数学物理学报, 2020, 40(2): 484-491 doi:

论文

Brown运动增量拟必然局部Strassen重对数律

李丰兵,, 刘永宏

Quasi Sure Local Strassen's Law of the Iterated Logarithm for Increments of a Brownian Motion

Li Fengbing,, Liu Yonghong

通讯作者: 刘永宏

收稿日期: 2018-11-20  

基金资助: 国家自然科学基金.  11661025
广西教育厅科研基金.  YB2014117
广西自然科学基金.  2018GXNSFBA281076
广西自然科学基金.  2017GXNSFBA198179

Received: 2018-11-20  

Fund supported: the NSFC.  11661025
the Science Research Foundation of Guangxi Education Department.  YB2014117
the NSF of Guangxi.  2018GXNSFBA281076
the NSF of Guangxi.  2017GXNSFBA198179

作者简介 About authors

李丰兵,E-mail:1286391997@qq.com , E-mail:1286391997@qq.com

摘要

该文建立了Brown运动增量的拟必然局部Strassen重对数律.利用这一结果,得到了Brown运动拟必然泛函连续模.

关键词: Brown运动 ; 增量 ; 局部重对数律 ; (r, p)-容度

Abstract

In this paper, we establish the quasi sure local Strassen's law of the iterated logarithm for increments of a Brownian motion. As an application, a quasi sure functional modulus of continuity for a Brownian motion is also derived.

Keywords: Brownian motion ; Increments ; Local law of the iterated logarithm ; (r, p)-Capacity

PDF (287KB) 元数据 多维度评价 相关文章 导出 EndNote| Ris| Bibtex  收藏本文

本文引用格式

李丰兵, 刘永宏. Brown运动增量拟必然局部Strassen重对数律. 数学物理学报[J], 2020, 40(2): 484-491 doi:

Li Fengbing, Liu Yonghong. Quasi Sure Local Strassen's Law of the Iterated Logarithm for Increments of a Brownian Motion. Acta Mathematica Scientia[J], 2020, 40(2): 484-491 doi:

1 引言和主要结果

$(B, H, \mu)$是抽象的Wiener空间.容度是$B$上的集合函数,在$\mu$ -零测集上有可能取正值.因此,一个有意义的问题是:布朗运动在几乎处处意义下成立的一些性质是否在容度意义下也成立?在本文中,我们将研究这一问题.

在近期文章[1]中,高付清等建立了Brown运动增量的局部泛函极限定理.在本文中,我们讨论类似结果,但是概率被容度取代.我们的结果也推广了王文胜[4]的相应结果.

让我们考虑经典的Wiener空间$(W, H, \mu)$如下:

$D^{r, p}, r>0, 1\leq p <\infty $是Sobolev空间,即:

其中$L^p$$(W, \mu)$上实值函数组成的$L^p$ -空间, ${\cal L} $$(W, H, \mu)$上的Ornstein-Uhlenbeck算子.对$r>0, p> 1$, $(r, p)$ -容度被定义如下:

对任意集$A\subset W$,

${\cal C}^D$表示从$[0, 1]$$R^D$的连续函数空间,赋予上确界范数$ \|f\|:=\sup\limits_{0\leq t \leq 1}|f(t)|$.并且记${\cal C}^D_0:=\{f\in {\mathcal C}^D; f(0)=0\}$, ${\cal H}^D:=\{f\in {\cal C}^D_0; f(t)=\int _0^t\dot{f}(s){\rm d}s, \|f\|_{{\cal H}^D}^2:= \int_0^1|\dot{f}(t)|^2{\rm d}t <\infty \}$, $K:=\{f\in{\mathcal{H}}^D; 2I(f)\leq 1\}$,其中

全文中,设$a_u, b_u $是从$(0, 1)$$(0, e^{-1})$的两个非减的连续函数,满足

(ⅰ) $a_u\leq b_u$,对任何$u\in(0, 1)$$\lim\limits_{u\to0}a_u=0$;

(ⅱ) $\frac{b_u}{a_u}$是非增的.

