数学物理学报  2016, Vol. 36 Issue (6): 1157-1164   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
邓继勤
邓子明
分数阶微分方程非局部柯西问题解的存在和唯一性
邓继勤, 邓子明     
湘潭大学数学与计算科学学院 湖南湘潭 411105
摘要:该文利用不动点定理和一个新的方法,研究了分数阶微分方程非局部问题解的存在和唯一性,并且获得了两个新的结果.
关键词C0半群     分数阶发展方程     适度解     非局部柯西问题     拉普拉斯变换     概率密度    
Existence and Uniqueness of Solutions for Nonlocal Cauchy Problem for Fractional Evolution Equations
Deng Jiqin, Deng Ziming     
School of Mathematics and Computational Science, Xiangtan University, Hunan Xiangtan 411105
Abstract: In this paper, by using the fixed point theorem and a new method, we study the existence and uniqueness of solutions for nonlocal Cauchy problem for fractional evolution equations with Caputo fractional derivative, and obtain two new results. Finally, to illustrate the theoretical results obtained, we give an example.
Key words: C0-Semigroup     Fractional evolution equations     Mild solution     Nonlocal Cauchy problem     Laplace transform     Probability density    
1 引言

非局部条件是由Byszewski在文献[20-21]中介绍的, 这些条件比通常的初值条件能够更好的描叙一些物理现象.

分数阶微分方程是描叙科学和工程诸多领域中很多现象的有效工具.这些领域包括:电化学、控制论、电磁学、多孔介质等.因此它们受得了广泛的关注.例如, 在文献[11-13, 20-22, 14-22]中, 作者们分别研究了分数阶发展方程, 分数阶扩散方程非局部问题解的存在和唯一性.

在这篇文章中, 我们研究下列分数阶微分方程非局部柯西问题

$ \begin{equation} ^{c}D^{\alpha}x(t)={\cal A}x(t)+f(t, x(t)), \ \ t\in I:=(0, T], \ 0<\alpha<1, \ x(0)+g(x)=x_{0}, \end{equation} $ (1.1)

其中$^{c}D^{\alpha}$表示$\alpha$阶的Caputo导数, $({\bf X}, \ \|\cdot\|)$是一个Banach空间, $f:\ I\times{\bf X}\rightarrow{\bf X}$$g:\ C:=C(I, \ {\bf X})\rightarrow{\bf X}$将在后面给出, ${\cal A}$${\bf X}$上算子半群$\{U(t)\}_{t\ge 0}$的无限维生成子并且$x_{0}\in{\bf X}$.

在以往的关于问题(1.1)的存在、唯一性结果中, 下列条件中的一些常被用到:

$(A)$ 对于每一个$t>0$, $U(t)$是紧算子;

$(F_{1})$ 对于每一个$t\in I$, 函数$f(t, \ \cdot):\ {\bf X}\rightarrow{\bf X}$是连续的, 对于每一个$x\in C:=C(I, \ {\bf X})$, 函数$f(\cdot, \ x):\ I\rightarrow{\bf X}$是强可测的并且存在$d, \ l>0$, 使得对于每一个$x\in{\bf X}$,

$ \|f(t, \ x)\|\le l\|x\|+d, ~~ t\in I ; $

$(F_{2})$ 存在$l>0$, 使得对于任意$x, \ y\in{\bf X}$,

$ \|f(t, \ x)-f(t, \ y)\|\le l\|x-y\|, \ \ t\in I; $

$(G_{1})$ $g\in C(C, \ {\bf X})$并且存在常数$L, \ D>0$使得对于每一个$x\in C$,

$ \|g(x)\|\le L\|x\|_{0}+D, $

其中$\|x\|_{0}=\sup\limits_{t\in I}\|x(t)\|$;

$(G_{2})$ 存在常数$L>0$使得对于任意$x, \ y\in C$,

$ \|g(x)-g(y)\|\le L\|x-y\|_{0}. $

但是不论存在性还是存在, 唯一性都必须满足下列条件

$(H')$

$ ML+\frac{lMT^{\alpha}}{\Gamma(1+\alpha)}<1, $

其中$M$见本文第2节.

虽然在文献[13]中, 一些新条件被首次使用, 但如果条件$(H')$不满足, 这些新条件一般也不可能满足.问题是:条件$(H')$能被改进吗?回答是肯定的.

