{[ promptMessage ]}

Bookmark it

{[ promptMessage ]}

m507b-hw6soln-s08

# m507b-hw6soln-s08 - MATH 507b ASSIGNMENT 6 SOLUTIONS SPRING...

This preview shows pages 1–2. Sign up to view the full content.

MATH 507b ASSIGNMENT 6 SOLUTIONS SPRING 2008 Prof. Alexander Chapter 7: (1.3) Let ξ be N (0 , t ), so Δ m,n has the distribution of 2 - n/ 2 ξ . We have 2 = t, E ( ξ 2 - t ) 2 = 2 t 2 , so the r.v.’s Δ 2 m,n - t 2 n are i.i.d. with mean 0. Therefore E 2 n X m =1 Δ 2 m,n - t ! 2 = E 2 n X m =1 Δ 2 m,n - t 2 n ! 2 (1) = 2 n X m =1 E Δ 2 m,n - t 2 n 2 = 2 n · 1 2 2 n E ( ξ 2 - t ) 2 = 2 t 2 2 n . Combining this with Chebyshev’s Inequality we obtain X n =1 P 2 n X m =1 Δ 2 m,n - t > ! X n =1 2 t 2 2 n 2 < , so by Borel-Cantelli, P 2 n X m =1 Δ 2 m,n - t > i.o. in n ! = 0 . Since is arbitrary this means 2 n m =1 Δ 2 m,n t a.s. (2.1) We have R = T 0 θ 1 + 1 so by the Markov property, P x ( R > 1 + t | F 1 ) = P x ( T 0 θ 1 > t | F 1 ) = P X 1 ( T 0 > t ) . Hence P x ( R > 1 + t ) = E x P X 1 ( T 0 > t ) = Z R p 1 ( x, y ) P y ( T 0 > t ) dy, since X 1 has the density p 1 ( x, y ) (as a function of y ) under P x . (2.3) By Theorem 2.6 (and its symmetric analog for “ B t < 0”), B t goes back and forth between positive and negative values infinitely often a.s. as t & 0. This means that almost surely, there exist s 1 > t 1 > s 2 > t 2 > . . . with B

This preview has intentionally blurred sections. Sign up to view the full version.

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

{[ snackBarMessage ]}