$w\in W$,对$ u\in (0, 1)$, $0\leq t\leq 1$, $\Delta(t, u)$记下面轨道:

本文主要结果陈述如下:

定理1.1  设$(r, p)\in [0, \infty)\times [1, \infty)$.若条件(ⅰ)和(ⅱ)成立,则

$ \limsup\limits_{u\rightarrow 0}\sup\limits_ {t\in[0, 1-\frac{a_u}{b_u}]}\|\beta_u\Delta(t, u)(\cdot)-K\|=0, ~~~~C_{r, p}\mbox{-q.s.}, $

$ \liminf\limits_{u\rightarrow 0}\inf\limits_ {t\in[0, 1-\frac{a_u}{b_u}]}\|\beta_u\Delta(t, u)(\cdot)-\varphi\|=0, ~\mbox{对所有}~\varphi \in K, ~~~~ C_{r, p}\mbox{-q.s.}. $

而且,若条件

(ⅲ) $\lim\limits_{u\rightarrow 0}\frac{\log{(b_u/{a_u})}}{\log\log b^{-1}_u}=\infty$也成立,则

$ \lim\limits_{u \to 0} \inf\limits_{t\in [0, 1-\frac{a_u}{b_u}]}\|\beta_u \Delta(t, u)(\cdot)-\varphi\|=0, ~\mbox{对任意}~\varphi \in K, ~~~~C_{r, p}\mbox{-q.s.}. $

推论1.1  设$M_{t, h}(x)=\frac{w(t+hx)-w(t)}{\sqrt{2h\log h^{-1}}}, 0\le x\le 1, 0\le t\le 1-h.$$(r, p)\in [0, \infty)\times [1, \infty)$,我们有

$ \label{cor1}\lim\limits_{h\rightarrow 0}\sup\limits_ {t\in[0, e^{-2}-h]}\|M_{t, h}(\cdot)-K\|=0, ~~~~C_{r, p}\mbox{-q.s.}, $

$ \label{cor2}\lim\limits_{h\rightarrow 0}\inf\limits_{t\in[0, e^{-2}-h]}\|M_{t, h}(\cdot)-\varphi\|=0~\mbox{对任何}~~\varphi \in K. ~~~~ C_{r, p}\mbox{-q.s.}. $

2 定理1.1的证明

引理2.1[4,引理2.4]   设$(W, H, \mu)$是经典Wiener空间,那么对任何闭子集$F\subset W$,任何$T>0$$0 <h\le T$,下式成立

引理2.2[3,引理2.2]   设$1\leq k\in Z$, $q_1, q_2, \in(1, \infty)$被给定,使得$\frac{1}{p}=\frac{1}{q_1}+\frac{1}{q_2}$.对任何$f\in K$,令

其中$0\leq t_i <\infty, h_i>0, i=1, 2, \cdot\cdot\cdot, n$.那么,存在一个常数$c=c(k, p, q_1, f, d)>0$,对任何$\delta\in(0, 1], \varepsilon\in(0, 1]$,我们有

引理2.3[2,引理4.4]  对$u \ge 3$,定义

(1)   对任何$(r, p) \in (0, \infty)\times (1, \infty)$,对$C_{r, p}\mbox{-q.s.} w$, $\{\xi _u(\cdot, w), u \ge 3\}$$W$中相对紧且其极限点集是$C: = \{f \in H; \| f\|_{H}^2 \le 1\}$$u\to\infty$.

(2)   对每个$f\in C$$\| f\|_{H}^2 < 1$,对任何$\eta>0$,存在$0 <c_\varepsilon < 1$使得对所有$0 <c < c_\varepsilon $,

引理2.4  存在非增序列$\{u_n\in(0, 1), n\in{\Bbb N}\}$$\lim\limits_{n\to\infty}u_n\to0$, $\lim\limits_{n\to \infty}\frac{a_{u_n}}{a_{u_{n+1}}}=1$,使得对任何$\varepsilon>0$,有

  设$ A=\{f\in C_0[0, 1]: \|f-K\|\geq \varepsilon \} $,显然, $A$是闭集.对任何$f\in A$,存在$\delta>0$使得

情形(Ⅰ)   $\limsup\limits_{u\to 0} \frac{\log \frac{b_u}{a_u}} {\log\log b_{u}^{-1}}<\infty$.

