数学物理学报, 2024, 44(4): 925-945

三维 Keller-Segel-Stokes 系统的快速信号扩散极限的收敛速率

喻婷,1,*, 冬英,2

1电子科技大学数学科学学院 成都 611731

2西华大学理学院 成都 610039

The Convergence Rate of the Fast Signal Diffusion Limit for a Three-Dimensional Keller-Segel-Stokes System

Yu Ting,1,*, Dong Ying,2

1School of Mathematical Sciences, University of Electronic Science and Technology of China, Chengdu 611731

2School of Science, Xihua University, Chengdu 610039

通讯作者: *喻婷, E-mail:yuting_pde@163.com

收稿日期: 2023-03-7   修回日期: 2023-10-16  

基金资助: 四川省自然科学基金(2022NSFSC1835)

Received: 2023-03-7   Revised: 2023-10-16  

Fund supported: Natural Science Foundation of Sichuan Province(2022NSFSC1835)

作者简介 About authors

冬英,E-mail:dyyd1208@163.com

摘要

该文通过对三维抛物-抛物型 Keller-Segel-Stokes 系统进行合适的能量迭代估计, 证明了当初始细胞质量很小时, 初边值问题的解在快速信号扩散极限过程中以代数速率收敛到相应的抛物-椭圆型 Keller-Segel-Stokes 系统.

关键词: Keller-Segel-Stokes; 快速信号扩散极限; 衰减估计; 收敛速率

Abstract

In this paper, We demonstrates that when the initial cell mass is small, the solution of the initial boundary value problem converges at an algebraic rate to the corresponding parabolic-elliptical Keller-Segel-Stokes system during the fast signal diffusion limit process by performing appropriate energy iterative estimation on the three-dimensional parabolic-parabolic Keller-Segel-Stokes system.

Keywords: Keller-Segel-Stokes; Fast signal diffusion limit; Decay estimate; Convergence rate

PDF (585KB) 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

喻婷, 冬英. 三维 Keller-Segel-Stokes 系统的快速信号扩散极限的收敛速率[J]. 数学物理学报, 2024, 44(4): 925-945

Yu Ting, Dong Ying. The Convergence Rate of the Fast Signal Diffusion Limit for a Three-Dimensional Keller-Segel-Stokes System[J]. Acta Mathematica Scientia, 2024, 44(4): 925-945

1 引言

趋化现象描述了细胞对环境变化的反应, 在诸如细胞聚集、胚胎发育和免疫反应等生物环境中都发挥着重要作用. 因此, 对趋化现象进行深入研究具有重要的实际意义和理论价值. 为了从数学上描述趋化现象, Keller 和 Segel[14] 于 1970 年提出了如下方程组

$\left\{\begin{split}&\partial_{t}n=\Delta n-\nabla\cdot\big(nS(x,n,c)\nabla c\big),\ \ &x\in\Omega,t>0,\\&\partial_{t}c=\Delta c-c+n,\ \ &x\in\Omega,t>0,\\\end{split}\right.$

其中 $n$$c$$S$ 分别表示细胞密度、化学物质浓度和趋化敏感性函数. 在过去 50 年间, Keller-Segel 模型得到了广泛而深入的研究, 包括方程解的整体适定性和大时间行为等[8,9]. 由文献[7,10,22,25,26,37]可知, 对 $S\equiv 1$, 系统 (1.1) 在一维空间情形时的齐次 Neumann 初边值问题的解整体存在且一致有界. 在 $\Omega=\{x\in\mathbb R^{2};\vert x\vert< L\}$ 和径向对称情形, 相应的初边值问题存在临界质量现象: 当初始细胞质量 $\int_{\Omega} n(x,0)\mathrm{d}x <8\pi$ 时, 其初边值问题的解一致有界; 当初始细胞质量 $\int_{\Omega} n(x,0)\mathrm{d}x >8\pi$ 时, 其初边值问题存在爆破解. 当 $\Omega=\{ x\in\mathbb R^{N};N\geq 3,\vert x\vert< L \}$ 时, 相应的齐次 Neumann 初边值问题存在在有限时刻爆破的解. 对于抛物-椭圆型 Keller-Segel 系统 (即系统 $(1.1)_{2}$$0=\Delta c-c+n$ 代替), 相关结果可参见文献[1,12,23,24]. 更一般地, 当趋化敏感性函数 $S\equiv S(n)$ 为一标量函数时, 其渐近行为决定了爆破现象是否发生. 例如, Horstmann 和 Winkler[11] 证明了当 $S(s)\leq C(1+s)^{-\alpha}$$\alpha>1-\frac{2}{N}$ 时, 系统 (1.1) 的齐次 Neumann 初边值问题的解整体存在且一致有界; 而当 $\Omega\subset\mathbb{R}^{N} (N\geq 2)$ 为一个球体, $S(s)>cs^{-\alpha}$$\alpha<1-\frac{2}{N}$ 时, 系统 (1.1) 存在爆破解. 这表明 $\alpha_{c}=1-\frac{2}{N}$ 为临界爆破指标.

悬浮在水滴中的枯草芽孢杆菌种群自发形成羽状聚集体的现象不仅说明了细胞之间存在相互作用, 还揭示了细胞和周围流体介质之间的相互影响. Tuval[29] 等提出用如下趋化-(Navier-)Stokes 系统

$\left\{\begin{split}&\partial_{t}n+u\cdot\nabla n=\Delta n-\nabla\cdot\big(nS\nabla c\big),\\&\partial_{t}c+u\cdot\nabla c=\Delta c-nf(c),\\&\partial_{t}u+\kappa(u\cdot\nabla)u+\nabla P=\Delta u+n\nabla \phi,\\&\nabla\cdot u=0\\\end{split}\right.$

来描述趋化、信号消耗、流体传输以及细胞对流体的相互作用, 其中 $u$$P$ 分别表示流体速度和压力, $f$$\phi$ 为给定参数函数, 分别代表信号消耗率和重力势函数, 系数 $\kappa\geq 0$ 与非线性流体对流强度有关. Duan[5] 和 Winkler[38]等人证实了在二维情形下该系统整体经典解的存在性; Cao 和 Lankeit[3] 提出了在三维和小初值情形下其经典解的存在性. 关于该系统及其变体的研究结果可参见文献[4,31,33,39,40,42,44,45].

当信号由细胞产生而非消耗时, 模型 (1.2) 可变为如下方程组

$\left\{\begin{split}&\partial_{t}n+u\cdot\nabla n=\Delta n-\nabla\cdot\big(nS\nabla c\big),\\&\partial_{t}c+u\cdot\nabla c=\Delta c-c+n,\\&\partial_{t}u+\kappa(u\cdot\nabla)u+\nabla P=\Delta u+n\nabla \phi,\\&\nabla\cdot u=0.\\\end{split}\right.$

根据模型 (1.1) 的研究结果, 有关系统 (1.3) 的大多研究都致力于在体积填充假设

$\vert S(s)\vert \leq \frac {C_{S}}{(1+s)^{\alpha}}, \ \ \alpha>1-\frac{2}{N}$

下考虑解的整体存在性与有界性, 其中 $S$ 可由一般矩阵值敏感性函数 $S(x,n,c)$ (参见文献[3,19,27,30,32,35,36,43]) 代替. 特别地, 在体积填充假设 (1.4) 下 Keller-Segel-Stokes 系统 (即在系统 (1.3) 中, $\kappa =0$) 的齐次 Neumann-Neumann-Dirichlet 初边值问题在二维空间情形存在整体有界经典解, 在三维空间情形存在整体弱解[13,32,35,36].

对于完全抛物型 Keller-Segel-(Navier-)Stokes 系统

$\left\{\begin{split}&\partial_t{n_{\epsilon}}+u_{\epsilon}\cdot\nabla n_{\epsilon}=\Delta n_{\epsilon}-\nabla\cdot\Big(n_{\epsilon}S(x,n_{\epsilon},c_{\epsilon})\cdot\nabla c_{\epsilon}\Big)+f(x,n_{\epsilon},c_{\epsilon}),&&x\in\Omega,\,t>0,\\&\epsilon\partial_t{c_{\epsilon}}+u_{\epsilon}\cdot\nabla c_{\epsilon}=\Delta{c_{\epsilon}}-c_{\epsilon}+n_{\epsilon},\,\,&&x\in\Omega,\,t>0,\\&\partial_t{u_{\epsilon}}+\kappa(u_{\epsilon}\cdot\nabla)u_{\epsilon}+\nabla P_{\epsilon}=\Delta{u_{\epsilon}}+n_{\epsilon}\nabla\phi,\,\,&&x\in\Omega,\,t>0,\\&\nabla\cdot\ {u_{\epsilon}}=0,&&x\in\Omega,\,t>0,\\&\Big(\nabla n_{\epsilon}-n_{\epsilon}S(x,n_{\epsilon},c_{\epsilon})\cdot \nabla c_{\epsilon}\Big)\cdot \nu=\nabla c_{\epsilon}\cdot\nu=0,u_{\epsilon}=0, &&x\in\partial\Omega,\,t>0,\\&n_{\epsilon}(x,0)=n_{0}(x), c_{\epsilon}(x,0)=c_{0}(x), u_{\epsilon}(x,0)=u_{0}(x),&&x\in\Omega,\\\end{split}\right.$

Wang[34] 等证明了当初始值 $(n_{0},c_{0},u_{0})$ 满足 $\Vert n_{0}\Vert_{L^{p}(\Omega)}\leq\delta,\ \Vert \nabla c_{0}\Vert_{L^{m}(\Omega)}\leq\delta,\ \Vert u_{0}\Vert_{L^{s}(\Omega)}\leq\delta$$f\equiv 0$ 时, 其中 $\delta>0$ 充分小, $p,\ m,\ s>N\geq 2$, 系统 (1.5) 的经典解整体存在; 当 $\sup\limits_{\epsilon}\Vert\nabla c_{\epsilon}\Vert_{L^{\lambda}{((0,T);L^{q}(\Omega))}}<\infty$, $\sup\limits_{\epsilon}\Vert u_{\epsilon}\Vert_{L^{\infty}{((0,T);L^{r}(\Omega))}}<\infty$ 时, 其中 $\lambda\in(2,\infty]$, $q>N\geq 2$, $r>\max\{2,N\}$, $\frac{1}{\lambda}+\frac{N}{2q}<\frac{1}{2}$, 解 $(n_{\epsilon},c_{\epsilon},u_{\epsilon},P_{\epsilon})$ 存在子序列收敛于相应的抛物-椭圆型系统

$\left\{\begin{split}&\partial_t{n}+u\cdot\nabla n=\Delta n-\nabla\cdot\Big(nS(x,n,c)\cdot\nabla c\Big)+f(x,n,c),&&x\in\Omega,\,t>0,\\&u\cdot\nabla c=\Delta{c}-c+n,\,\,&&x\in\Omega,\,t>0,\\&\partial_t{u}+\kappa(u\cdot\nabla u)+\nabla P=\Delta{u}+n\nabla\phi,\,\,&&x\in\Omega,\,t>0,\\&\nabla\cdot\ u=0,\,\,&&x\in\Omega,\,t>0,\\&\Big(\nabla n-nS(x,n,c)\cdot \nabla c\Big)\cdot \nu=\nabla c\cdot\nu=0,u=0,&&x\in\partial\Omega,\,t>0,\\&n(x,0)=n_{0}(x),u(x,0)=u_{0}(x),&&x\in\Omega\\\end{split}\right.$

的经典解 $(n,c,u,P)$. 基于这一结论, Li 和 Xiang 在文献[17]中证明了当 $f\equiv 0$, $\kappa=0$, $N=2$ 或 3 时, 系统 (1.5) 的快速信号扩散极限具有指数收敛速率; Li, Xiang 和 Zhou 在文献[18]中进一步研究了 $\kappa\neq 0$ 的情形. 对于系统 (1.5) 及其变体有关快速信号扩散极限的相关结果可参见文献[2,6,15,16,21,28].

2 主要结果

本文研究三维完全抛物型 Keller-Segel-Stokes 系统

$\left\{\begin{split}&\partial_t{n_{\epsilon}}+u_{\epsilon}\cdot\nabla n_{\epsilon}=\Delta n_{\epsilon}-\nabla\cdot\Big(n_{\epsilon}S(x,n_{\epsilon},c_{\epsilon})\cdot\nabla c_{\epsilon}\Big),&&x\in\Omega,\,t>0,\\&\epsilon\partial_t{c_{\epsilon}}+u_{\epsilon}\cdot\nabla c_{\epsilon}=\Delta{c_{\epsilon}}-c_{\epsilon}+n_{\epsilon},&&x\in\Omega,\,t>0,\\&\partial_t{u_{\epsilon}}+\nabla P_{\epsilon}=\Delta{u_{\epsilon}}+n_{\epsilon}\nabla\phi,\,\,&&x\in\Omega,\,t>0,\\&\nabla\cdot\ {u_{\epsilon}}=0,\,\,&&x\in\Omega,\,t>0,\\&\Big(\nabla n_{\epsilon}-n_{\epsilon}S(x,n_{\epsilon},c_{\epsilon})\cdot \nabla c_{\epsilon}\Big)\cdot \nu=\nabla c_{\epsilon}\cdot\nu=0,u_{\epsilon}=0,\,&&x\in\partial\Omega,\,t>0,\\&n_{\epsilon}(x,0)=n_{0}(x),c_{\epsilon}(x,0)=c_{0}(x),u_{\epsilon}(x,0)=u_{0}(x),&&x\in\Omega\\\end{split}\right.$

在初始细胞质量充分小时快速信号扩散极限的收敛速率, 其中 $\Omega\subset \mathbb R^{3}$ 为具有光滑边界的有界区域, $\epsilon\in(0,1)$, 趋化敏感性函数 $S=(S_{ij})_{3\times3}$ 满足

$S_{ij}(x,n_{\epsilon},c_{\epsilon})\in C^{2}(\overline \Omega\times[0,\infty)\times[0,\infty))$

和体积填充假设

$\vert S(x,n_{\epsilon},c_{\epsilon})\vert\leq\frac{C_{S}}{(1+n_{\epsilon})^{\alpha}},$

其中 $C_{S}$$\alpha$ 为常数且 $\alpha>\frac{1}{2}$, $\phi$ 为给定重力势函数且满足

$\phi\in W^{2,\infty}(\Omega).$

对于初始值 $n_{0}$$c_{0}$$u_{0},$ 我们假设

$\left\{\begin{split}&n_{0}\in W^{1,\infty}(\Omega),n_{0}\geq 0, \ n_{0}\not\equiv 0, \ x\in\overline\Omega,\\&c_{0}\in W^{1,\infty}(\Omega),c_{0}\geq 0, \ c_{0}\not\equiv 0, \ x\in\overline\Omega,\\&u_{0}\in W^{2,\infty}(\Omega;\mathbb R^{3}), \ \nabla \cdot u_{0}\equiv 0, \ x\in\Omega, \ u_{0}=0, \ x\in\partial\Omega.\\\end{split}\right.$

在本文中, 我们定义 Stokes 算子 $\mathcal{A}_{p}:=-\mathcal{P}\Delta$, 其定义域 $D(\mathcal{A}_{p}):=W^{2,p}(\Omega)\cap W_{0}^{1,p}(\Omega)\cap L_{\sigma}^{p}({\Omega})$, Helmholtz 投影 $\mathcal{P}:L^{p}(\Omega,\mathbb{R}^{N})\rightarrow L_{\sigma}^{p}({\Omega})$ 为有界线性算子, 无散度向量空间 $L_{\sigma}^{p}({\Omega}):=\overline {\{\varphi\in C_{0}^{\infty}(\Omega,\mathbb{R}^{N});\nabla \cdot \varphi=0 \}} ^{\Vert\cdot\Vert_{L^{p}(\Omega)}},$$p\in(1,\infty)$. 由于在 $D(\mathcal{A}_{p_{1}})\cap D(\mathcal{A}_{p_{2}})$ 上, $\mathcal{A}_{p_{1}}=\mathcal{A}_{p_{2}}$, 故在下文推导过程中省略了下标 $p$.

