Time-periodic boundary layer solutions to singularly perturbed parabolic problems (Q509751)

This paper deals with semilinear parabolic problems of the type \[ \mu\partial_tu(t,x)=\nu^2\partial_x^2 u(t,x)+f(t,x,u(t,x),\mu,\nu),\quad (t,x)\in \mathbb R\times (0,1) \] with homogeneous boundary conditions \(u(t,0)=u(t,1)=0\) and periodicity condition in time \(u(t+1,x)=u(t,x), \) where \(\mu,\nu\) are two independent small parameters, and \(f:\mathbb R\times [0,1]\times \mathbb R\times [0,1]^2 \to\mathbb R\) is \(C^3\)-smooth and 1-periodic with respect to \(t\). Moreover \(f\) is supposed to satisfy the following conditions:\newline 1) there exists a continuous 1-periodic (in \(t\)) function \(u^0:\mathbb R\times [0,1]\to \mathbb R\) such that \[ f(t,x,u^0(t,x),0,0)=0,\quad \partial_u f(t,x,u^0(t,x),0,0)<0, \] 2) there exist smooth maps \(v^0,w^0:\mathbb R\times [0,\infty)\to\mathbb R,\) 1-periodic in \(t,\) such that \[ \partial^2_yv^0(t,y)+f(t,0,u^0(t,0)+v^0(t,y),0,0)=0, \] \[ (v^0+u^0)(t,0)=v^0(t,\infty)=0, \] and \[ \partial^2_yw^0(t,y)+f(t,1,u^0(t,1)+w^0(t,y),0,0)=0, \] \[ w^0(t,0)+u^0(t,1)=w^0(t,\infty)=0, \] Moreover \(\partial_y v^0(t,0)\not= 0 \text{ and } \partial_y w^0(t,0)\not=0\). The authors are interested in the asymptotic behavior of solutions \(\hat{u}_{\mu,\nu}\) such that \[ \underset{(\mu,\nu)\to (0,0)}\lim \hat{u}_{\mu,\nu}(t,x) =u^0(t,x)\quad (t,x)\in\mathbb R\times (0,1). \] They introduce the approximate solution \[ u_{\nu}(t,x)=u^0(t,x)+v^0\left(t,\frac{x}{\nu}\right)+w^0\left(t,\frac{1-x}{\nu}\right). \] They prove the following. There exist \(\epsilon>0\) and \(c>0\) such that, for all \(\mu,\nu\in (0,\epsilon),\) there exists a solution \(\hat{u}_{\mu,\nu}\) with \[ ||\hat{u}_{\mu,\nu}-u_{\nu}||_{\infty}\leq c(\mu+\nu). \] Moreover, there exists \(\delta>0\) such that, for all \(\mu,\nu\in(0,\epsilon),\) the following is true: if \(u\) is a solution with \(||u-\hat{u}_{\mu,\nu}||_{\mu,\nu}\leq \delta,\) then \(u=\hat{u}_{\mu,\nu},\) where \(||\overset{.}{~}||_{\mu,\nu} \) is a suitable norm. The proof is based on an adapted implicit function theorem.