$\limsup\limits_{u\to 0} \frac{\log \frac{b_u}{a_u}} {\log\log b_{u}^{-1}}<\infty$,则存在$0 <M<\infty$,使得

因此$\frac{1}{\log b_u^{-1}}\le \frac{1}{\log a_u^{-1}-M\log\log a_u^{-1}}$.$u_n$使得$a_{u_n}=\exp\left(-\frac{n}{(\log n)^3}\right)$,那么

$ \label{eq-2-1}\frac{1}{\log b^{-1}_{u_n}}\leq\frac{(\log n)^3}{n-M(\log n)^4}. $

由引理2.1,我们有

$\begin{eqnarray}\label{22eq-2-2}&&C_{r, p}\Big(\sup\limits_{0\leq t\leqb_{u_n}-a_{u_{n+1}}}\|\beta_{u_{n}}(w(t+a_{u_n}\cdot)-w(t))-K\|\geq\varepsilon \Big)\\&=&C_{r, p}\Big(\sup\limits_{0\leq t\leqb_{u_n}-a_{u_{n+1}}}\left\|\frac{\Big(2\log \frac{b_{u_{n}}\log b_{u_{n}}^{-1}}{a_{u_{n}}}\Big)^{-1/2}}{\sqrt{a_{u_n}}}\Big(w(t+a_{u_n}\cdot)-w(t)\Big)-K\right\|\geq \varepsilon \Big)\\&=&C_{r, p}\Big(\bigcup\limits_{0\leq t\leq (b_{u_n}+a_{u_n}-a_{u_{n+1}})-a_{u_n}}\Big\{\frac{\Big(2\log \frac{b_{u_{n}}\log b_{u_{n}}^{-1}}{a_{u_{n}}}\Big)^{-1/2}}{\sqrt{a_{u_n}}}\Big(w(t+a_{u_n}\cdot)-w(t)\Big)\in A\Big\}\Big)\\&\leq& \frac{b_{u_{n}}+a_{u_n}-a_{u_{n+1}}}{a_{u_{n}}}\left(\frac{a_{u_n}}{b_{u_n}\log b^{-1}_{u_n}}\right)^{1+2\delta}.\end{eqnarray}$

结合(2.1)和(2.2)式,由Borel-Cantelli引理,我们能得到

情形(Ⅱ)    $\limsup\limits_{u\to 0} \frac{\log \frac{b_u}{a_u}} {\log\log b_{u}^{-1}}=\infty$.

$\limsup\limits_{u\to 0}\frac{\log \frac{b_u}{a_u}}{\log\log b_{u}^{-1}}=\infty$,则我们能选$u_n$使得$\frac{b_{u_n}}{a_{u_n}}=n^p$,其中$p>\frac{1}{\delta}$.和情形(Ⅰ)类似的方法,此引理的证明完成.

引理2.5   若条件(ⅰ)和(ⅱ)成立,则我们有

$ \label{22eq123a}\limsup\limits_{u\to 0}\sup\limits_{t\in [0, b_u-a_u]}\|\beta_u(w(t+a_u\cdot)-w(t))-K\|=0, ~~ C_{r, p}\mbox{-q.s.}. $

  设

情形(Ⅰ)   $\limsup\limits_{u\to 0} \frac{\log \frac{b_u}{a_u}} {\log\log b_{u}^{-1}}<\infty$.

$u_n$如引理2.4证明中情形(Ⅰ)定义,则对较小的$u\in (0, 1)$,存在整数$n$使得$u_{n+1} <u\le u_{n}$.注意到$\psi_{t, u}(s)=\frac{\beta_u}{\beta_{u_n}}\psi_{t, u_n}(\frac{a_u}{a_{u_n}}s)$,我们能得到