基于以上基本假设和定义, 本文的主要结论如下

定理 2.1 设 $\Omega\subset \mathbb R^{3}$ 为具有光滑边界的有界区域且条件 (2.2)-(2.5) 成立, 则存在充分小的 $\delta>0$, 使得当

$\Vert n_{0}\Vert_{L^{1}(\Omega)}<\delta$

时, 系统 (2.1) 的解 $(n_{\epsilon},c_{\epsilon},u_{\epsilon},P_{\epsilon})$ 对任意 $p>1$ 都存在常数 $\mu>0$$C_{1}>0$, 使得

$\Vert n_{\epsilon}(\cdot,t)-\overline n_{0}\Vert_{L^{\infty}(\Omega)}+\Vert c_{\epsilon}(\cdot,t)-\overline n_{0}\Vert_{W^{1,p}(\Omega)}+\Vert u_{\epsilon}(\cdot,t)\Vert_{L^{\infty}(\Omega)}\leq C_{1}{\rm e}^{-\mu t},\ \ \ t\in(0,\infty),$

其中 $\overline n_{0}:=\frac{1}{\vert \Omega\vert}\int_{\Omega}n_{0}(x)\mathrm{d}x$. 进而, 当 $\epsilon\rightarrow 0$ 时, 系统 (2.1) 的解 $(n_{\epsilon},c_{\epsilon},u_{\epsilon},P_{\epsilon})$ 将收敛到相应的抛物-椭圆型系统

$\left\{\begin{split}&\partial_t{n}+u\cdot\nabla n=\Delta n-\nabla\cdot\Big(nS(x,n,c)\cdot\nabla c\Big),&&x\in\Omega,\,t>0,\\&u\cdot\nabla c=\Delta{c}-c+n,\,\,&&x\in\Omega,\,t>0,\\&\partial_t{u}+\nabla P=\Delta{u}+n\nabla\phi,\,\,&&x\in\Omega,\,t>0,\\&\nabla\cdot\ u=0,\,\,&&x\in\Omega,\,t>0,\\&\Big(\nabla n-nS(x,n,c)\cdot \nabla c\Big)\cdot \nu=\nabla c\cdot\nu=0,u=0,\,&&x\in\partial\Omega,\,t>0,\\&n(x,0)=n_{0}(x),u(x,0)=u_{0}(x),&&x\in\Omega\\\end{split}\right.$

的解 $(n,c,u,P)$, 且存在常数 $C_{2}>0$, 使得

$\left\{\begin{split}&\Vert n_{\epsilon}(\cdot,t)-n(\cdot,t)\Vert_{L^{2}(\Omega)}+\Vert n_{\epsilon}(\cdot,s)-n(\cdot,s)\Vert_{L^{2}((0,t);W^{1,2}(\Omega))}\leq C_{2}(1+t)^{\frac{1}{2}}\epsilon^{\frac{1}{2}},\\&\Vert c_{\epsilon}(\cdot,s)-c(\cdot,s)\Vert_{L^{2}((0,t);W^{1,2}(\Omega))}\leq C_{2}(1+t)^{\frac{1}{2}}\epsilon^{\frac{1}{2}},\\&\Vert u_{\epsilon}(\cdot,t)-u(\cdot,t)\Vert_{W^{1,2}(\Omega)}+\Vert u_{\epsilon}(\cdot,s)-u(\cdot,s)\Vert_{L^{2}((0,t);W^{1,2}(\Omega))}\leq C_{2}(1+t)^{\frac{1}{2}}\epsilon^{\frac{1}{2}}.\\\end{split}\right.$

进一步, 对于任意的 $\theta\in(\frac{3}{4},1)$$p>2$, 存在常数 $C_{3}:=C_{3}(\theta)>0$$C_{4}:=C_{4}(p)>0$, 使得

$\Vert \mathcal{A}^{\theta}u_{\epsilon}(\cdot,t)-\mathcal{A}^{\theta}u(\cdot,t)\Vert_{L^{2}(\Omega)}\leq C_{3}(1+t)^{\frac{1}{2}}\epsilon^{\frac{1}{2}},\ \ \ t\in(0,\infty),$
$\Vert n_{\epsilon}(\cdot,t)-n(\cdot,t)\Vert_{L^{p}(\Omega)}\leq C_{4}(1+t)^{\frac{1}{2}}\epsilon^{\frac{2}{p^{2}}},\ \ \ t\in(0,\infty),\ \ \ 2<p\leq6,$
$\Vert n_{\epsilon}(\cdot,t)-n(\cdot,t)\Vert_{L^{p}(\Omega)}\leq C_{4}(1+t)^{\frac{1}{2}}\epsilon^{\frac{4}{3p^{2}-8p+12}},\ \ \ t\in(0,\infty),\ \ \ p>6.$

注 2.1 Li, Xiang 和 Zhou 在文献[18]中研究了二维 Keller-Segel-Navier-Stokes 系统在初始细胞质量 $\Vert n_{0}\Vert_{L^{1}(\Omega)}$ 充分小时, 其相应系统的快速信号扩散极限的收敛速率, 而本文针对三维 Keller-Segel-Stokes 系统建立了类似的结论.

注 2.2 由于在条件 (2.2)-(2.5) 下, 存在常数 $C>0$, 使得对所有的 $\epsilon\in(0,1)$$t\in(0,\infty)$, 有$\int_{0}^{t}\int_{\Omega}\partial_{t}{c_{\epsilon}c}\leq C(1+t) $ 成立(见下文引理 3.7), 故系统 (2.1) 的快速信号扩散极限的收敛速率可由指数增长改进为代数增长.

3 预备知识

引理 3.1[17] 假设条件 (2.2)-(2.5) 成立, 则对所有 $\epsilon\in(0,1)$, 在 $\Omega\times(0,\infty)$ 上都有 $n_{\epsilon}\geq 0$, $c_{\epsilon}\geq0$, 且

$\Vert n_{\epsilon}(\cdot,t)\Vert_{L^{1}(\Omega)}=\Vert n_{0}\Vert_{L^{1}(\Omega)}$

$\Vert c_{\epsilon}(\cdot,t)\Vert_{L^{1}(\Omega)}\leq\max\Big\{\Vert n_{0}\Vert_{L^{1}(\Omega)}, \ \Vert c_{0}\Vert_{L^{1}(\Omega)}\Big\}.$

引理 3.2[17] 假设条件 (2.2)-(2.5) 成立且 $\alpha>\frac{1}{2}$, 则对任意 $s>1$, 都存在常数 $C>0$, 使得

$\Vert n_{\epsilon}(\cdot,t)\Vert_{L^{s}(\Omega)}\leq C,\ \ \ t\in(0,\infty).$

引理 3.3[17] 假设条件 (2.2)-(2.5) 成立且 $\alpha>\frac{1}{2}$. 若存在常数 $p>1$$Q>0$ 使得对所有 $t\in(0,\infty)$, 都有 $\Vert n_{\epsilon}(\cdot,t)\Vert_{L^{p}(\Omega)}\leq Q$ 成立, 则存在只依赖于 $p$, $Q$$c_{0}$ 的常数 $C(p,Q,c_{0})>0$, 使得对任意 $\epsilon\in(0,1)$, 都有

$\Vert c_{\epsilon}(\cdot,t)\Vert_{L^{p}(\Omega)}\leq C(p,Q,c_{0}),\ \ \ t\in(0,\infty).$

引理 3.4 假设条件 (2.2)-(2.5) 成立且 $\alpha>\frac{1}{2}$, 则对任意 $s>1$, 都存在常数 $C>0$, 使得

$\Vert c_{\epsilon}(\cdot,t)\Vert_{L^{s}(\Omega)}\leq C,\ \ \ t\in(0,\infty).$

结合引理 3.2 和引理 3.3 可得.

引理 3.5[17] 假设条件 (2.2)-(2.5) 成立且 $\alpha>\frac{1}{2}$, 则存在常数 $C>0$, 使得

$\Vert\nabla c_{\epsilon}(\cdot,t)\Vert_{L^{2}(\Omega)}\leq C,\ \ \ t\in(0,\infty).$

引理 3.6[17] 假设条件 (2.2)-(2.5) 成立且 $\alpha>\frac{1}{2}$, 则对于任意 $q>1$, 都存在常数 $C>0$, 使得对任意 $\epsilon\in(0,1)$, 都有

$\Vert u_{\epsilon}(\cdot,t)\Vert_{L^{q}(\Omega)}\leq C,\ \ \ t\in(0,\infty).$

引理 3.7[17] 假设条件 (2.2)-(2.5) 成立, 则存在常数 $C>0$, 使得对所有 $\epsilon\in(0,1)$, 都有

$\int_{0}^{t}\int_{\Omega}\partial_{t}{c_{\epsilon}c}\leq C(1+t),\ \ \ t\in(0,\infty).$

引理 3.8 假设条件 (2.2)-(2.5) 成立, 则对任意 $m>1$, 都存在常数 $C>0$, 使得对所有 $\epsilon\in(0,1)$, 都有

$\Vert\nabla c_{\epsilon}(\cdot,t)\Vert_{L^{m}(\Omega)}\leq C,\ \ \ t\in(0,\infty).$

首先考虑 $m>\frac{3}{2}$ 的情形. 将 $\nabla$ 应用于方程 $(2.1)_{2}$, 并对所得方程两边同时乘以 $\vert\nabla c_{\epsilon}\vert^{2(m-1)}\nabla c_{\epsilon}$ 可得

$\begin{matrix}&\frac{\epsilon}{2m}\frac{\mathrm{d}}{\mathrm{d}t}\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2m}-\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2(m-1)}\nabla c_{\epsilon}\cdot\Delta\nabla c_{\epsilon}+\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2m} \\=&\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2(m-1)}\nabla c_{\epsilon}\cdot \nabla n_{\epsilon}-\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2(m-1)}\nabla c_{\epsilon}\cdot \nabla (u_{\epsilon}\cdot\nabla c_{\epsilon}).\end{matrix}$

利用 $\nabla c_{\epsilon}\cdot\nabla\Delta c_{\epsilon}=\frac{1}{2}\Delta\vert\nabla c_{\epsilon}\vert^{2}-\vert D^{2}c_{\epsilon}\vert^{2}$ 和分部积分公式可知, 对所有 $t\in(0,\infty)$, 有

$\begin{matrix}&\frac{\epsilon}{2m}\frac{\mathrm{d}}{\mathrm{d}t}\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2m}+\frac{m-1}{2}\int_{\Omega}\vert\nabla c_{\epsilon}|^{2(m-2)}\vert\nabla\vert\nabla c_{\epsilon}\vert^{2}|^{2}+\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2(m-1)}\vert D^{2}c_{\epsilon}\vert^{2}+\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2m} \\=&\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2(m-1)}\nabla n_{\epsilon}\cdot \nabla c_{\epsilon}+(m-1)\int_{\Omega}(u_{\epsilon}\cdot\nabla c_{\epsilon})\vert\nabla c_{\epsilon}\vert^{2(m-2)}\nabla c_{\epsilon}\cdot\nabla\vert\nabla c_{\epsilon}\vert^{2} \\& +\int_{\Omega}(u_{\epsilon}\cdot\nabla c_{\epsilon})\vert\nabla c_{\epsilon}\vert^{2(m-1)}\Delta c_{\epsilon}+\frac{1}{2}\int_{\partial \Omega}\vert\nabla c_{\epsilon}\vert^{2(m-1)}\frac{\partial \vert\nabla c_{\epsilon}\vert^{2}}{\partial \nu}.\end{matrix}$

对于 (3.2) 式右边的第 $1$ 项, 利用 $|\Delta c_{\epsilon}|^{2}\leq 3|D^{2}c_{\epsilon}|^{2}$, 分部积分公式和 Young 不等式可得

$\begin{matrix}& \, \int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2(m-1)}\nabla n_{\epsilon}\cdot \nabla c_{\epsilon} \\&=-\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2(m-1)} n_{\epsilon}\Delta c_{\epsilon}-(m-1)\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2(m-2)}n_{\epsilon}\nabla c_{\epsilon}\cdot\nabla\vert\nabla c_{\epsilon}\vert^{2} \\&\leq \sqrt{3}\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2(m-1)} n_{\epsilon}\vert D^{2} c_{\epsilon}\vert+(m-1)\int_{\Omega}n_{\epsilon}\vert\nabla c_{\epsilon}\vert^{2m-3}\vert\nabla\vert\nabla c_{\epsilon}\vert^{2}\vert \\&\leq \frac{1}{2}\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2(m-1)}\vert D^{2} c_{\epsilon}\vert^{2}+\frac{m-1}{4}\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2(m-2)}\vert\nabla\vert\nabla c_{\epsilon}\vert^{2}\vert^{2}+(m+\frac{1}{2})\int_{\Omega}n_{\epsilon}^{2}\vert\nabla c_{\epsilon}\vert^{2(m-1)}.\end{matrix}$

对于 (3.2) 式右边的第 $2$ 项和第 $3$ 项, 分别利用 Young 不等式可得

$\begin{matrix}&(m-1)\int_{\Omega}(u_{\epsilon}\cdot\nabla c_{\epsilon})\vert\nabla c_{\epsilon}\vert^{2(m-2)}\nabla c_{\epsilon}\cdot\nabla\vert\nabla c_{\epsilon}\vert^{2} \\\leq\ &\frac{m-1}{8}\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2(m-2)}\vert\nabla\vert\nabla c_{\epsilon}\vert^{2}\vert^{2}+2(m-1)\int_{\Omega}\vert u_{\epsilon}\vert^{2}\vert\nabla c_{\epsilon}\vert^{2m}\end{matrix}$

$\begin{matrix}\int_{\Omega}(u_{\epsilon}\cdot\nabla c_{\epsilon})\vert\nabla c_{\epsilon}\vert^{2(m-1)}\Delta c_{\epsilon}&\leq \sqrt{3}\int_{\Omega}\vert u_{\epsilon}\vert\vert\nabla c_{\epsilon}\vert^{2m-1} \vert D^{2} c_{\epsilon}\vert \\&\leq\frac{1}{2}\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2(m-1)} \vert D^{2} c_{\epsilon}\vert^{2}+\frac{3}{2}\int_{\Omega}\vert u_{\epsilon}\vert^{2}\vert\nabla c_{\epsilon}\vert^{2m}.\end{matrix}$

