Poisson PDF

Chapter 2.
Poisson Processes
Prof. Ai-Chun Pang

Graduate Institute of Networking and Multimedia,
Department of Computer Science and Information Engineering,
National Taiwan University, Taiwan
Outline
• Introduction to Poisson Processes
• Properties of Poisson processes
– Inter-arrival time distribution
– Waiting time distribution
– Superposition and decomposition
• Non-homogeneous Poisson processes (relaxing stationary)
• Compound Poisson processes (relaxing single arrival)
• Modulated Poisson processes (relaxing independent)
• Poisson Arrival See Average (PASTA)
Prof. Ai-Chun Pang, NTU 1

Introduction
Inter-arrival
n~(t )
time
~
x3
~ arrival epoch
x2
~
x1
~ ~ ~ ~ t ~
S0 S1 S2 S3 S4
(i) nth arrival epoch S̃n is

n
S̃n = x̃1 + x̃2 + . . . + x̃n = i=1 x̃i
S̃0 = 0
(ii) Number of arrivals at time t is: ñ(t). Notice that:
if f if f
{ñ(t) ≥ n} ⇔ {S̃n ≤ t}, {ñ(t) = n} ⇔ {S̃n ≤ t and S̃n+1 > t}

Introduction
Arrival Process: X = {x̃i , i = 1, 2, . . .}; x̃i ’s can be any

S = {S̃i , i = 0, 1, 2, . . .}; S̃i ’s can be any
N = {ñ(t), t ≥ 0}; −→ called arrival process
Renewal Process: X = {x̃i , i = 1, 2, . . .}; x̃i ’s are i.i.d.

S = {S̃i , i = 0, 1, 2, . . .}; S̃i ’s are general distributed
N = {ñ(t), t ≥ 0}; −→ called renewal process
Poisson Process: X = {x̃i , i = 1, 2, . . .}; x̃i ’s are iid exponential distributed

S = {S̃i , i = 0, 1, 2, . . .}; S̃i ’s are Erlang distributed
N = {ñ(t), t ≥ 0}; −→ called Poisson process

Counting Processes
• A stochastic process N = {ñ(t), t ≥ 0} is said to be a counting process
if ñ(t) represents the total number of “events” that have occurred up
to time t.
• From the definition we see that for a counting process ñ(t) must
satisfy:
1. ñ(t) ≥ 0.
2. ñ(t) is integer valued.
3. If s < t, then ñ(s) ≤ ñ(t).
4. For s < t, ñ(t) − ñ(s) equals the number of events that have
occurred in the interval (s, t].

Definition 1: Poisson Processes
The counting process N = {ñ(t), t ≥ 0} is a Poisson process with rate λ
(λ > 0), if:
1. ñ(0) = 0
2. Independent increments relaxed ⇒ Modulated Poisson Process
P [ñ(t) − ñ(s) = k1 |ñ(r) = k2 , r ≤ s < t] = P [ñ(t) − ñ(s) = k1 ]
3. Stationary increments relaxed ⇒ Non-homogeneous Poisson Process
P [ñ(t + s) − ñ(t) = k] = P [ñ(l + s) − ñ(l) = k]
4. Single arrival relaxed ⇒ Compound Poisson Process
P [ñ(h) = 1] = λh + o(h)
P [ñ(h) ≥ 2] = o(h)

Definition 2: Poisson Processes
The counting process N = {ñ(t), t ≥ 0} is a Poisson process with rate λ
(λ > 0), if:
1. ñ(0) = 0
2. Independent increments
3. The number of events in any interval of length t is Poisson distributed
with mean λt. That is, for all s, t ≥ 0
n
−λt (λt)
P [ñ(t + s) − ñ(s) = n] = e , n = 0, 1, . . .
n!