本文利用不动点定理和一个新的方法, 研究了分数阶微分方程非局部问题(1)的解的存在和唯一性, 并且获得了下列两个新的结果:定理3.1和3.2 (见第三节).在这些结果中, 我们改进了条件$(H')$ (见定理3.1中的条件$(H)$).相似的讨论参见文献[24-27].

这篇文章组织如下:在第二节中, 我们给出了一些记号, 定义和引理; 在第三节中, 我们给出并且证明了我们的主要结果:定理3.1和3.2;在第四节中, 我们给出了一个例子.

2 预备知识

$B({\bf X})$表示由XX的所有有界线性算子所组成的Banach空间, 其范数为

$ \|K\|_{B({\bf X})}=\sup\{|K(y)|:\ |y|=1\}, $

${\bf X}([a, \ b])$表示Banach空间$(C([a, \ b], \ {\bf X}), \ \|\cdot\|_{0})$, 其范数为

$ \|x\|_{0}=\max_{t\in[a, \ b]}\|x(t)\|, $

并且

$ M:=\sup_{t\in[0, \ \infty)}\|U(t)\|_{B({\bf X})}. $

定义2.1[9] 函数$x\in AC[0, \ \infty)$$\alpha$阶Riemann-Liouville分数阶积分定义为

$ I^{\alpha}x(t)=\int_{0}^{t}\frac{(t-s)^{\alpha-1}}{\Gamma(\alpha)}x(s){\rm d}s, \ \ t>0, \ \ 1<\alpha<1, $

只要上式右边在$[0, \ \infty)$上逐点有定义, 其中$\Gamma(\cdot)$$\Gamma$-函数.

定义2.2[9] 函数$x\in AC[0, \ \infty)$$\alpha$阶Riemann-Liouville分数阶导数定义为

$ D^{\alpha}x(t)=\frac{1}{\Gamma(1-\alpha)}\frac{\rm d}{{\rm d}t}\int_{0}^{t}(t-s)^{-\alpha}x(s){\rm d}s, \ \ t>0, \ \ 1 <\alpha <1. $

定义2.3[9] 函数$x\in AC[0, \ \infty)$$\alpha$阶Caputo分数阶导数定义为

$ ^{c}D^{\alpha}x(t)=D^{\alpha}[x(t)-x(0)], \ \ t>0, \ \ 1>\alpha>0. $

定义2.4[12] 称$x\in{\bf X}(I)$是问题(1.1)的适度解, 如果

$ \begin{eqnarray*} x(t)&=&\int_{0}^{\infty}h_{\alpha}(\theta)U(t^{\alpha}\theta)(x_{0} -g(x)){\rm d}\theta\\ &&+\alpha \int_{0}^{t}\int_{0}^{\infty}\theta(t-s)^{\alpha-1}h_{\alpha}(\theta)U((t-s)^{\alpha}\theta)f(s, \ x(s)){\rm d}\theta{\rm d}s, \ \ t\in I, \end{eqnarray*} $

其中$h_{\alpha}(\theta)=\frac{1}{\alpha}\theta^{-1-\frac{1}{\alpha}} \psi_{\alpha}(\theta^{-\frac{1}{\alpha}})$是定义在$(0, \infty)$上的概率密度函数并且对于每一个$\theta\in(0, \infty)$, $\psi_{\alpha}(\theta)=\frac{1}{\pi}\sum\limits_{n=1}^{\infty}(-1)^{n-1}\theta^{-\alpha n-1}\frac{\Gamma(n\alpha+1)}{n!}\sin(n\pi\alpha)$是单侧稳定的概率密度函数.

由文献[12]易知, 对于任意$x\in C$, $k:\ [0, \ \infty)\times{\bf X}\rightarrow{\bf X}$$\phi:\ C([0, \ \infty), \ {\bf X})\rightarrow{\bf X}$

$ \begin{equation} \bigg\|\int_{0}^{\infty}h_{\alpha}(\theta)U(t^{\alpha}\theta)\phi(x){\rm d}\theta\bigg\| \le M\|\phi(x)\|, \ \ t\in I, \end{equation} $ (2.1)

$ \begin{eqnarray} &&\alpha \bigg\|\int_{0}^{t}\int_{0}^{\infty}\theta(t-s)^{\alpha-1}h_{\alpha}(\theta)U((t-s)^{\alpha}\theta)k(s, \ x(s)){\rm d}\theta{\rm d}s\bigg\|\\ &\le& \frac{M}{\Gamma(\alpha)}\int_{0}^{t}(t-s)^{\alpha-1}\|k(s, \ x(s))\|{\rm d}s, \ \ t\in I. \end{eqnarray} $ (2.2)

引理2.1  (Schauder不动点定理) 如果$B$是Banach空间${\bf X}$中的一个闭的有界凸子集, 算子$N:\ B\rightarrow B$是全连续的, 则$N$$B$中有一个不动点.

