HWK3_sol - IEOR E4701 Assignment 3 Solutions Summer 2011...

Info iconThis preview shows pages 1–3. Sign up to view the full content.

View Full Document Right Arrow Icon

Info iconThis preview has intentionally blurred sections. Sign up to view the full version.

View Full DocumentRight Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: IEOR E4701 Assignment 3 - Solutions Summer 2011 Question 1 Consider the Markov chain whose transition probabilities are given by P = 0 1 2 1 2 1 2 1 2 1 3 2 3 3 4 1 4 a) Is the chain irreducible? why? b) Is the chain positive recurrent? why? c) Compute the equilibrium distribution in two di erent ways. Answer a) Notice that p ( i,j ) > for every i 6 = j . This means that every state in the states-space is reachable, independent of the starting point of the Markov chain. Therefore, the Markov chain is irreducible. b) Following the what we obseved in class, that every irreducible Markov chain is positively recurrent, we may conclude that the given Markov chain is positively recurrent. c) Since the given Markov chain is irreducible and positive recurrent we know that the equilibrium (steady- state) distribution, π , is unique. We will nd π using two di rent methods; the rst one via the formula π T = π T P and the second one through the connection π ( i ) = 1 E i τ i where τ i = min { n ≥ 1 : X n = i } , for i = 0 , 1 , 2 . Method 1: π T = π T P ( π (0) ,π (1) ,π (2)) = ( π (0) ,π (1) ,π (2)) 1 2 1 2 1 3 2 3 3 4 1 4 which leads to the following system of linear equations π (0) = 1 3 π (1) + 3 4 π (2) π (1) = 1 2 π (0) + 1 4 π (2) π (2) = 1 2 π (0) + 2 3 π (1) = ⇒ π = 4 11 , 3 11 , 4 11 . 1 IEOR E4701 Assignment 3 - Solutions Summer 2011 Method 2: π ( i ) = 1 E i τ i where τ i = min { n ≥ 1 : X n = i } , for i = 0 , 1 , 2 . For i = 0 . E τ = E ( τ | X = 0) = p (0 , 1) E ( τ | X = 0 , X 1 = 1) + p (0 , 2) E ( τ | X = 0 , X 1 = 2) = Markov property 1 2 E ( T | X = 1) + 1 2 E ( T | X = 2) where T = min { n ≥ 0 : X n = 0 } . Let g ( i ) = E ( T | X = i ) . Then, E τ = 1 2 g (1) + 1 2 g (2) . To nd g ( i ) we need to set up the following system of linear equations: For i = 0 : g (0) = 0 For i = 1 : g (1) = E ( T | X = 1) = E ( T | X = 1 , X 1 = 0) p (1 , 0) + E ( T | X = 1 , X 1 = 2) p (1 , 2) + 1 = Markov property 1 · 1 3 + g (2) · 2 3 + 1 = 2 3 g (2) + 4 3 For i = 2 : g (2) = E ( T | X = 2) = E ( T | X = 2 , X 1 = 0) p (2 , 0) + E ( T | X = 2 , X 1 = 1) p (2 , 1) + 1...
View Full Document

This note was uploaded on 10/30/2011 for the course IEOR 4701 taught by Professor Karlsigma during the Summer '10 term at Columbia.

Page1 / 9

HWK3_sol - IEOR E4701 Assignment 3 Solutions Summer 2011...

This preview shows document pages 1 - 3. Sign up to view the full document.

View Full Document Right Arrow Icon
Ask a homework question - tutors are online