对于 (3.2) 式右边的最后一项, 结合文献[20,引理 4.2], 即 $\nabla\vert\nabla c_{\epsilon}\vert^{2}\cdot\nu\leq 2C_{\Omega}\vert\nabla c_{\epsilon}\vert^{2}$, 其中 $C_{\Omega}$ 为区域 $\Omega$ 的边界曲率的上确界, 再运用迹定理、Gagliardo-Nirenberg 不等式和 Young 不等式以及引理 3.5 可知, 存在常数 $C_{1},C_{2},C_{3},C_{4}>0$, 使得

$\begin{matrix}\frac{1}{2}\int_{\partial \Omega}\vert\nabla c_{\epsilon}\vert^{2(m-1)}\frac{\partial \vert\nabla c_{\epsilon}\vert^{2}}{\partial \nu}&\leq C_{1}\int_{\partial \Omega}\vert\nabla c_{\epsilon}\vert^{2m}\leq C_{2}\Vert\vert\nabla c_{\epsilon}\vert^{m}\Vert_{W^{\frac{3}{4},2}(\Omega)}^{2} \\& \leq C_{3}\Vert\nabla\vert\nabla c_{\epsilon}\vert^{m}\Vert_{L^{2}(\Omega)}^{\frac{3(2m-1)}{3m-1}}\Vert\vert\nabla c_{\epsilon}\vert^{m}\Vert_{L^{\frac{2}{m}}(\Omega)}^{\frac{1}{3m-1}}+C_{3}\Vert\vert\nabla c_{\epsilon}\vert^{m}\Vert_{L^{\frac{2}{m}}(\Omega)}^{2} \\& \leq \frac{m-1}{4m^{2}}\int_{\Omega}\vert\nabla\vert\nabla c_{\epsilon}\vert^{m}\vert^{2}+C_{4}.\end{matrix}$

将 (3.3)、(3.4)、(3.5) 和 (3.6) 式代入 (3.2) 式可得, 存在常数 $C_{5}>0$, 使得

$\begin{matrix} &\epsilon\frac{\mathrm{d}}{\mathrm{d}t}\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2m}+\frac{m-1}{2m}\int_{\Omega}\vert\nabla\vert\nabla c_{\epsilon}\vert^{m}\vert^{2}+2m\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2m} \\\leq\ & C_{5}\int_{\Omega}\vert u_{\epsilon}\vert^{2}\vert\nabla c_{\epsilon}\vert^{2m}+C_{5}\int_{\Omega}n_{\epsilon}^{2}\vert\nabla c_{\epsilon}\vert^{2(m-1)}+C_{5}.\end{matrix}$

对于 (3.7) 式右边的第 $1$ 项, 运用 Hölder 不等式、引理 3.6、Gagliardo-Nirenberg 不等式、引理 3.5 和 Young 不等式可知, 存在常数 $C_{6},C_{7},C_{8}>0$, 使得

$\begin{matrix} C_{5}\int_{\Omega} \vert u_{\epsilon}\vert ^{2} \vert \nabla c_{\epsilon} \vert ^{2m}&\leq C_{5}\Vert \vert u_{\epsilon}\vert ^{2}\Vert_{L^{2}(\Omega)}\Vert \vert \nabla c_{\epsilon}\vert^{2m}\Vert_{L^{2}(\Omega)} \\&=C_{5}\Vert u_{\epsilon}\Vert_{L^{4}(\Omega)}^{2}\Vert \vert \nabla c_{\epsilon}\vert^{m}\Vert_{L^{4}(\Omega)}^{2}\leq C_{6}\Vert \vert \nabla c_{\epsilon}\vert^{m}\Vert_{L^{4}(\Omega)}^{2} \\&\leq C_{7}\Big(\Vert\nabla\vert\nabla c_{\epsilon}\vert^{m}\Vert_{L^{2}(\Omega)}^{\frac{3(2m-1)}{3m-1}}\Vert \vert \nabla c_{\epsilon}\vert^{m}\Vert_{L^{\frac{2}{m}}(\Omega)}^{\frac{1}{3m-1}}+\Vert \vert \nabla c_{\epsilon}\vert^{m}\Vert_{L^{\frac{2}{m}}(\Omega)}^{2}\Big) \\&\leq\frac{m-1}{4m}\Vert\nabla\vert\nabla c_{\epsilon}\vert^{m}\Vert_{L^{2}(\Omega)}^{2}+ C_{8}.\end{matrix}$

对于 (3.7) 式右边的第 $2$ 项, 运用 Hölder 不等式、引理 3.2、Gagliardo-Nirenberg 不等式、引理 3.5 和 Young 不等式可知, 存在常数 $C_{9}, C_{10}, C_{11}>0$, 使得

$\begin{matrix}C_{5}\int_{\Omega}n_{\epsilon}^{2}\vert\nabla c_{\epsilon}\vert^{2(m-1)}&\leq C_{5}\Vert n_{\epsilon}\Vert_{L^{4}(\Omega)}^{2}\Vert\nabla c_{\epsilon}\Vert_{L^{4(m-1)}(\Omega)}^{2(m-1)} \\&\leq C_{9}\Vert\nabla c_{\epsilon}\Vert_{L^{4(m-1)}(\Omega)}^{2(m-1)}=C_{9}\Vert\vert\nabla c_{\epsilon}\vert^{m}\Vert_{L^{\frac{4(m-1)}{m}}(\Omega)}^{\frac{2(m-1)}{m}} \\&\leq C_{10}\Big(\Vert\nabla\vert\nabla c_{\epsilon}\vert^{m}\Vert_{L^{2}(\Omega)}^{\frac{3(2m-3)}{3m-1}}\Vert \vert \nabla c_{\epsilon}\vert^{m}\Vert_{L^{\frac{2}{m}}(\Omega)}^{\frac{m+2}{m(3m-1)}}+\Vert \vert \nabla c_{\epsilon}\vert^{m}\Vert_{L^{\frac{2}{m}}(\Omega)}^{\frac{2(m-1)}{m}}\Big) \\&\leq \frac{m-1}{4m}\Vert\nabla\vert\nabla c_{\epsilon}\vert^{m}\Vert_{L^{2}(\Omega)}^{2}+ C_{11}.\end{matrix}$

将 (3.8)、(3.9) 式代入 (3.7) 式可得, 存在常数 $C_{12}>0$, 使得

$\epsilon\frac{\mathrm{d}}{\mathrm{d}t}\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2m}+2m\int_{\Omega}\vert\nabla c_{\epsilon}\vert^{2m}\leq C_{12}.$

综上可得, 对于任意 $m>\frac{3}{2}$, $\Vert \nabla c_{\epsilon}\Vert_{L^{2m}(\Omega)}$ 有界. 最后, 由 Hölder 不等式即可完成证明.

引理 3.9[17]$\Omega\subset \mathbb R^{3}$ 为具有光滑边界的有界区域且条件 (2.2)-(2.5) 成立, 则存在序列 $\big\{\epsilon _{i}\big\}^{\infty}_{i=1}$, 使得系统 (2.1) 的经典解 $(n,c,u,P)$ 和系统 (2.1) 的解 $(n_{\epsilon _{i}},c_{\epsilon _{i}},u_{\epsilon _{i}},P_{\epsilon _{i}})$ 满足如下性质: 当 $i\rightarrow\infty$ 时,

(1) 在 $C^{0}\big(\overline{\Omega}\times [0,\infty)\big)$ 中, $n_{\epsilon _{i}}\rightarrow n ;$

(2) 在 $L_{\rm loc}^{2}\big((0,\infty);W^{1,2}(\Omega)\big)$ 中, $n_{\epsilon_{i}}\rightharpoonup n ;$

(3) 在 $L_{\rm loc}^{\infty}\big((0,\infty);C^{0}(\overline{\Omega})\big)\bigcap L_{\rm loc}^{2}\big((0,\infty);W^{1,2}(\Omega)\big)$ 中, $c_{\epsilon_{i}}\rightarrow c ;$

(4) 在 $\bigcap _{m>3}L_{\rm loc}^{\infty}\big((0,\infty);W^{1,m}(\Omega)\big)\bigcap L_{\rm loc}^{\infty}\big(\Omega\times(0,\infty)\big)$ 中, $\nabla c_{\epsilon_{i}}\stackrel{*}{\rightharpoonup}\nabla c ;$

(5) 在 $C^{0}\big(\overline{\Omega}\times(0,\infty);\mathbb{R}^{3}\big)\bigcap C_{\rm loc}^{2,1}\big(\overline{\Omega}\times(0,\infty);\mathbb{R}^{3}\big)$ 中, $u_{\epsilon_{i}}\rightarrow u.$

4 衰减估计与收敛速率

引理 4.1 假设条件 (2.2)-(2.5) 成立且 $\epsilon\in(0,1)$, 则存在充分小的常数 $\delta>0$, 使得当$\Vert n_{0}\Vert_{L^{1}(\Omega)}\leq\delta$时, 存在常数 $C>0$$\mu>0$, 使得

$\Vert n_{\epsilon}(\cdot,t)-\overline n_{0}\Vert_{L^{2}(\Omega)}\leq C{\rm e}^{-\mu t},\ \ \ t\in(0,\infty),$

其中 $\overline n_{0}:=\frac{1}{\vert \Omega\vert}\int_{\Omega}n_{0}(x)\mathrm{d}x$.

证 首先, 在方程 $(2.1)_{1}$ 两边同时乘以 $n_{\epsilon}-\overline n_{0}$ 并在 $\Omega$ 上积分, 利用分部积分公式和 Young 不等式以及 (2.3) 式可知

$\begin{matrix}\frac{1}{2}\frac{\mathrm{d}}{\mathrm{d}t}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}+\Vert \nabla n_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}&=\int_{\Omega}\nabla n_{\epsilon}\cdot\Big(n_{\epsilon}S(x,n_{\epsilon},c_{\epsilon})\nabla c_{\epsilon}\Big) \\&\leq\frac{1}{2}\Vert \nabla n_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+\frac{C_{S}^{2}}{2}\int_{\Omega}n_{\epsilon}^{2}\vert \nabla c_{\epsilon}\vert^{2}.\end{matrix}$

再利用 Hölder 不等式可将上式化简为

$\begin{matrix} \frac{\mathrm{d}}{\mathrm{d}t}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}+\Vert \nabla n_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}&\leq C_{S}^{2}\int_{\Omega}n_{\epsilon}^{2}\vert \nabla c_{\epsilon}\vert^{2} \\ &\leq2C_{S}^{2}\int_{\Omega}(n_{\epsilon}-\overline n_{0})^{2}\vert \nabla c_{\epsilon}\vert^{2}+2C_{S}^{2}\int_{\Omega}\overline n_{0}^{2}\vert \nabla c_{\epsilon}\vert^{2} \\ &\leq2C_{S}^{2}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{4}(\Omega)}^{2}\Vert\nabla c_{\epsilon}\Vert_{L^{4}(\Omega)}^{2}+2C_{S}^{2}\overline n_{0}^{2}\Vert\nabla c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}. \end{matrix}$

$\overline n_{\epsilon}:=\frac{1}{\vert \Omega\vert}\int_{\Omega}n_{\epsilon}(\cdot,t)\mathrm{d}x$, 由引理 3.1 和 Poincaré 不等式可得, 存在常数 $C_{1}>0$, 使得

$\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}=\Vert n_{\epsilon}-\overline n_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}\leq C_{1}\Vert\nabla n_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}.$

对于 (4.2) 式右边的第 $1$ 项, 利用 Gagliardo-Nirenberg 不等式、Young 不等式、引理 3.8、插值不等式以及 Hölder 不等式可知, 存在常数 $C_{2},C_{3},C_{4},C_{5},C_{6},C_{7},C_{8}>0$, 使得

$\begin{matrix} & \ 2C_{S}^{2}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{4}(\Omega)}^{2}\Vert\nabla c_{\epsilon}\Vert_{L^{4}(\Omega)}^{2} \\ &\leq C_{2}\big(\Vert \nabla (n_{\epsilon}-\overline n_{0})\Vert_{L^{2}(\Omega)}^{\frac{9}{5}}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{1}(\Omega)}^{\frac{1}{5}}+\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{1}(\Omega)}^{2}\big)\Vert\nabla c_{\epsilon}\Vert_{L^{4}(\Omega)}^{2} \\ &\leq \frac{1}{2}\Vert\nabla n_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+C_{3}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{1}(\Omega)}^{2}\Vert\nabla c_{\epsilon}\Vert_{L^{4}(\Omega)}^{2} \\ &\leq \frac{1}{2}\Vert\nabla n_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+C_{4}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{1}(\Omega)}^{2}\Big(\Vert\nabla c_{\epsilon}\Vert_{L^{6}(\Omega)}^{\frac{3}{2}}\Vert\nabla c_{\epsilon}\Vert_{L^{2}(\Omega)}^{\frac{1}{2}}\Big) \\ &\leq \frac{1}{2}\Vert\nabla n_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+C_{5}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{1}(\Omega)}^{2}\Vert\nabla c_{\epsilon}\Vert_{L^{2}(\Omega)}^{\frac{1}{2}} \\ &\leq \frac{1}{2}\Vert\nabla n_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+C_{6}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{1}(\Omega)}^{\frac{8}{3}}+\frac{1}{8C_{1}}\Vert\nabla c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2} \\ &\leq\frac{1}{2}\Vert\nabla n_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+C_{7}\delta^{\frac{2}{3}}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{1}(\Omega)}^{2}+\frac{1}{8C_{1}}\Vert\nabla c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2} \\ &\leq\frac{1}{2}\Vert\nabla n_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+C_{8}\delta^{\frac{2}{3}}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}+\frac{1}{8C_{1}}\Vert\nabla c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2} \\ &\leq\frac{1}{2}\Vert\nabla n_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+\frac{1}{4C_{1}}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}+\frac{1}{8C_{1}}\Vert\nabla c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}, \end{matrix}$

这里假设 $\delta>0$ 充分小, 使得 $\delta<\min\Big\{\Big(\frac{1}{4C_{1}C_{8}}\Big)^{\frac{3}{2}},\Big(\frac{1}{8C_{1}C_{9}}\Big)^{\frac{1}{2}}\Big\}.$

将 (4.4) 式代入 (4.2) 式并结合 (4.3) 式可得, 存在常数 $C_{9}>0$, 使得

$\begin{matrix} \frac{\mathrm{d}}{\mathrm{d}t}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}+\frac{1}{4C_{1}}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}&\leq C_{9}\Vert n_{0}\Vert_{L^{1}(\Omega)}^{2}\Vert\nabla c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+\frac{1}{8C_{1}}\Vert\nabla c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}\\ &\leq \big(C_{9}\delta^{2}+\frac{1}{8C_{1}}\big)\Vert\nabla c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}\\ &\leq\frac{1}{4C_{1}}\Vert \nabla c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}. \end{matrix}$

为了处理 (4.5) 式右边的 $\frac{1}{4C_{1}}\Vert \nabla c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}$, 在方程 $(2.1)_{2}$ 两边同时乘以 $c_{\epsilon}-\overline n_{0}$ 并在 $\Omega$ 上积分, 利用分部积分公式、Hölder 不等式和 Young 不等式可得

$\begin{matrix} & \ \frac{\epsilon}{2}\frac{\mathrm{d}}{\mathrm{d}t}\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}+\Vert \nabla c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2} \\ &=\int_{\Omega}(n_{\epsilon}-\overline n_{0})(c_{\epsilon}-\overline n_{0}) \leq\frac{1}{2}\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}+\frac{1}{2}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}. \end{matrix}$

从而得到

$\epsilon\frac{\mathrm{d}}{\mathrm{d}t}\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}+2\Vert \nabla c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}\leq\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}.$

将 (4.7) 式和 (4.5) 式结合起来, 可得

$\frac{\mathrm{d}}{\mathrm{d}t}\big(\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}+\frac{\epsilon}{8C_{1}}\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}\big)+\frac{1}{8C_{1}}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}+\frac{1}{8C_{1}}\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}\leq 0.$

假设 $0<\epsilon<\min\{1,8C_{1}\}$, 若令

$h(t):=\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}+\frac{\epsilon}{8C_{1}}\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2},$

$h^{\prime}(t)+\frac{1}{8C_{1}}h(t)\leq 0.$

对 (4.8) 式关于时间在 $0$$t$ 上积分可得, 存在常数 $C_{10}>0$, 使得

$h(t)\leq {\rm e}^{-{\frac{t}{8C_{1}}}}h(0)\leq {\rm e}^{-{\frac{t}{8C_{1}}}}\big(\Vert n_{0}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}+\frac{1}{8C_{1}}\Vert c_{0}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}\big)\leq C_{10}{\rm e}^{-{\frac{t}{8C_{1}}}}.$

证毕.

