手机版
您的当前位置: 恒微文秘网 > 范文大全 > 专题范文 > 一类具有乘性噪声的时滞随机演化方程的随机中心流形的存在性与光滑性

一类具有乘性噪声的时滞随机演化方程的随机中心流形的存在性与光滑性

来源:专题范文 时间:2024-10-13 08:19:01

杨娟 龚佳鑫 吴隆钰 舒级

摘要:研究一类具有乘性噪声的时滞随机演化方程的随机中心流形的存在性与光滑性.由于时滞的影响,首先对带时滞的非线性项进行转化并处理由时滞影响产生的系数,从而得到中心流形的存在性,然后利用Lyapunov-Perron方法证明方程的中心流形的光滑性.

关键词:时滞随机演化方程; 随机中心流形; 乘性噪声; 存在性; 光滑性

中图分类号:O175.29  文献标志码:A  文章编号:1001-8395(2024)05-0696-12

doi:10.3969/j.issn.1001-8395.2024.05.016

本文研究如下一类具有乘性噪声的时滞随机演化方程的中心流形的存在性与光滑性:

dv(t)dt=Av(t)+H(v(t))+F(v(t-ρ))+v(t),(1)

其中,A是可分的实Hilbert空间X上的一簇自伴、稠定、具有紧预解集的线性算子,H(v(t))是从X到X的一簇Lipschitz连续项,F(v(t-ρ))是由X到X的一簇带时滞的非线性项,ρ是一个正常数,v(t)是乘性噪声,符号代表方程(1)在Stratonovich积分意义下成立.

本文主要证明方程(1)中心流形的存在性与光滑性.运用Lyapunov-Perron方法证明中心流形的存在性与光滑性,在时滞的影响下,需要对带有时滞的非线性项进行转化处理,确定新的谱间隙条件,从而得到具有乘性噪声的时滞随机演化方程的中心流形的存在性与光滑性.

不变流形在描述和理解非线性动力系统的动力学行为方面具有重要意义,并且时滞偏微分方程、随机偏微分方程在物理、力学、生物等相关领域受到人们的广泛关注,而实际问题中时滞因素与随机因素往往会同时出现.不变流形理论最先由Hadamard[1]和Lyapunov[2]以及Perron[3]分别用2种方法提出,这2种方法也为后期以及现代研究不变流形奠定了良好基础.

不变流形在数学理论中也有许多丰硕的成果,Carr[4]扩展了对有限维确定动力系统的不变流形的存在性以及分岔特征的研究,如文献[5-12]将不变流形的研究拓展到了无穷维随机动力系统.Chow和Lu[13-14]以及Duan等[15-16]利用Hadamard方法与Lyapunov-Perron方法探索了在无穷维动力系统上的不变流形的存在性与光滑性.随着对不变流形的研究,文献[17-26]拓展了对带有加性噪声或乘性噪声的随机偏微分方程的研究.此外,文献[27-31]也探究了时滞或延迟反应扩散方程的相关性质.Shi[32]研究了在不同相空间中具有乘性噪声的随机偏微分方程的中心流形的光滑收敛性.

本文将采用文献[32]的框架和研究方法,研究具有乘性噪声的时滞随机演化方程的中心流形.

1 预备知识

本节参考文献[16,33-34]给出随机动力系统与随机不变流形的基本知识,以及方程中每个算子的相关性质.

1.1 随机动力系统

设(Ω,F,P)是概率空间,X是具有范数‖·‖的可分Hilbert空间.用B(R),B(R+),B(X)分别表示R,R+,X的Borel集簇.

定义 1.1

如果(Ω,F,P,(θt)t∈R)满足:

(i) 映射θ:RΩ→Ω是(B(R)F,F)-可测的;

(ii) θ0=idΩ是Ω上的恒等算子,对所有t,s∈R,θt+s=θtθs;

(iii) 对所有t∈R,θtP=P;

则(Ω,F,P,(θt)t∈R)被称为度量动力系统.

定义 1.2

如果一个映射φ:R+×Ω×X→X,(t,ω,x)MT ExtraaA@φ(t,ω,x)满足:

(i) φ是(B(R+)FB(X),B(X))-可测的;

(ii) 映射φ(t,ω)=φ(t,ω,·):X→X在θt上形成一个余环:

φ(0,ω)=idX, ω∈Ω,φ(s+t,ω)=φ(t,θsω)φ(s,ω),s,t∈R+, ω∈Ω;

则φ被称为度量动力系统(Ω,F,P,(θt)t∈R)上的一个随机动力系统.若φ是随机动力系统并且对每个(t,ω)∈R+×Ω,映射

φ(t,ω):X→X, xMT ExtraaA@φ(t,ω)x

是Ck的,则φ被称为一个Ck光滑随机动力系统.

1.2 随机不变流形

如果非空闭集M(ω)X,ω∈Ω上的多值函数M=(M(ω))ω∈Ω满足

ωMT ExtraaA@infy∈Ω‖x-y‖

是对每个x∈X的随机变量,则M(ω)被称为随机集.若M(ω)是流形,则称M是随机流形;若M(ω)对每个ω是Ck光滑流形,则称M是Ck随机流形.

定义 1.3

若对于随机动力系统满足

(t,ω,M(ω))M(θtω), t≥0,则称M是随机流形.

1.3 Hilbert空间X上的无界线性算子A

设A是Hilbert空间X上的一簇线性算子,并且满足:

(i) A是稠定线性算子,并且在X上生成一个解析半群;

(ii) A是自伴的且具有紧预解集,

则σ(A)仅由多重有限的特征值{λn}∞n=1组成

λ1≥λ2≥…≥λn≥…→-∞,

其对应的特征向量{φn}∞n=1构成空间X的一组标准正交基.

将谱σ(A)写成

σ(A)=σu(A)∪σc(A)∪σs(A),

其中

σu(A):={λ∈σ(A)|λ≥λm},σc(A):={λ∈σ(A)|λm+l-1≤λ≤λm+1},σs(A):={λ∈σ(A)|λ≤λm+l}.

假设λm>0,λm+l<0以及σc(A)≠.用Xu、Xc、Xs分别表示σu、σc、σs对应的特征空间,则有

X=XuXcXs,

对应投影算子Pu:X→Xu,Pc:X→Xc,Ps:X→Xs.

由A生成的线性半群eAt有如下性质.

引理 1.4

对每个0<α

‖eAtPu‖L(X,X)≤eλmt, t≤0,‖eAtPc‖L(X,X)≤eα|t|, t∈R,‖eAtPs‖L(X,X)≤eλm+lt, t≥0.

