\documentclass[reqno]{amsart} \usepackage{hyperref} \AtBeginDocument{{\noindent\small \emph{Electronic Journal of Differential Equations}, Vol. 2014 (2014), No. 83, pp. 1--10.\newline ISSN: 1072-6691. URL: http://ejde.math.txstate.edu or http://ejde.math.unt.edu \newline ftp ejde.math.txstate.edu} \thanks{\copyright 2014 Texas State University - San Marcos.} \vspace{9mm}} \begin{document} \title[\hfilneg EJDE-2014/83\hfil Existence and stability of solutions] {Existence and stability of solutions to \\ nonlinear impulsive differential equations \\ in $\beta$-normed spaces} \author[J. Wang, Y. Zhang \hfil EJDE-2014/83\hfilneg] {Jinrong Wang, Yuruo Zhang} % in alphabetical order \address{Jinrong Wang \newline Department of Mathematics, Guizhou University, Guiyang, Guizhou 550025, China.\newline School of Mathematics and Computer Science, Guizhou Normal College, Guiyang, Guizhou 550018, China} \email{sci.jrwang@gzu.edu.cn} \address{Yuruo Zhang \newline Department of Mathematics, Guizhou University, Guiyang, Guizhou 550025, China} \email{yrzhangmath@126.com} \thanks{Submitted February 3, 2014. Published March 26, 2014.} \subjclass[2000]{34A37, 34D10} \keywords{Nonlinear impulsive differential equations; existence; stability} \begin{abstract} In this article, we consider nonlinear impulsive differential equations in $\beta$-normed spaces. We give new concepts of $\beta$-Ulam's type stability. Also we present sufficient conditions for the existence of solutions for impulsive Cauchy problems. Then we obtain generalized $\beta$-Ulam-Hyers-Rassias stability results for the impulsive problems on a compact interval. An example illustrates our main results. \end{abstract} \maketitle \numberwithin{equation}{section} \newtheorem{theorem}{Theorem}[section] \newtheorem{lemma}[theorem]{Lemma} \newtheorem{definition}[theorem]{Definition} \newtheorem{remark}[theorem]{Remark} \allowdisplaybreaks \section{Introduction} In the past decades, many researchers studied differential equations with instantaneous impulses of the type \begin{equation}\label{sy.0} \begin{gathered} x'(t)=f(t,x(t)),\quad t\in J':=J\setminus \{t_1,\dots,t_m\},\; J:=[0,T],\\ x(t^{+}_k)=x(t_k^-)+I_k(x(t^{-}_k)),\quad k=1,2,\dots,m. \end{gathered} \end{equation} where $f: J\times \mathbb{R}\to \mathbb{R}$ and $I_k: \mathbb{R}\to \mathbb{R}$ and $t_k$ satisfy $0=t_00$, $\psi\ge0$ and $\varphi\in PC(J,\mathbb{R}_+)$. We consider the following inequalities: \begin{equation}\label{U-H stable} \begin{gathered} |y'(t)-f(t,y(t))|\leq\epsilon,\quad