Stochastic

# 20 x eyk x y f y x using the law of large numbers for

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: that there are about n/Ey Ty returns by time n. Proof. We have already shown (1.15). To turn this into the desired result, we note that from the deﬁnition of R(k ) it follows that R(Nn (y )) n < R(Nn (y )+ 1). Dividing everything by Nn (y ) and then multiplying and dividing on the end by Nn (y ) + 1, we have n R(Nn (y ) + 1) Nn (y ) + 1 R(Nn (y )) < · Nn (y ) Nn (y ) Nn (y ) + 1 Nn (y ) Letting n ! 1, we have n/Nn (y ) trapped between two things that converge to Ey Ty , so n ! Ey T y Nn (y ) and we have proved the desired result. 43 1.7. PROOFS OF THE MAIN THEOREMS* Theorem 1.22. If p is an irreducible and has stationary distribution ⇡ , then ⇡ (y ) = 1/Ey Ty Proof. Suppose X0 has distribution ⇡ . From Theorem 1.21 it follows that Nn (y ) 1 ! n Ey T y Taking expected value and using the fact that Nn (y ) n, it can be shown that this implies E⇡ Nn (y ) 1 ! n Ey T y but since ⇡ is a stationary distribution E⇡ Nn (y ) = n⇡ (y ). Theorem 1.23. Suppose p is irreducible, has stationary distribution ⇡ , and P x |f (x)|⇡ (x) < 1 then n X 1X...
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