3 主要结果及证明

定理3.1 假若$(A)$, $(F_{1})$, $(G_{1})$和下列条件

$ (H)\;\;\;\;\;\;\;\;\;ML <1 $

成立, 则问题(1.1)至少存在一个解.

 让$\xi>0$, $k_{0}, \ \ i_{0}\in{\bf N}$是常数.它们满足

$ \begin{equation} \Big(\frac{i_{0}-2}{2}\Big)^{\alpha-1}\le\frac{1-ML}{2}, \end{equation} $ (3.1)
$ \begin{equation}(1+2\xi^{-\alpha}) \Big(ML+\frac{i_{0}Ml}{\xi^{\alpha}\Gamma(\alpha)}\Big)<\frac{1+5ML}{6}, \end{equation} $ (3.2)
$ \begin{equation}T+\frac{1}{\xi+k_{0}+2}>\sum_{i=1}^{k_{0}+1}\frac{1}{\xi+i}\ge T\end{equation} $ (3.3)

$ \begin{equation}k_{0}=[\xi {\rm e}^{T}-\xi+1]>i_{0}.\end{equation} $ (3.4)

定义${\bf X}([\delta_{0}, \ \delta_{i}])\ (i\in\{1, \ 2, \cdots, k_{0}\})$中的一个闭的有界凸子集$B_{i}$和算子$N_{i}$如下

$ \begin{eqnarray} B_{i}&=&\{x\in {\bf X}((\delta_{i-1}, \ \delta_{i}]\ \ \mbox{或}\ \ (\delta_{k_{0}}, \ T]):\ x(t)=x_{i-1}(t), \ \ t\in[0, \ \delta_{i-1}], \\ &&\ \|x(t)\|\le R {\rm e}^{(t-\delta_{i})^{\alpha}}, \ \ t\in(\delta_{i-1}, \ \delta_{i}]\ \ \mbox{或}\ \ (\delta_{k_{0}}, \ T]\}, \end{eqnarray} $ (3.5)

其中$i\in\{1, \cdots, k_{0}\}, \ x_{0}(0)=x_{0}-g(x)$, 并且, 对$t\in(0, \ \delta_{i}]\ (i\in\{1, \cdots, k_{0}\})$$(0, \ T]\ (i=k_{0}+1)$, 有

$ \begin{equation}R=\frac{3M}{1-ML} \Big(\|x_{0}\|+D+\frac{dT^{\alpha}} {\Gamma(1+\alpha)}\Big), \end{equation} $ (3.6)
$ \begin{eqnarray} N_{i}x(t)&=&\int_{0}^{\infty}h_{\alpha}(\theta)U(t^{\alpha}\theta) \bigg(x_{0} -\frac{1+{\rm sgn}(t-\delta_{i-1})}{2}g(x)-\frac{1+{\rm sgn}(\delta_{i-1}-t)}{2}g(x_{i-1}) \bigg){\rm d}\theta \\ &&+\alpha \int_{0}^{t}\int_{0}^{\infty}\theta(t-s)^{\alpha-1}h_{\alpha}(\theta)U((t-s)^{\alpha}\theta)f(s, \ x(s)){\rm d}\theta{\rm d}s. \end{eqnarray} $ (3.7)

下面我们将把证明划分成三步来进行.

第一步.分为两小步.

第1.1步 由有(2.1), (2.2), (3.1)-(3.7)式和条件$(F_{1})$, $(G_{1})$$H$, 易知, 对于每一个$x\in B_{i}\ (i\in\{1, \ 2, \cdots, i_{0}\})$,