引理 4.2 在引理 4.1 的假设条件下, 存在常数 $C>0$$\mu>0$, 使得

$||c_{\epsilon}(\cdot,t)-\overline n_{0}||_{L^{2}(\Omega)}\leq C{\rm e}^{-\mu t},\ \ \ t\in(0,\infty).$

由引理 4.1 证明过程中的 (4.7) 式并结合引理 4.1 可得, 存在常数 $C_{1}>0$, $\mu_{1}>0$, 使得

$\epsilon\frac{\mathrm{d}}{\mathrm{d}t}\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}+2\Vert \nabla c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}\leq\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2}\leq C_{1}{\rm e}^{-\mu_{1}t}.$

不妨设 $0<\epsilon<\min\{1,\frac{1}{2\mu_{1}}\}$. 若令$g(t):=\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{2},$$\epsilon g^{\prime}(t)+g(t)\leq C_{1}{\rm e}^{-\mu_{1}t}.$

直接计算得

$g(t)\leq g(0){\rm e}^{-{\frac{t}{\epsilon}}}+\frac{C_{1}{\rm e}^{-{\frac{t}{\epsilon}}}}{\epsilon}\int_{0}^{t}{\rm e}^{({\frac{1}{\epsilon}}-\mu_{1})s}\mathrm{d}s\leq g(0){\rm e}^{-{\frac{t}{\epsilon}}}+2C_{1}{\rm e}^{-\mu_{1}t}\leq C_{2}{\rm e}^{-\min\{\mu_{1},1\}t}.$

证毕.

推论 4.1 在引理 4.1 的假设条件下, 存在常数 $C>0$$\mu>0$, 使得对任意 $p>1$ 都有

$\Vert c_{\epsilon}(\cdot,t)-\overline n_{0}\Vert_{W^{1,p}(\Omega)}\leq C{\rm e}^{-\mu t},\ \ \ t\in(0,\infty).$

对任意固定的 $q>p>4$, 利用 Gagliardo-Nirenberg 不等式、引理 3.4、引理 3.8、引理 3.9 和引理 4.2 可得, 存在常数 $C_{1},C_{2},C_{3},\mu>0$, 使得

$\begin{matrix} \Vert c_{\epsilon}-\overline n_{0}\Vert_{W^{1,p}(\Omega)}&\leq C_{1}\Vert c_{\epsilon}-\overline n_{0}\Vert_{W^{2,q}(\Omega)}^{\beta}\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{1-\beta}+C_{1}\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)} \\ &\leq C_{1}\Big(\Vert c_{\epsilon}\Vert_{W^{2,q}(\Omega)}+\Vert\overline n_{0}\Vert_{W^{2,q}(\Omega)}\Big)^{\beta}\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{1-\beta}+C_{1}\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)} \\ &\leq C_{2}\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}^{1-\beta}+C_{1}\Vert c_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)} \\ &\leq C_{3}{\rm e}^{-\mu t}, \end{matrix}$

其中 $\beta={\frac{q(5p-6)}{p(7q-6)}}\in(\frac{1}{2},1)$. 对于 $1<p\leq 4$ 的情形, 由插值不等式并结合引理 3.4 和引理 3.8 可得. 证毕.

引理 4.3 在引理 4.1 的假设条件下, 存在常数 $C>0$$\mu>0$, 使得

$\Vert n_{\epsilon}(\cdot,t)-\overline n_{0}\Vert_{L^{\infty}(\Omega)}\leq C{\rm e}^{-\mu t},\ \ \ t\in(0,\infty)$

$\Vert u_{\epsilon}(\cdot,t)\Vert_{L^{\infty}}\leq C{\rm e}^{-\mu t},\ \ \ t\in(0,\infty).$

$({\rm e}^{t\Delta})_{t\geq 0}$ 为在 $\Omega$ 内的齐次 Neumann 热半群. 由

$n_{\epsilon}(\cdot,t)={\rm e}^{t\Delta}n_{0}-\int_{0}^{t}{\rm e}^{(t-s)\Delta}\Big(\nabla\cdot\big(n_{\epsilon}S(\cdot,n_{\epsilon},c_{\epsilon})\nabla c_{\epsilon}\big)+u_{\epsilon}\cdot\nabla n_{\epsilon}\Big)(\cdot,s)\mathrm{d}s$

并结合 ${\rm e}^{t\Delta}\overline n_{0}=\overline n_{0}$ 可得

$\begin{matrix} & \ \Vert n_{\epsilon}(\cdot,t)-\overline n_{0}\Vert_{L^{\infty}(\Omega)} \\ &\leq \Vert {\rm e}^{t\Delta}( n_{0}-\overline n_{0})\Vert_{L^{\infty}(\Omega)}+\int_{0}^{t}\Vert {\rm e}^{(t-s)\Delta}\nabla\cdot\Big(n_{\epsilon}S(\cdot,n_{\epsilon},c_{\epsilon})\nabla c_{\epsilon}\Big)(\cdot,s)\Vert_{L^{\infty}(\Omega)}\mathrm{d}s \\ &\ \ \ \ +\int_{0}^{t}\Vert {\rm e}^{(t-s)\Delta}u_{\epsilon}(\cdot,s)\cdot\nabla n_{\epsilon}(\cdot,s)\Vert_{L^{\infty}(\Omega)}\mathrm{d}s \\ &=:Q_{1}+Q_{2}+Q_{3}. \end{matrix}$

下面利用 Neumann 热半群理论 (参见文献[41,引理 1.3]) 分别估计 $Q_{1}$, $Q_{2}$$Q_{3}$.$\int_{\Omega}(n_{0}-\overline n_{0})\mathrm{d}x=0$ 可得, 存在常数 $C_{1},C_{2}>0$, 使得

$Q_{1}\leq C_{1} {\rm e}^{-\lambda_{1} t}\Vert n_{0}-\overline n_{0} \Vert_{L^{\infty}(\Omega)}\leq C_{2} {\rm e}^{-\lambda_{1} t}.$

对于 $Q_{2}$, 存在常数 $C_{3}>0$, 使得

$Q_{2}\leq C_{3}C_{S}\int_{0}^{t}\Big(1+(t-s)^{-\frac{1}{2}-\frac{3}{8}}\Big){\rm e}^{-\lambda_{1} (t-s)}\Vert n_{\epsilon}(\cdot,s)\cdot\nabla c_{\epsilon}(\cdot,s)\Vert_{L^{4}(\Omega)}\mathrm{d}s.$

对于 (4.14) 式右边的 $\Vert n_{\epsilon}(\cdot,s)\cdot\nabla c_{\epsilon}(\cdot,s)\Vert_{L^{4}(\Omega)}$, 利用 Hölder 不等式并结合引理 3.2 和推论 4.1 可得, 存在常数 $C_{4},\mu_{1}>0$, 使得

$\begin{matrix} \Vert n_{\epsilon}(\cdot,s)\cdot\nabla c_{\epsilon}(\cdot,s)\Vert_{L^{4}(\Omega)}&\leq \Vert n_{\epsilon}(\cdot,s)\Vert_{L^{6}(\Omega)}\Vert\nabla c_{\epsilon}(\cdot,s)\Vert_{L^{12}(\Omega)} \\ &=\Vert n_{\epsilon}(\cdot,s)\Vert_{L^{6}(\Omega)}\Vert\nabla (c_{\epsilon}(\cdot,s)-\overline n_{0})\Vert_{L^{12}(\Omega)}\leq C_{4}{\rm e}^{-\mu_{1}s}. \end{matrix}$

将 (4.15) 式代入 (4.14) 式可得, 存在常数 $C_{5},C_{6}>0$, 使得

$Q_{2}\leq C_{5}\int_{0}^{t}\Big(1+(t-s)^{-\frac{1}{2}-\frac{3}{8}}\Big){\rm e}^{-\lambda_{1} (t-s)}{\rm e}^{-\mu_{1}s}\mathrm{d}s\leq C_{6}{\rm e}^{-\min\{\lambda_{1},\mu_{1}\}t}.$

对于$Q_{3}$, 对任意固定的 $3<m<4$, 存在常数 $C_{7}>0$, 使得

$\begin{matrix} Q_{3}&=\int_{0}^{t}\Vert {\rm e}^{(t-s)\Delta}\nabla \cdot \Big((n_{\epsilon}-\overline n_{0})u_{\epsilon}\Big)(\cdot,s)\Vert_{L^{\infty}(\Omega)}\mathrm{d}s \\ &\leq C_{7}\int_{0}^{t}\Big(1+(t-s)^{-\frac{1}{2}-\frac{3}{2m}}\Big){\rm e}^{-\lambda_{1} (t-s)}\Vert u_{\epsilon}(\cdot,s)(n_{\epsilon}(\cdot,s)-\overline n_{0})\Vert_{L^{m}(\Omega)}\mathrm{d}s. \end{matrix}$

对于 (4.17) 式右边的 $\Vert u_{\epsilon}(\cdot,s)(n_{\epsilon}(\cdot,s)-\overline n_{0})\Vert_{L^{m}(\Omega)}$, 利用插值不等式、Hölder 不等式并结合引理 4.1、引理 3.6 和引理 3.2 可知, 存在常数 $C_{8},C_{9},\mu_{2}>0$, 使得

$\begin{matrix} & \ \Vert u_{\epsilon}(\cdot,s)(n_{\epsilon}(\cdot,s)-\overline n_{0})\Vert_{L^{m}(\Omega)}\\ &\leq \Vert u_{\epsilon}(\cdot,s)(n_{\epsilon}(\cdot,s)-\overline n_{0})\Vert_{L^{\frac{3}{2}}(\Omega)}^{\frac{12-3m}{5m}}\Vert u_{\epsilon}(\cdot,s)(n_{\epsilon}(\cdot,s)-\overline n_{0})\Vert_{L^{4}(\Omega)}^{\frac{8m-12}{5m}} \\ &\leq C_{8}(\Vert u_{\epsilon}(\cdot,s)\Vert_{L^{6}(\Omega)}\Vert n_{\epsilon}(\cdot,s)-\overline n_{0}\Vert_{L^{2}(\Omega)})^{\frac{12-3m}{5m}}(\Vert u_{\epsilon}(\cdot,s)\Vert_{L^{6}(\Omega)}\Vert n_{\epsilon}(\cdot,s)-\overline n_{0}\Vert_{L^{12}(\Omega)})^{\frac{8m-12}{5m}} \\ &\leq C_{9}{\rm e}^{-\mu_{2}s}. \end{matrix}$

将 (4.18) 式代入 (4.17) 式可得, 存在常数 $C_{10},C_{11}>0$, 使得

$Q_{3}\leq C_{10}\int_{0}^{t}\Big(1+(t-s)^{-\frac{1}{2}-\frac{3}{2m}}\Big){\rm e}^{-\lambda_{1} (t-s)}{\rm e}^{-\mu_{2}s}\mathrm{d}s\leq C_{11}{\rm e}^{-\min\{\lambda_{1},\mu_{2}\}t}.$

将 (4.13)、(4.16) 和 (4.19) 式代入 (4.12) 式, 即证得 (4.10) 式.

为证明 (4.11) 式, 利用 $u_{\epsilon}$ 的常数变易法并结合 $\mathcal{P}(\overline n_{0} \nabla \phi)=0$ 可得

$u_{\epsilon}(\cdot,t)={\rm e}^{-t\mathcal{A}}u_{0}+\int_{0}^{t}{\rm e}^{-(t-s)\mathcal{A}}\mathcal{P}\Big((n_{\epsilon}-\overline n_{0}) \nabla \phi\Big)(\cdot,s)\mathrm{d}s.$

进而可知, 存在常数 $C_{12},\mu_{3}>0$, 使得

$\begin{matrix} \Vert \mathcal{A}^{\beta}u_{\epsilon}\Vert_{L^{2}(\Omega)}&\leq\Vert \mathcal{A}^{\beta}{\rm e}^{-t\mathcal{A}}u_{0}\Vert_{L^{2}(\Omega)}+\int_{0}^{t}\Vert \mathcal{A}^{\beta}{\rm e}^{-(t-s)\mathcal{A}}\mathcal{P}\Big((n_{\epsilon}-\overline n_{0}) \nabla \phi\Big)(\cdot,s)\Vert_{L^{2}(\Omega)}\mathrm{d}s \\ &\leq C_{12}t^{-\beta}{\rm e}^{-\mu_{3}t}\Vert u_{0}\Vert_{L^{2}(\Omega)} \\ &\ \ \ \ +C_{12}\int_{0}^{t}(t-s)^{-\beta}{\rm e}^{-\mu_{3}(t-s)}\Vert\mathcal{P}\Big((n_{\epsilon}-\overline n_{0}) \nabla \phi\Big)(\cdot,s)\Vert_{L^{2}(\Omega)}\mathrm{d}s, \end{matrix}$

其中 $\beta\in(\frac{3}{4},1)$. 再结合引理 4.1 可得, 存在常数 $C_{13},C_{14},\mu_{4}>0$, 使得

$\begin{matrix} \Vert\mathcal{P}\Big((n_{\epsilon}-\overline n_{0}) \nabla \phi\Big)(\cdot,s)\Vert_{L^{2}(\Omega)} &\leq\Vert\Big((n_{\epsilon}-\overline n_{0}) \nabla \phi\Big)(\cdot,s)\Vert_{L^{2}(\Omega)} \\ &\leq C_{13}\Vert n_{\epsilon}-\overline n_{0}\Vert_{L^{2}(\Omega)}\Vert \nabla \phi\Vert_{L^{\infty}(\Omega)}\leq C_{14}{\rm e}^{-\mu_{4}s}. \end{matrix}$

将 (4.22) 式代入 (4.21) 式可得, 存在常数 $C_{15},C_{16},C_{17},\mu_{5}>0$, 使得

$\begin{matrix} \Vert \mathcal{A}^{\beta}u_{\epsilon}\Vert_{L^{2}(\Omega)}&\leq C_{12}t^{-\beta}{\rm e}^{-\mu_{3}t}\Vert u_{0}\Vert_{L^{2}(\Omega)}+C_{12}C_{14}\int_{0}^{t}(t-s)^{-\beta}{\rm e}^{-\mu_{3}(t-s)}{\rm e}^{-\mu_{4}s}\mathrm{d}s \\ &\leq C_{15}{\rm e}^{-\mu_{3}t}+C_{16}{\rm e}^{-\mu_{4}t} \leq C_{17}{\rm e}^{-\mu_{5}t},\quad{\quad{t\in(1,\infty)}}.\end{matrix}$

另一方面, 存在常数 $C_{18}>0$, 使得

$\Vert \mathcal{A}^{\beta}u_{\epsilon}\Vert_{L^{2}(\Omega)}\leq\Vert \mathcal{A}^{\beta}u_{0}\Vert_{L^{2}(\Omega)}+C_{16}{\rm e}^{-\mu_{4}t}\leq C_{18}{\rm e}^{-\mu_{5}t},\quad{\quad{t\in(0,1]}}.$

由 (4.23) 式和 (4.24) 式并利用 $D(\mathcal{A}^{\beta})\hookrightarrow L^{\infty}(\Omega)$, 其中 $\beta\in(\frac{3}{4},1)$, 得证.

