SchoolWork-LaTeX/随机过程/作业/第十五周作业.tex

81 lines
4.3 KiB
TeX
Raw Blame History

This file contains ambiguous Unicode characters

This file contains Unicode characters that might be confused with other characters. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

\documentclass[全部作业]{subfiles}
\input{mysubpreamble}
\begin{document}
\setcounter{chapter}{6}
\section{反射原理与极值分布}
\subsection{练习题\thesection}
\begin{enumerate}
\questionandanswerSolution[1]{
$B(t)=x$条件下$B^{*}(t)$的分布函数与概率密度函数。
}{
$(B^{*}(t),B(t))$的联合密度函数为$h(z,x) = \displaystyle \frac{2(2z-x)}{\sqrt{2 \pi t^{3}}} e^{- \frac{(2z - x)^{2}}{2t}}$($z\geqslant x, x\leqslant z$)$B(t)$的边际密度函数为$p(x)=\frac{1}{\sqrt{2\pi t}} e^{- \frac{x^{2}}{2t}} (x \in \mathbb{R})$,所以在$B(t)=t$的条件下,$B^{*}(t)$的条件密度函数为
$$
p_{B^{*}(t)|B(t)} (z|x) = \frac{h(z,x)}{p(x)} = \frac{2(2z-x)}{t} e^{- \frac{2(z^{2}-xz)}{t}} (z\geqslant x \lor 0)
$$
所以$B^{*}(t)$的条件分布函数为
$$
F_{B^{*}(t)|B(t)}(z|x)=\begin{cases}
\left. -e^{-\frac{2(u^{2}-xu)}{t}} \right|_{u=x \lor 0}^{z}=1-e^{- \frac{2(z^{2}-xz)}{t}},\quad & 当z\geqslant x\lor 0\\
0,\quad & 当z < x \lor 0\\
\end{cases}
$$
}
\questionandanswerSolution[2]{
$B^{*}(t)-B(t)$的概率密度函数。
}{
因为$B^{*}(t)-B(t) = \displaystyle \max_{0\leqslant u\leqslant t} [B(t-u)-B(t)]$,且$\{ B(t-u)-B(t)\ ;\ u \in [0,t] \}$为标准布朗运动,所以
$$
B^{*}(t) - B(t) \xlongequal{d} \max_{0\leqslant u\leqslant t} B(u) = B^{*}(t) \xlongequal{d} [B(t)]
$$
从而$B^{*}(t)-B(t)$的概率密度函数为$ \displaystyle \sqrt{\frac{2}{\pi t}}e^{- \frac{u^{2}}{2t}} (u\geqslant 0)$
}
\questionandanswerSolution[3]{
$P(B(1)\leqslant x | B(u)\geqslant 0, 0\leqslant u\leqslant 1)$
}{
根据对称性,
$$
\begin{aligned}
P(B(1)\leqslant x|B(u)\geqslant 0, 0\leqslant u\leqslant 1) & = P(B(1)\geqslant x |B(u)\leqslant 0, 0\leqslant u\leqslant 1) \\
& = P(B(1)\geqslant -x | B^{*}(1)=0) \\
\end{aligned}
$$ \\
$(B^{*}(1),B(1))$的联合密度函数为
$$
h(z,x)=\frac{2(2z-x)}{\sqrt{2 \pi}} e^{- \frac{(2z-x)^{2}}{2}}\quad (z\geqslant x \lor 0)
$$
$B^{*}(1)$的边际密度函数为 $\displaystyle p(z) = \sqrt{\frac{2}{\pi}} e^{- \frac{x^{2}}{2}} (z\geqslant 0)$,所以
$$
\text{原式}=\begin{cases}
- \int_{-x}^{0} u e^{- \frac{u^{2}}{2}} \mathrm{d}u = 1-e^{- \frac{x^{2}}{2}},\quad & 当 x\geqslant 0\\
0,\quad & 当 x<0\\
\end{cases}
$$
}
\questionandanswerProof[4]{
证明对任意$z\geqslant 0$$\tau_z$$\tau_{-z}$同分布。
}{
对任意$t\geqslant 0$,有
$$
P(\tau_{-z}> t) = P(B(u)>-z, 0\leqslant u\leqslant t) \xlongequal{对称性}P(B(u)<z, 0\leqslant u\leqslant t) = P(\tau_z>t)
$$
$\tau_z$$\tau_{-z}$同分布。
}
\questionandanswerProof[6]{
证明当$0<z \to \infty$时,$\tau_z \xrightarrow{a.s.} \infty$
}{
因为对任意$t>0$$z>0$
$$
0\leqslant P(\tau_z\leqslant t) = \int_{0}^{t} \frac{z}{\sqrt{2 \pi s^{3}}} e^{- \frac{z^{2}}{2s}} \mathrm{d}s \xlongequal{\text{} s=\frac{z^{2}}{u^{2}}} \sqrt{\frac{2}{\pi}} \int_{\frac{z}{\sqrt{t}}}^{+\infty} e^{- \frac{u^{2}}{s}} \mathrm{d}u \xrightarrow{z \to +\infty} 0
$$
所以 $\displaystyle \lim_{z \to +\infty} P(\tau_z\leqslant t)=0 (\forall t>0)$,又因为$\tau_z$关于$z$单调增,所以事件$\{ \tau_z\leqslant t \}$关于$z$单调减,从而对任意$t >0$
$$
0 = \lim_{n \to \infty} P(\tau_n\leqslant t) = P\left( \bigcap_{n = 1} ^{\infty} \{ \tau_n\leqslant t \} \right) =P\left( \lim_{n \to \infty} \tau_n\leqslant t \right) =P\left( \lim_{z \to +\infty} \tau_z\leqslant t \right)
$$
进而由
$$
0\leqslant P\left( \lim_{n \to +\infty} \tau_z<+\infty \right) =P\left( \bigcup_{m=1} ^{\infty} \left\{ \lim_{z \to +\infty} \tau_z\leqslant m \right\} \right) \leqslant \sum_{m=1}^{\infty} P \left( \lim_{z \to +\infty} \tau_z\leqslant m \right) =0
$$
$\displaystyle P\left( \lim_{z \to +\infty} \tau_z = +\infty \right) =1$,于是当$z \to +\infty$时有$\tau_z \xrightarrow{a.s.} +\infty$
}
\end{enumerate}
\end{document}