$ \begin{eqnarray} \|N_{i}x(t)\|&=&\bigg\|\int_{0}^{\infty}h_{\alpha}(\theta)U(t^{\alpha}\theta) \bigg(x_{0} -\frac{1+{\rm sgn}(t-\delta_{i-1})}{2}g(x)-\frac{1+ {\rm sgn}(\delta_{i-1}-t)}{2}g(x_{i-1})\bigg){\rm d}\theta \\ &&+\alpha \int_{0}^{t}\int_{0}^{\infty}\theta(t-s)^{\alpha-1}h_{\alpha}(\theta)U((t-s)^{\alpha}\theta)f(s, \ x(s)){\rm d}\theta{\rm d}s\bigg\| \\ &\le& M\|x_{0}\|+MLR{\rm e}^{\xi^{-\alpha}} +MD+\frac{M}{\Gamma(\alpha)} \int_{0}^{t}(t-s)^{\alpha-1}\|f(s, \ x(s))\|{\rm d}s\\ &\le& M[\|x_{0}\|+LR{\rm e}^{\xi^{-\alpha}} +D+\int_{0}^{t}\frac{(t-s)^{\alpha-1}}{\Gamma(\alpha)} (l\|x(s)\|+d){\rm d}s] \\ &\le& M[\|x_{0}\| +LR(1+2\xi^{-\alpha})+D+\frac{dT^{\alpha}}{\Gamma(\alpha+1)} +\frac{lR}{\Gamma(\alpha)}\int_{0}^{t}(t-s)^{\alpha-1} \|x(s)\|{\rm d}s] \\ &\le&\frac{1-ML}{3}R +MLR(1+2\xi^{-\alpha})+ \frac{MlR}{\Gamma(\alpha)} \bigg(\int_{\delta_{i-1}}^{t}(t-s)^{\alpha-1}{\rm e}^{(s-\delta_{i-1})^{\alpha}} {\rm d}s\\ &&+\sum_{h=1}^{i-1}\int_{\delta_{h-1}}^{\delta_{h}}(t-s)^{\alpha-1}{\rm e}^{(s-\delta_{h-1})^{\alpha}} {\rm d}s\bigg) \\ &\le&\frac{1-ML}{3}R \\ && +(1+2\xi^{-\alpha})\bigg[MLR+ \frac{MlR}{\Gamma(\alpha)}\bigg(\int_{\delta_{i-1}}^{t}(t-s)^{\alpha-1} {\rm d}s+\sum_{h=1}^{i-1}\int_{\delta_{h-1}}^{\delta_{h}}(t-s)^{\alpha-1} {\rm d}s\bigg)\bigg] \\ &\le& \frac{1-ML}{3}R +(1+2\xi^{-\alpha})\Big(ML+\frac{iMl}{\xi^{\alpha}\Gamma(\alpha)}\Big)R \\ & \le& \frac{1+ML}{2}R{\rm e}^{(t-\delta_{i-1})^{\alpha}}, \end{eqnarray} $ (3.8)

其中$t\in(\delta_{i-1}, \ \delta_{i}]\ (i\in\{1, \ 2, \cdots, i_{0}\})$, 并且$\sum\limits_{h=1}^{0}\int_{\delta_{h-1}}^{\delta_{h}}(t-s)^{\alpha-1} {\rm e}^{(s-\delta_{h-1})^{\alpha}} {\rm d}s=0$.结合(3.5)-(3.8)式, 易知$N_{i}: \ {\bf B}_{i}\rightarrow {\bf B}_{i}\ (i\in\{1, \ 2, \cdots, i_{0}\})$.

综合上面的讨论, 以及后面第二、第三步, 可知$N_{i}:\ B_{i}\rightarrow B_{i}$是一个全连续算子.从而, 由Schauder不动点定理, 可知:在${\bf B}_{i}\ (i\in\{1, \ 2, \cdots, i_{0}\})$中, (3.7)式至少有一个不动点$x_{i}$满足$\|x_{i}\|_{0}\le\frac{1+ML}{2} R{\rm e}^{(t-\delta_{i-1})^{\alpha}}$.

第1.2步 由(2.1), (2.2), (3.1)-(3.7)式, 条件$(F_{1})$, $(G_{1})$, $H$$\ln(1+t)=\sum\limits_{h=1}^{\infty}\frac{(-1)^{h-1}}{h}t^{h}$ $ (0<t<1)$易知, 对于每一个$x\in B_{i}\ (i\in\{i_{0}+1, \ i_{0}+2, \cdots, k_{0}\})$,

