WEBVTT
00:00:00.260 --> 00:00:03.800
In this lecture, we introduce
the Poisson process, which is
00:00:03.800 --> 00:00:07.320
a continuous time analog of
the Bernoulli process.
00:00:07.320 --> 00:00:10.690
One way of thinking about
it is as follows.
00:00:10.690 --> 00:00:14.610
Time is continuous, but
conceptually we divide it into
00:00:14.610 --> 00:00:16.640
a very large number of slots.
00:00:16.640 --> 00:00:19.090
And during each slot,
we have a tiny
00:00:19.090 --> 00:00:21.240
probability of an arrival.
00:00:21.240 --> 00:00:23.570
This probability is proportional
to the
00:00:23.570 --> 00:00:25.610
length of the slot.
00:00:25.610 --> 00:00:28.150
Furthermore, we have an
independence assumption for
00:00:28.150 --> 00:00:29.400
the different slots.
00:00:31.860 --> 00:00:35.560
The Poisson process is a very
elegant model of arrival
00:00:35.560 --> 00:00:37.870
processes in continuous time.
00:00:37.870 --> 00:00:40.850
It models many real-world
phenomena.
00:00:40.850 --> 00:00:44.390
And it also has a very clean
mathematical structure that
00:00:44.390 --> 00:00:46.970
allows us to calculate
practically
00:00:46.970 --> 00:00:50.300
every quantity of interest.
00:00:50.300 --> 00:00:53.820
Our development will parallel
our analysis of
00:00:53.820 --> 00:00:55.640
the Bernoulli process.
00:00:55.640 --> 00:00:59.830
For example, we will find the
PMF of the number of arrivals
00:00:59.830 --> 00:01:03.030
during a time interval
and the PDF of the
00:01:03.030 --> 00:01:05.930
time of the kth arrival.
00:01:05.930 --> 00:01:08.950
We will discuss the
memorylessness properties of
00:01:08.950 --> 00:01:10.900
the Poisson process.
00:01:10.900 --> 00:01:13.539
Similar to the case of the
Bernoulli process, this is
00:01:13.539 --> 00:01:16.450
just a consequence of the
independence assumptions that
00:01:16.450 --> 00:01:18.580
we are making.
00:01:18.580 --> 00:01:21.530
We will then exploit these
independence properties to
00:01:21.530 --> 00:01:25.580
argue that the interarrival
times are independent
00:01:25.580 --> 00:01:27.950
exponential random variables.
00:01:27.950 --> 00:01:30.480
And we will conclude with
a comprehensive example.