Theorem: Definitions 1 and 2 are equivalent.
Proof. We show that Definition 1 implies Definition 2. To start, fix u ≥ 0
and let
g(t) = E[e−uñ(t) ]
We derive a differential equation for g(t) as follows:
g(t + h) = E[e−uñ(t+h) ]

= E e−uñ(t) e−u[ñ(t+h)−ñ(t)]

−uñ(t) −u[ñ(t+h)−ñ(t)]
= E e E e by independent increments

−uñ(h)
= g(t)E e by stationary increments (1)

Conditioning on whether ñ(t) = 0 or ñ(t) = 1 or ñ(t) ≥ 2 yields

E e−uñ(h) = 1 − λh + o(h) + e−u (λh + o(h)) + o(h)
= 1 − λh + e−u λh + o(h) (2)
From (1) and (2), we obtain that
g(t + h) = g(t)(1 − λh + e−u λh) + o(h)
implying that
g(t + h) − g(t) −u o(h)
= g(t)λ(e − 1) +
h h

Letting h → 0 gives
g (t) = g(t)λ(e−u − 1)
or, equivalently,
g (t)
= λ(e−u − 1)
g(t)
Integrating, and using g(0) = 1, shows that
log(g(t)) = λt(e−u − 1)
or
−u −1)
g(t) = eλt(e → the Laplace transform of a Poisson r. v.
Since g(t) is also the Laplace transform of ñ(t), ñ(t) is a Poisson r. v.

Theorem: Definitions 1 and 2 are equivalent
Let Pn (t) = P [ñ(t) = n].
We derive a differential equation for P0 (t) in the following manner:
P0 (t + h) = P [ñ(t + h) = 0]
= P [ñ(t) = 0, ñ(t + h) − ñ(t) = 0]
= P [ñ(t) = 0]P [ñ(t + h) − ñ(t) = 0]
= P0 (t)[1 − λh + 0(h)]
Hence
P0 (t + h) − P0 (t) o(h)
= −λP0 (t) +
h h
Letting h → 0 yields
P0 (t) = −λP0 (t)
Since P0 (0) = 1, then
P0 (t) = e−λt
Theorem: Definitions 1 and 2 are equivalent
Similarly, for n ≥ 1
Pn (t + h) = P [ñ(t + h) = n]
= P [ñ(t) = n, ñ(t + h) − ñ(t) = 0]
+P [ñ(t) = n − 1, ñ(t + h) − ñ(t) = 1]
+P [ñ(t + h) = n, ñ(t + h) − ñ(t) ≥ 2]
= Pn (t)P0 (h) + Pn−1 (t)P1 (h) + o(h)
= (1 − λh)Pn (t) + λhPn−1 (t) + o(h)
Thus
Pn (t + h) − Pn (t) o(h)
= −λPn (t) + λPn−1 (t) +
h h
Letting h → 0,
Pn (t) = −λPn (t) + λPn−1 (t)

The Inter-Arrival Time Distribution
Theorem. Poisson Processes have exponential inter-arrival time
distribution, i.e., {x̃n , n = 1, 2, . . .} are i.i.d and exponentially
distributed with parameter λ (i.e., mean inter-arrival time = 1/λ).
Proof.
e−λt (λt)0
x̃1 : P (x̃1 > t) = P (ñ(t) = 0) = = e−λt
0!
·
· · x̃1 ∼ e(t; λ)
x̃2 : P (x̃2 > t|x̃1 = s)

= P {0 arrivals in (s, s + t]|x̃1 = s}
= P {0 arrivals in (s, s + t]}(by independent increment)
= P {0 arrivals in (0, t]}(by stationary increment)
= e−λt ·
· · x̃2 is independent of x̃1 and x̃2 ∼ exp(t; λ).
⇒ The procedure repeats for the rest of x̃i ’s.
The Arrival Time Distribution of the nth Event
Theorem. The arrival time of the nth event, S̃n (also called the waiting
time until the nth event), is Erlang distributed with parameter (n, λ).
Proof. Method 1 :
∞

· · e−λt (λt)k
· P [S̃n ≤ t] = P [ñ(t) ≥ n] =
k=n
k!
· λe−λt (λt)n−1
· · fS̃n (t) = (exercise)
(n − 1)!
Method 2 :
fS̃n (t)dt = dFS̃n (t) = P [t < S̃n < t + dt]
= P {n − 1 arrivals in (0, t] and 1 arrival in (t, t + dt)} + o(dt)
= P [ñ(t) = n − 1 and 1 arrival in (t, t + dt)] + o(dt)
= P [ñ(t) = n − 1]P [1 arrival in (t, t + dt)] + o(dt)(why?)
The Arrival Time Distribution of the nth Event
e−λt (λt)n−1
= λdt + o(dt)
(n − 1)!
f (t)dt λe −λt (λt)n−1
·
· · lim S̃n = fS̃n (t) =
dt→0 dt (n − 1)!

