数学杂志  2020, Vol. 40 Issue (5): 624-630   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
李宝麟
王转红
无限滞后脉冲测度微分方程解对参数的连续依赖性
李宝麟, 王转红    
西北师范大学数学与统计学院, 甘肃 兰州 730070
摘要:本文研究了解对参数连续依赖性的问题.利用Kurzweil-Stieltjes积分理论和正则函数的相关性质,获得了无限滞后脉冲测度微分方程解对参数的连续依赖性的结果.
关键词无限滞后脉冲测度微分方程    广义线性微分方程    参数连续依赖性    
CONTINUOUS DEPENDENCE OF SOLUTIOUS TO PARAMETER OF THE INFINITE DELAY IMPULSIVE MEASURE DIFFERENTIAL EQUATIONS
LI Bao-lin, WANG Zhuan-hong    
College of Mathematics and Statistics, Northwest Normal University, Lanzhou 730070, China
Abstract: In this paper, we study the problem of continuous dependence on parameters. By using the integration theory of Kurzweil-Stieltjes and the related properties of regular functions, we obtain the results of continuous dependence of solutions of differential equations with infinite delay impulse measures on parameters.
Keywords: impulsive measure differential equations with infinite delay     generalized linear differential equation     continuous dependence on parameter    
1 引言

Kurzweil[1]于1957年提出的广义常微分方程理论在处理常微分方程、脉冲微分方程、滞后型泛函微分方程及拓扑动力系统等问题时有重要作用, 已被许多作者进行深入广泛的研究, 并取得了一些新的成果[2-4].以下区别于文献[5], 没有利用常数变易公式而利用Kurzweil-Stieltjes积分理论和正则函数的性质, 讨论了无限滞后脉冲测度泛函微分方程

$ \left\{ \begin{array}{ll} Dy=\mathcal{G}(y_{s}, s)D\mu+f(t)D\mu, ~\\ \Delta y(t_{k})=I_{k}(y(t_{k})), ~~~k=1, 2, \cdots, n, \\ y_{t_{0}}=\varphi \end{array} \right. $ (1)

解对参数的连续依赖性, 所得结果是对文献[5-6]已有结果的推广.

方程(1)等价于积分方程

$ \left\{\begin{array}{ll} y(t)=\varphi+\displaystyle\int^{t}_{t_{0}}\mathcal{G}(y_{s}, s)\mathrm{d}\mu(s)+\int_{t_{0}}^{t}f(s)\mathrm{d}\mu(s)+\sum\limits_{k=1}^{n}I_{k}(y(t_{k})), \\ y_{t_{0}}=\varphi, \end{array} \right. $ (2)

其中$\mathcal{G}:O\times[t_{0}, t_{0}+\sigma]\rightarrow\mathrm{R}^{n}$关于第一个变量是线性的, $f:[t_{0}, t_{0}+\sigma]\rightarrow\mathrm{R}^{n}.$符号$y_{t}:(-\infty, 0]\rightarrow\mathrm{R}^{n}, y_{t}(\tau)=y(t+\phi), \phi\in(-\infty, 0]$表示滞后的长度, $\lambda_{0}\in R, \rho>0, \Lambda=\{\lambda_{0}\in R:\|\lambda-\lambda_{0}\|<\rho\}.$为了证明主要的结果, 先引入相关的定义和引理及一些符号的说明.

2 预备知识

$L(\bf{R}^{n})$表示由所有$n\times n$阶实矩阵构成的集合, $J\subset R$是一个有限或无限的区间, $\|\cdot\|$表示$L(\bf{R}^{n})$上的算子范数.对区间$[a, b]$的任何精细分划P, 使得$a=s_{0}<s_{1}<\cdots<s_{i+1}=b$, 若$\mathrm{var}_{a}^{b}A=\sup\{\sum\limits_{i=1}^{n(P)}\|A(s_{i})-A(s_{i-1})\|_{X}\}<\infty$, 则函数A(t)在$[a, b]$上是有界变差的, $BV([a, b], L(\bf{R}^{n}))$表示所有有界变差函数构成的全体. $G^{\ast}([a, b], L(\bf{R}^{n}))$表示所有正则函数$A(t):[a, b]\rightarrow L(\bf{R}^{n})$的全体.

