数学物理学报  2017, Vol. 37 Issue (6): 1129-1147   PDF    
扩展功能
加入收藏夹
复制引文信息
加入引用管理器
Email Alert
RSS
本文作者相关文章
郭庭光
徐志庭
带有扩散项和接种的传染病模型的行波解
郭庭光, 徐志庭     
华南师范大学数学科学学院 广州 510631
摘要:该文研究带有扩散项和接种的传染病模型的行波解存在性.首先建立一个带扩散项和接种的具有空间结构的传染病模型,并给出其解适定性.其次,构造一对向量型上、下解,应用Schauder不动点原理和Lyapunov函数方法得到此模型存在连接无病平衡点和有病平衡点的非平凡正行波解.利用稳定流形定理,得到行波指数衰减估计,进而,通过拉普拉斯变换,确定该模型行波解的不存在性.该文的研究技巧对建立高维非合作反应扩散系统行波解存在性提供了有效方法.
关键词行波解    传染病模型    Schauder不动点原理    向量型上、下解    拉普拉斯变换    
Existence of Traveling Waves in a Spatial Infectious Disease Model with Vaccination
Guo Tingguang, Xu Zhiting     
School of Mathematical Sciences, South China Normal University, Guangzhou 510631
Abstract: The current paper is devoted to investigate the existence of traveling waves in a spatial infectious disease model with vaccination. First, we propose a spatial infectious disease model with vaccination, and then study the well-posedness of it. Second, based on constructing a pair of the vector-value upper and lower solutions and the applications of Schauder's fixed point theorem, we show that the model admits nontrivial and positive traveling waves connecting the disease free equilibrium and the endemic equilibrium. Third, by Laplace transforms, we establish the non-existence of traveling waves for the model, in which the prior estimate of the exponential decay of the traveling wave solutions is obtained by the Stable Manifold Theorem. The approach in this paper provides an effective method to deal with more general high dimensional non-cooperative reaction-diffusion systems.
Key words: Traveling waves     Infectious disease model     Vector-value upper and lower solutions     Schauder's fixed point theorem     Laplace transform    
1 引言

众所周知, 传染病的传播很大程度上取决于易感染者的自身属性, 所处空间位置, 传播方式以及引起感染的病原体.有些群体(个体)相对于其他群体(个体)具有更高易感和传染疾病的风险度.因此, 传染病的感染更大程度上依赖于易感染者自身的风险度.众所周知, 通过注射疫苗控制传染病的传播是一种被广泛使用的重要策略之一.基于这一事实, 最近Rahman和Zou[1]提出一个结合疫苗接种的两种群传染病模型以研究疫苗接种策略的数学模型.

下面我们简要陈述文献[1] (关于模型(1.1)的精确的生物学背景和解释, 可见文献[1])所建立的数学模型.在此模型中, 总人口被分为两组:高风险组( $r$ )和临界组( $c$ ).高风险组是指在这组的个体有更高的感染率.而临界组是指在这组的个体在其社会活动行为中表现出更有意识或更少接触感染个体, 这使得其自身具有更低的感染率.因此, 根据他们的风险度, 这两组在总人口中被划分为两个不同的组.而每个组又都被分为易感个体( $S$ )和感染个体( $I$ )两个子类.假设每个易感个体若受感染, 则其一生只能留在感染个体的子类里.每组的易感个体以一定的接种率接种疫苗并转化为一个接种疫苗的子类( $V$ ).为了更好地描述个体行为的变化和感染的机理, Rahman和Zou[1]选择形如 $h(I)=\frac{I}{1+\alpha I}$ ( $\alpha>0$ )的饱和发生率[4].由此, Rahman和Zou[1]建立如下带接种疫苗的数学模型

$ \begin{equation}\label{1.1} \left\{\begin{array}{l} \dot S^r=\Lambda^r-\Big(\frac{\beta_r^rI^r}{1+\alpha_rI^r} +\frac{\beta_c^rI^c}{1+\alpha_cI^c}\Big)S^r-(\mu+\theta^r)S^r, \\[3mm] \dot S^c=\Lambda^c-\Big(\frac{\beta_r^cI^r}{1+\alpha_rI^r} +\frac{\beta_c^cI^c}{1+\alpha_cI^c}\Big)S^c-(\mu+\theta^c)S^c, \\[3mm] \dot I^r=\Big(\frac{\beta_r^rI^r}{1+\alpha_rI^r} +\frac{\beta_c^rI^c}{1+\alpha_cI^c}\Big)S^r-(\mu+v^r)I^r, \\[3mm] \dot I^c=\Big(\frac{\beta_r^cI^r}{1+\alpha_rI^r} +\frac{\beta_c^c I^c}{1+\alpha_cI^c}\Big)S^c-(\mu+v^c)I^c, \\[2mm] \dot V =\theta^rS^r+\theta^cS^c-\mu V, \end{array}\right. \end{equation} $ (1.1)

其中 $S^r(t)$ , $S^c(t)$ , $I^r(t)$ , $I^c(t)$ 以及 $V(t)$ 分别表示在时刻 $t$ 时, $r$ 组中易感个体的密度, $c$ 组中易感个体的密度, $r$ 组中感染个体的密度, $c$ 组中感染个体的密度以及已经接种疫苗的个体密度. $\Lambda^r$ $\Lambda^c$ 是组 $r$ 和组 $c$ 的易感个体的招募率. $\beta_i^j$ $(i, j=r, c)$ 是易感个体和感染个体之间的接触率. $v^i$ $(i=r, c)$ 是由疾病引起的死亡率, $\mu$ 表示自然死亡率. $\theta^r$ $\theta^c$ 分别表示组 $r$ $c$ 的疫苗接种率. $\alpha_i$ $(i=r, c)$ 表示当感染人口数很大时的饱和度.

由于系统(1.1)中最后一个方程独立于其他四个方程, Rahman和Zou[1]仅分析模型(1.1)中前四个方程的全局动力学性质.他们证明, 当基本再生数小于 $1$ 时, 这种疾病会随着时间消失殆尽, 当基本再生数大于 $1$ 时, 疾病将持续流行而感染个体数会接近一个正常数.

事实上, 普遍认为空间结构对刻画传染病的传播和控制具有重要的作用[15].因此, 在设计预防和控制传染病传播的策略时, 应充分考虑某些空间(诸如移民、个体运动、边境的管制以及隔离等)因素的影响.正是基于这一认知, 在本文, 对应模型(1.1), 我们考虑如下具有空间结构的带扩散和接种的传染病数学模型

$ \begin{equation}\label{1.2} \left\{\begin{array}{l} \frac{\partial S^r}{\partial t} =d_1\frac{\partial ^2S^r}{\partial x^2}+\Lambda^r-\Big(\frac{\beta_r^rI^r}{1+\alpha_rI^r} +\frac{\beta_c^rI^c}{1+\alpha_cI^c}\Big)S^r-(\mu+\theta^r)S^r, \\[3mm] \frac{\partial S^c}{\partial t} =d_2\frac{\partial ^2S^c}{\partial x^2}+\Lambda^c-\Big(\frac{\beta_r^cI^r}{1+\alpha_rI^r} +\frac{\beta_c^cI^c}{1+\alpha_cI^c}\Big)S^c-(\mu+\theta^c)S^c, \\[3mm] \frac{\partial I^r}{\partial t} =d_3\frac{\partial ^2I^r}{\partial x^2}+\Big(\frac{\beta_r^rI^r}{1+\alpha_rI^r} +\frac{\beta_c^rI^c}{1+\alpha_cI^c}\Big)S^r-(\mu+v^r)I^r, \\[3mm] \frac{\partial I^c}{\partial t} =d_4\frac{\partial ^2I^c}{\partial x^2}+\Big(\frac{\beta_r^cI^r}{1+\alpha_rI^r} +\frac{\beta_c^cI^c}{1+\alpha_cI^c}\Big)S^c-(\mu+v^c)I^c, \end{array}\right. \end{equation} $ (1.2)

其中 $S^r(t, x)$ , $S^c(t, x)$ , $I^r(t, x)$ $I^c(t, x)$ 分别表示在时刻 $t$ 和位置 $x\in{\Bbb R}$ 时, $r$ 组中易感个体的密度, $c$ 组中易感个体的密度, $r$ 组中感染个体的密度, $c$ 组中感染个体的密度以及已接种疫苗的个体密度. $d_1$ , $d_2$ , $d_3$ , $d_4$ $>0$ 是其对应的扩散率.参数 $\beta_i^j$ , $\Lambda^i$ , $\mu$ , $\theta^i$ , $v^i$ , $\alpha_i$ $(i, j=r, c)$ 与模型(1.1)具有相同的生物学含义.

具空间结构的传染病模型的行波解刻画了传染病从最初的无病平衡点到最终有病平衡点的演变过程, 同时还描述传染病在空间以常数速度传播并且保持形状不变[15].行波解研究有助于人们认识传染病在空间传播的模式, 进而为人们评估传染病对人类自身健康的影响和危害提供理论依据.

由于系统(1.2)不满足比较原理和单调原理, 因此, 不能直接应用由文献[10, 12-13, 20, 25]所建立的关于单调系统行波解理论.本文受工作[2-3, 6, 8, 11, 18-19, 21, 26, 28]的诱发, 构造一对向量值型的上、下解, 并由其定义一个不变集.进而, 在此变集上引入系统(1.2)对应的波方程(3.1)的解映射, 应用Schauder不动点定理建立系统(3.1)的非负解的存在性.应当指出, 由于系统(3.1)有四个方程组成, 不同于文献[2-3, 6, 8, 11, 18-19, 21, 26, 28], 我们的方法的一个重要特征就是需要构造关于系统(3.1)的向量值型的上、下解.而这种思想是受文献[22]的启发, 或见文献[7, 17].除此之外, 我们还运用双边拉普拉斯变换研究系统(2.1)的行波解的不存在性.应用拉普拉斯变换, 其关键之处需要对行波作指数衰减的先验估计[2, 5, 18-19, 21, 26].再注意到(3.1)模型是四维的系统, 文献[2, 5, 18-19, 21, 26]的研究方法并不能给出系统(3.1)解的的先验估计.因此, 我们使用专著[16]的稳定流形定理(或见文献[28])得到系统(3.1)解的的先验估计.

依所构造的上、下解, 不难得到系统(3.1)的解 $u(t)$ ${\Bbb R}$ 上非负且满足 $\lim\limits_{t\to-\infty}u(t)=E^0$ .但就此并不能确定系统(3.1)的行波是否连接有病平衡点 $E^*$ .为证明系统(3.1)行波连接两个平衡点 $E^0$ $E^*$ , 我们还需要证明 $\lim\limits_{t\to+\infty}u(t)=E^*$ .为此, 受文献[1, 6, 8, 11]工作的启发, 我们构造Lyapunov函数并证得 $\lim\limits_{t\to+\infty}u(t)=E^*$ .值得指出的是, 由于系统(3.1)是由四个二阶微分方程组成, Lyapunov函数的构造是非平凡的.本文的研究技巧对建立高维非合作(或不满足比较原理)反应扩散系统提供了一个有效可行的方法.

本文由四节组成.第二节研究系统(2.1)的适定性.第3节给出一些辅助结果.第4节建立系统(3.1)的行波解的存在性与不存在性.