Conditional Distribution of the Arrival Times
Theorem. Given that ñ(t) = n, the n arrival times S̃1 , S̃2 , . . . , S̃n have
the same distribution as the order statistics corresponding to n i.i.d.
uniformly distributed random variables from (0, t).
...........................................................................
Order Statistics. Let x̃1 , x̃2 , . . . , x̃n be n i.i.d. continuous random
variables having common pdf f . Define x̃(k) as the kth smallest value
among all x̃i ’s, i.e., x̃(1) ≤ x̃(2) ≤ x̃(3) ≤ . . . ≤ x̃(n) , then x̃(1) , . . . , x̃(n)
are known as the “order statistics” corresponding to random variables
x̃1 , . . . , x̃n . We have that the joint pdf of x̃(1) , x̃(2) , . . . , x̃(n) is
fx̃(1) ,x̃(2) ,...,x̃(n) (x1 , x2 , . . . , xn ) = n!f (x1 )f (x2 ) . . . f (xn ),
where x1 < x2 < . . . < xn (check the textbook [Ross]).

...........................................................................
Proof. Let 0 < t1 < t2 < . . . < tn+1 = t and let hi be small enough so that
ti + hi < ti+1 , i = 1, . . . , n.
· ·
· P [ti < S̃i < ti + hi , i = 1, . . . , n|ñ(t) = n]
⎛ ⎞
exactly one arrival in each [ti , ti + hi ]
P⎝ ⎠
i = 1, 2, . . . , n, and no arrival elsewhere in [0, t]
=
P [ñ(t) = n]
(e−λh1 λh1 )(e−λh2 λh2 ) . . . (e−λhn λhn )(e−λ(t−h1 −h2 ...−hn ) )
=
e−λt (λt)n /n!
n!(h1 h2 h3 . . . hn )
=
tn
· P [ti < S̃i < ti + hi , i = 1, . . . , n|ñ(t) = n] n!
· · = n
h 1 h 2 . . . hn t

Taking lim ( ), then
hi →0,i=1,...,n
n!
fS̃1 ,S̃2 ,...,S̃n |ñ(t) (t1 , t2 , . . . , tn |n) = n , 0 < t1 < t2 < . . . < tn .
t

Example (see Ref [Ross], Ex. 2.3(A) p.68). Suppose that travellers arrive
at a train depot in accordance with a Poisson process with rate λ. If
the train departs at time t, what is the expected sum of the
ñ(t)
waiting times of travellers arriving in (0, t)? That is, E[ i=1 (t− S̃i )] =?
.......... t
~ ~ ~ ~~
S1 S2 S3 S n (t )

Answer. Conditioning on ñ(t) = n yields
ñ(t)

n
E[ (t − S̃i )|ñ(t) = n] = nt − E[ S̃i ]
i=1 i=1

n
= nt − E[ ũ(i) ] (by the theorem)
i=1
n
n
n
= nt − E[ ũi ] (· · · ũ(i) = ũi )
i=1 i=1 i=1
t nt t
= nt − · n = (· · · E[ũi ] = )
2 2 2
ñ(t)

To find E[ (t − S̃i )], we should take another expectation
i=1
ñ(t)
t λt2
· E[ (t − S̃i )] = · E[ñ(t)] =
· ·
i=1
2 2
=λt

Superposition of Independent Poisson Processes
Theorem. Superposition of independent Poisson Processes

N
(λi , i = 1, . . . , N ), is also a Poisson process with rate λi .
1
Poisson λ1
Poisson λ2
Poisson
N
rate = ∑ λi
Poisson λN 1
<Homework> Prove the theorem (note that a Poisson process must

satisfy Definitions 1 or 2).

