This preview shows page 1. Sign up to view the full content.
Unformatted text preview: = πi−1 =
3 i≥1 i≥0 Thus,
π0 =
πi = 1
6
��
5 2i
, for i ≥ 1.
12 3 b) The transition rates are qij = Pij vi = Pij 2i . Therefore, q01 = 1, qi,i+1 = (2/5)2i and
qi,i−1 = (3/5)2i for i = 1, 2, ....The steady state probabilities pi for the Markov process are
proportional to πi /vi = (5/12)(1/3)i for i ≥ 1 and π0 /v0 = 1/6. Normalizing so that pi ’s
sum to one,we get: 6 p0 =
pi = 4
9
��
10 1 i
, for i ≥ 1
93 The pi ’s decay faster than πi ’s. This is because the transition rate vi increases with i.
So, the mean time until a transition from the current state decreases with i.
c) Since vi is growing unboundedly with increasing i and Pi,i+1 and Pi,i−1 is constant for
all i > 0, qij = Pij vi is also growing unboundedly with i for j = i + 1 and j = i − 1. Thus
for any δ > 0, and for large enough n, the transit...
View
Full
Document
This document was uploaded on 03/19/2014 for the course EECS 6.262 at MIT.
 Spring '11
 RobertGallager

Click to edit the document details