$\hat{n}:=n_{\epsilon}-n$, $\hat{c}:=c_{\epsilon}-c$, $\hat{u}:=u_{\epsilon}-u$, $\hat{P}:=P_{\epsilon}-P$, 则 $(\hat{n},\hat{c},\hat{u},\hat{P})$ 满足

$\left\{ \begin{split} \begin{aligned} &\partial_t{\hat{n}}=\Delta \hat{n}-u_{\epsilon}\cdot\nabla \hat{n}-\hat{u}\cdot\nabla n-\nabla \cdot\Big(\hat{n}S(x,n_{\epsilon},c_{\epsilon})\cdot\nabla c_{\epsilon}+nS(x,n_{\epsilon},c_{\epsilon})\cdot\nabla \hat{c}\\ &\ \ \ \ \ \ \ \ \ \ +n\Big(S(x,n_{\epsilon},c_{\epsilon})-S(x,n,c)\Big)\cdot\nabla c\Big),\\ &\epsilon\partial_t{c_{\epsilon}}=\Delta\hat{c}-u_{\epsilon}\cdot\nabla\hat{c}-\hat{u}\cdot\nabla c-\hat{c}+\hat{n},\\ &\partial_t{\hat{u}}=\Delta\hat{u}-\nabla\hat{P}+\hat{n}\nabla\phi,\\ &\nabla\cdot \hat{u}=0 \end{aligned} \end{split} \right.$

且满足初边值条件

$\left\{ \begin{split} \begin{aligned} &\Big(\nabla\hat{n}-\hat{n}S(x,n_{\epsilon},c_{\epsilon})\cdot\nabla c_{\epsilon}-nS(x,n_{\epsilon},c_{\epsilon})\cdot\nabla \hat{c}\\ &\ \ \ \ \ -n\Big(S(x,n_{\epsilon},c_{\epsilon})-S(x,n,c)\Big)\cdot\nabla c\Big)\cdot\nu=0,\\ &\nabla\hat{c}\cdot\nu=0,\hat{u}=0,\quad{x\in \partial \Omega,t>0,}\\ &\hat{n}(x,0)=0,\hat{u}(x,0)=0,\quad{x\in\Omega.} \end{aligned} \end{split} \right.$

引理 4.4[17] 假设条件 (2.2)-(2.5) 成立, 存在常数 $C>0$, 使得对任意 $\epsilon\in(0,1)$ 都有

$\Vert\hat{n}(\cdot,t)\Vert_{L^{2}(\Omega)}+\Vert\hat{u}(\cdot,t)\Vert_{L^{2}(\Omega)}\leq C{\rm e}^{Ct}\epsilon^{\frac{1}{2}},\ \ \ t\in(0,\infty)$

$\Vert\hat{n}\Vert_{L^{2}((0,t);W^{1,2}(\Omega))}+\Vert\hat{c}\Vert_{L^{2}((0,t);W^{1,2}(\Omega))}+\Vert\hat{u}\Vert_{L^{2}((0,t);W^{1,2}(\Omega))}\leq C{\rm e}^{Ct}\epsilon^{\frac{1}{2}},\ \ \ t\in(0,\infty).$

引理 4.5 在引理 4.1 的假设条件下, 存在常数 $C>0$, 使得对任意 $\epsilon\in(0,1)$, 都有

$\Vert\hat{n}(\cdot,t)\Vert_{L^{2}(\Omega)}+\Vert\hat{u}(\cdot,t)\Vert_{L^{2}(\Omega)}\leq C(1+t)^{\frac{1}{2}}\epsilon^{\frac{1}{2}},\ \ \ t\in(0,\infty)$

$\Vert\hat{n}\Vert_{L^{2}((0,t);W^{1,2}(\Omega))}+\Vert\hat{c}\Vert_{L^{2}((0,t);W^{1,2}(\Omega))}+\Vert\hat{u}\Vert_{L^{2}((0,t);W^{1,2}(\Omega))}\leq C(1+t)^{\frac{1}{2}}\epsilon^{\frac{1}{2}},\ \ \ t\in(0,\infty).$

分析泛函

$Q\Vert\hat{n}(\cdot,t)\Vert_{L^{2}(\Omega)}^{2}+\epsilon\Vert c_{\epsilon}(\cdot,t)\Vert_{L^{2}(\Omega)}^{2}+\Vert\hat{u}(\cdot,t)\Vert_{L^{2}(\Omega)}^{2},$

其中 $Q$ 待确定. 首先, 在方程 $(4.25)_{1}$ 两边分别乘以 $\hat{n}$ 并在 $\Omega$ 上积分, 利用分部积分公式、Young 不等式和 Hölder 不等式可得: 存在常数 $C_{1}>0$, 使得

$\begin{matrix} \frac{1}{2}\frac{\mathrm{d}}{\mathrm{d}t}\Vert \hat{n}\Vert_{L^{2}(\Omega)}^{2}+\Vert \nabla\hat{n}\Vert_{L^{2}(\Omega)}^{2}&=\int_{\Omega}\Big((n-\overline n_{0})\hat{u}+\hat{n}S(x,n_{\epsilon},c_{\epsilon})\nabla(c_{\epsilon}-\overline n_{0})\Big)\cdot\nabla\hat{n} \\ & \ \ \ +\int_{\Omega}\Big((n-\overline n_{0})S(x,n_{\epsilon},c_{\epsilon})\nabla\hat{c}+\overline n_{0}S(x,n_{\epsilon},c_{\epsilon})\nabla\hat{c}\Big)\cdot\nabla\hat{n} \\ & \ \ \ +\int_{\Omega}n\Big(S(x,n_{\epsilon},c_{\epsilon})-S(x,n,c)\Big)\cdot\nabla(c-\overline n_{0})\cdot\nabla\hat{n} \\ &\leq\frac{1}{4}\Vert \nabla\hat{n}\Vert_{L^{2}(\Omega)}^{2}+C_{1}\Vert n-\overline n_{0}\Vert_{L^{\infty}(\Omega)}^{2}\Vert\hat{u}\Vert_{L^{2}(\Omega)}^{2} \\ & \ \ \ +C_{1}\Vert\nabla(c_{\epsilon}-\overline n_{0})\Vert_{L^{4}(\Omega)}^{2}\Vert\hat{n}\Vert_{L^{4}(\Omega)}^{2} \\ & \ \ \ +C_{1}\Vert n-\overline n_{0}\Vert_{L^{\infty}(\Omega)}^{2}\Vert\nabla\hat{c}\Vert_{L^{2}(\Omega)}^{2}+C_{1}\overline n_{0}^{2}\Vert\nabla\hat{c}\Vert_{L^{2}(\Omega)}^{2} \\ & \ \ \ +C_{1}\int_{\Omega}n^{2}\vert S(x,n_{\epsilon},c_{\epsilon})-S(x,n,c)\vert^{2}\vert\nabla(c-\overline n_{0})\vert^{2}. \end{matrix}$

对于(4.27) 式右边的第 $3$ 项, 利用 Gagliardo-Nirenberg 不等式可得: 存在常数 $C_{2},C_{3}>0$, 使得

$\begin{matrix} & \ C_{1}\Vert\nabla(c_{\epsilon}-\overline n_{0})\Vert_{L^{4}(\Omega)}^{2}\Vert\hat{n}\Vert_{L^{4}(\Omega)}^{2} \\ &\leq C_{2}\Vert\nabla(c_{\epsilon}-\overline n_{0})\Vert_{L^{4}(\Omega)}^{2}\Big(\Vert\nabla\hat{n}\Vert_{L^{2}(\Omega)}^{\frac{3}{4}}\Vert\hat{n}\Vert_{L^{2}(\Omega)}^{\frac{1}{4}}+\Vert\hat{n}\Vert_{L^{2}(\Omega)}\Big)^{2} \\ &\leq\frac{1}{8}\Vert\nabla\hat{n}\Vert_{L^{2}(\Omega)}^{2}+C_{3}\Big(\Vert\nabla(c_{\epsilon}-\overline n_{0})\Vert_{L^{4}(\Omega)}^{8}+\Vert\nabla(c_{\epsilon}-\overline n_{0})\Vert_{L^{4}(\Omega)}^{2}\Big)\Vert\hat{n}\Vert_{L^{2}(\Omega)}^{2}. \end{matrix}$

对于 (4.27) 式右边的第 $6$ 项, 利用微分中值定理、Hölder 不等式和 Gagliardo-Nirenberg 不等式可得: 存在常数 $C_{4},C_{5},C_{6}>0$, 使得

$\begin{matrix} & \ C_{1}\int_{\Omega}n^{2}\vert S(x,n_{\epsilon},c_{\epsilon})-S(x,n,c)\vert^{2}\vert\nabla(c-\overline n_{0})\vert^{2} \\ &\leq2C_{1}\int_{\Omega}n^{2}\Big(\vert S(x,n_{\epsilon},c_{\epsilon})-S(x,n,c_{\epsilon})\vert^{2}+\vert S(x,n,c_{\epsilon})-S(x,n,c)\vert^{2}\Big)\vert\nabla(c-\overline n_{0})\vert^{2} \\ &\leq2C_{1}\int_{\Omega}n^{2}\Big(\vert\nabla S(x,\eta,c_{\epsilon})\vert^{2}\vert\hat{n}\vert^{2}+\vert\nabla S(x,n,\xi)\vert^{2}\vert\hat{c}\vert^{2}\Big)\vert\nabla(c-\overline n_{0})\vert^{2} \\ &\leq C_{4}\Vert\nabla(c-\overline n_{0})\Vert_{L^{4}(\Omega)}^{2}\Big(\Vert\hat{n}\Vert_{L^{4}(\Omega)}^{2}+\Vert\hat{c}\Vert_{L^{4}(\Omega)}^{2}\Big) \\ &\leq C_{5}\Vert\nabla(c-\overline n_{0})\Vert_{L^{4}(\Omega)}^{2}\Big(\Vert\nabla\hat{n}\Vert_{L^{2}(\Omega)}^{\frac{3}{4}}\Vert\hat{n}\Vert_{L^{2}(\Omega)}^{\frac{1}{4}}+\Vert\hat{n}\Vert_{L^{2}(\Omega)}\Big)^{2} \\ &\ \ \ \ +C_{5}\Vert\nabla(c-\overline n_{0})\Vert_{L^{4}(\Omega)}^{2}\Big(\Vert\nabla\hat{c}\Vert_{L^{2}(\Omega)}^{\frac{3}{4}}\Vert\hat{c}\Vert_{L^{2}(\Omega)}^{\frac{1}{4}}+\Vert\hat{c}\Vert_{L^{2}(\Omega)}\Big)^{2} \\ &\leq \frac{1}{8}\Vert\nabla\hat{n}\Vert_{L^{2}(\Omega)}^{2}+\rho\Vert\nabla\hat{c}\Vert_{L^{2}(\Omega)}^{2} \\ &\ \ \ \ + C_{6}\Big(\Vert\nabla(c-\overline n_{0})\Vert_{L^{4}(\Omega)}^{8}+\Vert\nabla(c-\overline n_{0})\Vert_{L^{4}(\Omega)}^{2}\Big)\Big(\Vert\hat{n}\Vert_{L^{2}(\Omega)}^{2}+\Vert\hat{c}\Vert_{L^{2}(\Omega)}^{2}\Big), \end{matrix}$

其中 $\eta$ 存在于 $n_{\epsilon}$$n$ 之间, $\xi$ 存在于 $c_{\epsilon}$$c$ 之间, $\rho>0$ 待确定. 结合 (4.27)、(4.28) 和 (4.29) 式并利用推论 4.1、引理 4.3 和引理 3.9 可得: 存在常数 $C_{7},\mu_{1}>0$, 使得

$\begin{matrix} \frac{\mathrm{d}}{\mathrm{d}t}\Vert \hat{n}\Vert_{L^{2}(\Omega)}^{2}+\Vert \nabla\hat{n}\Vert_{L^{2}(\Omega)}^{2}&\leq2\rho\Vert\nabla\hat{c}\Vert_{L^{2}(\Omega)}^{2}+C_{7}\Big({\rm e}^{-\mu_{1}t}+\overline n_{0}^{2}\Big)\Vert\nabla\hat{c}\Vert_{L^{2}(\Omega)}^{2} \\ & \ \ \ +C_{7}{\rm e}^{-\mu_{1}t}\Vert \hat{c}\Vert_{L^{2}(\Omega)}^{2}+C_{7}{\rm e}^{-\mu_{1}t}\Big(\Vert \hat{n}\Vert_{L^{2}(\Omega)}^{2}+\Vert \hat{u}\Vert_{L^{2}(\Omega)}^{2}\Big). \end{matrix}$

其次, 在方程 $(4.25)_{2}$ 两边分别乘以 $\hat{c}$ 并在 $\Omega$ 上积分, 利用分部积分公式、Young 不等式和 Hölder 不等式可得: 存在常数 $C_{8},C_{9}>0$, 使得