t\in (s_i,t_{i+1}],\; i=0,1,2,\dots,m,\\ |y(t)-g_i(t,y(t))|\leq \epsilon,\quad t\in (t_i,s_{i}],\; i=1,2,\dots,m, \end{gathered} \end{equation} and \begin{equation}\label{generalized U-H-R stable} \begin{gathered} |y'(t)-f(t,y(t))|\leq\varphi(t),\quad t\in (s_i,t_{i+1}],\; i=0,1,2,\dots,m,\\ |y(t)-g_i(t,y(t))|\leq \psi,\quad t\in (t_i,s_{i}],\; i=1,2,\dots,m, \end{gathered} \end{equation} and \begin{equation}\label{U-H-R stable} \begin{gathered} |y'(t)-f(t,y(t))|\leq\epsilon\varphi(t),\quad t\in (s_i,t_{i+1}],\; i=0,1,2,\dots,m,\\ |y(t)-g_i(t,y(t))|\leq\epsilon \psi,\quad t\in (t_i,s_{i}],~i=1,2,\dots,m. \end{gathered} \end{equation} Next, our aim is to find a solution $y(\cdot)$ close to the measured output $x(\cdot)$ and whose closeness is defined in the sense of $\beta$-Ulam's type stabilities. \begin{definition}\label{def1} \rm Equation \eqref{sy.1-im-no} is $\beta$-Ulam-Hyers stable if there exists a real number $c_{f,\beta,g_i,\varphi}>0$ such that for each $\epsilon>0$ and for each solution $y\in PC^1(J,\mathbb{R})$ of \eqref{U-H stable} there exists a solution $x\in PC^1(J,\mathbb{R})$ of \eqref{sy.1-im-no} with \[ |y(t)-x(t)|^\beta \leq c_{f,\beta,g_i,\varphi}\epsilon^\beta ,\quad t\in J. \] \end{definition} \begin{definition}\label{def2} Equation \eqref{sy.1-im-no} is generalized $\beta$-Ulam-Hyers stable if there exists $\theta_{f,\beta,g_i,\varphi}\in C(\mathbb{R}_+,\mathbb{R}_+)$, $\theta_{f,\beta,g_i,\varphi}(0)=0$ such that for each solution $y\in PC^1(J,\mathbb{R})$ of \eqref{U-H stable} there exists a solution $x\in PC^1(J,\mathbb{R})$ of \eqref{sy.1-im-no} with \[ |y(t)-x(t)|^\beta \leq\theta_{f,\beta,g_i,\varphi}(\epsilon^\beta),~t\in J. \] \end{definition} \begin{definition}\label{def3} \rm Equation \eqref{sy.1-im-no} is $\beta$-Ulam-Hyers-Rassias stable with respect to $(\varphi,\psi)$ if there exists $c_{f,\beta,g_i,\varphi}>0$ such that for each $\epsilon>0$ and for each solution $y\in PC^1(J,\mathbb{R})$ of \eqref{U-H-R stable} there exists a solution $x\in PC^1(J,\mathbb{R})$ of \eqref{sy.1-im-no} with \[ |y(t)-x(t)|^\beta \leq c_{f,\beta,g_i,\varphi}\epsilon^\beta (\psi^\beta+\varphi^\beta(t)),~t\in J. \] \end{definition} \begin{definition}\label{def4} \rm Equation \eqref{sy.1-im-no} is generalized $\beta$-Ulam-Hyers-Rassias stable with respect to $(\varphi,\psi)$ if there exists $c_{f,\beta,g_i,\varphi}>0$ such that for each solution $y\in PC^1(J,\mathbb{R})$ of \eqref{generalized U-H-R stable} there exists a solution $x\in PC^1(J,\mathbb{R})$ of \eqref{sy.1-im-no} with \[ |y(t)-x(t)|^\beta \leq c_{f,\beta,g_i,\varphi}(\psi^\beta+\varphi^\beta(t)),\quad t\in J. \] \end{definition} Obviously, (i) Definition \ref{def1} implies Definition \ref{def2}; (ii) Definition \ref{def3} implies Definition \ref{def4}; (iii) Definition \ref{def3} for $\varphi (\cdot)=\psi=1$ implies Definition \ref{def1}; (iv) Definitions \ref{def1}-\ref{def4} become to Ulam's stability concepts in Wang et al.~\cite{Wang-JMAA} when $\beta=1$ and $s_i=t_i$. \begin{remark}\label{remark1} \rm A function $y\in PC^{1}(J,\mathbb{R})$ is a solution of \eqref{generalized U-H-R stable} if and only if there is $G\in PC(J,\mathbb{R})$ and a sequence $G_i$, $i=1,2,\dots,m$ (which depend on $y$) such that \begin{itemize} \item[(i)] $|G(t)|\leq \varphi(t)$, $t\in J$ and $|G_i|\le \psi$, $i=1,2,\dots,m$; \item[(ii)] $y'(t)=f(t,y(t))+G(t)$, $t\in (s_i,t_{i+1}]$, $i=0,1,2,\dots,m$; \item[(iii)] $y(t)=g_i(t,y(t))+G_i$, $t\in (t_i,s_{i}]$, $i=1,2,\dots,m$. \end{itemize} \end{remark} By Remark \ref{remark1} we get the following results. \begin{remark} \rm If $y\in PC^{1}(J,\mathbb{R})$ is a solution of \eqref{generalized U-H-R stable} then $y$ is a solution of the integral inequality \begin{equation}\label{est1} \begin{gathered} |y(t)-g_i(t,y(t))|\leq \psi,\quad t\in(t_i,s_i],\; i=1,2,\dots,m;\\ \big|y(t)-y(0)-\int_0^{t}f(s,y(s))ds\big| \leq \int_0^{t}\varphi(s)ds,\quad t\in[0,t_1];\\ \big|y(t)-g_i(s_i,y(s_i))-\int_{s_i}^{t}f(s,y(s))ds\big| \leq \psi+\int_{s_i}^{t}\varphi(s)ds,\\ t\in [s_i,t_{i+1}],\; i=1,2,\dots,m. \end{gathered} \end{equation} \end{remark} We can give similar remarks for the solutions of the inequalities \eqref{U-H stable} and \eqref{U-H-R stable}. To study Ulam's type stability, we need the following integral inequality results (see \cite[Theorem 16.4]{Bainov92}). \begin{lemma}\label{Gronwall-class} (i) Let the following inequality holds \[ u(t)\leq a(t)+\int_0^{t}b(s)u(s) ds,\quad t\geq 0, \] where $u$, $a,\in PC(\mathbb{R}_{+},\mathbb{R}_{+})$, $a$ is nondecreasing and $b(t)> 0$. Then, for $t\in \mathbb{R}_{+}$, \[ u(t)\leq a(t)\exp\Big(\int_0^{t}b(s) ds\Big). \] (ii) Assume \[ % \label{im-in-in} u(t)\leq a(t)+\int_0^{t}b(s)u(s) ds +\sum_{0 0$, $\beta_k>0$, $k\in \{1,\dots,m\}$. Then, for $t\in \mathbb{R}_{+}$, \[ %\label{inqq} u(t)\leq a(t)(1+\beta)^{k}\exp\Big(\int_0^{t}b(s) ds\Big),\quad t\in (t_k,t_{k+1}],\; k\in\{1,\dots,m\}, \] where $\beta=\sup_{k\in \{1,\dots,m\}}\{\beta_k\}$. \end{lemma} \section{Main results} We use the following assumptions: \begin{itemize} \item[(H1)] $f\in