2 适定性

这一节研究系统(2.1)解的适定性.为符号简单, 令 $u_1(x, t)=S^r(x, t), \ u_2(x, t)=S^c(x, t), $ $ u_3(x, t)=I^r(x, t), \ u_4(x, t)=I^c(x, t)$ , $\beta_{11}=\beta_r^r, \ \beta_{12}=\beta _c^r, \ \beta_{21}=\beta_r^c, \ \beta_{22}=\beta_c^c, \ \Lambda_1=\Lambda^r, $ $ \Lambda_2=\Lambda^c, $ $\theta_1=\mu+\theta^r, \ \theta_2=\mu+\theta^c, \ \theta_3=\mu+\nu^r, \ \theta_4=\mu+\nu^c$ , 则系统(1.2)变形为

$ \begin{equation}\label{2.1} \left\{\begin{array}{l} \frac{\partial u_1}{\partial t}=d_1\frac{\partial^2u_1}{\partial x^2} +\Lambda_1-[\beta_{11}f_1(u_3)+\beta_{12}f_2(u_4)]u_1-\theta_1u_1, \\[3mm] \frac{\partial u_2}{\partial t}=d_2\frac{\partial^2u_2}{\partial x^2} +\Lambda_2-[\beta_{21}f_1(u_3)+\beta_{22}f_2(u_4)]u_2-\theta_2u_2, \\[3mm] \frac{\partial u_3}{\partial t}=d_3\frac{\partial^2u_3}{\partial x^2} + [\beta_{11}f_1(u_3)+\beta_{12}f_2(u_4)]u_1-\theta_3u_3, \\[3mm] \frac{\partial u_4}{\partial t}=d_4\frac{\partial^2u_4}{\partial x^2} + [\beta_{21}f_1(u_3)+\beta_{22}f_2(u_4)]u_2-\theta_4u_4, \end{array}\right. \end{equation} $ (2.1)

其中 $(t, x)\in(0, \infty)\times\mathbb{R}$ , $f_i(x)=\frac{x}{1+\alpha_ix}$ , $i=1, 2$ .对系统(2.1), 考虑如下初值条件

$ \begin{equation}\label{2.2} u_i(0, x)=\varphi_i(x)\geq0, \ \forall x\in\mathbb{R}, \ i=1, 2, 3, 4, \ \mbox{但不恒为0. } \end{equation} $ (2.2)

$u_1^0:=\frac{\Lambda_1}{\theta_1}$ , $u_2^0:=\frac{\Lambda_2}{\theta_2}$ .依文献[1], 定义(2.1)的基本再生数 ${\bf R}_0$

$ {{\bf R}_0}:=\frac{1}{2}\bigg(\frac{\beta_{11}}{\theta_3}u_1^0 +\frac{\beta_{22}}{\theta_4}u_2^0 +\sqrt{\Big(\frac{\beta_{11}}{\theta_3}u_1^0 -\frac{\beta_{22}}{\theta_4}u_2^0\Big)^2 +\frac{4\beta_{12}\beta_{21}}{\theta_3\theta_4}u_1^0u_2^0}\ \bigg). $

由于系统(2.1)和系统(1.1)具有同样的常数平衡点, 由文献[1, 定理3.4], 得到下列引理2.1.

引理2.1 (1) 若 ${{\bf R}}_0<1$ , 则系统(2.1)有且仅有一个无病平衡点 $E^0=(u_1^0, u_2^0, 0, 0)$ .

(2) 若 ${{\bf R}}_0>1$ , 则系统(2.1)存在有病平衡点 $E^*=(u_1^*, u_2^*, u_3^*, u_4^*)$ .

注意到 $\Lambda_1=[\beta_{11}f_1(u^*_3)+\beta_{12}f_2(u^*_4)]u^*_1+\theta_1u^*_1 >\theta_1u^*_1$

$ \theta_3u^*_3=[\beta_{11}f_1(u^*_3)+\beta_{12}f_2(u^*_4)]u^*_1 <\Big(\frac{\beta_{11}}{\alpha_1}+\frac{\beta_{12}}{\alpha_2}\Big)u^*_1. $

易知

$ 0<u_1^*<u_1^0, \quad 0<u_3^*<\frac{u_1^0}{\theta_3} \Big(\frac{\beta_{11}}{\alpha_1}+\frac{\beta_{12}}{\alpha_2}\Big). $

类似地, 可到

$ 0<u_2^*<u_2^0, \quad 0<u_4^*<\frac{u_2^0}{\theta_4} \Big(\frac{\beta_{21}}{\alpha_1}+\frac{\beta_{22}}{\alpha_2}\Big). $

下面, 讨论系统(2.1)-(2.2)的初值问题的解的正不变性和一致有界性.

在本文, 锥体 ${\Bbb R}_+^4$ 表示 ${\Bbb R}^4$ 中向量 $x\geq 0$ 的子集.设 ${{\bf X}}:=C({\Bbb R}, {\Bbb R}^4)$ 表示以 $\|\cdot\|_{{\bf X}}$ 为上确界范数的巴拿赫空间.令 ${{\bf X}^+}=C(\mathbb{R}, {\Bbb R}_+^4)$ , 于是 $({\bf X}, {\bf X}^+)$ 是一个强有序空间.此外, 我们在 ${\Bbb R}^4$ ${\bf X}$ 中均使用标准的偏序.令 ${\bf X}_{{\bf M}}=\big\{\varphi=(\varphi_1, \varphi_2, \varphi_3, \varphi_4) \in{{\bf X}}^+:{{\bf 0}}\leq\varphi(x)\leq{{\bf M}}, \ \forall x\in{{\Bbb R}}\big\}$ .其中 ${{\bf 0}}:=(0, 0, 0, 0)$ , 且

$ {{\bf M}}:=\Big(u_1^0, u_2^0, \frac{u_1^0}{\theta_3} \Big(\frac{\beta_{11}}{\alpha_1}+\frac{\beta_{12}}{\alpha_2}\Big), \frac{u_2^0}{\theta_4}\Big(\frac{\beta_{21}} {\alpha_1}+\frac{\beta_{22}}{\alpha_2}\Big)\Big). $

对任意的 $\varphi=(\varphi_1, \varphi_2, \varphi_3, \varphi_4)\in{{\bf X}}_{{\bf M}}$ $x\in{\Bbb R}$ , 定义 $F=(F_1, F_2, F_3, F_4):{\bf X}_{{\bf M}}\to {\bf X}$ ,

$ \begin{eqnarray*} F_1(\varphi)(x):&=&\Lambda_1-[\beta_{11}f_1(\varphi_3)(x) +\beta_{12}f_2(\varphi_4)(x)]\varphi_1(x)-\theta_1\varphi_1(x), \\ F_2(\varphi)(x):&=&\Lambda_2-[\beta_{21}f_1(\varphi_3)(x) +\beta_{22}f_2(\varphi_4)(x)]\varphi_2(x)-\theta_2\varphi_2(x), \\ F_3(\varphi)(x):&=&[\beta_{11}f_1(\varphi_3)(x)+\beta_{12}f_2(\varphi_4)(x)] \varphi_1(x)-\theta_3\varphi_3(x), \\ F_4(\varphi)(x):&=&[\beta_{21}f_1(\varphi_3)(x)+\beta_{22}f_2(\varphi_4)(x)] \varphi_2(x)-\theta_4\varphi_4(x). \end{eqnarray*} $

$F$ ${{\bf X}}_{{\bf M}}$ 的任意有界子集上局部Lipschitz连续, 且系统(2.1)和(2.2)可以写成如下的抽象微分方程

$ \left\{\begin{array}{l} \frac{{\rm d}{{\bf u}}}{{\rm d}t}= A{{\bf u}}+ F({{\bf u}}), \ t >0, \\[2mm] {{\bf u}}_0=\varphi\in{{\bf X}}_{{\bf M}}, \end{array}\right. $

其中 ${{\bf u}}=(u_1, u_2, u_3, u_4)$ , $A{{\bf u}}:=(d_1\Delta u_1, d_2\Delta u_2, d_3\Delta u_3, d_4\Delta u_4)^T$ , $\varphi=(\varphi_1, \varphi_2, \varphi_3, \varphi_4)$ .

定理2.1 对任意给定的初值 $\varphi=(\varphi_1, \varphi_2, \varphi_3, \varphi_4)\in{{\bf X}}_{{\bf M}}$ , 系统(2.1)在 $[0, \infty)\times{{\Bbb R}}$ 上存在唯一的非负解 ${{\bf u}}(t, x;\varphi)$ 并且满足 ${{\bf u}}(t, \cdot;\varphi)=\varphi$ .另外, 对任意 $t\geq0$ , ${{\bf u}}(t, \cdot;\varphi)\in{{\bf X}}_{{\bf M}}$ , 且对任意的 $(t, x)\in[0, \infty)\times\mathbb{R}$ , ${{\bf u}}(t, \cdot;\varphi)$ 是系统(2.1)的一个古典解.

 对任意的 $\varphi=(\varphi_1, \varphi_2, \varphi_3, \varphi_4)\in{{\Bbb X}}_{{\bf M}}$ 和任意的 $\kappa\geq 0$ , 有

$ \varphi(x)+\kappa F(\varphi)(x) =\left(\begin{array}{c} \kappa \Lambda_1-\kappa[\beta_{11}f_1(\varphi_3)(x)+\beta_{12}f_2(\varphi_4)(x)] \varphi_1(x)+(1-\kappa\theta_1)\varphi_1(x) \\ \kappa \Lambda_2-\kappa[\beta_{21}f_1(\varphi_3)(x)+\beta_{22}f_2(\varphi_4)(x)] \varphi_2(x)+(1-\kappa\theta_2)\varphi_2(x)\\ \kappa[\beta_{11}f_1(\varphi_3)(x)+\beta_{12}f_2(\varphi_4)(x)]\varphi_1(x) +(1-\kappa\theta_3)\varphi_3(x) \\ \kappa[\beta_{21}f_1(\varphi_3)(x)+\beta_{22}f_2(\varphi_4)(x)]\varphi_2(x) +(1-\kappa\theta_4)\varphi_4(x) \end{array} \right). $

$ \eta_1=\frac{u_1^0}{\theta_3}\Big(\frac{\beta_{11}}{\alpha_1} +\frac{\beta_{12}}{\alpha_2}\Big), \quad \eta_2=\frac{u_2^0}{\theta_4}\Big(\frac{\beta_{21}}{\alpha_1} +\frac{\beta_{22}}{\alpha_2}\Big). $

则对任意 $0\leq\kappa<\min\big\{\frac{1}{\theta_3}, \frac{1}{\theta_4}, \kappa_1, \kappa_2\big\}$ , 其中

$ \kappa_1=\frac{1}{\theta_1+\beta_{11}f_1(\eta_1)+\beta_{12}f_2(\eta_2)}, \quad \kappa_2=\frac{1}{\theta_2+\beta_{21}f_1(\eta_1) +\beta_{22}f_2(\eta_2)}, $

$ \varphi(x)+hF(\varphi)(x) \geq\left(\begin{array}{c} [1-\kappa\theta_1-\kappa \beta_{11}f_1(\eta_1) -\kappa\beta_{12}f_2(\eta_2)]\varphi_1(x)\\ [1-\kappa\theta_2-\kappa\beta_{21}f_1(\eta_1)-\kappa\beta_{22} f_2(\eta_2)]\varphi_2(x)\\ (1-\kappa\theta_3)\varphi_3(x)\\ (1-\kappa\theta_4)\varphi_4(x) \end{array} \right) \geq\left(\begin{array}{c} 0\\ 0\\ 0\\ 0 \end{array} \right). $

另一方面, 对 $\kappa<\min\big\{\frac{1}{\theta_3}, \frac{1}{\theta_4}, \kappa_1, \kappa_2\big\}$ , 有

$ \varphi(x)+\kappa F(\varphi)(x) \leq\left(\begin{array}{c} \kappa \Lambda_1+(1-\kappa\theta_1)u_1^0 \\ \kappa \Lambda_2+(1-\kappa\theta_2)u_2^0 \\[2mm] \kappa\Big(\frac{\beta_{11}}{\alpha_1}+\frac{\beta_{12}}{\alpha_2}\Big)u_1^0 +(1-\kappa\theta_3)\eta_1 \\[3mm] \kappa\Big(\frac{\beta_{21}}{\alpha_1}+\frac{\beta_{22}}{\alpha_2}\Big)u_2^0 +(1-\kappa\theta_4)\eta_2\\ \end{array} \right) =\left(\begin{array}{c} {u_1^0} \\ {u_2^0} \\[2mm] \frac{1}{\theta_3}\Big(\frac{\beta_{11}}{\alpha_1} +\frac{\beta_{12}}{\alpha_2}\Big)u_1^0 \\[3mm] \frac{1}{\theta_4}\Big(\frac{\beta_{21}}{\alpha_1} +\frac{\beta_{22}}{\alpha_2}\Big)u_2^0 \\ \end{array} \right). $

于是 $\varphi(x)+\kappa F(\varphi)(x)\in{{\bf X}}_{{\bf M}}$ .则

$ \lim\limits_{\kappa\to 0^+}\frac{1}{\kappa} \mbox{dist}\big(\varphi(x)+\kappa F(\varphi)(x), \ {{\bf X}}_{{\bf M}}\big)=0, \ \forall \varphi \in{{\bf X}}_{{\bf M}}. $

依文献[14, Corollary 4] (令时滞为0), 系统(2.1)和(2.2)在 $[0, \infty)\times\mathbb{R}$ 上存在唯一温和解 ${\bf u}(t, x;\varphi)$ 满足 ${{\bf u}}(t, \cdot;\varphi)=\varphi$ ${{\bf u}}(t, \cdot;\varphi)\in{{\bf X}}_{{\bf M}}$ 对于任意 $t\geq0$ .进而, 运用文献[24, Corollary 2.2.5], 当 $t\geq0$ 时此温和解还是古典解.证毕.