1.4 非线性算子H与F

考虑非线性算子H:X→X,F:X→X.假设H和F是全局Lipschitz连续并且一致Lipschitz有界,即

L1=supx≠y,x,y∈X‖H(x)-H(y)‖X‖x-y‖X<∞,(2)

L2=supx≠y,x,y∈X‖F(x)-F(y)‖X‖x-y‖X<∞,(3)

假设H(0)=0,F(0)=0.

为了研究Ck范数下中心流形的光滑性,进一步假设H、F从X到X是Ck的,其中整数k≥1,即H、F是Ck可微的并且每阶导数DiH、DiF是一致有界的,且导数一致连续的,Li(X,X)是从X到X的所有i阶有界线性算子的空间.

设C([-ρ,0],X)是从[-ρ,0]到X的所有连续函数的空间,其中ρ>0,且有范数

‖φ‖C([-ρ,0],X)=sup-ρ≤s≤0{‖φ(s)‖X:s∈[-ρ,0],φ∈C([-ρ,0],X)}.

为了简洁,用Cρ表示C([-ρ,0],X).

2 中心流形的存在性

本节证明以下方程的随机中心流形存在:

dv(t)dt=Av(t)+H(v(t))+F(v(t-ρ))+v(t),(4)

其中,A、H(v(t))、F(v(t-ρ))在前面已给出,W(t)是一维标准Wiener过程,表示一般的白噪声,v(t)是Stratonovich微分.但是,由参考文献[35]的第七章可知,不变流形存在性理论通常应用于It方程.随机演化方程(4)的等价It方程为

dv(t)=Av(t)dt+(H(v(t))+F(v(t-ρ)))dt+v(t)2dt+v(t)d.(5)

为了研究由方程(4)的解生成的随机动力系统,考虑一维线性随机微分方程

dz+zdt=dW,(6)

这个方程的解被称为Ornstein-Uhlenbeck过程,并且满足以下性质.

引理 2.1[16] 上述过程具有以下性质:

(i) 存在全测度{θt}t∈R不变集

Ω1∈B(C0(R,R)),其具有次线性增长

limt→±∞|ω(t)||t|=0, ω∈Ω1,

在P-测度1下.

(ii) 对于ω∈Ω1的随机变量

z(ω)=-∫0-∞eτω(τ)dτ

存在并生成了方程(6)的唯一稳态解,表示为

Ω1×R瘙綍(ω,t)→z(θtω)=-∫0-∞eτθtω(τ)dτ=-∫0-∞eτω(τ+t)dτ+ω(t),

且映射t→z(θtω)是连续的.

(iii) 特别地,有

limt→±∞|z(θtω)||t|=0, ω∈Ω1.

(iv) 除此之外,有

limt→±∞1t∫t0z(θτω)dτ=0, ω∈Ω1.

对余下部分,限制θt在全测度不变集Ω1上,而不是Ω上,定义对应的概率空间(Ω1,F,P),但依旧表示为(Ω,F,P).

接下来,研究方程(4)的解定义一个随机动力系统,为了证明这一点,考虑如下随机偏微分方程

du(t)dt=Au(t)+z(θtω)u(t)+e-z(θtω)H(ez(θtω)u(t))+e-z(θtω)F(ez(θt-ρω)u(t-ρ)),(7)

初值条件为

u(s)=x(s), s∈[-ρ,0],(8)

其中x(s)∈Cρ.与原始随机微分方程相比,此方程未出现随机积分.通过解的存在唯一性定理,该方程对每个ω∈Ω都有唯一解,则解映射

(t,ω,x)MT ExtraaA@u(t,ω,x)

生成一个随机动力系统,即u是

B(R+)FB(Cρ)可测,且生成余环:

u(0,ω,x)=x, ω∈Ω,u(t+s,ω,x)=u(t,θsω,·)u(s,ω,x),s,t∈R+, ω∈Ω, x∈Cρ.

对每个x∈Cρ,ω∈Ω,引入下列随机变换

T(ω,x)=e-z(ω)x.

显然,对固定的ω∈Ω,其逆变换是

T-1(ω,x)=ez(ω)x.

下面这个命题给出了原始随机方程(4)的解生成一个随机动力系统.

命题 2.2

假设u是方程(7)解生成的一个随机动力系统,则

(t,ω,x)MT ExtraaA@T-1(θtω,·)u(t,ω,T(ω,x))=:v(t,ω,x)(9)

是一个随机动力系统.对于任意x∈Cρ,这个过程是方程(4)的一个解并生成一个随机动力系统.

证明

这个证明过程与文献[32]相似,所以此处省略.

注意到,谱σ(A)仅由特征值{λn}∞n=1组成,且有

λ1≥λ2≥…≥λn≥…→-∞.

设0<α

Cη={φ:C(R,X)|supt∈Re-η|t|-∫t0z(θrω)dr‖φ(t)‖X<∞},

且具有范数

‖φ‖Cη=supt∈Re-η|t|-∫t0z(θrω)dr‖φ(t)‖X.

设L是一个正常数且满足

L1+NL2≤L<∞,(10)

其中N=max{Nu,Nc,Ns},这些系数将在后文中具体给出.为了保证中心流形的存在性和Ck光滑性,假设

L(1η-α+1λm-η+1-λm+l-η)<1,(11)

谱间隙条件kη

L(1iη-α+1λm-iη+1-λm+l-iη)<1,1≤i≤k.(12)

为了研究中心流形的Ck光滑性,还需要选择κ>0,使得有

L(1i(η±κ)-α+1λm-i(η±κ)+1-λm+l-i(η±κ))<1, 1≤i≤k.(13)

现在考虑方程(7)的随机不变流形.设

Mc(ω)={x|u(·,ω,x)∈Cη},

当它是流形时,则Mc(ω)被称为中心流形.显然,0∈Mc(ω),它是非空的且是不变的.

以下引理表明Mc(ω)中的点可以由一个积分方程所确定.

引理 2.3

对于x∈Mc(ω)当且仅当存在一个初值为u(0)=x的函数u(·)∈Cη,且满足

u(t)=eAt+∫t0z(θrω)drξ+∫t0eA(t-s)+∫tsz(θrω)drPc[e-z(θsω)H(ez(θsω)u(s))+e-z(θsω)F(ez(θs-ρω)u(s-ρ))]ds+∫t+∞eA(t-s)+∫tsz(θrω)drPu[e-z(θsω)H(ez(θsω)u(s))+e-z(θsω)F(ez(θs-ρω)u(s-ρ))]ds+∫t-∞eA(t-s)+∫tsz(θrω)drPs[e-z(θsω)H(ez(θsω)u(s))+e-z(θsω)F(ez(θs-ρω)u(s-ρ))]ds,(14)

其中ξ=Pcx.

