Durrett 3.8 Counter processes. Suppose that arrivals at a counter come at times of a Poisson process with rate . An arriving particle that finds the counter free gets registered and then locks the counter for an amount of time . Particles that arrive while the counter is locked are not counted and have no effect. (a) Find the limiting probability that the counter is locked at time . (b) Compute the limiting fraction of particles that get registered.
Durrett 3.21 The city of Ithaca, NY, allows for two-hour parking in all downtown spaces. Methodical parking officials patrol the downtown area, passing the same point every 2h. When an official encounters a car, he marks it with chalk. If the car is still there 2h later, a ticket is written. Suppose that you park your car for a random amount of time that is distributed uniformly over hours. What is the probability you will get a ticket?
Durrett 4.2 A small computer store has room to display up to three computers for sale. Customers come at times of a Poisson arrival process at a rate of 2 per week. If a customer arrives and a computer is available, the customer will buy it. When the store has only one computer left they place an order for two more computers. The order takes an exponentially distributed amount of time to be completed, with a rate of 1 per week. Of course, while the store is waiting for delivery, sales may reduce inventory from 1 to 0. (a) Write a matrix for the transition rates and solve to find the stationary distribution. (b) At what rate does the store makes sales?
Durrett 4.9 A hemoglobin molecule can carry one oxygen or one carbon monoxide molecule. Suppose that the two types of gases arrive at rates 1 and 2 and attach for an exponential amount of time with rates 3 and 4, respectively. Formulate a Markov chain model with state space where denotes an attached oxygen molecule, denotes an attached carbon monoxide molecule, and 0 a free hemoglobin molecule. Find the long-run fraction of time the hemoglobin molecule is in each of its three states.
Autocorrelation of a Stochastic Process (DUE WED APR 23)
Consider a sequence of random variables that take the values 5 and 1 with asymptotic frequencies and respectively. We will consider some stochastic processes that produce such sequences and compute the one-step autocorrelation of these sequences.
The autocovariance function of a sequence is defined to be
- iid model. Suppose that the sequence is iid with and . Compute for .
- Deterministic switching. Consider the sequence . Compute the “empirical” autocovariance function
- Markov chain model. Suppose that is a Markov chain that takes the values 1 and 5 whose transition probabilities are given by the transition matrix
- For what value of will satisfy the asymptotic frequencies given above?
- Compute the one-step autocovariance function for this process assuming that the initial condition is drawn from the stationary distribution.