Notes on nonlocal dispersal equations in a periodic habitat
Extracted source text ยท arXiv 1711.00978
\documentclass[11pt]{article}
\usepackage{amssymb}
\usepackage{amsmath}
\usepackage{amsthm}
\usepackage{color}
\usepackage{ulem}
\renewcommand \baselinestretch{1.5}
\numberwithin{equation}{section}
\newtheorem{theorem}{Theorem}[section]
\newtheorem{lemma}{Lemma}[section]
\newtheorem{proposition}{Proposition}[section]
\newtheorem{definition}{Definition}[section]
\newtheorem{example}{Example}[section]
\newtheorem{corollary}{Corollary}[section]
\newtheorem{remark}{Remark}[section]
\date{}
\title{\bf Notes on nonlocal dispersal equations in a periodic habitat}
\author{Jian Fang
\\Department of Mathematics\\
Harbin Institute of Technology\\
Harbin 150001, China\\
E-mail:\, jfang@hit.edu.cn
\and
Xiao-Qiang Zhao\\
Department of Mathematics and Statistics\\
Memorial University of Newfoundland\\
St. John's, NL A1C 5S7, Canada\\
E-mail:\, zhao@mun.ca}
\begin{document}
\maketitle
%\section{Introduction}
Assume that $\int_{\mathbb{R}}J(y)dy=1$. We consider the linear
equation
\begin{equation}\label{E1}
u_t=J*u-u, \quad t>0, \, x\in \mathbb{R}.
\end{equation}
Let $b>0$ be given, and $\mathcal{C}_b:=C(\mathbb{R},[0,b])$.
For any $\phi, \psi\in \mathcal{C}_b$, we define
$$
d(\phi,\psi)=\sum_{k=1}^{\infty}\frac 1{2^k}\max_{-k\leq x\leq k}|\phi(x)-\psi(x)|.
$$
Then $(\mathcal{C}_b, d)$ is a complete metric space, and the metric $d$
induces the compact open topology on $\mathcal{C}_b$. Let $T(t):\mathcal{C}_b\to
\mathcal{C}_b$ be the solution maps associated with \eqref{E1}.
\
\noindent {\bf Lemma 1.} For each $t>0$, the map $T(t):
\mathcal{C}_b\to \mathcal{C}_b$ is an $\alpha$-contraction,
where $\alpha$ is the Kuratowski measure of noncompactness in the metric
space $(\mathcal{C}_b, d)$.
\noindent {\bf Proof.} Letting $n=1$ in (3.2) and (3.5) of
\cite{WZ} with $P(x)=J(x)$, we see that
\begin{equation}
T(t)\phi(x)=e^{-t}\sum_{k=0}^{\infty}\frac{t^k}{k!}a_k(\phi)(x),
\quad \forall t\geq 0, \, x\in \mathbb{R},
\end{equation}
where
$$
a_0(\phi)(x)=\phi(x), \quad
a_k(\phi)(x)=\int_{\mathbb{R}}J(x-y)a_{k-1}(\phi)(y)dy,\, \forall
k\geq 1.
$$
By induction, we have $a_k(\phi)(x)\leq b, \, \forall x\in \mathbb{R}, \, \phi\in \mathcal{C}_b$.
Define
$$g(x)=\int_{\mathbb{R}}|J(x+z)-J(z)|dz.$$
Then we obtain
$$
|a_k(\phi)(x_1)-a_k(\phi)(x_2)|\leq bg(x_1-x_2), \forall x_1, x_2\in \mathbb{R}, \, \phi\in \mathcal{C}_b,
\, k\geq 1.
$$
Let
$$T_1(t)\phi(x)=e^{-t}\phi(x),\quad
T_2(t)\phi(x)=e^{-t}\sum_{k=1}^{\infty}\frac{t^k}{k!}a_k(\phi)(x).
$$
Clearly, $T(t)=T_1(t)+T_2(t)$. Since
$e^t=\sum_{k=0}^{\infty}\frac{t^k}{k!}$, we have
$$
|T_2(t)\phi(x_1)-T_2(t)\phi(x_2)|\leq e^{-t}
bg(x_1-x_2)\sum_{k=1}^{\infty}\frac{t^k}{k!}\\
\leq bg(x_1-x_2),
$$
for all $x_1, x_2\in \mathbb{R}, \, \phi\in \mathcal{C}_b$. By
Ascoli's theorem, it then easily follows that for each $t>0$, the
map $T_2(t)$ is compact with respect to the compact open topology.
