Scribd
Upload
160 views

Final Exam, Stochastic Processes: Hai Le, ID: 998010705

1) The document is the final exam for a stochastic processes course, containing two math problems involving stochastic processes and Ito's formula. 2) For problem 2.5, the student shows that a stochastic process is a martingale and uses the optional stopping theorem to show the expected value of the process at a bounded stopping time is equal to the initial value. 3) For problem 6.5, the student applies Ito's formula to two stochastic processes involving functions of Brownian motion to derive stochastic differential equations.

Uploaded by

Hai Le
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
160 views

Final Exam, Stochastic Processes: Hai Le, ID: 998010705

1) The document is the final exam for a stochastic processes course, containing two math problems involving stochastic processes and Ito's formula. 2) For problem 2.5, the student shows that a stochastic process is a martingale and uses the optional stopping theorem to show the expected value of the process at a bounded stopping time is equal to the initial value. 3) For problem 6.5, the student applies Ito's formula to two stochastic processes involving functions of Brownian motion to derive stochastic differential equations.

Uploaded by

Hai Le
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 2

P ENNSYLVANIA S TATE U NIVERSITY

Final Exam, Stochastic Processes

Hai Le, ID: 998010705


June 27, 2017

Problem 2.5. Since T is bounded, the optional stopping theorem is usable. First we notice that
{Xt } is a martingale. On the other hand, the process {Yt : t o} = {Xt2 t : t 0} is as well
a martingale adapted to the natural filtration of Brownian motion. Indeed, for s < t, we have
E[Xt2 ] = t < and

E[Xt2 t|Fs ] = E[(Xt Xs + Xs )2 t|Fs ]


= E[(Xt Xs )2 |Fs ] + 2E[Xs (Xt Xs )|Fs ] + E[Xs2 |Fs ] t
= (t s) + 0 + Xs2 t
= Xs2 s.
Applying the optional stopping theorem, we obtain E[XT t ] = E[X0 ] = 0 for any t 0. We
also have that for any t 0, |XT t | max{a, b}. Hence, letting t and applying the domi-
nated convergence theorem yield E[XT ] = 0. Similarly, we get that E[XT2 T ] = 0, which yields
E[XT2 ] = ET .

We notice that, if X N(0,t), then a common result says that E(exp( X)) = exp( 2t/2) < .
We then have, for s < t,
E(exp( Xt )|Fs ) = E(exp( Xs ) exp( (Xt Xs ))|Fs )
= exp( Xs )E(exp( (Xt Xs ))|Fs )
= exp( Xs )E(exp( (Xt Xs )))
= exp( Xs ) exp( 2 (t s)/2).
Rearranging the terms yields

E(exp( Xt 2t/2)|Fs ) = exp( Xs 2 s/2).

1
Applying the optional stopping theorem yields the desired result. Similarly, we can show that
{exp(i XT + 2 T /2)} yields a martingale by noticing that for X N(0,t), E(exp(i X)) =
exp( 2t/2) < . Then the optional stopping theorem finishes the proof. 

Problem 6.5. Let G(x,t) = g(t + x,t). We have that

Gx = gx (t + x,t), Gx x = 2 gxx (t + x,t),

Gt = gx (t + x,t) + gt (t + x,t).
Applying Itos formula to G, we obtain
2
 
dYt = gx (Xt ,t)dBt + gx (Xt ,t) + gt (Xt ,t) + gxx (Xt ,t) dt.
2
Using dXt = dt + dBt , we get the desired result
2
 
dYt = gx (Xt ,t)dXt + gt (Xt ,t) + gxx (Xt ,t) dt.
2
Let K(x,t) = g(exp(t + x,t),t). We have that

Kx (x,t) = exp(t + x)gx (exp(t + x,t),t),

Kxx (x,t) = 2 exp(t + x)gx (exp(t + x,t),t) + 2 exp(t + x)2 gxx (exp(t + x,t),t),
Kt (x,t) = exp(t + x,t)gx (exp(t + x,t),t) + gt (exp(t + x,t),t).
Applying Itos formula to K, we obtain
n 2  o
dYt = Gt gx (Gt ,t)dBt + Gt gx (Gt ,t) + gt (Gt ,t) + 2 Gt gx (Gt ,t) + Gt2 gxx (Gt ,t) dt.

Problem 5.4 For t > s, we have that


N(t) t
!
E(Yt |Fs ) = E W j Z( j ) EW Z()d|Fs
j=1 0
N(s) s N(t) t
!
= W j Z( j ) EW 0
Z()d+E W j Z( j ) EW
s
Z()d|Fs
j=1 j=N(s)+1
N(s) s N(t) t
!
= W j Z( j ) EW 0
Z()d+E W j Z( j ) EW
s
Z()d
j=1 j=N(s)+1
N(s) s
= W j Z( j ) EW 0
Z()d
j=1

= Ys .


You might also like