This preview shows pages 1–3. Sign up to view the full content.
This preview has intentionally blurred sections. Sign up to view the full version.
View Full Document
Unformatted text preview: STAT 333 Assignment 3 SOLUTIONS 1. Consider a discretetime Markov Chain with state space S = {0, , 7} and transition matrix P = 1/2 1/4 1/4 1/4 3/4 1/3 2/3 1/5 1/5 1/5 1/5 1/5 1/6 1/3 1/6 1/6 1/6 1 1/4 3/4 2/3 1/3 a. Determine the classes of this chain, which are open or closed, write P in simplified form, and find the period of each closed class. C 1 = {0, 1, 6} closed so positive recurrent aperiodic C 2 = {2, 7} closed so positive recurrent aperiodic C 3 = {3} open so transient C 4 = {4} open so transient C 5 = {5} closed so positive recurrent aperiodic P sim = 1/2 1/4 1/4 1/4 3/4 1/4 3/4 1/3 2/3 2/3 1/3 1 1/5 1/5 1/5 1/5 1/5 1/6 1/6 1/6 1/6 1/3 with new order of states {0, 1, 6, 2, 7, 5, 3, 4} b. Find the equilibrium distribution corresponding to each closed class. Write down the general form of all stationary distributions for this chain. Within C 2 , we can see the submatrix is doubly stochastic so equilibrium is ( , ) Within C 5 , there is only one state, so equilibrium is (1) Within C 1 , we get + 1 + 6 = 0 = 1 + 1 + 6 = 1 solving gives (1/3, 1/12, 7/12) The general form is: (1/3, 1/12, 0, 0, 0, 0, 7/12, 0) + (0, 0, , 0, 0, 0, 0, ) + (1 )(0, 0, 0, 0, 0, 1, 0, 0) where , , and (1 ) are between 0 and 1. c. Find the absorption probabilities from each transient state into each closed class. A 4, C1 = P 4, C1 + P 4, 4 *A 4, C1 = (1/6 + 1/6) + 1/3 *A 4, C1 So A 4, C1 = A 4, C2 = P 4, C2 + P 4, 4 *A 4, C2 = 1/6 + 1/3 *A 4, C2 So A 4, C1 = So A 4, C5 = 1 A 4, C1 A 4, C2 = A 3, C1 = P 3, C1 + P 3, 3 *A 3, C1 + P 3, 4 *A 4, C1 = 1/5 + 1/5 *A 3, C1 + 1/5*1/2 So A 3, C1 = 3/8 A 3, C2 = P 3, C2 + P 3, 3 *A 3, C2 + P 3, 4 *A 4, C2 = 1/5 + 1/5 *A 3, C1 + 1/5*1/4 So A 3, C1 = 5/16 So A 3, C5 = 1 A 3, C1 A 3, C2 = 5/16 d. Describe the longrun behaviour of the chain if X = 0. Do the same if X = 3. If X0 = 0, we are starting in closed class C1, so the longrun behaviour is (1/2, 1/12, 0, 0, 0, 0, 7/12, 0) If X0 = 3, the longrun behavior is (3/8)(1/3, 1/12, 0, 0, 0, 0, 7/12, 0)+(5/16)(0, 0, , 0, 0, 0, 0, )+(5/16)(0, 0, 0, 0, 0, 1,0,0) 2. Consider a discretetime Markov Chain with five transient states (1, 2, 3, 4, and 5) and two recurrent states (6 and 7). The chain is equally likely to start in any of the transient states....
View
Full
Document
This note was uploaded on 12/12/2010 for the course STAT 333 taught by Professor Chisholm during the Spring '08 term at Waterloo.
 Spring '08
 Chisholm

Click to edit the document details