Note that $T_1(t)$ is an $\alpha$-contraction with the contraction
coefficient being $e^{-t}$. Consequently, $T(t)$ is also an
$\alpha$-contraction with the contraction coefficient being
$e^{-t}$.
\
From the proof of Lemma 1, we have the following observation.
\
\noindent {\bf Lemma 2.} For any given nonempty and bounded interval $I:=[a,b]\subset\mathbb{R}$ and $t\geq 0$, there holds
$\alpha ((T(t)\mathcal{U})_I)\leq e^{-t}\alpha ((\mathcal{U})_I)$
for each set $\mathcal{U}\subset \mathcal{C}_b$,
where $\alpha$ is the Kuratowski measure of noncompactness in the Banach
space $C([a,b],\mathbb{R})$.
\
By the basic properties of the noncompactness measure, it is easy to prove
the following result (see, e.g., the proof of \cite[Lemma 5]{Banas1980}).
\
\noindent
{\bf Lemma 3.} Let $a<b$ be two real numbers and $X$ be a Banach space.
For a bounded set $B\subset C([a,b],X)$, define
$B(s):=\{g(s): g\in B\}, \, s\in [a, b]$, and $\int_a^b B(s){\rm d}s:=\{\int_a^b g(s){\rm d}s: g\in B\}$. If $B$ is equicontinuous on $[a,b]$,
then $\alpha\left(\int_a^b B(s){\rm d}s\right)\leq \int_a^b \alpha(B(s)){\rm d}s$.
\
Next we consider the nonlinear equation
\begin{equation}\label{nonL}
u_t=J*u-u +f(x,u) \quad t>0, \, x\in \mathbb{R}.
\end{equation}
Assume that $f(x,0)\equiv 0$ and $f(x+L,u)=f(x,u)$ for some $L>0$, system \eqref{nonL} has a positive $L$-periodic steady state $\beta(x)$, and there exists a real number $k_f>0$ such that
\begin{equation}
\label{Lip}
|f(x,u_1)-f(x,u_2)|\leq k_f|u_1-u_2|, \,
\forall x\in [0,L], u_1, u_2\in [0,\beta (x)].
\end{equation}
Let $Q(t)$ be the solution semiflow of
system \eqref{nonL} on $\mathcal{C}_{\beta}$. Then we have the
following result.
\
\noindent {\bf Theorem 1.} For any given nonempty interval $I=[a,b]\subset \mathbb{R}$ and $t\geq 0$, there holds
$\alpha ((Q(t)B)_I)\leq e^{(k_f-1)t}\alpha (B_I)$
for each set $B\subset \mathcal{C}_{\beta}$,
where $\alpha$ is the Kuratowski measure of noncompactness in the Banach
space $C([a,b],\mathbb{R})$.
\noindent{\bf Proof.} Define $\hat f (\phi)(x)=f(x,\phi(x)), \, \forall x\in \mathbb{R}$.
By the constant-variation formula, it follows that
\[
Q(t)\phi=T(t)\phi +\int_0^tT(t-s)\hat f(Q(s)\phi)ds, \, \, \forall t>0.
\]
Define $Q_t\phi=Q(t)\phi$ and let $B\subset \mathcal{C}_{\beta}$ be given.
In view of Lemmas 2 and 3, we then have
\begin{eqnarray*}
\alpha( (Q_tB)_I)&&\le \alpha ((T(t)B)_I)+\int_0^t\alpha((T(t-s)\hat f(Q_sB))_I)ds\\
&& \le e^{- t}\alpha(B_I)+\int_0^t e^{-(t-s)}\alpha((\hat f(Q_sB))_I)ds\\
&& \le e^{- t}\alpha(B_I)+\int_0^t e^{-(t-s)}k_f\alpha((Q_sB)_I)ds,
\end{eqnarray*}
and hence,
\[
e^{t} \alpha((Q_tB)_I)= \alpha(B_I)+\int_0^t k_f e^{ s}\alpha((Q_sB)_I)ds.
\]
By Grownwall's inequality, it follows that
\[
e^{t} \alpha((Q_tB)_I)\le \alpha(B_I)e^{\int_0^t k_f ds}=\alpha(B_I)e^{k_f t}.
\]
This implies that $\alpha(Q_t(B)_I)\le e^{(k_f-1)t}\alpha(B_I)$.