$O\subset G^{\ast}((-\infty, t_{0}+\sigma], \mathrm{R}^n)$是一个开集且具有以下性质(延拓性质):如果$y=y(t), $ $t\in[t_0, t_{0}+\sigma]$$\bar{t}\in[t_0, t_{0}+\sigma], $那么$\bar{y}(t)\in O$

$ {\bar{y}(t)}=\left\{ \begin{array}{ll} y(t), ~~~~t \in [~t_{0}, \bar{t}~], \\ y(\bar{t}+), ~~~~~t\in (~\bar{t}, t_{0}+\sigma]. \end{array} \right. $

显然, 当$y\in O\subset G^{\ast}((-\infty, t_{0}+\sigma], \mathrm{R}^n)$时, 对任意的$t\in[t_0, t_{0}+\sigma]$都有$x_{t}\in G^{\ast}([-r, 0], \mathrm{R}^n)$.

定义2.1$G:\Omega \rightarrow \bf{R}^{n}$, $\Omega\subseteq O\times[t_0, t_0+\sigma].$称函数$x:[\alpha, \beta]\rightarrow\bf{R}^{n}$为Kurzweil广义常微分方程

$ \frac{\mathrm{d}x}{\mathrm{d}\tau}=DG(x, t) $ (3)

在区间$[\alpha, \beta]$上的一个解是指对所有的$t\in[\alpha, \beta], (x(t), t)\in G, $

$ x(s_2)-x(s_1)=\int_{s_1}^{s_2}DG(x(\tau), t), ~~s_1, s_2\in[\alpha, \beta] $ (4)

成立, 其中右端积分是函数$U(\tau, t)=G(x(\tau), t)$$[s_1, s_2]$上的Kurzweil积分.当(3)式中的$G(x, t)=A(t)x+g(t)$时, 称为Kurzweil广义线性常微分

$ \frac{\mathrm{d}x}{\mathrm{d}\tau}=D[A(t)x+g(t)]. $ (5)

用更为传统的符号$\displaystyle\int_{a}^{b}\mathrm{d}[A(\tau)]x(\tau)$来代替$\displaystyle\int_{a}^{b}D[A(t)x(\tau)], $其中$A:[a, b]\rightarrow L(\bf{R}^{n})$$[a, b]$上的$n\times n$矩阵值函数, $g:[a, b]\rightarrow\bf{R}^{n}$上的$n$维列向量函数, 且$A, g$$[a, b]$上是有界变差的.将(5)式简记为

$ \mathrm{d}x=\mathrm{d}[A]x+\mathrm{d}g. $ (6)

引理2.2[2]$A\in BV([a, b], L(\mathrm{R}^{n})), ~g\in G^{\ast}([a, b], \mathrm{R}^{n})$, 称$x:[a, b]\rightarrow \mathrm{R}^{n}$为广义线性微分方程

$ \frac{\mathrm{d}x}{\mathrm{d}t}=\mathrm{d}[A(t)x(t)+g(t)] $ (7)

满足初始条件$x(t_{0})=x_{0}\in X$在区间$[a, b]$上的解, 是指对任意的$t\in[a, b], $

$ x(t)=x_{0}+\int_{a}^{t}\mathrm{d}[A(t)]x+g(t)-g(t_{0}) $ (8)

$x\in G^{\ast}([a, b], \mathrm{R}^{n}).$

引理2.3[5]$g, g_{n}\in G^{\ast}([a, b], \mathrm{R}^{n}), A, A_{k}\in BV([a, b], L(\mathrm{R}^{n})), $对每个$k\in N$

$ \lim\limits_{k\rightarrow\infty}\|g_{k}-g\|_{\infty}=0, $ (9)
$ \begin{array}{l} \alpha^{\ast}=\sup\{\mathrm{var}_{a}^{b}A_{k}:k\in N\}<\infty, \\ \lim\limits_{k\rightarrow\infty}\|A_{k}-A\|_{\infty}=0, \end{array} $ (10)