$\begin{eqnarray}\label{2a2}\sup\limits_{t\in [0, b_u-a_u]}\|\psi_{t, u}(\cdot)-K\|&=&\sup\limits_{t\in [0, b_u-a_u]}\inf\limits_{f\in K}\|\psi_{t, u}(\cdot)-f(\cdot)\|\\&\leq& \sup\limits_{t\in [0, b_{u_n}-a_{u_{n+1}}]}\inf\limits_{f\in K}\bigg\|\psi_{t, u_n}(\frac{a_u}{a_{u_n}}\cdot)-f(\frac{a_u}{a_{u_n}}\cdot)\bigg\|\\&&+\left|\frac{\beta_u}{\beta_{u_n}}-1\right|\sup\limits_{t\in [0, b_{u_n}-a_{u_{n+1}}]}\|\psi_{t, u_n}(\cdot)\|+\sup\limits_{f\in K}\bigg\|f(\frac{a_u}{a_{u_n}}\cdot)-f(\cdot)\bigg\|.\end{eqnarray}$

因为$\beta_u, \log\frac{b_u\log b_u^{-1}}{a_u}$非增,故我们有

$ \label{eqa}\frac{\beta_u}{\beta_{u_{n+1}}}\geq\frac{\beta_{u_n}}{\beta_{u_{n+1}}}=\left(\frac{a_{u_{n+1}}\log\frac{b_{u_{n+1}}\log b_{u_{n+1}}^{-1}}{a_{u_{n+1}}}}{a_{u_n}\log\frac{b_{u_n}\log b_{u_n}^{-1}}{a_{u_n}}}\right)^{1/2}\geq\left(\frac{a_{u_{n+1}}}{a_{u_n}}\right)^{1/2}. $

$ \label{22eq}\bigg \|f(\cdot)-f(\frac{a_u}{a_{u_n}}\cdot)\bigg\|\leqslant 2\bigg|1-\frac{a_{u_{n+1}}}{a_{u_n}}\bigg|^{\frac{1}{2}}. $

$a_{u_n}=\exp\big(-\frac{n}{(\log n)^3}\big)$,容易看出$\frac{a_{u_n}}{a_{u_{n+1}}}\to 1$ ($n\to\infty$),由引理2.4, (2.4)-(2.6)式,完成了(2.3)式的证明.

情形(Ⅱ)   $\limsup\limits_{u\to 0} \frac{\log \frac{b_u}{a_u}}{\log\log b_{u}^{-1}}=\infty$.

$\limsup\limits_{u\to 0}\frac{\log\frac{b_u}{a_u}}{\log\log b_u^{-1}}=\infty$,如引理2.4证明中情形(Ⅱ),我们能选非增序列$\{u_n; n\geq 1\}$使得$\frac{b_{u_n}}{a_{u_n}}=n^p, p>\frac{1}{\delta}$.$h(n)=\frac{\log \frac{b_{u_n}}{a_{u_n}}}{\log\log b_{u_n}^{-1}}=\frac{p\log n}{\log\log b_{u_n}^{-1}}$,则$b_{u_n}^{-1}=\exp\{n^{\frac{p}{h(n)}}\}$$h(n)\to\infty, n\to\infty $.而且,对任何小的$\alpha>0, \frac{(n+1)^\alpha}{\log b_{u_{n+1}}^{-1}}\to \infty $,且

$ \label{eq-2-4r}1\leq \frac{b_{u_n}}{b_{u_{n+1}}}=\exp\Big((n+1)^{\frac{p}{h(n+1)}}-n^{\frac{p}{h(n)}}\Big)\to 1, \quad\frac{a_{u_n}}{a_{u_{n+1}}}\to 1, \quad n\to\infty. $

由情形(Ⅰ)的证明中类似方法,我们能完成(2.3)式的证明.

引理2.6   若条件(ⅰ)和(ⅱ)成立,则对任何$f\in K$,我们有

$ \label{33eqa121}\liminf\limits_{u\to 0}\|\beta_u\Delta(1-\frac{a_u}{b_u}, u)-f\|=0, ~~ C_{r, p}\mbox{-q.s.} $

  设$\rho=\lim\limits_{u\to 0}\frac{a_u}{b_u}.$$\rho<1$$b_u\to b\neq 0, (u\to 0)$,则$\lim\limits_{u\to 0}\frac{\log\frac{b_u}{a_u}}{\log\log b_u^{-1}}=\infty$,在此情形, (2.8)式能从下面引理2.8得到.

这里,我们仅考虑两种情形: (Ⅰ) $\rho<1$$b_u\to 0, (u\to 0)$; (Ⅱ) $\rho=1$.