证明

设τ,t∈R,x∈Mc(ω),因为对所有t∈R方程的解u(t;x,ω)存在,则有

u(t;x,ω)=eA(t-τ)+∫tτz(θrω)dru(τ;x,ω)+∫tτeA(t-s)+∫tsz(θrω)dr

[e-z(θsω)H(ez(θsω)u(s;x,ω))+e-z(θsω)F(ez(θs-ρω)u(s-ρ;x,ω))]ds.(15)

取τ=0,用投影算子Pc作用于方程(15),得到

Pcu(t;x,ω)=eAt+∫t0z(θrω)drPcx+∫t0eA(t-s)+∫tsz(θrω)drPc

[e-z(θsω)H(ez(θsω)u(s;x,ω))+e-z(θsω)F(ez(θs-ρω)u(s-ρ;x,ω))]ds.(16)

用投影算子Pu作用于方程(15)有

Puu(t;x,ω)=eA(t-τ)+∫tτz(θrω)drPuu(τ;x,ω)+∫tτeA(t-s)+∫tsz(θrω)drPu[e-z(θsω)H(ez(θsω)u(s;x,ω))+e-z(θsω)F(ez(θs-ρω)u(s-ρ;x,ω))]ds.(17)

由引理1.4,当τ>max{t,0}时,且τ→+∞,可得

‖eA(t-τ)+∫tτz(θrω)drPuu(τ;x,ω)‖X

≤eλm(t-τ)+∫tτz(θrω)dr+η|τ|+∫τ0z(θrω)dr‖u(·;x,ω)‖Cη≤

eλmt+∫t0z(θrω)dr+(η-λm)τ‖u(·;x,ω)‖Cη→0.

当τ→+∞时,对方程(17)取极限,有

Puu(t;x,ω)=∫t+∞eA(t-s)+∫tsz(θrω)drPu[e-z(θsω)H(ez(θsω)u(s;x,ω))+e-z(θsω)F(ez(θs-ρω)u(s-ρ;x,ω))]ds.(18)

类似地,得到

Psu(t;x,ω)=∫t-∞eA(t-s)+∫tsz(θrω)drPs[e-z(θsω)H(ez(θsω)u(s;x,ω))+e-z(θsω)F(ez(θs-ρω)u(s-ρ;x,ω))]ds.(19)

综合(16)、(18)以及(19)式,得到方程(14).由简单计算可以得到逆向结论.

接下来证明方程(7)的中心流形的存在性.

定理 2.4

假设(11)式成立,则有:

(i) 对每个ξ∈Xc,方程(14)有唯一解

u(·;ξ,ω)∈Cη±κ,且满足

‖u(·;ξ,ω)-u(·;,ω)‖Cη≤(1-L(1η-α+1λm-η+1-λm+l-η))-1‖ξ-‖Xc;

(ii) Mc(ω)能由Lipschitz映射

hc(·,ω):Xc→XuXs的图表示,即

Mc(ω)={ξ+hc(ξ,ω)|ξ∈Xc},(20)

其中

hc(ξ,ω)=Puu(0;ξ,ω)+Psu(0;ξ,ω),hc(0,ω)=0.

证明

首先证明(i),即证明当ξ∈Xc时,方程(14)在Cη上有唯一解u=u(·;ξ,ω),且满足Lipschitz连续.用Qc(u;ξ,ω)表示方程(14)的右边.通过引理1.4和(10)式,得到

e-η|t|-∫t0z(θrω)dr‖Qc(u;ξ,ω)‖Xc≤‖e-η|t|+Atξ‖Xc+‖∫t0e-η|t|+A(t-s)+∫0sz(θrω)drPc[e-z(θsω)H(ez(θsω)u(s))+e-z(θsω)F(ez(θs-ρω)u(s-ρ))]ds‖Xc+‖∫t+∞e-η|t|+A(t-s)+∫0sz(θrω)drPu[e-z(θsω)H(ez(θsω)u(s))+e-z(θsω)F(ez(θs-ρω)u(s-ρ))]ds‖Xc+‖∫t-∞e-η|t|+A(t-s)+∫0sz(θrω)drPs[e-z(θsω)H(ez(θsω)u(s))+e-z(θsω)F(ez(θs-ρω)u(s-ρ))]ds‖Xc≤‖e-η|t|+Atξ‖Xc+{‖∫t0e-η|t|+α|t-s|+∫0sz(θrω)dre-z(θsω)H(ez(θsω)u(s))ds‖Xc+‖∫t-ρ-ρe-η|t|+α|t-s-ρ|+∫0s+ρz(θrω)dr×e-z(θs+ρω)F(ez(θsω)u(s))ds‖Xc}+{‖∫t+∞e-η|t|+λm(t-s)+∫0sz(θrω)dre-z(θsω)H(ez(θsω)u(s))ds‖Xc+‖∫t-ρ+∞e-η|t|+λm(t-s-ρ)+∫0s+ρz(θrω)dr×e-z(θs+ρω)F(ez(θsω)u(s))ds‖Xc}+{‖∫t-∞e-η|t|+λm+l(t-s)+∫0sz(θrω)dr×e-z(θsω)H(ez(θsω)u(s))ds‖Xc+‖∫t-ρ-∞e-η|t|+λm(t-s-ρ)+∫0s+ρz(θrω)dr×e-z(θs+ρω)F(ez(θsω)u(s))ds‖Xc}≤‖ξ‖Xc+{1η-αL1‖u(·)‖Cη+1η-αL2Nc‖u(·)‖Cη}+{1λm-ηL1‖u(·)‖Cη+1λm-ηL2Nu‖u(·)‖Cη}+{1-λm+l-ηL1‖u(·)‖Cη+1-λm+l-ηL2Ns‖u(·)‖Cη}≤‖ξ‖Xc+L(1η-α+1λm-η+1-λm+l-η)‖u(·)‖Cη,

其中

Nc=sups∈R 2e2ηρe∫ss+ρz(θrω)dre-z(θs+ρω)ez(θsω),Nu=sups∈R 2e2λmρe∫ss+ρz(θrω)dre-z(θs+ρω)ez(θsω),Ns=sups∈R 2e-2λm+lρe∫ss+ρz(θrω)dre-z(θs+ρω)ez(θsω).

这意味着映射Qc(u;ξ,ω)是从Cη到自身的.