$\begin{matrix} & \ \frac{\epsilon}{2}\frac{\mathrm{d}}{\mathrm{d}t}\Vert c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+\Vert \nabla\hat{c}\Vert_{L^{2}(\Omega)}^{2}+\Vert \hat{c}\Vert_{L^{2}(\Omega)}^{2} \\ &=\epsilon\int_{\Omega}\partial_{t}c_{\epsilon}c+\int_{\Omega}\Big(\hat{n}-\hat{u}\cdot\nabla(c-\overline n_{0})\Big)\hat{c} \\ &\leq \epsilon\int_{\Omega}\partial_{t}c_{\epsilon}c+\Vert \hat{n}\Vert_{L^{2}(\Omega)}\Vert \hat{c}\Vert_{L^{2}(\Omega)}+\Vert \hat{u}\cdot\nabla(c-\overline n_{0})\Vert_{L^{2}(\Omega)}\Vert \hat{c}\Vert_{L^{2}(\Omega)} \\ &\leq\epsilon\int_{\Omega}\partial_{t}c_{\epsilon}c+\Vert \hat{n}\Vert_{L^{2}(\Omega)}\Vert \hat{c}\Vert_{L^{2}(\Omega)}+\Vert \hat{u}\Vert_{L^{4}(\Omega)}\Vert\nabla(c-\overline n_{0})\Vert_{L^{4}(\Omega)}\Vert \hat{c}\Vert_{L^{2}(\Omega)} \\ &\leq\epsilon\int_{\Omega}\partial_{t}c_{\epsilon}c+\frac{1}{2}\Vert \hat{c}\Vert_{L^{2}(\Omega)}^{2}+\Vert \hat{n}\Vert_{L^{2}(\Omega)}^{2}+\Vert \hat{u}\Vert_{L^{4}(\Omega)}^{2}\Vert\nabla(c-\overline n_{0})\Vert_{L^{4}(\Omega)}^{2} \\ &\leq \epsilon\int_{\Omega}\partial_{t}c_{\epsilon}c+\frac{1}{2}\Vert \hat{c}\Vert_{L^{2}(\Omega)}^{2}+\Vert \hat{n}\Vert_{L^{2}(\Omega)}^{2} \\ &\ \ \ +C_{8}\Vert\nabla(c-\overline n_{0})\Vert_{L^{4}(\Omega)}^{2}\Big(\Vert\nabla\hat{u}\Vert_{L^{2}(\Omega)}^{\frac{3}{4}}\Vert\hat{u}\Vert_{L^{2}(\Omega)}^{\frac{1}{4}}+\Vert\hat{u}\Vert_{L^{2}(\Omega)}\Big)^{2} \\ &\leq\epsilon\int_{\Omega}\partial_{t}c_{\epsilon}c+\frac{1}{2}\Vert \hat{c}\Vert_{L^{2}(\Omega)}^{2}+\Vert \hat{n}\Vert_{L^{2}(\Omega)}^{2}+\frac{1}{4}\Vert\nabla\hat{u}\Vert_{L^{2}(\Omega)}^{2} \\ &\ \ \ +C_{9}\Big(\Vert\nabla(c-\overline n_{0})\Vert_{L^{4}(\Omega)}^{8}+\Vert\nabla(c-\overline n_{0})\Vert_{L^{4}(\Omega)}^{2}\Big)\Vert\hat{u}\Vert_{L^{2}(\Omega)}^{2}. \end{matrix}$

$\int_{\Omega}\hat{n}=0$ 并根据 Poincaré 不等式、推论 4.1 和引理 3.9 可知, 存在常数 $C_{10},\mu_{2}>0$, 使得

$\begin{matrix} &\epsilon\frac{\mathrm{d}}{\mathrm{d}t}\Vert c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+2\Vert \nabla\hat{c}\Vert_{L^{2}(\Omega)}^{2}+\Vert \hat{c}\Vert_{L^{2}(\Omega)}^{2} \\ \leq\ &2\epsilon\int_{\Omega}\partial_{t}c_{\epsilon}c+\frac{1}{2}\Vert\nabla\hat{u}\Vert_{L^{2}(\Omega)}^{2}+C_{10}\Vert \nabla\hat{n}\Vert_{L^{2}(\Omega)}^{2}+C_{10}{\rm e}^{-\mu_{2}t}\Vert \hat{u}\Vert_{L^{2}(\Omega)}^{2}. \end{matrix}$

最后, 在方程 $(4.25)_{3}$ 两边分别乘以 $\hat{u}$ 并在 $\Omega$ 上积分, 利用分部积分公式、Hölder 不等式、Poincaré 不等式和 Young 不等式可得, 存在常数 $C_{11},C_{12}>0$, 使得

$\begin{matrix} \frac{1}{2}\frac{\mathrm{d}}{\mathrm{d}t}\Vert \hat{u}\Vert_{L^{2}(\Omega)}^{2}+\Vert \nabla\hat{u}\Vert_{L^{2}(\Omega)}^{2}=\int_{\Omega}\hat{n}\nabla\phi\cdot\hat{u}&\leq\Vert\nabla\phi\Vert_{L^{\infty}(\Omega)}\Vert \hat{n}\Vert_{L^{2}(\Omega)}\Vert \hat{u}\Vert_{L^{2}(\Omega)} \\ &\leq C_{11}\Vert \nabla\hat{u}\Vert_{L^{2}(\Omega)}\Vert \nabla\hat{n}\Vert_{L^{2}(\Omega)} \\ &\leq\frac{1}{2}\Vert \nabla\hat{u}\Vert_{L^{2}(\Omega)}^{2}+C_{12}\Vert \nabla\hat{n}\Vert_{L^{2}(\Omega)}^{2}. \end{matrix}$

从而存在常数 $C_{13}>0$, 使得

$\frac{\mathrm{d}}{\mathrm{d}t}\Vert \hat{u}\Vert_{L^{2}(\Omega)}^{2}+\Vert \nabla\hat{u}\Vert_{L^{2}(\Omega)}^{2}\leq C_{13}\Vert \nabla\hat{n}\Vert_{L^{2}(\Omega)}^{2}.$

现取 $T_{1}>0$ 充分大使得

$C_{7}{\rm e}^{-\mu_{1}t}<\frac{1}{6(C_{10}+C_{13})}, \quad{\ \ t\in(T_{1},\infty),}$

并取 $\delta>0$ 充分小使得

$C_{7}\overline n_{0}^{2}<\frac{1}{6(C_{10}+C_{13})}, \quad{\ \ t\in(T_{1},\infty).}$

$\rho=\frac{1}{12(C_{10}+C_{13})},\ \ \ \ \ Q=2(C_{10}+C_{13}).$

由 (4.30)、(4.32) 和 (4.34) 式可得: 存在常数 $C_{14}>0$, 使得

$\begin{matrix} &\frac{\mathrm{d}}{\mathrm{d}t}\Big(Q\Vert\hat{n}\Vert_{L^{2}(\Omega)}^{2}+\epsilon\Vert c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+\Vert\hat{u}\Vert_{L^{2}(\Omega)}^{2}\Big)+\frac{Q}{2} \Vert\nabla\hat{n}\Vert_{L^{2}(\Omega)}^{2}+\Vert\nabla\hat{c}\Vert_{L^{2}(\Omega)}^{2} \\& +\frac{1}{2}\Vert\nabla\hat{u}\Vert_{L^{2}(\Omega)}^{2}+\frac{2}{3}\Vert\hat{c}\Vert_{L^{2}(\Omega)}^{2} \\ \leq\ &2\epsilon\int_{\Omega}\partial_{t}c_{\epsilon}c+C_{14}{\rm e}^{-\mu t}\Big(\Vert \hat{u}\Vert_{L^{2}(\Omega)}^{2}+\Vert \hat{n}\Vert_{L^{2}(\Omega)}^{2}\Big), \end{matrix}$

其中 $\mu:=\min\{\mu_{1},\mu_{2}\}$. 再由 Poincaré 不等式可得, 存在常数 $C_{15},C_{16}>0$, 使得

$\begin{matrix} &\frac{\mathrm{d}}{\mathrm{d}t}\Big(Q\Vert\hat{n}\Vert_{L^{2}(\Omega)}^{2}+\epsilon\Vert c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+\Vert\hat{u}\Vert_{L^{2}(\Omega)}^{2}\Big)+C_{15}\Big(Q\Vert\hat{n}\Vert_{L^{2}(\Omega)}^{2}+\epsilon\Vert c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+\Vert\hat{u}\Vert_{L^{2}(\Omega)}^{2}\Big) \\ &\ \ \ +\frac{1}{4}\Big(Q\Vert\nabla\hat{n}\Vert_{L^{2}(\Omega)}^{2}+\Vert \nabla\hat{c} \Vert_{L^{2}(\Omega)}^{2}+\Vert\nabla\hat{u}\Vert_{L^{2}(\Omega)}^{2}+\Vert\hat{c}\Vert_{L^{2}(\Omega)}^{2}\Big) \\ &\leq C_{16}{\rm e}^{-\mu t}\Big(Q\Vert\hat{n}\Vert_{L^{2}(\Omega)}^{2}+\epsilon\Vert c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+\Vert\hat{u}\Vert_{L^{2}(\Omega)}^{2}\Big)+\epsilon\Big(C_{15}\Vert c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+2\int_{\Omega}\partial_{t}c_{\epsilon}c\Big). \end{matrix}$

现令

$\begin{aligned} g(t):=&Q\Vert\hat{n}\Vert_{L^{2}(\Omega)}^{2}+\epsilon\Vert c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+\Vert\hat{u}\Vert_{L^{2}(\Omega)}^{2},\\ h(t):=&\frac{1}{4}\Big(Q\Vert\nabla\hat{n}\Vert_{L^{2}(\Omega)}^{2}+\Vert \nabla\hat{c}\Vert_{L^{2}(\Omega)}^{2}+\Vert\nabla\hat{u}\Vert_{L^{2}(\Omega)}^{2}+\Vert\hat{c}\Vert_{L^{2}(\Omega)}^{2}\Big),\\ y(t):=&C_{15}\Vert c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}+2\int_{\Omega}\partial_{t}c_{\epsilon}c.\nonumber \end{aligned}$

从而 (4.36) 式转化为

$g^{\prime}(t)+C_{15}g(t)+h(t)\leq C_{16}{\rm e}^{-\mu t}g(t)+\epsilon y(t), \quad{\ \ t\in(T_{1},\infty).}$

现取 $T_{2}>0$ 充分大使得对所有 $t>T_{2}$, 有 $2C_{16}{\rm e}^{-\mu t}<C_{15}$. 从而

$g^{\prime}(t)+\frac{C_{15}}{2}g(t)+h(t)\leq \epsilon y(t), \quad{\ \ t\in(T_{3},\infty),}$

其中 $T_{3}:=\max\big\{T_{1},T_{2}\big\}$. 再由引理 3.7 和引理 4.4 以及引理 3.9 可知: 对所有 $t\in(T_{3},\infty)$, 存在常数 $C_{17},C_{18}>0$, 使得

$\begin{matrix} g(t)&\leq g(T_{3})+\epsilon\int_{T_{3}}^{t}y(s)\mathrm{d}s=g(T_{3})+\epsilon\Big(C_{15}\int_{T_{3}}^{t}\Vert c_{\epsilon}\Vert_{L^{2}(\Omega)}^{2}\mathrm{d}s+2\int_{T_{3}}^{t}\int_{\Omega}\partial_{t}c_{\epsilon}c\mathrm{d}s\Big) \\ &\leq \Big((Q+1)C_{17}{\rm e}^{C_{17}T_{3}}\epsilon+\epsilon\Vert c_{\epsilon}(\cdot,T_{3})\Vert_{L^{2}(\Omega)}^{2}\Big)+\epsilon C_{17}(1+t):=C_{18}(1+t)\epsilon. \end{matrix}$

进而得到

$\int_{T_{3}}^{t} h(s)\mathrm{d}s\leq g(T_{3})+\epsilon\int_{T_{3}}^{t}y(s)\mathrm{d}s\leq C_{18}(1+t)\epsilon, \quad{\ \ t\in(T_{3},\infty).}$

再由 Poincaré 不等式可得 $t\in(T_{3},\infty)$ 时该引理的结论. 对于 $t\in(0,T_{3})$ 的情形可直接由引理 4.4 得到相应结论. 证毕.

推论 4.2 在引理 4.1 的假设条件下, 存在常数 $C>0$, 使得对任意 $\epsilon\in(0,1)$, 都有

$\Vert\nabla\hat{u}(\cdot,t)\Vert_{L^{2}(\Omega)}\leq C(1+t)^{\frac{1}{2}}\epsilon^{\frac{1}{2}},\ \ \ t\in(0,\infty).$

在方程 $(4.25)_{3}$ 两边同时运用 Helmholtz 投影 $\mathcal{P}$, 再将所得方程两边同时乘以 $\mathcal{A}\hat{u}$, 利用 Hölder 不等式和 Young 不等式可得, 存在常数 $C_{1},C_{2},C_{3}>0$, 使得

$\begin{matrix} \frac{1}{2}\frac{\mathrm{d}}{\mathrm{d}t}\Vert \nabla \hat{u}\Vert_{L^{2}(\Omega)}^{2}+\Vert \mathcal{A} \hat{u}\Vert_{L^{2}(\Omega)}^{2}&=\int_{\Omega}\mathcal{P}(\hat{n}\nabla\phi)\cdot \mathcal{A}\hat{u} \\ &\leq C_{1}\Vert \hat{n}\Vert_{L^{2}(\Omega)}\Vert \nabla\phi\Vert_{L^{\infty}(\Omega)}\Vert \mathcal{A} \hat{u}\Vert_{L^{2}(\Omega)} \\ &\leq C_{2}\Vert \hat{n}\Vert_{L^{2}(\Omega)}\Vert \mathcal{A} \hat{u}\Vert_{L^{2}(\Omega)} \\ &\leq \Vert \mathcal{A} \hat{u}\Vert_{L^{2}(\Omega)}^{2}+C_{3}\Vert \hat{n}\Vert_{L^{2}(\Omega)}^{2}. \end{matrix}$

结合引理 4.5 可知: 存在常数 $C_{4}>0$, 使得

$\Vert \nabla \hat{u}(\cdot,t)\Vert_{L^{2}(\Omega)}^{2}\leq 2C_{3}\int_{0}^{t}\Vert \hat{n}(\cdot,s)\Vert_{L^{2}(\Omega)}^{2}\mathrm{d}s\leq C_{4}(1+t)\epsilon.$

证毕.

引理 4.6 在引理 4.1 的假设条件下, 对任意 $\theta\in (\frac{3}{4},1)$, 都存在常数 $C(\theta)>0$, 使得对任意 $\epsilon\in(0,1)$, 都有

$\Vert \mathcal{A}^{\theta}\hat{u}(\cdot,t)\Vert_{L^{2}(\Omega)}\leq C(\theta)(1+t)^{\frac{1}{2}}\epsilon^{\frac{1}{2}},\ \ \ t\in(0,\infty).$

利用 $\hat{u}$ 的常数变易公式可知

$\mathcal{A}^{\theta}\hat{u}(\cdot,t):=\int_{0}^{t}\mathcal{A}^{\theta}{\rm e}^{-(t-s)\mathcal{A}}\mathcal{P}(\hat{n}\nabla\phi)(\cdot,s)\mathrm{d}s.$

若令 $g(\cdot,s):=\mathcal{P}(\hat{n}\nabla\phi)(\cdot,s),$ 则存在常数 $C_{1}>0$, 使得

$\Vert g(\cdot,s)\Vert_{L^{2}(\Omega)}\leq \Vert \hat{n}\Vert_{L^{2}(\Omega)}\Vert\nabla\phi\Vert_{L^{\infty}(\Omega)}\leq C_{1}\Vert \hat{n}\Vert_{L^{2}(\Omega)}.$

结合引理 4.5 可知, 存在常数 $C_{2},C_{3},C_{4}>0$, 使得

$\begin{matrix} \Vert \mathcal{A}^{\theta}\hat{u}(\cdot,t) \Vert_{L^{2}(\Omega)}&\leq C_{2}\int_{0}^{t}(t-s)^{-\theta}{\rm e}^{-\lambda(t-s)}\Vert \hat{n}\Vert_{L^{2}(\Omega)} \mathrm{d}s \\ &\leq C_{3}\epsilon^{\frac{1}{2}}\int_{0}^{t}(1+s)^{\frac{1}{2}}(t-s)^{-\theta}{\rm e}^{-\lambda(t-s)}\mathrm{d}s \\ &\leq C_{4}(1+t)^{\frac{1}{2}}\epsilon^{\frac{1}{2}}. \end{matrix}$

证毕.

