返回题库

命中 1 与 -2:概率与期望时间

Brownian motion hits 1

专题
Finance / 金融
难度
L4

题目详情

For a Brownian motion WtW_{t} , what is the probability that it hits 1 before it hits - 2? What is the expected value of the stopping time to hit either of these levels? Repeat both questions for the stochastic process XtX_{t} satisfying dXt=μdt+dWtd X_{t} = \mu d t + d W_{t} .

解析

先看标准布朗运动 WtW_t,令

τ=inf{t0:Wt{1,2}}.\tau=\inf\{t\ge 0: W_t\in\{1,-2\}\}.

(1) 命中 1 的概率:由 WtW_t 是鞅且可选停止

0=E[Wτ]=1P(Wτ=1)+(2)P(Wτ=2),0=\mathbb{E}[W_\tau]=1\cdot\mathbb{P}(W_\tau=1)+(-2)\cdot\mathbb{P}(W_\tau=-2),

并且两概率和为 1,解得

P(Wτ=1)=23.\boxed{\mathbb{P}(W_\tau=1)=\frac{2}{3}}.

(2) 期望停时Wt2tW_t^2-t 是鞅,故

0=E[Wτ2τ]E[τ]=E[Wτ2]=1223+(2)213=2.0=\mathbb{E}[W_\tau^2-\tau]\Rightarrow \mathbb{E}[\tau]=\mathbb{E}[W_\tau^2]=1^2\cdot\frac23+(-2)^2\cdot\frac13=2.

E[τ]=2.\boxed{\mathbb{E}[\tau]=2}.

再看带漂移过程 Xt=μt+WtX_t=\mu t+W_t,同样令 τ\tau 为首次命中 {1,2}\{1,-2\}

用指数鞅 e2μXte^{-2\mu X_t}(当 μ0\mu\ne 0)并可选停止:

1=E[e2μXτ]=P(Xτ=1)e2μ+(1P(Xτ=1))e4μ,1=\mathbb{E}[e^{-2\mu X_\tau}]=\mathbb{P}(X_\tau=1)e^{-2\mu}+\bigl(1-\mathbb{P}(X_\tau=1)\bigr)e^{4\mu},

解得

P(Xτ=1)=1e4μe2μe4μ=1e4μ1e6μ.\boxed{\mathbb{P}(X_\tau=1)=\frac{1-e^{4\mu}}{e^{-2\mu}-e^{4\mu}}=\frac{1-e^{-4\mu}}{1-e^{-6\mu}}}.

XtμtX_t-\mu t 是鞅,故 0=E[Xτμτ]0=\mathbb{E}[X_\tau-\mu\tau],因此

E[τ]=E[Xτ]μ=3P(Xτ=1)2μ(μ0).\boxed{\mathbb{E}[\tau]=\frac{\mathbb{E}[X_\tau]}{\mu}=\frac{3\mathbb{P}(X_\tau=1)-2}{\mu}}\quad(\mu\ne 0).

μ0\mu\to 0 可回到标准布朗运动情形。