则有

$ \lim\limits_{k\rightarrow\infty}\Big\|\int_{a}^{t}\mathrm{d}[A_{k}]g_{k}-\int_{a}^{t}\mathrm{d}[A]g\Big\|_{\infty}=0. $ (11)

因为正则函数是有限阶梯函数的一致极限, 所以对任意的$\varepsilon>0, $存在有限阶梯函数$\tilde{g}:[a, b]\rightarrow\mathrm{R}^{n}$$n_{0}\in N, $使得$\parallel g-\tilde{g}\parallel_{\infty}<\varepsilon$及对任意的$k\geq n_{0}, $$\parallel g_{k}-\tilde{g}\parallel_{\infty}<\varepsilon, \\ \parallel A_{k}-A\parallel_{\infty}<\varepsilon.$由文献[3, 命题10]及文献[7, 引理3.1], 对$t\in[a, b], k\geq n_{0}, $则有

$ \begin{align*} \Big\|\int_{a}^{t}\mathrm{d}[A_{k}]g_{k}-\int_{a}^{t}\mathrm{d}[A]g\Big\|_{\infty}&=\Big\|\int_{a}^{t}\mathrm{d}[A_{k}](g _{k}-\tilde{g})+\int_{a}^{t}\mathrm{d}[A_{k}-A]\tilde{g}+\int_{a}^{t}\mathrm{d}[A](\tilde{g}-g)\Big\|_{\infty}\\ &\leq\Big\|\int_{a}^{t}\mathrm{d}[A_{k}](g _{k}-\tilde{g})\Big\|_{\infty}+\Big\|\int_{a}^{t}\mathrm{d}[A_{k}-A]\tilde{g}\Big\|_{\infty} +\Big\|\int_{a}^{t}\mathrm{d}[A](\tilde{g}-g)\Big\|_{\infty}\\ &\leq(2\alpha^{\ast}+2\parallel\tilde{g}\parallel_{BV}+\mathrm{var}_{a}^{b}A)\cdot\varepsilon \end{align*} $

$\varepsilon$的任意性可得(11)式成立.

3 解对参数的连续依赖性

下面主要介绍无限滞后脉冲测度微分方程

$ \left\{\begin{array}{ll} y_{k}(t)=y_{k}(t_{0})+\displaystyle\int^{t}_{t_{0}}\mathcal{G}_{k}((y_{k})_{s}, s)\mathrm{d}\mu_{k}(s)+\int^{t}_{t_{0}}f_{k}(s)\mathrm{d}\mu_{k}(s)+\sum\limits_{k=1}^{n}I_{k}^{p}(y(t_{k})), \\ (y_{k})_{t_{0}}=\varphi_{k} \end{array} \right. $ (12)

解对参数的连续依赖性.其中$\mu_{k}:[t_{0}, t_{0}+\sigma]\rightarrow\mathrm{R}$是不减的, $\mathcal{G}_{k}:O\times[t_{0}, t_{0}+\sigma]\rightarrow\mathrm{R}^{n}$关于第一个变量是线性的, $f_{k}:[t_{0}, t_{0}+\sigma]\rightarrow\mathrm{R}^{n}$$\varphi_{k}\in O.$$p=1, 2, \cdots$$y\in S$, 定义如下函数

$ \text{(i)}~~ {F_{k}(y, t)}(v)=\left\{ \begin{array}{ll} 0, ~~~~~~-\infty<v\leq t_0, \\ \displaystyle\int_{t_0}^{v}\mathcal{G}_{k}(y_{s}, s)\mathrm{d}\mu_{k}(s), ~~~t_0\leq v\leq t \leq t_0+\sigma, \\ \displaystyle\int_{t_0}^{t}\mathcal{G}_{k}(y_{s}, s)\mathrm{d}\mu_{k}(s), ~~~t_0\leq t\leq v\leq t_0+\sigma, \end{array} \right. $ (13)
$ \text{(ii)}~~ {(g_{k}(t))}(v)=\left\{ \begin{array}{ll} 0, ~~~~~~-\infty< v\leq t_0, \\ \displaystyle\int_{t_0}^{v}f_{k}(s)\mathrm{d}\mu_{k}(s), ~~~t_0\leq v\leq t \leq t_0+\sigma, \\ \displaystyle\int_{t_0}^{t}f_{k}(s)\mathrm{d}\mu_{k}(s), ~~~t_0\leq t\leq v\leq t_0+\sigma. \end{array}~~~~~~~ \right. $ (14)