3 辅助性结果
3.1 波方程

这一节研究波方程的特征值问题.在系统(2.1)中, 设 $u_i(t, x)=\tilde u_i(x+ct)$ , $i=1, 2, 3, 4$ .其中 $(\tilde u_1, \tilde u_2, \tilde u_3, \tilde u_4)$ , $\xi:=x+ct$ $c$ 分别称为波轮廓, 波坐标和波速.为简单起见, 仍用 $u_1, \ u_2, \ u_3, \ u_4, \ t$ 代替 $\tilde u_1$ , $\tilde u_2$ , $\tilde u_3$ , $\tilde u_4$ , $\xi$ , 由此得到下列波方程

$ \begin{equation}\label{3.1} \left\{\begin{array}{l} cu_1'=d_1u_1''+\Lambda_1-[\beta_{11}f_1(u_3)+\beta_{12}f_2(u_4)]u_1 -\theta_1u_1, \\ cu_2'=d_2u_2''+\Lambda_2-[\beta_{21}f_1(u_3)+\beta_{22}f_2(u_4)]u_2 -\theta_2u_2, \\ cu_3'=d_3u_3''+[\beta_{11}f_1(u_3)+\beta_{12}f_2(u_4)]u_1-\theta_3u_3, \\ cu_4'=d_4u_4''+[\beta_{21}f_1(u_3)+\beta_{22}f_2(u_4)]u_2-\theta_4u_4, \end{array}\right. \end{equation} $ (3.1)

且满足边界条件

$ \begin{equation}\label{3.2} \lim\limits_{t\to-\infty}(u_1(t), u_2(t), u_3(t), u_4(t))=(u_1^0, u_2^0, 0, 0), \end{equation} $ (3.2)
$ \begin{equation}\label{3.3} \lim\limits_{t\to+\infty}(u_1(t), u_2(t), u_3(t), u_4(t))=(u_1^*, u_2^*, u_3^*, u_4^*). \end{equation} $ (3.3)

对系统(3.1)的第三和第四个方程在 $E^0(u_1^0, u_2^0, 0, 0)$ 处线性化, 得到

$ \begin{equation}\label{3.4} \left\{\begin{array}{l} cu_3'=d_3u_3''+(\beta_{11}u_1^0-\theta_3)u_3+\beta_{12}u_1^0u_4, \\ cu_4'=d_4u_4''+(\beta_{22}u_2^0-\theta_4)u_4+\beta_{21}u_2^0u_3. \end{array} \right. \end{equation} $ (3.4)

$u_3=\nu_1{\rm e}^{\lambda t}$ , $u_4=\nu_2{\rm e}^{\lambda t}$ 代入系统(3.4), 得下列特征值问题

$ \begin{equation} \mbox{det} A(\lambda)=0, \quad A(\lambda)\left(\begin{array}{c} \nu_1 \\ \nu_2 \end{array} \right)=0, \quad \mbox{其中}\quad A(\lambda)=\left(\begin{array}{cc} h_1(\lambda)~~&\beta_{12}u_1^0 \\ \beta_{21}u_2^0~~& h_2(\lambda) \end{array} \right). \end{equation} $

$h_1(\lambda)=d_3\lambda^2-c\lambda+\beta_{11}u_1^0-\theta_3$ , $h_2(\lambda)=d_4\lambda^2-c\lambda+\beta_{22}u_2^0-\theta_4$ .

$\beta_{11}u_1^0<\theta_3$ $\beta_{22}u_2^0<\theta_4$ , 设

$ \lambda_3^\pm:=\frac{c\pm\sqrt{c^2-4d_3(\beta_{11}u_1^0-\theta_3)}}{2d_3}, \quad \lambda_4^\pm:=\frac{c\pm\sqrt{c^2- 4d_4(\beta_{22}u_2^0-\theta_4)}}{2d_4}, $

且设 $\lambda_M^\pm:=\max\{\lambda_3^\pm, \lambda_4^\pm\}$ $\lambda_m^\pm:=\min\{\lambda_3^\pm, \lambda_4^\pm\}$ .这里, $h_1(\lambda_3^\pm)=0$ $h_2(\lambda_4^\pm)=0$ .

注意到 $\beta_{11}u_1^0<\theta_3$ $\beta_{22}u_2^0<\theta_4$ .由 ${{\bf R}}_0>1$ 可知 $\det A(0)<0$ , 则方程 $\mbox{det} A(\lambda)=0$ 至少有一个正根.

在本文, 我们总体假设条件 $({\bf H}_0)$ 成立.

$({\bf H}_0)$ $\beta_{11}u_1^0<\theta_3$ , $\beta_{22}u_2^0<\theta_4$ ${{\bf R}}_0>1$ .

类似于文献[11, Lemma 2.1]的证明, 得到如下引理.

引理3.1 设条件 $({\bf H}_0)$ 成立.则存在一个正常数 $c^*$ 使得下列结论成立.

(1) 若 $c>c^*$ , 则特征方程 $\mbox{det} A(\lambda)=0$ 有三个正根 $0<\lambda_1<\lambda_2<\lambda_3$ , 和一个负根 $\lambda_4 <0$ , 且满足 $\lambda_1, \lambda_2\in(0, \lambda_m^+), \lambda_3\in(\lambda_M^+, +\infty), \ \lambda_4\in(-\infty, \lambda_m^-)$ .进一步, 有 $\mbox{det} A(\lambda)>0$ 对任意 $\lambda\in(\lambda_1, \lambda_2)$ ; $\mbox{de}t A(\lambda)<0$ 对任意 $\lambda\in(0, \lambda_m^+) \backslash(\lambda_1, \lambda_2)$ .

(2) 若 $c=c^*$ , 则特征方程 $\det A(\lambda)=0$ 有两个互异的正根和一个负根.

(3) 若 $0<c<c^*$ , 则特征方程 $\det A(\lambda)=0$ 有一个正根, 一个负根和一对具有正实部的复根.

3.2 向量值上、下解

若无特别说明, 总体假设 ${{\bf R}}_0>1$ $c>c^*$ 成立.令 $\lambda_1$ 为引理3.1(1)中定义的最小特征值且 $(\nu_1, \nu_2)\gg 0$ 是其对应的特征向量.为符号简化, 设向量 $(\nu_1, \nu_2)=(\nu_1, 1)\gg (0, 0)$ , 其中 $\nu_1$ 满足

$ \begin{equation}\label{3.5} h_1(\lambda_1)\nu_1+\beta_{12}u_1^0=0, \quad \beta_{21}u_2^0\nu_1+h_2(\lambda_1)=0. \end{equation} $ (3.5)

再由引理3.1(1)知对足够小的 $\eta\in(0, \lambda_2-\lambda_1)$ , 有

$ \mbox{det} A(\lambda_1+\eta)=h_1(\lambda_1+\eta)h_2(\lambda_1+\eta) -\beta_{12}\beta_{21}u_1^0u_2^0>0 $

$h_1(\lambda_1+\eta)<0, \quad h_2(\lambda_1+\eta)<0$ .由于

$ \frac{h_1(\lambda_1+\eta)}{\beta_{12}u_1^0} <\frac{\beta_{21}u_2^0}{h_2(\lambda_1+\eta)}<0, $

于是可选取常数 $h>0$ 使得

$ \begin{equation}\label{3.6} -\frac{\beta_{21}u_2^0}{h_2(\lambda_1+\eta)} <h<-\frac{h_1(\lambda_1+\eta)}{\beta_{12}u_1^0}. \end{equation} $ (3.6)

受文献[7, 17, 22]诱发, 构造系统(3.1)的向量值型上、下解.为此, 对 $t\in{\Bbb R}$ , 定义如下八个连续函数.

$ \begin{array}{*{4}{l}} & \bar u_1(t)=u_1^0=\frac{\Lambda_1}{\theta_1}, \\[3mm] & \bar u_2(t)=u_2^0=\frac{\Lambda_2}{\theta_2}, \\[3mm] &\bar u_3(t)=\min\{\nu_1{\rm e}^{\lambda_1t}, K_1\}, \\[2mm] &\bar u_4(t)=\min\{{\rm e}^{\lambda_1t}, K_2\}, \end{array} \quad\mbox{和}\quad \begin{array}{*{4}{l}} & \underline{u}_1(t)=\max\Big\{\frac{\alpha_1\alpha_2\Lambda_1} {\beta_{11}\alpha_2+\beta_{12}\alpha_1+\theta_1\alpha_1\alpha_2}, u_1^0-\frac{1}{\sigma}{\rm e}^{\sigma t}\Big\}, \\[3mm] & \underline{u}_2(t)=\max\Big\{\frac{\alpha_1\alpha_2\Lambda_2} {\beta_{21}\alpha_2+\beta_{22}\alpha_1+\theta_2\alpha_1\alpha_2}, u_2^0-\frac{1}{\sigma}{\rm e}^{\sigma t}\Big\}, \\[3mm] &\underline{u}_3(t)=\max\Big\{0, \nu_1{\rm e}^{\lambda_1t}-q{\rm e}^{(\lambda_1+\eta)t}\Big\}, \\[2mm] &\underline{u}_4(t)=\max\Big\{0, {\rm e}^{\lambda_1t}-qh{\rm e}^{(\lambda_1+\eta)t}\Big\}, \end{array} $

其中

$ K_1=\Big(\frac{\beta_{11}}{\alpha_1}+\frac{\beta_{12}}{\alpha_2}\Big) \frac{u_1^0}{\theta_3}, \quad K_2=\Big(\frac{\beta_{21}}{\alpha_1} +\frac{\beta_{22}}{\alpha_2}\Big)\frac{u_2^0}{\theta_4}, $

这里 $\sigma$ , $\eta$ , $q$ 是待定的正常数.

引理3.2 下列不等式成立.

$ \begin{eqnarray} N_1(\bar u_1)(t):&=&d_1\bar u''_1(t)-c\bar u'_1(t)+\Lambda_1 -[\beta_{11}f_1(\underline{u}_3(t))+\beta_{12}f_2(\underline{u}_4(t))] \bar u_1(t)\nonumber\\ &&-\theta_1\bar u_1(t)\leq0, \ \forall t \in {\Bbb R};\label{3.7} \end{eqnarray} $ (3.7)
$ \begin{eqnarray} N_2(\bar u_2)(t):&=&d_2\bar u''_2(t)-c\bar u'_2(t)+\Lambda_2 -[\beta_{21}f_1(\underline{u}_3(t))+\beta_{22}f_2(\underline{u}_4(t))] \bar u_2(t)\nonumber \\ &&-\theta_2\bar u_2(t) \leq0, \ \forall t \in {\Bbb R};\label{3.8} \end{eqnarray} $ (3.8)
$ \begin{eqnarray} N_3(\bar u_3)(t):&=&d_3\bar u''_3(t)-c\bar u'_3(t)+[\beta_{11}f_1(\bar u_3(t)) +\beta_{12}f_2(\bar u_4(t))]\bar u_1(t) \nonumber\\ &&-\theta_3\bar u_3(t) \leq0, \ \forall t \ne t_0:=\frac{1}{\lambda_1}\ln K_1;\label{3.9} \end{eqnarray} $ (3.9)
$\begin{eqnarray} N_4(\bar u_4)(t):&=& d_4\bar u''_4(t)-c\bar u'_4(t)+[\beta_{21}f_1(\bar u_3(t)) +\beta_{22}f_2(\bar u_4(t))]\bar u_2(t) \nonumber\\ &&-\theta_4\bar u_4(t) \leq0, \ \forall t \ne {t_1}:=\frac{1}{\lambda_1} \ln K_2.\label{3.10} \end{eqnarray} $ (3.10)

这里函数 $\bar u(t)=(\bar u_1(t), \bar u_2(t), \bar u_3(t), \bar u_4(t))$ 称作系统(3.1)的一个上解.

 注意到 $\bar u_1(t)=\frac{\Lambda_1}{\theta_1}$ , $\bar u_2(t)=\frac{\Lambda_2}{\theta_2}$ $\underline{u}_3(t)$ , $\underline{u}_4(t)$ 对任意 $t\in{\Bbb R}$ 是非负的, 则(3.7)式和(3.8)式成立.

由于(3.10)式的证明与(3.9)式类似, 下面仅验证(3.9)式成立.事实上, 当 $t<t_0$ 时, $\bar u_3(t)=\nu_1{\rm e}^{\lambda_1t}$ .而 $\bar u_1(t)=u_1^0$ $\bar u_4(t)\leq {\rm e}^{\lambda_1t}$ 对任意 $t\in\mathbb{R}$ 成立.于是

$ \begin{eqnarray*} N_3(\bar u_3)(t)&\leq& d_3\bar u''_3(t)-c\bar u'_3(t) +[\beta_{11}\bar u_3(t)+\beta_{12}\bar u_4(t)]\bar u_1(t)-\theta_3\bar u_3(t)\\ &\leq& [h_1(\lambda_1)\nu_1+\beta_{12}u^0_1]{\rm e}^{\lambda_1t} = 0. \end{eqnarray*} $

$t>t_0$ 时, $\bar u_1(t)=u_1^0$ , $\bar u_3(t)=K_1$ , 则

$ N_3(\bar u_3)(t)\leq\Big(\frac{\beta_{11}}{\alpha_1} +\frac{\beta_{12}}{\alpha_2}\Big)u_1^0-\theta_3K_1=0. $

故(3.9)式成立.证毕.