Decomposition of a Poisson Process
Theorem.
• Given a Poisson process N = {ñ(t), t ≥ 0};
• If ñi (t) represents the number of type-i events that occur by time
t, i = 1, 2;
• Arrival occurring at time s is a type-1 arrival with probability p(s),
and type-2 arrival with probability 1 − p(s)
⇓then
• ñ1 , ñ2 are independent,

• ñ1 (t) ∼ P (k; λtp), and
t
1
• ñ2 (t) ∼ P (k; λt(1 − p)), where p = p(s)ds
t 0

p(s) Poisson ( k ; λ ∫ p ( s )ds )

Poisson λ
1-p(s) Poisson ( k ; λ ∫ [1 − p( s )]ds )
special case: If p(s) = p is constant, then
p Poisson rate λp
Poisson λ
1-p Poisson rate λ (1 − p )

Proof. It is to prove that, for fixed time t,
P [ñ1 (t) = n, ñ2 (t) = m] = P [ñ1 (t) = n]P [ñ2 (t) = m]

e−λpt (λpt)n e−λ(1−p)t [λ(1 − p)t]m
= ·
n! m!
...........................................................................
P [ñ1 (t) = n, ñ2 (t) = m]

∞

= P [ñ1 (t) = n, ñ2 (t) = m|ñ1 (t) + ñ2 (t) = k] · P [ñ1 (t) + ñ2 (t) = k]
k=0
= P [ñ1 (t) = n, ñ2 (t) = m|ñ1 (t) + ñ2 (t) = n + m] · P [ñ1 (t) + ñ2 (t) = n + m]

• From the “condition distribution of the arrival times”, any event
occurs at some time that is uniformly distributed, and is independent
of other events.
• Consider that only one arrival occurs in the interval [0, t]:
P [type - 1 arrival|ñ(t) = 1]
t
= P [type - 1 arrival|arrival time S̃1 = s, ñ(t) = 1]
0
×fS̃1 |ñ(t) (s|ñ(t) = 1)ds
t t
1 1
= P (s) · ds = P (s)ds = p
0 t t 0

· P [ñ1 (t) = n, ñ2 (t) = m]

· ·
= P [ñ1 (t) = n, ñ2 (t) = m|ñ1 (t) + ñ2 (t) = n + m] · P [ñ1 (t) + ñ2 (t) = n + m]
⎛ ⎞
n+m e −λt (λt)n+m
= ⎝ ⎠ pn (1 − p)m ·
n (n + m)!
−λt
(n + m)! n m e (λt)n+m
= p (1 − p) ·
n!m! (n + m)!
e−λpt (λpt)n e−λ(1−p)t [λ(1 − p)t]m
= ·
n! m!

Example (An Infinite Server Queue, textbook [Ross]).
G
Poisson λ departure
• Gs̃ (t) = P (S̃ ≤ t), where S̃ = service time

• Gs̃ (t) is independent of each other and of the arrival process
• ñ1 (t): the number of customers which have left before t;
• ñ2 (t): the number of customers which are still in the system at
time t;
⇒ ñ1 (t) ∼? and ñ2 (t) ∼?
Answer.
ñ1 (t): the number of type-1 customers
ñ2 (t): the number of type-2 customers
type-1: P (s) = P (finish before t)

= P (S̃ ≤ t − s) = Gs̃ (t − s)
type-2: 1 − P (s) = Ḡs̃ (t − s)

· 1 t
· · ñ1 (t) ∼ P k; λt · Gs̃ (t − s)ds
t 0

1 t
ñ2 (t) ∼ P k; λt · Ḡs̃ (t − s)ds
t 0

t
· 1
· · E[ñ1 (t)] = λt · G(t − s)ds
t 0
0 t−s=y
= λ G(y)(−dy)
t s=t−y
t
= λ G(y)dy
0
As t → ∞, we have
t
lim E[ñ2 (t)] = λ Ḡ(y)dy = λE[S̃] (Little’s formula)
t→∞ 0

Non-homogeneous Poisson Processes
• The counting process N = {ñ(t), t ≥ 0} is said to be a non-stationary
or non-homogeneous Poisson Process with time-varying intensity
function λ(t), t ≥ 0, if:
1. ñ(0) = 0
2. N has independent increments
3. P [ñ(t + h) − ñ(t) ≥ 2] = o(h)
4. P [ñ(t + h) − ñ(t) = 1] = λ(t) · h + o(h)
t
• Define “integrated intensity function” m(t) = λ(t )dt .
0
Theorem.
e−[m(t+s)−m(t)] [m(t + s) − m(t)]n
P [ñ(t + s) − ñ(t) = n] =
n!
Proof. < Homework >.
Example. The “output process” of the M/G/∞ queue is a
non-homogeneous Poisson process having intensity function
λ(t) = λG(t), where G is the service distribution.
Hint. Let D(s, s + r) denote the number of service completions in the
interval (s, s + r] in (0, t]. If we can show that
s+r
• D(s, s + r) follows a Poisson distribution with mean λ s G(y)dy,
and
• the numbers of service completions in disjoint intervals are
independent,
then we are finished by definition of a non-homogeneous Poisson
process.