情形(Ⅰ)   $\rho<1$$b_u\to 0, (u\to 0)$.

我们选$u_n$使得

$k=[r]+1$,由引理2.2,有

$\begin{eqnarray}\label{qqq22eq} &&C_{r, p}\left(\bigcap\limits_{m=l}^n\Big(\|\beta_{u_m}\Delta(1-\frac{a_{u_m}}{b_{u_m}})-f\|\ge 2\varepsilon\Big)\right)^{1/p}\\&=& C_{r, p}\left(\bigcap\limits_{m=l}^n\Big(\|\frac{w(b_{u_m}-a_{u_m}+a_{u_m}\cdot)-w(b_{u_m}-a_{u_m})}{\sqrt{2a_{u_m}\ell_{u_m}}}-f\|\ge 2\varepsilon\Big)\right)^{1/p}\\&\le & cn^{k}\varepsilon^{-2k^2-k}\mu\left(\bigcap\limits_{m=l}^n\Big(\|\frac{w(b_{u_m}-a_{u_m}+a_{u_m}\cdot)-w(b_{u_m}-a_{u_m})}{\sqrt{2a_{u_m}\ell_{u_m}}}-f\|\ge \varepsilon\Big)\right)^{1/q_2}\\&=& cn^{k}\varepsilon^{-2k^2-k}\prod\limits_{m=l}^n\left\{1-\mu\left(\Big(\|\frac{w(b_{u_m}-a_{u_m}+a_{u_m}\cdot)-w(b_{u_m}-a_{u_m})}{\sqrt{2a_{u_m}\ell_{u_m}}}-f\|< \varepsilon\Big)\right)\right\}^{1/q_2}\\&=& cn^{k}\varepsilon^{-2k^2-k}\prod\limits_{m=l}^n\left\{1-\mu\left(\frac{1}{\sqrt{2\ell_{u_m}}}w(\cdot)\in A\right)\right\}^{1/q_2}, \end{eqnarray}$

其中$A=\{g; \|g-f\|<\varepsilon\}$.因为$2\inf \limits_{g\in A}I(g)<1$,我们能选$\eta>0$,使得$\eta_0=2\inf\limits_{g\in A}I(g)+\eta<1$.而且,由大偏差

$ \label{32q22eq} \mu\left(\frac{1}{\sqrt{2\ell_{u_m}}}w(\cdot)\in A\right)\ge \left(\frac{a_{u_m}}{b_{u_m}\log b_{u_m}^{-1}}\right)^{\eta_0}, $

我们有

$\begin{eqnarray}\label{eq12345} &&C_{r, p}\left(\bigcap\limits_{m=l}^n\Big(\|\beta_{u_m}\Delta(1-\frac{a_{u_m}}{b_{u_m}})-f\|\ge 2\varepsilon\Big)\right)^{1/p}\\&\le & cn^{k}\varepsilon^{-2k^2-k}\exp\left(-\frac{1}{q_2}\sum\limits_{m=l}^n\left(\frac{a_{u_m}}{b_{u_m}\log b_{u_m}}\right)^{\eta_0}\right).\end{eqnarray}$

注意到,存在$A=A(l)>0$,使得

$n$足够大时,对常数$c'>0$,我们能证明

因此,我们有

$\begin{eqnarray}\label{aceq12345} &&C_{r, p}\left(\bigcap\limits_{m=l}^n\Big(\|\beta_{u_m}\Delta(1-\frac{a_{u_m}}{b_{u_m}})-f\|\ge 2\varepsilon\Big)\right)^{1/p}\\&\le & c_0n^{2k^2}\exp\left(-\frac{A}{q_2}n^{1-\eta_0}\right)\to 0, ~ n\to \infty, \end{eqnarray}$

其中$c_0$是常数.我们得到

因此

$ \label{33abeq}\liminf\limits_{u\to 0}\|\beta_u\Delta(1-\frac{a_u}{b_u}, u)-f\|=0, ~~ C_{r, p}\mbox{-q.s.}. $

情形(Ⅱ)   $\rho=1$.

$\rho=1$,则$a_u=b_u$.见引理2.3.