引理3.3 下列不等式成立.

$ \begin{eqnarray} N_1(\underline{u}_1)(t):&=& d_1\underline{u}_1''(t)-c\underline{u}_1'(t) +\Lambda_1-[\beta_{11}f_1(\bar u_3(t))+\beta_{12}f_2(\bar u_4(t))] \underline{u}_1(t)\nonumber\\ &&-\theta_1\underline{u}_1(t) \geq0;\ t \ne t_2:=\frac{1}{\sigma}\ln(u_1^0\sigma);\label{3.11} \end{eqnarray} $ (3.11)
$ \begin{eqnarray} N_2(\underline{u}_2)(t):&=& d_2\underline{u}_2''(t)-c\underline{u}_2'(t) +\Lambda_2-[\beta_{21}f_1(\bar u_3(t))+\beta_{22}f_2(\bar u_4(t))] \underline{u}_2(t)\nonumber\\ &&-\theta_2\underline{u}_2(t)\geq0;\ t\ne t_3:=\frac{1}{\sigma}\ln(u_2^0\sigma);\label{3.12} \end{eqnarray} $ (3.12)
$ \begin{eqnarray} N_3(\underline{u}_3)(t):&=& d_3\underline{u}_3''(t)-c\underline{u}_3'(t)+ [\beta_{11}f_1(\underline{u}_3(t))+\beta_{12}f_2(\underline{u}_4(t))] \underline{u}_1(t)\nonumber\\ &&-\theta_3\underline{u}_3(t) \geq0, \ \forall t \ne t_4:=\frac{1}{\eta}\ln\frac{\nu_1}{q};\label{3.13} \end{eqnarray} $ (3.13)
$ \begin{eqnarray} N_4(\underline{u}_4)(t):&=& d_4\underline{u}_4''(t)-c\underline{u}_4'(t)+ [\beta_{21}f_1(\underline{u}_3(t))+\beta_{22}f_2(\underline{u}_4(t))] \underline{u}_2(t)\nonumber\\ &&-\theta_4\underline{u}_4(t) \geq0, \ \forall t \ne t_5:=\frac{1}{\eta}\ln\frac{1}{qh};\label{3.14} \end{eqnarray} $ (3.14)

其中

$ \begin{eqnarray*} &0<&\sigma<\min\Big\{\frac{\lambda_1}{2}, \frac{1}{u_1^0}, \frac{1}{u_2^0}, \frac{c}{d_1+(\beta_{11}\nu_1+\beta_{12})(u_1^0)^2}, \frac{c}{d_2+(\beta_{21}\nu_1+\beta_{22})(u_2^0)^2}\Big\}, \\ & 0<&\eta<\min\Big\{\lambda_2-\lambda_1, \frac{\sigma}{2}, \frac{\lambda_1}{2} \Big\}, \\ &q >&\max\Big\{\nu_1, \frac{1}{h}, -\frac{(\alpha_1\beta_{11}\nu_1^2 +\alpha_2\beta_{12})u_1^0 +\frac{1}{\sigma}(\beta_{11}\nu_1+\beta_{12})} {h_1(\lambda_1+\eta)+h\beta_{12}u_1^0}, \\ &&\hspace{2cm}-\frac{(\alpha_1\beta_{21}\nu_1^2+\alpha_2\beta_{22})u_2^0 +\frac{1}{\sigma} (\beta_{21}\nu_1+\beta_{22})}{hh_2(\lambda_1+\eta)+\beta_{21}u_2^0}\Big\}. \end{eqnarray*} $

这里函数 $\underline{u}(t):=(\underline{u}_1(t), \underline{u}_2(t), \underline{u}_3(t), \underline{u}_4(t))$ 称作系统(3.1)的一个下解.

 由于 $\underline{u}_1(t)=\frac{\alpha_1\alpha_2\Lambda_1}{\beta_{11}\alpha_2 +\beta_{12}\alpha_1+\theta_1\alpha_1\alpha_2}$ $t>t_2$ , 则

$ \begin{eqnarray*} {N_1}(\underline{u}_1)(t)&\geq& d_1\underline{u}''_1(t)-c\underline{u}_1'(t) +\Lambda_1-\Big(\frac{\beta_{11}}{\alpha_1}+\frac{\beta_{12}}{\alpha_2}\Big) \underline{u}_1(t)-\theta_1\underline{u}_1(t)\\ &=& \Lambda_1-\Big(\frac{\beta_{11}}{\alpha_1}+\frac{\beta_{12}}{\alpha_2}\Big) \underline{u}_1(t)-\theta_1\underline{u}_1(t)\\ &=& 0. \end{eqnarray*} $

注意到 $\underline{u}_1(t)=u_1^0-\frac{1}{\sigma}{\rm e}^{\sigma t}$ $t<t_2$ 成立, 而 $\bar u_3(t)\leq \nu_1{\rm e}^{\lambda_1t}$ , $\bar u_4(t)\leq {\rm e}^{\lambda_1t}$ 对任意 $t\in\mathbb{R}$ 也成立, 并且

$ {\rm e}^{(\lambda_1-\sigma)t}<{\rm e}^{(\lambda_1-\sigma)\frac{1}{\sigma}\ln(u_1^0\sigma)} =(u_1^0\sigma)^{\frac{(\lambda_1-\sigma)}{\sigma}}<u_1^0\sigma, \ \ \forall t<t_2<0. $

$ \begin{eqnarray*} N_1(\underline{u}_1)(t)&\geq& d_1\underline{u}_1''(t)-c\underline{u}_1'(t) +\Lambda_1-[\beta_{11}\bar u_3(t)+\beta_{12}\bar u_4(t)]\underline{u}_1(t) -\theta_1\underline{u}_1(t) \\ &\geq& -d_1\sigma {\rm e}^{\sigma t}+ c{\rm e}^{\sigma t}+\Lambda_1 -\beta_{11}\nu_1{\rm e}^{\lambda_1t}\Big(u_1^0-\frac{1}{\sigma}{\rm e}^{\sigma t}\Big)\\ && -\beta_{12}{\rm e}^{\lambda_1t}\Big(u_1^0-\frac{1}{\sigma}{\rm e}^{\sigma t}\Big) -\theta_1\Big(u_1^0-\frac{1}{\sigma}{\rm e}^{\sigma t}\Big) \\ &>& {\rm e}^{\sigma t}[-d_1\sigma+c-\beta_{11}u_1^0\nu_1{\rm e}^{(\lambda_1-\sigma)t} -\beta_{12}u_1^0{\rm e}^{(\lambda_1-\sigma)t}]\\ &>& {\rm e}^{\sigma t}[-d_1\sigma+c-\sigma(\beta_{11}\nu_1+\beta_{12})(u_1^0)^2] \\ &>& 0. \end{eqnarray*} $

故(3.11)式成立.同理可证(3.12)式成立.

下面验证(3.13)式成立.事实上, 对任意 $t>t_4$ , $\underline{u}_3(t)=0$ , 有

$ N_3(\underline{u}_3)(t)=\beta_{12}f_2(\underline{u}_4(t))\underline{u}_1(t)\geq0. $

$t<t_4<0$ , $\underline{u}_1(t)\geq u_1^0-\frac{1}{\sigma}{\rm e}^{\sigma t}$ , $\underline{u}_3(t)=\nu_1{\rm e}^{\lambda_1t}- q{\rm e}^{(\lambda_1+\eta)t}$ 以及 $\underline{u}_4(t)\geq {\rm e}^{\lambda_1t}- qh{\rm e}^{(\lambda_1+\eta)t}$ .则, 由 $\nu_1{\rm e}^{-\eta t}>\nu_1{\rm e}^{-\eta \frac{\ln\frac{\nu_1}{q}}{\eta}}=q$ 可知

$ \begin{eqnarray*} &&d_3\underline{u}_3''(t)-c\underline{u}_3'(t)-\theta_3\underline{u}_3(t)\\ &=&\nu_1{\rm e}^{(\lambda_1+\eta)t}{\rm e}^{-\eta t}[d_3\lambda_1^2-c\lambda_1 -\theta_3]-q{\rm e}^{(\lambda_1+\eta)t}[d_3(\lambda_1+\eta)^2 -c(\lambda_1+\eta)-\theta_3]\\ &>&q{\rm e}^{(\lambda_1+\eta)t}[h_1(\lambda_1)-\beta_{11}u_1^0] -q{\rm e}^{(\lambda_1+\eta)t}[h_1(\lambda_1+\eta)-\beta_{11}u_1^0] \\ &=&-q{\rm e}^{(\lambda_1+\eta)t}[h_1(\lambda_1+\eta)-h_1(\lambda_1)] \end{eqnarray*} $

$ \begin{eqnarray*} \beta_{11}f_1(\underline{u}_3(t))\underline{u}_1(t) &\geq&\beta_{11}\underline{u}_1(t)\underline{u}_3(t)(1-\alpha_1\underline{u}_3(t))\\ &\geq&\beta_{11}\Big(u_1^0 -\frac{1}{\sigma}{\rm e}^{\sigma t}\Big) (\nu_1{\rm e}^{\lambda_1t}- q{\rm e}^{(\lambda_1+\eta)t}) -\alpha_1\beta_{11}u_1^0\nu_1^2{\rm e}^{2\lambda_1t} \\ &=&\beta_{11}u_1^0(\nu_1{\rm e}^{\lambda_1t}-q{\rm e}^{(\lambda_1+\eta)t}) -\frac{\beta_{11}}{\sigma}\nu_1{\rm e}^{(\lambda_1+\sigma)t}\\ &&+\frac{\beta_{11}}{\sigma}q{\rm e}^{(\lambda_1+\eta +\sigma)t} -\alpha_1\beta_{11}u_1^0\nu_1^2{\rm e}^{2\lambda_1t}, \end{eqnarray*} $

以及

$ \begin{eqnarray*} \beta_{12}f_2(\underline{u}_4(t))\underline{u}_1(t) &\geq& \beta_{12}\underline{u}_1(t)\underline{u}_4(t) -\alpha_2\beta_{12}\underline{u}_1(t)\underline{u}_4^2(t) \\ &\geq& \beta_{12}\Big(u_1^0-\frac{1}{\sigma}{\rm e}^{\sigma t}\Big) ({\rm e}^{\lambda_1t}- qh{\rm e}^{(\lambda_1+\eta)t})-\alpha_2\beta_{12}u_1^0{\rm e}^{2\lambda_1t}\\ &=& \beta_{12}u_1^0({\rm e}^{\lambda_1t}-qh{\rm e}^{(\lambda_1+\eta)t}) -\frac{\beta_{12}}{\sigma}{\rm e}^{(\lambda_1+\sigma)t}\\ &&+\frac{qh}{\sigma}\beta_{12}{\rm e}^{(\lambda_1+\eta+\sigma)t} -\alpha_2\beta_{12}u_1^0{\rm e}^{2\lambda_1t} . \end{eqnarray*} $

注意到 ${\rm e}^{(\lambda_1-\eta)t}<1$ ${\rm e}^{(\sigma-\eta)t}<1$ $t<t_4$ 成立.于是, 由(3.5)和(3.6)式, 得到

$ \begin{eqnarray*} N_3(\underline{u}_3)(t)&\geq& -q{\rm e}^{(\lambda_1+\eta)t} [h_1(\lambda_1+\eta)-h_1(\lambda_1)]+\beta_{11}u_1^0(\nu_1{\rm e}^{\lambda_1t} -q{\rm e}^{(\lambda_1+\eta)t})\\ &&-\frac{\beta_{11}}{\sigma}\nu_1{\rm e}^{(\lambda_1+\sigma)t} +\frac{\beta_{11}}{\sigma} q{\rm e}^{(\lambda_1+\eta+\sigma)t}-\alpha_1\beta_{11}u_1^0\nu_1^2{\rm e}^{2\lambda_1t}\\ &&+\beta_{12}u_1^0({\rm e}^{\lambda_1t}-qh{\rm e}^{(\lambda_1+\eta)t}) -\frac{\beta_{12}}{\sigma}{\rm e}^{(\lambda_1+\sigma)t} +\frac{qh}{\sigma}\beta_{12}{\rm e}^{(\lambda_1+\eta+\sigma)t} -\alpha_2\beta_{12}u_1^0{\rm e}^{2\lambda_1t}\\ &\geq& -{\rm e}^{(\lambda_1+\eta)t}\Big[q(h_1(\lambda_1+\eta)+\beta_{12}hu_1^0) +\frac{1}{\sigma}(\beta_{11}\nu_1+\beta_{12}){\rm e}^{(\sigma-\eta)t}\\ &&+(\alpha_1\beta_{11}\nu_1^2 +\alpha_2\beta_{12})u_1^0 {\rm e}^{(\lambda_1-\eta)t}\Big]\\ &>& -{\rm e}^{(\lambda_1+\eta)t}\Big[q(h_1(\lambda_1+\eta)+\beta_{12}hu_1^0) +\frac{1}{\sigma}(\beta_{11}\nu_1+\beta_{12})+(\alpha_1\beta_{11}\nu_1^2 +\alpha_2\beta_{12})u_1^0\Big]\\ &>& 0. \end{eqnarray*} $

故(3.13)式成立.同理可证(3.14)式成立.证毕.

