Nonlocal and nonlinear evolution equations in perforated domains (Q1995878)

The authors describe the behavior when \(\varepsilon \rightarrow 0\) of the solution to a nonlocal evolution equation written as \[ u_{t}(x,t)=\int_{ \mathbb{R}^{N}\backslash (\Omega \backslash \Omega ^{\varepsilon })}J(x-y)u(y,t)dy-h_{\varepsilon }(x)u(x,t)+f(x,u(x,t)), \] with \((x,t)\in \Omega ^{\varepsilon }\times (0,\infty )\), where \(\Omega ^{\varepsilon }\) is included in a fixed and bounded domain \(\Omega \subset \mathbb{R}^{N}\) and is such that its characteristic function \(\chi _{\varepsilon }\) weakly\( ^{\ast }\) converges in \(L^{\infty }(\Omega )\) to some function \(\chi \in L^{\infty }(\mathbb{R}^{N})\) which is positive in \(\Omega \). With this hypothesis, the authors cover more situations than that of a perforated domain. Dirichlet or Neumann boundary conditions are added, together with the initial data \(u(x,0)=u_{0}(x)\). In the above equation, \(J\in C(\mathbb{R} ^{N},\mathbb{R})\) is a smooth, non-negative and non-singular kernel which satisfies \(J(0)>0\), \(J(-x)=J(x)\) and \(\int_{\mathbb{R}^{N}}J(x)dx=1\). The function \(f:\Omega ^{\varepsilon }\times L^{1}(\Omega ^{\varepsilon })\rightarrow \mathbb{R}\) is defined through \(f(x,u)=(g\circ m_{\Omega ^{\varepsilon }})(x,u)\), where \(g:\mathbb{R}\rightarrow \mathbb{R}\) is a smooth and globally Lipschitz function, and \(m_{\Omega ^{\varepsilon }}:\Omega ^{\varepsilon }\times L^{1}(\Omega ^{\varepsilon })\rightarrow \mathbb{R}\) is defined as \(m_{\Omega ^{\varepsilon }}=\frac{1}{\left\vert B_{\delta }(x)\cap \Omega ^{\varepsilon }\right\vert }\int_{B_{\delta }(x)\cap \Omega ^{\varepsilon }}u(y)dy\). The first main result deals with the case where \(h_{\varepsilon }=1\) and proves the existence of a unique limit \(u^{\ast }:\mathbb{R}\times \mathbb{R}^{N}\rightarrow \mathbb{R}\) with \(u^{\ast }(x,t)\equiv 0\) in \(\mathbb{R}^{N}\backslash \Omega \) and \(u^{\ast }\in C^{1}([a,b],L^{2}(\mathbb{R}^{N}))\) for any closed interval \( [a,b]\subset \mathbb{R}^{N}\), such that \(u_{\varepsilon }\rightarrow u^{\ast }\) weakly\(^{\ast}\) in \(L^{\infty }([a,b];L^{2}(\Omega ))\), as \(\varepsilon \rightarrow 0\). Moreover, this limit \(u^{\ast }\) is the solution to the nonlocal equation \[ u_{t}(x,t)=\chi (x)\int_{\mathbb{R} ^{N}}J(x-y)(u(y,t)-u(x,t))dy+\chi (x)f_{\chi }(x,u(x,t))+(\chi (x)-1)u(x,t), \] in \(\Omega \times (0,\infty )\), with the initial condition \(u(x,0)=\chi (x)u_{0}(x)\). Here \(f_{\chi }(x,u)=(g\circ m_{\chi })(x,u)\), with \(m_{\chi }(x,u)=\frac{1}{\int_{B_{\delta }(x)}\chi (y)dy}\int_{B_{\delta }(x)}u(y)dy\). The second main result deals with the case where \[ h_{\varepsilon }(x)=\int_{\mathbb{R}^{N}\backslash (\Omega \backslash \Omega ^{\varepsilon })}J(x-y)dy \text{ for } x\in \mathbb{R}^{N}. \] The authors here prove a similar asymptotic behavior, but with a slightly different limit nonlocal equation. The limit solutions depend on \(\chi \) and on the small parameter \(\delta \). The third main result of the paper describes the asymptotic behavior of these limit solutions as \(\delta \rightarrow 0\), under further hypotheses on the data. For the proofs of these results, the authors first prove the existence of a unique solution \(u^{\varepsilon }\in C^{1}([a,b];L^{2}(\Omega ^{\varepsilon }))\) to the original problem on which they establish uniform estimates, considering the first eigenvalue of the problem. They also establish properties of the limit problem for which they prove the existence of a unique solution. The convergence results are obtained through direct computations and using the dominated convergence theorem.