接着,证明解是唯一的.对u,∈Cη,有

‖Qc(u;ξ,ω)-Qc(;ξ,ω)‖Cη≤supt∈Re-η|t|-∫t0z(θrω)dr{‖∫t0eA(t-s)+∫tsz(θrω)drPc×[e-z(θsω)H(ez(θsω)u(s))-e-z(θsω)H(ez(θsω)(s))+e-z(θsω)F(ez(θs-ρω)u(s-ρ))-e-z(θsω)F(ez(θs-ρω)(s-ρ))]ds‖Xc+‖∫t+∞eA(t-s)+∫tsz(θrω)drPu[e-z(θsω)H(ez(θsω)u(s))-e-z(θsω)H(ez(θsω)(s))+e-z(θsω)F(ez(θs-ρω)u(s-ρ))-e-z(θsω)F(ez(θs-ρω)(s-ρ))]ds‖Xc+‖∫t-∞eA(t-s)+∫tsz(θrω)drPs[e-z(θsω)H(ez(θsω)u(s))-e-z(θsω)H(ez(θsω)(s))+e-z(θsω)F(ez(θs-ρω)u(s-ρ))-e-z(θsω)F(ez(θs-ρω)(s-ρ))]ds‖Xc}≤{‖∫t0e-η|t|+α|t-s|+∫0sz(θrω)dr[e-z(θsω)H(ez(θsω)u(s))-e-z(θsω)H(ez(θsω)(s))]ds‖Xc+‖∫t-ρ-ρe-η|t|+α|t-s-ρ|+∫0s+ρz(θrω)dr×[e-z(θs+ρω)F(ez(θsω)u(s))-e-z(θs+ρω)F(ez(θsω)(s))]ds‖Xc}+{‖∫t+∞e-η|t|+λm(t-s)+∫0sz(θrω)dr[e-z(θsω)×H(ez(θsω)u(s))-e-z(θsω)H(ez(θsω)(s))]ds‖Xc+‖∫t-ρ+∞e-η|t|+λm(t-s-ρ)+∫0s+ρz(θrω)dr×[e-z(θs+ρω)F(ez(θsω)u(s))-e-z(θs+ρω)F(ez(θsω)(s))]ds‖Xc}+{‖∫t-∞e-η|t|+λm+l(t-s)+∫0sz(θrω)dr[e-z(θsω)×H(ez(θsω)u(s))-e-z(θsω)H(ez(θsω)(s))]ds‖Xc+‖∫t-ρ-∞e-η|t|+λm(t-s-ρ)+∫0s+ρz(θrω)dr×[e-z(θs+ρω)F(ez(θsω)u(s))-e-z(θs+ρω)F(ez(θsω)(s))]ds‖Xc}≤L(1η-α+1λm-η+1-λm+l-η)‖u(·)-(·)‖Cη.

由(11)式可得Qc(·;ξ,ω)是关于(ξ,ω)一致收缩的.根据压缩映射原理,对每个ξ∈Xc,Qc(·;ξ,ω)都有唯一不动点u(·;ξ,ω)∈Cη.

类似地,对所有ξ,∈Xc,有

‖u(t;ξ,ω)-u(t;,ω)‖Cη≤‖ξ-‖Xc+L(1η-α+1λm-η+1-λm+l-η)‖u(·;ξ,ω)-u(·;,ω)‖Cη.

因此

‖u(t;ξ,ω)-u(t;,ω)‖Cη≤(1-L(1η-α+1λm-η+1-λm+l-η))-1‖ξ-‖Xc.(21)

此外,因为u(·;ξ,ω)可以是压缩映射Qc从0开始迭代的ω-向极限,并且将一个F-可测函数映射到一个可测函数,所以u(·;ξ,ω)关于ω是F-可测的.另一方面,因为u(·;ξ,ω)关于ξ是Lipschitz连续的,所以u(·;ξ,ω)关于(ξ,ω)是可测的.

下面证明(ii).设

hc(ξ,ω)=Puu(0;ξ,ω)+Psu(0;ξ,ω),

则有

hc(ξ,ω)=∫0+∞e-As+∫0sz(θrω)drPu[e-z(θsω)H(ez(θsω)u(s))+e-z(θsω)F(ez(θs-ρω)u(s-ρ))]ds+∫0-∞e-As+∫0sz(θrω)drPs[e-z(θsω)H(ez(θsω)u(s))+e-z(θsω)F(ez(θs-ρω)u(s-ρ))]ds,

且hc(0,ω)=0.

运用引理1.4和(10)式,对每个ξ,∈Xc,有

‖hc(ξ,ω)-hc(,ω)‖Xc≤(1λm-η+1-λm+l-η)×L1-L(1η-α+1λm-η+1-λm+l-η)‖ξ-‖Xc,

而且hc关于(ξ,ω)是可测的.

根据引理2.3和hc(ξ,ω)的定义,得到

Mc(ω)={ξ+hc(ξ,ω)|ξ∈Xc}.

现在证明Mc(ω)是随机集,即对任意x∈Cρ,有

ωMT ExtraaA@infy∈X|x-(Pcy+hc(Pcy,ω))|(22)

是可测的.令是可分空间X的可数稠密集,则(22)式的右边等价于

infy∈|x-(Pcy+hc(Pcy,ω))|,

这说明hc(·,ω)是连续的.由于对任意y∈Rn,ωMT ExtraaA@hc(Pcy,ω)是可测的,所以(22)式中下确界的任意表达式都是可测的.

3 中心流形的光滑性

下面证明中心流形的光滑性,运用归纳法和导数定义,首先证明方程解是一阶可微的,其次证明一阶导数存在且连续,最后证明k阶导数存在且连续.

定理 3.1

假设H和F关于u是Ck的,谱间隙条件kη

证明

归纳法证明.首先,当k=1时,根据(11)式,存在一个较小的数κ>0,使得α<η-2κ,且满足对所有0≤δ≤2κ,有

L(1(η-δ)-α+1λm-(η-δ)+1-λm+l-(η-δ))<1.