$ \begin{eqnarray}\|N_{i}x(t)\| &\le&\frac{1-ML}{3}R +MLR(1+2\xi^{-\alpha})+ \frac{MlR}{\Gamma(\alpha)} \bigg(\int_{\delta_{i-1}}^{t}(t-s)^{\alpha-1}{\rm e}^{(s-\delta_{i-1})^{\alpha}} {\rm d}s \\ &&+\sum_{1\le h\le i-2, \ h\neq i-i_{0}}\int_{\delta_{h-1}}^{\delta_{h}}(t-s)^{\alpha-1}{\rm e}^{(s-\delta_{h-1})^{\alpha}} {\rm d}s\bigg) \\ && +\frac{(1+L)lR}{2\Gamma(\alpha)}\int_{\delta_{i-2}}^{\delta_{i-1}}(t-s)^{\alpha-1}{\rm e}^{(s-\delta_{i-2})^{\alpha}} \\ && + \frac{MlR}{\Gamma(\alpha)}\int_{\delta_{i-i_{0}-1}}^{\delta_{i-i_{0}}}(t-s)^{\alpha-1}{\rm e}^{(s-\delta_{i-i_{0}-1})^{\alpha}} {\rm d}s \\ &\le&\bigg(\frac{1+ML}{2}-\frac{(1+2\xi^{-\alpha})(1-ML)Ml}{2\xi^{\alpha}\Gamma(\alpha)}\bigg) R{\rm e}^{(t-\delta_{i-1})^{\alpha}} \\ &&+ \frac{(1+2\xi^{-\alpha})MlR}{\Gamma(\alpha)}\int_{\delta_{i-i_{0}-1}}^{\delta_{i-i_{0}}} \bigg(\sum_{h=1}^{i_{0}}\frac{1}{\xi+i-i_{0}-1+h}-s\bigg)^{\alpha-1} {\rm d}s\\ &\le& \bigg(\frac{1+ML}{2}-\frac{(1+2\xi^{-\alpha})(1-ML)Ml}{2\xi^{\alpha}\Gamma(\alpha)}\bigg) R{\rm e}^{(t-\delta_{i-1})^{\alpha}} \\ &&+ \frac{(1+2\xi^{-\alpha})MlR}{\Gamma(\alpha)}\int_{\delta_{i-i_{0}-1}}^{\delta_{i-i_{0}}} \Big(\ln\frac{\xi+i}{\xi+i-i_{0}}-s\Big)^{\alpha-1} {\rm d}s\\ &\le&\bigg(\frac{1+ML}{2}-\frac{(1+2\xi^{-\alpha})(1-ML)Ml}{2\xi^{\alpha}\Gamma(\alpha)}\bigg)R{\rm e}^{(t-\delta_{i-1})^{\alpha}} \\ &&+ \frac{(1+2\xi^{-\alpha})MlR}{\Gamma(\alpha)}\int_{\delta_{i-i_{0}-1}}^{\delta_{i-i_{0}}} \Big(\frac{i_{0}}{2(\xi+i-i_{0})}-s\Big)^{\alpha-1} {\rm d}s \\ &\le& \bigg(\frac{1+ML}{2}-\frac{(1+2\xi^{-\alpha})(1-ML)Ml}{2\xi^{\alpha}\Gamma(\alpha)}\bigg) R{\rm e}^{(t-\delta_{i-1})^{\alpha}} \\ &&+ \frac{(1+2\xi^{-\alpha})MlR}{\Gamma(\alpha)}\int_{\delta_{i-i_{0}-1}}^{\delta_{i-i_{0}}} \Big(\frac{i_{0}-2}{2}\Big)^{\alpha-1}s^{\alpha-1} {\rm d}s\\ & \le&\frac{1+ML}{2}R{\rm e}^{(t-\delta_{i-1})^{\alpha}}, \ \ t\in(\delta_{i-1}, \ \delta_{i}]\ (i\in\{i_{0}+1, \ i_{0}+2, \cdots, k_{0}\}). \end{eqnarray} $ (3.9)

结合(3.5)-(3.7)式和(3.9)式, 易知$N_{i}: \ {\bf B}_{i}\rightarrow {\bf B}_{i}\ (i\in\{i_{0}+1, \ i_{0}+2, \cdots, k_{0}\}). $

综合上面的讨论, 以及后面第二、第三步和Arzela-Ascoli定理, 可知$N_{i}:\ B_{i}\rightarrow B_{i}$是一个全连续算子.从而, 由Schauder不动点定理, 可知:在${\bf B}_{i}\ (i\in\{i_{0}+1, \ i_{0}+2, \cdots, k_{0}\})$中, (3.7)式至少有一个不动点$x_{i}$满足$\|x_{i}\|_{0}\le\frac{1+ML}{2} R{\rm e}^{(t-\delta_{i-1})^{\alpha}}$.