3.3 Schauder不动点定理条件的验证

在这一小节, 我们将使用在3.2节里构造的上、下解验证Schauder不动点定理条件成立.令 $K^*=\max\{K_1, K_2\}$ $u=(u_1, u_2, u_3, u_4)\in C({{\Bbb R}}, \Omega)$ , 其中 $\Omega:=[0, u_1^0]\times[0, u_2^0]\times[0, K^*]\times[0, K^*]$ .定义以下四个常数

$ \beta_1\geq\beta_{11}f_1(K^*)+\beta_{12}f_2(K^*)+\theta_1, \beta_2\geq\beta_{21}f_1(K^*)+\beta_{22}f_2(K^*)+\theta_2, \beta_3\geq\theta_3, \beta_4\geq\theta_4, $

使得

$ H_1(u)(t):=\beta_1u_1(t)+\Lambda_1-[\beta_{11}f_1(u_3(t)) +\beta_{12}f_2(u_4(t))]u_1(t)-\theta_1u_1(t) $

对任意 $t\in{{\Bbb R}}$ , 关于 $u_1$ 单调递增, 关于 $u_3$ $u_4$ 单调递减;

$ H_2(u)(t):=\beta_2u_2(t)+\Lambda_2-[\beta_{21}f_1(u_3(t)) +\beta_{22}f_2(u_4(t))]u_2(t)-\theta_2u_2(t) $

对任意 $t\in{\Bbb R}$ , 关于 $u_2$ 单调递增, 关于 $u_3$ $u_4$ 单调递减;

$ H_3(u)(t):=\beta_3u_3(t)+[\beta_{11}f_1(u_3(t)) +\beta_{12}f_2(u_4(t))]u_1(t)-\theta_3u_3(t) $

对任意 $t\in{\Bbb R}$ , 关于 $u_i$ , $i=1, 3, 4$ 单调递增, 以及

$H_4(u)(t):=\beta_4u_4(t)+[\beta_{21}f_1(u_3(t)) +\beta_{22}f_2(u_4(t))]u_2(t)-\theta_4u_4(t) $

对任意 $t\in{{\Bbb R}}$ , 关于 $u_i$ , $i=2, 3, 4$ 单调递增.很明显, 系统(3.1)可写成

$ \begin{eqnarray}\label{3.15} \left\{\begin{array}{l} d_1u_1''(t)-cu_1'(t)-\beta_1u_1(t)+H_1(u)(t)=0, \\ d_2u_2''(t)-cu_2'(t)-\beta_2u_2(t)+H_2(u)(t)=0, \\ d_3u_3''(t)-cu_3'(t)-\beta_3u_3(t)+H_3(u)(t)=0, \\ d_4u_4''(t)-cu_4'(t)-\beta_4u_4(t)+H_4(u)(t)=0 . \end{array}\right. \end{eqnarray} $ (3.15)

定义集合 $\Gamma=\big\{(u_1, u_2, u_3, u_4)\in C({{\Bbb R}}, \Omega): \underline{u}_i(t)\leq u_i(t)\leq \bar u_i(t), \ \forall t\in\mathbb{R}, \ i= 1, 2, 3, 4\big\}$ .显然, $\Gamma$ 非空, 且在 $C({{\Bbb R}}, \Omega)$ 上是闭和紧的.定义映射 $F=(F_1, F_2, F_3, F_4):\Gamma\to C({{\Bbb R}}, \Omega)$ 满足

$ F_i(u)(t)=\frac{1}{D_i}\left(\int_{-\infty}^t{{\rm e}^{\lambda_{i1}(t-s)}} +\int_t^{+\infty}{{\rm e}^{\lambda_{i2}(t-s)}}\right)H_i(u)(s){\rm d}s, $

其中 $\lambda_{i1}<0<\lambda_{i2}$ $d_i\lambda^2-c\lambda-\beta_i=0$ 的根, 且 $D_i=d_i(\lambda_{i2}-\lambda_{i1})$ , $i=1, 2, 3, 4$ .

显然, $F$ 的不动点是系统(3.15)的一个解.由此, 系统(3.15)解的存在性的证明转化成验证映射 $F$ 满足Schauder不动点定理条件.为此, 需要下面三个引理.

引理3.4 F是从 $\Gamma$ $\Gamma$ 的映射.

 给定 $u=(u_1, u_2, u_3, u_4)\in\Gamma$ .显然, 仅需证明

$ \underline{u}_i(t)\leq F_i(u)(t)\leq\bar u_i(t), \ t\in{{\Bbb R}}, \ i=1, 2, 3, 4. $

基于常数 $\beta_i$ ( $i=1, 2, 3, 4$ )的选取, 对任意 $t\in{\Bbb R}$ , 只需证明下列不等式成立

$ \begin{eqnarray*} \underline{u}_1(t)\leq F_1(\underline{u}_1, \underline{u}_2, \bar u_3, \bar u_4)(t) \leq F_1(u)(t)\leq F_1(\bar u_1, \bar u_2, \underline{u}_3, \underline{u}_4)(t)\leq \bar u_1(t);\\ \underline{u}_2(t)\leq F_2(\underline{u}_1, \underline{u}_2, \bar u_3, \bar u_4)(t) \leq F_2(u)(t)\leq F_2(\bar u_1, \bar u_2, \underline{u}_3, \underline{u}_4)(t)\leq \bar u_2(t);\\ \underline{u}_3(t)\leq F_3(\underline{u}_1, \underline{u}_2, \underline{u}_3, \underline{u}_4)(t)\leq F_3(u)(t)\leq F_3(\bar u_1, \bar u_2, \bar u_3, \bar u_4)(t)\leq \bar u_3(t);\\ \underline{u}_4(t)\leq F_4(\underline{u}_1, \underline{u}_2, \underline{u}_3, \underline{u}_4)(t)\leq F_4(u)(t)\leq F_4(\bar u_1, \bar u_2, \bar u_3, \bar u_4)(t)\leq\bar u_4(t). \end{eqnarray*} $

首先, 考虑 $F_1(u)(t)$ $t \geq t_2$ .由(3.11)式, 有

$ \begin{eqnarray*} &&F_1(\underline{u}_1, \underline{u}_2, \bar u_3, \bar u_4)(t)\\ &=& \frac{1}{D_1}\Big(\int_{-\infty}^t {\rm e}^{\lambda_{11}(t-s)} +\int_t^{+\infty}{\rm e}^{\lambda_{12}(t-s)}\Big)H_1(\underline{u}_1, \underline{u}_2, \bar u_3, \bar u_4)(s){\rm d}s \\ &\geq& \frac{1}{D_1}\Big(\int_{-\infty}^t{\rm e}^{\lambda_{11}(t-s)} +\int_t^{+\infty}{\rm e}^{\lambda_{12}(t-s)}\Big)\big[-d_1\underline{u}_1''(s) +c\underline{u}_1'(s)+\beta_1\underline{u}_1(s)\big]{\rm d}s\\ &\geq&\underline{u}_1(t)+\frac{d_1}{D_1}{\rm e}^{\lambda_{11}(t-t_1)} \big[\underline{u}_1'(t_1+ 0)-\underline{u}_1'(t_1-0)\big]\\ &\geq&\underline{u}_1(t), \end{eqnarray*} $

最后不等式是由 $\underline{u}_1'(t_1+0)=0$ $\underline{u}_1'(t_1-0)<0$ 得到.

同理, 当 $t<t_2$ 时, 也可证明 $F_1(\underline{u}_1, \underline{u}_2, \bar u_3, \bar u_4)(t) \geq \underline{u}_1(t)$ 成立.再由 $F_1(\underline{u}_1, \underline{u}_2, \bar u_3, \bar u_4)(t)$ $\underline{u}_1(t)$ 关于 $t$ 的连续性, 可知 $F_1(\underline{u}_1, \underline{u}_2, \bar u_3, \bar u_4)(t) \geq \underline{u}_1(t)$ 对任意 $t\in {{\Bbb R}}$ 成立.由(3.7)式, 对任意 $t\in {\Bbb R}$ , 知

$ H_1(\bar u_1, \bar u_2, \underline{u}_3, \underline{u}_4)(t) \leq \beta_1\bar u_1(t)+ c\bar u_1'(t)-d_1\bar u_1''(t) \leq \beta_1u_1^0. $

$ F_1(\bar u_1, \bar u_2, \underline{u}_3, \underline{u}_4)(t) \leq \frac{\beta_1u_1^0}{D_1}\Big(\int_{-\infty}^t{{\rm e}^{\lambda_{11}(t-s)}} +\int_t^{+\infty}{{\rm e}^{\lambda_{12}(t-s)}}\Big){\rm d}s =u_1^0. $

$F_i(u)(t)$ ( $i=2, 3, 4$ )的证明与 $F_1(u)(t)$ 是类似的, 故省略.证毕.

$\sigma\in\big(0, \min\limits_{1\leq i\leq4}\{-\lambda_{i1}\}\big)$ .对 $\phi=(\phi_1, \phi_2, \phi_3, \phi_4)\in C({\Bbb R}, \Omega)$ , 定义

$ \|\phi\|_\sigma=\max\limits_{i=1, 2, 3, 4}\big\{\sup\limits_{t\in{\Bbb R}} |\phi_i(t)|{\rm e}^{-\sigma|t|}\big\}\quad \mbox{和}\quad B_\sigma({\Bbb R}, \Omega)=\big\{\phi\in C({\Bbb R}, \Omega): \|\phi\|_\sigma<\infty\big\}. $

容易验证 $B_\sigma({{\Bbb R}}, \Omega)$ 是以 $\|\cdot\|_\sigma$ 为范数的Banach空间.

引理3.5  $F=(F_1, F_2, F_3, F_4):\Gamma\to\Gamma$ 是关于范数 $\|\cdot\|_\sigma$ 连续的映射.

引理3.5的证明类似于文献[26, Lemma 3.6]或[28, Lemma 2.7], 从略.

引理3.6  $F=(F_1, F_2, F_3, F_4):\Gamma\to\Gamma$ 是关于范数 $\|\cdot\|_\sigma$ 的紧映射.

引理3.6的证明类似于类似于文献[26, Lemma 3.7]或[28, Lemma 2.8], 从略.

4 行波解的存在性及不存在性
4.1 行波解的存在性

先给出系统(3.1)的解的一些性质.

性质4.1 设条件 $({\bf H}_0)$ 成立.对任意 $c>c^*$ , 系统(3.1)存在一个非平凡正解 $u(t)=(u_1(t), u_2(t)$ , $u_3(t), u_4(t))$ 满足

$ \begin{equation}\label{4.1} \lim\limits_{t\to-\infty}(u_1(t), u_2(t), u_3(t), u_4(t))=(u_1^0, u_2^0, 0, 0), \end{equation} $ (4.1)

$ \begin{equation}\label{4.2} 0<u_1(t)\leq u_1^0, \ 0<u_2(t)\leq u_2^0, \ 0<u_3(t)\leq K_1, \ 0<u_4(t)\leq K_2, \ \forall t\in{\Bbb R} \end{equation} $ (4.2)

$ \begin{equation}\label{4.3} \lim\limits_{t\to-\infty}{\rm e}^{-\lambda_1t}u_3(t)=\nu_1, \ \lim\limits_{t\to-\infty}{\rm e}^{-\lambda_1t}u_4(t)=1. \end{equation} $ (4.3)

 由引理3.4, 3.5和3.6, 根据Schauder不动点定理, 映射 $F$ 存在一个不动点 $(u_1, u_2, $ $u_3, $ $u_4) \in\Gamma$ .从而, 函数 $(u_1(t), u_2(t), u_3(t), u_4(t))$ 是系统(3.1)的一个解, 且有

$ 0\leq u_1(t)\leq u_1^0, \ 0\leq u_2(t) \leq u_2^0, \ 0\leq u_3(t) \leq K_1, \ 0\leq u_4(t)\leq K_2, \ \forall t \in {\Bbb R}. $

我们断言:对任意 $t\in{\Bbb R}$ , $u_i(t)>0$ ( $i=1, 2, 3, 4$ ).事实上, 注意到 $u=(u_1, u_2, u_3, u_4)\in\Gamma$ 是映射 $F$ 的一个不动点, 则 $u_i(t)=F_i(u(t))$ , $i=1, 2, 3, 4$ .显然, 对任意的 $t\in{\Bbb R}$ , 有 $u_1(t)>0$ $u_2(t)>0$ .于是, 仅需证明, 对任意 $t\in {\Bbb R}$ , $u_3(t)>0$ $u_4(t)>0$ .事实上, 若存在 $T_0\in {{\Bbb R}}$ 使得 $u_3(T_0)=0$ , 则存在常数 $T_1, T_2\in {{\Bbb R}}$ 使得 $T_1<\frac{1}{\eta}\ln\frac{\nu_1}{q}\leq T_2$ $T_0\in(T_1, T_2)$ .对任意 $t\in[T_1, T_2]$ , $u_3(t)$ $(T_1, T_2)$ 内可以取到最小值.则由系统(3.1)的第三个方程, 可知

$ -d_3u_3''(t)+cu_3'(t)+\theta_3u_3(t)\geq 0, \ \ t\in[T_1, T_2]. $

利用强极大值原理[27, 引理2.1.2], 知对 $t\in[T_1, T_2], $ $u_3(t)\equiv0 $ .然而, 根据引理3.3, $u_3(t)>0$ 对于 $t\in[T_1, \frac{1}{\eta}\ln\frac{\nu_1}{q})$ .矛盾.同理可证 $u_4(t)>0$ 对任意 $t\in {\Bbb R}$ 成立.