给定一个$t_{k}^{c}\in[t_{0}, \infty)$, 定义

$ ~~~{\int_{k}^{c}(t)}=\left\{ \begin{array}{ll} 0, ~~~t_{0}\leq t\leq t_{k}^{c}, \\ 1, ~~~~t_{k}^{c}<t. \end{array} \right. $

则方程(2)中的脉冲项可定义为

$ \text{(iii)}~~ {J_{k}(y, t)}(v)=\sum\limits_{k=1}^{n}\int_{k}^{c}(t)\int_{k}^{c}(v)I_{k}^{p}(y(t_{k})). $ (15)

方程(12)右端的积分是关于不减函数$\mu_{k}:[t_{0}, t_{0}+\sigma]\rightarrow \mathrm{R}$的Kurzweil-Stieltjes积分, 而且$\mathcal{G}_{k}, f_{k}, I_{k}^{p}:\mathrm{R}^{n}\rightarrow\mathrm{R}^{n}$需满足以下条件

$(\mathrm{H_{1}})$对任意的$y\in O, ~t\in[t_{0}, t_{0}+\sigma]$, 使得$\displaystyle\int^{t}_{t_{0}}\mathcal{G}(y_{s}, s)\mathrm{d}\mu_{k}(s)$存在, $k\in N$;

$(\mathrm{H_{2}})$对任意的$y, z\in O$, $s_{1}, s_{2}\in[t_{0}, t_{0}+\sigma], $存在一个函数$M_{k}:[t_{0}, t_{0}+\sigma]\rightarrow\mathrm{R}^{+}$关于$\mu_{k}$是Kurzweil-Stieltjes积分, 使得

$ \Big|\Big|\int^{s_{2}}_{s_{1}}(\mathcal{G}_{k}(y_{s}, s)-\mathcal{G}_{k}(z_{s}, s))\mathrm{d}\mu_{k}(s)\Big|\Big| \leq \int^{s_{2}}_{s_{1}}M_{k}(t)\|y_{s}-z_{s}\|_{\infty}\mathrm{d}\mu_{k}(s); $

$(\mathrm{H_{3}})$对任意的$k\in N$, 积分$\displaystyle\int^{s_{2}}_{s_{1}}f_{k}(s)d\mu_{k}(s)$存在;

$(\mathrm{H_{4}})$对任意的$k\in N, $存在一个函数$N_{k}:[t_{0}, t_{0}+\sigma]\rightarrow\mathrm{R}^{+}$关于$\mu_{k}$是Kurzweil-Stieltjes积分, 使得

$ \Big|\Big| \int^{s_{2}}_{s_{1}}f_{k}(s)\mathrm{d}\mu_{k}(s)\Big|\Big|\leq \int^{s_{2}}_{s_{1}}N_{k}(s)\mathrm{d}\mu_{k}(s); $

$(\mathrm{H_{5}})$对任意的$k\in\{1, 2, \cdots, m\}, p=0, 1, \cdots, x\in\mathrm{R}^n, Q_{1}>0, $

$ \|I_{k}^{p}(x)\|\leq Q_{1}; $

$(\mathrm{H_{6}})$对任意的$k\in\{1, 2, \cdots, m\}, x, y\in\mathrm{R}^n, Q_{2}>0, $

$ \|I_{k}^{p}(x)-I_{k}^{p}(y)\|\leq Q_{2}|x-y|. $

定理3.1 对任意的$k\in N, ~p=0, 1, \cdots, $定义$A_{k}(t)y=F_{k}(y, t)+J_{k}(y, t).$假设$\mu_{k}:[t_{0}, t_{0}+\sigma]\rightarrow \mathrm{R}$是不减的左连续函数, $\mathcal{G}_{k}, f_{k}, I_{k}^{p}$满足条件(H$_{1}$)-(H$_{6})$, 且满足以下条件

