This preview shows page 1. Sign up to view the full content.
Unformatted text preview: x + x2 = (1 x)2 i.e., a double root at x = 1. In general when µ = 1, the graph of is tangent
to the diagonal (x, x) at x = 1. This slows down the convergence of ⇢n to 1 so
that it no longer occurs exponentially fast.
In more advanced treatments, it is shown that if the o↵spring distribution
has mean 1 and variance 2 > 0, then
P1 (Xn > 0) ⇠ 2
n2 This is not easy even for the case of binary branching, so we refer to reader to
Section 1.9 of Athreya and Ney (1972) for a proof. We mention the result here
because it allows us to see that the expected time for the process to die out
P
n P1 (T0 > n) = 1. If we modify the branching process, so that p(0, 1) = 1
then in the modiﬁed process
If µ < 1, 0 is positive recurrent 59 1.11. CHAPTER SUMMARY
If µ = 1, 0 is null recurrent
If µ > 1, 0 is transient Our ﬁnal example gives an application of branching processes to queueing
theory.
Example 1.54. M/G/1 queue. We will not be able to explain the name
of this example until we consider Poisson processes in Chapter 2. However,
imagine a que...
View
Full
Document
This document was uploaded on 03/06/2014 for the course MATH 4740 at Cornell University (Engineering School).
 Spring '10
 DURRETT
 The Land

Click to edit the document details