Stochastic

Lemma 43 is a stationary distribution if and only if q

This preview shows page 1. Sign up to view the full content.

This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: ts into two at rate , so q (i, i + 1) = i. To ﬁnd the transition probability of the Yule process we will guess and verify that pt (1, j ) = e t (1 ) for j tj 1 e i.e., a geometric distribution with success probability e To explain the mean we note that d EX (t) = E X (t) dt j implies 1 t (4.13) and hence mean e t . E1 X (t) = e t . To check (4.13), we will use the forward equation (4.9) to conclude that if 1 then p0 (1, j ) = j pt (1, j ) + (j 1)pt (1, j 1) (4.14) t where pt (1, 0) = 0. The use of the forward equation here is dictated by the fact that we are only writing down formulas for pt (i, j ) when i = 1. To check the proposed formula for j = 1 we note that p0 (1, 1) = t e t = pt (1, 1) Things are not so simple for j > 1: p0 (1, j ) = t t e +e t (1 (j ) e tj 1 1)(1 e ) tj 2 (e t ) 127 4.3. LIMITING BEHAVIOR Recopying the ﬁrst term on the right and using e t = (1 the second, we can rewrite the right-hand side of the above as e t (1 +e ) t (j 1)(1 tj 2 (j t (1 e t )+ in 1) tj 1 e e e )...
View Full Document

This document was uploaded on 03/06/2014 for the course MATH 4740 at Cornell.

Ask a homework question - tutors are online