(1) 对任意的$y\in O, $ $\lim \limits_{k\rightarrow\infty}\sup\limits_{t\in[t_{0}, t_{0}+\sigma]}\Big\|\displaystyle\int^{t}_{t_{0}}\mathcal{G}_{k}(y_{s}, s)\mathrm{d}\mu_{k}(s)-\int^{t}_{t_{0}}\mathcal{G}_{0}(z_{s}, s)\mathrm{d}\mu_{0}(s)\Big\|=0.$

(2) $\lim \limits_{k\rightarrow\infty}\sup\limits_{t\in[t_{0}, t_{0}+\sigma]}\Big\|\displaystyle\int^{t}_{t_{0}}f_{k}(s)\mathrm{d}\mu_{k}(s)-\int^{t}_{t_{0}}f_{0}(s)\mathrm{d}\mu_{0}(s)\Big\|=0.$

(3) 存在一个常数$\eta>0, $使得$\displaystyle\int^{t}_{t_{0}}M_{k}(t)\mathrm{d}\mu_{k}(s)\leq\eta, $$k\in N.$

(4) 对任意的$x\in\mathrm{R}^n, k=1, 2, \cdots, m, $使得$\lim\limits_{p\rightarrow\infty}I_{k}^{p}(x)=I_{k}^{0}(x).$

$y_{k}:[t_{0}, t_{0}+\sigma]\rightarrow\mathrm{R}^{n}, k=1, 2, \cdots$是方程(12)在$[t_{0}, t_{0}+\sigma]$上的一个解, 使得

$ \lim\limits_{k\rightarrow\infty}y_{k}(s)=y_{0}(s). $ (16)

$y_{0}:[t_{0}, t_{0}+\sigma]\rightarrow\mathrm{R}^{n}$是方程

$ \left\{ \begin{array}{ll} Dy=\mathcal{G}_{0}(y_{s}, s)D\mu(s)+f_{0}(t)D\mu(s), \\ \Delta y(t_{k})=I_{0}(y(t_{k})), ~~~k=1, 2, \cdots, m, \\ y_{t_{0}}=\varphi_{0} \end{array} \right. $ (17)

的一个解.

$A_{k}(t)y=F_{k}(y, t)+J_{k}(y, t), $其中$F_{k}, J_{k}, g_{k}$分别由(13), (15), (14)式给出, $k\in N.$由文献[8, 定理3.2]知, $A_{k}, g_{k}$是正则且左连续函数.对区间$[t_{0}, t_{0}+\sigma]$的任意分划$t_{0}=s_{0}<s_{1}<\cdots<s_{k}=t_{0}+\sigma, $$y\in G^{\ast}((-\infty, t_{0}], L(\mathrm{R}^{n})), $

$ \begin{align*} \|A_{k}(t)y\|_{\infty}&\leq\sup\limits_{v\in[t_{0}, t_{0}+\sigma]}|F_{k}(y, t)(v)|+\sup\limits_{v\in[t_{0}, t_{0}+\sigma]}|J_{k}(y, t)(v)|\\ &\leq\sup\limits_{v\in[t_{0}, t]}\bigg\|\int_{t_{0}}^{v}\mathcal{G}_{k}(y_{s}, s)\mathrm{d}\mu_{k}(s)\bigg\|+I_{k}^{p}(y(t_{k}))\int_{t_{k}}(v)\mathrm{var}_{t_{0}}^{t}(y)\\ &\leq\|y\|_{\infty}\int_{t_{0}}^{t}M_{k}(s)\mathrm{d}\mu_{k}(s)+Q_{1}\|y\|_{\infty}, \end{align*} $

则由函数$\mathcal{G}$的线性知, $A_{k}(t)\in L(\mathrm{R}^{n})$.即对$t_{0}\leq s_{i-1}<s_{i}\leq t_{0}+\sigma$$y\in G^{\ast}((-\infty, t_{0}+\sigma], L(\mathrm{R}^{n}))$, 由条件(H$_{2})$, 有