因为Cη-δCη,对任意0≤δ≤2κ,Qc(·;ξ,ω)在Cη-δ中也是一致压缩的,所以

u(·;ξ,ω)∈Cη-δ.令

S=eAt+∫t0z(θrω)dr,

对∈Xc,有

Tv=∫t0eA(t-s)+∫tsz(θrω)drPcDu×[e-z(θsω)H(ez(θsω)u(s;,ω))+e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω))]vds+∫t+∞eA(t-s)+∫tsz(θrω)drPuDu×[e-z(θsω)H(ez(θsω)u(s;,ω))+e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω))]vds+∫t-∞eA(t-s)+∫tsz(θrω)drPsDu×[e-z(θsω)H(ez(θsω)u(s;,ω))+e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω))]vds,

其中v∈Cη-κ.显然地,S是从X到Cη-κ的有界线性算子.类似于Qc的证明,可知T是一个由Cη-κ到自身的有界线性算子,且有

‖T‖≤L(1(η-κ)-α+1λm-(η-κ)+1-λm+l-(η-κ))<1,

这说明id-T在Cη-κ上是可逆的.对ξ,∈Xc,令

I=∫t0eA(t-s)+∫tsz(θrω)drPc{[e-z(θsω)H(ez(θsω)u(s;ξ,ω))-e-z(θsω)H(ez(θsω)u(s;,ω))+e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω))-e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω))]-[Du(e-z(θsω)H(ez(θsω)u(s;,ω)))(u(s;ξ,ω)-u(s;,ω))+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω)))×(u(s-ρ;ξ,ω)-u(s-ρ;,ω))]}ds+∫t+∞eA(t-s)+∫tsz(θrω)drPu{[e-z(θsω)H(ez(θsω)u(s;ξ,ω))-e-z(θsω)H(ez(θsω)u(s;,ω))+e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω))-e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω))]-[Du(e-z(θsω)H(ez(θsω)u(s;,ω)))(u(s;ξ,ω)-u(s;,ω))+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω)))×(u(s-ρ;ξ,ω)-u(s-ρ;,ω))]}ds+∫t-∞eA(t-s)+∫tsz(θrω)drPs{[e-z(θsω)H(ez(θsω)u(s;ξ,ω))-e-z(θsω)H(ez(θsω)u(s;,ω))+e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω))-e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω))]-[Du(e-z(θsω)H(ez(θsω)u(s;,ω)))×(u(s;ξ,ω)-u(s;,ω))+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω)))×(u(s-ρ;ξ,ω)-u(s-ρ;,ω))]}ds.

断言称,当ξ→时,‖I‖Cη-κ=(‖ξ-‖Xc).根据断言可得

u(·;ξ,ω)-u(·;,ω)-T(u(s;ξ,ω)-u(s;,ω))=S(ξ-)+I=S(ξ-)+(‖ξ-‖Xc),

等价于

u(·;ξ,ω)-u(·;,ω)=(id-T)-1S(ξ-)+(‖ξ-‖Xc),

即知u(·;ξ,ω)关于ξ是可微的.

下面证明上述断言是成立的.令

B=[e-z(θsω)H(ez(θsω)u(s;ξ,ω))-e-z(θsω)H(ez(θsω)u(s;,ω))+e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω))-e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω))]-[Du(e-z(θsω)H(ez(θsω)u(s;,ω)))×(u(s;ξ,ω)-u(s;,ω))+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω)))×(u(s-ρ;ξ,ω)-u(s-ρ;,ω))].

对I的第一个积分项,设Z1是一个足够大的正数,当t≥Z1时,令

I11=e-(η-κ)|t|-∫t0z(θrω)dr‖∫Z10eA(t-s)+∫tsz(θrω)drPcBds‖X,I12=e-(η-κ)|t|-∫t0z(θrω)dr‖∫tZ1eA(t-s)+∫tsz(θrω)drPcBds‖X;

当t≤-Z1时,令

I13=e-(η-κ)|t|-∫t0z(θrω)dr‖∫-Z10eA(t-s)+∫tsz(θrω)drPcBds‖X,I14=e-(η-κ)|t|-∫t0z(θrω)dr‖∫t-Z1eA(t-s)+∫tsz(θrω)drPcBds‖X;

当-Z1≤t≤Z1时,令

I15=e-(η-κ)|t|-∫t0z(θrω)dr‖∫Z1-Z1eA(t-s)+∫tsz(θrω)drPcBds‖X.

对于I的第二项积分,设Z2是一个足够大的正常数,当t≥Z2时,令

I21=e-(η-κ)|t|-∫t0z(θrω)dr‖∫t+∞eA(t-s)+∫tsz(θrω)drPuBds‖X;

当-Z2≤t≤Z2时,令

I22=e-(η-κ)|t|-∫t0z(θrω)dr‖∫Z2+∞eA(t-s)+∫tsz(θrω)drPuBds‖X,I23=e-(η-κ)|t|-∫t0z(θrω)dr‖∫tZ2eA(t-s)+∫tsz(θrω)drPuBds‖X;

当t≤-Z2时,令

I24=e-(η-κ)|t|-∫t0z(θrω)dr‖∫Z2+∞eA(t-s)+∫tsz(θrω)drPuBds‖X,I25=e-(η-κ)|t|-∫t0z(θrω)dr‖∫-Z2Z2eA(t-s)+∫tsz(θrω)drPuBds‖X,I26=e-(η-κ)|t|-∫t0z(θrω)dr‖∫t-Z2eA(t-s)+∫tsz(θrω)drPuBds‖X.

对于I的第三项积分,设Z3是一个足够大的正常数,当t≥Z3时,令

I31=e-(η-κ)|t|-∫t0z(θrω)dr‖∫-Z3-∞eA(t-s)+∫tsz(θrω)drPsBds‖X,I32=e-(η-κ)|t|-∫t0z(θrω)dr‖∫Z3-Z3eA(t-s)+∫tsz(θrω)drPsBds‖X,I33=e-(η-κ)|t|-∫t0z(θrω)dr‖∫tZ3eA(t-s)+∫tsz(θrω)drPsBds‖X;

当-Z3≤t≤Z3时,令

I34=e-(η-κ)|t|-∫t0z(θrω)dr‖∫-Z3-∞eA(t-s)+∫tsz(θrω)drPsBds‖X,I35=e-(η-κ)|t|-∫t0z(θrω)dr‖∫t-Z3eA(t-s)+∫tsz(θrω)drPsBds‖X;

当t≤-Z3时,令

I36=e-(η-κ)|t|-∫t0z(θrω)dr‖∫t-∞eA(t-s)+∫tsz(θrω)drPsBds‖X.