综合上面第1.1, 1.2步的讨论, 定义2.4和(3.7)式, 可知(1.1)在${\bf X}(I)$中至少有一个解.

第二步 由$(F_{1})$$(G_{1})$, 易知:映射$N$是连续的.

第三步 要证$N$是一致连续的.

$t_{1}, \ t_{2}\in I$, $t_{1}<t_{2}$并且$x\in B$.则, 由$(F_{1})$和(3.7)式, 我们得到

$ \begin{eqnarray*} \|Nx(t_{2})-Nx(t_{1})\|&\le& M \bigg\|\int_{0}^{t_{2}}\frac{(t_{2}-s)^{\alpha-1}} {\Gamma(\alpha)}f(s, \ x(s)){\rm d}s-\int_{0}^{t_{1}}\frac{(t_{1}-s)^{\alpha-1}}{\Gamma(\alpha)}f(s, \ x(s)){\rm d}s\bigg\| \\ &\le & M\bigg[\int_{t_{1}}^{t_{2}}\frac{(t_{2}-s)^{\alpha-1}}{\Gamma(\alpha)}\|f(s, \ x(s))\|{\rm d}s \\ &&+\int_{0}^{t_{1}}\frac{(t_{1}-s)^{\alpha-1}-(t_{2}-s)^{\alpha-1}}{\Gamma(\alpha)}\|f(s, \ x(s))\|{\rm d}s\bigg]\\ &\le& M(2lR+d)\bigg(\int_{t_{1}}^{t_{2}}\frac{(t_{2}-s)^{\alpha-1}}{\Gamma(\alpha)}{\rm d}s +\int_{0}^{t_{1}}\frac{(t_{1}-s)^{\alpha-1}-(t_{2}-s)^{\alpha-1}}{\Gamma(\alpha)}{\rm d}s\bigg) \\ &\le& \frac{2M(2lR+d)}{\Gamma(1+\alpha)}(t_{2}-t_{1})^{\alpha}, \end{eqnarray*} $

即, $N$是一致连续的.定理3.1证毕.

定理3.2 假若$(A)$, $(F_{1})$, $(G_{1})$$(H)$成立, 则问题(1.1)有一个唯一的解.

 首先, 由$(F_{2})$$(G_{2})$, 易知, 如果取$d=\|f(t, \ 0)\|_{0}, \ D=\|g(0)\|$, 则, 定理3.1中的条件被满足.因此, 在定理3.2的条件下, 问题(1.1)至少有一个解.

$\xi, \ i_{0}, \ k_{0}$$\delta_{i}$如在定理3.1的证明中.现在我们来证明唯一性.

假设问题(1.1)有两个解$x$$y$满足

$ \begin{equation}\|x-y\|_{0}>0\end{equation} $ (3.10)

而且存在$\zeta\in I$, 使得$\zeta\in(\delta_{n_{0}-1}, \ \delta_{n_{0}}]$

$ \begin{equation}\|x(\zeta)-y(\zeta)\|>R{\rm e}^{(t-\delta_{n_{0}-1})^{\alpha}}, \end{equation} $ (3.11)

其中$R=\frac{1+ML}{2}\|x-y\|_{0}$.

$ \begin{equation} \zeta_{i}=\min\{t\in(\delta_{i-1}, \delta_{i}]:\ \|x(t)-y(t)\|> R{\rm e}^{(t-\delta_{i-1})^{\alpha}}, t\in(\delta_{i-1}, \delta_{i}]\}, \ i\in\{1, 2, \cdots, k_{0}\}, \end{equation} $ (3.12)
$ \begin{equation} \zeta_{k_{0}+1}=\min\{t\in(\delta_{k_{0}}, T]:\ \|x(t)-y(t)\|> R{\rm e}^{(t-\delta_{k_{0}})^{\alpha}}, \ \ t\in(\delta_{k_{0}}, T]\}\end{equation} $ (3.13)

$ \begin{equation} \zeta_{0}=\min_{1\le i\le k_{0}+1}\zeta_{i}.\end{equation} $ (3.14)

如果$\phi$是一个空集, 我们设$\min\phi=T$.则, 由条件$(G_{2})$和(3.11)-(3.14)式, 有

$ \begin{equation}\|x(\zeta_{0})-y(\zeta_{0})\|=R{\rm e}^{(\zeta_{0}-\delta_{n_{1}-1})^{\alpha_{1}}}\end{equation} $ (3.15)

$ \begin{equation}0<\zeta_{0}<\zeta, \end{equation} $ (3.16)

其中$\zeta_{0}\in(\delta_{n_{1}-1}, \ \delta_{n_{1}-1}]$.因此, 由(3.16)式及相似与(3.8)或(3.9)式的计算, 有

$ \begin{equation} \|x(\zeta_{0})-y(\zeta_{0})\|\le\frac{3+ML}{4}R .\end{equation} $ (3.17)

此式与(3.15)式矛盾.因此$\|x-y\|_{0}=0$.定理3.2证毕.