C(J\times\mathbb{R},\mathbb{R})$. \item[(H2)] There exists a positive constant $L_f$ such that \[ |f(t,u_1)-f(t,u_2)|\leq L_f|u_1-u_2|, \] for each $t\in J$ and all $u_1,u_2 \in \mathbb{R}$. \item[(H3)] $g_i\in C([t_i,s_{i}]\times\mathbb{R},\mathbb{R})$ and there are positive constants $L_{g_i}$, $i=1,2,\dots,m$ such that \[ |g_i(t,u_1)-g_i(t,u_2)|\leq L_{g_i}|u_1-u_2|, \] for each $t\in [t_i,s_{i}]$ and all $u_1,u_2 \in \mathbb{R}$. \item[(H4)]: Let $\varphi\in C(J,\mathbb{R_{+}})$ be a nondecreasing function. There exists $c_{\varphi}>0$ such that \[ \int_0^{t}\varphi(s)ds\leq c_{\varphi}\varphi(t), \] for each $t\in J$. \end{itemize} Concerning the existence and uniqueness result for the solutions to \eqref{sy.1-im-Cauchy}, we give the following theorem. \begin{theorem}\label{theorem-existence} Assume that {\rm (H1)--(H3)} are satisfied. Then \eqref{sy.1-im-Cauchy} has a unique solution $x$ provided that \begin{equation}\label{contract-} \varrho:=\max\{L_{g_i}^\beta+L_f^\beta(t_{i+1}-s_i)^\beta,L_f^\beta t_1^\beta: i=1,2,\dots,m\}<1. \end{equation} \end{theorem} \begin{proof} Consider a mapping $F: PC(J,\mathbb{R})\to PC(J,\mathbb{R})$ defined by \begin{gather*} (Fx)(0)= x_0;\\ (Fx)(t)= g_i(t,x(t)),\quad t\in(t_i,s_i],~i=1,2,\dots,m;\\ (Fx)(t)= x_0+\int_0^{t}f(s,x(s))ds,\quad t\in [0,t_1];\\ (Fx)(t)= g_i(s_i,x(s_i))+\int_{s_i}^{t}f(s,x(s))ds,\quad t\in (s_i,t_{i+1}],\; i=1,2,\dots,m. \end{gather*} Obviously, $F$ is well defined. For any $x,y\in PC(J,\mathbb{R})$ and $t\in (s_i,t_{i+1}]$, $i=1,2,\dots,m$, we have \begin{align*} |(Fx)(t)-(Fy)(t)| &\leq L_{g_i}|x(s_i)-y(s_i)|+L_f\int_{s_i}^{t}|x(s)-y(s)|ds\\ &\leq L_{g_i}\|x-y\|_{C}+L_f\int_{s_i}^{t}\max_{t\in [s_i,t_{i+1}]}|x(s)-y(s)|ds\\ &\leq L_{g_i}\|x-y\|_{C}+L_f(t_{i+1}-s_i)\|x-y\|_{PC}, \end{align*} which implies \[ |(Fx)(t)-(Fy)(t)|^\beta \leq L^\beta_{g_i}\|x-y\|_{P\beta}+L^\beta_f(t_{i+1}-s_i)^\beta\|x-y\|_{P\beta}. \] This reduces to \[ \|Fx-Fy\|_{P\beta} \leq \big(L^\beta_{g_i}+L^\beta_f(t_{i+1}-s_i)^\beta\big)\|x-y\|_{P\beta}, \quad t\in (s_i,t_{i+1}]. \] Proceeding as above, we obtain that \begin{gather*} \|Fx-Fy\|_{P\beta} \leq L^\beta_ft_1^{\beta}\|x-y\|_{P\beta},\quad t\in [0,t_1],\\ \|Fx-Fy\|_{P\beta}\leq L^\beta_{g_i}\|x-y\|_{P\beta},\quad t\in(t_i,s_i],\; i=1,2,\dots,m. \end{gather*} From the above facts, we have \[ \|Fx-Fy\|_{P\beta}\leq \varrho\|x-y\|_{P\beta}, \] where $\varrho$ is defined in \eqref{contract-}. Finally, we can deduce that $F$ is a contraction mapping. Then, one can derive the result