引理2.7  若条件(ⅲ)也成立,则对任何$f\in K$,存在子列$\{u_n, n\ge 1\}$,使得

$ \label{33eqacc}\limsup\limits_{n\to \infty}\inf\limits_{t\in [0, b_{u_{n+1}}-a_{u_n}]}\|\beta_{u_n}(w(t+a_{u_n}\cdot)-w(t))-f\|_q=0, ~~ C_{r, p}\mbox{-a.s.}. $

  因为$\lim\limits_{u\to 0}\frac{\log\frac{b_u}{a_u}}{\log\log b^{-1}_u}=\infty$,故存在子列$\{u_n, n\ge 1\}$使得$\frac{b_{u_n}}{a_{u_n}}=n^{d'}$.$t_i=ia_{u_n}, $$i=0, 1, 2, \cdot\cdot\cdot$, $k_n=[\frac{b_{u_{n+1}}}{a_{u_n}}]-1$.$t_i=ia_{u_n}, i=0, 1, 2, \cdot\cdot\cdot$, $k_n=[\frac{b_{u_{n+1}}}{a_{u_n}}]-1$$h(n)=\frac{\log\frac{b_{u_n}}{a_{u_n}}}{\log\log b^{-1}_{u_{n}}}=\frac{\log n^{d'}}{\log\log b^{-1}_{u_n}}$.我们有$b^{-1}_{u_n}=\exp(n^{\frac{d'}{h(n)}})$$h(n)\to \infty$ ($n\to \infty$).而且,对任何小$\alpha>0$, $\frac{n^\alpha}{\log b^{-1}_{u_{n}}}\to \infty$, $1\le \frac{b_{u_n}}{b_{u_{n+1}}}=\exp\{(n+1)^{\frac{d'}{h(n+1)}}- n^{\frac{d'}{h(n)}}\}\le\exp(n^{\frac{d'}{h(n)}-1})\to 1, (n\to \infty).$$k=[r]+1$,对$n$足够大,我们有

$\begin{eqnarray}\label{33qeq12345} &&C_{r, p}\left(\inf\limits_{t\in [0, b_{u_{n+1}}-a_{u_n}]}\Big(\|\beta_{u_n}\Big(w(t+a_{u_n}\cdot)-w(t)\Big)-f\|\ge 2\varepsilon\Big)\right)^{1/p}\\&\le & c(1+k_n)^k\varepsilon^{-k^2-k}\mu(\min\limits_{0\le i\le k_n}\|\beta_{u_n}(w(t_i+a_{u_n}\cdot)-w(t_i))-f\|\ge \varepsilon)^{1/q_2}\\&=&c(1+k_n)^k\varepsilon^{-k^2-k}\left\{1-\mu\left(\frac{1}{\sqrt{2\ell_{u_n}}}w(\cdot)\in A\right)\right\}^{\frac{k_n+1}{q_2}}, \end{eqnarray}$

其中$A=\{a; \|g-f\|<\varepsilon\}$.因为$2\inf\limits_{g\in A}I(g)<1$,选$\delta>0$,使得$\mu:=2\inf\limits_{g\in A}I(g)+2\delta<1$.$n$足够大,由大偏差

因此

$\begin{eqnarray}\label{33qq**}& &C_{r, p}\left(\inf\limits_{t\in [0, b_{u_{n+1}}-a_{u_n}]}\Big(\|\beta_{u_n}\Big(w(t+a_{u_n}\cdot)-w(t)\Big)-f\|\ge 2\varepsilon\Big)\right)^{1/p}\\&\le & c(1+k_n)^k\varepsilon^{-k^2-k}\mu(\min\limits_{0\le i\le k_n}\|\beta_{u_n}(w(t_i+a_{u_n}\cdot)-w(t_i))-f\|\ge \varepsilon)^{1/q_2}\\&\le &c'n^{k^2d'}\exp\left\{-\frac{1}{q_2}\left(\frac{a_{u_n}}{b_{u_n}\log b^{-1}_{u_n}}\right)^{\mu}\left[\frac{b_{u_{n+1}}}{a_{u_n}}\right]\right\}, \end{eqnarray}$

其中$c, c'$是常数.若选适当的$d'$,则

由Borel-Cantelli引理

引理2.7证毕.