4 例

例4.1 考虑下列问题[12]

$ \begin{gathered} \partial _t^\alpha u(t, \;z) = \partial _z^2u(t, \;z) + {\partial _z}G(t, \;u(t, \;z)), \;\;0 < t \leqslant d, \;\;0 \leqslant z \leqslant \pi, \\ u(t, \;0) = u(t, \;\pi ) = 0, \;\;0 \leqslant t \leqslant d, \\ u(0, \;z) + \sum\limits_{i = 0}^n {\int_0^\pi k } (z, \;\xi )u({t_i}, \;\xi ){\text{d}}\xi = {u_0}(z), \;\;0 \leqslant z \leqslant \pi, \\ \end{gathered} $ (4.1)

其中$\partial^{\alpha}_{t}$$\alpha$阶Caputo分数阶偏导数, $1>\alpha>0, \ d>0$, $G$是给定的函数, $n$是正整数, $0<t_{0}<t_{1}<\cdots<t_{n}<d$, $u_{0}(z)\in{\bf X}=L^{2}([0, \pi], \ {\Bbb R})$$k(z, \xi)\in L^{2}([0, \pi]\times[0, \pi], \ [0, \infty))$.

首先, 设

$ \begin{gathered} x(t)(z) = u(t, \;z), \;\;0 \leqslant t \leqslant d, \;\;0 \leqslant z \leqslant \pi, \\ f(t, \;x(t))(z) = {\partial _z}G(t, \;u(t, \;z)), \;\;0 \leqslant t \leqslant d, \;\;0 \leqslant z \leqslant \pi, \\ g(x)(z) = \sum\limits_{i = 0}^n {{K_g}} x({t_i})(z), \\ \end{gathered} $

其中$K_{g}v(z)=\int_{0}^{\pi}k(z, \ \xi)v(\xi){\rm d}\xi$ ($v\in{\bf X}$), $z\in[0, \ \pi]$.

如果让$ D({\cal A})=\{v(\cdot)\in{\bf X}:\ v\ \ v'$绝对连续, $v''\in{\bf X}, \ \ v(0)=v(\pi)=0\}$$ {\cal A}v=v'', $则(4.1)式等价于(1.1)式.其次, 取

$ \alpha=\frac{1}{2}, \ \ f(t, \ x(t))=\frac{100\sin x(t)}{t^{1/3}}, ~~ L=(n+1)\bigg[\int_{0}^{\pi}\int_{0}^{\pi}k^{2}(z, \ \xi){\rm d}\xi {\rm d}z\bigg]^{\frac{1}{2}}. $

则, 当$ML<1$并且$(H)$被满足时, 由定理3.2, 易知问题(4.1)在$I$上有一个唯一的解(注意$(H')$不被满足).因此我们的结果是新的.