Answer.
• An arrival at time y is called a type-1 arrival if its service
completion occurs in (s, s + r].
• Consider three cases to find the probability P (y) that an arrival at
time y is a type-1 arrival:
Case 1 Case 2 Case 3
s s+r t
y y y
– Case 1: y ≤ s.
P (y) = P {s − y < S̃ < s + r − y} = G(s + r − y) − G(s − y)
– Case 2: s < y ≤ s + r.
P (y) = P {S̃ < s + r − y} = G(s + r − y)

– Case 3: s + r < y ≤ t.
P (y) = 0
• Based on the decomposition property of a Poisson process, we
conclude that D(s, s + r) follows a Poisson distribution with mean
t
λpt, where p = (1/t) 0 P (y)dy.
t s s+r
P (y)dy = [G(s + r − y) − G(s − y)]dy + G(s + r − y)dy
0 0 s
t
+ (0)dy
s+r
s+r s
= G(s + r − y)dy − G(s − y)dy
0 0
s+r s s+r
= G(z)dz − G(z)dz = G(z)dz
0 0 s

• Because of
– the independent increment assumption of the Poisson arrival

process, and
– the fact that there are always servers available for arrivals,
⇒ the departure process has independent increments

Compound Poisson Processes
• A stochastic process {x̃(t), t ≥ 0} is said to be a compound Poisson
process if
– it can be represented as
ñ(t)

x̃(t) = ỹi , t≥0
i=1
– {ñ(t), t ≥ 0} is a Poisson process

– {ỹi , i ≥ 1} is a family of independent and identically distributed
random variables which are also independent of {ñ(t), t ≥ 0}
• The random variable x̃(t) is said to be a compound Poisson random
variable.
• E[x̃(t)] = λtE[ỹi ] and V ar[x̃(t)] = λtE[ỹi2 ].

Compound Poisson Processes
• Example (Batch Arrival Process). Consider a parallel-processing
system where each job arrival consists of a possibly random number
of tasks. Then we can model the arrival process as a compound
Poisson process, which is also called a batch arrival process.
• Let ỹi be a random variable that denotes the number of tasks
comprising a job. We derive the probability generating function
Px̃(t) (z) as follows:

Px̃(t) (z) = E z x̃(t) = E E z x̃(t) |ñ(t) = E E z ỹ1 +···+ỹñ(t) |ñ(t)

ỹ1 +···+ỹñ(t)
= E E z (by independence of ñ(t) and {ỹi })

= E E z ỹ1 · · · E z ỹñ(t) (by independence of ỹ1 , · · · , ỹñ(t) )

ñ(t)
= E (Pỹ (z)) = Pñ(t) (Pỹ (z))

Modulated Poisson Processes
• Assume that there are two states, 0 and 1, for a “modulating process.”
0 1
• When the state of the modulating process equals 0 then the arrive rate
of customers is given by λ0 , and when it equals 1 then the arrival rate
is λ1 .
• The residence time in a particular modulating state is exponentially
distributed with parameter μ and, after expiration of this time, the
modulating process changes state.
• The initial state of the modulating process is randomly selected and is
equally likely to be state 0 or 1.
• For a given period of time (0, t), let Υ be a random variable that
indicates the total amount of time that the modulating process has
been in state 0. Let x̃(t) be the number of arrivals in (0, t).
• Then, given Υ, the value of x̃(t) is distributed as a non-homogeneous
Poisson process and thus
(λ0 τ + λ1 (t − τ ))n e−(λ0 τ +λ1 (t−τ ))
P [x̃(t) = n|Υ = τ ] =
n!
• As μ → 0, the probability that the modulating process makes no
transitions within t seconds converges to 1, and we expect for this case
that
n −λ0 t n −λ1 t
1 (λ0 t) e (λ1 t) e
P [x̃(t) = n] = +
2 n! n!