引理 4.7 在引理 4.1 的假设条件下, 对任意 $p>2$, 存在常数 $C(p)>0$, 使得对任意 $\epsilon\in(0,1)$, 都有

$\Vert\hat{n}(\cdot,t)\Vert_{L^{p}(\Omega)}\leq C(p)(1+t)^{\frac{1}{2}}\epsilon^{\frac{2}{p^{2}}},\ \ \ t\in(0,\infty),\ \ \ 2<p\leq 6$

$\Vert\hat{n}(\cdot,t)\Vert_{L^{p}(\Omega)}\leq C_{4}(1+t)^{\frac{1}{2}}\epsilon^{\frac{4}{3p^{2}-8p+12}},\ \ \ t\in(0,\infty), \ \ \ p>6.$

在方程 $(4.25)_{1}$ 两边分别乘以 $\hat{n}^{r-1}$ (其中 $r>2$) 并在 $\Omega$ 上积分, 利用分部积分公式、Hölder 不等式、Young 不等式和微分中值定理可知, 存在常数 $C_{1}>0$, 使得

$\begin{matrix} & \ \frac{1}{r}\frac{\mathrm{d}}{\mathrm{d}t}\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{2}(\Omega)}^{2}+\frac{4(r-1)}{r^{2}}\Vert \nabla\hat{n}^{\frac{r}{2}}\Vert_{L^{2}(\Omega)}^{2} \\ &=(r-1)\int_{\Omega}\Big((n-\overline n_{0})\hat{n}^{r-2}\hat{u}+\hat{n}^{r-1}S(x,n_{\epsilon},c_{\epsilon})\nabla(c_{\epsilon}-\overline n_{0})\Big)\cdot\nabla\hat{n} \\ &\ \ \ +(r-1)\int_{\Omega}(n-\overline n_{0})\hat{n}^{r-2}S(x,n_{\epsilon},c_{\epsilon})\nabla\hat{c}\cdot\nabla\hat{n}+(r-1)\int_{\Omega}\overline n_{0}\hat{n}^{r-2}S(x,n_{\epsilon},c_{\epsilon})\nabla\hat{c}\cdot\nabla\hat{n} \\ &\ \ \ +(r-1)\int_{\Omega}n\hat{n}^{r-2}\Big(S(x,n_{\epsilon},c_{\epsilon})-S(x,n,c)\Big)\nabla(c-\overline n_{0})\cdot\nabla\hat{n} \\ &\leq\frac{r-1}{4}\int_{\Omega}\hat{n}^{r-2}\vert\nabla\hat{n}\vert^{2}+C_{1}\Vert n-\overline n_{0}\Vert_{L^{\infty}(\Omega)}^{2}\int_{\Omega}\hat{n}^{r-2}\vert\hat{u}\vert^{2}+C_{1}\int_{\Omega}\hat{n}^{r}\vert\nabla(c_{\epsilon}-\overline n_{0})\vert^{2} \\ &\ \ \ +C_{1}\Vert n-\overline n_{0}\Vert_{L^{\infty}(\Omega)}^{2}\int_{\Omega}\hat{n}^{r-2}\vert\nabla\hat{c}\vert^{2}+C_{1}\overline n_{0}\int_{\Omega}\hat{n}^{r-2}\vert\nabla\hat{c}\vert^{2} \\ &\ \ \ +C_{1}\Vert n\Vert_{L^{\infty}(\Omega)}^{2}\int_{\Omega}\hat{n}^{r-2}\Big(\hat{n}^{2}+\hat{c}^{2}\Big)\vert\nabla(c-\overline n_{0})\vert^{2}. \end{matrix}$

结合引理 4.3、引理 3.9 和推论 4.1 可知, 存在常数 $C_{2},\mu>0$, 使得对任意 $p>1$$t\in(0,\infty)$, 有

$\Vert n-\overline n_{0}\Vert_{L^{\infty}(\Omega)}^{2}\leq C_{2}{\rm e}^{-\mu t}$

$\Vert \nabla(c-\overline n_{0})\Vert_{L^{p}(\Omega)}^{2}\leq C_{2}{\rm e}^{-\mu t}.$

对于 (4.41) 式右边的最后一项, 对任意 $m>r>2$, 利用 Hölder 不等式、插值不等式、(4.43) 式、(4.42) 式、引理 3.9、引理 4.3 和 Young 不等式可得, 存在常数 $C_{3},C_{4},C_{5}>0$, 使得

$\begin{matrix} & \ \int_{\Omega}\hat{n}^{r-2}(\hat{n}^{2}+\hat{c}^{2})\vert\nabla(c-\overline n_{0})\vert^{2} \\ &\leq \Big(\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{2}+\Vert \hat {c}\Vert_{L^{r}(\Omega)}^{2}\Big)\Vert \hat {n}\Vert_{L^{m}(\Omega)}^{r-2}\Vert \nabla(c-\overline n_{0}) \Vert_{L^{\frac{2mr}{(r-2)(m-r)}}(\Omega)}^{2} \\ &\leq\Big(\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{2}+\Vert \hat {c}\Vert_{L^{r}(\Omega)}^{2}\Big)\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{\frac{r(r-2)}{m}}\Vert n_{\epsilon}-n\Vert_{L^{\infty}(\Omega)}^{\frac{(r-2)(m-r)}{m}}\Vert \nabla(c-\overline n_{0}) \Vert_{L^{\frac{2mr}{(r-2)(m-r)}}(\Omega)}^{2} \\ &\leq C_{3}{\rm e}^{-\mu t}\Big(\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{2}+\Vert \hat {c}\Vert_{L^{r}(\Omega)}^{2}\Big)\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{\frac{r(r-2)}{m}} \\ &\leq C_{4}{\rm e}^{-\mu t}\Big(\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{2}+\Vert \hat {c}\Vert_{L^{r}(\Omega)}^{2}\Big)\Big(\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{r-2}+1\Big) \\ &\leq C_{5}{\rm e}^{-\mu t}\Big(\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{r}+\Vert \hat {c}\Vert_{L^{r}(\Omega)}^{r}\Big)+C_{5}{\rm e}^{-\mu t}\Big(\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{2}+\Vert \hat {c}\Vert_{L^{r}(\Omega)}^{2}\Big). \end{matrix}$

类似地, 对于 (4.41) 式右边的第 $3$ 项, 对任意 $m>r>2$, 利用 Hölder 不等式、插值不等式、推论 4.1、(4.42) 式、引理 4.3、引理 3.9 和 Young 不等式可得, 存在常数 $C_{6},C_{7},C_{8}>0$, 使得

$\begin{matrix} \int_{\Omega}\hat{n}^{r}\vert\nabla(c_{\epsilon}-\overline n_{0})\vert^{2}&=\int_{\Omega}\hat{n}^{r-2}\hat{n}^{2}\vert\nabla(c_{\epsilon}-\overline n_{0})\vert^{2} \\ &\leq \Vert \hat {n}\Vert_{L^{r}(\Omega)}^{2}\Vert \hat {n}\Vert_{L^{m}(\Omega)}^{r-2}\Vert \nabla(c_{\epsilon}-\overline n_{0}) \Vert_{L^{\frac{2mr}{(r-2)(m-r)}}(\Omega)}^{2} \\ &\leq\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{2}\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{\frac{r(r-2)}{m}}\Vert n_{\epsilon}-n\Vert_{L^{\infty}(\Omega)}^{\frac{(r-2)(m-r)}{m}}\Vert \nabla(c_{\epsilon}-\overline n_{0}) \Vert_{L^{\frac{2mr}{(r-2)(m-r)}}(\Omega)}^{2} \\ &\leq C_{6}{\rm e}^{-\mu t}\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{2}\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{\frac{r(r-2)}{m}} \\ &\leq C_{7}{\rm e}^{-\mu t}\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{2}\Big(\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{r-2}+1\Big) \\ &\leq C_{8}{\rm e}^{-\mu t}\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{r}+C_{8}{\rm e}^{-\mu t}\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{2}. \end{matrix}$

结合 (4.44)、(4.45) 和 (4.41) 式并利用 Young 不等式可得, 存在常数 $C_{9}>0$, 使得

$\begin{matrix} &\frac{\mathrm{d}}{\mathrm{d}t}\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{2}(\Omega)}^{2}+\frac{3(r-1)}{r}\Vert \nabla\hat{n}^{\frac{r}{2}}\Vert_{L^{2}(\Omega)}^{2} \\ \leq\ & C_{9}{\rm e}^{-\mu t}\Big(\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{r}+\Vert \hat {u}\Vert_{L^{r}(\Omega)}^{r}\Big)+C_{9}\overline n_{0}\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{r}+ C_{9}\Big({\rm e}^{-\mu t}+\overline n_{0}\Big)\Vert \hat {c}\Vert_{W^{1,r}(\Omega)}^{r} \\ &+C_{9}{\rm e}^{-\mu t}\Vert \hat {n}\Vert_{L^{r}(\Omega)}^{2}+C_{9}{\rm e}^{-\mu t}\Vert \hat {c}\Vert_{L^{r}(\Omega)}^{2}. \end{matrix}$

现假设 $C_{9}>1$$\delta>0$ 充分小使得 $\overline n_{0}< 1$, 利用 Gagliardo-Nirenberg 不等式和 Young 不等式可知, 存在常数 $C_{10},C_{11},C_{12},C_{13}>0$, 使得

$\begin{matrix} \Vert \hat {n}\Vert_{L^{r}(\Omega)}^{r}=\Vert\hat{n}^{\frac{r}{2}}\Vert_{L^{2}(\Omega)}^{2}&\leq C_{10}\Big(\Vert \nabla\hat{n}^{\frac{r}{2}}\Vert_{L^{2}(\Omega)}^{\frac{2(3r-6)}{3r-2}}\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{\frac{4}{r}}(\Omega)}^{\frac{8}{3r-2}}+\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{\frac{4}{r}}(\Omega)}^{2}\Big) \\ &\leq\frac{r-1}{C_{9}r}\Vert \nabla\hat{n}^{\frac{r}{2}}\Vert_{L^{2}(\Omega)}^{2}+C_{11}\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{\frac{4}{r}}(\Omega)}^{2} \end{matrix}$

$\begin{matrix} \Vert \hat {n}\Vert_{L^{r}(\Omega)}^{2}=\Vert\hat{n}^{\frac{r}{2}}\Vert_{L^{2}(\Omega)}^{\frac{4}{r}}&\leq C_{12}\Big(\Vert \nabla\hat{n}^{\frac{r}{2}}\Vert_{L^{2}(\Omega)}^{\frac{3r-6}{3r-2}}\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{\frac{4}{r}}(\Omega)}^{\frac{4}{3r-2}}+\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{\frac{4}{r}}(\Omega)}\Big)^{\frac{4}{r}} \\ &\leq\frac{r-1}{2C_{9}r}\Vert \nabla\hat{n}^{\frac{r}{2}}\Vert_{L^{2}(\Omega)}^{2}+C_{13}\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{\frac{4}{r}}(\Omega)}^{\frac{16}{3r^{2}-8r+12}}+C_{13}\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{\frac{4}{r}}(\Omega)}^{\frac{4}{r}}. \end{matrix}$

将 (4.47) 和 (4.48) 式代入 (4.46) 式并结合引理 4.5 可得, 存在常数 $C_{14},C_{15}>0$, 使得

$\begin{matrix} & \ \frac{\mathrm{d}}{\mathrm{d}t}\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{2}(\Omega)}^{2}+\frac{1}{2}\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{2}(\Omega)}^{2} \\ &\leq\frac{\mathrm{d}}{\mathrm{d}t}\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{2}(\Omega)}^{2}+\frac{r-1}{2r}\Vert \nabla\hat{n}^{\frac{r}{2}}\Vert_{L^{2}(\Omega)}^{2}+\frac {C_{11}}{2}\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{\frac{4}{r}}(\Omega)}^{2} \\ &\leq \Big(C_{9}C_{11}{\rm e}^{-\mu t}+C_{9}C_{11}+\frac {C_{11}}{2}\Big)\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{\frac{4}{r}}(\Omega)}^{2}+C_{9}{\rm e}^{-\mu t}\Vert \hat {u}\Vert_{L^{r}(\Omega)}^{r}+C_{9}\Big({\rm e}^{-\mu t}+1\Big)\Vert \hat {c}\Vert_{W^{1,r}(\Omega)}^{r} \\ &\ \ \ +C_{9}C_{13}{\rm e}^{-\mu t}\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{\frac{4}{r}}(\Omega)}^{\frac{16}{3r^{2}-8r+12}}+C_{9}C_{13}{\rm e}^{-\mu t}\Vert \hat{n}^{\frac{r}{2}}\Vert_{L^{\frac{4}{r}}(\Omega)}^{\frac{4}{r}}+C_{9}{\rm e}^{-\mu t}\Vert \hat {c}\Vert_{L^{r}(\Omega)}^{2} \\ &\leq C_{14}(1+t)^{\frac{r}{2}}\epsilon^{\frac{r}{2}}+C_{14}(1+t)^{\frac{4r}{3r^{2}-8r+12}}\epsilon^{\frac{4r}{3r^{2}-8r+12}}+C_{14}(1+t)\epsilon \\ & \ \ \ +C_{9}\Big({\rm e}^{-\mu t}+1\Big)\Vert \hat {c}\Vert_{W^{1,r}(\Omega)}^{r}+C_{9}{\rm e}^{-\mu t}\Vert \hat {c}\Vert_{L^{r}(\Omega)}^{2} \\ &\leq 3C_{14}(1+t)^{\frac{r}{2}}\epsilon^{\frac{4r}{3r^{2}-8r+12}}+C_{9}\Big({\rm e}^{-\mu t}+1\Big)\Vert \hat {c}\Vert_{W^{1,r}(\Omega)}^{r}+C_{9}{\rm e}^{-\mu t}\Vert \hat {c}\Vert_{L^{r}(\Omega)}^{2}, \end{matrix}$

其中利用了

$\Vert \hat {u}\Vert_{L^{r}(\Omega)}^{r}\leq C_{15}\Big(\Vert \nabla\hat{u}\Vert_{L^{2}(\Omega)}^{\frac{3r-6}{2}}\Vert \hat{u}\Vert_{L^{2}(\Omega)}^{\frac{6-r}{2}}+\Vert \hat{u}\Vert_{L^{2}(\Omega)}^{r}\Big).$

$h(t):=\Vert\hat{n}(\cdot,t)\Vert_{L^{r}(\Omega)}^{r}=\Vert \hat{n}^{\frac{r}{2}}(\cdot,t)\Vert_{L^{2}(\Omega)}^{2}$, 则存在常数 $C_{16}>0$, 使得

$h^{\prime}(t)+\frac{1}{2}h(t)\leq C_{16}(1+t)^{\frac{r}{2}}\epsilon^{\frac{4r}{3r^{2}-8r+12}}+C_{16}\Vert \hat {c}\Vert_{W^{1,r}(\Omega)}^{r}+C_{16}\Vert \hat {c}\Vert_{L^{r}(\Omega)}^{2}.$

