Canards and bifurcation delays of spatially homogeneous and inhomogeneous types in reaction-diffusion equations (Q1038573)

The authors first give an informal description of the notions of canard-solution and of exit time by considering the van der Pol equation. They then extend and apply these concepts to certain PDE's of parabolic type. The first case to be considered is a linear PDE of the following form: \[ \begin{aligned} & \varepsilon(u_t-u_{xx})= a(x,t,\varepsilon)u,\quad u(x,0)=u_0(x)\\ & u_x(0,t)=u_x(1,t)=0.\end{aligned}\tag{1} \] Here, \(a\) is smooth and defined on a rectangle \([0,1]\times [0,t_m]\times [-\varepsilon_0,\varepsilon_0]\), some \(t_m,\varepsilon_0>0\). Moreover it is assumed that there exists a smooth curve \(t_*(x)\in (a,t_m]\), \(x\in (0,1]\) such that \[ \begin{aligned} & a(x,t,0)<0\text{ for }t<t_*(x)\\ & a(x,t,0)\text{ for }t_*(x)<t.\end{aligned}\tag{2} \] Let \(A(t)=\int^t_0\sup_xa(x,s,0)\,ds.\) As to (1), the following result holds. Let (2) be satisfied and let \(u_0(x)\) in (1) be strictly positive. Given \(\delta>0\) there is \(\varepsilon_1\in(0,\varepsilon_0]\) and constants \(L_0\), \(U_0>0\) such that for \(\varepsilon\in(0,\varepsilon_1]\) there is a solution \(u(x,t;\varepsilon)\) of (1) such that \[ L_0\exp \varepsilon^{-1}(A(t)-\delta)\leq u(x,t;\varepsilon)\leq U_0\exp\varepsilon^{-1}(A(t)+\delta). \] It follows that there is a unique time \(t_E\) called ``exit time'' such that \(A(t_\varepsilon)=0\). The authors also consider nonlinear PDE's of the form \[ \begin{aligned} & \varepsilon u_t-\varepsilon^\alpha u_{xx}=f(u,t,x,\varepsilon)u,\quad u(x,0)=u_0(x) \\ & u_x(0,t)=u_x(x_1,t)=0\end{aligned}\tag{4} \] where \(f\) is smooth and with \(a(x,t,\varepsilon)=f(0,t,x,\varepsilon)\) subject to (2), moreover one assumes \(\varepsilon,\alpha>0\). Theorems 3, 4 now address the existence of an exit time \(t_\varepsilon\). Thus theorem 3 states that if \(0<\alpha<2\), then \(t_E\) exists and is determined by the condition \[ \int^{t_\varepsilon}_0\sup_x f(a,x,t,0)\,dt=0. \] Theorem 5 finally guarantees the existence of positive solutions to (4); the proof is based on the method of sub- and supersolutions.