immediately. \end{proof} Next, we discuss hte stability of \eqref{sy.1-im-no} by using the concept of generalized $\beta$-Ulam-Hyers-Rassias in the above section. \begin{theorem}\label{theorem-U-H-R} Assume that {\rm (H1)-(H4)} and \eqref{contract-} are satisfied. Then \eqref{sy.1-im-no} is generalized $\beta$-Ulam-Hyers-Rassias stable with respect to $(\varphi,\psi)$. \end{theorem} \begin{proof} Let $y\in PC^1(J,\mathbb{R})$ be a solution of \eqref{generalized U-H-R stable}. Denote by $x$ the unique solution of the impulsive Cauchy problem \begin{equation}\label{sy.1-ref} \begin{gathered} x'(t)=f(t,x(t)),\quad t\in (s_i,t_{i+1}],\; i=0,1,2,\dots,m,\\ x(t)=g_i(t,x(t)),\quad t\in (t_i,s_{i}],\; i=1,2,\dots,m,\\ x(0)=y(0). \end{gathered} \end{equation} Then we obtain \[ x(t)=\begin{cases} g_i(t,x(t)), & t\in(t_i,s_i],\; i=1,2,\dots,m;\\ y(0)+\int_0^{t}f(s,x(s))ds, &t\in [0,t_1];\\ g_i(s_i,x(s_i))+\int_{s_i}^{t}f(s,x(s))ds, & t\in (s_i,t_{i+1}],\; i=1,2,\dots,m. \end{cases} \] Keeping in mind \eqref{est1}, for each $t\in (s_i,t_{i+1}]$, $i=1,2,\dots,m$, we have \[ \big|y(t)-g_i(s_i,y(s_i))-\int_{s_i}^{t}f(s,y(s))ds\big| \leq \psi+\int_{s_i}^{t}\varphi(s)ds \leq \psi+ c_{\varphi}\varphi(t), \] and for $t\in(t_i,s_i]$, $i=1,2,\dots,m$, we have \[ |y(t)-g_i(t,y(t))|\leq \psi, \] and for $t\in [0,t_1]$, we have \[ \big|y(t)-y(0)-\int_0^{t}f(s,y(s))ds\big| \leq c_{\varphi}\varphi(t). \] Hence, for each $t\in (s_i,t_{i+1}]$, $i=1,2,\dots,m$, we have \begin{align*} &|y(t)-x(t)|\\ &= \big|y(t)-g_i(s_i,x(s_i))-\int_{s_i}^{t}f(s,x(s))ds\big|\\ &\leq \big|y(t)-g_i(s_i,y(s_i))-\int_{s_i}^{t}f(s,y(s))ds\Big|\\ &\quad +\big|g_i(s_i,y(s_i))-g_i(s_i,x(s_i))\Big| +\Big(\int_{s_i}^{t}|f(s,y(s))-f(s,x(s))|ds\Big)\\ &\leq (1+c_{\varphi})[\psi+\varphi(t)]+L_{g_i}|y(s_i)-x(s_i)| +\int_{s_i}^{t}L_f|y(s)-x(s)|ds\\ &\leq (1+c_{\varphi})[\psi+\varphi(t)]+\sum_{0< s_i< t}L_{g_i}|y(s_i)-x(s_i)|+\int_0^{t}L_f|y(s)-x(s)|ds. \end{align*} Clearly, $ a(t):=(1+c_{\varphi})[\psi+\varphi(t)]$, $t\in (s_i,t_{i+1}], $ is nondecreasing and $a \in PC(\mathbb{R}_{+},\mathbb{R}_{+})$. By Lemma \ref{Gronwall-class} (ii), we obtain \begin{align*} |y(t)-x(t)| &\leq (1+c_{\varphi})[\psi+\varphi(t)](1+L_g)^i\exp\Big(\int_0^{t}L_f ds\Big)\\ &\leq (1+c_{\varphi})[\psi+\varphi(t)](1+L_g)^i\exp\big(L_ft_{i+1}\big) \end{align*} where $L_g=\max\{L_{g_1},L_{g_2},\dots,L_{g_m}\}$. Thus, \begin{equation}\label{UHR-E1} \begin{aligned} |y(t)-x(t)|^\beta &\leq \big[(1+c_{\varphi})[\psi+\varphi(t)](1+L_g)^i\exp\big(L_ft_{i+1}\big) \big]^\beta\\ &\leq \big[(1+c_{\varphi})(1+L_g)^i\exp\big(L_ft_{i+1}\big)\big]^\beta [\psi+\varphi(t)]^\beta\\ &\leq \big[(1+c_{\varphi})(1+L_g)^i\exp\big(L_ft_{i+1}\big)\big]^\beta (\psi^\beta+\varphi(t)^\beta), \end{aligned} \end{equation} for $t\in (s_i,t_{i+1}]$, $i=1,2,\dots,m$. Further, for $t\in(t_i,s_i]$, $i=1,2,\dots,m$, we have \begin{align*} |y(t)-x(t)|^\beta &\leq |y(t)-g_i(t,x(t))|^\beta\\ &\leq |y(t)-g_i(t,y(t))|^\beta+ |g_i(t,y(t))-g_i(t,x(t))|^\beta \\ &\leq \psi^\beta +L_{g_i}^\beta |y(t)-x(t)|^\beta, \end{align*} which yields \begin{equation}\label{UHR-E2} |y(t)-x(t)|^\beta \leq \frac{1}{1-L_{g_i}^\beta} \psi^\beta.\quad \text{(\eqref{contract-} implies $L_{g_i}^\beta<1)$} \end{equation} Moreover, for $t\in [0,t_1]$, we have \begin{align*} |y(t)-x(t)| &= \big|y(t)-y(0)-\int_0^tf(s,x(s))ds\big|\\ &\leq \big|y(t)-y(0)-\int_0^tf(s,y(s))ds\big| +\Big(\int_0^t|f(s,y(s))-f(s,x(s))|ds\Big)\\ &\leq c_{\varphi}\varphi(t)+ \int_0^{t}L_f|y(s)-x(s)|ds. \end{align*} By Lemma \ref{Gronwall-class} (i), we obtain \begin{align*} |y(t)-x(t)| &\leq c_{\varphi}\varphi(t)\exp\Big(\int_0^{t}L_f ds\Big)\\ &\leq c_{\varphi}\varphi(t)\exp\big(L_ft_1\big). \end{align*} Thus, we obtain \begin{equation}\label{UHR-E3} \begin{aligned} |y(t)-x(t)|^\beta &\leq \big[c_{\varphi}\varphi(t)\exp\big(L_ft_1\big)\big]^\beta\\ &\leq \big[c_{\varphi}\exp\big(L_ft_1\big)\big]^\beta\varphi(t)^\beta, \quad t\in [0,t_1]. \end{aligned} \end{equation} Summarizing, we combine \eqref{UHR-E1}, \eqref{UHR-E2} and \eqref{UHR-E3} and derive that \begin{align*} |y(t)-x(t)|^\beta &\leq \Big(\big[(1+c_{\varphi})(1+L_g)^i\exp\big(L_ft_{i+1}\big)\big]^\beta\\ &\quad +\frac{1}{1-L_{g_i}^\beta} +\big[c_{\varphi}\exp\big(L_ft_1\big)\big]^\beta\Big)(\psi^\beta+\varphi^\beta(t))\\ &:= c_{f,\beta,g_i,\varphi}(\psi^\beta+\varphi^\beta(t)),\quad t\in J, \end{align*} which implies that \eqref{sy.1-im-no} is generalized $\beta$-Ulam-Hyers-Rassias stable with respect to $(\varphi,\psi)$. The proof is complete. \end{proof} \section{An example} Consider the nonlinear differential equation, without instantaneous impulses, \begin{equation}\label{E4.1} \begin{gathered} x'(t)=\frac{1}{1+15e^{t}}\arctan (t^2+x(t)), \quad t\in (0,1],\\ x(t)=\frac{1}{15+t^2}\ln( x(t)+1),\quad t\in (1,2], \end{gathered} \end{equation} and inequalities \begin{equation}\label{E4.2} \begin{gathered} \big|y'(t)-\frac{1}{1+15e^{t}}\arctan (t^2+y(t))\big| \leq e^t, \quad t\in (0,1],\\ \big|y(t)-\frac{1}{15+t^2}\ln(y(t)+1)\big|\leq 1,\quad t\in (1,2]. \end{gathered} \end{equation} Set $J=[0,2]$, $0=s_0