另一方面, 由于 $u\in\Gamma$ , 则, 对于 $t\in {\Bbb R}$ , 下列不等式成立.

$ u_1^0-\frac{1}{\sigma}{\rm e}^{\sigma t}\leq u_1(t)\leq u_1^0, \quad u_2^0-\frac{1}{\sigma}{\rm e}^{\sigma t}\leq u_2(t)\leq u_2^0 $

$ \nu_1{\rm e}^{\lambda_1t}-q{\rm e}^{(\lambda_1+\eta)t}\leq u_3(t)\leq\nu_1{\rm e}^{\lambda_1t}, \quad {\rm e}^{\lambda_1t}-qh{\rm e}^{(\lambda_1+\eta)t}\leq u_4(t)\leq {\rm e}^{\lambda_1t}, $

从而

$ \lim\limits_{t\to-\infty}u_1(t)=u_1^0, \ \lim\limits_{t\to-\infty}u_2(t)=u_2^0, \ \lim\limits_{t\to-\infty} u_3(t)=0, \ \lim\limits_{t\to-\infty}u_4(t)=0 $

$ \lim\limits_{t\to-\infty}{\rm e}^{-\lambda_1t}u_3(t)=\nu_1, \quad \lim\limits_{t\to-\infty}{\rm e}^{-\lambda_1t}u_4(t)=1. $

证毕.

下面,构造Lyapunov函数, 证明行波解在 $t=+\infty$ 时收敛于无病平衡点 $E^*$ .为此, 先给出系统(3.1)的解 $u(t)=(u_1(t)$ , $u_2(t), u_3(t), u_4(t))$ 有界性的估计.

引理4.1 设 $u(t)=(u_1(t), u_2(t), u_3(t), u_4(t))$ 是系统(3.1)满足条件(4.2)的解.则存在正常数 $M_{ij}$ , $i=1, 2, 3, 4$ , $j=1, 2$ , 使得下列不等式成立.

$ -M_{i1}u_i(t)<u'_i(t)< M_{i2}u_i(t), \ \ i=1, 2, 3, 4, \forall t\geq 0. $

引理4.1的证明是初等的, 可仿文献[8, Lemma 3.1]证得, 从略.

为证明下面的性质4.2, 还需要以下引理.

引理4.2 对于任意正数 $x^*$ , 不等式

$ F(x):=g\Big(\frac{f(x)}{f(x^*)}\Big)-g\Big(\frac{x}{x^*}\Big)\leq0, \ \ \forall x>0 $

成立, 其中 $g(x)=x-1-\ln x$ 以及 $f(x)=\frac{x}{1+\alpha x}$ , $\alpha>0$ .

对函数 $F(x)$ 求出最大值点, 即可证得, 从略.

性质4.2 设 $u(t)=(u_1(t), u_2(t), u_3(t), u_4(t))$ 是系统(3.1)满足条件(4.2)的解, 则

$ \lim\limits_{t\to+\infty}(u_1(t), u_2(t), u_3(t), u_4(t))=(u_1^*, u_2^*, u_3^*, u_4^*). $

 定义开的有界集 ${\Bbb D}$ 如下

$ {{\bf D}}=\left\{u\in C([0, \infty), \Omega): \begin{array}{l} -M_{i1}u_i(t)<u'_i(t)<M_{i2}u_i(t)\ \mbox{对 } t\geq 0, M_{i1}, \\ M_{i2}, i=1, 2, 3, 4, \mbox{ 由引理4.1给定} \end{array} \right\} $

定义如下Lyapunov函数 $V:{\Bbb D}\to\mathbb{R}$ :

$ V(u):= ad_1u_1'\Big(\frac{u_1^*}{u_1}- 1\Big)+bd_2u_2'\Big(\frac{u_2^*}{u_2}-1\Big) +ad_3u_3'\Big(\frac{u_3^*}{u_3}-1\Big)+bd_4u_4'\Big(\frac{u_4^*}{u_4}- 1\Big) +cV_g(u), $

其中

$ V_g(u):=au_1^*g\Big(\frac{u_1}{u_1^*}\Big) +bu_2^*g\Big(\frac{u_2}{u_2^*}\Big)+au_3^*g\Big(\frac{u_3}{u_3^*}\Big) +bu_4^*g\Big(\frac{u_4}{u_4^*}\Big) $

$ a:=\beta_{21}f_1(u_3^*)u_2^*, b:=\beta_{12}f_2(u_4^*)u_1^*, g(x)=x-1-\ln x, \ \mbox{对于任意}\ x>0. $

$V$ 沿着系统(3.1)的解 $u(t)$ 求导, 得到

$ \begin{eqnarray*} \frac{{\rm d}V}{{\rm d}t}&=& a\Big(1-\frac{u_1^*}{u_1}\Big)[d_1u_1''+\Lambda_1-\beta_{11} f_1(u_3)u_1-\beta_{12}f_2(u_4)u_1-\theta_1u_1] \\ &&+b\Big(1-\frac{u_2^*}{u_2}\Big)[d_2u_2''+\Lambda_2-\beta_{21}f_1(u_3)u_2 -\beta_{22}f_2(u_4)u_2-\theta_2u_2]\\ && +a\Big(1-\frac{u_3^*}{u_3}\Big)[d_3u_3''+\beta_{11}f_1(u_3)u_1 +\beta_{12}f_2(u_4)u_1-\theta_3u_3]\\ &&+b\Big(1-\frac{u_4^*}{u_4}\Big)[d_4u_4''+\beta_{21}f_1(u_3)u_2 +\beta_{22}f_2(u_4)u_2-\theta_4u_4] \\ && +ad_1\Big[u_1''\Big(\frac{u_1^*}{u_1}-1\Big) -u_1^*\Big(\frac{u_1'}{u_1}\Big)^2\Big] +bd_2\Big[u_2''\Big(\frac{u_2^*}{u_2}-1\Big) -u_2^*\Big(\frac{u_2'}{u_2}\Big)^2\Big] \\ && +ad_3\Big[u_3''\Big(\frac{u_3^*}{u_3}-1\Big) -u_3^*\Big(\frac{u_3'}{u_3}\Big)^2\Big] +bd_4\Big[u_4''\Big(\frac{u_4^*}{u_4}-1\Big) -u_4^*\Big(\frac{u_4'}{u_4}\Big)^2\Big]. \end{eqnarray*} $

把平衡点 $E^*$ 所满足的等式代入上式并简化, 得到

$ \begin{eqnarray*} \frac{{\rm d}V}{{\rm d}t}&=& -a\Big[d_1u_1^*\Big(\frac{u_1'}{u_1}\Big)^2 +d_3u_3^*\Big(\frac{u_3'}{u_3}\Big)^2\Big] -b\Big[d_2u_2^*\Big(\frac{u_2'}{u_2}\Big)^2+d_4u_4^*\Big(\frac{u_4'}{u_4}\Big)^2\Big]\\ &&+a\beta_{11}f_1(u_3^*)u_1^*\Big[2-\frac{u_1^*}{u_1} +\frac{f_1(u_3)}{f_1(u_3^*)}-\frac{f_1(u_3)u_1u_3^*}{f_1(u_3^*)u_1^*u_3} -\frac{u_3}{u_3^*}\Big]\\ &&+ab\Big[2-\frac{u_1^*}{u_1}+\frac{f_2(u_4)}{f_2(u_4^*)} -\frac{f_2(u_4)u_1u_3^*}{f_2(u_4^*)u_1^*u_3}-\frac{u_3}{u_3^*}\Big] \\ &&+ab\Big[2-\frac{u_2^*}{u_2}+\frac{f_1(u_3)}{f_1(u_3^*)} -\frac{f_1(u_3)u_2u_4^*}{f_1(u_3^*)u_2^*u_4}-\frac{u_4}{u_4^*}\Big] \\ &&+b\beta_{22}f_2(u_4^*)u_2^*\Big[2-\frac{u_2^*}{u_2} +\frac{f_2(u_4)}{f_2(u_4^*)}-\frac{f_2(u_4)u_2u_4^*}{f_2(u_4^*)u_2^*u_4} -\frac{u_4}{u_4^*}\Big]\\ &&-a\theta_1\frac{(u_1-u_1^*)^2}{u_1}-b\theta_2 \frac{(u_2-u_2^*)^2}{u_2}. \end{eqnarray*} $

由引理4.2知

$ \begin{eqnarray*} &&2-\frac{u_1^*}{u_1}+\frac{f_1(u_3)}{f_1(u_3^*)} -\frac{f_1(u_3)u_1u_3^*}{f_1(u_3^*)u_1^*u_3}-\frac{u_3}{u_3^*}\\ &= &-g\Big(\frac{u_1^*}{u_1}\Big) -g\Big(\frac{f_1(u_3)u_1u_3^*}{f_1(u_3^*)u_1^*u_3}\Big) +g\Big(\frac{f_1(u_3)}{f_1(u_3^*)}\Big)-g\Big(\frac{u_3}{u_3^*}\Big)\\ &\leq& 0, \\[2mm] &&2-\frac{u_1^*}{u_1}+\frac{f_2(u_4)}{f_2(u_4^*)} -\frac{f_2(u_4)u_1u_3^*}{f_2(u_4^*)u_1^*u_3}-\frac{u_3}{u_3^*} +2-\frac{u_2^*}{u_2}+\frac{f_1(u_3)}{f_1(u_3^*)} -\frac{f_1(u_3)u_2u_4^*}{f_1(u_3^*)u_2^*u_4}-\frac{u_4}{u_4^*}\\ &=&-g\Big(\frac{u_1^*}{u_1}\Big)-g\Big(\frac{f_2(u_4)u_1u_3^*} {f_2(u_4^*)u_1^*u_3}\Big) -g\Big(\frac{u_2^*}{u_2}\Big)-g\Big(\frac{f_1(u_3)u_2u_4^*}{f_1(u_3^*)u_2^*u_4}\Big)\\ &&+g\Big(\frac{f_2(u_4)}{f_2(u_4^*)}\Big)-g\Big(\frac{u_4}{u_4^*}\Big) +g\Big(\frac{f_1(u_3)}{f_1(u_3^*)}\Big)-g\Big(\frac{u_3}{u_3^*}\Big)\\ &\leq& 0, \\ &&2-\frac{u_2^*}{u_2}+\frac{f_2(u_4)}{f_2(u_4^*)}-\frac{f_2(u_4)u_2u_4^*} {f_2(u_4^*)u_2^*u_4}-\frac{u_4}{u_4^*}\\ &=&-g\Big(\frac{u_2^*}{u_2}\Big) -g\Big(\frac{f_2(u_4)u_2u_4^*}{f_2(u_4^*)u_2^*u_4}\Big) +g\Big(\frac{f_2(u_4)}{f_2(u_4^*)}\Big)-g\Big(\frac{u_4}{u_4^*}\Big)\\ &\leq& 0, \end{eqnarray*} $

因此, $\frac{{\rm d}V}{{\rm d}t}\leq0$ $t\geq 0$ , 即 $V(t)$ $t\geq 0$ 单调递减, 且

$ \begin{equation}\label{4.4} \frac{{\rm d}V}{{\rm d}t}=0, \ \mbox{当且仅当}\ u_i=u_i^*, \ i=1, 2, 3, 4. \end{equation} $ (4.4)

取单调递增的常数列 $\{t_n\}$ 满足 $\lim\limits_{n\to\infty}t_n=+\infty$ , 并定义

$ \{u_i^{(n)}(t)\}_{n=1}^\infty:=\{u_i(t+t_n)\}_{n=1}^\infty, \ i= 1, 2, 3, 4. $

注意到 $\{u_i^{(n)}(t)\}_{n=1}^\infty$ 是有界的(由(4.2)式).则存在一个子序列(仍记为为 $\{u_i^{(n)}(t)\}_{n = 1}^\infty$ )使得 $\lim\limits_{n\to\infty}u_i^{(n)}(t):=u_i^{(\infty)}(t)$ .