对于I11,取固定的Z1,有

I11=e-(η-κ)|t|-∫t0z(θrω)dr×‖∫Z10eA(t-s)+∫tsz(θrω)drPc{[e-z(θsω)H(ez(θsω)u(s;ξ,ω))-e-z(θsω)H(ez(θsω)u(s;,ω))+e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω))-e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω))]-[Du(e-z(θsω)H(ez(θsω)u(s;,ω)))(u(s;ξ,ω)-u(s;,ω))+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω)))×(u(s-ρ;ξ,ω)-u(s-ρ;,ω))]}ds‖X≤|∫Z10e-(η-κ)|t|+α|t-s|+(η-κ)|s|×{∫10|Du(e-z(θsω)H(ez(θsω)(τu(s;ξ,ω)+(1-τ)u(s;,ω))))-Du(e-z(θsω)H(ez(θsω)))(u(s;ξ,ω))|dτ}ds|×‖u(·;ξ,ω)-u(·;,ω)‖Cη-κ+|∫Z1-ρ-ρe-(η-κ)|t|+α|t-s-ρ|+(η-κ)|s|+∫ss+ρz(θrω)dr×{∫10|Du(e-z(θs+ρω)F(ez(θsω)(τu(s;ξ,ω)+(1-τ)u(s;,ω))))-Du(e-z(θs+ρω)F(ez(θsω)))(u(s;ξ,ω))|dτ}ds|×‖u(·;ξ,ω)-u(·;,ω)‖Cη-κ≤|∫Z10e-αs+(η-κ)s×{∫10|Du(e-z(θsω)H(ez(θsω)(τu(s;ξ,ω)+(1-τ)u(s;,ω))))-Du(e-z(θsω)H(ez(θsω)))(u(s;ξ,ω))|dτ}ds|×(1-L(1(η-κ)-α+1λm-(η-κ)+1-λm+l-(η-κ)))-1‖ξ-‖Xc+|∫Z1-ρ-ρe-αs-αρ+(η-κ)s+∫ss+ρz(θrω)dr×{∫10|Du(e-z(θs+ρω)F(ez(θsω)(τu(s;ξ,ω)+(1-τ)u(s;,ω))))-Du(e-z(θs+ρω)F(ez(θsω)))(u(s;ξ,ω))|dτ}ds|×(1-L(1(η-κ)-α+1λm-(η-κ)+1-λm+l-(η-κ)))-1‖ξ-‖Xc.

由[0,Z1],[-ρ,Z1-ρ]是紧闭区间,以及积分关于(s;ξ)的连续性知,存在β1>0,使得当

‖ξ-‖Xc≤β1时,有

supt≥Z1  I11≤114ζ‖ξ-‖Xc.

根据引理1.4,(2)、(3)、(10)以及(21)式,可得

I12≤2|∫tZ1e-(η-κ)|t|+α|t-s|+(η-κ)|s|e-κ|s|L1ds|‖u(·;ξ,ω)-u(·;,ω)‖Cη-2κ+2|∫t-ρZ1-ρe-(η-κ)|t|+α|t-s-ρ|+(η-κ)|s|e-κ|s|+∫ss+ρz(θrω)dre-z(θs+ρω)L2ez(θsω)ds|‖u(·;ξ,ω)-u(·;,ω)‖Cη-2κ≤2Le-κ(Z1-ρ)(η-κ-α)[1-L(1(η-2κ)-α+1λm-(η-2κ)+1-λm+l-(η-2κ))]‖ξ-‖Xc,

对任意ζ>0,存在Z1足够大时,使得

2Le-κ(Z1-ρ)(η-κ-α)[1-L(1(η-2κ)-α+1λm-(η-2κ)+1-λm+l-(η-2κ))]≤114ζ,

则有

supt≥Z1I12≤114ζ‖ξ-‖Xc.

以上几项均可分成这2种情况,则当‖ξ-‖Xc≤β1时,有

supt≥Z1{I11+I12}+supt≤-Z1{I13+I14}+sup-Z1≤t≤Z1I15≤37ζ‖ξ-‖Xc.

类似地,通过选择足够大的Z2、Z3,以及充分小的β2>0,使得当‖ξ-‖Xc≤β2时,有

supt≥Z2I21+sup-Z2≤t≤Z2{I22+I23}+supt≤-Z2{I24+I25+I26}≤27ζ‖ξ-‖Xc,

以及

supt≥Z3{I31+I32+I33}+sup-Z3≤t≤Z3{I34+I35}+supt≤-Z3I36≤27ζ‖ξ-‖Xc.

令=min{β1,β2},使得当‖ξ-‖Xc≤时,则得到

‖I‖Cη-κ≤ζ‖ξ-‖Xc.

因此,当ξ→时,‖I‖Cη-κ=(‖ξ-‖Xc),即知u(·;ξ,ω)关于ξ是可微的,且导数满足Dξu(t;ξ,ω)∈L(Xc,Cη-κ)以及

Dξu(t;ξ,ω)=eAt+∫t0z(θrω)dr+∫t0eA(t-s)+∫tsz(θrω)drPc[Du(e-z(θsω)×H(ez(θsω)u(s;ξ,ω)))Dξu(s;ξ,ω)+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω)))×Dξu(s-ρ;ξ,ω)]ds+∫t+∞eA(t-s)+∫tsz(θrω)drPu[Du(e-z(θsω)×H(ez(θsω)u(s;ξ,ω)))Dξu(s;ξ,ω)+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω)))×Dξu(s-ρ;ξ,ω)]ds+∫t-∞eA(t-s)+∫tsz(θrω)drPs[Du(e-z(θsω)×H(ez(θsω)u(s;ξ,ω)))Dξu(s;ξ,ω)+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω)))×Dξu(s-ρ;ξ,ω)]ds.

对于ξ,∈Xc,有

Dξu(t;ξ,ω)-Dξu(t;,ω)=∫t0eA(t-s)+∫tsz(θrω)drPc{[Du(e-z(θsω)×H(ez(θsω)u(s;ξ,ω)))Dξu(s;ξ,ω)+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω)))×Dξu(s-ρ;ξ,ω)]-[Du(e-z(θsω)H(ez(θsω)u(s;,ω)))Dξu(s;,ω)+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω)))×Dξu(s-ρ;,ω)]}ds+∫t+∞eA(t-s)+∫tsz(θrω)drPu{[Du(e-z(θsω)×H(ez(θsω)u(s;ξ,ω)))Dξu(s;ξ,ω)+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω)))×Dξu(s-ρ;ξ,ω)]-[Du(e-z(θsω)H(ez(θsω)u(s;,ω)))Dξu(s;,ω)+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω)))×Dξu(s-ρ;,ω)]}ds+∫t-∞eA(t-s)+∫tsz(θrω)drPs{[Du(e-z(θsω)×H(ez(θsω)u(s;ξ,ω)))Dξu(s;ξ,ω)+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω)))×Dξu(s-ρ;ξ,ω)]-[Du(e-z(θsω)H(ez(θsω)u(s;,ω)))Dξu(s;,ω)+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω)))×Dξu(s-ρ;,ω)]}ds=∫t0eA(t-s)+∫tsz(θrω)drPc×[Du(e-z(θsω)H(ez(θsω)u(s;ξ,ω)))×(Dξu(s;ξ,ω)-Dξu(s;,ω))+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω)))×(Dξu(s-ρ;ξ,ω)-Dξu(s-ρ;,ω))]ds+∫t+∞eA(t-s)+∫tsz(θrω)drPu×[Du(e-z(θsω)H(ez(θsω)u(s;ξ,ω)))×(Dξu(s;ξ,ω)-Dξu(s;,ω))+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω)))×(Dξu(s-ρ;ξ,ω)-Dξu(s-ρ;,ω))]ds+∫t-∞eA(t-s)+∫tsz(θrω)drPs×[Du(e-z(θsω)H(ez(θsω)u(s;ξ,ω)))×(Dξu(s;ξ,ω)-Dξu(s;,ω))+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω)))×(Dξu(s-ρ;ξ,ω)-Dξu(s-ρ;,ω))]ds+,

