WEBVTT

00:00:00.460 --> 00:00:03.970
In this lecture, we look in
some depth into various

00:00:03.970 --> 00:00:06.690
properties of Poisson
processes.

00:00:06.690 --> 00:00:10.030
These properties would be quite
hard to study if one

00:00:10.030 --> 00:00:12.390
were to proceed just
analytically by

00:00:12.390 --> 00:00:14.620
manipulating formulas.

00:00:14.620 --> 00:00:17.930
However, by using memorylessness
and our

00:00:17.930 --> 00:00:20.650
intuitive understanding of
what the Poisson process

00:00:20.650 --> 00:00:23.770
really is, they become
quite simple.

00:00:23.770 --> 00:00:26.550
And the mathematical
manipulations can be avoided

00:00:26.550 --> 00:00:29.690
almost entirely.

00:00:29.690 --> 00:00:33.150
We will start by arguing that
the sum of independent Poisson

00:00:33.150 --> 00:00:36.080
random variables is Poisson.

00:00:36.080 --> 00:00:40.250
But we will then establish the
much stronger statement that

00:00:40.250 --> 00:00:44.650
if we merge two independent
Poisson processes, we, again,

00:00:44.650 --> 00:00:47.500
obtain a Poisson process.

00:00:47.500 --> 00:00:50.250
We will see that we can exploit
this fact to solve

00:00:50.250 --> 00:00:54.130
some problems that would be
quite difficult otherwise.

00:00:54.130 --> 00:00:58.820
Once more, intuitive reasoning
is the key.

00:00:58.820 --> 00:01:02.040
Finally, we will spend some time
discussing a phenomenon

00:01:02.040 --> 00:01:06.160
that goes under the name
of random incidence.

00:01:06.160 --> 00:01:08.590
The Poisson process
has been running.

00:01:08.590 --> 00:01:10.890
You show up at a certain time.

00:01:10.890 --> 00:01:14.110
You look at the size of the
inter-arrival interval during

00:01:14.110 --> 00:01:15.940
which you show up.

00:01:15.940 --> 00:01:19.450
It turns out that this
inter-arrival interval that

00:01:19.450 --> 00:01:22.810
you get to observe is
not a typical one.

00:01:22.810 --> 00:01:25.260
It tends to be larger
than the typical

00:01:25.260 --> 00:01:27.940
inter-arrival interval.

00:01:27.940 --> 00:01:30.930
We will understand what exactly
is going on, build

00:01:30.930 --> 00:01:34.240
some intuition, and realize
that this is a general

00:01:34.240 --> 00:01:37.770
phenomenon that also shows up
in many other contexts.