• As μ → ∞, then the modulating process makes an infinite number of
transitions within t seconds, and we expect for this case that
(βt)n e−βt λ0 + λ1
P [x̃(t) = n] = , where β =
n! 2
• Example (Modeling Voice).
– A basic feature of speech is that it comprises an alternation of
silent periods and non-silent periods.
– The arrival rate of packets during a talk spurt period is Poisson
with rate λ1 and silent periods produce a Poisson rate with λ0 ≈ 0.
– The duration of times for talk and silent periods are exponentially
distributed with parameters μ1 and μ0 , respectively.
⇒ The model of the arrival stream of packets is given by a modulated
Poisson process.
Poisson Arrivals See Time Averages (PASTA)
• PASTA says: as t → ∞
Fraction of arrivals who see the system in a given state

upon arrival (arrival average)
= Fraction of time the system is in a given state (time average)
= The system is in the given state at any random time
after being steady
• Counter-example (textbook [Kao]: Example 2.7.1)

1/2 1/2
service time = 1/2

inter-arrival time = 1
1 1
0 1 2 3 4 5

– Arrival average that an arrival will see an idle system = 1
– Time average of system being idle = 1/2
• Mathematically,
– Let X = {x̃(t), t ≥ 0} be a stochastic process with state space S,
and B ⊂ S
– Define an indicator random variable
⎧
⎨ 1, if x̃(t) ∈ B
ũ(t) =
⎩ 0, otherwise
– Let N = {ñ(t), t ≥ 0} be a Poisson process with rate λ denoting the

arrival process
then,

t t
ũ(s)dñ(s) ũ(s)ds
0 0
lim = lim
t→∞ ñ(t) t→∞ t
(arrival average) (time average)
• Condition – For PASTA to hold, we need the lack of anticipation

assumption (LAA): for each t ≥ 0,
– the arrival process {ñ(t + u) − ñ(t), u ≥ 0} is independent of
{x̃(s), 0 ≤ s ≤ t} and {ñ(s), 0 ≤ s ≤ t}.
• Application:
– To find the waiting time distribution of any arriving customer
– Given: P[system is idle] = 1 − ρ; P[system is busy] = ρ

Case 1: system is idle
Poisson
Case 2: system is busy

Poisson
⇒ P (w̃ ≤ t) = P (w̃ ≤ t|idle) · P (idle upon arrival)

+ P (w̃ ≤ t|busy) · P (busy upon arrival)

Memoryless Property of the Exponential Distribution
• A random variable x̃ is said to be without memory, or memoryless, if
P [x̃ > s + t|x̃ > t] = P [x̃ > s] for all s, t ≥ 0 (3)
• The condition in Equation (3) is equivalent to

P [x̃ > s + t, x̃ > t]
= P [x̃ > s]
P [x̃ > t]
or
P [x̃ > s + t] = P [x̃ > s]P [x̃ > t] (4)
• Since Equation (4) is satisfied when x̃ is exponentially distributed (for
e−λ(s+t) = e−λs e−λt ), it follows that exponential random variable are
memoryless.
• Not only is the exponential distribution “memoryless,” but it is the
unique continuous distribution possessing this property.
Comparison of Two Exponential Random Variables
Suppose that x̃1 and x̃2 are independent exponential random variables
with respective means 1/λ1 and 1/λ2 . What is P [x̃1 < x̃2 ]?
∞
P [x̃1 < x̃2 ] = P [x̃1 < x̃2 |x̃1 = x]λ1 e−λ1 x dx
0∞
= P [x < x̃2 ]λ1 e−λ1 x dx
0∞
= e−λ2 x λ1 e−λ1 x dx
0∞
= λ1 e−(λ1 +λ2 )x dx
0
λ1
=
λ1 + λ2

Minimum of Exponential Random Variables
Suppose that x̃1 , x̃2 , · · · , x̃n are independent exponential random variables,
with x̃i having rate μi , i = 1, · · · , n. It turns out that the smallest of the x̃i
is exponential with a rate equal to the sum of the μi .
P [min(x̃1 , x̃2 , · · · , x̃n ) > x] = P [x̃i > x for each i = 1, · · · , n]

n
= P [x̃i > x] (by independence)
i=1
n
= e−μi x
i=1
n

= exp − μi x
i=1
How about max(x̃1 , x̃2 , · · · , x̃n )? (exercise)

Poisson PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Poisson PDF

Uploaded by

Copyright:

Available Formats

Chapter 2.