直接计算得

$\begin{matrix} h(t)&\leq C_{16}\epsilon^{\frac{4r}{3r^{2}-8r+12}}\int_{0}^{t}(1+s)^{\frac{r}{2}}{\rm e}^{-\frac{(t-s)}{2}}{\mathrm{d}s}+C_{16}\int_{0}^{t}{\rm e}^{-\frac{(t-s)}{2}}\Vert \hat {c}(\cdot,s)\Vert_{W^{1,r}(\Omega)}^{r}{\mathrm{d}s} \\ &\ \ \ +C_{16}\int_{0}^{t}{\rm e}^{-\frac{(t-s)}{2}}\Vert \hat {c}(\cdot,s)\Vert_{L^{r}(\Omega)}^{2}{\mathrm{d}s} \\ &=:M_{1}+M_{2}+M_{3}. \end{matrix}$

下面分别估计 $M_{1}$, $M_{2}$$M_{3}$. 对于 $M_{1}$,

$ M_{1}\leq C_{16}\epsilon^{\frac{4r}{3r^{2}-8r+12}}(1+t)^{\frac{r}{2}}\int_{0}^{t}{\rm e}^{-\frac{(t-s)}{2}}{\mathrm{d}s}\leq C_{16}\epsilon^{\frac{4r}{3r^{2}-8r+12}}(1+t)^{\frac{r}{2}}.$

对于 $M_{2}$, 设 $p>r+2>4$, 利用插值不等式、引理 3.8 、引理 4.5、引理 3.9 和引理 3.4 可知: 存在常数 $C_{17},C_{18},C_{19},C_{20}>0$, 使得

$\begin{matrix} M_{2}&\leq C_{17}\int_{0}^{t}{\rm e}^{-\frac{(t-s)}{2}}\Vert \hat {c}(\cdot,s)\Vert_{W^{1,2}(\Omega)}^{\frac{2(p-r)}{p-2}}\Vert \hat {c}(\cdot,s)\Vert_{W^{1,p}(\Omega)}^{\frac{p(r-2)}{p-2}}{\mathrm{d}s} \\ &\leq C_{18}\int_{0}^{t}{\rm e}^{-\frac{(t-s)}{2}}\Vert \hat {c}(\cdot,s)\Vert_{W^{1,2}(\Omega)}^{\frac{2(p-r)}{p-2}}{\mathrm{d}s} \\ &\leq C_{19}\Big(\int_{0}^{t} \Vert\hat{c}(\cdot,s)\Vert_{W^{1,2}(\Omega)}^{2}{\mathrm{d}s}\Big)^{\frac{p-r}{p-2}}\Big(\int_{0}^{t}{\rm e}^{-\frac{p-2}{r-2}\frac{(t-s)}{2}}{\mathrm{d}s}\Big)^{\frac{r-2}{p-2}} \\ &\leq C_{20}(1+t)^{\frac{p-r}{p-2}}\epsilon^{\frac{p-r}{p-2}}. \end{matrix}$

类似地, 对于 $M_{3}$, 利用插值不等式、Hölder 不等式、引理 4.5 和引理 3.9 可知: 存在常数 $C_{21}, C_{22}$, $C_{23}>0$, 使得

$\begin{matrix} M_{3}&\leq C_{21}\int_{0}^{t}{\rm e}^{-\frac{(t-s)}{2}}\Vert \hat {c}(\cdot,s)\Vert_{L^{2}(\Omega)}^{\frac{4}{r}}\Vert \hat {c}(\cdot,s)\Vert_{L^{\infty}(\Omega)}^{\frac{2(r-2)}{r}}{\mathrm{d}s} \\ &\leq C_{22}\Big(\int_{0}^{t}\Vert \hat {c}(\cdot,s)\Vert_{L^{2}(\Omega)}^{2}{\mathrm{d}s}\Big)^{\frac{2}{r}}\Big(\int_{0}^{t}{\rm e}^{-\frac{r}{r-2}\frac{(t-s)}{2}}{\mathrm{d}s}\Big)^{\frac{r-2}{r}} \\ &\leq C_{23}(1+t)^{\frac{2}{r}}\epsilon^{\frac{2}{r}}. \end{matrix}$

综上可得, 存在常数 $C_{24},C_{25}>0$, 使得

$\begin{align*} h(t)&\leq C_{24}(1+t)^{\frac{r}{2}}\epsilon^{\frac{4r}{3r^{2}-8r+12}}+C_{24}(1+t)^{\frac{p-r}{p-2}}\epsilon^{\frac{p-r}{p-2}}+C_{24}(1+t)^{\frac{2}{r}}\epsilon^{\frac{2}{r}}\\ &\leq C_{25}(1+t)^{\frac{r}{2}}\epsilon^{\frac{2}{r}},\ 2<r\leq 6 \end{align*}$

$\begin{align*}h(t)&\leq C_{24}(1+t)^{\frac{r}{2}}\epsilon^{\frac{4r}{3r^{2}-8r+12}}+C_{24}(1+t)^{\frac{p-r}{p-2}}\epsilon^{\frac{p-r}{p-2}}+C_{24}(1+t)^{\frac{2}{r}}\epsilon^{\frac{2}{r}}\\ &\leq C_{25}(1+t)^{\frac{r}{2}}\epsilon^{\frac{4r}{3r^{2}-8r+12}}, \ r>6. \end{align*}$

证毕.

定理 2.1 的证明 结合推论 4.1、引理 4.3、引理 4.5、引理 4.6 和引理 4.7 可得.

参考文献

Biler P.

Local and global solvability to some parabolic-elliptic systems of chemotaxis

Adv Math Sci Appl, 1998, 8: 715-743

[本文引用: 1]

Biler P, Brandolese L.

On the parabolic-elliptic limit of the doubly parabolic Keller-Segel system modelling chemotaxis

Stud Math, 2009, 193: 241-261

[本文引用: 1]

Cao X, Lankeit J.

Global classical small-data solutions for a three-dimensional chemo-taxis Navier-Stokes system involving matrix-valued sensitivities

Calc Var Partial Dif, 2016, 55: Article 107

[本文引用: 2]

Duan R, Li X, Xiang Z.

Global existence and large time behavior for a two-dimensional chemotaxis-Navier-Stokes system

J Diff Equ, 2017, 263: 6284-6316

[本文引用: 1]

Duan R, Lorz A, Markowich P A.

Global solutions to the coupled chemotaxis-fluid equations

Comm Part Diff Eqs, 2010, 35: 1635-1673

[本文引用: 1]

Freitag M.

The fast signal diffusion limit in nonlinear chemotaxis systems

Disc Cont Dyn Syt-B, 2020, 25: 1109-1128

[本文引用: 1]

Herrero M A, Velazquez J J L.

A blow up mechanism for a chemotaxis system

Ann Scuola Norm Sup Pisa Cl Sci, 1997, 24: 633-683

[本文引用: 1]

Hillen T, Painter K J.

A user's guide to PDE models for chemotaxis

J Math Biol, 2009, 58(1): 183-217

[本文引用: 1]

Horstmann D.

From 1970 until present: The Keller-Segel model in chemotaxis and its consequences I.

Jahresber Dtsch Math Ver, 2003, 105: 103-165

[本文引用: 1]

Horstmann D, Wang G.

Blow-up in a chemotaxis model without symmetry assumptions

European J Appl Math, 2001, 12: 159-177

[本文引用: 1]

Horstmann D, Winkler M. Boundedness vs.

blow-up in a chemotaxis system

J Diff Equ, 2005, 215(1): 52-107

[本文引用: 1]

Jager W, Luckhaus S.

On explosions of solutions to a system of partial differential equations modelling chemotaxis

Trans Amer Math Soc, 1992, 329: 819-824

[本文引用: 1]

Ke Y, Zheng J.

An optimal result for global existence in a three-dimensional Article Keller-Segel-Navier-Stokes system involving tensor-valued sensitivity with saturation

Calc Var Partial Dif, 2019, 58: Article 109

[本文引用: 1]

Keller E F, Segel L A.

Initiation of slime mold aggregation viewed as an instability

J Theor Biol, 1970, 26: 399-415

PMID:5462335      [本文引用: 1]

Kurokiba M, Ogawa T.

Singular limit problem for the Keller-Segel system and drift-diffusion system in scaling critical spaces

J Evol Equ, 2020, 20: 421-457

[本文引用: 1]

Lemarie Rieusset P G.

Small data in an optimal Banach space for the parabolic-parabolic and parabolic-elliptic Keller-Segel equations in the whole space

Adv Differential Equations, 2013, 18: 1189-1208

[本文引用: 1]

Li M, Xiang Z.

The convergence rate of the fast signal diffusion limit for a Keller-Segel-Stokes system with large initial data

Proc R Soc Edinburgh Sect A Math, 2021, 151: 1972-2012

[本文引用: 9]

Li M, Xiang Z, Zhou G.

The stability analysis of a 2D Keller-Segel-Navier-Stokes system in fast signal diffusion

Euro J of Applied Mathematics, 2023, 34: 160-209

[本文引用: 2]

Li X, Wang Y, Xiang Z.

Global existence and boundedness in a 2D Keller-Segel-Stokes system with nonlinear diffusion and rotational flux

Commun Math Sci, 2016, 14: 1889-1910

[本文引用: 1]

Mizoguchi N, Souplet P.

Nondegeneracy of blow-up points for the parabolic Keller-Segel system

Ann I H Poincaré AN, 2014, 31: 851-875

[本文引用: 1]

Mizukami M.

The fast signal diffusion limit in a Keller-Segel system

J Math Anal Appl, 2019, 472: 1313-1330

DOI:10.1016/j.jmaa.2018.11.077      [本文引用: 1]

This paper deals with convergence of a solution for the parabolic-parabolic Keller-Segel system {(u(lambda))(t) = Delta u(lambda) - chi del. (u(lambda) del v(lambda)) in Omega x (0, infinity), lambda(v(lambda))(t) = Delta v(lambda) - v(lambda) + u(lambda) in Omega x (0, infinity), where Omega is a bounded domain in R-n (n >= 2) with smooth boundary, chi, lambda > 0 are constants, to that for the parabolic-elliptic-Keller-Segel system (u(t) = Delta u - chi del. (u del v) in Omega x (0, infinity), 0 = Delta v - v + u in Omega x (0, infinity) as lambda SE arrow 0. (C) 2018 Elsevier Inc. All rights reserved.

Nagai T.

Global existence of solutions to a parabolic system for chemotaxis in two space dimensions

Nonlinear Analysis Theory Methods Applications, 1997, 30: 5381-5388

[本文引用: 1]

Nagai T.

Blowup of nonradial solutions to parabolic-elliptic systems modeling chemotaxis in two-dimensional domains

J Inequal Appl, 2001, 6: 37-55

[本文引用: 1]

Nagai T, Senba T.

Behavior of radially symmetric solutions of a system related to chemo-taxis

Rims Kokyuroku, 1996, 973: 32-39

[本文引用: 1]

Nagai T, Senba T, Yoshida K.

Application of the Trudinger-Moser inequality to a parabolic system of chemotaxis

Funkcial Ekvac Ser Int, 1997, 40: 411-433

[本文引用: 1]

Osaki K, Yagi A.

Finite dimensional attractor for one-dimensional Keller-Segel equations

Funkcial Ekvac, 2001, 44: 441-469

[本文引用: 1]

Peng Y, Xiang Z.

Global existence and boundedness in a 3D Keller-Segel-Stokes system with nonlinear diffusion and rotational flux

Z Angew Math Phys, 2017, 68: Article 68

[本文引用: 1]

Raczynski A.

Stability property of the two-dimensional Keller-Segel model

Asympt Anal, 2009, 61: 35-59

[本文引用: 1]

Tuval I, Cisneros L, Dombrowski C, et al.

Bacterial swimming and oxygen transport near contact lines

Proc Nat Acad Sci USA, 2005, 102: 2277-2282

[本文引用: 1]

Wang Y.

Global weak solutions in a three-dimensional Keller-Segel-Navier-Stokes system with subcritical sensitivity

Math Models Methods Appl Sci, 2017, 27: 2745-2780

[本文引用: 1]

Wang Y, Winkler M, Xiang Z.

A smallness condition ensuring boundedness in a two-dimensional Chemotaxis-Navier-Stokes system involving Dirichlet boundary conditions for the signal

Acta Mathematica Sinica, 2022, 38: 985-1001

[本文引用: 1]

Wang Y, Winkler M, Xiang Z.

Global classical solutions in a two-dimensional chemotaxis-Navier-Stokes system with subcritical sensitivity

Ann Sc Norm Super Pisa Cl Sci, 2018, 18: 421-466

[本文引用: 2]

Wang Y, Winkler M, Xiang Z.

The small-convection limit in a two-dimensional chemotaxis-Navier-Stokes system

Math Z, 2018, 289: 71-108

[本文引用: 1]

Wang Y, Winkler M, Xiang Z.

The fast signal diffusion limit in Keller-Segel(-fluid) systems

Calc Var Partial Dif, 2019, 58: Article 196

[本文引用: 1]

Wang Y, Xiang Z.

Global existence and boundedness in a Keller-Segel-Stokes system involving a tensor-valued sensitivity with saturation

J Diff Equ, 2015, 259: 7578-7609

[本文引用: 2]

Wang Y, Xiang Z.

Global existence and boundedness in a Keller-Segel-Stokes system involving a tensor-valued sensitivity with saturation: the 3D case

J Diff Equ, 2016, 261: 4944-4973

[本文引用: 2]

Winkler M.

Finite-time blow-up in the higher-dimensional parabolic-parabolic Keller-Segel system

J Math Pures Appl, 2013, 100: 748-767

[本文引用: 1]

Winkler M.

Global large-data solutions in a chemotaxis-(Navier-)Stokes system modeling cellular swimming in fluid drops

Comm Part Diff Eqs, 2012, 37: 319-351

[本文引用: 1]

Winkler M.

Global weak solutions in a three-dimensional chemotaxis-Navier-Stokes system

Ann I H Poincaré AN, 2016, 33: 1329-1352

[本文引用: 1]

Winkler M.

Stabilization in a two-dimensional chemotaxis-Navier-Stokes system

Arch Rational Mech Anal, 2014, 211: 455-487

[本文引用: 1]

Winkler M. Aggregation vs.

global diffusive behavior in the higher-demensional Keller-Segel model

J Diff Equ, 2010, 248: 2889-2905

[本文引用: 1]

Wu C, Xiang Z.

Saturation of the signal on the boundary: Global weak solvability in a chemotaxis-Stokes system with porous-media type cell diffusion

J Diff Equ, 2022, 315: 122-158

[本文引用: 1]

Yu H, Wang W, Zheng S.

Global classical solutions to the Keller-Segel-Navier-Stokes system with matrix-valued sensitivity

J Math Anal Appl, 2018, 461: 1748-1770

[本文引用: 1]

Zhang Q, Li Y.

Global weak solutions for the three-dimensional chemotaxis-Navier-Stokes system with nonlinear diffusion

J Diff Equ, 2015, 259: 3730-3754

[本文引用: 1]

Zheng J.

Global existence and boundedness in a three-dimensional chemotaxis-Stokes system with nonlinear diffusion and general sensitivity

Annali di Matematica Pura ed Applicata, 2022, 201: 243-288

[本文引用: 1]

/