其中

=∫t0eA(t-s)+∫tsz(θrω)drPc×{[Du(e-z(θsω)H(ez(θsω)u(s;ξ,ω)))-Du(e-z(θsω)H(ez(θsω)u(s;,ω)))]Dξu(s;,ω)+[Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω)))-Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω)))]×Dξu(s;,ω)}ds+∫t+∞eA(t-s)+∫tsz(θrω)drPu×{[Du(e-z(θsω)H(ez(θsω)u(s;ξ,ω)))-Du(e-z(θsω)H(ez(θsω)u(s;,ω)))]Dξu(s;,ω)+[Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω)))-Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω)))]×Dξu(s;,ω)}ds+∫t-∞eA(t-s)+∫tsz(θrω)drPs×{[Du(e-z(θsω)H(ez(θsω)u(s;ξ,ω)))-Du(e-z(θsω)H(ez(θsω)u(s;,ω)))]Dξu(s;,ω)+[Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω)))-Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;,ω)))]×Dξu(s;,ω)}ds,

则有

‖Dξu(t;ξ,ω)-Dξu(t;,ω)‖L(Xc,Cη)≤‖‖L(Xc,Cη)1-L(1(η-κ)-α+1λm-(η-κ)+1-λm+l-(η-κ)).

利用上述断言类似地证明,可得当ξ→时,

‖‖L(Xc,Cη)=(1),

则Dξu(·;ξ,ω)从Xc到L(Xc,Cη)是连续的.因此,u(·;·,ω)从Xc到Cη是C1的.

其次,由归纳假设可知,u从Xc到C(k-1)η是Ck-1的,且(k-1)阶导数Dk-1ξu(t;ξ,ω)满足

Dk-1ξu(t;ξ,ω)=∫t0eA(t-s)+∫tsz(θrω)drPc[Du(e-z(θsω)×H(ez(θsω)u(s;ξ,ω)))Dk-1ξu(s;ξ,ω)+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω)))×Dk-1ξu(s-ρ;ξ,ω)]ds+∫t+∞eA(t-s)+∫tsz(θrω)drPu[Du(e-z(θsω)×H(ez(θsω)u(s;ξ,ω)))Dk-1ξu(s;ξ,ω)+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω)))×Dk-1ξu(s-ρ;ξ,ω)]ds+∫t-∞eA(t-s)+∫tsz(θrω)drPs[Du(e-z(θsω)×H(ez(θsω)u(s;ξ,ω)))Dk-1ξu(s;ξ,ω)+Du(e-z(θsω)F(ez(θs-ρω)u(s-ρ;ξ,ω)))×Dk-1ξu(s-ρ;ξ,ω)]ds+∫t0eA(t-s)+∫tsz(θrω)drPc[Rk-1(s;ξ,ω)+k-1(s-ρ;ξ,ω)]ds+∫t+∞eA(t-s)+∫tsz(θrω)drPu[Rk-1(s;ξ,ω)+k-1(s-ρ;ξ,ω)]ds+∫t-∞eA(t-s)+∫tsz(θrω)drPs[Rk-1(s;ξ,ω)+k-1(s-ρ;ξ,ω)]ds,

其中

Rk-1(s;ξ,ω)=∑k-3i=0k-2 iDk-2-iξ×Du(e-z(θsω)H(ez(θsω)u(s;ξ,ω)))Di+1ξu(s;ξ,ω),k-1(s-ρ;ξ,ω)=∑k-3i=0k-2 iDk-2-iξDu(e-z(θsω)F(ez(θs-ρω)×u(s-ρ;ξ,ω)))Di+1ξu(s-ρ;ξ,ω).

最后,证明k阶成立,根据归纳假设知,当i=1,2,…,k-1时,Diξu∈Ciη成立.由于H、F是Ck的,则

Rk-1(s;ξ,ω),k-1(s-ρ;ξ,ω)∈Lk-1(Xc,C(k-1)η)

且关于ξ是C1的,其中Lk-1(Xc,C(k-1)η)是从Xc到C(k-1)η的所有(k-1)阶有界线性算子的空间.又因为谱间隙条件成立,即对任意1≤i≤k,α

L(1iη-α+1λm-iη+1-λm+l-iη)<1.

利用证明k=1时的方法,可知Dk-1ξu(·;·,ω)从Xc到Lk(Xc,Ckη)是C1的.

参考文献

[1] HADAMARD J. Sur literation et les solutions asymptotiques des equations differentielles[J]. Bulletin de la Société Mathématique de France,1901,29:224-228.

[2] LYAPUNOV A. Problème géneral de la stabilité du mouvement[M]. Ewing:
Princeton University Press,1947.

[3] PERRON O. ber stabilitt und asymptotisches verhalten der integrale von differentialgleichungs systemen[J]. Mathematische Zeitschrift,1929,29(1):129-160.

[4] CARR J. Application of centre manifold theory[M]. Berlin:
Springer-Verlag,1981.

[5] BATES P, LU K, ZENG C. Persistence of overflowing manifolds for semiflow[J]. Communications on Pure and Applied Mathematics,1999,52(8):983-1046.

[6] BATES P, LU K, ZENG C. Existence and persistence of invariant manifolds for semiflows in Banach space[J]. Memoirs of the American Mathematical Society,1998,135(645):1-132.

[7] BATES P, LU K, ZENG C. Invariant foliations near normally hyperbolic invariant manifolds for semiflows[J]. Trans Amer Math Soc,2000,352:4641-4676.

[8] BATES P, LU K, ZENG C. Approximately invariant manifolds and global dynamics of spike states[J]. Inventiones Mathematicae,2008,174(2):355-433.

[9] CHUESHOV I, GIRYA T. Inertial manifolds and stationary measures for stochastically perturbed dissipative dynamical systems[J]. Sbornik:
Mathematics,1995,186(1):29-45.

