The probability of more than one event in the interval [, is a function of a higher
order term of 2 t for 2t 0.
tt 2 t
Events in disjoint intervals are mutually independent.
The Poisson process may be defined completely by its intensity ( ) t :
0
() lim^1
t 2t 2 t P (one event in [,tt 2 t ) (2.51)
If ( ) t is constant in time the Poisson process is said to be homogeneous, otherwise it is
inhomogeneous.
In general the probability of n events in the interval [0, [t of a Poisson process with intensity
()t can be shown to be given as:
^0
0
()
exp ( )
!
t n
t
n
d
Pt d
n
3 3
3 3
!
"#!
$%"
$%
# (2.52)
with mean value ENt () and variance Var N t ():
0
() () ( )
t
ENtVarNt 3 3d (2.53)
The probability of no events in the interval [0 i.e. is especially interesting considering
reliability problems. This probability may be determined directly from Equation
,[t Pt 0 ()
(2.52) as:
0
0
exp ( )
t
Pt 3 3d
!
"
$%
# (2.54)
implying that the time till and between events are Exponential distributed.
From Equation (2.54) the cumulative distribution function of the waiting time till the first
event , i.e. 1 may be straightforwardly derived. Recognising that the probability of
is there is:
T 1
t
FtT() 1
)
t
T 1 & Pt 0 (
1
1 1
0
FtT 1-exp 3 3( )d
!
""
$%
## (2.55)
Consider now the sum of independent and exponential distributed n waiting times Tgiven as:
TTT 12 ... Tn (2.56)
It can be shown (see the lecture notes Basic Theory of Probability and Statistics in Civil
Engineering, Faber, 2006) that T is Gamma distributed:
( )(-1)exp( )
()
(1)!
n
T
ft t
n
t