Prof. Ai-Chun Pang

Prof. Ai-Chun Pang, NTU 1

(i) nth arrival epoch S̃n is

Prof. Ai-Chun Pang, NTU 2

Arrival Process: X = {x̃i , i = 1, 2, . . .}; x̃i ’s can be any

Renewal Process: X = {x̃i , i = 1, 2, . . .}; x̃i ’s are i.i.d.

Poisson Process: X = {x̃i , i = 1, 2, . . .}; x̃i ’s are iid exponential distributed

Prof. Ai-Chun Pang, NTU 3

Prof. Ai-Chun Pang, NTU 4

P [ñ(t) − ñ(s) = k1 |ñ(r) = k2 , r ≤ s < t] = P [ñ(t) − ñ(s) = k1 ]

3. Stationary increments relaxed ⇒ Non-homogeneous Poisson Process

P [ñ(t + s) − ñ(t) = k] = P [ñ(l + s) − ñ(l) = k]

4. Single arrival relaxed ⇒ Compound Poisson Process

Prof. Ai-Chun Pang, NTU 5

Prof. Ai-Chun Pang, NTU 6

Prof. Ai-Chun Pang, NTU 7

From (1) and (2), we obtain that

g(t + h) = g(t)(1 − λh + e−u λh) + o(h)

Prof. Ai-Chun Pang, NTU 8

Since g(t) is also the Laplace transform of ñ(t), ñ(t) is a Poisson r. v.

Prof. Ai-Chun Pang, NTU 9

Prof. Ai-Chun Pang, NTU 11

x̃2 : P (x̃2 > t|x̃1 = s)

Prof. Ai-Chun Pang, NTU 14

fx̃(1) ,x̃(2) ,...,x̃(n) (x1 , x2 , . . . , xn ) = n!f (x1 )f (x2 ) . . . f (xn ),

where x1 < x2 < . . . < xn (check the textbook [Ross]).

Prof. Ai-Chun Pang, NTU 16

Prof. Ai-Chun Pang, NTU 17

Prof. Ai-Chun Pang, NTU 18

Prof. Ai-Chun Pang, NTU 19

<Homework> Prove the theorem (note that a Poisson process must

Prof. Ai-Chun Pang, NTU 20

• ñ1 , ñ2 are independent,

Prof. Ai-Chun Pang, NTU 21

p(s) Poisson ( k ; λ ∫ p ( s )ds )

1-p(s) Poisson ( k ; λ ∫ [1 − p( s )]ds )

special case: If p(s) = p is constant, then

1-p Poisson rate λ (1 − p )

Prof. Ai-Chun Pang, NTU 22

P [ñ1 (t) = n, ñ2 (t) = m] = P [ñ1 (t) = n]P [ñ2 (t) = m]

P [ñ1 (t) = n, ñ2 (t) = m]

Prof. Ai-Chun Pang, NTU 23

Prof. Ai-Chun Pang, NTU 24

· P [ñ1 (t) = n, ñ2 (t) = m]

Prof. Ai-Chun Pang, NTU 25

• Gs̃ (t) = P (S̃ ≤ t), where S̃ = service time

type-1: P (s) = P (ﬁnish before t)

Prof. Ai-Chun Pang, NTU 27

Prof. Ai-Chun Pang, NTU 28

Prof. Ai-Chun Pang, NTU 30

P (y) = P {s − y < S̃ < s + r − y} = G(s + r − y) − G(s − y)

P (y) = P {S̃ < s + r − y} = G(s + r − y)

Prof. Ai-Chun Pang, NTU 32

– the independent increment assumption of the Poisson arrival

⇒ the departure process has independent increments

Prof. Ai-Chun Pang, NTU 33

– {ñ(t), t ≥ 0} is a Poisson process

Prof. Ai-Chun Pang, NTU 34

Prof. Ai-Chun Pang, NTU 35

Prof. Ai-Chun Pang, NTU 37

Fraction of arrivals who see the system in a given state

• Counter-example (textbook [Kao]: Example 2.7.1)