we
can find the distribution of arrival times:

More specifically, Tn is the sum of n independent Exponential(λ) random variables.

As if Tn=X1+X2+⋯+Xn, where the Xi's are independent Exponential(λ) random
variables, then Tn∼Gamma(n,λ). The Gamma(n,λ) is also called Erlang distribution,
i.e, we can write

The PDF of Tn, for n=1,2,3,⋯, is given by

16
Since Tn=X1+X2+⋯+Xn, we conclude that

Arrival Times for Poisson Processes

If N(t) is a Poisson process with rate λ, then the arrival times T1, T2, ⋯ have
Gamma (n,λ) distribution. In particular, for n=1,2,3,⋯, we have

17
 Merging and Splitting Poisson Processes
Merging Independent Poisson Processes:
Let N1(t) and N2(t) be two independent Poisson processes with rates λ1 and λ2
respectively. Let us define N(t)=N1(t)+N2(t). That is, the random process N(t) is
obtained by combining the arrivals in N1(t) and N2(t) (Figure 11.5). We claim
that N(t) is a Poisson process with rate λ=λ1+λ2.
To see this, first note that
N(0)=N1(0)+N2(0)=0+0=0.

Figure 11.5 - Merging two Poisson processes N1(t) and N2(t).

18
The numbers of arrivals in I associated with N1(t) and N2(t) are Poisson(λ1τ) and
Poisson(λ2τ) and they are independent. Therefore, the number of arrivals in I
associated with N(t) is Poisson((λ1+λ2)τ) (sum of two independent Poisson random
variables).
Splitting a Poisson Processes
Let N(t) be a Poisson process with rate λ. Here, we divide N(t) to two processes
N1(t) and N2(t) in the following way (Figure 11.6). For each arrival, a coin with
P(H)=p is tossed. If the coin lands heads up, the arrival is sent to the first process
(N1(t)), otherwise it is sent to the second process. The coin tosses are independent
of each other and are independent of N(t). Then,
 N1(t) is a Poisson process with rate λp;
 N2(t) is a Poisson process with rate λ(1-p);
 N1(t) and N2(t) are independent.

19
Figure 11.6 - Splitting a Poisson process to two independent Poisson processes.
Problem 1
Let {N(t),t∈[0,∞)} be a Poisson process with rate λ=0.5.
Find the probability of no arrivals in (3,5].
Find the probability that there is exactly one arrival in each of the following
intervals: (0,1], (1,2], (2,3], and (3,4].

Answer: 1. If Y is the number arrivals in (3,5], then Y ∼Poisson(μ=0.5×2).

Therefore,

2. Let Y1, Y2, Y3 and Y4 be the numbers of arrivals in the intervals (0,1], (1,2], (2,3], and
(3,4]. Then Yi∼Poisson(0.5) and Yi's are independent, so

P(Y1=1,Y2=1,Y3=1,Y4=1) = P(Y1=1)⋅P(Y2=1)⋅P(Y3=1)⋅P(Y4=1)

20
Problem 2
Let N1(t) and N2(t) be two independent Poisson processes with rates λ1=1 and λ2=2,
respectively. Let N(t) be the merged process N(t)=N1(t)+N2(t).
Find the probability that N(1)=2 and N(2)=5.
Given that N(1)=2, find the probability that N1(1)=1.

Answer: N(t) is a Poisson process with rate λ=1+2=3.

b.

21
 Nonhomogeneous Poisson Processes
Let N(t) be the number of customers arriving at a fast food restaurant by time t. We
think that the customers arrive somewhat randomly, so we might want to model N(t)
as a Poisson process. However, we notice that this process does not have stationary
increments.

For example, we note that the arrival rate of customers is larger during lunch time
compared to, say, 4 p.m. In such scenarios, we might model N(t) as a
nonhomogeneous Poisson process. Such a process has all the properties of a Poisson
process, except for the fact that its rate is a function of time, i.e., λ=λ(t).

Nonhomogeneous Poisson Process

Let λ(t):[0,∞)↦[0,∞) be an integrable function. The counting process
{N(t),t∈[0,∞)} is called a nonhomogeneous Poisson process with rate λ(t) if all the
following conditions hold.
N(0)=0;
N(t) has independent increments;

22
For a nonhomogeneous Poisson process with rate λ(t), the number of arrivals in any
interval is a Poisson random variable; however, its parameter can depend on the
location of the interval. More specifically, we can write

23