引理2.8   若条件(ⅲ)也成立,则对任何$f\in K$,我们有

$ \label{33eqa**}\limsup\limits_{u\to 0}\inf\limits_{t\in [0, b_u-a_u]}\|\beta_u(w(t+a_u\cdot)-w(t))-f\|=0, ~~ C_{r, p}\mbox{-q.s.}. $

  设$\phi_{t, u}(s)=\beta_u(w(t+a_us)-w(t)), s\in[0, 1], t\in [0, b_u-a_u], $$u_n$如引理2.7的证明中定义.因为$\phi_{t, u}(s)=\frac{\beta_u}{\beta_{u_n}}\phi_{t, u_n}(\frac{a_u}{a_{u_n}}s)$,我们有

$\begin{eqnarray}\label{33qqbcb}& &\inf\limits_{t\in [0, b_u-a_u]}\|\phi_{t, u}(\cdot)-f(\cdot)\|=\inf\limits_{t\in [0, b_u-a_u]}\|\frac{\beta_u}{\beta_{u_n}}\phi_{t, u_n}(\frac{a_u}{a_{u_n}}\cdot)-f(\cdot)\|\\&\le & \inf\limits_{t\in [0, b_{u_{n+1}}-a_{u_n}]}\|\phi(\frac{a_u}{a_{u_n}}\cdot)-f(\frac{a_u}{a_{u_n}}\cdot)\|\\&&+ \sup\limits_{t\in [0, b_{u_{n+1}}-a_{u_n}}\left|\frac{\beta_u}{\beta_{u_n}}-1\right|\left\|\phi_{t, u_n}\Big(\frac{a_u}{a_{u_n}}\cdot\Big)\right\|+\|f(\frac{a_u}{a_{u_n}})-f(\cdot)\|.\end{eqnarray}$

而且

$ \label{133qqbcb} \frac{\beta_u}{\beta_{u_n}}\le \frac{\beta_{u_{n+1}}}{\beta_{u_n}}= \left(\frac{a_{u_n}\log\frac{b_{u_n}\log b^{-1}_{u_n}}{a_{u_n}}}{a_{u_{n+1}}\log\frac{b_{u_{n+1}} \log b^{-1}_{u_{n+1}}}{a_{u_{n+1}}}}\right)^{1/2}\le\left(\frac{a_{u_n}}{a_{u_{n+1}}}\right)^{1/2}, $

$ \label{233qqbcb} \left\|f(\frac{a_u}{a_{u_n}}\cdot)-f(\cdot)\right\|\le 2\left|1-\frac{a_{u_{n+1}}}{a_{u_n}}\right|^{1/2}. $

我们从(2.18)-(2.20)式和引理2.7能得出(2.17)式.

参考文献

Gao F , Liu Y .

Local functional limit theorems of increments for Brownian motion

Acta Mathematica Sinica, 2018, 34B (7): 1074- 1086

URL     [本文引用: 1]

刘永宏, 高付清.

扩散过程的拟必然局部Strassen重对数律

数学物理学报, 2004, 24A (2): 231- 237

DOI:10.3321/j.issn:1003-3998.2004.02.016      [本文引用: 1]

Liu Y , Gao F .

Quasi sure local Strassen's law of the iterated logarithm for diffusion processes

Acta Math Sci, 2004, 24A (2): 231- 237

DOI:10.3321/j.issn:1003-3998.2004.02.016      [本文引用: 1]

Liu Y , Li L .

The rate of quasi sure convergence in the functional limit theorem for increments of a Brownian motion

J Math Anal Appl, 2009, 356, 21- 29

DOI:10.1016/j.jmaa.2009.02.036      [本文引用: 1]

Wang W .

A generalization of functional law of the iterated logarithm for (r, p)-capacities on the Wiener space

Stochastic Process Appl, 2001, 96, 1- 16

DOI:10.1016/S0304-4149(00)00096-X      [本文引用: 2]

Yoshida N .

A large deviation principle for (r, p)-capacities on the Wiener space

Probab Theory Relat Fields, 1993, 94, 473- 488

DOI:10.1007/BF01192559     

/