Determine Cov[X(t), X(t+s)] when {X(t), t ⩾0] is a compound Poisson process.

Name: Problem 87, Chapter 5: Introductory to Probability Models
Uploaded: 2020-12-18T22:14:36-08:00
Duration: 13 min 11 s
Channel: Robin Corrigan
Description: Determine Cov[X(t), X(t+s)] when {X(t), t ⩾0] is a compound Poisson process.

Determine Cov[X(t), X(t+s)] when {X(t), t ⩾0] is a compound...

Key Concepts

Variance Calculation for Compound Processes

The variance of a compound Poisson process involves both the rate of the underlying Poisson process and the variance (or second moment) of the jump sizes. Specifically, when calculating the covariance between X(t) and X(t+s), the result is linked to the variance of X(t), which is governed by the product of the time parameter and the second moment of the jump distribution.

Covariance in Stochastic Processes

Covariance measures the degree to which two random variables vary together. In the context of a compound Poisson process, the covariance between values at two different times simplifies due to the independent increments property, often reducing to the variance of the process evaluated at the earlier time point.

Compound Poisson Process

A compound Poisson process is a stochastic process defined by the sum of a random number of independent and identically distributed jumps, where the number of jumps up to time t is itself a Poisson process. This process is particularly used for modeling random events that occur sporadically over time, such as in risk theory or queueing systems.

Independent Increments

This property implies that the increments of the process over non-overlapping time intervals are independent of each other. For a compound Poisson process, this greatly simplifies the analysis of covariance because the value of the process at a later time can be decomposed into the sum of its value at an earlier time and an independent increment.

Transcript

00:01 This question is about the poisson telegraphic process, which is a process with a random variable, n at t.

00:10 And the way the process works is that at time zero, n at t is assigned either plus one or minus one randomly with probability of 50 % each.

00:22 And then events occur according to a poisson process.

00:26 And each time an event occurs, the parity of n at t switches.

00:32 So for example, if n at t is plus 1 and an event occurs, then it switches to minus 1 and vice versa.

00:41 We are also told that the probability of an even number of events occurring in the time interval from 0 to t is 0 .5.

00:52 And so for part a, we are asked to explain why for time greater than 0, given this statement here, we are asked to explain why the following is true.

01:05 So if we look at the first of these two, it's explained why the conditional probability that n at t equals plus 1, given n at 0 is equal to plus 1, equals p.

01:21 And remember, p is the probability of an even number of events in time 0 to t.

01:30 So if we think of our time interval, let's say t is here.

01:41 And as events occur according to a post -one process, let's say an event happens here, we'll call this capital t sub 1 for the first event, t sub 2 for the second, and so on.

01:58 So we can say that ti is the ieth event in the interval from 0 to t.

02:17 So then if n at 0 is plus 1, then we have n at the time of the first event will be minus 1, and at the time of the second event plus 1, which means that n at t is positive only for even i so for an even number of events so that n at t will be plus one only if we have an even number of events so this means that the probability of n at t being plus one given that we start at n at zero is plus one is simply this probability here the probability of the statement being true which we know to be p, the probability of an even number of events in 0 to t.

04:34 And now for the second statement we can use the same logic.

04:40 So this time if n at 0 is equal to minus 1, that's the condition, then n at the time of the first event is equal to plus 1, and at the time of the second event is minus 1, and so on.

05:12 And therefore you can see that n at t is plus one only when the number of events on the interval is odd numbered.

05:27 I can say that if we start with n at 0 equal to minus 1, n at t is equal to plus 1, only if there is an odd number of events on the time interval from 0 to t.

06:03 So therefore we can say that the probability that n at t is equal to minus 1.

06:12 Or rather plus one, given that we start at n at zero, equal to minus one, is the truth of this statement.

06:30 And the probability of an odd number of events is 1 minus p.

06:34 Remember that p is the probability of an even number of events.

06:38 So therefore we have explained why these statements in part a are true.

06:50 So now for part b, we are asked to use our results from part a and the law of total probability to show that the probability that n at t is equal to plus 1 is equal to 0 .5, or all time greater than or equal to 0.

07:14 Using the law of total probability, the probability that n at t equals plus 1 is expressed as the probability that n at t equals plus 1, given that n at 0 equals 1.

07:40 Times the probability that n at 0 equals 1...

Please give Ace some feedback