另一方面, 由引理4.1, 易知 $V(t)$ ${{\bf D}}$ 上有下界, 即存在常数 $C\in{\Bbb R}$ 使得 $V(t)\geq C$ 对任意 $t\geq 0$ 成立.再注意到 $V(t)$ 关于 $t\geq0$ 是单调递减.因此, 对任意 $n\in{{\bf N}}$ , 有

$ C\leq V(u_1^{(n)}, u_2^{(n)}, u_3^{(n)}, u_4^{(n)})(t)= V(u_1, u_2, u_3, u_4)(t+t_n) \leq V(u_1, u_2, u_3, u_4)(t). $

则存在 $V_0\in{\Bbb R}$ 使得, 对任意 $t\geq 0$ , 有

$ \lim\limits_{n\to\infty}V(u_1^{(n)}, u_2^{(n)}, u_3^{(n)}, u_4^{(n)})(t) =\lim\limits_{t+t_n\to\infty}V(u_1^{(n)}, u_2^{(n)}, u_3^{(n)}, u_4^{(n)})(t+t_n) ={V_0}. $

$ V_0=\lim\limits_{n\to\infty}V(u_1^{(n)}, u_2^{(n)}, u_3^{(n)}, u_4^{(n)})(t) =V(u_1^{(\infty)}, u_2^{(\infty)}, u_3^{(\infty)}, u_4^{(\infty)})(t). $

于是

$ \frac{\rm d}{{\rm d}t}V(u_1^{(\infty)}, u_2^{(\infty)}, u_3^{(\infty)}, u_4^{(\infty)})(t)=0. $

再由(4.4)式, 可知

$ \lim\limits_{t\to+\infty}(u_1(t), u_2(t), u_3(t), u_4(t))=(u_1^*, u_2^*, u_3^*, u_4^*). $

证毕.

由性质4.1和4.2, 可得系统(2.1)的行波解的存在性.

定理4.1 设条件 $({\bf H}_0)$ 成立.对任意 $c>c^*$ , 系统(2.1)存在一个非平凡正的行波解 $(u_1(x+ct), u_2(x+ct), u_3(x+ct), u_4(x+ct))$ 满足边界条件(3.2)-(3.3)和(4.2)-(4.3).

4.2 行波解的不存在性

这一小节研究系统(2.1)的行波解的不存在性.先给出行波解的先验估计.

引理4.3 设条件 $({\bf H}_0)$ 成立.若 $u(t)=(u_1(t), u_2(t), u_3(t), u_4(t))$ 是系统(3.1)满足条件(3.2)和(3.3)的一个非平凡且非负解, 则存在正常数 $\omega$ , $\rho_1$ $\rho_2$ 使得

$ \begin{equation}\label{4.5} \sup\limits_{t\in{{\Bbb R}}}\big\{u_i(t){\rm e}^{-\omega t}\big\}<+\infty, \ \sup\limits_{t\in{{\Bbb R}}}\big\{|u_i'(t)|{\rm e}^{-\omega t}\big\}<+\infty, \ \sup\limits_{t\in{{\Bbb R}}}\big\{|u_i ''(t)|{\rm e}^{-\omega t}\big\}<+\infty, \ i=3, 4, \end{equation} $ (4.5)
$ \begin{equation} \sup\limits_{t\in{{\Bbb R}}}\big\{|u_1^0u_3(t)-f_1(u_3(t))u_1(t) |{\rm e}^{-(\omega+\rho_1)t}\big\}<+\infty, \label{4.6} \end{equation} $ (4.6)
$ \begin{equation} \sup\limits_{t\in{{\Bbb R}}}\big\{|u_1^0u_4(t)-f_2(u_4(t))u_1(t)| {\rm e}^{-(\omega+\rho_1)t}\big\}<+\infty, \label{4.7} \end{equation} $ (4.7)
$ \begin{equation} \sup\limits_{t\in{{\Bbb R}}}\big\{|u_2^0u_3(t)-f_1(u_3(t))u_2(t)| {\rm e}^{-(\omega+\rho_2)t}\big\}<+\infty, \label{4.8} \end{equation} $ (4.8)
$ \begin{equation} \sup\limits_{t\in{{\Bbb R}}}\big\{|u_2^0u_4(t)-f_2(u_4(t))u_2(t)| {\rm e}^{-(\omega+\rho_2)t}\big\}<+\infty.\label{4.9} \end{equation} $ (4.9)

 由于 $u(t)=(u_1(t), u_2(t), u_3(t), u_4(t))$ 是系统(3.1)满足条件(3.2)和(3.3)的非平凡且非负解, 则 $u(t)$ 满足

$ \begin{eqnarray}\label{4.10} \begin{array}{l} d_3u_3''(t)-cu_3'(t)+[\beta_{11}f_1(u_3(t))+\beta_{12}f_2(u_4(t))]u_1(t) -\theta_3u_3(t)=0, \\ d_4u_4''(t)-cu_4'(t)+[\beta_{21}f_1(u_3(t))+\beta_{22}f_2(u_4(t))]u_2(t) -\theta_4u_4(t)=0. \end{array} \end{eqnarray} $ (4.10)

$u_3'=w_3$ , $u_4'=w_4$ , 则(4.10)式可以写成

$ y'=By+f(t, y), $

其中

$ y=\left(\begin{array}{c} u_3 \\ w_3 \\ u_4 \\ w_4 \\ \end{array}\right), \quad B=\left(\begin{array}{cccc} 0& 1& 0&~~ 0 \\[2mm] {\frac{\theta_3}{d_3}-u_1^0\frac{\beta_{11}}{d_3}} & ~~ {\frac{c}{d_3}}~~& {-u_1^0\frac{\beta_{12}}{d_3}}& ~~0 \\[2mm] 0& 0& 0& ~~1 \\[2mm] {-u_2^0\frac{\beta_{21}}{d_4}}&0 & {\frac{\theta_4}{d_4}-u_2^0\frac{\beta_{22}}{d_4}} & ~~{\frac{c}{d_4}} \end{array}\right), $
$ f(t, y)=\left({\begin{array}{c} 0 \\[2mm] \frac{\beta_{11}}{d_3}u_1^0u_3(t)-\frac{\beta_{11}}{d_3}f_1(u_3(t))u_1(t) +\frac{\beta_{12}}{d_3}u_1^0u_4(t)-\frac{\beta_{12}}{d_3}f_2(u_4(t))u_1(t) \\[2mm] 0 \\[2mm] \frac{\beta_{21}}{d_4}u_2^0u_3(t)-\frac{\beta_{21}}{d_4}f_1(u_3(t))u_2(t) +\frac{\beta_{22}}{d_4}u_2^0u_4(t)-\frac{\beta_{22}}{d_4}f_2(u_4(t))u_2(t) \end{array}} \right). $

易知矩阵 $B$ 的特征方程就是 $\mbox{det}A(\lambda)$ .由引理3.1(3)知, 方程 $\mbox{det}A(\lambda)=0$ 有一个正的特征值, 一个负的特征值以及一对具有正实部的共轭复特征值.因此, 初始平衡点 $(u_1^0, u_2^0, 0, 0)$ 双曲的.根据稳定流形定理[16, p107] (或文献[28, Lemma 3.1]的证明), 知存在一个正常数 $\omega\in{{\Bbb R}}$ 使得

$ \sup\limits_{t\in{{\Bbb R}}}\{u_i(t){\rm e}^{-\omega t}\}<+\infty, \quad \sup\limits_{t\in{{\Bbb R}}}\{|u_i'(t)|{\rm e}^{-\omega t}\}<+\infty, \ i=3, 4. $

注意到 $(u_1(t), u_2(t), u_3(t), u_4(t))$ 满足条件(3.2)和(3.3), 同时由(4.10)式, 可得

$ \sup\limits_{t\in{{\Bbb R}}}\{|u_i''(t)|{\rm e}^{-\omega t}\}<+\infty, \ i=3, 4. $

下面, 考虑 $u_1^0u_3(t)-f_1(u_3(t))u_1(t)$ 指数估计.显然

$ \begin{equation}\label{4.11} u_1^0u_3(t)-f_1(u_3(t))u_1(t)=\alpha_1u_1^0\frac{u_3^2(t)}{1+\alpha_1u_3(t)} +(u_1^0-u_1(t))\frac{u_3(t)}{1+\alpha_1u_3(t)}. \end{equation} $ (4.11)

注意到, 对任意 $t\in{{\Bbb R}}$ , 有 $\frac{u_3^2(t)}{1+\alpha_1u_3(t)}\leq u_3^2(t)$ , 再由(4.5)式, 可知

$ \begin{equation}\label{4.12} \sup\limits_{t\in{{\Bbb R}}}\Big\{\frac{u_3^2(t)}{1+\alpha_1u_3(t)} {\rm e}^{-2\omega t}\Big\}<+\infty. \end{equation} $ (4.12)

$U_1(t)=u_1^0-u_1(t)$ .则由定理2.1, 对任意 $t\in{\Bbb R}$ , $U_1(t)\geq0$ .对系统(3.1)的第一个方程从 $-\infty $ $t<0$ 积分, 得到

$ \begin{equation}\label{4.13} d_1U'_1(t)-cU_1(t)=-f(t), \quad t<0. \end{equation} $ (4.13)

由于 $0\leq u_1(t)\leq u^0_1$ , 对任意 $t\in{\Bbb R}$ , 于是

$ \begin{eqnarray*} f(t)&=&\int_{-\infty}^t\big[(\beta_{11}f_1(u_3(s))+\beta_{12}f_2(u_4(s)))u_1(s) -\Lambda_1+\theta_1u_1(s)\big]{\rm d}s\\ &\leq& \int_{-\infty}^t \big[\beta_{11}f_1(u_3(s))+\beta_{12}f_2(u_4(s))\big] u_1(s){\rm d}s\nonumber \\ &\leq& C_0 {\rm e}^{\omega t}, \ \ C_0>0. \end{eqnarray*} $

即有 $f(t)=O({\rm e}^{\omega t})$ $t\to -\infty$ .求解(4.13)式, 得到

$ U_1(t)=U_1(0){\rm e}^{\frac{c}{d_1}t}+\frac{1}{d_1}{\rm e}^{\frac{c}{d_1}t} \int_t^0 {\rm e}^{-\frac{c}{d_1}s}f(s){\rm d}s, \ t<0. $

$\rho_1<\min\{\frac{c}{d_1}, \frac{c}{d_2}, \omega\}$ , 得到

$ U_1(t){\rm e}^{-\rho_1 t}=U_1(0){\rm e}^{(\frac{c}{d_1}-\rho_1)t}+\frac{1}{d_1} {\rm e}^{(\frac{c}{d_1}-\rho_1)t}\int_t^0 {\rm e}^{-\frac{c}{d_1}s}f(s){\rm d}s, \ t<0. $

注意到, 当 $t\to -\infty$ 时, $f(t)=0({\rm e}^{\omega t})$ , 易知 $U_1(t)=0({\rm e}^{\rho_1 t})$ $t\to-\infty$ .再利用 $0\leq U_1(t)\leq u_1^0$ 对任意 $t\in{{\Bbb R}}$ 成立, 以及 $\lim\limits_{t\to+\infty} U_1(t){\rm e}^{-\rho_1t}=0$ , 可得

$ \begin{equation}\label{4.14} \sup\limits_{t\in{{\Bbb R}}}\big\{U_1(t){\rm e}^{-\rho_1t}\big\}<+\infty . \end{equation} $ (4.14)

于是, 由(4.11), (4.12)和(4.14)式, 知(4.6)式成立.同理可证(4.7)式成立.类似于(4.6)和(4.7)式的证明, 可知存在正常数 $\rho_2<\min\{\frac{c}{d_3}, \frac{c}{d_4}, \omega\}$ 使得(4.8)和(4.9)式成立.证毕.

定理4.2 设条件 $({\bf H}_0)$ 成立.则对任意 $c\in(0, c^*)$ , 系统(2.1)不存在非平凡且非负行波解 $(u_1(x+ct), u_2(x+ct), u_3(x+ct), u_4(x+ct))$ 满足条件(3.2)和(3.3).

