
Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
Material Type: Assignment; Professor: Hayat; Class: Probability Theory and Stochastic Processes; Subject: Electrical & Computer Engineer; University: University of New Mexico; Term: Fall 2008;
Typology: Assignments
1 / 1
This page cannot be seen from the preview
Don't miss anything!

ECE 541, Probability and Stochastic Processes Fall 2008 Homework Assignment 9 Due date: Thursday, Dec. 4, 2008
Problem 1. Consider the Gauss-Markov model
Bn = ρBn− 1 + W (n), n > 0 , 0 < |ρ| < 1 ,
where the initial state B 0 is a zero-mean Gaussian random variable with variance σ^2 B 0. Assume that W (1), W (2),... , are iid, Gaussian, zero mean, and independent of B 0. Let σ^2 W denote the variance of W (n). Derive a difference equation characterizing the autocorrelation function RB (m+n, n) =^4 E[Bn+mBn]. Solve the equation and comment on your result. Be sure to specify any initial conditions needed. Hint: First find a recursion for RB (n, n), and then find a recursion for RB (m+n, n) (in the variable m) for each fixed n.
Problem 2. (From Hoel, Port, and Stone) Consider an irreducible birth and death chain on the nonnegative integers with
qi/pi = ( i i + 1
Show that the chain is transient.
Problem 3. (From Hoel, Port, and Stone) Consider a Markov chain on the nonnegative integers having transition probabilities given by pi,i+1 = p and pi, 0 = 1 − p, where 0 < p < 1. Show that this chain has a unique stationary distribution π and find it.
Problem 4. There are n nodes in a certain cluster of computers that are initially operating at time k = 0. During each unit of time, each node may fail with probability p (independently of previous times and other processors) and remains dysfunctional from that point and on. For k = 0, 1 , 2 ,.. ., let Nk denote the number of processors that are up at time k.
a) Argue that Nk is a Markov chain and derive its transition probability matrix IP.
b) Use conditional expectations to derive a recursion for E[Nk] and solve for E[Nk], k ≥ 0.
c) Determine a recursion for the second moment of Nk, specify the initial condition, and solve for the second moment.
d) Calculate P{Nk = 0}.
e) Let T be the first time when all nodes have failed, that is, T = min{i : Ni = 0}. Show that
P{T = k} =
( 1 + [p − 1](1 − p)k−^1
)n −
( 1 − (1 − p)k−^1
)n .
Hint: {T = k} = {Nk− 1 > 0 } ∩ {Nk = 0}.
f) Is the chain irreducible? Justify your answer.
g) Which sates are recurrent and which states are transient? Justify your answer.