$ \begin{align*} \|[A_{k}(s_{2})-A_{k}(s_{1})]y\|_{\infty}&\leq\sup\limits_{v\in[t_{0}, t_{0}+\sigma]}\parallel([A_{k}(s_{2})-A_{k}(s_{1})]y)(v)\parallel\\ &\leq\sup\limits_{v\in[s_{i-1}, s_{i}]}\Big\|\int_{s_{i-1}}^{v}\mathcal{G}_{k}(y_{s}, s)\mathrm{d}\mu_{k}(s)\Big\|+I_{k}^{p}(y(t_{k}))\int_{t_{k}}(v)\mathrm{var}_{t_{0}}^{t}(y)\\ &\leq\|y\|_{\infty}\int_{s_{i-1}}^{s_{i}}M_{k}(s)\mathrm{d}\mu_{k}(s)+Q_{1}\|y\|_{\infty}. \end{align*} $

因此

$ \|A_{k}(s_{i})-A_{k}(s_{i-1})\|_{\infty}\leq\int_{s_{i-1}}^{s_{i}} M_{k}(s)\mathrm{d}\mu_{k}(s)+Q_{1}\leq\eta+Q_{1}. $

对任何$i=1, 2, \cdots, n$, 又有

$ \sum\limits_{i=1}^{n}\|A(s_{i})-A(s_{i-1})\|\leq\eta+Q_{1}. $

在上式右端对$[t_{0}, t_{0}+\sigma]$的所有分划取上确界得$\mathrm{var}_{a}^{b}A\leq\gamma, $其中$\gamma=\eta+Q_{1}.$$t\in[t_{0}, t_{0}+\sigma], $

$ {x_{k}(t)}(\vartheta)=\left\{ \begin{array}{ll} y_{k}(\vartheta), ~~~~\vartheta\in [t_{0}-r, t], \\ y_{k}(t), ~~~~~\vartheta\in [t, t_0+\sigma], \end{array} \right. $

$k=1, 2, \cdots$

$ {x_{0}(t)}(\vartheta)=\left\{ \begin{array}{ll} y_{0}(\vartheta), ~~~~\vartheta\in [t_{0}-r, t], \\ y_{0}(t), ~~~~~\vartheta\in [t, t_0+\sigma]. \end{array} \right. $

则对$k\in N, $由引理2.2知$x_{k}:[t_{0}, t_0+\sigma]\rightarrow O$是方程(8)的一个解.

$y\in O, $$A_{k}(t)y, ~~k\in N$的定义有

$ \begin{align*} \parallel A_{k}(t)y-A_{0}(t)y \parallel_{\infty} \leq&\sup\limits_{v\in[t_{0}, t_{0}+\sigma]}\parallel[A_{k}(t)y-A_{0}(t)y](v)\parallel\\ \leq&\sup\limits_{v\in[t_{0}, t_{0}+\sigma]}\bigg\|\int^{v}_{t_{0}}\mathcal{G}_{k}(y_{s}, s)\mathrm{d}\mu_{k}(s)-\int^{v}_{t_{0}}\mathcal{G}_{0}(z_{s}, s)\mathrm{d}\mu_{0}(s)\bigg\|\\ &+\int_{t_{k}}(v)\big\|I_{k}^{p}(y)-I_{k}^{0}(y)\big\|. \end{align*} $

对任意的$t\in[t_{0}, t_{0}+\sigma].$因此

$ \begin{align*} \sup\limits_{t\in[t_{0}, t_{0}+\sigma]}\parallel A_{k}(t)y-A_{0}(t)y \parallel_{\infty} \leq&\sup\limits_{v\in[t_{0}, t_{0}+\sigma]}\bigg\|\int^{v}_{t_{0}}\mathcal{G}_{k}(y_{s}, s)\mathrm{d}\mu_{k}(s)-\int^{v}_{t_{0}}\mathcal{G}_{0}(z_{s}, s)\mathrm{d}\mu_{0}(s)\bigg\|\\ &+\int_{t_{k}}(v)\big\|I_{k}^{p}(y)-I_{k}^{0}(y)\big\|. \end{align*} $