参考文献
[1] Diethelm K, Freed A D. On the solution of nonlinear fractional order differential equations used in the modeling of viscoelasticity//Keil F, Mackens W, Voss H, et al. Scientific Computing in Chemical Engineering II-Computational Fluid Dynamics, Reaction Engineering and Molecular Properties. Heidelberg:Springer-Verlag, 1999:217-224
[2] Gaul L, Klein P, Kempfle S. Damping description involving fractional operators. Mech Syst Signal Process, 1991, 5: 81–88. DOI:10.1016/0888-3270(91)90016-X
[3] Glockle W G, Nonnenmacher T F. A fractional calculus approach of self-similar protein dynamics. Biophys J, 1995, 68: 46–53. DOI:10.1016/S0006-3495(95)80157-8
[4] Hilfer R. Applications of Fractional Calculus in Physics. Singapore: World Scientific, 2000.
[5] Mainardi F. Fractional calculus:Some basic problems in continuum and statistical mechanics//Carpinteri A, Mainardi F. Fractals and Fractional Calculus in Continuum Mechanics. New York:Springer-Verlag, 1997:291-348 http://www.scirp.org/reference/ReferencesPapers.aspx?ReferenceID=999090
[6] Metzler F, Schick W, Kilian H G, Nonnenmacher T F. Relaxation in filled polymers:A fractional calculus approach. J Chem Phys, 1995, 103: 7180–7186. DOI:10.1063/1.470346
[7] Kilbas A A, Srivastava Hari M, Juan Trujillo J. Theory and Applications of Fractional Differential Equations. Amsterdam: Elsevier Science, 2006.
[8] Miller K S, Ross B. An Introduction to the Fractional Calculus and Differential Equations. New York: John Wiley, 1993.
[9] Podlubny I. Fractional Differential Equations. San Diego: Academic Press, 1999.
[10] Lakshmikantham V, Leela S, Vasundhara Devi J. Theory of Fractional Dynamic Systems. Cambridge: Cambridge Academic Publishers, 2009.
[11] Chang Y K, Nieto J J. Existence of solutions for impulsive neutral integrodifferential inclusions with nonlocal initial conditions via fractional operators. Numerical Func Anal Optim, 2009, 30: 227–244. DOI:10.1080/01630560902841146
[12] Zhou Y, Jiao F. Nonlocal Cauchy problem for fractional evolution equations. Nonlinear Anal:RWA, 2010, 11(5): 4465–4475. DOI:10.1016/j.nonrwa.2010.05.029
[13] Zhou Y, Shen X H, Zhang L. Cauchy problem for fractional evolution equations with Caputo derivative. Eur Phys J Special Topics, 2013, 222: 1749–1765. DOI:10.1140/epjst/e2013-01961-5
[14] El-Sayed A M A. Fractional order diffusion-wave equations. Internat J Theoret Phys, 1996, 35: 311–322. DOI:10.1007/BF02083817
[15] Mainardi F, Paradisi P, Gorenflo R. Probability distributions generated by fractional diffusion equations//Kertesz J, Kondor I. Econophysics:An Emerging Science. Dordrecht:Kluwer, 2000 http://www.oalib.com/references/13877293
[16] Meerschaert M M, Benson D A, Scheffler H, Baeumer B. Stochastic solution of space-time fractional diffusion equations. Phys Rev E, 2012, 65: 1103–1106.
[17] Schneider W R, Wayes W. Fractional diffusion and wave equation. J Math Phys, 1989, 30: 134–144. DOI:10.1063/1.528578
[18] Pazy A. Semigroups of Linear Operators and Applications to Partial Differential Equations. New York:Springer, 1983 http://www.sciepub.com/reference/177277
[19] Baeumer B, Kurita S, Meerschaert M M. Inhomogeneous fractional diffusion equations. J Frac Appl Anal, 2005, 8: 375–397.
[20] Byszewski L. Theorems about existence and uniqueness of solutions of a semi-linear evolution nonlocal Cauchy problem. J Math Anal Appl, 1991, 162: 494–505. DOI:10.1016/0022-247X(91)90164-U
[21] Byszewski L, Lakshmikantham V. Theorem about the existence and uniqueness of a solution of a nonlocal abstract Cauchy problem in a Banach space. Appl Anal, 1991, 40: 11–19. DOI:10.1080/00036819008839989
[22] Fu X, Ezzinbi K. Existence of solutions for neutral differential evolution equations with nonlocal conditions. Nonlinear Anal, 2003, 54: 215–227. DOI:10.1016/S0362-546X(03)00047-6
[23] El-Borai M M. Some probability densities and fundamental solutions of fractional evolution equations. Chaos Solutions Fractals, 2004, 149: 823–831.
[24] Deng J Q, Ma L F. Existence and uniqueness of solutions of initial value problems for nonlinear fractional differential equations. Appl Math Letter, 2010, 23: 676–680. DOI:10.1016/j.aml.2010.02.007
[25] Deng J Q, Qu H L. New uniqueness results of solutions for fractional differential equations with infinite delay. Comput Math Appl, 2010, 60: 2253–2259. DOI:10.1016/j.camwa.2010.08.015
[26] Deng J Q, Deng Z M, Yang L X. An existence and uniqueness result for fractional order semilinear functional differential equations with nondense domain. Appl Math Letters, 2011, 24: 924–928. DOI:10.1016/j.aml.2010.12.051
[27] Deng J Q, Deng Z M. Existence of solutions of initial value problems for nonlinear fractional differential equations. Appl Math Lett, 2014, 32: 6–12. DOI:10.1016/j.aml.2014.02.001