 设 $u(t)=(u_1(t), u_2(t), u_3(t), u_4(t))$ 是系统(3.1)满足条件(3.2)和(3.3)的一个非平凡且非负解.令 $\rho=\min\{\rho_1, \rho_2\}$ , 则由(4.6)-(4.9)式, 可得

$ \sup\limits_{t\in{\Bbb R}}\big\{\big[\beta_{11}u_1^0u_3(t)+\beta_{12}u_1^0u_4(t) -[\beta_{11}f_1(u_3(t))+\beta_{12}f_2(u_4(t))]u_1(t)\big]{\rm e}^{-(\omega+\rho)t}\big\} <+\infty $

$ \sup\limits_{t\in{\Bbb R}}\big\{\big[\beta_{21}u_2^0u_3(t)+\beta_{22}u_2^0u_4(t) -[\beta_{21}f_1(u_3(t))+\beta_{22}f_2(u_4(t))]u_2(t)\big]{\rm e}^{-(\omega+\rho)t}\big\} <+\infty. $

对任意 $\lambda\in{{\bf C}}$ 满足 $0<{\rm Re}\lambda<\omega$ , 关于函数 $u_i(t)$ ( $i=3, 4$ )的双边Laplace变换定义如下

$ L_i(\lambda)=\int_{-\infty}^{+\infty}{\rm e}^{-\lambda t}u_i(t){\rm d}t, \ i=3, 4. $

由Laplace变换性质[23] $L_i(\lambda)$ ( $i=3, 4$ )关于 $\lambda\in{{\bf C}}$ 解析, 且满足 $0<{\rm Re}\lambda_i<\lambda_i^*<+\infty$ , 其中 $\lambda=\lambda_i^*$ $L_i(\lambda)$ 的一个奇点, 或者 $\lambda_i^*=+\infty$ .

下面断言: $\lambda_3^*=\lambda_4^*<+\infty$ .事实上, 由引理4.3, 易知 $\lambda_i^*\geq\omega, \ i=3, 4$ .改写系统(3.1)的第三和第四个方程为

$ \begin{eqnarray}\label{4.15} &&d_3u_3''(t)-cu_3'(t)+(\beta_{11}u^0_1-\theta_3)u_3(t) +\beta_{12}u_1^0u_4(t)\nonumber\\ &=&\beta_{11}u_1^0u_3(t)-\beta_{11}f_1(u_3(t))u_1(t) +\beta_{12}u_1^0u_4(t)-\beta_{12}f_2(u_4(t))u_1(t) \end{eqnarray} $ (4.15)

$ \begin{eqnarray}\label{4.16} &&d_4u_4''(t)-cu_4'(t)+(\beta_{22}u^0_2-\theta_4) u_4(t)+\beta_{21}u_2^0u_3(t)\nonumber\\ &=&\beta_{21}u_2^0u_3(t)-\beta_{21}f_1(u_3(t))u_2(t) +\beta_{22}u_2^0u_4(t)-\beta_{22}f_2(u_4(t))u_2(t). \end{eqnarray} $ (4.16)

对方程(4.15)和(4.16)使用双边Laplace变换, 可得

$ \begin{equation}\label{4.17} h_1(\lambda)L_3(\lambda)+\beta_{12}u_1^0L_4(\lambda)=J_1(\lambda) \end{equation} $ (4.17)

$ \begin{equation}\label{4.18} h_2(\lambda)L_4(\lambda)+\beta_{21}u_2^0L_3(\lambda)=J_2(\lambda), \end{equation} $ (4.18)

其中

$ J_1(\lambda)=\int_{-\infty}^{+\infty}{\rm e}^{-\lambda t}\big[\beta_{11}u_1^0u_3(t) -\beta_{11}f_1(u_3(t))u_1(t) +\beta_{12}u_1^0u_4(t)-\beta_{12}f_2(u_4(t))u_1(t)\big]{\rm d}t $

$ J_2(\lambda)=\int_{-\infty}^{+\infty}{\rm e}^{-\lambda t}\big[\beta_{21}u_2^0u_3(t) -\beta_{21}f_1(u_3(t))u_2(t) +\beta_{22}u_2^0u_4(t)-\beta_{22}f_2(u_4(t))u_2(t)\big]{\rm d}t. $

首先证明 $\lambda_i^*<+\infty, \ i=3, 4$ .事实上, 由(4.17)式, 可知

$ \Delta(\lambda):=\big(d_3\lambda^2-c\lambda-\theta_3\big)L_3(\lambda) +\int_{-\infty}^{+\infty}{\rm e}^{-\lambda t}\big[\beta_{11}f_1(u_3(t)) +\beta_{12}f_2(u_4(t))\big]u_1(t){\rm d}t=0. $

注意到 $L_3(\lambda)>0$

$ \int_{-\infty}^{+\infty}{\rm e}^{-\lambda t}[\beta_{11}f_1(u_3(t)) +\beta_{12}f_2(u_4(t))]u_1(t){\rm d}t>0, \ \lambda\in[0, \lambda_3^*], $

$\lambda_3^*=+\infty$ , 则 $\Delta (+\infty)=+\infty $ , 矛盾.于是, $\lambda_3^*<+\infty$ .同理可证 $\lambda_4^*<+\infty$ .

往下再证明 $\lambda_3^*=\lambda_4^*$ .事实上, 若 $\lambda_3^*<\lambda_4^*$ , 则

$ \lim\limits_{\lambda\to\lambda_3^*-0} L_3(\lambda)=+\infty, \quad \lim\limits_{\lambda\to \lambda_3^*-0} L_4(\lambda)={L_4}(\lambda_3^*)<+\infty. $

由于 $J_1(\lambda_3^*)<+\infty$ , 这与(4.17)式矛盾.于是 $\lambda_3^*\geq\lambda_4^*$ .同理可证 $\lambda_4^*\geq\lambda_3^*$ .因此 $\lambda^*:=\lambda_3^*=\lambda_4^*$ , 这证明了 $\lambda_3^*=\lambda_4^*<+\infty$ .进一步, 若 $h_1(\lambda^*)\geq0$ , 则

$ h_1(\lambda^*)L_3(\lambda^*)+\beta_{12}u_1^0L_4(\lambda^*)=+\infty >J_1(\lambda^*), $

这与(4.17)式矛盾.于是 $h_1(\lambda^*)<0$ .同理可证 $h_2(\lambda^*)<0$ .由引理3.1(3)知特征方程 $\mbox{det}A(\lambda)=0$ 只有一个正根 $\lambda_0$ 满足 $\lambda_0>\lambda_M^{\pm}$ .则 $\lambda^*<\lambda_0$ .因此, $\mbox{det}A(\lambda^*)<0$ .然而

$ \mbox{det}A(\lambda)L_3(\lambda)L_4(\lambda)= J_1(\lambda)J_2(\lambda) -\beta_{21}u_2^0L_3(\lambda)J_1(\lambda)-\beta_{12}u_1^0L_4(\lambda)J_2(\lambda), $

于是

$ \mbox{det}A(\lambda)=\frac{J_1(\lambda)J_2(\lambda)}{L_3(\lambda)L_4(\lambda)} -\beta_{21}u_2^0\frac{J_1(\lambda)}{L_4(\lambda)} -\beta_{12}u_1^0\frac{J_2(\lambda)}{L_3(\lambda)}. $

$\mbox{det}A(\lambda^*)=\lim\limits_{\lambda\to\lambda^*-0}\mbox{det}A(\lambda)=0$ , 这与 $\mbox{det}A(\lambda^*)<0$ 矛盾.证毕.

参考文献
[1] Rahman S M A, Zou X. Modelling the impact of vaccination on infectious diseases dynamics. J Biol Dyns, 2015, 9(suppl 1): 307–320.
[2] Bai Z, Wu S L. Traveling waves in a delayed SIR epidemic model with nonlinear incidence. Appl Math Comput, 2015, 263: 221–232.
[3] Berestycki H, Hamel F, Kiselev A, Ryzhik L. Quenching and propagation in KPP reaction-diffusion equations with a heat loss. Arch Ration Mech Anal, 2005, 178: 57–80. DOI:10.1007/s00205-005-0367-4
[4] Capasso V, Serio G. A generalization of the Kermack-Mckendrick deterministic epidemic model. Math Biosci, 1978, 42: 43–61. DOI:10.1016/0025-5564(78)90006-8
[5] Carr J, Chmaj A. Uniquence of the traveling waves for nonlocal monostable equations. Proc Amer Math Soc, 2004, 132: 2433–2439. DOI:10.1090/S0002-9939-04-07432-5
[6] Ducrot A, Langlais M, Magal P. Qualitative analysis and travelling wave solutions for the SI model with vertical transmission. Commun Pure Appl Anal, 2012, 11: 97–113.
[7] Fang J, Zhao X Q. Monotone wavefronts for partially degenerate reaction-diffusion systems. J Dynam Differential Equations, 2009, 21: 663–680. DOI:10.1007/s10884-009-9152-7
[8] Fu S C. Traveling waves for diffusive SIR model with delay. J Math Anal Appl, 2016, 435: 20–37. DOI:10.1016/j.jmaa.2015.09.069
[9] Hsu C H, Yang T S. Existence, uniqueness, monotonicity and asymptotic behaviour of travelling waves for epidemic models. Nonlinearity, 2013, 26: 121–139. DOI:10.1088/0951-7715/26/1/121
[10] Huang J, Zou X. Existence of traveling wavefronts of delayed reaction diffusion systems without monotonicity. Discrete Contin Dyn Sys, 2003, 9: 925–936. DOI:10.3934/dcdsa
[11] Li Y, Li W T, Lin G. Traveling waves of a delayed diffusive SIR epidemic model. Comm Pure Appl Anal, 2015, 14: 1001–1022. DOI:10.3934/cpaa
[12] Liang X, Zhao X Q. Asymptotic speeds of spread and traveling waves for monotone semiflows with applications. Comm Pure Appl Math, 2007, 60:1-40; Erratum:Comm Pure Appl Math, 2008, 61:137-138
[13] Ma S. Traveling wavefronts for delayed reaction-diffusion systems via a fixed point theorem. J Differential Equations, 2001, 171: 294–314. DOI:10.1006/jdeq.2000.3846
[14] Martin R, Smith H. Absract functional differential equations and reaction-diffusion systems. Trans Amer Math Soc, 1990, 321: 1–44.
[15] Murray J D. Mathematical Biology Ⅰ, Ⅱ. New York: Springer, 2002.
[16] Perko L. Differential Equations and Dynamical Systems. New York: Springer, 2001.
[17] Wang H. Spreading speeds and traveling waves for non-cooperative reaction-diffusion stsyems. J Nonlinear Sciences, 21:747-783 https://link.springer.com/article/10.1007/s00332-011-9099-9
[18] Wang H, Wang X S. Travelling waves phenomena in a Kermack-McKendrick SIR model. J Dynam Differential Equations, 2016, 28: 143–166. DOI:10.1007/s10884-015-9506-2
[19] Wang X S, Wang H, Wu J. Travelling waves of diffusive predator-prey systems:disease outbreak propagation. Discrete Contin Dyn Sys, 2012, 32: 3303–3324. DOI:10.3934/dcdsa
[20] Wang Z C, Li W T, Ruan S G. Traveling wave fronts in reaction-diffusion systems with spatio-temporal delays. J Differential Equations, 2006, 222: 185–232. DOI:10.1016/j.jde.2005.08.010
[21] Wang Z C, Wu J. Travelling waves of a diffusive Kermack-McKendrick epidemic model with non-local delayed transmission. Proc R Soc A, 2010, 466: 237–261. DOI:10.1098/rspa.2009.0377
[22] Weng P, Zhao X Q. Spreading speed and traveling waves for a multi-type SIS epidemics model. J Differential Equations, 2006, 229: 270–296. DOI:10.1016/j.jde.2006.01.020
[23] Widder D V. The Laplace Transform. Princeton: Princeton University Press, 1941.
[24] Wu J. Theory and Applications of Partial Functional Differential Equations. New York: Springer, 1996.
[25] Wu J H, Zou X F. Traveling wave fronts of reaction-diffusion systems with delays. J Dynam Differential Equations, 2001, 13:651-687; Erratum:J Dynam Differential Equations, 2008, 20:531-533
[26] Xu Z. Traveling waves in a Kermack-Mckendrick epidemic model with diffusion and latent period. Nonlinear Analysis, 2014, 111: 66–81. DOI:10.1016/j.na.2014.08.012
[27] Ye Q, Li Z. Introduction to Reaction-Diffusion Equations. Beijing: Science Press, 1990.
[28] Zhang T, Wang W. Existence of thaveling wave solutions for influenza model with treatment. J Math Anal Appl, 2014, 419: 469–495. DOI:10.1016/j.jmaa.2014.04.068