$ \lim\limits_{k\rightarrow\infty}\sup\limits_{t\in[t_{0}, t_{0}+\sigma]}\parallel A_{k}(t)y-A_{0}(t)y \parallel_{\infty}=0, ~~y\in O. $

类似地, 有

$ \|g_{k}(t)-g_{0}(t)\|_{\infty}\leq\sup\limits_{t\in[t_{0}, t_{0}+\sigma]}\bigg\|\int^{v}_{t_{0}}f_{k}(s)\mathrm{d}\mu_{k}(s)-\int^{v}_{t_{0}}f_{0}(s)\mathrm{d}\mu_{0}(s)\bigg\|, ~~t\in[t_{0}, t_{0}+\sigma], $

$\lim\limits_{k\rightarrow\infty}\|g_{k}-g_{0}\|_{\infty}=0$成立.

由(16)式很容易证明$\lim\limits_{k\rightarrow\infty}\|x_{k}-x_{0}\|_{\infty}=0.$由文献[6, 定理19, 定理20], 对$k\in N, $函数

$ {y_{k}}(\vartheta)=\left\{ \begin{array}{ll} x_{k}(t_{0})(\vartheta), ~~~~\vartheta\in [t_{0}-r, t_{0}], \\ x_{k}(\vartheta)(\vartheta), ~~~~~\vartheta\in [t_{0}, t_0+\sigma] \end{array} \right. $

是方程(12)的解, 且

$ \begin{align*} \parallel y_{k}(t)-y_{0}(t)\parallel&=\parallel x_{k}(t)(t)-x_{0}(t)(t)\parallel\\ &\leq\parallel x_{k}(t)-x_{0}(t)\parallel_{\infty}\\ &\leq\sup\limits_{\vartheta\in[t_{0}, t_0+\sigma]}\| x_{k}(\vartheta)-x_{0}(\vartheta)\|_{\infty}. \end{align*} $

$ \lim\limits_{k\rightarrow\infty}y_{k}(s)=y_{0}(s). $

$y_{0}$为方程(17)的解, 从而定理得证.

定理3.1主要是对无限滞后脉冲测度微分方程解对参数的连续依赖性结果的证明, 其结果是文[5-6]中相应结果的推广.当然, 我们也可借助$\Phi$-有界变差函数理论与Kurzweil方程理论建立方程(2)的$\Phi$-有界变差解对参数的连续依赖性定理.

参考文献
[1] Kurzweil J. Generalized ordinary differential equations and continuous dependence on a parameter[J]. Czech. Math., 1957, 7(82): 418–448.
[2] Schwabik Š. Generalized ordinary differential equations[M]. Singapore: World Scientific, 1992.
[3] Schwabik Š. Abstract perron-stieltjes integral[J]. Mach. Bohem., 1996, 121(4): 425–447. DOI:10.21136/MB.1996.126036
[4] Federson M, Schwabik Š. Generalized ODE approach to impulsive retarded functional differential equation[J]. Journal of Differential and Integral Equation, 2006, 19(11): 1201–1234.
[5] Halas Z, Tvrd M. Continuous dependence of solutions of generalized linear ordinary differential equations on a parameter[J]. Functional Differential Equations, 2009, 16(2): 199–213.
[6] Lgobi D K, Abasiekwere U. Results on uniqueness of solution of nonhomogeneous impulsive retarded equation using the generalized ordinary differential equation[J]. International Journal of Differential Equations, 2019, 10(2): 1–9.
[7] Monteiro G A, Tvrdy M. On kurzweil-stieltjes integral in banach[EB]. http://dml.cn/dmlcz/142992.
[8] Monteiro G A, Slavik A. Linear measure functional equations with infinite delay[J]. Math. Nachr., 2014, 287(11-12): 1363–1382. DOI:10.1002/mana.201300048
[9] Schwabik Š. Linear stieltjes integral equations in banach spaces[J]. Math. Bohem., 1999, 124(4): 433–457.
[11] Federson M, Taboas P Z. Topological dynamics of retarded functional differential equations[J]. J. Differential Equations, 2003, 195(2): 313–331. DOI:10.1016/S0022-0396(03)00061-5
[11] Hale J. Theory of functional differential equations[M]. New York: Springer-Verlag, 1977.