关键词 > MATH47101/67101
MATH47101/67101 STOCHASTIC CALCULUS 2021
发布时间:2022-01-17
Hello, dear friend, you can consult us at any time if you have any questions, add WeChat: daixieit
MATH47101/67101
STOCHASTIC CALCULUS
2021
1. Let B = (Bt )t≥0 be a standard Brownian motion started at zero, and let (Ft(B))t≥0 denote the natural filtration of B .
(1.1) State the definition of B .
(1.2) Determine whether ~t1/4Bt1/2│t≥0 defines a standard Brownian
motion. Explain your answer.
(1.3) Show that τ = inf {t ≥ 0 :)Bt)= 1+It } is a stopping time with
respect to (Ft(B))t≥0 .
(1.4) Show that ~Bt +e一2Bt 一2t│t≥0 is a martingale with respect to (Ft(B))t≥0 .
(1.5) Set Mt = Bt +e一2Bt 一2t for t ≥ 0 . Compute E(Mσ ) and E(e一4σ)
when σ = inf {t ≥ 0 : Bt = -1 } .
2. Let X = (Xt )t≥0 be a continuous semimartingale with values in 皿 , let It = 0(t) eXsds for t ≥ 0 , and let F : 皿+ ×皿×皿+ 二 皿 be a C1,2,1 function.
(2.1) Apply Itˆo’s formula to F (t, Xt , It ) for t ≥ 0 . Determine a continuous
local martingale (Mt )t≥0 starting at 0 and a continuous bounded
variation process (At )t≥0 such that F (t, Xt , It ) = Mt +At for t ≥ 0 .
Let B = (Bt )t≥0 be a standard Brownian motion started at zero, let X = (Xt )t≥0 be a non-negative stochastic process solving
dXt = dt + dBt (X0 = 0)
and let F (t, x) = It x2 for t ≥ 0 and x e 皿+ .
(2.2) Explain why Itˆo’s formula can be applied to F (t, Xt ) for t ≥ 0 .
(2.3) Apply Itˆo’s formula to F (t, Xt ) for t ≥ 0 . Determine a continuous
local martingale (Mt )t≥0 starting at 0 and a continuous bounded variation process (At )t≥0 such that F (t, Xt ) = Mt +At for t ≥ 0 .
(2.4) Show that (Mt )t≥0 in (2.3) is a martingale and compute 〈M, M〉t
for t ≥ 0 .
(2.5) Compute E(τ ) when τ = inf {t e [0, 7] : Xt = I7 -t } .
3. Let B = (Bt )t≥0 be a standard Brownian motion started at zero, let St = sup 0≤s≤t Bs and It = 0(t) eBsds for t ≥ 0 , and let F : 皿+ ×皿×皿+ ×皿+ 二 皿 be a C1,2,1,1 function.
(3.1) Explain why Itˆo’s formula can be applied to F (t, Bt , St , It ) for t ≥ 0 .
(3.2) Apply Itˆo’s formula to F (t, Bt , St , It ) for t ≥ 0 . Determine a continuous
local martingale (Mt )t≥0 starting at 0 and a continuous bounded variation process (At )t≥0 such that F (t, Bt , St , It ) = Mt +At for t ≥ 0 .
(3.3) Show that if Ft (t, x, s, i)+Fxx (t, x, s, i) + ex Fi (t, x, s, i) = 0 for all
(t, x, s, i) with Fs (t, x, s, i) = 0 for x = s , then F (t, Bt , St , It ) is a continuous local martingale for t ≥ 0 .
(3.4) Show that (St -Bt )2 + -
It - t is a martingale for t ≥ 0 .
4. Let B = (Bt )t≥0 be a standard Brownian motion started at zero, and let M = (Mt )t≥0 be a stochastic process defined by
′et 一1
0
for t ≥ 0 .
(4.1) Show that M is a standard Brownian motion.
(4.2) Compute E~(1+Mt )2 0(t)(1+Ms )2 ds │ for t ≥ 0 .
(4.3) Compute E~(1+Mt )2 0(t)(1+Ms )2 dMs │ for t ≥ 0 .
(4.4) Consider the process Z = (Zt )t≥0 defined by
Zt = f(t)B′et 一1
for t ≥ 0 where f : 皿+ 二 皿 is a C1 function. Using Itˆo’s formula examine whether f can be chosen so that Z solves
dZt = g(t)Zt dt + dMt (Z0 = 0)
for some function g . If this is possible, determine f and g explicitly.
5. Let B = (Bt )0≤t≤T be a standard Brownian motion started at zero under a probability measure P , and let = (
t )0≤t≤T be a stochastic process defined by
t = Bt - t e一Bs I(Bs ≥1) ds
for t e [0, T] , where T > 0 is a given and fixed constant.
(5.1) Determine a probability measure under which
is a standard
Brownian motion. [7 marks]
(5.2) Compute ~[ 0(t)
s dBs +
t -0(t)
s
I(Bs ≥1)ds]2 │ for t ≥ 0 . [4 marks]
(5.3) Compute E(eτ ) when τ = inf {t ≥ 0 : Bt = -2 or Bt = 1 } . [7 marks]
(5.4) Compute ~e3 ′2(Bσ 一 0σ
I(Bs≥1) ds)一9σ│ and
~e一9σ│ when
σ = inf {t ≥ 0 : Bt = I5 + 0(t) I(Bs ≥1) ds } . [7 marks]
(Recall that denotes expectation under
, and E denotes expectation under P .)
6. Let B = (Bt )t≥0 be a standard Brownian motion started at zero. Consider the stochastic differential equation
dXt = (1 -Xt ) dt + (1+2Xt ) dBt
for a stochastic process X = (Xt )t≥0 where it is assumed that X0 = 1 .
(6.1) Show that there exists a unique strong solution X to this equation.
(6.2) Verify by Itˆo’s formula that this solution is given by
Xt = Yt ╱ 1 - 0 t ds + 0 t
dBs ←
for t ≥ 0 , where the process Y = (Yt )t≥0 solves
dYt = -Yt dt + 2Yt dBt
with Y0 = 1 .
(6.3) Compute E~〈X, Y〉t │ for t ≥ 0 .
(6.4) Show that the following identity in law holds
Xt - Yt ╱ 1 + 0 t dBs ←
Ys ds
for each t ≥ 0 given and fixed.