This preview shows page 1. Sign up to view the full content.
Unformatted text preview: and letting t ! 1, we get (since, derivative of a constant is zero), 0 = Q
with the requirement that probabilities sum up to 1 1 + 2 + ::: = 1:
CS 522, v 0.94, d.medhi, W’99 8 if e is a column vector of the same dimension as with each component being 1, we can write this in
compact form as e = 1.
Note that 0 = Q e = 1 6 is a linear system of equation which has a unique solution (under certain conditions).
To illustrate, suppose that we have a system that takes three values (0, 1, 2). The probabilities are: = 0 1 2
and Q = 0 is 2 ,q q
0
01
0 1 2 4 q10 ,q1
q20 and e = 1 is q21 3 q02
q12 5 =
,q2 000 0 + 1 + 2 = 1: If we know q ’s, then we can solve for . 5. BirthandDeath Process
This is a special case of continuoustime Markov chain. You can only transition to a neighbor (one
step away) or stay where you are in an inﬁnitesimal time period, and the rate is statedependent, one deﬁned
for arrival and the other for departure. =
,1 =
q =0 qi i+1 i qi i i ij i=0 1 2 :::
i = 1 2 :::
for j 6= i + 1 i , 1 7
i: Due to this situation and the requirement that the row sum to be zero (from (3)), we have qi = ,qii = i + i for i not in boundary state : Now, what d...
View
Full
Document
 Fall '08
 Staff
 Computer Networks

Click to edit the document details