[10] DA PRATO G, DEBUSSCHE A. Construction of stochastic inertial manifolds using backward integration[J]. Stochastics and Stochastic Reports,1996,59(3/4):305-324.

[11] HENRY D. Geometric theory of semilinear parabolic equations[M]. Berlin:
Springer-Verlag,1981.

[12] VANDERBAUWHEDE A, VAN GILS S A. Center manifolds and contractions on a scale of Banach spaces[J]. Journal of Functional Analysis,1987,72(2):209-224.

[13] CHOW S N, LU K N. Ck centre unstable manifolds[J]. Proceedings of the Royal Society of Edinburgh:
Section A Mathematics,1988,108(3/4):303-320.

[14] CHOW S N, LU K N. Invariant manifolds for flows in Banach spaces[J]. Journal of Differential Equations,1988,74(2):285-317.

[15] DUAN J Q, LU K N, SCHMALFUSS B. Invariant manifolds for stochastic partial differential equations[J]. The Annals of Probability,2003,31(4):2109-2135.

[16] DUAN J Q, LU K N, SCHMALFUSS B. Smooth stable and unstable manifolds for stochastic evolutionary equations[J]. Journal of Dynamics and Differential Equations,2004,16(4):949-972.

[17] BENSOUSSAN A, FLANDOLI F. Stochastic inertial manifold[J]. Stochastics and Stochastic Reports,1995,53(1/2):13-39.

[18] CARABALLO T, CHUESHOV I, LANGA J A. Existence of invariant manifolds for coupled parabolic and hyperbolic stochastic partial differential equations[J]. Nonlinearity,2005,18(2):747-767.

[19] CARABALLO T, DUANY J, LU K N, et al. Invariant manifolds for random and stochastic partial differential equations[J]. Advanced Nonlinear Studies,2010,10(1):23-52.

[20] LU K N, SCHMALFU  B. Invariant manifolds for stochastic wave equations[J]. Journal of Differential Equations,2007,236(2):460-492.

[21] LU K N, SCHMALFU B. Invariant foliations for stochastic partial differential equations[J]. Stochastics and Dynamics,2008,8(3):505-518.

[22] MOHAMMED S E A, SCHEUTZOW M K R. The stable manifold theorem for stochastic differential equations[J]. The Annals of Probability,1999,27(2):615-652.

[23] MOHAMMED S E A, ZHANG T S, ZHAO H Z. The stable manifold theorem for semilinear stochastic evolution equations and stochastic partial differential equations[J]. Memoirs of the American Mathematical Society,2008,196(917):1-47.

[24] SCHMALFUSS B. A random fixed point theorem and the random graph transformation[J]. Journal of Mathematical Analysis and Applications,1998,225(1):91-113.

[25] SANTAMARA E. Distance of attractors of evolutionary equations[D]. Madrid:
Universidad Complutense de Madrid,2014.

[26] VARCHON N. Domain perturbation and invariant manifolds[J]. Journal of Evolution Equations,2012,12(3):547-569.

[27] LI D S, SHI L. Upper semicontinuity of random attractors of stochastic discrete complex Ginzburg-Landau equations with time-varying delays in the delay[J]. Journal of Difference Equations and Applications,2018,24(6):872-897.

[28] LI D S, LU K N, WANG B X, et al. Limiting dynamics for non-autonomous stochastic retarded reaction-diffusion equations on thin domains[J]. Discrete & Continuous Dynamical Systems Series-A,2019,39(7):3717-3747.

[29] LI D S, SHI L, ZHAO J. Regular random attractors for non-autonomous stochastic evolution equations with time-varying delays on thin domains[J]. Journal of Mathematical Physics,2020,61(11):112702.

[30] SHI L, WANG R H, LU K N, et al. Asymptotic behavior of stochastic FitzHugh-Nagumo systems on unbounded thin domains[J]. Journal of Differential Equations,2019,267(7):4373-4409.

[31] SHI L, LI X L. Limiting behavior of non-autonomous stochastic reaction-diffusion equations on unbounded thin domains[J]. Journal of Mathematical Physics,2019,60(8):082702.

[32] SHI L. Smooth convergence of random center manifolds for SPDEs in varying phase spaces[J]. Journal of Differential Equations,2020,269(3):1963-2011.

[33] ARNOLD L. Random dynamical system[M]. Berlin:
Springer-Verlag,1998.

[34] ARRIETA J, SANTAMARA E. Estimates on the distance of inertial manifolds[J]. Discrete & Continuous Dynamical Systems Series-A,2014,34(10):3921-3944.

[35] DA PRATO G, ZABCZYK J. Stochastic equations in infinite dimension[M]. Berlin:
Springer-Verlag,1992.

The Existence and Smoothness of Random Center Manifolds for a Class of Delay Stochastic Evolutionary Equations with Multiplicative Noise

YANG Juan1,2, GONG Jiaxin1,2, WU Longyu1,2, SHU Ji1,2

(1. School of Mathematical Science, Sichuan Normal University, Chengdu 610066, Sichuan;

2. V.C. & V.R. Key Lab, Sichuan Normal University, Chengdu 610066, Sichuan)

Abstract:We study the existence and smoothness of random center manifolds for a class of delay stochastic evolutionary equations with multiplicative noise. Due to the effect of delay, we first transform the nonlinear terms with delay and deal with the coefficients generated by the effect of delay, thus the existence of center manifolds is obtained. Then, we use the Lyapunov-Perron method to investigate the smoothness of center manifolds for the equations with delay.

Keywords:delay stochastic evolutionary equations; random center manifolds; multiplicative noise; existence; smoothness2020 MSC:37L25; 60H15

(编辑 余 毅)

Invariant foliations of overflowing manifolds for semiflows in Banach space[J]. Bifurcation Theory and Its Numerical Analysis,1999:1-12.

基金项目:国家自然科学基金(12326414)和四川省科技厅应用基础计划项目(2016JY0204)

*通信作者简介:舒 级(1976—),男,教授,博士,主要从事随机动力系统和偏微分方程的研究,E-mail:shuji2008@hotmail.com

引用格式:杨娟,龚佳鑫,吴隆钰,等. 一类具有乘性噪声的时滞随机演化方程的随机中心流形的存在性与光滑性[J]. 四川师范大学学报(自然科学版),2024,47(5):696-707.

推荐内容

恒微文秘网 https://www.sc-bjx.com Copyright © 2015-2024 . 恒微文秘网 版权所有

Powered by 恒微文秘网 © All Rights Reserved. 备案号:蜀ICP备15013507号-1

Top