\
\noindent
{\bf Remark 1.} From the proof of Theorem 1, it is easy to see that the conclusion
of Theorem 1 also holds if the condition \eqref{Lip} is replaced by the
assumption that $\alpha((\hat f(B))_I)\leq k_f\alpha((B)_I)$ for any $B\subset \mathcal{C}_{\beta}$.
\
\noindent
{\bf Remark 2.} Let $Q_t$ be the solution maps of $u_t=D(J\ast u-u)+f(x,u)$.
By the proof of Theorem 1, we see that for each $t>0$, $Q_t$ is an $\alpha$-contraction relative to $I$ provided that $D> k_f$. It follows that
the large dispersal rate implies the $\alpha$-contraction property of solution
maps for this class of nonlinear nonlocal dispersal equations.
\
\noindent
{\bf Remark 3.} The spreading speed and traveling waves were
obtained in \cite{ShenZhangJDE2010,ShenZhangCANA2012,CovPoincare2013,RSZDCDSA2015}
for nonlocal dispersal equation \eqref{nonL} under the condition that $\frac{f(x,u)}{u}$ is strictly
decreasing in $u>0$, which guarantees the linear determinacy of the spreading speed.
In the case where $k_f<1$, Theorem 1 implies that the solution
map $Q_t$ satisfies the $\alpha$-contraction assumption (E3) in \cite{LZ}
for each $t>0$. Thus, one can use \cite[Theorems 5.2 and 5.3]{LZ}
to establish the existence of the spreading speed and their coincidence
with the minimum wave speed for monotone spatially $L$-periodic traveling
waves of nonlocal equation \eqref{nonL} no matter whether
the spreading speed is linearly determinate.
\
\noindent
{\bf Remark 4.} It is easy to generalize Theorem 1 and Remark 2 to
nonlocal dispersal systems in a periodic habitat. By the results and arguments in \cite{LZ,FZ2014}, it then follows that the nonlocal dispersal
Lotka-Volterra competition system, as studied in \cite{KRS2015, BaoLiShen2016}, admits a spatially periodic traveling
wave with the wave speed being the spreading speed under the condition that the dispersal rates are large. However, it remains an
open problem whether this system admits such a critical wave without assuming the large dispersal rates.
\begin{thebibliography}{99}
\bibitem{Banas1980} J. Bana\'{s}, On measures of noncompactness in Banach spaces,
{\it Comment. Math. Univ. Carolin.}, 21(1980), 131--143.
\bibitem{BaoLiShen2016} X. Bao, W.-T. Li and W. Shen, Traveling wave solutions of Lotka-Volterra competition systems with nonlocal dispersal in periodic habitats, {\it Journal of Differential Equations}, 260(2016), 8590--8637.
\bibitem{CovPoincare2013} J. Coville, J. Davila and S. Martinez, Pulsating fronts for nonlocal dispersion and KPP nonlinearity, {\it Ann. Inst. H. Poincare Anal. Non Lineaire}, 30(2013), 179--223.
\bibitem{FZ2014} J. Fang and X.-Q. Zhao, Traveling waves for monotone semiflows with weak compactness,
{\it SIAM J. Math. Anal.}, 46(2014), 3678--3704.
\bibitem{KRS2015} L. Kong, N. Rawal and W. Shen, Spreading speeds and linear determinacy for two species competition systems with nonlocal dispersal in periodic habitats, {\it Math. Model. Nat. Phenom.},
10(2015), 113--141.
\bibitem{LZ}X. Liang and X.-Q. Zhao, Spreading speeds and
traveling waves for abstract monostable evolution systems, {\it J. Funct. Anal.}, (259)(2010), 857--903.
\bibitem{RSZDCDSA2015}N. Rawal, W. Shen and A. Zhang, Spreading speeds and traveling waves of nonlocal monostable equations in time
and space periodic habitats, {\it Discrete Contin. Dyn. Syst.}, 35(2015), 1609--1640.
\bibitem{ShenZhangJDE2010} W. Shen and A. Zhang, Spreading speeds for monostable equations with nonlocal dispersal
in space periodic habitats, {\it J. Differential Equations}, 249(2010), 747--795.
\bibitem{ShenZhangCANA2012} W. Shen and A. Zhang, Traveling wave solutions of spatially periodic nonlocal monostable
equations, {\it Comm. Appl. Nonlinear Anal.}, 19 (2012), 73--101.
\bibitem{WZ} P. Weng and X.-Q. Zhao, Spreading speed and traveling waves for a
multi-type SIS epidemic model, {\it J. Differential Equations}, 229(2006),
270--296.
\end{thebibliography}
\end{document}